Kinetic resolution of chiral α-olefins using optically active ansa-zirconocene polymerization catalysts

Published on Web 06/09/2004

Kinetic Resolution of Chiral r-Olefins Using Optically Active

ansa-Zirconocene Polymerization Catalysts

Cliff R. Baar, Christopher J. Levy, Endy Y.-J. Min, Lawrence M. Henling,

Michael W. Day, and John E. Bercaw*

Contribution from the Arnold and Mabel Beckman Laboratories of Chemical Synthesis,

California Institute of Technology, Pasadena, California 91125

Received January 16, 2004; E-mail: bercaw@caltech.edu

Abstract: A series of enantiopure C₁-symmetric metallocenes, {(SiMe₂)₂[η⁵-C₅H(CHMe₂)₂][η⁵-C₅H₂((S)-

CHMeCMe₃)]}ZrCl₂, (S)-2, {(SiMe₂)₂[η⁵-C₅H(CHEt₂)₂][η⁵-C₅H₂((S)-CHMeCMe₃)]}ZrCl₂, (S)-6, and {(SiMe₂)₂-

[η⁵-C₅HCy₂][η⁵-C₅H₂((S)-CHMeCMe₃)]}ZrCl₂, (S)-7 (Cy ) cyclohexyl), zirconocene dichlorides that have

an enantiopure methylneopentyl substituent on the “upper” cyclopentadienyl ligand, and diastereomerically

pure precatalysts, {(SiMe₂)₂[η⁵-C₅H((S)-CHMeCy)(CHMe₂)][η⁵-C₅H₃]}ZrCl₂, (S)-8a and (S)-8b, which have

an enantiopure, 1-cyclohexylethyl substituent on the “lower” cyclopentadienyl ligand, has been synthesized

for use in the polymerization of chiral R-olefins. When activated with methylaluminoxane, these metallocenes

show unprecedented activity for the polymerization of bulky racemic monomers bearing substitution at the

3- and/or 4-positions. Due to the optically pure nature of these single site catalysts, they effect kinetic

resolution of racemic monomers: the polymeric product is enriched with the faster reacting enantiomer,

while recovered monomer is enriched with the slower reacting enantiomer. The two components are easily

separated. For most olefins surveyed, a partial kinetic resolution was achieved (s ) k_faster/k_slower≈ 2), but,

in one case, the polymerization of 3,4-dimethyl-1-pentene, high levels of separation were obtained (s >

15). ¹³C NMR spectroscopy of poly(3-methyl-1-pentene) produced with (S)-2 indicates that the polymers

are highly isotactic materials. X-ray crystal structure determinations for (S)-2, {(SiMe₂)₂[η⁵-C₅H(CHMe₂)₂]-

[η⁵-C₅H₂((S)-CHMeCMe₃)]}Zr(SC₆H₅)₂, (S)-6, and (S)-7 have been used in combination with molecular

mechanics calculations to examine the prevailing steric interactions expected in the diastereomeric transition

states for propagation during polymerization. Precatalysts (S)-8a and (S)-8b are less selective polymerization

catalysts for the kinetic resolution of 3-methyl-1-pentene than are (S)-2, (S)-6, and (S)-7.

Introduction

pate in substrate-directed catalysis, primarily through chelation

to the catalytically active metal center. There are a few examples

Simple nonracemic, chiral olefins represent a highly versatile

substrate class for asymmetric synthesis and are potential

monomers for the development of polymeric materials with

previously inaccessible optical or physical properties. For these

reasons, efficient routes to enantiopure alkenes are highly

desirable. Kinetic resolution, especially catalytic kinetic resolu-

tion, is a particularly attractive approach because many racemic

alkenes are readily available, while methods for the direct

synthesis of enantiopure chiral alkenes are few.

Despite the growing number of both enzymatic and metal-

mediated kinetic resolutions of a wide range of synthetically

useful racemic substrates, the practical kinetic resolution of

simple racemic alkenes (i.e., without heteroatom substituents)

has not been realized.¹This stems from the difficulty in

producing diastereomeric transition states with sufficient energy

differences due to the alkene’s lack of functionality. Conse-

quently, most successful attempts to kinetically resolve alkenes

have focused on more or less functionalized substrates such as

allylic alcohols,²allylic ethers,³and dienes,⁴which can partici-

where the antipodes of simple chiral alkenes can be separated

by kinetic resolution. For example, the osmium tetraoxide-

cinchona alkaloid system developed by Sharpless mediates the

dihydroxylation of axially disymmetric internal olefins with

modest efficiency.^5,6In a related more recent report, asymmetric

dihydroxylation was used to resolve 2,6-dimethylbenzylidenecy-

clohexane.⁷In contrast, reductive kinetic resolution strategies

have not been reported for unfunctionalized olefins.⁸In addition

to the limitations discussed above, reductive strategies for simple

(2) (a) Gao, Y.; Hanson, R. M.; Klunder, J. M.; Ko, S. Y.; Masamune, H.;

Sharpless, K. B. J. Am. Chem. Soc. 1987, 109, 5765. (b) Martin, V. S.;

Woodard, S. S.; Katsuki, T.; Yamada, Y.; Ikeda, M.; Sharpless, K. B. J.

Am. Chem. Soc. 1981, 103, 6237. (c) Kitamura, M.; Kasahara, I.; Manbe,

K.; Noyori, R.; Takaya, H. J. Org. Chem. 1988, 53, 708.

(3) (a) Adams, J. A.; Ford, J. G.; Stamatos, P. J.; Hoveyda, A. H. J. Org.

Chem. 1999, 64, 9690. (b) Morken, J. P.; Didiuk, M. T.; Visser, M. S.;

Hoveyda, A. H. J. Am. Chem. Soc. 1994, 116, 3123.

(4) (a) La, D. S.; Alexander, J. B.; Cefalo, D. R.; Graf, D. D.; Hoveyda, A.

H.; Schrock, R. R. J. Am. Chem. Soc. 1998, 120, 9720.

(5) VanNieuwenhze, M. S.; Sharpless, K. B. J. Am. Chem. Soc. 1993, 115,

7864.

(6) (a) Hamon, D. P. G.; Tuck, K. L.; Christie, H. S. Tetrahedron 2001, 57,

9499. (b) Christie, H. S.; Hamon, D. O. G.; Tuck, K. L. Chem. Commun.

1999, 1989.

(1) (a) Keith, J. M.; Larrow, J. F.; Jacobsen, E. N. AdV. Synth. Catal. 2001,

343, 5. (b) Hoveyda, A. H.; Didiuk, M. T. Curr. Org. Chem. 1998, 2, 489.

(7) Gardiner, J. M.; Norret, M.; Sadler, I. H. Chem. Commun. 1996, 2709.

9

8216

J. AM. CHEM. SOC. 2004, 126, 8216-8231

Kinetic Resolution of Chiral R-Olefins

A R T I C L E S

Scheme 1

chiral alkenes suffer from the added difficulty of separating

unreacted alkene from product alkane fractions. In an attempt

to address limitations in the application of kinetic resolution to

unfunctionalized olefins, we have explored olefin polymerization

catalysis as a possible alternative.

Ziegler-Natta and metallocene catalysts often exhibit very

high levels of enantiofacial selectivity in the polymerization of

prochiral olefins and can produce polymer with a well-defined

microstructure or tacticity.⁹Moreover, these catalysts are

extremely active, producing in many cases >10³kg of polymer/g

of metal‚h.¹⁰Hence, the prospects of using enantiopure Ziegler-

Natta or metallocene catalysts as kinetic resolving agents to

preferentially polymerize one enantiomer of a chiral alkene,

leaving the less reactive enantiomer behind (eq 1), are particu-

larly attractive. The enantio-enriched olefin should be recover-

able by simple filtration. Moreover, a new class of polymer,

one that is optically active by virtue of enantiopure substituents

off the main chain, may likewise be isolated.

very high syndiospecificities and with extremely high activi-

ties.¹⁶Modification of this catalyst system with a racemic 3,3-

dimethyl-2-butyl (“methylneopentyl”) substituent has also been

accomplished (2).¹⁷

That enantiopure, chiral sites in heterogeneous Ziegler-Natta

systems are sensitive to preexisting chirality in the olefin

monomer has been established by a number of research

groups.^11,12However, due to the racemic nature of the active

sites on a macroscopic level, resolution was not possible. A

soluble enantiopure single site catalyst, (S)-ethylene-bis(4,5,6,7-

tetrahydro-1-η⁵-indenyl)Zr(O-acetyl-(R)-mandelate)₂/MAO (MAO

) methylaluminoxane), has been used to effect a low efficiency

resolution of 4-substituted chiral olefins such as 4-methyl-1-

hexene (s ) k_faster/k_slower) 1.4).^13,14Unfortunately, poor catalyst

activity prohibited the polymerization of chiral R-olefins with

substituents in the 3-position. Although 3-methyl-1-pentene can

be successfully polymerized with related metallocene catalysts,

prior to the present work, only C_s- and racemic C₂-symmetric

catalysts have been used, precluding any possible kinetic

resolution.¹⁵

We now report that, with an enantiopure methylneopentyl

substituent, partial kinetic resolution of simple racemic R-olefins

can be carried out and that, for 3,4-dimethyl-1-pentene, syntheti-

cally useful degrees of separation can be achieved. This substrate

represents one of the simplest organic molecules to be kinetically

resolved to date.

Results and Discussion

Synthesis of {(SiMe₂)₂[η⁵-C₅H(CHMe₂)₂][η⁵-C₅H₂((S)-

CHMeCMe₃)]}ZrCl₂, (S)-2. Installation of a racemic methyl-

neopentyl substituent in rac-2 has been accomplished by

methylation of 4-tert-butylfulvene.^16,17As there are no prece-

dents for highly asymmetric fulvene alkylation or reduction, a

different route to the enantiopure metallocene was necessary

(Scheme 1).

