Studies of the Reactivity of Permethylzirconocene Complexes with Binuclear Transition Metal Carbonyl Compounds Citation Berry, Donald Harry (1984) Studies of the Reactivity of Permethylzirconocene Complexes with Binuclear Transition Metal Carbonyl Compounds. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/yahb-n031. https://resolver.caltech.edu/CaltechTHESIS:10242018-111901394 Abstract The low valent permethylzirconocene complex {Cp* 2 ZrN 2 } 2 N 2 (Cp*≡η 5 -C 5 Me 5 ) reacts with binuclear transition metal carbonyl compounds of the general formula Cp 2 M 2 (CO) x (Cp≡η 5 -C 5 H 5 ) to produce trinuclear complexes. Two major modes of reactivity are observed. Compounds with M-M bonds that are easily reduced such as (CpM(CO) 3 ) 2 (M = W, Mo) react with {Cp* 2 ZrN 2 } 2 N 2 to produce trinuclear species in which two CpM(CO) 3 anions are connected to a Cp* 2 Zr dication via µ 2 -(C-η 1 ; O-η 1 ) carbonyl bridges. The structure of the molybdenum complex has been determined by single crystal X-ray diffraction methods. These trinuclear complexes are also produced in the reaction of the Cp* 2 ZrRR'(R,R' = H,alkyl) with two equivalents of the hydride complexes CpM(CO) 3 H· The dinuclear complexes Cp* 2 M(H)(CpMo(CO) 3 )(M = Zr, L = PMe 3 ; M = Hf, L = CO) have been similarly prepared by treatment of the zirconium or hafnium dihydride with the corresponding molybdenum hydride. The trinuclear complexes react with HX (X = Cl, OH) to produce CpM(CO) 3 H and Cp* 2 ZrX 2 . The other mode of reactivity between {Cp* 2 ZrN 2 } 2 N 2 and metal dimers is observed with (CpM(CO) 2 ) 2 (M = Fe, Ru). In these cases trinuclear complexes are produced in which two carbonyl ligands are reductively coupled yielding an O 2 C 2 moiety bridging the three metals. This coupling is reversible and the Zr(II) intermediate can be trapped with CO or other metal dimers. The iron analog is also produced in the reaction of Cp* 2 ZrI 2 with Na(CpFe(CO) 2 ). The reaction of {Cp* 2 ZrN 2 } 2 N 2 with the bridging methylene complex cis-Cp 2 Ru 2 (CO) 3 (CH 2 ) does not result in carbonyl coupling, yielding instead what appears on the basis of spectroscopic data to be a π-complex between permethylzirconocene and cis-Cp 2 Ru 2 (CO) 3 (CH 2 ) held together only by µ 2 -(C-η 1 ; C,O-η 2 ) bridging carbonyl ligands. The ruthenium carbonyl coupled product, Cp* 2 Zr(O 2 C 2 )Ru 2 Cp 2 (CO) 2 , reacts under H 2 to produce the bridging zirconoxy carbene complex trans-Cp 2 Ru 2 (CO) 3 (C(H)OZr(H)Cp* 2 ). This complex is fluxional on the 1 H and 13 C NMR timescales. ΔG ‡ for this process is 14.0(2) kcal·mol -1 at -4°C. The fluxionality is best explained by a process involving a terminal carbine intermediate. The carbene formation is reversible, and treatment with D 2 results in label incorporation in both the hydride and carbene hydrogen positions. Reaction with CO produces Cp* 2 Zr(CO) 2 , (CpRu(CO) 2 ) 2 and presumably H 2 . Treatment with BF 3 or HBF 4 results in C-O scission and produces the bridging methylene complex Cp 2 Ru 2 (CO) 3 (CH 2 ). Labeling studies indicate both methylene hydrogens originate on zirconium. This bridging methylene complex can be synthesized in 60% yield by treatment of (CpRu(CO) 2 ) 2 with LiBEt 3 H in toluene. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Collins, Terrence J. Thesis Committee: Bercaw, John E. (chair) Grubbs, Robert H. Beauchamp, Jesse L. Collins, Terrence J. Defense Date: 31 January 1984 Record Number: CaltechTHESIS:10242018-111901394 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:10242018-111901394 DOI: 10.7907/yahb-n031 ORCID: Author ORCID Berry, Donald Harry 0000-0001-6732-3726 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 11245 Collection: CaltechTHESIS Deposited By: INVALID USER Deposited On: 24 Oct 2018 22:50 Last Modified: 29 Oct 2025 23:35 Thesis Files Preview PDF - Final Version See Usage Policy. 30MB Repository Staff Only: item control page

Studies of the Reactivity of Permethylzirconocene Complexes with Binuclear Transition Metal Carbonyl Compounds Thesis by Donald H. Berry In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1984 (Submitted January 31, 1984) ii To My Pa!r.e.n.t-6 iii ACKNOWLEDGMENTS My stay at Caltech has been immeasurably enriched, both intellectually and personally, by the people I've come to call ~y friends. The Bercaw group members, past and present, have provided a rowdy, enjoyable atmosphere over the years for which I am grateful. From my introduction to vacuum line technique at the hands of Pete, Rich and Jimmy, to John's attempts to weed out the group in the Sierras, there are countless experiences I would not have had elsewhere. Each group member (including the one that use to hang from the ceiling) has added his or her own particular touch to the office spirit. I am especially grateful to Dean, Nancy, Allan, Francis, Maureen and Martha for their friendship. Of course the 3 R's certainly have made things exciting in lab. I feel very lucky to have spent my graduate career with John Bercaw, whose enthusiasm and remarkable insight into science have never ceased to amaze me. Finally, I thank Sum a for making life so pleasant and for helping me keep my perspective on those things which are truly important. iv ABSTRACT The low valent permethylzirconocene complex {Cp*2ZrN2} 2N2 (Cp*:n5-C5Me5) reacts with binuclear transition metal carbonyl compounds of the general formula Cp2M2(CO)x (Cp:n5-C5H5) to produce trinuclear complexes. Two major modes of reactivity are observed. Compounds with M-M bonds that are easily reduced such as ( CpM(C0)3 )2 (M = W, Mo) react with {Cp*2ZrN2 }2N2 to produce trinuclear species in which two CpM(C0)3 anions are connected to a Cp*2Zr dication via ll2-(C- nl; 0- nl) carbonyl bridges. The structure of the molybdenum complex has been determined by single crystal X-ray diffraction methods. These trinuclear complexes are also produced in the reaction of the Cp*2ZrRR' (R,R' = H, alkyl) with two equivalents of the hydride complexes CpM(C0)3H· The dinuclear complexes Cp*2M(H) ( CpMo(C0)3) (M = Zr, L = PMe3; M = Hf, L =CO) have been similarly prepared by treatment of the zirconium or hafnium dihydride with the corresponding molybdenum hydride. The trinuclear complexes react with HX (X = Cl, OH) to produce CpM(C0)3H and Cp*2ZrX2. The other mode of reactivity between {Cp*2ZrN2 }2N2 and metal dimers is observed with ( CpM(C0)2 )2 (M = Fe, Ru). In these cases trinuclear complexes are produced in which two carbonyl ligands are reductively coupled yielding an 02C2 moiety bridging the three metals. This coupling is reversible and the Zr(II) intermediate can be trapped with CO or other metal dimers. The iron analog is also produced in the reaction of Cp*2Zrl2 with Na ( CpFe(C0)2 ). v The reaction of { Cp*2ZrN2} 2N2 with the bridging methylene complex cis-Cp2RU2(CO)J(CH2) does not result in carbonyl coupling, yielding instead what appears on the basis of spectroscopic data to bean-complex between permethylzirconocene and cis-Cp2Ru2(CO)J(CH2) held together only by ll2(C-n 1; C,O-n2) bridging carbonyl ligands. The ruthenium carbonyl coupled product, Cp*2Zr(02C2)Ru2CP2(C0)2, reacts under H2 to produce the bridging zirconoxy carbene complex transCp2Ru2(CO)J( C(H)OZr(H)Cp*2). This complex is fluxional on the 1H and 13c NMR timescales. ~G~ for this process is 14.0(2) kcal·mol-1 at -4oc. The fluxionality is best explained by a process involving a terminal carbene intermediate. The carbene formation is reversible, and treatment with D2 _results in label incorporation in both the hydride and carbene hydrogen positions. Reaction with CO produces Cp*2Zr(C0)2, ( CpRu(C0)2 )2 and presumably H2. Treatment with BF3 or HBF4 results in C-0 scission and produces the bridging methylene complex Cp2Ru2(CO)J(CH2). Labeling studies indicate both methylene hydrogens originate on zirconium. This bridging methylene complex can be synthesized in 60% yield by treatment of ( CpRu(C0)2 )2 with LiBEt3H in toluene. vi TABLE OF CONTENTS CHAPTER I Complexes of Permethylzirconocene With Group VI Cyclopentadieny1 Anions. Reduction of a MetalMetal Bond and Formation of 112-(C- n1; 0-n 1) Carbonyl Bridges. • • • • • • • • • • • • · • · • • · • • • • • • • • • • • • • • • • • • • • 1 CHAPTERD Reductive Coupling of Carbonyl Ligands in Binuclear Transition Metal Complexes. Reversible C-C Bond Formation in Reactions of Permethylzirconocene. • • · · • • • • • • 32 CHAPTERm Reversible Reduction of Coordinated Carbon Monoxide by Permethylzirconocene Hydride Complexes. Characterization and Reactivity of a Bridging Diruthenium Zirconoxy Car bene Complex. • • • • • • • • • • • • • • • · • 74 -1- CHAPTER I Complexes of Permethylzirconocene With Group VI Cyclopentadienyl Anions. Reduction of a Metal-Metal Bond and Formation of ~ 2-( C- n1; 0- n 1) Car bony 1 Bridges -2- Introduction Despite active efforts by several research groups, unequivocal examples of complexes with early/late transiton metal M-M bonds remain relatively uncommon 1,2 compared with the occurrence of M-M bonds between late metal pairs. The comparison is even more stark when only Group IV /late metal pairs are considered.2 One such complex, Cp*2Zr(C0)2CoCp ( Cp*:n5C5(CHJ)5), in which the Zr-Co bond is supported by two carbonyl bridges, was characterized in these laboratories.2a Thus, we were particularly interested in preliminary reports that the complexes Cp2Zr(CH3)( CpMo(CO)J )3a and Cp2Zr( C(O)CHJ) ( CpMo(CO)J )2b contained analogous structures. These compounds were prepared with the elimination of methane from Cp2ZrMe2 or Cp2ZdCH3)( C(O)CHJ) and CpMo(CO)JH. We have found that a related complex, Cp*2Zr ( CpMo(CO)J) 2' can be prepared by elmination of H2 and alkane from Cp*2Zr(H)(R) (R =H, alkyl) and two equivalents of CpMo(CO)JH. Subsequent structural characterization of Cp2Zr( C(O)CHJ) ( CpMo(CO)J) has shown that there is no Zr-Mo bond, and that the metals are connected by a ll2-(C- nl; 0- nl) carbonyl or "isocarbonyl" bridge. 4 This type of -3- interaction is prevalent in complexes between early transition metals and carbonyl compounds.3 The structure of the dimolybdenum carbonylate species reported in this chapter also exhibits this bonding mode. It was the trinuclear aspect of Cp*2Zr( CpMo(C0)3 )2, however, that presented us with additional questions. In an effort to understand how permethylzirconocene interacts with complexes containing metal-metal bonds we undertook the studies described in this and subsequent chapters. Results {Cp*2ZrN2} 2N2, 1, reacts rapidly and cleanly with two equivalents of ( CpM(C0)3 )2, 2a,b (a: M = Mo, b: M =W) in toluene at 25°C to yield {Cp* ZrN } N 2 2 2 2 1 + 2 (CpM(C0) ) 3 2 2a,b Cp* Zr(CpMo(C0) ) 2 3 2 Ja,b + 3 N 2 (1) red-brown crystals of Cp*2Zr( CpM(C0)3 )2, 3a,b. The 1H NMR spectrum of 3a exhibits singlets in the Cp and Cp* regions, while the 13c NMR spectrum shows two signals for carbonyls at 231 and 242 ppm in a 2:1 ratio. Five bands are observed for C-O stretches in the infrared spectrum: 2025, 1945, 1860, 1578 and 1533 cm-1. The latter two, extremely low energy bands are due to the 1-1 2-< C- n1; 0- n1) bridging car bony Is. -4- Compound 3a is unreactive towards both CO and H2 at 4 atm and sooc. Reaction with HX (X = Cl and OH) in benzene at 250C rapidly yields Cp*2ZrX2 and two equivalents of CpMo(C0)3H, 6. Over days at 80°C, reaction with CH3l in benzene leads to primarily Cp*2Zrl2 and CpMo(C0)3Me. In addition to direct metal-metal bond reduction, several other routes lead to 3. At sooc in benzene, Cp*2Zr(C0)2 reacts + 80°C benzene Cp* Zr(CpMo(C0) ) 2 3 2 3a ( 2) with ( CpMo(C0)2 )2 (M:::M), 4, to quantitatively yield 3a. Reaction of 1 with 4 gives purple black solutions of a compound with the probable composition Cp*2Zr (CpMo(C0)2 )2, .5, based on the single Cp and Cp* resonances in a 1:3 ratio observed in the 1H NMR spectrum. Compound 5 decomposes upon standing in solution to black insoluble material, and has thus far eluded all attempts at isolation. Compound 5 generated in situ reacts with CO to yield primarily 3a. Dinuclear elimination of isobutane and hydrogen from Cp*2Zr(H)(CH2CHMe2) and two equivalents of 6 also produces 3a. In the reaction of the labeled compound Cp*2Zr(D)(CH2CDMe2) with 6 (eq. 3), >9596 of the hydrocarbon formed is mono-deuterated as determined by 1H NMR and GC- -5- + 2 CpMo (CO) H 3 6 Cp* Zr(CpMo(C0) ) 2 3 2 3a + HD + (CH ) CD 3 3 ( 3) MS analysis. This rules out intramolecular reductive elimination followed by reaction of 6 with permethylzirconocene. Reaction of one equivalent of 6 with Cp*2HfH2 at -78°C yields Cp*2Hf(H)( CpMo(C0)3), 7. The 1H NMR spectrum of 7 exhibits single Cp and Cp* resonances in addition to the hydride signal at 12.28 ppm. In the infrared spectrum the terminal CO stretches are found at 1938 and 1850 cm-1, with the bridging CO stretch appearing at 1561 cm-1. In addition, the hafnium hydride stretch is found at 1669 cm-1. Stirring 7 in toluene with 1 atm of D2 at 25°C for 24 h results in deuteration at the hydride position with concurrent loss of the IR and 1H NMR signals arising from this proton. 2H NMR spectra show a single enriched resonance at 12.26 ppm. An analogous mono-molybdenum carbonylate complex of zirconium, Cp*2Zr(H)( CpMo(C0)2(PMe3)), 8, is formed by the slow addition of CpMo(C0)2(PMe3)H to Cp*2ZrH2 at -78°C. In this case competing formation of the dicarbonylate species is a greater problem than in the synthesis of 7. The lower selectivity found for Cp*2ZrH2 versus the hafnium dihydride is in keeping with observations that zirconium hydrides are kinetically more reactive than hafnium hydrides.5 The spectral properties Table 1. NMR~ and IR~ Data. Compound Cp*2Zr ( CpMo(C0)3 )2 3a IR v(CO): 2025(w), 1945(vs), Assignment 1H NMR 1860(vs), 1578(vs), 1533(vs) Chemical Shift C5(C!!3)5 1. 