We recently reported that doubly bridged ansa-zirconocene

catalysts (1) activated with MAO polymerize propylene with

(8) Reductive strategies have been successfully employed for the resolution

of allylic alcohols: Kitamura, M.; Kashara, I.; Manabe, K.; Noyori, R.;

Takaya, H. J. Org. Chem. 1988, 53, 710.

(9) (a) Brintzinger, H. H.; Fischer, D.; Mu¨lhaupt, R.; Rieger, B.; Waymouth,

R. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143. (b) Britovsek, G. J.

P.; Gibson, V. C.; Wass, D. F. Angew. Chem., Intl. Ed. 1999, 38, 428. (c)

Coates, G. W. Chem. ReV. 2000, 100, 1223. (d) Resconi, L.; Cavallo, L.;

Fait, A.; Piemontesi, F. Chem. ReV. 2000, 100, 1253.

(10) Facts and Figures for the Chemical Industry. Chem. Eng. News 2001, 79,

42.

(11) Pino, P. AdV. Polym. Sci. 1965, 4, 393.

(12) (a) Carlini, C.; Altomare, A.; Menconi, F.; Ciardelli, F. Macromolecules

1987, 20, 464. (b) Vizzini, J.; Ciardelli, F.; Chien, J. C. W. Macromolecules

1992, 25, 108. (c) Zambelli, A.; Proto, A.; Pasquale, L. Zeigler Catalysis;

Springer-Verlag: New York, 1995; p 218. (d) Zambelli, A.; Ammendola,

P.; Sacchi, M. C.; Locatelli, P.; Zannoni, G. Macromolecules 1983, 16,

341.

(13) (a) Ciardelli, F.; Carlini, C.; Altomare, A. Zeigler Catalysis; Springer-

Verlag: New York, 1995; p 455. (b) Chien, J. C. W.; Vizzini, J. C.;

Kaminsky, W. Makromol. Chem., Rapid Commun. 1992, 13, 479.

(14) s ) (rate of fast reacting enantiomer)/(rate of slow reacting enantiomer). s

may be determined using Kagan’s equation, s ) ln[(1 - c)(1 - ee)]/ln[(1

- c)(1 + ee)], where ee is the enantiomeric excess of the recovered olefin

and c is fraction conversion. Eliel, A. L.; Wilen, S. H.; Mander, L. N.

Stereochemistry of Organic Compounds; Wiley & Sons: New York, 1994.

(15) (a) Oliva, L.; Longo, P.; Zambelli, A. Macromolecules 1996, 29, 6383. (b)

Sacchi, M. C.; Barsties, E.; Tritto, I.; Locatelli, P.; Brintzinger, H. H.;

Stehling, U. Macromolecules 1997, 30, 1267.

Access to metallocene (S)-2 involves, as the first step,

asymmetric borane reduction of pinacolone with an oxazaboro-

lidine catalyst to give (R)-3,3-dimethylbutan-2-ol (ee_R

>

95%).^18,19The alcohol was converted to a tosylate, followed

by nucleophilic attack with cyclopentadienide anion.²⁰Remark-

ably, the reaction occurred with near perfect inversion of the

(16) Herzog, T. A.; Zubris, D. L.; Bercaw, J. E. J. Am. Chem. Soc. 1996, 118,

11988.

(17) Veghini, D.; Henling, L. M.; Burkhardt, T. J.; Bercaw, J. E. J. Am. Chem.

Soc. 1999, 121, 564.

(18) Corey, E. J.; Helal, C. J. Chem. ReV. 1998, 37, 1986.

(19) The enantiomeric excess of the alcohol was determined by chiral GC of

the trifluoroacetyl derivative.

(20) Although the success of this synthetic strategy might be surprising, it has

been shown that nucleophilic attack of cyclopentadienide anion on some

hindered secondary tosylates can proceed with inversion of configuration:

Giardello, M. A.; Conticello, V. P.; Sabat, M.; Rheingold, A. L.; Stern, C.

L.; Marks, T. J. J. Am. Chem. Soc. 1994, 116, 10212.

9

J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004 8217

A R T I C L E S

Baar et al.

Scheme 2

carbon stereocenter. The expected pathway, elimination, does

not dominate, despite the highly hindered nature of the (R)-

3,3-dimethylbutyl-2-p-toluenesulfonate substrate. The product

forms as a mixture of double bond isomers, but deprotonation

with n-BuLi yields a single species. The enantiomeric excess

of the methylneopentyl substituent was determined by synthesis

(eq 2) and NMR analysis of diastereomeric ferrocene derivatives.

Figure 1. View of (S)-2.

In the reaction of FeCl₂with Li[C₅H₄(rac-CHMeCMe₃)],

essentially equimolar amounts of the racemic (R,R/S,S) and meso

(R,S) products were obtained, while reaction with enantio-

enriched Li[C₅H₄((S)-CHMeCMe₃)] gave only a small amount

1

of the meso isomer (5.1% by H NMR spectroscopy). This

observed diastereomeric excess (de) corresponds to a 94% ee

for Li[C₅H₄((S)-CHMeCMe₃)], approximating the ee of the

starting alcohol.

Synthesis of the desired precatalyst (S)-2 then follows the

procedures already established for rac-methylneopentyl-ThpZrCl₂

(Scheme 2).^16,17Following the isolation of enantio-enriched (S)-

2, we found that a single recrystallization from toluene could

be carried out to give optically pure material. The optical purity

was assayed by reaction of (S)-2 with 2 equiv of lithium (S)-

3-methyl-1-butanethiolate (eq 3). The ¹H and ¹³C NMR spectra

of (S,S,S)-{(SiMe₂)₂[η⁵-C₅H(CHMe₂)₂][η⁵-C₅H₂(CHMeCMe₃)]}-

Zr(SCH₂CH(CH₃)CH₂CH₃)₂((S,S,S)-4) were compared to the

NMR data obtained for a 1:1 mixture of diastereomers resulting

from the reaction of rac-2 with the (S)-thiolate. No resonances

corresponding to the (R,S,S) diastereomer were observed,

indicative of >98% optically pure material.

X-ray Crystal Structure Determinations for (S)-2 and

{(SiMe₂)₂[η⁵-C₅H(CHMe₂)₂][η⁵-C₅H₂((S)-CHMeCMe₃)]}Zr-

(SC₆H₅)₂, (S)-5. Single crystals of (S)-2 were grown from a

saturated toluene solution. For comparison purposes, single

crystals of {(SiMe₂)₂[η⁵-C₅H(CHMe₂)₂][η⁵-C₅H₂((S)-CHMe-

Figure 2. View (S)-5.

CMe₃)]}Zr(SC₆H₅)₂, (S)-5, prepared from the reaction of (S)-2

with 2 equiv of Li(SC₆H₅), were grown from hexamethyldisi-

loxane. The molecular structures of (S)-2 and (S)-5 show that

the large tert-butyl group of the methylneopentyl substituent is

located above the upper cyclopentadienyl ligand of the metal-

locene (Figures 1 and 2). Newman projections (Figure 3) down

the C15a-C4a bond of (S)-2 and the C15-C4 bond of (S)-5

indicate that the tert-butyl fragment is directed from the methine

carbon C15 nearly perpendicular to the cyclopentadienyl plane

9

8218 J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004

Kinetic Resolution of Chiral R-Olefins

A R T I C L E S

Table 1. Kinetic Resolutions of Chiral, 3- and 4-Methyl-

Substituted R-Olefins Using (S)-2

Figure 3. Newman projections showing the orientation of the methylneo-

pentyl group in (S)-2 and (S)-5.

and that the methyl group is forced down into the wedge of the

metallocene, effectively differentiating left from right in the

molecule. The isopropyl substituents on the lower cyclopenta-

dienyl assume a conformation in which the methine proton is

directed back toward the bulky dimethylsilylene linker groups

as has been observed for the structures of similarly substituted

metallocenes.^16,17,21In the case of (S)-5, the steric requirements

of the thiophenoxy groups have rotated the C16 methyl group

out of the wedge to some extent, but it is the relative position

of each thiophenoxy ligand that is particularly noteworthy. The

phenyl ligand attached to S1 occupies an open region of space

in the upper left quadrant of the molecular structure, avoiding

nonbonded contacts with C16 (Figure 2) and the left isopropyl

group on the lower cyclopentadienyl ligand. As a consequence,

the neighboring phenyl group attached to S2 is oriented down

and toward the lower right quadrant. This anti arrangement

clearly minimizes steric interactions between the thiophenoxy

phenyls in the metallocene wedge. Deflection of the phenyl

group into the lower half of the wedge may also prevent a direct

steric interaction between the phenyl attached to S2 and C16.

Finally, the plane defined by S1-Zr-S2 is 6.9° off the

perpendicular to the ring centroid-Zr-ring centroid plane. For

(S)-2, a smaller deviation of 3.1° of the Cl1a-Zr1a-Cl2a in

the opposite direction is observed. For (S)-2, the distortion is

in the direction expected to minimize interaction of the

methylneopentyl methyl (C16a) with chlorine (Cl2a); for (S)-

5, the distortion is in the opposite direction, probably because

the phenyl ligand is forced down and experiences unfavorable

steric interactions with the isopropyl substituent on the lower

cyclopentadienyl ligand.

Polymerization of Chiral r-Olefins with {(SiMe₂)₂[η⁵-

C₅H(CHMe₂)₂][η⁵-C₅H₂((S)-CHMeCMe₃)]}ZrCl₂, (S)-2. When

activated with MAO, (S)-2 catalyzes the polymerization of a

variety of chiral R-olefins with moderate to good kinetic

resolution. Polymerizations were carried out in tetradecane,

which acts as both a solvent and an internal standard for GC

analysis. The high boiling point of tetradecane also allows for

facile recovery of volatile, unreacted monomer by vacuum

transfer. When desired, the polymers were more conveniently

isolated using toluene as solvent. All of the polymeric materials

were isolated as white powders.