75 (s, 30H) C5!:!5 5.29 (s, 10H) 13c { 1H }NMR c 5(cH 3)5 11 C5H5 90 Mo-CO 231 Mo-COZr 242 C5(CH3)5 1 • 83 (s, 30H) C5H5 5 • 22 (s, l.OH) C5(C!!3)5 1 • 88 (s, 30H) C5!:!5 5.23 (s, 5H) Hf-H 12.28 (s, 1H) I Cp*2Zr ( CpW(C0)3 )2 3b Cp*2Hf(H) ( CpMo(C0)3) 7 v(CO): 2022(w), 1940(vs), 1H NMR 1852(vs), 1588(vs), 1542(vs) v(CO): 1938(s), 1849(s), 1542(s) 'V(Hf-H): 1668(m) 1H NMR ~ I Table I. Continued. Compound Assignment IR 1H NMR Cp*2Zr(H)( CpMo(C0)2(PMeJ)) 8 Chemical Shift P(C!!J)J 1.28 (d, 9H, Jp_H =7.5 Hz) C5(C!!3)5 1.95 (s, 30H) C5H5 5.22 (s, 5H) Zr-H 6.45 (s, 1H) v(CO): 1856(s), 1477(s) I I """"" ~MR spectra in benzene-d6 at ambient temperature. Chemical shifts reported in o relative to internal TMS. ~R spectra obtained as nufol mulls. Values reported in cm-1. -8- Table 2. Cp*2Zr( CpMo(C0)3)) 2 Formula weight = 851.72 monoclinic space group C2/c 0 MoKa 0.71069 A 1.0 < 28 < 55.00 - - 0 a = 20.009(9) A 0 b = 12.14 5( 4) A 0 c = 14.595(5) A 8 = 105.18(3)0 0 Cell volume = 3423(2) A3 Z=4 P calc = 1.65 g/cc l..l = 25.3 cm-1 R =0.095 (full data set, N0 =3523 reflections) R = 0.053 ( F~ ~ 3cr(F~) only, N0 = 2113 reflections) Goodness of fit* = 1.34 Np = 298 Cl6 Zr Mo I \0 I c(;Q CIO Figure 1. Molecular Configuratioln of Cp*2Zr [ CpMo(C0)3] 2, Ja. 02 1 I .._. 0 I Figure 2. Skeletal view of Cp*2Zr [CpMo(C0)3] 2' 3a. Mo Figure 3. Views of CpMo(C0)2(ll-CO)Zr units in Cp*2Zr[ CpMo(C0)3 J2, 3a. -120 Table 3. Bond distances for Cp*2Zr ( CpMo(C0)3 )2, 3a (A). ATOM ATOM DISTANCE ATOM ATOM DISTANCE Zrl Zrl Zrl Zrl Zrl Zrl Mol Mol Mol Mol Mol Mol Mol Mol Cl C2 01 Cl C2 C3 C4 C5 Cll Cl2 Cl3 Cl4 Cl5 Cl6 Cl7 Cl8 C6 Cl 2.078 (5) 2.532 (7) 2.485 (7) 2.479 (7) 2.528 (7) 2.534 (7) 1.868 (7) 1. 927 (8) 1. 958 (8) 2.346 (14) 2.354 (12) 2.380 (13) 2.359 (12) 2.354 (11) 1.507 (10) 1.408 (9) C3 C3 C4 C5 C5 C7 C8 Cll C9 Cl4 Cl4 Cl5 Cl6 Cl7 02 03 C2 C4 C5 Cl ClO C2 C3 01 C4 C15 Cl6 Cl8 Cl7 Cl8 Cl2 Cl8 1.394 (10) 1.406 (9) 1.402 (9) 1.421 (9) 1.501 (10) 1.500 (11) 1. 501 ( 10) 1.217 (9) 1.503 (10) 1.359 (18) 1.379 (19) 1.383 (16) 1.384 (17) 1.358 (16) 1.167 (10) 1.133 (10) Table 4. Bond Angles for Cp*2Zr ( CpMo(C0)3 )2, 3a (deg). ATOM ATOM ATOM ANGLE ATOM ATOM ATOM ANGLE 01 Cl2 Cl3 Cl3 Cll C5 C6 C6 Cl Cl C7 C2 C2 Zrl Mol Mol Mol 01 Cl Cl Cl C2 C2 C2 C3 C3 01' Cll Cll Cl2 Zrl C2 C2 C5 C3 C7 C3 C4 C8 90.7 (2) 86.9 (3) 89.9 (3) 91.3 (3) 170.7 (5) 107.1 (6) 125.3 ( 6) 126.3 (6) 109.0 (6) 125.2 ( 6) 125.7 (6) 107.7 (6) 126.4 (6) C4 C3 C5 C5 Cl Cl ClO 01 Cl6 Cl8 Cl7 Cl8 Cl7 C3 C4 C4 C4 C5 C5 C5 Cll Cl4 Cl5 Cl6 Cl7 Cl8 C8 C9 C3 C9 C4 ClO C4 Mol Cl5 Cl4 Cl4 Cl6 Cl5 125.8 (6) 127.1 (6) 108.5 ( 6.) 124.3 (6) 107.7 (6) 126.7 (6) 124.7 (6) 177.8 (6) 109.4 (12) 106.8 ( 11) 106.8 ( 11) 107.9 ( 11) 109.1 (10) -13- of 8 are similar to those of 7. NMR and IR data are presented in Table 1. The structure of 3a was confirmed by single crystal X-ray diffraction ana1ysis.6 Crystal data and details of the structure determination are given in Table 2. The structure of 3a is shown in Figures 1 and 2. The molybdenum units are very regular "three-legged piano stools". as shown in Figure 3. The molybdenum and zirconium atoms are connected via ll2-(C- n1; 0- n1) carbonyl, or "isocarbonyl" linkages. The ( CpMo(C0)3) units and Cp* rings are related by a two-fold crystallographic axis containing the zirconium atom, with one molybdenum slightly above the equatorial plane of the Cp*2Zr unit, and one slightly below. At 170.7(5)0, the Zr-0-C angles are nearly linear, as are the Mo-C-0 angles. Complete lists of bond angles and distances are found in Tables 3 and 4. Discussion Of the great number of organometallc compounds containing more than a single transition metal, a very large proportion contain coordinated carbon monoxide. In a large number of these complexes carbon monoxide bridges two or more metals in one of several modes.7 While the most common binuclear bridging mode, ll2-n2, has been recognized since the structure determination of Fe2(C0)9 in 1939,8 it has been only within the last decade that other possibilities have been observed. This is no doubt partially due to the advances in crystallography as well as the tremendous number of new compounds synthesized since the explosion of interest in organo-transition metal chemistry. -14- The four primary binuclear bridging modes for C<? are shown in Figure 4. The first three modes, ll2-n2, semi-bridging and ll2-(c-n1; c,o-n2) are always found bridging two metals that are directly bonded. The ll2-(o-n1; C-n1) bonding mode, however, always bridges two metals that are not otherwise connected. Several examples of this mode have been reported.3 As might be expected, the different ends of carbon monoxide show different preferences for metal center types. The carbon-donor end binds to low valent, electron-rich metals in much the same manner as do terminal carbonyls. The oxygen-donor end, by contrast, is more basic and inevitably binds to Lewis acidic centers. In fact, this bonding mode is commonly found between transition metal carbonyl anions and Lewis acidic main group cations such as Al(III), 9 Mg(II), 10 and Na(I).11 The major resonance structures for the ll2-(c-n1; 0-nl) carbonyl bridges in 3 are shown below. A involves a dative interaction Zr ..-:Q=C-M A .. ..,_. Zr ....:Q=C =M ..-. Zr-0-C=M .. B c between an anionic metal carbonyl complex and a cationic zirconium center. In B the dative interaction still persists, but there is some 1T -backbonding. C represents the structure with complete 1T -backbonding from the metal to the bridging ligand. -15- symmetrical bridge II 11 Semibridging 11 •• , •• M==C= O.. Figure 4. Principal binuclear bridging modes of the carbonyl ligand. Ml -16- Based on resonance structures A and B, one might expect that the metal centers could be solvated in coordinating media. Norton et al. reported that Cp2Zr( C(O)Me) ( CpMo(C0)3) appears to be a 1:1 electrolyte in acetonitrile. 4 In addition, they report that the infrared spectrum of this compound in acetonitrile is consistent with free ( CpMo(C0)3)anion (eq. 4). A similar result has been reported in the case of ( 4) Cp*2Ti( CpMo(C0)3) 2 in DME.12 Compounds 3a and 3b are not stable in DME or acetonitrile, but the solution infrared spectra taken with less than 50% decomposition show the absence of the pair of strong bands ca. 1550 cm-1, and the appearance of a single, weak band at 1590 cm-1. Several new bands are found between 2000 and 1700 cm-1. This is consistent with dissociation of at least one CpM(C0)3 anion, although Cp*2Zrn+ (n = 1 or 2) appears to be a strong enough Lewis acid to undergo subsequent further reactions with DME. In C on the other hand, the backbonding results in an "oxy-carbyne" structure with no formal charge separation. This implies a shortened M-C distance and a lengthened C-0 distance. In 3a, the latter distance is clearly 0 longer than in the terminal car bony 1 ligands ( 1.217(9) vs 1.150(11) A (average)), but this is still ca. 0.2 ~ shorter than a typical C-0 single bond.l3 Stuckey has reported the structure of a cationic titanium-DME -17- complex, ( Cp2TiDME) 2+( Zn2Cl6) 2-, in which the C-0 single bond in the 0 bound DME is 1.305(8) A.14 The Mo-C distance is significantly 0 shorter than in the terminal carbonyls ( 1.868(7) vs 1.942(8) A (average)). There are data available for only one molybdenum carbyne, Re(C0)9Mo(CPh).15 0 0 At 1.835(25) A the M-C triple bond is slightly longer than the 1.829 A calculated from average Mo-Mo and C-C triple bonds.15 Reported W-C triple 0 0 bond distances vary over the range 1.775 to 1.842 A, or at least 0.1 A shorter than typical W-C and Mo-C distances in carbonyls.l6 These comparisons do not unambiguously demonstrate participation of structure C in the bonding picture of 3a, but the trends in bond length are in the correct direction. Recently Casey and coworkers reported a new structural pattern for compounds containing both zirconium and a second transition meta1.2c The complexes Cp2Zr(R)( CpM(C0)2) (R =Cl, Me, 0-_!-Bu; M =Fe, Ru) represent rare examples of unbridged early/late transition metal M-M bonds. The electron density on the metals in the CpM(C0)2 anions is considerably higher than on the molybaenum in CpMo(C0)3- which has one additional carbonyl and two fewer d-electrons. The iron and ruthenium anions are also much less stericaUy encumbered. Both factors clearly favor direct metal-zirconium interactions. The absence of low energy bands in theIR make assignment of this unbridged structure relatively simple. The problem arises in structures containing bonds between zirconium and a metal with type III bridging carbonyls. Distinguishing between type III and type IV bridging carbonyls on the basis of C-O stretching frequencies is not a straightforward matter. Type IV ll-CO stretches have been reported -18- between 1690 and 1533 cm-1.3 By comparison, type III 11-CO stretches range from 1683 to 1560 em -1.1 ,2 Our original intention in synthesizing 8 was to increase electron density of the molybdenum sufficiently to disfavor a type IV bridge and favor a Mo-Zr bond. The infrared spectrum of 8 shows bands at 1843 and 1477 cm-1. While the structural assignment is open to debate, we believe that the low energy band at 1477 cm-1 is due to a type IV bridge, analogous to 3a. It is likely that the combination of Cp* rings and the large anion precludes the close approach of the directly bonded structure on steric grounds. In keeping with the increased electron density in the phosphine substituted case, the CO stretching frequencies in 8 are all shifted to lower energy. That the bridging CO band shifts to a greater extent is to be expected, as the zirconium should increase the n -acceptor properties of this ligand as suggested by resonance structure C. The formation of 3, 7, and 8 by dinuclear elimination is well precedented.2b,3a It has quite reasonably been suggested that later metal hydrides react with zirconium alkyls via a proton transfer mechanism. 4 The reduction of the metal-metal bond in 2 by 1 bears further comment. Using the bonding description A, the net result of the reaction of 1 with 2 is oxidation of zirconium from the divalent to the tetravalent state with concomitant reduction of the M-M bond yielding two CpM(CO)J anions. This is comparable to the two electron reduction of 2a,b with Mg/Hg.l 0 In addition, recent reports describe Cp*2Yb3f and Cp2Ti(C0)23a as one electron reductants with Co2(C0)8 and 4. The reduction of 2 may be envisioned as proceeding via outer sphere electron transfer to 2 from 1 prior to dissociation of the dinitrogen ligands, or alternatively, by coordination of the molybdenum -19- dimer to permethylzirconocene followed by inner sphere electron transfer. The inner sphere process could proceed via loss of the N2 ligands in l, followed by formation of the transient zirconium(II) complex D. Electron I Q_'c 'MCp(C0)2 cp;zr \ cf'c /MCp(C0)2 D transfer would then yield 3. Type III (n-bound) carbonyl bridges are fairly common among early transition metal complexes, although a structure held together only by such linkages has no precedent.17 Outer sphere reduction of 2 by divalent zirconium is indicated in the formation of 3a in a different system. Recent studies of the carbonyl coupled compound Cp*2Zr(02C2)Fe2CP2(C0)2, 9, have shown that this complex undergoes several substitution reactions through a Zr(II) intermediate.l8 While the exchange of ( Cp'Fe(C0)2) 2 (Cp':::C 5H4Me) for ( CpFe(C0)2) 2 in 9 takes >one week at 25oc (eq. 5), 9 reacts with 2a to produce 3a and Cp* Zr(o c )Fe cp (C0) 2 2 2 2 2 2 + 9 >1 week Cp* Zr(o c )Fe cp• (C0) 2 2 2 2 2 2 9 + -20- Cp* Zr(o c )Fe Cp (C0) 2 2 2 2 2 2 9 seconds benzene + (CpFe(C0) ) 2 2 (CpMo(C0) ) 3 2 2a + Cp* Zr(CpMo(C0) ) 2 3 2 3a (6) ( CpFe(C0)2) 2 within seconds (eq. 6). This has been explained in terms of an outer sphere electron transfer process, i.e., without coordination of 2a.l8 In lieu of some method for distinguishing between inner and outer sphere processes in the reaction of 1 with 2, both mechanisms must be considered viable. Merola has proposed that the reaction has proposed that the reaction of Cp*2Zr(C0)2 with the unsaturated dimer 4 proceeds via initial CO transfer to yield 2a and titanocene, "Cp2Ti", which subsequently form the observed product, ( Cp2THCpMo(C0)3)) 2·3e Our findings are in accord with this proposal. Cp*2Zr(C0)2 reacts with 4 to produce 3a (eq. 2) but under much more forcing conditions than required by reaction of 1 with 2a. -21- Experimental Section General Considerations. All of the manipulations were carried out using either high vacuum line or glove box techniques. Solvents were purified by vacuum transfer first from LiAlH4, then from "titanocene," prepared as described ear Her .19 Benzene-~6 and toluene-~g (Stohler, Inc.) were vacuum transferred from "titanocene." Hydrogen and deuterium (Matheson) were 0 purified by passage over activated 4 A molecular sieves and MnO on vermiculite.20 Carbon monoxide and hydrogen chloride (MCB) were used directly from the cylinder. 1H NMR spectra were recorded on Varian EM-390 and JEOL FX-90Q spectrometers. The latter machine was used to measure 13c spectra. Infrared spectra were measured on a Beckman IR-4240 spectrophotometer. { Cp*2ZrN2} 2N2,21 Cp*2ZrH2,22 Cp*2Zr(D)(iso-C4HgD),22 Cp*2Zr(C0)2,23 Cp*2Zr(CH3)2,22 Cp*2HfH2,5 ( CpMo(C0)3 )2, 24 ( CpMo(C0)2) 2,25 and CpMo(C0)3H26 were prepared as described previously. Elemental analyses were performed by Bernhardt Microanalytical Laboratory, Dornis and Kolbe Microanalytical Laboratory, or by Larry Henling at the CIT analytical service. Several reactions were carried out in sealed NMR tubes and monitored by NMR spectrometry. A typical example is the reaction of Cp*2Zr(C0)2 with ( CpMo(C0)2 )2: 20 mg (0.048 mmol) Cp*2Zr(C0)2 and 20 mg (0.046 mmol) ( CpMo(C0)2 )2 were placed in an NMR tube sealed to a ground glass joint and fitted with a Teflon needle valve adapter. Benzene-~6 (ca. 0.3 mL was distilled into the tube at -780C and the tube was sealed with a torch. -22- Cp*2Zr( CpMo(CO)J) 2' 3a. Onto 300 mg (0.37 mmol) 1 and 365 mg (0.75 mmol) 2a was distilled 50 mL toluene at -1960C. The mixture was aJlowed to warm to 250C with stirring. Within 20 min the solution had gone from deep purple to dark orange-red. After an additional 2 h stirring the solvent was stripped off and the evolved gas was collected and analyzed via Toepler pump. Non-combustible gases totaled 1.06 mmol, or 95% of the expected N2. The flask containing the residue was attached to a small swivel frit. Recrystallization from toluene yielded 340 mg (54%) of red-brown micro-crystalline solid. A second crop of 120 mg ( 19%) was obtained in the same manner. Anal. calcd. for C36H4oMo206Zr: C, 50.76; H, 4.73; Mo, 22.53; 0, 11.26; Zr, 10.71. Found: C, 50.52; H, 4.85; Mo, 22.69; Zr, 10.88. Cp*2Zr( CpW(CO)J)2, 3b. Onto 410 mg (0.51 mmol) 1 and 610 mg (0.92 mmol) 2b was distilled 30 mL toluene at -1960C. The mixture was allowed to warm to 250C with stirring. After 1 h the red-brown solid was collected on the frit, washed once with 5 mL toluene and dried in vacuo, yielding 655 mg (70%). Anal. calcd. for C3oH4Q06W2Zr: C, 42.07; H, 3.92; O, 9.34; W, 35.78; Zr, 8.88. Found: C, 42.68; H, 4.07; Zr, 9.38. Cp*2Hf(H)( CpMo(CO)J), 7. Onto 450 mg (1.00 mmol) Cp*2HfH2 and 245 mg (1.00 mmol) 6 was distilled 20 mL toluene at -1960C. The mixture was allowed to warm to -780C with stirring, at which point the solution began to turn light yellow with gas evolution. After 1 h gas evolution had ceased. The yellow solution was then filtered, and the solvent stripped off to leave a yellow oil. Trituration with 20 mL petroleum ether, followed by filtration -23- and a single washing with petroleum ether yielded 490 mg (71 %) of yellow crystalline solid. A second crop of 90 mg (13%) was obtained in the same manner. Anal. calcd. for C28H36HfMo03: C, 48.39; H, 5.22; Hf, 25.68; Mo, 13.81; 0, 6.90. Found: C, 48.21; H, 5.10; Hf, 2.5.91. Cp*2Zr(H)( CpMo(C0)2{PMeJ)), 8. A solution containing 85 mg (0.03 mmol) CpMo(C0)2(PMe3)H in 15 mL petroleum ether was added dropwise via syringe to 110 mg (0.03 mmol) Cp*2ZrH2 in 15 mL petroleum ether at -78°C. During addition the reaction mixture went from pale yellow to red-orange. The solution was allowed to warm to 25oc, and was filtered. Solvent volume was reduced to 10 mL and cooled slowly to -780C. Filtration and removal of solvent yielded 105 mg (55%) of yellow solid. 1H NMR of the filtrate showed a mixture of mono- and di-molybdate species. Anal. calcd. for C3QH45Mo02PZr: C, 54.94; H, 6.92; Mo, 14.63; 0, 4.88; P, 4.72; Zr, 13.91. Found: . C, 54.87; H, 7.01; P, 4.86; Zr, 14.12. Exchange of ll2 With Cp*2Hf(H)( CpMo(CO)J), 7. A solution of 50 mg 7 in 25 mL petroleum ether was stirred under 1 atm D2 for 24 h at 25°C. The solvent was stripped off, and 45 mg yellow solid scraped from the flask. 1H NMR showed an unchanged spectrum of 7 minus the hydride signal at 12.28 ppm. 2H NMR showed a single peak at 12.26 ppm. Reaction of Cp*2Zr(DXiso-C4HgD) With CpMo(CO)JH. Cp*2Zr(D)(isoC4H8D) (35 mg, 0.08 mmol), 6, (40 mg, 0.16 mmol), 0.3 mL of benzene-2_6, and TMS were sealed in an NMR tube. Within 1 min of reaching 25°C the mixture was deep red-orange. The 1H NMR spectrum indicated quantitative conversion to Ja and (CH3)3CD 0:1:1 triplet at 0.85 ppm, JH-D = 1Hz). -24- GC-MS of residual gas in NMR tube indicates less than 5% of (CH3)2(CH2D)CD (mass 60). Reaction of { Cp*2ZrN2}2N2 With ( CpMo(C0)2)2· Compound 1 (20 mg, 0.03 mmol), 4 (22 mg, 0.05 mmol), 0.3 mL of benzene-d6, and TMS were sealed in an NMR tube. Within 1 min of reaching 25oc the mixture was black purple. The 1H NMR spectrum indicated a nearly quantitative reaction with two new singlets at 5.13 and 1.79 ppm in a ratio of 1:3. After one week at 250C the spectrum showed complete decomposition. Separate attempts to isolate this material at low temperature were not successful. Reaction of { Cp*2ZrN2}2N2 With (1) ( CpMo(C0)2)2 (2) CO. Onto 1 (20 mg, 0.03 mmol) and 4 (22 mg, 0.05 mmol) was distilled 10 mL of toluene at -780C. After stirring at 250C for 10 min the purplish black solution was cooled to -780C and 400 torr CO introduced into the flask. After stirring at 25oc for 1 h the dirty red-orange solution was stripped of solvent. 1H NMR of the residue indicated that while the reaction was not quantitative, the majority of the product was 3a; Reaction of Cp*2Zr(C0)2 with ( CpMo(C0)2)2. Cp*2Zr(C0)2 (20 mg, 0.05 mmol), ( CpMo(C0)2) 2' (20 mg, 0.05 mmol), 0.3 mL of benzene-~6, and TMS were sealed in an NMR tube and the reaction was periodically monitored by 1H NMR spectrometry. Conversion to 3a was complete after 12 hat 80°C. Reaction of Cp*2Zr(CHy2 with CpMo(C0)3H. Cp*2Zr(CH3)2 (32 mg, 0.08 mmol), CpMo(C0)3H (40 mg, 0.16 mmol), 0.3 mL of benzene-~6, and TMS were sealed in an NMR tube. Reaction of initially one, followed by a second equivalent of CpMo(C0)3H was followed by 1H NMR spectrometry. -25- Quantitative conversion to 3a was complete after ten days at 250C Attempted Reaction of 3a with H2. Compound 3a (10 mg, 0.01 mmol), H2 (0.07 mmol), 0.3 mL of benzene-2_6, and TMS were sealed in an NMR tube at -1960C. No reaction was observed by 1H NMR spectrometry after three months at 800C. Attempted Reaction of 3a with CO. Compound 3a (1 0 mg,O.O 1 mmol), CO (0.04 mmol), 0.3 mL of benzene-Q6, and TMS were sealed in an NMR tube. No reaction was observed by 1H NMR spectrometry after one week at 800C. Reaction of 3a with HCJ. Compound 3a (10 mg, 0.01 mmol) HCl (38 torr, 19.8 mL at 25oc, 0.04 mmol), 0.3 mL of benzene-Q6, and TMS were seale~ in an NMR tube. Immediately upon warming to 25oc the mixture became light yellow. The 1H NMR spectrum indicated quantitative conversion to Cp*2ZrCl2 and CpMo(C0)3H, by comparison with the known compounds. Reaction of 3a with H20. Compound 3a ( 10 mg, 0.01 mmol), H20 (0.02 mmol), 0.3 mL of benzene-Q6, and TMS were sealed in an NMR tube. Immediately upon warming to 25oc the mixture was light yellow. The 1H NMR spectrum indicated quantitative conversion to CpMo(C0)3H and Cp*2Zr(OH)2· Reaction of 3a with CHJI. Compound 3a (10 mg, 0.01 mmol), CH3I (22 torr, 19.8 mL at 25°C, 0.02 mmol), 0.3 mL of benzene-Q6, and TMS were sealed in an NMR tube. Reaction was complete after three days at 800C. The 1H NMR spectrum indicated the major Zr containing product was Cp*2Zri2. Some CpMo(C0)3CH3 was observed along with an equal amount of ( CpMo(C0)3 )2. -26- structure Determination of Cp*2Zr( CpMo(CO)J) 2· A series of oscillation and Weissenberg photographs (Co Ka) indicated a monoclinic ceJJ with systematic absence hkl, b. + ~odd and b_O!, ! odd, consistent with space groups Cc or C2/c. A crystal with dimensions 0.4 x 0.2 x 0.2 mm was mounted in a glass capillary under N2 with its long axis slightly skewed to the cp-axis of a Syntex P21 diffractometer. Unit cell dimensions were obtained by a least-squares fit of the sin2e /X 2 values of 15 reflections measured on the diffractometer. The calculated density of 1.65 g/cc (assuming Z =4) is reasonable for organometaJJic compounds of this type. Using Mo Ka radiation monochromated with a graphite crystal, one hemisphere of data was measured between 1.0 and 55.00 in 29, using a a-29 scan technique with a scan range of 1o and a scan rate of 1Of min. Background counts were taken for 15 seconds before and after each scan. Intensities of three check reflections measured every 100 reflections showed a systematic rise of ca. 3%. This was attributed to the X-ray tube current rising from 20.0 to 20.6 rnA over the course of 4443 reflections. The intensities were scaled to account for this rise over the collection time. Intensities and their variances were corrected for Lorentz and polarization effects. No absorption correction was applied (ll = 25.3 cm-1 ). The data set was placed on a absolute scale by means of a Wilson plot using scattering factors for Zr and Mo calculated from Cromer and Mann,27 and for C,H, and 0 from Cromer and Waber.28 The Wilson statistics clearly indicated a centric space group. Thus the correct space group is C2/c, as confirmed by the successful structure solution and refinement. Systematic absences and one reflection having an extremely unsymmetrical background were deleted, leaving 3523 reflections -27- for the working data set. Solution and Refinement.29 The locations of the three metal atoms were determined from a three-dimensional Patterson map. This shows that the molecule sits on a two-fold rotation axis, making only half of the atoms unique. A structure factors computation using these positions gave an R index (1: II F0 1-1 Fe II /1: I F 0 I) of 0.506. The Fourier map showed the positions of all of the non-hydrogen atoms. Several cycles of least-squares refinement using isotropic Gaussian amplitude parameters led to a R index of 0.165. At this point anisotropic Gaussian amplitudes were introduced. Two cycles of least-squares refinement yielded an R index of 0.106. Since the hydrogen atom positions could not be reliably determined from the difference Fourier map, they were placed at positions calculated using C-H bond lengths 0 of 0.9 A and tetrahedral and trigonal geometries around Cp* and Cp carbons, respectively. This allowed further refinement of the thermal parameters for these carbon atoms. Two cycles of refinement, leaving the hydrogen positions and their Gaussian amplitudes unrefined, and including the scale factor, gave an R index of 0.09 5 with a goodness of fit of 1.34. Further leastsquares refinements yielded no decrease in the R index. A structure factors calculation using only data with F~ ~ 3cr(F~) gave an R index of 0.053 (2113 reflections). Difference Fourier maps generated in the plane of the Cp ring and in the planes perpendicular to the ring carbon-methyl carbon bonds in the Cp* did not help in locating the true positions of the hydrogen atoms. Relevant data concerning the structure are presented in Tables 2-5. Supplementary material including calculated hydrogen atom positions, structure factor amplitudes, positional and thermal parameters is available. -28- Table 5. Least-Squares Planes 0 Atom Deviation From Plane (A) C1 C2 C3 C4 C5 0.0078 -0.0119 -0.0006 0.0114 -0.0066 C14 C15 C16 C17 C18 -0.0063 0.0019 0.0081 -0.0070 0.0033 -29- References and Notes (1) Examples of Group V/late metal M-M bonded complexes: (a) Pasynskii, A. A.; Skripkin, Y. V.; Suvorova, K. M.; Kuzmichea, 0. N.; Kalinnikov Izv. Akad. SSSR, Ser Khim. 1978, 1226. (b) Wong, K. S.; Labinger, J. A. J. Am. Chern. Soc. 1980, 102, 3652. (c) Labinger, J. A.; Wong, K. S.; Scheidt, W. R Ibid. 1978, 100, 3254. (d) Douglas, W. E.; Green, M. L. H.; Prout, C. K.; Rees, G. V. J. Chern. Soc., Chern. Commun. 1971, 896. (2) Examples of Group IV/late metal M-M bonded complexes: (a) Barger, P. T.; Bercaw, J. E. J. Organomet. Chern. 1980, 201, C39-C44. (b) Longato, B.; Norton, J. R.; Huffman, J. C.; Marsella, J. A.; Caulton, K. G. J. Am. Chern. Soc. 1981, 103, 209. (c) Casey, C. P.; Jordan, R. F.; Rheingold, A. L. Ibid. 1983, 105, 665-667. (d) Schmid, G.; Stutte, B.; Boese, R. Chern. Ber. 1978, 111, 1239. (3) (a) Renaut, P .; Tainturier, G.; Gautheron, B. J. Organomet. Chern. 1970, 150, C9-C10. (b) Schneider, M.; Weiss, E. Ibid. 1976, 121, 305371. (c) Crease, A. E.; Legzdins, P. J. Chern. Soc., Dalton 1973, 1501-1507. (d) Hamilton, D. M., Jr.; Willts, W. S.; Stucky, G. D. l! Am. Chern. Soc. 1981, 103, 4255. (e) Merola, J. S.; Gentile, R. A.; Ansell, G. B.; Modrick, M.A.; Zentz, S. Oganometallics 1982, !, 173117 33. (f) Tilley, T. D.; Anderson, R. A. J. Chern. Soc., Chern. Commun. 1981' 985-986. (4) Marsella, J. A.; Huffman, S. C.; Caulton, K. G.; Longato, B.; Norton, J. R. J. Am. Chern. Soc. 1982, 104, 6360. (5) Seidler, P. F.; Frysnk, M. D.; Roddick, D. M.; Hillhouse, G. L.; -30- Bercaw, J. E. Manuscript in prepartion. (6) Carried out at the CIT X-ray facility with the invaluable assistance of Dr. William P. Schaefer. (7) A recent search of the Cambridge Crystallographic Database turned up 1545 compounds containing M-M bonds,. of which 927 contained carbonyl ligands. Of these, 228 contained binuclear bridging carbonyls. (8) Powell, H. M.; Ewens, R. V. G. J. Chern. Soc. 1939, 286. (9) (a) Peterson, R. B.; Stezowski, J. J.; Wan, C.; Burlitch, J. M.; Hughes, R. E. J. Am. Chern. Soc. 1971, 93, 3532-3533. (b) Conway, A. J.; Gainsford, G. J.; Schrieke, R. R.; Smith, J. D. J. Chern. Soc., Dalton 1975, 2499-2507. (10) Ulmer, S. W.; Skarstad, P.M.; Barlitch, J. M.; Hughes, R. E. J. Am. Chern. Soc. 1973, 95, 4469-4471. (11) Cooper, M. K.; Duckworth, P. A.; Henrick, K.; McPartlin, M. J. Organomet. Chern. 1981, 212, 10. (12) Merola, J. S.; Camp, K. S.; Gentile, R. A.; Modrick, M. A.; Zentz, S. Organometallics, in press. (13) Sutton, L. E., Ed. "Tables of Interatomic Distances and Configuration in Molecules and Ions"; The Chemical Society: London, 19 58, S 17. (14) Sekutowski, D. G.; Stuckey, G. D. Inorg. Chern. 1975, ~' 2192-2199. ( 15) Fischer, E. 0.; Huttner, G.; Lindner, T. L.; Frank, A.; Kreiss!, F. R. Angew. Chern., Int. Ed. Engl. 1976, .!2, 157. (16) (a) Neugebauer, D.; Fischer, E. 0.; Dao, N. Q.; Schubert, U. J. Organomet. Chern. 1978, 153, C41. (b) Fischer, E. 0.; Gammel, F. J.; Neugebauer, D. Chern. Ber. 1980, 113, 1010. -31- ( 17) Such a bonding mode is suggested by the spectral properties of a complex of cis-Cp2Ru2(C0)2(ll-CO)(ll-CH2) with permethylzirconocene.18 ( 18) Chapter II of this thesis. (19) Marvich, r. H.; Brintzinger, H. H. J. Am. Chern. Soc. 1971, 94, 2046. (20) Brown, T. L.; Dickerhoof, D. W.; Bafus, D. A.; Morgan, G. L. Rev. Sci. Instrum. 1962, 33, 491. (21) Manriquez, J. M.; Bercaw, J. E. J. Am. Chern. Soc. 1974, 96, 6229. (22) Manriquez, J. M.; McAlister, D. R.; Sanner, R. D.; Bercaw, J. E. J. Am. Chern. Soc. 1978, 100, 2716. (23) Manriquez, J. M.; McAlister, D. R.; Sanner, R. D.; Bercaw, J. E. J. Am. Chern. Soc. 1976, 98, 6733. (24) Birdwhistell, R.; Hackett, P.; Manning, A. R. J. Organomet. Chern. 1978, 157, 239. (25) Klinger, R. J.; Butler, W.; Curtis, M. D. J. Am. Chern. Soc. 197.5, 97, 3535. (26) Piper, T. S.; Wilkinson, G. J. lnorg. Nucl. Chern. 19.56, 1, 104. (27) Cromer, D. T.; Mann, B. Acta Cryst. 1978, 24a, 321. (28) Cromer, D. T.; Waber, J. T. Acta. Cryst. 