Olefin enantioassay required derivatization procedures to

achieve useful separation by enantioselective GC. The recovered

monomer is first converted to a carboxylic acid having one less

carbon atom using Ru-catalyzed NaIO₄oxidation. The acid is

then treated with a BF₃/MeOH solution to give a methyl ester

that is used for the enantiomeric excess (ee) assay. Because

possible racemization during derivatization was a concern, a

control experiment was carried out in which enantiopure (S)-

3-methyl-1-pentene²²was converted to the corresponding methyl

ester with the same derivatization procedure. The resulting chiral

GC trace showed a single peak with no evidence for racem-

ization (ee_S> 98%), demonstrating that the enantioassay is valid

for 3-methyl-1-pentene. The polymerization results obtained

with the first-generation catalyst (S)-2 are outlined in Table

1.^14,23

The data show small but significant s values, even for the

simplest possible chiral R-olefin, 3-methyl-1-pentene (3-MP1).

(22) (a) Fu, S. C. J.; Birnbaum, S. M.; Greenstein, J. P. J. Am. Chem. Soc.

1954, 76, 6054. (b) Schurig, V.; Leyer, U.; Wistuba, D. J. Org. Chem.

1986, 51, 242. (c) Millar, J. G.; Underhill, E. W. J. Org. Chem. 1986, 51,

4726. (d) Wood, N. F.; Chang, F. C. J. Org. Chem. 1965, 30, 2054.

(23) For polymerizations with the (S)-2/MAO, the prevailing absolute config-

uration in the recovered monomers was determined by comparison of their

optical rotations with literature reports. For 3,5,5-trimethyl-1-hexene, the

rotation of enriched material has never been reported. The (S)-mandelic

ester derivatives of enriched 3-methyl-1-pentene and 3,5,5-trimethyl-1-

hexene were made as outlined in the Experimental Section. Both products

show less intensity for the upfield signal of the diastereotopic methyl groups

(R to the methyl mandelic ester moiety), arguing for a common absolute

configuration in these olefins. The same is true for the ester derived from

3, 4-dimethyl-1-pentene. With these assignments in hand, the relative

retention times for olefin derivatives as obtained from enantioselective GC

were used to assign enantiomer selectivity with other catalyst systems. For

literature reports on the optical rotation of olefins 3-methyl-1-pentene,

4-methyl-1-hexene, 3,5,5-trimethyl-1-hexene, and 3,4-dimethyl-1-pentene,

see: (a) Schurig, V.; Leyrer, U.; Wistuba, D. J. Org. Chem. 1986, 51,

242. (b) Lazzaroni, R.; Salvadori, P.; Bertucci, C.; Veracini, C. A. J.

Organomet. Chem. 1975, 99, 475. (c) Lardicci, L.; Caporusso, A. M.;

Giacomelli, G. J. Organomet. Chem. 1974, 70, 333. (d) Lazzaroni, R.;

Salvadori, P.; Pino, P. Tetrahedron Lett. 1968, 2507. (e) Pino, P.; Lardicci,

L.; Centoni, L. Gazz. Chem. Ital. 1961, 91, 428. During our analysis of the

relevant literature, we found a discrepancy in the assignment of the absolute

configuration and the optical rotation for enantiomers of 3,4-dimethyl-1-

pentene. For the (+) specific rotation, one series of papers assigned the

enantiomer (S)-3,4-DMP1 (ref 23d; Lardicci, L.; Menicagli, R.; Caporusso,

A. M.; Giacomelli, G. Chem. Ind. 1973, 4, 184). However, another series

of articles has correlated the (+) specific rotation to the enantiomer (R)-

3,4-DMP1 (ref 23b,c; Caporusso, A. M.; Giacomelli, G. P.; Lardicci, L.

Atti Soc. Tosc. Sci. Nat., Mem. 1973, 80, 40). Due to this conflict in the

literature, it was necessary to unambiguously assign each enantiomer its

proper optical rotation. The literature is in agreement regarding the absolute

configuration of structurally related (-)(R)-2,3-dimethylbutyric acid (in this

case, the absolute configuration has been established by comparison to well-

established steroid natural products) (Sakai, K.; Tsuda, K. Chem. Pharm.

Bull. 1963, 11, 650. Tarzia, G.; Tortorella, V.; Romeo, A. Gazz. Chim.

Ital. 1967, 97, 102. di Maio, G.; Romeo, A. Gazz. Chim. Ital. 1959, 89,

1627). Because this acid is the product of 3,4-DMP1 oxidation with NaIO₄

and RuCl₃, and because the stereocenter is retained during the oxidation

process (vide infra), determination of the absolute configuration of 3,4-

DMP1 follows directly. Hence, the 3,4-DMP1 monomer recovered fol-

lowing polymerization with (S)-2 was converted to 2,3-dimethylbutyric acid,

and the optical rotation was measured: λ ) 598^NaDnm, l ) 50 mm, c )

0.9860, T ) 25.9 °C, specific rotation ) -18.1156. The specific rotation

is consistent with (-)(R)-2,3-dimethylbutyric acid, identifying the recovered

olefin as (-)(R)-3,4-dimethyl-1-pentene. The S monomer then is prefer-

entially polymerized.

(21) Veghini, D.; Day, M. W.; Bercaw, J. E. Inorg. Chim. Acta 1998, 280, 226.

9

J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004 8219

A R T I C L E S

Baar et al.

Figure 4. ¹³C NMR spectrum (125.72 MHz) of poly(3-methyl-1-pentene) made with catalyst (S)-2/MAO.

Scheme 3

Scheme 4

metallocene wedge) during each enchainment event, consistent

with the highly isotactic poly-(3-methyl-1-pentene).

Although the differences in transition state energies are clearly

small, one can rationalize the enantiomeric preferences on the

basis of relative steric interactions. For these 3-methyl-

substituted R-olefins, uptake of the S enantiomer on the right

side of the wedge is favored. If one assumes conformational

freedom for the incoming olefin, then for the olefin adduct (a

first-order model of the insertion transition state) S-monomer

binding places the larger 3-position substituent (R ) Et,

isopropyl, neopentyl vs Me) in the more open region between

the lower cyclopentadienyl ligand isopropyl groups. For the

R-monomer, the olefin R substituent is directed more toward

the lower right [CHMe₂] cyclopentadienyl substituent (Scheme

4). Hence, the antipode selectivity of this system is the result

of an indirect mechanism. An interaction between the chiral

methylneopentyl substituent of the “upper” cyclopentadienyl

dictates the preferred side of the wedge for the growing polymer

chain, which in turn allows the incoming olefin access to only

one side of the metallocene wedge. The steric interactions of

the side chain of the monomer with the chiral pocket provided

by the “lower” cyclopentadienyl ligand, not the cyclopentadienyl

directly bearing the enantiopure methylneopentyl substituent,

lead to selectivity!

Most interestingly, a synthetically useful s value is obtained

for an olefin bearing substitution at both the 3- and the

4-positions; 3,4-dimethyl-1-pentene (3,4-DMP1) was resolved

with s > 15. Increasing the substitution at the 5-position, as in

3,5,5-trimethyl-1-hexene (3,5,5-TMH1), has a less dramatic

effect. Very poor resolution was observed for 4-methyl-1-hexene

(4-MH1) in which the stereogenic center is separated from the

olefin moiety by a methylene unit. In all cases, (S)-2 preferen-

tially incorporates the S antipode of the racemic olefin into the

polymer. Although we have measured some selectivity for

3-phenyl-1-hexene polymerization (s ) 1.3), the antipode

selectivity remains uncertain at this time.

Significantly, the ¹³C NMR spectrum of poly(3-methyl-1-

pentene) made with (S)-2 was consistent with a predominantly

isotactic microstructure (Figure 4).^12c,15aAnalysis of the prevail-

ing steric interactions in the active site of the catalyst suggests

that isotactic polymer results from a rapid site epimerization

process (polymer chain swinging to the opposite side of

metallocene wedge after each migratory insertion of monomer),

as has been established for rac-2/MAO-catalyzed propylene

polymerization under dilute monomer conditions.¹⁷Hence, to

minimize steric interactions, the growing polymer chain swings

to the more open (left) side of the wedge, avoiding contact with

the methyl group of the chiral substituent (Scheme 3). Because

the chiral olefins polymerized with enantiopure 2 are bulkier

than propylene, and are therefore enchained more slowly, it is

likely that the rate of site epimerization greatly exceeds the rate

of propagation. With a high rate of site epimerization (relative

to olefin uptake and insertion), the olefin encounters the same

steric environment (the chiral pocket on the right side of the

To support these conclusions, molecular mechanics calcula-

tions (CAChe) were carried out for olefin binding into the right

side of the wedge using [{(SiMe₂)₂[η⁵-C₅H(CHMe₂)₂][η⁵-C₅H₂-

((S)-CHMeCMe₃)]}ZrCH₃]⁺as a model for the active catalyst.²⁴

Consistent with the experimental results, for 3-MP1, S antiopode

binding is slightly favored (∆E ) 0.8 kcal mol^-1). On the other

hand, care should be exercised when using these computational

(24) Molecular mechanics calculations were carried out using the CAChe

software package (v. 4.9). A model complex {(SiMe₂)₂[η⁵-C₅H(CHMe₂)₂]-

[η⁵-C₅H₂((S)-CHMeCMe₃)]}ZrMe(η²-olefin)}⁺was used for the active site.

The Zr atom, the methyl group, and the olefinic carbons were forced to be

coplanar. The R-substituent was fixed at 90° to the plane containing the Zr

atom and the olefin carbons.