196.5, _!!, 104. (29) Calculations with the CR YM system of computer programs. -32- CHAPTER II Reductive Coupling of Carbonyl Ligands in Binuclear Transition Metal Complexes. Reversible C-C Bond Formation in Reactions of Permethylzirconocene. -33- Introduction There has been a great deal of interest recently in complexes containing both early and late transition metals. The relative paucity of such complexes combined with the belief that they might behave quite differently from complexes containing more similar metals has sparked many research efforts in this area. Our original involvement with heteronuclear complexes of zirconium stemmed from studies of the reduction of transition metalbound carbon monoxide by Cp*2ZrH2 ( Cp*:n5-C5(CH3)5) .1 The formation of zirconoxy carbene complexes by hydride transfer to terminal carbonyl ligands has been shown to be a fairly general process.2 However, it was observed that in certain instances hydride transfer did not occur .3 In these cases reductive elimination was induced and complexes of permethylzirconocene and the transition metal carbonyl compounds resulted. For example, Cp*2ZrH2 reacts with CpCo(C0)2 to produce H2 and Cp*2Zr(C0)2CoCp, in which a Zr-Co bond is supported by two different types of bridging car bony1 ligands. 3 A natural extension of this work was the study of reactive Zr(II) complexes with transition metal dimers. In the initial studies it was found that divalent zirconium in the form of the dinitrogen complex {Cp*2ZrN2} 2N2, 1, reduces ( CpMo(C0)3) 2 to form Cp*2Zr ( CpMo(C0)3) 2 with evolution of N2. This work is described in detail in the preceding chapter. In our survey of the reactions of 1 with other metal dimers a second, more interesting mode of reactivity was discovered. Iron and ruthenium carbonyl dimers react with 1 to yield trinuclear complexes in which two carbonyl ligands are reductively coupled producing a new carbon-carbon bond -34- in the form of a 02C2 moiety linking the three metals. 4 The carbonyl coupled products are of interest both in terms of conversion of carbon monoxide into a two-carbon fragment and also as a new structural mode for early/late transition metal combination complexes. In this chapter the synthesis, characterization and reactivity of these and related complexes are discussed. Results { Cp*2ZrN2} 2N2, l, reacts with the transition metal carbonyl dimers ( (C 5H4R)M(C0)2) 2 (2a: M = Fe, R =H; 2b: M =Fe, R = Me; 2c: M =Ru, R = H) in toluene to yield nitrogen and complexes with the stoichiometries Cp*2Zr(02C2)M2(C0)2(C5H4R)2 (2a: M =Fe, R =H; 2b: M =Fe, R =Me; 2c: M = Ru, R = H). Spectral data for these compounds are listed in table 1. The l 2a,b,c 3/2 N 2 + Cp* Zr(o c )((C H R)M(C0)) 2 2 2 2 5 4 (1) Ja,b,c spectral properties of Ja are representative of the series. The 1H NMR spectrum of Ja exhibits two resonances at o 4.81 (lOH) and 1.69 (30H) attributable to Cp and Cp* protons. In the 13c NMR spectrum resonances at o 268 and 305 are observed in addition to signals for the Cp and Cp* ligands. The infrared spectrum for 3a exhibits four bands due to v(CO): 1768(w), 1729(s), 1295(ms), and 1282(ms) cm-1 ( v(13co): 1730, 1636, and 1260 cm-1 Table 1. NMR~ and IRE Data. Compound Infrared v(CO) Cp*2Zr(02C2)Fe2Cp2(C0)2 17 68( w), 1729(s), Assignment 13c{ 1H} NMR ).J-0 2 ~ 129 5(ms), 1282(ms) 3a 13co: 1730, 1686, 1260 (br) Chemical Shift 305 ).J-CO 268 ~(CH3)5 125 ~H5 91 C5(CH3)5 11 C5!::!5 4.81 (s, 10H) C5(C!:!3)5 1.69 (s, 30H) I 1H NMR Cp*2Zr(02C2)Fe2CP'2(C0)2 3b 17 55(w), 1713(s), 1287(s,br) 13c{ 1H} NMR ll-0 2c 2 \.,..) \.11 I 304 ll-CO 269 C5(CH3)5 124 f5(CH3)H4 102, 92, 90 C5(CH3)H4 13 C5(CH3)5 11 Table l. Continued. -Compound Infrared 'J(CO) Assignment 1H NMR Cp*2Zr(02C2)Ru2CP2(C0)2 17 40(s), 1720(s), 1300(s) C5(CH3)H4 4.79 (m, 8H) C 5( C!:!J)H4 2.12 (s, 6H) C5(C!:!J)5 1.67 (s, 30H) 13c{ lH} NMR lJ-0 2 ~ ll-CO 3c Chemical Shift 289 250 I \.,.) 0'\ I lH NMR Cp*2Zr ( Cp2Ru2(C0)3(CH2)) 9 1920(w), 1837(s), 1765(m), lH NMR £5(CH3)5 121 ~H5 91 C5(CH3)5 11 C5!:!5 5.38 (s, 10H) C5(C!:!3)5 1.68 (s, 30H) C!i2 9.26 8.85 (s, 1H) (s, 1H) C5H5 5.35 (s, 10H) C5(C!:!3)5 1.67 1.60 (s, 15H) (s, 15H) 17 59(s) Table l. Continued. Infrared v(CO) Compound {Cp*2Zr}204C4{CpFe}2 1329(s), 1250(m), 1034(s) Assignment 1H NMR (s, 10H, V(fwhm) = 33 Hz) C5H5 5.68 C5(CH3)5 1. 39 (s, 60H, V(fwhm) = 6 Hz) lOa 13co: 1300, 1235, 1010 Chemical Shift I ~ 1H NMR {Cp*2Zr}2o4c4{CpFe}2 lOb 1330(s), 1258(m), 1020(s) C5!::!5 4.47 (s, 10H) C5(C!:!J)5 2.06 (s, 60H) Chemical shifts reported in o relative to internal TMS. ~R spectra obtained as nujol mulls. Values reported in cm-1. ~NMR spectra in benzene-~6 at ambient temperature. "' I -38- (br, unresolved)). The structure of 3a as determined by single crystal X-ray diffraction methods is shown in Figure 1 and the skeletal view is shown in Figure 2. 4,5 Coupling of two carbonyl Jigands has given rise to the Zr02C2 unit bridging the iron atoms (A). These seven atoms all lie approximately in the same A plane with a perpendicular crystallographic mirror plane containing the zirconium atom bisecting the C-C and Fe-Fe bonds. Bond angles and distances are listed in Tables 2 and 3. Treatment of 3a with CO ( 4 atm, 250C) in benzene produces 2a and Cp*2Zr(C0)2, 4, quantitatively in about one week. Under the same conditions the ruthenium analog 3c yields 2c and 4 in under 10 min. When 13co is used, >9096 of the isotopic label is found as Cp*2Zr( 13co)2, as determined by 13c NMR. Cp* Zr(o c )Cp Ru (C0) 2 2 2 2 2 2 + 13co 10 min 3c Cp*2Zr(13C0)2 If ( 2) + 2c CIO' I VJ \.0 I Cl6 Figure 1. Molecular Configuration of Cp*2Zr(02C2) [ CpFe(CO)] 2, Ja. Cl8 1 -2.535(7) 1.89(3) I 4==" 0 I Figure 2. Skeletal view of Cp*2Zr(02C2)[ CpFe(CO)] 2' 3a. -41- 0 Table 2. Bond Distances in 3a (A). ATOM ATOM DISTANCE Zr Zr Zr Fe Fe C1 C2 C3 C4 01 RING1 RING2 RING3 Fe' C2 C3 C4 2.11(1) 2.23 2.27 1.76 2.535(7) 1.43(3) 1.43(3) 1.41(3) 1.36(3) 1.39(3) 1.47(3) 1.34(3) 1.42( 4) 1.62(5) 1.56(3) 1.62(4) 1.46(3) 1.36(3) 1.43(4) 1.56(5) 1.57(3) 1.55(3) 1.57(4) 1.82(2) 1.32(2) 1.91(3) 1.21(4) 1.89(3) 1.21(4) C5 C6 C7 C8 C6 C7 C8 C12 C13 C14 C12 C13 C14 C18 C18 C18 C19 C19 C20 C20 C5 C1 C7 C8 C8' C9 C10 C11 C13 C14 C14' C15 C16 C17 C18' Fe 01 02 02 Fe 03 *Ring refers to the centroid of an n5-c5 ring. * * * -42- Table 3. Bond Angles in 3a (deg). ATOM ATOM ATOM ANGLE C5 Cl C2 C3 C4 C9 C7 ClO ClO C6 Cll Cll C7 Cl5 Cl3 C16 Cl6 C12 Cl7 C17 Cl3 Fe Fe Cl8 Fe Fe Fe Fe RING! 01 C18 Cl8 C19 Cl C2 C3 C4 C5 C6 C6 C7 C7 C7 C8 C8 C8 Cl2 C12 C13 Cl3 Cl3 C14 C14 Cl4 C18 C18 Cl8' Cl9 Cl9 C20 C20 Zr Zr Fe Fe Fe C2 C3 C4 C5 Cl C7 C7' C6 C8 C8 C7 C8' C8' C13 C13' Cl2 C14 C14 C13 Cl4' Cl4' C18' 01 01 Fe 02 Fe 03 RING2 01' Cl9 C20 C20 116(2) 102(2) 106(2) 115(3) 10 1(3) 128(1) 103(3) 124(2) 126(2) 109(2) 123(2) 127(1) 109(1) 128(2) 101(3) 123(2) 125(2) 111(2) 124(2) 128(1) 108(1) 105(1) 141(1) 113(1) 83(1) 138(1) 84(1) 138(1) 140 77(1) 86(1) 85(1) 95(1) *Ring refers to the centroid of an n5-c 5 ring. * -43- In benzene solution a mixture of 3a and 2b slowly reaches equilibrium (ca. 1 week at 25°C) with 3b and 2a (Scheme I). Over longer periods of time ( >2 weeks) new 1H NMR signals consistent with CpCp'Fe2(C0)4, and Cp*2Zr(02C2)Fe2CpCp'(C0)2 (Cp'::C5H4Me) are also observed. Reaction of 3c with 2a results in complete conversion to 3a and 2c. The unbridged molybdenum dimer ( CpMo(C0)3 )2, 5 reacts with 3a to produc 2a and Cp*2Zr ( CpMo(C0)3 )2, 6, a complex in which there is no Mo-Mo bond and the two CpMo(C0)3 moieties are bound to a Cp*2Zr center via 112-(C- n1; 0- n1) carbonyl bridges.6 Cp* Zr(o c )Cp Fe (C0) 2 2 2 2 2 2 3a ( CpMo (CO) + seconds ) 3 2 5 Cp* Zr(CpMo(C0) ) 2 3 2 6 + (CpFe(C0) ) 2 2 ( 3) 2a While treatment 3a with H2 (4 atm, 25°C) in benzene results in decomposition to a mixture of unidentified products over one week, 3c cleanly forms the bridging zirconoxy carbene complex Cp2Ru2(C0)2( ll-CO)( ll-C(H)OZrHCp*2 ), 7 within 10 min with H2 under the same conditions. This compound has been prepared previously from the reaction of Cp*2ZrH2 and 2c (see Chapter III). Cp* Zr(o c )Cp Ru (C0) 2 2 2 2 2 2 + 10 min 3c Cp Ru (C0) (ll-C0)( 11 -C(H)OZrHCp* ) 2 2 2 2 7 ( 4) -44- Scheme I Cp I OXFeI'jf.o Cpz*Z( 0 0 Fe I 1 c~Fe Cp•zr 2 ~ \ \ 0~ Cp CP 0~ B 3a -2o lr \ Cp +2o (Cp 2*Zr] +2b c· I P OXFeI'ff.o 0 Cpz*Z( 0 Fe \ 3b I Cp o 1~ f/0 c~Fe lf -2b -45- Treatment of 3a or 3c with CHJI, HCl, H20, Me3SiCl, MeS02CF3 or dimethyl acetylenedicarboxylate results in rapid decomposition. Varying amounts of 2 are observed among the products. Compound 3a is also produced by treatment of Cp*2Zrl2 with Na( CpFe(C0)2) within 2 hat 250C in benzene. No reaction is observed between Cp*2ZrCl2 and Na( CpFe(C0)2) in benzene or DME after weeks at 25oc. 2 Na(CpFe(C0) ) 2 + 2 Nal ( 5) + 3a Proton NMR studies show that when a mixture of cis and trans isomers 8, is treated with 1, only cis-8 is initially consumed and a new product, Cp*2Zr( Cp2Ru2(C0)3(CH2)), 9, is formed within 20 min at 25°C (Scheme II). Excess 8 regains a cis-trans ratio of ca. 2:1 within 30 min. The 1H NMR spectrum of 9 shows methylene resonances at o 9.26 (lH) and 8.85 (lH), a Cp singlet at 5.35 (lOH) and two Cp* singlets at 1.67 (15H) and 1.60 (15H). Four bands attributable to \)(CO) are observed in the infrared spectrum (nujol mull) of 9: 1920(w), 1837(s), 1765(m) and 1759(s) cm-1. The absence of low energy C-O stretching bands implies a different structure for 9 than found for 2a. Compound 9 is only slightly soluble in hydrocarbon solvents, and while apparently stable indefinitely in the solid state, solutions of 9 decompose completely within two days in benzene at 250C to a myriad of products. Quantitative production of 4 and 8 from 9 and CO (4 atm, 250C) -46- Scheme n ll + C~Zr [CP2R~(C0)3(CH2)] 9 -47- requires less than 30 min and appears to be limited mainly by the rate of Cp* Zr(Cp Ru (C0) (CH )) 2 2 2 3 2 + <30 min co 9 Cp* zr (CO) 2 2 + Cp Ru (C0) (a-I ) 2 2 2 3 4 ( 6) 8 dissolution of 9 in benzene. Similarly, treatment of 9 with 2c yields 3c and 3 in <5 h at 250C (eq. 7). Cp* Zr(Cp Ru (C0) (CH )) 2 2 2 2 3 9 <5 h + 2c 3c 8 Compound 3a reacts with 1 in toluene to produce a paramagnetic complex { Cp*2Zr} 204C4 { FeCp} 2, lOa. This complex exhibits broad 1H NMR signals at o 5.68 ( 10H, V(fwhm) = 33Hz) and 1.39 ( 60H, V(fwhm) = 6Hz) and bands in the infrared at 1329 (s), 1250 (m), and 1034 (s) cm-1 (13co: 1300, 1235, and 1010 cm-1). The molecular weight of lOa as determined by ebulliometry is 1040 AMU (calc d. 1077). Magnetic susceptibility measurements show an effective magnetic moment, lleff, of 3.26 B.M. per Zr2Fe2 unit at 280 °K ( S = 1, ll (spin only) = 2.83 B.M.). Variable temperature measurements indicate slightly antiferromagnetic behaviour with a Nee! -48- temperature of ca. 30°K. The diamagnetic ruthenium analog, { Cp*2Zr} 204C4{ RuCp} 2, lOb, can be prepared from 1 and 3c. Both lOa and lOb can be prepared by treatment of one equivalent of the dimeric 1 with the dimers 2a and 2c, respectively. Compound 10 reacts with CO to produce insoluble products and trace amounts of 2 and 4. These reactions are currently under investigation. Discussion Structure of 3a. The Zr02C2 unit bridging the two iron atoms is clearly the most interesting feature of the structure of 3a. The bond distances are in accord with the dioxozirconacyclopenta-3,4-diylidene diiron formuo lation indicated in A. The C-C bond distance ( 1.57(4) A) is similar to the value found for C-C single bonds in simple hydrocarbons.? While the 0 C-0 distance ( 1.32(2) A) is rather short for a single bond,7 it is comparable with the C-O single bond length found for the DME molecule coordinated to a 0 Ti(III) cation in ( Cp2Ti(DME)) +2( Zn2Cl6) 2- ( 1.305(8) A) .8 That the 0 Fe-C(carbene) distance for 3a is 0.09 A shorter than the Fe-C(terminal carbonyl) for cis-2a is also reasonable, in view of the order of bond lengths typically found for "Fischer-type" carbenes ( M-C(carbene) > M-C(carbonyl)) .9 The coordination about the zirconium is normal, with the exception of the small 0-Zr-0 angle of 77(1)0 , imposed by the five-membered ring. Stuckey's titaniumDME cation exhibits a similar angle of 76.6(2)0.8 An alternative resonance structure for the bridging moiety, C, in which there are Fe-Fe and C-C double bonds is not consistent with the -49- 0 observed bond lengths. The Fe-Fe bond ( 2.535(7) A) is no shorter than in 0 cis-2a ( 2.531(2) A). Indeed, none of the bond distances for the CpFe(ll-C0)2FeCp moiety are statistically different from those in cis-2a.l0 p Fe Zr D Fe c The spectral features for 3a are readily interpretable in light of its structure. Thus the infrared bands at 1768 and 1729 cm-1 are attributed to the bridging carbonyls of the Cp2Fe2(ll-C0)2 unit (cf. 1800 and 1765 cm-1 for cis-2a)10 and the bands at 1295 and 1282 cm-1 (principally) to the \)(CO) modes of the dioxozirconacyclopenta-3,4-diylidene moiety. 13c NMR spectra (benzene-2_6) for 3a show two downfield resonances, at o 305 and 268 (1:1). 13c NMR spectra of cis-2a at low temperatures exhibit resonances at 0 212 and 262 due to terminal and bridging carbonyls, respectively.ll Hence the most straightforward assignment for 3a is that the 0 268 and 305 resonances are due to the bridging carbonyl and carbene carbons. Compound 3c, the ruthenium analog of 3a, has a similar car bony!coupled structure. As in 3a, a broad low energy band at 1300 cm-1 in the infrared spectrum is found for the 02C2 bridge, in addition to stretches at 1740 and 1720 cm-1 for the carbonyls bridging the two ruthenium atoms. The -50- downfield resonances at o 289 and 250 in the 13c NMR spectra for 3c can be assigned to the 02C2 and bridging carbonyl carbons.l2 The Carbonyl Coupling Reaction. The formation of 3 may be viewed as a reductive coupling of the two terminal carbonyl ligands of 2 with oxidation of zirconium from the divalent to tetravalent state. In this respect reaction 1 may be likened to the reductive coupling of aldehydes4 and terminal olefins14 by permethylzirconocene. While radial pathways cannot be excluded, it is likely that all of these processes involve intramolecular coupling of two 1T -bound ligands. Indeed, there is precedent for "side-on" 1T -bonding between early transition metals and bridging carbonyl ligands.3,15 While a complex held together by only~ -(C-n 1; C,O-n 2) bridging carbonyls as in 8 is not precedented, the structure of 9, the product of the reaction of 1 with the bridging methylene complex cis-8, raises some intriguing questions. It is the absence of the low energy v(CO) bands in the infrared spectrum of 9 that rules out a structure analogous to 3a. The 1H NMR spectrum of 9 exhibits signals due to inequivalent methylene protons, a single Cp resonance, and a pair of inequivalent Cp* signals. The four bands between 1920 and 17 59 em -1 in the infrared spectrum are not especially helpful, except in eliminating a 02C2 bridge from consideration. We favor the structure analogous to 8 shown below. In this complex a zirconium(II) center is coordinating to the two terminal carbonyl ligands of cis-8 in a 1T -fashion. Infrared bands for 1T -bound carbony Is have been reported from 1560 to 1683 cm-1.3,15 While the stretches observed for 9 fall outside of this domain, there is reason to believe the Zr-carbonyl interactions may be weaker than in these cases. In aU of the previously observed structures of 1T -bound -51- carbonyls direct bonds between the metals add to the stability of the carbonyl bridge. In addition, the wide "bite" of the cis-Ru2(C0)2 chelate in D may result in less than optimal Zr-carbonyl interactions. Such a weakened interaction would result in higher stretching frequencies for the car bony 1 ligands in D. D It is not possible to rule out less symmetrical structures which are fluxional on the NMR time scale. The low solubility of 9 has prevented low temperature studies. In considering such structures, it must be taken into account that 1 reacts only with cis-8. In addition, the rapid substitution chemistry of 9 requires a structure which readily releases 8. StructureD seems the most straightforward proposal in all of these regards. It is significant that permethylzirconocene forms a complex with an apparently intact metal carbonyl dimer such as cis-8 without carbonyl coupling. It is difficult to explain this fact on the basis of structural -52- differences between cis-8 and 2c. The structure of a related 1-1-alkylidene complex of ruthenium, cis-Cp2Ru2(CO)J(CMe2)16 exhibits a Ru-Ru bond length ( 2.712(1) K) essentially. identical to that found in trans-2c (2.72 X>.17 In addition, the disposition of the cis terminal carbonyls in this compound is not significantly different from that found in cis~2a.l0 On the other hand, the different structures of 3 and 9 can be justified on the basis of electronic arguments. Replacement of one carbonyl by a methylene group in 8 results in additional lT-backbonding to the remaining carbonyl ligands. The increased electron density on the terminal carbonyls would render the reductive coupling less favorable. In keeping with this proposal, the observed carbonyl stretching frequencies for 8 are found at lower energies than in 2c.18 Other processes involving the reductive coupling of ligands have been reported. Lippard and coworkers have studied the coupling of adjacent isocyanides in a seven-coordinate molybenum complex.19 In this case the reductant is zinc metal and the product upon hydrolysis is a molybdenumbound bis(alkylamino)acetylene. A similar process has been suggested in the formation of a diniobium complex. The structure of Nb2Cl6(t-BuNC)4(1J-t-BuNCCN-t-Bu), 11, reported by Cotton amd Roth,20 apparently arises from the coupling of cis isocyanides. While the mechanism of formation is uncertain, the C2(NR)2Nb moiety bears a strong resemblance to the C202Zr unit in 3a. Bond orders >1 along the N- C-C- N frame, however, suggest significant differences. The structure analogous to A in the dinuclear 11 would be a metallacyclopropadiene. It is not surprising that this hypothetical ring is not observed and that the electrons are distributed in a different manner along the reduced ligand framework than in A. However, comparison -53- 11 with A does raise the interesting question as to whether an unsaturated metal center would insert into a Nb-C bond in 11 to produce a dimetallacyclobutadiene structure. It is also significant that the two niobium atoms have such different environments. The NbCl4 unit should be Lewis acidic in much the same manner as is a zirconium(IV) center. The NbCl2(CNR)4 center, however, is relatively electron rich. This observation is perhaps the only factor in common between 3, these two coupled-isocyanide cases and two additional reports of carbon-carbon bond formation via reduction of ligands. Lukehart21 and Cramer22 have published work describing C-C bond formation in boron-iron and uranium-iron complexes. The mechanisms of these transformation are unclear, but both products contain oxygen substituted allyl ligands bound to iron centers. The presence of five-membered rings containing Lewis acids bound to heteroatoms in 3, 11 and these systems suggests that chelation may play an important role in positioning ligands prior to reductive bond -54- formation. Discussion of one other related system is in order. Casey and coworkers have recently reported the crystal structures of Cp2Zr ( CpM(C0)2) 2 (12a: M =Fe, 12b: M =Ru).23 These compounds contain two separate CpM(C0)2 units bound via unbridged M-M bonds to the zirconium. Compound 12 is prepared by treatment of Cp2Zri2 with two equivalents of Na ( CpM(C0)2). This is in striking co':ltrast to the analogous reaction in the Cp* system (eq. 5) in which 3a is produced. These different_ results can be understood by consideration of the steric and electronic differences between Cp2Zr and Cp2*Zr systems. The larger steric bulk of the Cp* rings would favor a larger separation between the metals and associated ligands. This is the case in 3 where there are three intervening atoms between the metals. In addition, the greater electron donating properties of the permethyl ri~gs will result in a nigher electron density on the zirconium in the Cp*2Zr fragment than in Cp2Zr. The lower oxidation state in 12 (formally Zr(II)) is thus less well suited to the permethyl ring system. The interesting question of which structural mode will be adopted by the complexes with partially methylated rings is currently being studied by Casey and coworkers.23 Reversibility of the Coupling Process. In an effort to exploit the C-C bond formation in the carbonyl coupling process we have explored the reactivity of these compounds. The first experiment that often comes to mind in these laboratories is treatment of a new compound with carbon monoxide. Coupling of carbenes with CO is a well documented process,24-26 and since the diferracyclobutadiene moiety may very well be a strained species, we had hoped to find incorporation of CO into 3 with release of that strain. In -55- fact, 3, reacts with CO to cleanly produce Cp*2Zr(C0)2, 4, and 2. Compound 4 contains formally divalent zirconium, thus this transformation requires transfer of two electrons back to the zirconium from the 02C2Fe2 unit. The fact that labeled CO is incorporated only into 4 shows that the 02C2 bridge uncouples and remains on the iron atoms as carbonyl ligands. These observations suggest that the formation of 3 from 2 and Cp*2Zr(II) is reversible. In agreement, we find that an equilibrium between 2a, 3a, 2b and 3b is slowly established in benzene solution ( 25oc, > 1 week (Scheme I)). The observation of the mixed complexes Cp'CpFe2(C0)4 and Cp*2Zr(02C2)Fe2CpCp'(C0)2 after longer reaction times may not be due to any role played by the zirconium species. A 1:1 mixture of 2a and 2b scrambles to a statistical mixture of 2a, 2b and Cp'CpFe2(C0)4 within 30 days at 250C in benzene. This scrambling is almost four-fold faster in the presence of ca. 0.1 equivalent of Na( CpFe(C0)2 )·THF. No efforts were made to exclude ambient light from any of these experiments. The equilibrium between 3a and the n-complex Bin Scheme I bears a resemblance to the rapid metalJacyclopentane and bis(olefin) complex interconversions observed in nickel, 27 titanium, 27 and zirconium 14 systems. In the latter, Erwin and Bercaw14 observed the stepwise reaction of ethylene *cP2zrX (8) -56- and 3,4-dlethylzirconacyclopentane, Cp*2ZrCH2CHEtCHEtCH2 to Cp*2ZrCH2CH2CHEtCH2 and subsequently to Cp*2ZrCH2CH2CH2CH2 with release of 1-butene (eq. 8). It is significant that 2a will completely displace the ruthenium dimer in .3c to form 3a, indicating that 3a is the thermodynamic product (eq. 9). The reasons for this preference may be two-fold: (1) the M-M bond in 2c is 3c 2a Cp* Zr(o c )Fe cp (C0) + 2 2 2 2 2 2 3a (CpRu(C0) ) 2 2 ( 9) 2c 0 ca. 0.2 A longer than in 2a or 3a. If this distance persists in 3c it might further destabilize the dimetallacyclobutadiene fragment, and (2) any process which increases electron density at the metals would be less favorable in the case of ruthenium vs. iron. While the reductive coupling of the terminal carbonyls does not formally change the oxidation states of the metals in the dimer, an effective increase in electron density occurs with the loss of some of the n-accepting properties of the carbonyls. This is evidenced in the infrared spectra of 3a and .3c by decreases in the energies of the bridging carbonyl stretches of ca. 30 cm-1 relative to the parent dimers. Both this electronic effect and the longer M- M distance in 2c are in accord with the greater stability of the iron product. There are two likely steps in the exchange of dimers in 3 (Scheme I and eq. 9). First the 02C2 unit must uncouple, and second the dimer must -57- dissociate from the zirconium. The rate of dimer substitution is roughly the same for the exchange of 2b with 3a and 2a with 3c, despite the driving force for the latter. This suggests that uncomplexed permethylzirconocene is produced at comparable rates for 3a and 3c. In contrast to the dimer exchange reactions, 3c reacts with CO to produce 4 and 2c ca. 1o3 times faster than the analogous carbonylation of 3a, suggesting that the rapid carbonylation of 3c (and possif?ly the slower reaction of 3a) does not proceed via trapping of uncomplexed permethylzirconocene by CO. While it does not seem likely that CO would accelerate the rate of the uncoupling step, an associative substitution process on the 7T-complex B is feasible. Based on the data it is not possible to conclude whether the uncoupling step is faster in 3c or if 2a and 2c have significantly different ligating abilities in the 'IT-complex. While the carbonylation may proceed associatively, the steric constraints of the Cp* rings should preclude interaction of a second dimer with B, rendering the dimer exchange process dissociative. The reversibility of the carbonyl coupling process is manifested in other reactions of 3. While 3a decomposes slowly under H2 (4 atm, 25°C), 3c is cleanly converted to the bridging zirconoxy carbene complex Cp2Ru2(C0}2( 11-CO) ( ll-C(H)OZrHCp*2 ),28 7, in <10 min. The reaction of Cp*2ZrH2 with 2c also yields 7. Thus it seems likely that H2 reacts with 3c to yield Cp*2ZrH2 and 2c which then react further to produce 7. However, as in the carbonylation, the short reaction time implies that dissociation of 2c from 3c is promoted by H2. The reaction of the iron complex 3a with CO, H2, and 2b are all relatively slow, requiring ca. one week at 25oc. A notable deviation from -58- this pattern is found in the reaction of 3a with ( CpMo(CO)J) 2, 5. In this instance production of 2a and Cp*2Zr( CpMo(CO)J) 2' 6, in solution occurs within seconds. Compound 6, prepared previously from the reaction of 1 with 5,6 is best described as a Cp*2Zr dication coordinated to the carbonyl oxygens of two CpMo(CO)J anions. Thus the reaction of 1 with .5 results in reduction of the Mo-Mo bond, rather than reductive carbonyl coupling. While it seems feasible that small molecules such as H2 and CO might associate with 3c whereas the dimer 2a is too large, it is difficult to rationalize the extremely rapid reaction of 3a with 5 in terms of an associative substitution process. As an alternative explanation we favor an outer sphere electron transfer from either 3a or the 1T -bound intermediate B to the molybdenum dimer to produce ( CpMo(C0)3) 2- and a zirconium(III) cation. Rapid dissociation of 2a, perhaps assisted by the opposite charge on the molybdenum species, followed by a second electron transfer and collapse would lead to 6 (Scheme IV). Outer sphere transfer of two electrons to yield a zirconium(IV) dication and two CpMo(CO)J anions is also a possibility, although this is less attractive as it requires formation of greater charge separation in a solvent with a relatively low dielectric constant such as benzene. Reaction of Other Metal Carbonyl Complexes with 1. In an attempt to extend this carbonyl coupling process beyond ( CpM(C0)2) 2 complexes, we have studied the reactions of 1 with other metal carbonyl compounds. The complexes M3(COh2 (M =Fe, Ru, Os) were treated with 1 in hydrocarbon solvents under various conditions. In all cases light orange powders were -59- isolated which exhibited a myriad of Cp* resonances in the 1H NMR spectra. Infrared spectra of these materials showed several broad bands clustered around 2000 and 1700 cm-1. No lower energy bands attributable to carbonyl coupled products were observed. H20s3(CO) 10 and ( CpNi(CO)) 2 reacted with 1 similarly. None of these reaction products were pursued further. The observation was made, however, that when excess 1 was used in preparing 3, yields and purity were diminished. When Ja is treated with 1 in toluene at 25°C, dark crystals of { Cp*2Zr} 204C4 { CpFe} 2, lOa, can be isolated in moderate yield from the brown solution. These crystals dissolve to give red-brown solutions which exhibit broad singlets at o 5.68 (1 OH) and 1.39 (60H) in the 1H NMR spectrum. Three significant bands are observed in the infrared spectrum: 1329(s), 1250(m), 1034(s) cm-1 (13co: 1300, 1235, 1/2{Cp* ZrN } N 2 2 2 2 + Ja,c 1 ( 1 0) + 10a,b and 1010 cm-1). No higher energy v(CO) bands are observed. The extremely broad 1H NMR resonances ( Cp: v(fwhm) = 33Hz, Cp*: v(fwhm) = 6 Hz) found for lOa suggest that this complex is paramagnetic. This is confirmed by susceptibililty measurements indicating