9

8220 J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004

Kinetic Resolution of Chiral R-Olefins

A R T I C L E S

Scheme 5

predictions with such a simple model (e.g., for 3,4-DMP1, the

R antipode is favored by more than 1.5 kcal mol^-1, contrary to

experiment!), especially when the transition state energy dif-

ferences are so very small, as indicated by the relatively small

s values. Moreover, it is likely that the ∆E of the diastereomeric

transition states for an alkyl migration process, rather than the

∆E for a discrete olefin intermediate, as in these approximations,

determines the relative rates of enantiomer polymerization in

the current system. Finally, several recent lines of evidence (vide

infra) suggest that additional interactions arising from the chiral

monomer and asymmetric zirconocene fragment with the

asymmetric â and γ centers of the migrating polymeryl chain

are also important and should therefore be included in the

transition state computations.

Figure 5. View of (S)-6.

Synthesis and Kinetic Resolution Trials with {(SiMe₂)₂-

[η⁵-C₅H(CHEt₂)₂][η⁵-C₅H₂((S)-CHMeCMe₃)]}ZrCl₂, (S)-6,

and {(SiMe₂)₂[η⁵-C₅HCy₂][η⁵-C₅H₂((S)-CHMeCMe₃)]}ZrCl₂,

(S)-7. Because the steric interactions between substituents on

the 3 carbon of the incoming chiral olefin and the isopropyl

substituents on the “lower” cyclopentadienyl ligand for (S)-2

appear to be key factors in differentiating the diastereotopic

transition states for propagation, we next sought to examine the

effect of changing the nature of the 3,5-substituents on the lower

ligand of the zirconocene catalyst. For (S)-2, the incorporation

of two isopropyl groups involves successive condensation

reactions of cyclopentadiene with acetone to give as a primary

ligand fragment, 1,3-(CHMe₂)₂C₅H₄.^16,25A similar synthetic

route was taken here for the incorporation of 3-pentyl substit-

uents (Scheme 5). The preparation of 1,3-Cy₂C₅H₄was carried

out according to the literature procedure.^25b

Starting from the appropriately substituted cyclopentadiene

and using a procedure analogous to that used in the preparation

of (S)-2, 3-pentyl or cyclohexyl substitution was introduced at

the 3,5-positions to give enantiopure precatalysts (S)-6 and (S)-

7, respectively.^16,17

Figure 6. View of (S)-7.

the X-ray data for complexes (S)-2 and (S)-5 (Table 4), the

methylneopentyl groups adopt a conformation that orients the

methyl group toward one side of the metallocene wedge. Figure

5 also shows that the 3-pentyl substituents at the 3,5-positions

of the “lower” cyclopentadienyl ligand adopt a staggered

conformation and are rotated so that the methine proton is

directed back toward the SiMe₂groups, much like the isopropyl

substituents of (S)-2. When comparing the molecular structures

of (S)-2 (Figure 1) and (S)-6 (Figure 5), it is evident that for

each catalyst the open areas between the 3,5-substituents on

the “lower” cyclopentadienyl ligand are not greatly different.

The molecular structure of (S)-6 is shown in Figure 5, and

that for (S)-7 is shown in Figure 6. As expected on the basis of

(25) (a) Stone, K. J.; Little, R. D. J. Org. Chem. 1984, 49, 1849. (b) Clark, T.

J.; Killian, C. M.; Luthra, S.; Nile, T. A. J. Organomet. Chem. 1993, 462,

247.

9

J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004 8221

A R T I C L E S

Baar et al.

Table 2. Comparisons of the Kinetic Resolutions of Chiral, 3- and

4-Methyl-Substituted R-Olefins Using (S)-2, (S)-6, and (S)-7

chiral information of the C₁-symmetric metallocene framework

to the olefin, repeatedly approaching from the same side of the

metallocene wedge. For this second-generation catalyst, a more

direct interaction was anticipated between the incoming chiral

monomer and the chiral substituent in the ligand framework.

The choice of enantiopure 1-cyclohexylethyl as the chiral

substituent requires as a key component 1-((S)-1-cyclohexyl-

ethyl)-3-(isopropyl)cyclopentadienide. The successful strategy

follows the sequence shown in Scheme 6. As before with the

methylneopentyl substituent, access to enantiopure metallocene

begins with an optically pure alcohol. A known, enzymatically

catalyzed kinetic resolution of 1-cyclohexylethanol was ame-

nable to scale-up and could be used to give both (R)- and (S)-

1-cyclohexylethanol in high enantiopurity (ee > 96%).^19,26

Conversion of the alcohol to mesylate and nucleophilic attack

by cyclopentadienide anion affords (S)-1-cyclohexylethylcy-

clopentadiene. Pyrrolidine-catalyzed condensation with acetone

gives 6,6-dimethyl-(S)-1-cyclohexylethylfulvene as a single

regioisomer.²⁵Reduction with LiAlH₄gives the desired cyclo-

pentadiene as a mixture of double bond isomers, and subsequent

deprotonation with n-butyllithium gives a single product. At

this stage, synthesis of ferrocene derivatives was again used to

establish that (near) perfect inversion of the carbon stereocenter

had once again occurred during the S_N2 reaction (ee_S> 96%):

although ¹H NMR was not very informative due to strong

overlap of reasonances for rac and meso isomers in all regions

of the spectrum, examination of the ¹³C NMR spectrum showed

the presence of only the S,S diastereomer (de > 99%) in the

reaction of FeCl₂with Li[C₅H₄((S)-1-cyclohexylethyl)].

Table 3. Kinetic Resolutions of Chiral, 3- and 4-Methyl-

Substituted R-Olefins Using Diastereomer 8a

The cyclohexyl groups of precatalyst (S)-7 take eclipsed, chair

conformations, leading to an even smaller change in steric bulk

relative to (S)-2 (Figure 6). Overall then, the prevailing steric

interactions presented by complexes (S)-2, (S)-6, and (S)-7 are

rather similar.

From this point on, a modified version of the procedures used

for preparing the ligand for (S)-2 was employed; however, due

to the nonequivalence of the 1,3-substituents, two products are

possible on reaction with SiCl₂Me₂, and both are observed

(Scheme 6). As expected, a complicated ¹H NMR spectrum was

obtained due the presence of double bond isomers, but depro-

tonation gave a more easily interpreted spectrum, and the

presence of two regioisomers in an approximate 1:1 ratio could

be identified. Reaction with Li(C₅H₅) followed by deprotonation

gives a 1:1 mixture of the singly bridged linkage isomers.

Addition of SiCl₂Me₂to the doubly deprotonated mixture of

regioisomers then yields a single species through formation of

what appears to be the stable tricyclic ligand arrangement. A

final deprotonation with 2 equiv of n-butyllithium gives the

dilithio salt, “(S)-1-cyclohexylethyl-ThpLi₂”, in high yield.

Activation of (S)-6 or (S)-7 with MAO affords highly active

catalysts suitable for the polymerization of 3-substituted olefins.

The results of the kinetic resolution are given in Table 2 together

with the average s values for (S)-2.²³Readily apparent is that,

relative to (S)-2, only minor changes in s are obtained for most

of the olefins surveyed. In the few deviations of note, precatalyst

(S)-6 showed a small increase in s for the polymerization of

4-methyl-1-pentene and more significantly a 3-fold increase for

3,5,5-trimethyl-1-hexene (s ) 6) relative to (S)-2 and (S)-7 (s

) 2). This increase in selectivity is consistent with a “fine-

tuning” of the catalyst chiral pocket for interaction with 4-MH1

and 3,5,5-TMH1. For 3,4-DMP1, there is a slight decrease in s

relative to that for (S)-2. The variation in turnover numbers is

not readily understood, but may reflect different percentages

of active zirconium centers and/or differing catalyst stabilities.

Overall, the general trends in the observed s values for (S)-2,

(S)-6, and (S)-7 track well with expectations based on the

molecular structures of (S)-2 and (S)-6 (see above).

{(SiMe₂)₂[η⁵-C₅H((S)-CHMeCy)(CHMe₂)][η⁵-C₅H₃]}Zr-

Cl₂(8a), a Precatalyst with an Enantiopure Substituent on

the “Lower” Cyclopentadienyl Ligand. In addition to rela-

tively simple extensions that involve (minor) variances in the

steric interactions at the 3,5-substituents of the lower cyclo-

pentadienyl ligand, a rather different strategy was explored:

introduction of an enantiopure chiral group at one of these

positions. Catalysts based on (S)-2, for which the enantiopure

group dictates the orientation of the polymer chain, relay the

Metalation of the above ligand was not straightforward. The

lack of symmetry of the ligand allows for two possible

diastereomers on metalation. Although some diastereoselectivity

was expected, to our surprise, reaction of (S)-1-cyclohexylethyl-

ThpLi₂with ZrCl₄gave an essentially 1:1 ratio of diastereomers

(Scheme 7). It is likely that the diastereotopic faces of the less

substituted “upper” cyclopentadienyl ligand encounter the

zirconium metal first in a nonselective coordination event,

followed by a rapid coordination of the bulky “lower” cyclo-

pentadienyl ligand, giving roughly equal amounts of each

diastereomer. Although the salt metathesis route is nonselective,

a single diastereomer could be isolated by washing the mixture

with cold petroleum ether to remove the much more soluble

(26) (a) Frykman, H.; O¨ hrner, N.; Norin, T.; Hult, K. Tetrahedron Lett. 1993,

34, 1367. (b) O¨ hrner, N.; Martinelle, M.; Mattson, A.; Norin, T.; Hult, K.

Biotechnol. Lett. 1992, 14, 263.

9

8222 J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004

Kinetic Resolution of Chiral R-Olefins

A R T I C L E S

Table 4. X-ray Experimental Data^a

compound

(S)-2

(S)-5

C₃₈H₅₂S₂Si₂Zr

(S)-6‚toluene

(S)-7‚dichloromethane

C_33.32H_52.64Cl_4.64Si₂Zr

[C₃₂H₅₀Cl₂Si₂Zr‚1.32(CH₂Cl₂)]

765.15 [653.04‚1.32(84.93)]

monoclinic

P2₁(No. 4)

13.2833(7)

19.3805(10)

14.7651(8)

90

107.195(1)

90

3631.2(3)

4

formula

C₂₆H₄₂Cl₂Si₂Zr

C_33.50H₅₄Cl₂Si₂Zr

[C₃₀H₅₀Cl₂Si₂Zr‚¹/₂(C₇H₈)]

675.09 [629.02‚¹/₂(92.14)]

triclinic

P1 (No. 1)

10.3602(5)

13.1543(7)

14.0843(7)

97.199(1)

104.068(1)

106.275(1)

1748.26(15)

2

formula weight

crystal system

space group

a, Å

572.90

720.32

tetragonal

P4₃2₁2 (No. 96)

12.6459(8)

47.844(4)

90

monoclinic

P2₁(No. 4)

8.8568(8)

19.7932(18)

16.5826(15)

90

b, Å

c, Å

R, deg

â, deg

105.522(1)

90

γ, deg

90

volume, Å³

Z

2801.0(4)

4

1.359

0.68

1200

7651.1(10)

8

1.251

0.48

F

_calc, g/cm³

1.282

0.56

714

1.400

0.73

1598

µ, mm^-1

F₀₀₀

3040

crystal shape

crystal color

crystal size, mm

T, K

blade

irregular block

pale yellow

0.17 × 0.22 × 0.26

98

plate

colorless

plate

colorless

0.16 × 0.23 × 0.27

98

colorless

0.09 × 0.20 × 0.26

100

0.15 × 0.22 × 0.24

100

θ range, deg

1.6, 28.3

-11, 11; -26, 25; 21,21

48 755

1.7, 28.6

-16, 16; -16, 16; -64, 63

75 962

1.6, 28.2

-13, 13; -17, 17; -17, 18

39 577

1.4, 28.4

-17, 16; -25, 25; -19, 19

54 234

h,k,l limits

data measured

unique data

12 773

9368

15 367

16 726

R_int

0.066

0.082

0.041

0.064

data, F_o> 4σ(F_o)

parameters/restraints

R1,^bwR2;^call data

R1,^bwR2;^cF_o> 4σ(F_o)

GOF^don F ²

∆F_max,min, e Å^-3

11 952

583/1

0.038, 0.058

0.034, 0.058

1.21

7859

544/0

0.057, 0.060

0.043, 0.059

1.55

13 420

716/3

0.036, 0.063

0.030, 0.062

1.45

12 842

765/1

0.059, 0.064

0.037, 0.060

1.13

1.26, -0.48

0.98, -0.72

0.60, -0.31

1.17, -0.76

^aAll data were collected on a Bruker SMART 1000 ccd with graphite monochromated Mo KR radiation (λ ) 0.71073 Å). ^bR1 ) ∑||F_o| - |F_c||/∑|F_o|.

2

^cwR2 ) {∑[w(F_o- F_c)²]/∑[w(F_o)²]}^1/2

.

^dGOF ) S ) {∑[w(F_o- F_c)²]/(n - p)}^1/2

.

diastereomer 8b. The identity of the less soluble diastereomer

fully reversible amine elimination procedure recently developed

for the improved synthesis of rac-(EBTHI)ZrCl₂complexes

(EBTHI ) ethylene-bis(4,5,6,7-tetrahydro-1-η⁵-indenyl)).²⁹For

the present system, reaction of the diprotio form of the ligand

with Zr(NMe₂)₄gives a e1:10 ratio of 9a:9b at equilibrium

(Scheme 8).

1

as 8a was established by H and ¹³C NMR spectroscopy.

The ¹H NMR spectrum of 8a shows two resonances at higher

field corresponding to “axial” methyl groups of the isopropyl

and cyclohexylethyl substituents, respectively (see below), those

lying almost perpendicular to and below the plane of the “lower”

Cp ligand. A single resonance appearing at lower field corre-

sponds to an equatorial methyl group (also part of the isopropyl

substituent) which is approximately coplanar with the “lower”

cyclopentadienyl ligand. The chemical shift assignments are

consistent with a greater ring current effect on the axial methyl

Reaction of the diastereomeric mixture of diamides with

Me₃SiCl gives the desired precatalyst 8b as the main species

(>90%). The ¹H NMR spectrum of 8b showed the presence of

two downfield resonances at δ 1.26 and δ 1.34 for methyl

groups that are approximately coplanar with the “lower” Cp. A

single upfield resonance for an axial methyl group is present at

1

hydrogens²⁷and have been corroborated by comparison to H

1

NMR data obtained for a related, X-ray structurally characterized

zirconocene dichloride also bearing isopropyl and cyclohexyl-

ethyl substituents at the 3,5-positions.²⁸As a consequence, the

large cyclohexyl group points forward and out of the metal-

locene wedge in the prevailing conformation, while the methine

proton is directed back toward the dimethylsilylene group,

consistent with the molecular structures of (S)-2 and (S)-6 and

(S)-7 (Figures 1, 5, and 6). Molecular mechanics calculations

were carried out for 8a, and the minimized structure is fully

consistent with the cyclohexylethyl group favoring the confor-

mation shown in Scheme 7.

δ 0.90. These H NMR parameters are in contrast to those

observed for 8a and indicate that the large cyclohexyl group is

rotated out of plane in diastereomer 8b. The minimized structure

obtained from molecular mechanics calculations is consistent

with this conclusion. Unfortunately, despite repeated efforts, the

extreme solubility of 8b prevented successful crystallization,

and 8b could not be isolated in analytically pure form (impurities

1

<10% by H NMR spectroscopy). Despite this difficulty, we

considered the complex to be of sufficient purity for use in

kinetic resolution procedures, and some data were acquired (vide

infra).

Molecular mechanics calculations also predict that diastere-

omers 8a and 8b are of substantially different energy with 8b

favored by 1.7 kcal mol^-1. Salt metathesis, then, appears to give

a kinetic mixture of products rather than a thermodynamically

established one. Isolation of 8b was accomplished using the

The kinetic resolution results for 8a are given in Table 3.²³

The data show that the s values are markedly diminished relative

to those with (S)-2, (S)-6, and (S)-7. Indeed, the only olefin to

show even a small degree of selectivity was 3,4-DMP1 (s )

2.6). Significantly, in all cases, the R-monomer was polymerized

(27) Friebolin, H. Basic One- and Two-Dimensional NMR Spectroscopy, 3rd

ed.; Wiley-VCH: New York, 1998.

(28) Baar, C. R.; Day, M. W.; Bercaw, J. E., manuscript in preparation.

(29) Diamond, G. M.; Rodewald, S.; Jordan, R. F. Organometallics 1995, 14,

5.

9

J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004 8223

A R T I C L E S

Baar et al.

Scheme 6

Figure 7. ¹³C NMR spectrum (125.72 MHz) of polypropylene made with

catalyst 8a/MAO.

Scheme 8

Scheme 7

orientation, and hence its minimal role in steric interactions with

the bound olefin or with the growing polymer chain. Diastere-

omer 8b then behaves much like an achiral, C_s-symmetric

system, precluding selectivity for kinetic resolutions. For this

reason, other, less bulky, olefins were not examined.

On the basis of our steric model, a predominantly syndiotactic

enchainment is perhaps expected with catalysts derived from

8a and 8b, because, in the absence of an “upper” methyl-

neopentyl substituent, the polymeryl chain can access open

regions of space on either side of the metallocene wedge.³⁰This

expectation is borne out for polymerization of (liquid) propylene

with 8a, which gives rise to a mainly syndiotactic polymer, [r]

) 91.5% (Figure 7).^30aIn contrast, the ¹³C NMR spectrum

obtained for poly(3-methyl-1-pentene) was consistent with a

prevailingly isotactic polymer (Figure 8).^12c,15

at a higher rate than the S-monomer, opposite to that observed

for (S)-2, (S)-6, and (S)-7. With diastereomer 8b, 3,4-DMP1

was polymerized with very low selectivity, s ) 1.4, again with

a preference for the R antipode. This diminished selectivity is

almost certainly a consequence of the cyclohexyl group’s axial

(30) In a C_S-symmetric catalyst, syndiotactic polymer microstructure arises from

propylene insertion from regularly alternating sides of the metallocene

wedge: (a) Resconi, L.; Cavallo, L.; Fait, A.; Piemontesi, F. Chem. ReV.

2000, 100, 1223. (b) Coates, G. W. Chem. ReV. 2000, 100, 1253.

9

8224 J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004

Kinetic Resolution of Chiral R-Olefins

A R T I C L E S

Scheme 10

polymer that is predominantly isotactic (with isolated r diads)

will be formed, but it will contain stereoblocks of opposite

absolute configuration at the carbons within the polymer

backbone (-RRRRR- and -SSSSS-).

Figure 8. ¹³C NMR spectrum (125.72 MHz) of poly(3-methyl-1-pentene)

made with catalyst 8a/MAO.

Scheme 9

Although in Scheme 9 we have indicated that the stereose-

lectivities for a single antipode of racemic 3MP1 are very high

for each of the three manifolds, this need not be the case. The

observed s value of 1.1 indicates that if stereoselectivities are

high, then there are nearly equal rates for isotactic enchainment

manifolds A and B. Alternatively (and more likely), the

stereoselectivities for both manifolds are low.