lJeff = 3.26 B.M. per Zr2Fe2 unit at 280°K ( (S = 1, lJ (spin only) = 2.83 B.M.). The susceptibility and low temperature antiferromagnetic behavior29 are not -60- readily interpretable without further structural information for lOa. Green crystals of { Cp*2Zr} 204C4{ RuCp} 2, lOb, can be isolated in 30-4096 yield from the reaction of 1 with 3c. The 1H NMR spectrum of lOb exhibits singlets at 0 4.47 (lOH) and 2.06 (60H). In this case, however, both resonances are quite sharp ( v(fwhm) < 1 Hz) and susceptibility measurements show that the complex is diamagnetic. The infrared spectrum of lOb is virtually identical with lOa, with three bands at 1330(s), 1258(m), and 1020(s) em -1. Single crystals of 1Ob have been grown, but they do not appear to diffract in an X-ray beam. Attempts to grow crystals of lOa and lOb suitable for a structure determination are proceeding. In lieu of better structural information, it is necessary to speculate on the basis of spectral data. The singlets observed in the 1H NMR spectra of 10 for the Cp* protons broaden somewhat at lower temperatures, but no splitting is observed down to -1oooc. This indicates that either the four Cp* rings have equivalent environments, or that a very facile fluxional process is occurring. The same is true for the Cp protons. From the absence of any bands above 1300 em -1 in the infrared spectra it can be concluded that the complexes contain no typical terminal or bridging carbonyl ligands. Three reasonable structures (Fig. 4) can be drawn for complexes derived from 3 via a second carbonyl coupling step. Structure I is analogous to 3, except that both pairs of carbonyl ligands have been coupled to form two planar dimetallacyclobutadiene rings. In structure ll the 02C2 moieties are turned 900 to yield tetrahedral bridges similar to the arrangement often found for acetylenes bound to metal dimers. Structure m contains one of each 02C2 bridge geometry, and phenomenologically is the result of coupling the two bridge carbonyls in 3 with ••••••••••••••• •••••••••• • • • • - 1: (1) - ~ ~ - - CD I 0'\ 3: - I 0.0 0 50 100 150 Temperature 200 (•K) Figure 3. Plot of lJ(eff) vs. T for {Cp*2Zr }04C4 { CpFe} 2, lOa. 250 300 II ...... II >< II 3 ,..... II CD 8 3 c: II ! II 3 II 0 - 10 CD II . /111111 0 !1!111 a" II II II II I 0'\ II N I !JII!I a· all II 5 100 150 200 Temperature (°K) Figure 4. Plot of 1/X vs. T for { Cp*2Zr} 04C4 { CpFe} 2, lOa. I 250 I 300 -63- no other major rearrangement. This structure would be expected to have inequivalent pairs of Cp* ligands, although the bridges might rapidly interconvert. Alternatively, the chemical shift difference might be too small to detect, although this seems unlikely. While the reactions of 10 have not been fully explored, preliminary results indicate significant differences from 3. In contrast to the substitution reactions found for 3, 10 reacts rapidly under CO to produce dark insoluble solids. Only a small amount of 4 and 2 are observed in solution. The insoluble product mixture from the carbonylation of lOa exhibits bands at 1725(s), 1212(m), 1195(s), 1156(m), 1130(s), and 1092(2) cm-1. As yet we have not determined if this is a single product, or even the stoichiometry of the reaction. This could be a productive area for further study. @7 * 0 c~ I M @7 .cp* ~Zr \Zr...,./o:(I)Ko, I ·,. ..· · . ./ Zr o/ ~~-cp* cp1: ~ c( '/XIX', 'o :-...M 0 ~ ~ ~Zr ~o Cp* c~ O" \ M ~ Cp* II I 0'\ ~ @7 c~* o Zr ,~~ 'o Cp* I I I c* P M '/ J:IJ<o'z / ~ M , ~ r if \ Cp* Ill Figure .5. Three possible structures for {Cp*2Zr l04C4 { CpM }2, 10. -65- Experimental Section General Considerations. AJl manipulations were performed using either glove box or high vacuum line techniques. Hydrocarbon solvents were purified by vacuum transfer first from LiAlH4 and then from "titanocene".31 NMR solvents, benzene-d6 and toluene-d8 were also purified by transfer - - from "titanocene". Hydrogen was passed through MnO on vermiculite and 0 then over 4 A molecular sieves.32 Carbon monoxide (MCB) was used directly from the cylinder. { Cp*2ZrN2} 2N2,33 ( CpFe(C0)2) 2,34 ( CpRu(C0)2) 2,35,36 Cp2Ru2(C0)3(CH2),28 ( CpMo(C0)3) 2,37 and ( (C5H4Me)Fe(C0)2) 234 were prepared as described previously. Infrared spectra were measured on a Beckman 4240 spectrophotometer. 1H NMR spectra were obtained using Varian EM-390 or JEOL FX-90Q spectrometers. 13c NMR spectra were obtained using the JEOL FX-90Q instrument. Samples for 13c NMR experiments were generaJly prepared with 13co enriched ( CpRu(C0)2 )238 and ( CpFe(C0)2 )2 and ca. 0.1 Madded Cr(acac)311 to facilitate observation of the carbonyl carbons. Elemental analyses were performed by Bernhardt Microanalytical Laboratory, Dornis and Kolbe Microanalytical Laboratory, or by Larry Henling at the CIT analytical service. and 2a (250 mg, 0.707 mmol) in 50 mL toluene was stirred under vacuum for 2 h at 25oc. Toepler pump analysis of the gases generated showed 1.04 mmol (98.2%) of non-combustible gas, presumed to be dinitrogen. Recrystallization of the residue from toluene/petroleum ether yielded 265 mg (53%) dark brown crystals. -66- Anal. calcd. for C34H4oFe204Zr: C, 57.07; H, 5.64; Zr, 12.75. Found: C, 56.94; H, 5.71; Zr, 12.54. Cp*2Zr(<>2C2)Ru2CpiC0)2, k. A solution of 1 (264 mg, 0.33 mmol) and 2c (255 mg, 0.57 mmol) in 40 mL benzene was stirred for 1 h at 250C. The resulting metallic orange solid was collected by filtration and washed once with 3 mL benzene, yielding 144 mg (31 %) of solid. The orange-black residue was recrystallized from toluene/petroleum ether to yield a second crop of 180 mg (39%). 1H NMR indicates these products contain ca. 0.5 moles of benzene or toluene as solvent of crystallization, which can be removed by stirring a slurry 3c in petroleum ether overnight, filtering and drying in vacuo. Anal. calcd. for C34H4o04Ru2Zr: C, 50.66; H, 5.00; Zr, 11.32. Found: C, 50.92; H, 5.20; Zr, 10.80. Cp*2ZrRu2CpiCO)J(CH2), 9. A slurry of 1 (205 mg, 0.254 mmol) and 8 (209 mg, 0.486 mmol) in 10 mL toluene was stirred 2 hat 25°C, cooled to -80°C and filtered. The solid was washed with 5 mL petroleum ether and dried in vacuo, yielding 287 mg (75%) of red-orange solid. Anal caJcd. for C34H4203Ru2Zr: C, 51.56; H, 5.35. Found: C, 52.05; H, 5.48. Cp*2Zrl2. A slurry of Cp*2ZrCl2 (2.3 g, 5.3 mmol) and BI3 (1.4 g, 3.6 mmol) in 20 mL toluene was stirred for 4 hat 25oc. The solvent was then removed and the yellow solid dried in vacuo for 1 h. Recrystallization from diethyl ether yielded 2.73 g (88%) of pure, crystalline product. Anal. calcd. for C2oH3oi2Zr: C, 39.03; H, 4.91. Found: C, 39.27; H, 4.85. -67- Reaction of Cp*2Zri2 with Na( CpFe(C0)2)·THF. Cp*2Zrl2 (32 mg, 0.052 mmol), Na( CpFe(C0)2) ·THF (30 mg, 0.110 mmol) and 0.3 mL benzene~6 were added to an NMR tube and the tube sealed with a torch. Formation of .3a over a period of 2 h was accompanied by a color change from bright yellow to red-brown and formation of precipitate. Reaction of 3c with 13co. Compound 3c ( 18 mg, 0.022 mmol) and benzene-~6 (0.3 mL) were ad~ed to an NMR tube, 13co (120 torr, 770K, ca. 2 mL) introduced and the tube sealed. A lH NMR spectrum observed after 10 min showed only Cp*2Zr(C0)2 and 2c. A 13c NMR spectrum showed >90% enrichment in the carbonyl carbons of Cp*2Zr(C0)2 (o 275). Reaction of 3c with H2. Compound 3c (1 0 mg, 0.012 mmol) and benzene-~6 (0.3 mL) were added to an NMR tube, H2 (700 torr, 77°K, ca. 2 .mL) introduced and the tube sealed. A lH NMR spectrum observed after 10 min showed complete conversion to Cp2Ru2(C0)3 ( C(H)OZrHCp*2), as confirmed by comparison with a spectrum of an authentic sample. Reaction of 3a with ( CpMo(C0)3)2· Compound 3a (15 mg, 0.021 mmol) and ( CpMo(C0)3) (10 mg, 0.020 mmol) were added to an NMR tube. Upon addition of benzene-~6 (0.3 mL) a rapid reaction ensued with a color change from brown to dark red. The tube was sealed with a latex septum. The 1H NMR spectrum showed conversion to 2a and Cp*2Zr ( CpMo(C0)3) 2' as confirmed by comparison with spectra of the authentic compounds. Reaction of 3c with 2a. Compound 3c (14 mg, 0.017 mmol), 2a (6 mg, 0.017 mmol) and benzene-~6 were sealed in an NMR tube. After ten days the 1H NMR spectrum showed conversion to 3a and 2c. Reaction of 3a with 2b. Compound .3a (10 mg, 0.014 mmol), 2b (5 mg, -68- 0.013 mmol) and benzene-£6 were sealed in an NMR tube. After one week the 1H NMR spectrum showed a mixture of Ja, Jb, 2a, and 2b. After two weeks additional signals for Cp*2Zr(02C2)Fe2(C0)2CpCp' and CpCp'Fe(C0)4 were observed. { Cp*2Zr} 2<)4C4{ CpFe} 2, lOa. A solution of 1 (490 mg, 0.607 mmol) and 2a (210 mg, 0.587 mmol) in 40 mL toluene was stirred for 2 hat 25oc, filtered and the solvent volume reduced to 10 mL. The solution was stirred at -800C for 1 h, filtered and the residue washed once with 2 mL toluene, once with petroleum ether and then dried in vacuo for 3 h. 286 mg (4596) of black crystals were collected. Proton NMR analysis indicated the product is solvated with toluene (1: 1). Analysis of the filtrate showed a ca. 1:1 mixture of 3a and 12a. Additional 1 (100 mg, 0.124 mmol) was dissolved with the filtrate in 30 mL toluene and the solution was stirred for 1 h. Removal of solvent followed by addition of 10 mL petroleum ether yielded a red solution with brown-black solid. The solid as collected on a frit and washed three times with 10 mL petroleum ether, yielding 193 mg (3196) of product. Anal. calcd. for C54H7oFe204Zr2·C7H8: C, 62.65; H, 6.72. Found: C, 62.35; H, 6.52. {CP*2Zr}2<)4C4{CpRu}2' lOb. A slurry of 1 (570 mg, 0.644 mmol) and 2c (230 mg, 0.518 mmol) in 10 mL toluene was stirred at 250C for 5 h. Removal of solvent and addition of petroleum ether followed by filtration, yield gray crude product. Recrystallization from toluene yielded 245 mg (4196) of green solid. Proton NMR analysis indicated somewhat less than 1 toluene of crystallization. Anal. calcd. for C54H7004Ru2Zr2·C7H8: C, 58.15; H, 6.24. Found: -69- C, 57 .33; H, 6.1 0. Magnetic Susceptibility Measurements. Variable temperature magnetic susceptibility measurements were made at the University of Southern California SQUID-based S.H.E. Corp. magnetic susceptometer. A diamagnetic correction for lOa of Xd = -1.21 x 1o- 3 was applied. This value was calculated from the measured diamagnetic susceptibility of the ruthenium analog lOb (Xcf = -1.17 x 1Q-3), corrected for Ru-Fe differences found for 2a and 2c. The susceptibility of lOa was also measured at 2980K (2.98 B.M.) using the Evans method.39 -70- References and Notes (1) Wolczanski, P. T.; Bercaw, J. E. Accts. Chern. Res. 1980, _!l, 121127. (2) (a) Wolczanski, P. T.; Threlkel, R. S.; Bercaw, J. E. J. Am. Chern. Soc. 1979, 101, 218-220. (b) Barger, P. T.; Santarsiero, B. D.; Armantrout, J.; Bercaw, J. E. Ibid. in press. (c) Barger, P. T.; Bercaw, J. E. Organometallics in press. (d) Threlkel, R. S.; Bercaw, J. E. J. Am. Chern. Soc. 1981, 103, 2650-2659. (3) Barger, P. T.; Bercaw, J. E. J. Organomet. Chern. 1980, 201, C39C44. (4) Berry, D. H.; Bercaw, J. E.; Jircitano, A. J.; Mertes, K. B. J. Am. Chern. Soc. 1982, 104, 4712-4715. (5) The crystal structure of 3a was carried out by Kristin Mertes at the University of Kansas, Lawrence. 4 (6) Chapter I of this thesis. (7) Sutton, L. E., Ed. "Tables of Interatomic Distances and Configuration in Molecules and Ions"; The Chemical Society: London, 1958, S17. (8) Sekutowski, D. G.; Stuckey, G. D. lnorg. Chern. 197.5, ~' 2192-2199. (9) Fischer, E. 0. Adv. Organomet. Chern. 1976, ~' 1. (1 0) Bryan, R. F.; Green, P. T .; Newland, N. J .; Field, D. S. J. Chern. Soc. A 1970, 3068. (11) Gansow, 0. A.; Burke, A. R.; Vernon, W. D. J. Am. Chern. Soc. 1972, 94, 2550-2552. (12) The terminal and bridging carbonyl 13c NMR resonances of 2c at -1360C are found at 0 199 and 253.13 -71- (13) Gansow, 0. A.; Burke, A. R.; Vernon, W. D. J. Am. Chern. Soc. 1976, 98, 5817-5825. (14) Erwin, D. K. Ph.D. Thesis, California Institute of Technology, Pasadena, California, 1979. (15) (a) Longato, B.; Norton, J. R.; Huffman, J. C.; Marsella, J. A.,; Caul ton, K. G. J. Am. Chern. Soc. 1981, 103, 209. (b) Pasynskii, A. A.; Skripkin, Y. V.; Suvorova, K. M.; Kuzmichea, 0. N.; Kalinnikov Izv. Akad. SSSR, Ser Khim. 1978, 1226. (16) Dyke, A. F.; Knox, S. A. R.; Mead, R. A.; Woodward, P. J. Chern. Soc., Chern. Commun. 1981, 861-862. (17) Mills, 0. S.; Nice, J. P. J. Organomet. Chern. 1967, .2_, 339-344. (18) Cis and trans 8: v(CO) = 1980(s), 1943 (s), 1924(m), 1789(w), 1751(s). Cis and trans 2c: v(CO) = 20 19(m), 2009(m), 1972(vs), 1964(s), 1942(vs), 1792(s). (19) (a) Lam, C. T .; Corfield, P. W. R.; Lippard, S. J. J. Am. Chern. Soc. 1977, 99, 617-618. (b) Giandomenico, C. M.; Lam, C. T.; Lippard, S. J. Ibid. 1982, 104, 1263. (20) Cotton, F. A.; Roth, W. J. J. Am. Chern. Soc. 1983, 105, 3734-3735. (21) Lukehart, C. N.; Srinivasan, K. J. Am. Chern. Soc. 1981, 103, 41664170. (22) Cramer, R. E.; Higa, K. T .; Pruskin, S. L.; Gilje, J. W. J. Am. Chern. Soc. 1983, 105, 6749-6750. (23) Casey, C.P. Private communication. (24) Herrmann, W. A.; Planke, J. Angew. Chern. Int. Ed., Engl. 1978, .!Z., 525-526. -72- (25) Bodnar, T. W.; Cutler, A. R. J. Am. Chern. Soc. 1983, 105, 5926-5928. (26) Kreiss!, F. R.; Uedelhoven, W.; Eber, K. Angew. Chern. Int. Ed., Engl. 1978, !Z, 859-860. (27) Grubbs, R. H.; Miyashita, A. J. Am. Chern. Soc. 1978, 100, 1300. (28) Chapteriii of this thesis. (29) Carried out at the University of Southern California SQUID-based S.H.E. Corp. magnetic susceptometer. (30) Hoffman, D. M.; Hoffman, R.; Fisel, C. R. J. Am. Chern. Soc. 1982, 104, 3858-387 5. (31) Marvich, R. H.; Brintzinger, H. H. J. Am. Chern. Soc. 1971, 94, 2046. (32) Brown, T. L.; Dickerhoof, D. W.; Bafus, D. A.; Morgan, G. L. Rev. Sci. Instrum. 1962, 33, 491. (33) Manriquez, J. M.; McAlister, D. R.; Sanner, R. D.; Bercaw, J. E. J. Am. Chern. Soc. 1978, 100, 2716. (34) Angelici, R. "Synthesis and Techniques in Inorganic Chemistry"; W. B. Saunders Co.: Philadelphia, 1977, p. 151. (35) Humphries, A. P .; Knox, S. A. R. J. Chern. Soc., Dalton 197.5, 17101714. (36) Gibson, D. H.; Hsu, W.; Steinmetz, A. L. J. Organomet. Chern. 1981, 208, 89-102. (37) BirdwhisteJJ, R.; Hackett, P.; Manning, A. R. J. Organomet. Chern. 1978, 157, 239. (38) Roberts, J. D.; Rosenberg, E.; Harris, D. C. J. Chern. Soc., Dalton 1974, 2398-2403. (39) Evans, D. F. J. Chern. Soc. 19.59, 2003-2005. -73- CHAPTERID Reversible Reduction of Coordinated Carbon Monoxide by Permethylzirconocene Hydride Complexes. Characterization and Reactivity of a Bridging Diruthenium Zirconoxy Carbene Complexes. -74- Introduction Our interest in complexes containing the permethylzirconocene moiety and later transition metals has led us to several new structural and reactivity modes for carbonyl complexes. In the course of investigating the reaction chemistry of Cp*2Zr(02C2)Cp2Ru2(C0)2 ( Cp*:n 5-C5(CH3)5) , a complex in which the terminal carbonyl ligands have been reductively coupled, an interesting reaction was observed.la Reaction of the ruthenium complex with hydrogen quantitatively yielded a new complex in which a carbonyl ligand has been reduced by addition of a zirconium hydride. Previous studies have shown that the carbonyl coupled products are in equilibrium with zirconocene and simple dinuclear carbonyl complexes.! Combined with the fact that permethylzirconocene rapidly reacts with hydrogen to form a dihydride, this suggested to us that a study of the reactions of zirconium hydrides with ( CpRu(C0)2) 2 was in order. Following the discovery in these laboratories that (C5Me5)2ZrH2, 1, reacts with CO to yield a variety of products, studies were undertaken to elucidate the mechanism of what was at the time the only homogeneous reduction of CO by a transition metal hydride complex. These studies have been summarized elsewhere,2 but a key finding was the fact that 1 will react with terminal carbonyl ligands to produce "zirconoxy" carbenes (eq. 1). Zirconoxy carbenes have been observed in the reaction of zirconium ( 1) -75- hydride complexes with Cp2M(CO) (M =Cr, Mo, W),3 Cp2NbR(CO) (R =H, alkyl, phenyl),3 Cp2Zr(PMeJ)(Co)4 and CpM(C0)2 (M =Co, Rh).5 Reported in this chapter are the synthesis and characterization of zirconoxy carbenes which bridge two ruthenium atoms. This is the first instance of a zirconoxy carbene which occupies a bridging position. The addition of the zirconium hydride across the bound carbonyl is shown to be reversible. Furthermore, the fluxionality of the diruthenium fragment is proposed to occur through an intermediate in which the carbene occupies a terminal position. Lewis acids are shown to induce C-0 bond cleavage in the carbene, with subsequent transfer of a second zirconium hydride and production of a methylene bridge. These studies suggested the conditions in which LiHBEt3 can be used to effect the transformation of a bridging carbonyl to a bridging methylene ligand directly. Results Cp*2ZrH2, 1, reacts with ( CpRu(C0)2) 2, 2, rapidly at 25°C in hydrocarbon solvents to yield trans-Cp2Ru2(C0)2(ll-CO)( ll-C(H)OZr(H)Cp*2), 3. This compound is fluxional and exhibits temperature dependent 1H NMR + 2 Cp Ru (C0) (C(H)OZrHCp* ) 2 2 2 3 3 ( 2) oc ~p ~ 'Ru" ~ ~ ~ c~ '\_ ,,,,.H H-.,,,,C/ C=O /z''o~ ". / Cp* Ru ~~ ~ Cp 0 ~c * I "'J ~ I * * * ~I I I 13 12 I I r I 1 I s I s I I I I 4 3 2 1 Figure l. 500 MHz lH NMR spectrum of Cp2Ru2(C0)3 [C(H)OZrHCp*2], 3, at 172 K, toluenedg solvent. Asterisks denote resonances due to solvent. I o Table 1. NM~ and IR~ Data. Compound IR Chemical Shift ZrOC(H) 12.75 (s, 1H) V(Co): 1929(s), 177l(s), Zr-H 6.10 (s, 1H) 1108(s) C5!!5 5.04 (s, 1OH) C.5(C!:!3)5 2.01 (s, 30H) ZrOC(H) 12.67 (s, 1H) Zr-H 5.94 (s, 1H) C5!!5 4.98 4.85 (s, 5H) (s, 5H) C5(CH3)5 2.04 1. 95 (s, 15H) (s, 15H) ll-CO 248 Ru-CO 203 201 ZrO-C(H) 192 fj(CH3)5 118 117 1H NMR£ Cp2Ru2(C0)3 ( C(H)OZrHCp*2) 3 Assignment 1H NMR2. I ........ ........ I 13c{ 1H}~ (JcH = 153) Table I. Continued. Compound IR Assignment Chemical Shift ~H5 Cp2Ru2(C0)3( C(H)ZrC1Cp*2 )~ 1H NMR~ .5 95 92 C5(CH3)5 12 ZrOC(H) 12.60 (s, 1H) C5H5 5.14 (s, 10H) I 1H NMR£J Cp2Ru2(C0)3 ( C(H)0Zr(OH)Cp*2) 6 v(Co): 1925(s), 1770(s), 1HNM~ C5(C!:!3)5 1.94 (s, 30H) ZrOC(H) 12.73 12.59 (s, 1H)b. (s, 1H)& C5!:!5 5.37 5.28 (s, 5H)& ~s, 5H~b. s, 5H g 4.86 4.79 (s, 5H)b. C5(CH3)5 1.85 (s, broad)&!! ZrOC(H) 12.65 (s, 1H) C5H5 5.15 (s, 10H) '-I 00 I Table 1. Continued. Compound IR Assignment 1100(s,br) ZrOH 4. 25 (s, 1H) Cj(C!:!J)5 1 • 86 (s, 30H) V(OH): 367 5(w) ~Chemical shifts reported in Chemical Shift orelative to internal TMS. ~Spectra obtained as nujol mulls. Values reported in cm-1. £Toluene-~g solvent, ambient temperature. ~Toluene-~g solvent, -7ooc. ~Compound .5 was not isolated. !At -70°C .5 exists as a mixture of two isomers in a 3:1 ratio, denoted by g and !!· &Major isomer. !!Minor isomer. I """ \.0 I -80- spectra. In the low temperature limit (Fig. 1) resonances are found at 12.67 and 5.90 ppm for the carbene and zirconium hydride protons, with two resonances each for inequivalent Cp and Cp* ligands. These signals coalesce into single Cp and Cp* resonances upon warming. Based upon a coalescence temperature of -40C for the Cp at 90MHz, ~G ~ for ring equilibration is 14.0(2) kcal·mol-1. In the hafnium analog, 4, synthesized from 2 and Cp*2HfH2, ~G ~ is 14.4(2) kcal·mol-1 at .5 oc. The infrared spectrum of 3 (nujol mull) shows strong bands at 1929, 1771, and 1108 cm-1. The 13c NMR spectrum observed at -7ooc shows carbonyl resonances at 201, 203 and 248 ppm. The latter is consistent with a bridging carbonyl and the former two with terminal ligands. In the 1H coupled spectru, the carbene carbon appears as a doublet at 192 ppm (JcH = 153 Hz) • .In addition, pairs of Cp and Cp* resonances are observed. Upon warming the three carbonyl signals coalesce, disappearing into the baseline. While the carbene signal broadens significantly, then sharpens above room temperature, the C-H coupling constant is virtually the same at 50°C (JcH = 154Hz) as it is at -700C. Cp*2ZrHCl reacts with 2 in about one day at 25°C to form 5, the chloride analog of 3. Compound 5 also exhibits a temperature dependent 1H NMR, with the Cp appearing as a very broad signal at 25oc, separating into two signals below 1ooc. Upon further cooling each Cp resonance separates into two peaks in roughly 1:3 ratios. Similarly, the carbene proton resonance separates into peaks at 12.75 and 12.59 (ca. 3:1). Threlkel reported a ·s imilar observation in the low temperature 13c NMR spectrum of Cp2Nb(H) (C(H)OZrHCp*2) and ascribed it to two different rotomers about -81- the C-O bond of the zirconoxy carbene.6 A second equivalent of 1 can attack 3 to give what appears to be Cp2Ru2(C0)2( ll-C(H)OZrHCp*2) 2· This compound is insoluble in hydrocarbon solvents, but exhibits bands in the infrared at 1913 and 1058 cm-1, consistent with a trans structure with two bridging carbenes and two terminal carbonyls. No further characterization has been attempted. As a result of this side reaction, preparation of 3 is best carried out with a slow addition of 1 to a solution of 2. This is most conveniently accomplished by the action of hydrogen on the carbonyl coupled product Cp*2Zr(02C2)Cp2Ru2(C0)2.l This produces 3 cleanly with no excess 1 present. Compound 3 decomposes in benzene solution at 250C over a period of weeks to many unidentified products. This process takes hours at 8QOC. In the presence of 4 atm of H2 at 25oc neither the rate nor products of this decomposition are significantly changed. With D2 under the same conditions, Cp Ru (C0) (C(H)OZrHCp* ) 2 2 3 2 3 Cp Ru (C0) (C(D)OZrDCp* ) 2 2 2 3 (3) ~2-3 however, both the hydride and carbene positions are deuterated (t~ -3 h). Compound 3 reacts under 4 atm of carbon monoxide at 25°C in 14 hours to co Cp Ru (C0) (C(H)OZrHCp* ) 2 2 3 2 3 Cp* Zr(C0) 2 2 + (CpRu(C0) 2 ) 2 2 + H 2 (4) -82- yield Cp*2Zr(C0)2, 2 and presumably H2. Reaction with CH3I results in rapid formation of methane and the iodide analog of 3, which then reacts further with CH3l to produce Cp*2Zrl2 and 2 over a period of several days. Compound 3 reacts with H20 in benzene to yield H2 and the hydroxy complex Cp2Ru2(C0)2(ll-CO)( ll-C(H)OZr(OH)Cp*2), 6. The 1H NMR spectrum of 6 at 250C shows resonances at 12.64 and 4.35 ppm attributable to carbene and hydroxide protons, as well as singlets for the Cp and Cp* ligands. The 0-H stretch is observed at 3675 cm-1. Compound 6 is considerably more thermally stable than 3, with complete decomposition requiring heating at 55°C for 2.5 weeks in benzene. While the 1H NMR spectrum of this pyrolysis shows a myriad of Cp* signals, greater than 95% of the ruthenium is present as 2 and CpRu(C0)2H in a 1:1.25 molar ratio. While protonation of 3 with CF3COOH results in formation of reaction of CF3COOH with the latter species is nearly as fast as with 3. Compound 3 reacts with HBF4·Me20 in benzene at 25°C to rapidly yield the known bridging methylene complex Cp2Ru2(C0)3(CH2), 7, as a Cp Ru (C0) (C(H)OZrHCp* ) 2 2 3 2 3 • + HBF •Me o 2 4 Cp Ru (C0) (CH ) 2 2 3 2 7 + mixture of cis and trans isomers. A single Cp* containing product, as yet unidentified, is also obtained. Using 2.2-3, this reaction produces 9_2-7 (5) -83- initially. Compound 7 is also formed from 3 and BF3·Et20 in benzene. Compound 7 is the major product in the reaction of 2 and LiHBEt3 in (CpRu(C0) ) 2 2 + LiHBEt toluene 3 2 Cp Ru (C0) (01 ) 2 2 2 3 7 (6) toluene, and can be isolated in 60% yield. No 7 is formed if this reaction is run in THF instead of toluene. The known iron analog of 7 could not be synthesized in this manner. Discussion Structure and Fluxional Behavior of Cp2RuiCO)ip.-COXp.C(H)OZr(H)Cp*2). On the basis of NMR data, 3 can be shown to have the trans structure pictured below. Low temperature 13c NMR data (Table 1) are particularly indicative in this determination: (1) Assignment of the zirconoxy trans-3 -84- carbene to the bridging position is supported by the presence of resonances for two terminal carbonyls (20 1 and 203 ppm) and only a single bridging carbonyl (248 ppm).7 (2) The trans disposition of the Cp rings and terminal carbonyls about the Ru-Ru bond is mandated by the low temperature NMR data also. In the cis form the two Cp rings would be equivalent, appearing as a single resonance. In fact, two singlets are observed in both the 13c and 1H NMR spectra at low temperatures, as expected for the trans arrangement. The chemical shifts for the zirconium hydride and carbene proton signals are near or within the characteristic ranges.3 The C-H coupling constant of 153 Hz for the carbene is somewhat larger than the 136 Hz reported by Threlkel for a terminal niobium zirconoxy carbene,3 implying more p character in the C-H bond of the latter. That 3 exhibits fluxional behavior is not surprising in light of similar observations for bridging alkylidenes of iron,8 ruthenium,9 and cobalt.IO In a logical extension of the exhaustive studies of Cotton,11 Roberts, 12, Gansow13 and coworkers, Knox has suggested the bridged terminal equilibrium shown in Scheme I to explain the cis-trans isomerization of a bridging isopropylidene complex.9 While it is reasonable that 3 is undergoing such a process, one distinction must be made. In Knox's example both cis and trans isomers are observed in equilibrium, while cis-3 has never been detected. The process under consideration is the equilibration of the inequivalent Cp rings, in which two bridge openings and rotations are required (Scheme II). Alternatively, rotation about the M-C double bond of the terminal carbene followed by closure would result in Cp equilibration with only a single opening step (Scheme III). Such rotations in mono-Cp carbene -85- Schemel /- 0.. (.) -86- Scheme n ~ts 00 .,,,,,,,, ' =- -oo a: *N a. u ::r: ~ N 0 II -87- complexes appear to be facile. Fischer and coworkers report that while in the solid state structure of CpMn(C0)2(=CMe2) the isopropylidene ligand is in the "vertical" configuration with inequivalent methyl groups, only single resonances are observed in the 1H and 13c NMR spectra at -55oc)4,15 This implies rapid rotation around the Mn-C double bond, although there is insufficient information to quantify the barrier. Rapid rotation at -550C is in agreement with theoreti~al work published by Fenske and Kostic for which the two-fold alkylidene rotation barrier in CpMn(C0)2(=CMe2) was calculated to be ca. 9 kcal•mol-1.16 In the case of the purely carbonyl system 2, Gansow has shown that pairwise carbonyl bridge opening is rate limiting with ~Gf = 7.6 kcal•mol-1.13 In the same study the rotational barrier around the Ru-Ru bond was estimated at less than 5 kcal·mol-1. Knox reports an activation energy of ca. 20 kcal·mol-1 at 1oooc for cis-trans isomerization.9 In addition, Bergman and Theopold report activation energies of ca. 17 kcal·mol-1 for the cis-trans isomerizations of Cp2Co2(C0)2( ll-C(H)(neo-C5H11> and Cp2CoRh(C0)2(ll-CH2).10,17 The lower values at ca. ooc of 14.0 for 3 and 14.4 kcal·mol-1 for the hafnium analog 4 may in part be due to stabilization of the terminal carbene intermediate relative to the bridging carbene by the zirconoxy substituent. Such stabilization would lower the bridged-terminal carbene exchange barrier in both Schemes II and III. In addition, the barrier to carbene rotation should also be lower in the case of a zirconoxy carbene, due to the decrease in M-C bond order associated with resonance forms available to heteroatom carbenes. A third mechanism for Cp equilibration in 3 is shown in Scheme IV. A hydrogen shift from the carbene back to the zirconium produces a symmetrical -88- Scheme Ill *N Q. u J: ~ N 0 fl a::: -89- intermediate. Transfer of the other hydride to the carbonyl yields the carbene in which the environments of the Cp rings have been exchanged. Alternatively, the hydrogen shift might produce free 1 and 2 which then recombine with the same net randomization of Cp rings. The reaction chemistry of 3 provides ample evidence for the reversibility of the addition of hydride across the carbonyl C-0 bond. For example, reaction of 3 with CO to form 2 and Cp*2Zr(C0)2 is most easily explained by the presence of a small equilibrium concentration of 1 which under CO reductively eliminates H2 and