Conclusions

We have developed a series of enantiopure C₁-symmetric

metallocene catalysts that polymerize relatively bulky, racemic

3-methyl-substituted R-olefins with unprecedented activities. In

addition, the enantiomers react at different rates, allowing for

the partial kinetic resolution of simple racemic alkenes. For 3,4-

dimethyl-1-pentene, s > 15 has been achieved, marking the first

practically useful kinetic resolution of such a simple unfunc-

tionalized monomer. We have proposed a general mechanism

to explain both the isotacticity of the resulting polymers as well

as the preferred antipodal selectivity for catalysts (S)-2, (S)-6,

and (S)-7. We note, however, that the situation is significantly

more complicated, considering the number of stereocenters in

the vicinity of the active site (Scheme 10).

That chain-end chirality at the γ stereocenter may be

important in determining the relative rates of enantiomer

polymerization was recently demonstrated using rac-C₂-sym-

metric and C_s-symmetric metallocene catalysts.¹⁵Similarly,

polymerization of 3-methyl-1-pentene with 8a gives isotactic

polymer due to a chain-end-mediated site epimerization process.

This process is thought to be controlled by chirality at the

â-stereocenter.^31,32The extent to which the chiral polymer chain

end influences site epimerization in catalysts (S)-2, (S)-6, and

(S)-7 remains unknown. Considering catalyst (S)-2, for example,

the influence of the methylneopentyl group on the preferred

orientation of the polymer chain, that is, right or left in the

zirconocene wedge, could work in concert with or against any

preference for the chain to be right or left due to influences

from the chiral polymer chain end itself. Similarly, during

enchainment, the antipodal preference for chiral monomer will

be a function of both the enantiomorphic catalyst site and the

polymer chain-end chirality (largely at â and γ positions).

Future directions for the project will focus on copolymeri-

zation experiments with mixtures of chiral and achiral olefins

Initially, it seemed difficult to rationalize a highly isotactic

polymer in combination with low resolving power for 8a (Table

3): if isotacticity was the result of the polymer chain lying

predominantly on one side of the metallocene wedge (as is the

case for (S)-2), then one would expect at least a similar degree

of resolution, because the sterics of the 3,5-substituents of the

“lower” cyclopentadienyl ligand are at least as large in the

present catalyst as in the original system. Instead, we propose

that the polymer chain does have access to both sides of the

metallocene wedge through the mechanism shown in Scheme

9. A similar mechanism was proposed by Zambelli to explain

the formation of isotactic blocks during the polymerization of

bulky 3-substituted olefins (e.g., 3-methyl-1-butene) using an

achiral C_s-symmetric metallocene catalyst.³¹In this mechanism,

a chain-end-mediated site epimerization controls which diaster-

eomeric active site is favored for a â-R or â-S chain end,

respectively (equilibrium A S B in Scheme 9). If the energy

difference between diastereomeric transition states using either

A or B is large, then one expects site epimerization to be fast

relative to olefin uptake and insertion. As a result, enchainment

occurs mainly from one side of the metallocene wedge,

producing an isotactic block until a syndiotactic enchainment

error occurs (by olefin insertion prior to site epimerization). A

syndiotactic insertion error, followed by rapid site epimerization,

switches olefin uptake from one side of the wedge to the other.

Once again, isotactic enchainment occurs (but with opposite

enantiofacial selectivity) until another syndiotactic insertion error

switches the side of olefin uptake yet again. In this way, a

(32) Borriello, A.; Busico, V.; Cipullo, R.; Chadwick, J. C.; Sudmeijer, O.

(31) Grisi, F.; Longo, P.; Zambelli, A.; Ewen, J. A. J. Mol. Catal. A 1999, 225.

Macromol. Chem., Rapid Commun. 1996, 589.

9

J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004 8225

A R T I C L E S

Baar et al.

mL) in a 250 mL round-bottom flask. Pyridine (35.5 mL, 440 mmol)

and DMAP (250 mg) were added, and the reaction mixture was stirred

for 48 h at room temperature. Thin-layer chromatography (4:1 hexanes:

ethyl acetate eluant) indicated that the reaction had gone to completion.

The reaction mixture was subjected to subsequent washings with water,

1 N potassium bisulfate, saturated sodium bicarbonate, and brine. The

organic layer was dried over anhydrous MgSO₄, and the solvent was

removed in vacuo to give 35.0 g of a thick light-yellow oil (99.6%

from the alcohol). ¹H NMR (CDCl₃): δ ) 0.70 (s, 9H), 1.42 (d, 3H),

4.38 (q, 1H, J ) 7.8 Hz), 7.35 (d, 2H), 7.75 (d, 2H, J ) 8.4 Hz).

(rac)-3,3-Dimethylbutyl-2-p-toluenesulfonate was synthesized according

to the same procedure (90.3%).

that are intended to indirectly probe the role that the stereogenic

centers of the polymer chain end play in the resolution process.

We are also developing new C₁-symmetric catalysts that contain

multiple enantiopure substituents, nonracemic C₂-symmetric

metallocenes, and we are exploring nonmetallocene strategies

in an effort to design more general and selective kinetic

resolution catalysts.

Experimental Section

General Considerations. All air- and/or moisture-sensitive com-

pounds were manipulated using standard high-vacuum line, swivel frit

assembly, Schlenk and cannula techniques, or in a glovebox under a

nitrogen atmosphere as described previously.³³Argon was purified and

dried by passage through columns of MnO on vermiculite and activated

4 Å molecular sieves. All solvents were stored under vacuum over

sodium benzophenone ketyl, titanocene, or calcium hydride prior to

use. Pinacolone was dried over anhydrous CaSO₄and distilled prior to

use. Methylaluminoxane (MAO) was purchased from Albemarle, and

all volatiles were removed in vacuo to give a white powder. The CBS

catalyst (S)-tetrahydro-1-methyl-3,3-diphenyl-1H,3H-pyrrolo[1,2-c]-

[1,3,2]oxazaborole-borane and the Novozyme-435 enzyme were pur-

chase from Aldrich. Olefins were purchased from Chemsampco and

stored under vacuum over lithium aluminum hydride. (S)-2-Methylbu-

tane-1-thiol³⁴and Li[(rac)-MNCp]¹⁶(MN ) methylneopentyl) were

synthesized according to literature methods. (S)-2 was synthesized by

the route previously reported for the racemic counterpart, except that

[((S)-CHMeCMe₃)C₅H₄]Li was used in place of the racemate.¹⁷

Mandelic esters of alcohols and carboxylic acids were synthesized on

the basis of literature procedures.³⁵(S)-Ethylthiooctanoate was made

according to the literature procedure.^26a

(S)-Methylneopentylcyclopentadiene. Lithium cyclopentadienide

(41.1 g, 571 mmol) was weighed into a 1 L flask in a nitrogen glovebox.

Dry tetrahydrofuran (500 mL) was added, and the flask was cooled in

an ice bath. (R)-3,3-Dimethylbutyl-2-p-toluenesulfonate (94.1 g, 393

mmol) was added as a 1:1 solution with THF. N,N,N′,N′-Tetrameth-

ylethylenediamine (172 mL, 1140 mmol) was added, and the solution

was allowed to warm to room temperature. The solution was refluxed

for 2 days under an argon atmosphere. An aliquot, taken after 24 h,

showed (GC) a 1:1 ratio of (S)-methylneopentylcyclopentadiene:

tosylate; after 48 h, the ratio was 14:1. Low boiling petroleum ether

(250 mL) was added, and excess LiCp was quenched by the slow

addition of water (10 mL). The solution was treated to washings with

water, 1 N KHSO₄, saturated NaHCO₃, and brine. The solution was

dried over anhydrous MgSO₄, and the solvent was removed in vacuo.

The (S)-methylneopentylcyclopentadiene was then vacuum transferred

from the reaction mixture to give 38.3 g of a colorless liquid (57%

from tosylate). As expected, the ¹H NMR spectrum was complex,

indicating a mixture of double bond isomers.

Li[(S)-methylneopentylcyclopentadienide]. (S)-Methylneopentyl-

cyclopentadiene (19.2 g, 128 mmol) was dissolved in dry diethyl ether

(250 mL) in a 500 mL flask attached to a swivel frit assembly. The

flask was cooled to -78 °C, and n-butyllithium (1.6 M in hexanes,

95.6 mL, 153 mmol) was added by syringe via the sidearm of the swivel

frit. The cold bath was removed, and the solution was allowed to warm

to room temperature under an atmosphere of argon open to a mercury

bubbler. The reaction was stirred for 12 h at room temperature, and

the solvent was then removed in vacuo. Petroleum ether (200 mL) was

then vacuum transferred onto the solid material, the resulting suspension

was filtered, and the solid was washed twice with petroleum ether.

The white solid was dried in vacuo and stored in a nitrogen glovebox.

NMR spectra were recorded on a General Electric QE300 (300 MHz

1

for H), Bruker AMX (500 MHz for H), JEOL Delta (400 MHz for

1

¹H), or Varian Mercury (300 MHz for H) spectrometer. Gas chro-

matographs were obtained on an Agilent 6890 Series gas chromatograph

using a 30 m × 0.25 mm, γ-cyclodextrin trifluoroacetyl “Chiraldex-

TA” column from Advanced Separations Technology. Optical rotations

were measured on a Jasco P1030 polarimeter at ambient temperature.

A 1 mL cell with a 1 dm path length was used.

(R)-3,3-Dimethyl-2-butanol. The borane adduct of the (S)-CBS

catalyst (14.6 g, 50 mmol) was placed in a 500 mL two-neck flask

along with a magnetic stir bar, under a nitrogen atmosphere in a

glovebox. Dry methylene chloride (50 mL) and borane-dimethyl

sulfide complex (100 mL, 1.00 mol) were added to the argon-flushed

flask, and the mixture was cooled to -20 °C with an ethylene glycol/

water slush bath. Pinacolone (125.0 mL, 1.00 mol) was added over a

4 h period via a syringe pump. Stirring at -20 °C was continued for

4 h after the addition was complete. The reaction was quenched by

dropwise addition of methanol in a 1 L beaker cooled in an ice bath.

After addition of the reaction mixture, the solution was allowed to warm

to room temperature and was stirred for 1 h, after which time no further

hydrogen evolution was evident. The solution was concentrated to 150

mL by distillation, 250 mL of methanol was added, and the process

was repeated to remove the volatile boron compounds from the reaction

mixture. Distillation was continued until all methanol was removed,

and the product alcohol was then vacuum distilled. (rac)-3,3-Dimethyl-

2-butanol was synthesized by the LiAlH₄reduction of pinacolone using

standard conditions.

1

Yield: 17.4 g, 87% from methylneopentylcyclopentadiene. H NMR

(THF-d₈): δ ) 0.79 (s, 9H), 1.13 (d, 3H), 2.39 (q, 1H, J ) 7.14 Hz),

5.47 (d, 2H), 5.49 (d, 2H, J ) 2.5 Hz).

(rac/meso)-Iron Bis(methylneopentylcyclopentadienide), 3. Iron-

(II) chloride (81 mg, 0.64 mmol) and lithium (rac)-methylneopentyl-

cyclopentadienide (200 mg, 1.28 mmol) were weighed into a 50 mL

flask under a nitrogen atmosphere. The flask was evacuated, and dry

THF (15 mL) was vacuum transferred into the flask at -78 °C. The

flask was allowed to warm to room temperature and stirred for 3 h.

The contents of the flask were then poured into water (50 mL), and

petroleum ether (50 mL) was added to aid in separation of the organic

and aqueous layers. The aqueous layer was washed twice with diethyl

ether, and the organic fractions were combined and dried with

anhydrous MgSO₄. The resulting orange solution was filtered through

silica, and the solvent was removed in vacuo to give a viscous orange-

1

brown liquid (188 mg, 91.2%). The H NMR spectrum shows a 1:1

(R)-3,3-Dimethylbutyl-2-p-toluenesulfonate. (S)-3,3-Dimethyl-2-

butanol (15.0 g, 147 mmol) was dissolved in methylene chloride (100

ratio of two compounds, the racemic and meso forms of the ferrocene.

¹H NMR (meso, CDCl₃): δ ) 0.77 (s, 18H), 1.33 (d, 6H), 2.22 (q,

2H, J ) 7.12 Hz), 3.98 (m, 8H). ¹³C NMR (meso, CDCl₃): δ ) 16.00,

(33) Burger, B. J.; Bercaw, J. E. In New DeVelopments in the Synthesis,

Manipulation and Characterization of Organometallic Compounds; Wayda,

A., Darensbourg, M. Y., Eds.; American Chemical Society: Washington,

DC, 1987; Vol. 357.

(34) Vasi, I. G.; Desai, K. R. J. Indian Chem. Soc. 1975, 52, 837.

(35) (a) Parker, D. J. Chem. Soc., Perkin Trans. 2 1983, 83. (b) Chataigner, I.;

Lebreton, J.; Durand, D.; Guingant, A.; Villie´ras, J. Tetrahedron Lett. 1998,

39, 1759.

1

28.08, 33.97, 43.48, 67.55, 67.65, 67.99, 71.60, 93.39. H NMR (rac,

CDCl₃): δ ) 0.77 (s, 18H), 1.31 (d, 6H), 2.22 (q, 2H, J ) 7.09 Hz),

3.97 (m, 8H). ¹³C NMR (rac, CDCl₃): δ ) 15.86, 28.06, 34.02, 43.74,

67.31, 67.75, 71.38, 93.01. (S,S)-Iron bis(methylneopentylcyclopen-

tadienide) was synthesized by the same procedure using (S)-lithium

9

8226 J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004

Kinetic Resolution of Chiral R-Olefins

A R T I C L E S

1

methylneopentylcyclopentadienide as the starting material (yield 99%).

The ¹H NMR (CDCl₃) spectrum shows 5.1% meso compound, as

compared to (S,S)-3 which makes up 94.9% of the observed signal.

{(SiMe₂)₂[η⁵-C₅H(CHMe₂)₂][η⁵-C₅H₂((R/S)-CHMeCMe₃)]}Zr[(S)-

2-methylbutane-1-thiolate]₂, (R/S,S,S)-4. rac-2 (100 mg, 0.197 mmol)

and lithium (S)-2-methylbutane-1-thiolate (43.3 mg, 0.393 mmol),

prepared by reaction of (S)-2-methylbutane-1-thiol with 1 equiv of

butyllithium (1.6 M in hexanes) in diethyl ether and precipitated by

addition of petroleum ether, were placed in a 25 mL flask and attached

to a swivel frit assembly. Diethyl ether (15 mL) was vacuum transferred

onto the solids at -78 °C. The flask was allowed to warm to room

temperature and was stirred for 3 h, resulting in a yellow solution and

a white precipitate. The solvent was removed in vacuo, and petroleum

ether (15 mL) was transferred into the flask. The suspension was

filtered, the solid was washed with petroleum ether, and the solvent

was removed in vacuo to give a yellow oil. Hexamethyldisiloxane (10

mL) was added and removed in vacuo, leaving the product as a yellow

Torr) to give 31 g (83%) of a pale yellow oil. H NMR (300 MHz,

benzene-d₆): δ 0.80, 0.87 (t’s, 6H, CH₃), 1.3, 1.5 (m’s, 4H, CH₂), 2.2

(m, 1H, CH), 2.64, 2.8, 5.93, 6.15, 6.23, 6.34, 6.45 (m’s, 5H, C₅H₅).

3-(3-Pentyl)-6,6-diethylfulvene. The desired product was prepared

by reaction of (3-pentyl)CpH (15.52 g, 114 mmol), 3-pentanone (10.79

g, 126 mmol), and pyrrolidine (10.54 g, 148 mmol) in 100 mL of

methanol. The reaction was refluxed overnight. The workup was the

same as that for 6,6-diethylfulvene. 17.22 g (74%) of bright orange oil

was obtained via Kugel-Ro¨hr distillation (80 °C, <10^-3Torr). ¹H NMR

(300 MHz, benzene-d₆): δ 0.93 (q, J ) 7.8 Hz, 12H, CH₃), 1.58 (m,

4H, (CH₂)₂CH), 2.25 (m, 1H, CH), 2.25 (m, 1H, CH), 2.3 (q, J ) 7.8

Hz, 4H, (CH₂) ₂Cd), 6.24 (t, J ) 2.4 Hz, 1H, C₅H₃), 6.46 (dd, J ) 5.4,

1.8 Hz, 1H, C₅H₃), 6.57 (dd, J ) 5.4, 1.8 Hz, 1H, C₅H₃).

1,3-Bis(3-pentyl)cyclopentadiene Isomers, “(3-pentyl)₂CpH”. The

product was obtained by reaction of 3-(3-pentyl)-6,6-diethylfulvene

(12.54 g, 61.04 mmol) in ether (80 mL) and the 150 mL ether slurry

of LiAlH₄(3.48 g, 91.7 mmol). The reaction was quenched with 3.5 g

of water, followed by 3.5 g of 15% NaOH and 10.5 g of water.

Filtration, concentration, and Kugel-Ro¨hr distillation (90 °C, <10^-3

1

solid (73% isolated). The H NMR spectrum shows a 1:1 mixture of

1

(S,S,S)-4 and (R,S,S)-4. H NMR [(S,S,S), benzene-d₆]: δ ) 0.56 (s,

1

3H), 0.58 (s, 3H), 0.59 (s, 3H), 0.64 (s, 3H), 0.91 (s, 9H), 0.82-1.02

(m, 6H), 0.94 (d, 3H, J ) 5.13 Hz), 1.04 (d, 3H, J ) 7.19 Hz), 1.05

(d, 3H, J ) 5.95 Hz), 1.06 (d, 3H, J ) 5.13 Hz), 1.28 (m, 2H), 1.38

(d, 3H, J ) 6.60), 1.60 (d, 3H, J ) 6.59 Hz), 1.67 (m, 2H), 1.90 (d,

3H, J ) 7.03 Hz), 2.8-3.5 (m, 9H), 6.62 (s, 1H), 6.70 (d, 1H), 6.77

Torr) yielded 10.47 g (83%) of pale yellow oil. H NMR (300 MHz,

benzene-d₆): δ 0.82 (t, J ) 7.1 Hz, 6H, CH₃), 0.90 (t, J ) 7.4 Hz, 6H,

CH₃), 1.35-1.50 (m, 8H, CH₂), 2.12 (m, 1H, CH), 2.55, 2.7, 5.87,

6.09, 6.12 (s’s, 4H, C₅H₄).

Lithium [1,3-Bis(3-pentyl)cyclopentadienide]‚DME, “Li[(3-pentyl)-

₂Cp]‚DME”. Pentane (125 mL) was vacuum transferred onto (3-

pentyl)₂CpH (20.5 g, 98.7 mmol) in a swivel frit assembly. n-BuLi

(67.9 mL, 1.6 M in hexane, 108.6 mmol) was added via syringe over

5 min at -78 °C. The solution was allowed to warm slowly to room

temperature and was stirred for 6 h. DME (11.3 mL, 108.6 mmol) was

vacuum transferred onto the reaction mixture to facilitate the precipita-

tion of the salt product. After being stirred for 2 h at room temperature,

the solution was cooled with an ice bath, and the resulting white solid

was filtered and washed with pentane (3 × 75 mL). The white solid

was dried in vacuo resulting in a fine white powder; 21.9 g (99%). ¹H

NMR (300 MHz, THF-d₈): δ 0.81 (t, J ) 7.5 Hz, 12H, CH₃), 1.45,

1.58 (m’s, 8H, CH₂), 2.19 (m, 2H, CH), 3.27 (s, 6H, DME CH₃), 3.43

(s, 4H, DME CH₂), 5.40 (m, 3H, C₅H₃).