forms the carbonyl complex.l8 Wolczanski reported a similar transformation of a tungsten zirconoxy carbene, albeit under considerably more forcing conditions.3 Deuterium exchange with the hydride positions in 1 has been shown to be remarkably facile even at -8ooc.l9 While exchange with the mono-hydride in 3 would not be unprecedented, deuteration of the carbene position reasonably requires some sort of hydride-carbene interchange. Similarly, reaction of the iodide and trifluoroacetate derivatives of 3 with Mel and CF3COOH to form 2 and Cp*2Zrl2 and Cp*2Zr(OOCCFJ)2 most likely proceeds through intermediate Cp*2ZrHX (X = I, CFJCOO) generated by a hydrogen shift from carbon to zirconium. The faster rate of reaction for addition of a second equivalent of CF3COOH implies that production of Cp*2ZrHX (X = CFJCOO) is considerably faster than when X is OH or I. This might be due to destabilization of the trifluoroacetate derivative of 3 by the steric bulk of the acetate group. The decomposition of 6 to 2 and CpRu(C0)2H is another case in point. A hydrogen shift from the carbene back to the zirconium would produce 2 and OC Cp ~ ~ ~ ~ ' ~=0 ~ ~ ~ Cp* Ru 3 ~ < '\_ ,,,,.H H-.,,,,C/ /zr,0 ~ ""- / * Ru 1\ Cp c Cp 0 ~ 0 c Cp - * '\_ ,,.H / ' CP ~ ~ ~ ~ Ru' C=O /Zr~:O=C" / Cp* 0 C Cp ', z ,,,,,,,..0...,,,, / 'Ru/ .,. ' Cp* r, cl/ '''C H H,- " "'\. /C=O R~ I Cp ~c ~ ~ 0 / ~ I \0 0 1\ Ru Cp C 0 I -91- Cp*2Zr(H)(OH).20 Under pyrolysis conditions the latter could decompose in any number of ways to produce H2, which could react with 2 to form the observed hydride. An independent experiment confirms that 2 will react with H2 to yield CpRu(C0)2H· Cp Ru (C0) (C(H)OZr(OH)Cp* ) 2 2 2 3 6 + ( 7) + Cp* zr (H) (OH) 2 (CpRu(C0) ) 2 2 H 2 + + ( 8) ( Zr) 2 CpRu(C0) H 2 (9) 2 While the chemical reactivity of 3 appears to ensure that the process in Scheme IV is occurring, the dominant mechanism for non-rigid behavior in 3 most likely involves bridge-terminal exchange (Schemes II and III). From 1H coalescence data the rate for Cp ring equilibration at -4oc is 26 sec-1. However, the failure to observe magnetization transfer at 280C between the carbene proton and the zirconium hydride positions sets an upper limit of ca. 1 sec-1 for the mechanism in Scheme IV. The hydrogen shift mechanism, therefore, is not predominant. Furthermore, the undiminished C-H coupling constant of the carbene carbon at high temperatures is consistent with this proposal. It is not possible, however, to distinguish between the mechanisms in Schemes II and III on the basis of these data. -92- Reaction of 3 with Lewis Acids. While protonation with an acid such as water or CF3COOH with 3 results in reaction at the hydride position and formation of hydrogen and substituted 3, stronger acids attack the oxygen of the zirconoxy carbene. This is similar to the action of Lewis acids on the iron carbene Cp2Fe2(C0)3( C(H)(OCH3)) which yield ( Cp2Fe2(C0)3(CH)) + (eq. 10).8b Both Casey and Pettit have shown that this bridging methyJidyne ( 1 0) H ( 11) is attacked by hydrides to produce Cp2Fe2(C0)3(CH2) (eq. 11).8 The presence of an internal hydride source in 3 suggests that a reaction sequence similar to equations 10 and 11 occurs in the reaction of 3 with HBF4·Me20 or BF3·Et20 (Scheme V). The formation of Q2-7 from HBF 4·Me20 and Q2-3 confirms that the second methylene proton is transferred from the zirconium. Cp Ru (C0) (C(D)OZrDCp* ) 2 2 2 3 9.2-3 + HBF •Me o 4 2 Cp Ru (C0) (CD ) 2 2 2 3 9.2-7 + ( 12) - 93- Scheme v + ~~ ~ '0 m ~~ N *Q./ ' *Q. u u -94- The synthesis of 7 directly from 2 is an attractive alternative to the published procedure, which requires two steps with a 3596 overall yield.21 Use of 1 as a reducing agent is not practical, of course, so the commercially available LiBEt3H, 8, was tried. Gladysz reported several treating metal carbonyl dimers including ( CpFe(C0)2) 2 with 8 in a general procedure for preparing lithium salts of the metal anions.22 He noted, however, that while THF was a suitable solvent for most dimers, a more polar mixture such as 5096 THF/HMPA was required to cleanly reduce ( CpFe(C0)2) 2 to Li( CpFe(C0)2). Several groups have reported reduction of terminal carbonyls to formyl, hydroxymethyl, and methyl groups, with non-polar solvents resulting in the most reduced products.23-25 The present studies show that while 7 can be synthesized in 6096 isolated yield from 2 and 8 in toluene, no 7 is produced when the reaction is run in THF. Clearly the greater Lewis acidity of BEt3 in non-coordinating solvents is important in the C-0 scission step in the reduction of carbonyl ligands. -95- Experimental Section General Considerations. All manipulations were performed using either glove box or high vacuum line techniques. Hydrocarbon solvents were purified by vacuum transfer first from LiAlH4 and then from "titanocene".26 NMR solvents, toluene-Qg and benzene-Q6 were a~so purified by transfer from "titanocene". Hydrogen was passed through MnO on vermiculite and then 0 over 4 A molecular sieves.27 Carbon monoxide (MCB) was used directly from the cylinder. Cp*2ZrH2,28 Cp*2HfH2,29 Cp*2ZrHCl,5 ( CpRu(C0)2) 230,31 and Cp*2Zr(02C2)Cp2Ru2(C0)21a were synthesized as described previously. Infrared spectra were measured on a Beckman 4240 spectrophotometer. 1H NMR spectra were obtained using Varian EM-390 and JEOL FX-90Q spectrometers. 13c NMR spectra were obtained using the JEOL FX-90Q instrument. Samples for 13c NMR experiments were generally prepared with 13co enriched ( CpRu(C0)2) 212 and ca. 0.1 M added Cr(acac)332 to facilitate observation of the carbonyl carbons. Coupling constants were measured from either gated spectra or spectra obtained using the INEPT pulse sequence. The magnetization transfer experiments were run as described elsewhere.33 Elemental analyses were performed by Bernhardt Microanalytical Laboratory, Dornis and Kolbe Microanalytical Laboratory or Larry Henling at the CIT Analytical service. Cp2RuiCO)J( C(H)OZrHCp*2 ), 3. a) From 1 and 2. A solution of 1 (330 mg,0.909 mmol) and 2 (300 mg, 0.676 mmol) in 30 mL toluene was stirred for 30 min at -sooc, warmed to 250C and stirred for an additional 1 h. After filtration the solvent volume was reduced to 5 mL and 5 mL petroleum ether added. Stirring for 30 min produced a bright yellow slurry, which upon -96- filtration yielded 393 mg (73%) of yellow solid. Anal. calcd. for C34H4204Ru2Zr: C, 50.54; H, 5.24; Ru, 25.02; Zr, 11.28. Found: C, 51.21; H, 5.40; Ru, 24.39; Zr, 10.93. b) From Cp*2Zr(02C2)CP2Ru2<C0)2 and H2. A slurry of Cp*2Zr(02C2)Cp2Ru2(C0)2·0.5 C7H8 (500 mg, 0.587 mmol) and 20 mL petroleum ether was stirred under 1 atm H2 for 14 h at 25oc. Filtration yielded 393 mg (83%) of yellow solid. Compound 3-2.2 was prepared by method (b) using D2· Cp2Ru2(C0)3 (C(H)OHfHCp*2 ), 4, was prepared by method (a) using Cp*2HfH2. Anal. calcd. for C34H42Hf04Ru2: C, 45.61; H, 4.73. Found: C, 45.85; H, 4.75. Reaction of 3 Under CO. Compound 3 (10 mg, 0.012 mmol) and benzene-~6 (0.3 mL) were added to an NMR tube, CO (0.33 mmol) was introduced at -1960C and the tube sealed. After 14 h the 1H NMR spectrum showed complete conversion to 2 and Cp*2Zr(C0)2. Reaction of 3 with CH3I. Compound 3 (15 mg, 0.019 mmol) and benzene-~6 (0.3 mL) were added to an NMR tube into which was distilled CH3I (0.036 mmol) at -1960C. The tube was sealed and the course of the reaction monitored by 1H NMR. After 20 min at 250C conversion of 3 to a new compound with resonances at o5.12 (s, 10H, Cp), 2.00 (s, 30H, Cp*) and 12.80 (s, 1H, CH) was ca. 70% complete. A small quantity of 2 was also observed. After 15 h no 3 remained, and conversion to 2 was about 60% complete. Cp*2Zrl2 and CH4 were also observed. Reaction of 3 with CF3COOH. Compound 3 (17 mg, 0.020 mmol) and benzene-~6 (0.3 mL) were added to an NMR tube and sealed with a latex -97- septum. CF3COOH (1 lJL, 0.013 mmol) was injected via syringe. A 1H NMR spectrum measured after ca. 30 sec showed 2, 3 and a new product ( o 5.10 (s, Cp), 2.78 (s, Cp*)) in a ratio of ca. 2:6:3. After addition of a second 1 lJL aliquot of CF3COOH, the ratio was 6:2:5. A third aliquot was added, and after 15 min only 2 and Cp*2Zr(OOCCF3)2 were observed. Reaction of 3 with BF3·Me2<>. Compound 3 (15 mg, 0.019 mmol) and benzene-s!_6 were added to an NMR tube and sealed with a latex septum. BF 3•Me20 (2 lJL, 0.011 mmol) was injected via syringe. The resulting brown solid was centrifuged to the top of the tube. A 1H NMR spectrum showed a single Cp* resonance at o 1.87 and signals for a mixture of cis and trans isomers of Cp2Ru2(C0)3(CH2) (70% conversion based on integrated intensities relative to internal TMS). Reaction of 3 with HBF4·Et2<>. Treatment of 3 (15 mg, 0.019 mmol) with HBF4•Et20 (4 lJL, 0.030 mmol) as above resulted in 54% conversion to Cp2Ru2(C0)3(CH2). As with BF3·Me20, the only CpRu containing species remaining in solution besides Cp2Ru2(C0)3(CH2) was a small amount of 2. Synthesis of Cp2RuiCO)J(CH2) from 2. Compound 8 (2.5 mL, 1 M in THF) was added to a solution of 2 (50.5 mg, 1.14 mmol) in 30 mL toluene and the reaction mixture was stirred for 2 hat 25oc. Solvent was removed to yield a dark oil which contained 2 and 7 in a 1:2 ratio ( 1H NMR). The oil was dissolved in 30 mL toluene and an additional 1 mL 8 added. After 2 hat 25° the reaction mixture was opened to the air and poured into 50 mL H20. The organic phase was separated and washed twice with H20, dried over MgS04 and column chromatographed (Si02/C6H6). A single yellow band was -98- collected and the solvent flash evaporated. Recrystallization from toluene/petroleum ether yielded 280 mg (57%) of yellow crystals, pure by 1H NMR analysis. Cp2RuiC0)2 ( C(H)OZrHCp*2 )2. A mixture of 3 (111 mg, 0.14 mmol) and 1 (55 mg, 0.15 mmol) in 10 mL toluene was stirred for 2 hat 25oc. The resulting mixture was filtered and dried to yield 134 mg (82%) of the yellow solid. Anal. calcd. for C54H7404Ru2Zr2: C, 55.35; H, 6.37. Found: C, 55.71; H, 6.31. Cp2RuiC0)3 ( C(H)OZr(OH)Cp*2 ), 6. Into a solution of 3 (142 mg, 0.176 mmol) in benzene (25 mL) was injected H20 (3.5 1JL, 0.194 mmol) at 25oc. The yellow solution was stirred for 4 h, filtered, then the solvent volume reduced to 5 mL. Petroleum ether (1 0 mL) was added and the slurry was stirred an additional 45 min followed by filtration to yield 100 mg (70%) of yellow solid. Anal. calcd. for C34H4205Ru2Zr: C, 49.55; H, 5.14. Found: C, 49.43; H, 5.15. Reaction of 2 with Cp*2ZrHCl. Compound 2 (24 mg,0.054 mmol) Cp*2ZrHCl (24 mg, 0.063 mmol) and toluene-2_8 (0.3 mL) were sealed in an NMR tube. The course of the reaction was monitored over three days, at which time formation of 5 was complete. -99- References and Notes (1) (a) Berry, D. H.; Bercaw, J. E. Manuscript in preparation. (b) Berry, D. H.; Bercaw, J. E.; Jircitano, A. J.; Mertes, K. B. J. Am. Chern. Soc. 1982, 104, 4712-4715. (2) Wolczanski, P. T.; Bercaw, J. E. Accts. Chern. Res. 1980, Q, 121127. (3) Wolczanski, P. T.; Threlkel, R. S.; Bercaw, J. E. J. Am. Chern. Soc. 1979, 101, 218-220. (4) Barger, P. T.; Santarsiero, B. D.; Armantrout, J.; Bercaw, J. E. J. Am. Chern. Soc. in press. (5) Barger, P. T .; Bercaw, J. E. Organometallics in press. (6) Threlkel, R. S.; Bercaw, J. E. J. Am. Chern. Soc. 1981, 103, 26502659. (7) Reasonable structures for 3 would contain either zero or two bridging ligands. (8) (a) Casey, C. P.; Fagan, P. J.; Miles, W. H. J. Am. Chern. Soc. 1982, 104, 1134-1136. (b) Rao, S. C.; Lu, P. P. Y .; Pettit, R. Organometallics 1982, 1, 911-918. (9) Dyke, A. F.; Knox, S. A. R.; Mead, R. A.; Woodward, P. J. Chern. Soc., Chern. Commun. 1981, 861-862. (10) Theopold, K. H.; Bergman, R. G. J. Am. Chern. Soc. 1983, 105, 464475. ( 11) Adams, R. D.; Cotton, F. A. "Dynamic Nuclear Magnetic Resonance Spectroscopy"; Academic Press: New York, 197 5, Ch. 12 and references therein. -100- (12) Roberts, J. D.; Rosenberg, E.; Harris, D. C. J. Chern. Soc., Dalton 1974, 2398-2403. (13) Gansow, 0. A.; Burke, A. R.; Vernon, W. D. J. Am. Chern. Soc. 1976, 98, 5817-5825. (14) Fischer, E.O.; Clough, R. L.; Bess!, G.; Kreiss!, F. R. Angew. Chern. Int. Ed., Engl. 1976, _!2, 543-544. (15) Friedrich, P.; Besl, G.~ Fischer, E. 0.; Huttner, G. J. Organomet. Chern. 1977, 139, C68-C72. (16) Kostic, N. M.; Fenske, R. F. J. Am. Chern. Soc. 1982, 104, 3879-3884. (17) These authors also propose equilibration of the diastereotopic methylene protons in trans-Cp2CoRh(C0)2( ll-CH2) via a symmetrical intermediate in which the methylene and both carbonyl ligands are bridging. While such a mechanism could be proposed for 3, we believe that requisite intermediate, Cp2Ru2( 1J-C0)3 ( 1J-C(H)OZrHCp*2 ), is a rather unlikely species. (18) Marsella, J. A.; Curtis, C. J.; Bercaw, J. E.; Coulton, K. G. J. Am. Chern. Soc. 1980, 102, 7244-7246. (19) Bercaw, J. E. Advances in Chemistry Series 1978, 167, 136-148. (20) Hillhouse, G. L.; Bercaw, J. E. J. Am. Chern. Soc., submitted for publication. (21) Davies, D. L.; Dyke, A. F.; Knox, S. A. R.; Morris, M. J. J. Organomet. Chern. 1981, 215, C30-C32. (22) Gladysz, J. A.; Williams, G. M.; Tam, W.; Johnson, D. L. J. Organomet. Chern. 1972, 140, C1-C6. (23) Grahmn, W. A. G.; Sweet, J. R. J. Am. Chern. Soc. 1982, 104, 28112815. -101- (24) Casey, C. P.; Andrews, M.A.; McAlister, J. R. J. Am. Chern. Soc. 1980, 102, 1927. (25) Astruc, D.; Lapinty, C. J. Chern. Soc., Chern. Commun. 1983, 430431. (26) Marvich, R. H.; Brintzinger, H. H. J. Am. Chern. Soc. 1971, 94, 2046. (27) Brown, T. L.; Dicker hoof, D. W.; Bafus, D. A.; Morgan, G. L. Rev. Sci. lnstrum. 1962, 33, 491. (28) Manriquez, J. M.; McAlister, D. R.; Sanner, R. D.; Bercaw, J. E. J. Am. Chern. Soc. 1978, 100, 2716. (29) Seidler, P. F.; Frysuk, M. D.; Roddick D. M.; Hillhouse, G. L.; Bercaw, J. E. Manuscript in preparation. (30) Humphries, A. P .; Knox, S. A. R. J. Chern. Soc., Dalton 197.5, 17101714. (31) Gibson, D. H.; Hsu, W.; Steinmetz, A. L. J. Organomet. Chern. 1981, 208, 89-102. (32) Gansow, 0. A.; Burke, A. R.; Vernon, W. D. J. Am. Chern. Soc. 1972, 94, 2550-2552. (33) Perkins, T. Ph.D. Thesis, California Institute of Technology, 1981.