{C₅H₃(CHEt₂)₂}SiMe₂Cl Isomers. THF (150 mL) was vacuum

transferred onto Li[(3-pentyl)₂Cp]‚DME (10.0 g, 43.0 mmol) in a swivel

frit assembly. While the solution was cooling at -78 °C, dichlorodi-

methylsilane (9.0 mL, 74.2 mmol) was added by vacuum transfer. The

reaction mixture was allowed to warm slowly to room temperature

overnight. The reaction was stirred at 25 °C for another 24 h. The

solvent was removed in vacuo, leaving a white paste. Petroleum ether

(150 mL) was added by vacuum transfer, and the product was extracted

away from LiCl salts. The solvent was removed in vacuo, leaving a

yellow oil (11.69 g, 92%) which was used in the next step without

1

(d, 1H, J ) 1.95 Hz). H NMR [(R,S,S), benzene-d₆]: δ ) 0.56 (s,

3H), 0.58 (s, 3H), 0.59 (s, 3H), 0.64 (s, 3H), 0.91 (s, 9H), 0.82-1.02

(m, 6H), 0.94 (d, 3H, J ) 5.13 Hz), 1.04 (d, 3H, J ) 7.19 Hz), 1.05

(d, 3H, J ) 5.95 Hz), 1.06 (d, 3H, J ) 5.13 Hz), 1.28 (m, 2H), 1.38

(d, 3H, J ) 6.60), 1.60 (d, 3H, J ) 6.59 Hz), 1.67 (m, 2H), 1.90 (d,

3H, J ) 7.03 Hz), 2.85-3.40 (m, 9H), 6.61 (s, 1H), 6.69 (d, 1H), 6.75

(d, 1H, J ) 1.47 Hz). (S,S,S)-4 (70% yield) was made using the same

procedure.

{(SiMe₂)₂[η⁵-C₅H(CHMe₂)₂][η⁵-C₅H₂((S)-CHMeCMe₃)]}Zr(SC₆-

H₅)₂, (S)-5. Compound 5 was made analogously to 4, but from the

reaction of (S)-2 with 2 equiv of lithium thiophenoxide (74% yield).

¹H NMR (benzene-d₆): δ ) 0.08 (s, 3H), 0.31 (d, 3H, J ) 6.60 Hz),

0.54 (s, 3H), 0.55 (s, 3H), 0.67 (s, 9H), 0.69 (s, 3H), 0.88 (d, 3H, 7.03

Hz), 1.05 (d, 3H, J ) 7.03), 1.49 (d, 3H, J ) 7.03 Hz), 1.59 (d, 3H,

J ) 6.60 Hz), 2.85-3.0 (m, 2H), 3.28 (quintet, 1H), 6.45 (s, 1H), 6.55

(s, 1H), 6.63 (s, 1H), 6.75-6.98 (m, 4H), 7.05 (t, 2H, J ) 7.47 Hz),

7.53 (s, 1H), 7.55 (s, 1H), 8.03 (s, 1H), 8.06 (s, 1H).

6,6-Diethylfulvene. Into a 500 mL flask containing a methanol (250

mL) solution of CpH (50 g, 751 mmol) were added 3-pentanone (65

g, 751 mmol) and pyrrolidine (8.01 g, 113 mmol) at room temperature.

The bright yellow solution was allowed to stir overnight at room

temperature. The resulting dark orange reaction mixture was neutralized

with 50% (v/v) acetic acid and was transferred to a separatory funnel

containing 100 mL of water. After the layers were separated, the

aqueous layer was extracted with ether (3 × 100 mL). The combined

organic layers were washed with water (2 × 100 mL), sat. NaHCO₃(1

× 100 mL), and brine (1 × 100 mL). The organic layer was dried

over anhydrous MgSO₄, concentrated, and Kugel-Ro¨hr distilled (25

°C, <10^-3Torr) to yield 68.14 g (68%) of bright yellow oil. ¹H NMR

(300 MHz, benzene-d₆): δ 0.9 (t, J ) 7.8 Hz, 6H, CH₃), 2.3 (q, J )

7.8 Hz, 4H, CH₂), 6.5 (s, 4H, C₅H₄).

1

further purification. H NMR (300 MHz, benzene-d₆): δ 0.15 (s, 3H,

Si(CH₃)₂), 0.30 (s, 3H, Si(CH₃)₂), 0.88 (m, 12H, CH₃), 1.50 (m, 8H,

CH₂), 2.16 (m, 1H, CH), 2.60 (m, 1H, CH), 3.35, 5.87, 6.2 (s’s, 3H,

C₅H₃).

{C₅H₃(CHEt₂)₂}SiMe₂{C₅H₄((S)-CHMeCMe₃)}. THF (100 mL)

was vacuum transferred onto Li (S)-MNCp (6.17 g, 39.5 mmol) in a

swivel frit assembly. THF (25 mL) was vacuum transferred onto {C₅H₃-

(CHEt₂)₂}SiMe₂Cl in a 50 mL flask. The solution of the silyl compound

was added to the Li[(S)-MNCp] solution via cannula transfer at 0 °C.

The mixture was stirred at room temperature overnight. Solvent was

removed in vacuo, and petroleum ether was vacuum transferred onto

the white paste. The product was extracted away from the LiCl with

petroleum ether (3 × 100 mL). The solvent was removed in vacuo,

resulting in an orange oil, which was Kugel-Ro¨hr distilled (110 °C,

1-(3-Pentyl)cyclopentadiene and Isomers, “(3-pentyl)CpH”. Un-

der an atmosphere of Ar, an ether (330 mL) slurry of LiAlH₄(1 equiv)

was prepared and was equipped with a condenser and an addition

funnel. 6,6-Diethylfulvene (30 g, 223 mmol) was dissolved in 100 mL

of ether, and the solution was added dropwise to the LiAlH₄slurry

over 1 h. The solvent refluxed very gently during the addition. The

reaction was stirred at room temperature for 4 h. While the flask was

cooled with a dry ice bath, the reaction was quenched slowly with 9 g

of water, followed by 9 g of 15% NaOH and 27 g of water. On warming

to room temperature, the gray slurry became white in color. The solution

was filtered, concentrated, and Kugel-Ro¨hr distilled (30 °C, <10^-3

1

<10^-3Torr) to give 16.10 g (99%) of product. H NMR (300 MHz,

benzene-d₆): δ -0.14, -0.04 (br s, 6H, Si(CH₃)₂), 0.924 (m, 12H,

CH₂CH₃), 0.98 (s, 9H, C(CH₃)₃), 1.20 (d, J ) 7.2 Hz, 3H, CH₃), 1.56

(m, 8H, CH₂), 2.12-2.58 (m, 3H, CH), 2.96-3.6, 6.01-6.61 (m, 7H,

C₅H₃, C₅H₄).

9

J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004 8227

A R T I C L E S

Baar et al.

Li₂[{C₅H₂(CHEt₂)₂}SiMe₂{C₅H₃((S)-CHMeCMe₃)}]‚DME. Ether

(150 mL) was vacuum transferred onto {C₅H₃(CHEt₂)₂}SiMe₂{C₅H₄-

((S)-CHMeCMe₃)} (16.1 g, 39.2 mmol) in a swivel frit assembly.

n-BuLi (54.0 mL, 1.6 M in hexane, 86.4 mmol) was syringed onto the

solution while it was cooled at -78 °C. The clear solution was stirred

overnight at room temperature. The solvent was removed in vacuo,

and petroleum ether (125 mL) and DME (9.0 mL, 86.4 mmol) were

added by vacuum transfer, producing a copious amount of white

precipitate. After cooling in an ice bath, the solution was filtered. The

solid was dried in vacuo, resulting in a fine white powder (32.5 g,

benzene-d₆): δ 1.1-2.3 (m, 22H, CH₂), 2.72, 2.81, 5.84, 6.15, 6.22

(m, 3H, C₅H₃).

Li[(C₅H₃Cy₂)]‚DME. The synthesis of the target compound from

Cy₂CpH was analogous to that used in the preparation of Li[(3-

1

pentyl)₂Cp]‚DME (85%). H NMR (300 MHz, THF-d₈): δ 1.30 (m,

10H, Cy CH₂), 1.6 (m, 10H, Cy CH₂), 1.9 (m, 10H, Cy CH₂), 2.35 (m,

1H, Cy CH₂), 3.3 (s, 6H, DME CH₃), 3.4 (s, 4H, DME CH₂), 5.35, 5.4

(m’s, 3H, C₅H₃).

(C₅H₃Cy₂)SiMe₂Cl Isomers. The desired compound was prepared

as a white fluffy solid in a manner similar to that used for {C₅H₃-

1

90%). H NMR (300 MHz, THF-d₈): δ 0.29, 0.30 (s, 6H, Si(CH₃)₂),

(CHEt₂)₂}SiMe₂Cl (90%). H NMR (300 MHz, benzene-d₆): δ 0.10,

0.81 (t, J ) 7.2 Hz, 12H, CH₂CH₃), 0.85 (s, 9H, C(CH₃) ₃), 1.21 (d, J

) 7.2 Hz, 3H, MN CH₃), 1.36-1.64 (m, 8H, CH₂), 2.21 (quint, J )

7.2 Hz, 1H, CH(CH₂)₂), 2.47 (q, J ) 7.2 Hz, 1H, MN CH), 2.67 (quint,

J ) 7.2 Hz, 1H, CH(CH₂)₂), 3.27 (s, 6H, DME CH₃), 3.43 (s, 4H,

DME CH₂), 5.55 (d, J ) 3.0 Hz, 1H, C₅H₂), 5.67 (t, J ) 2.7 Hz, 1H,

C₅H₃), 5.71 (d, J ) 2.7 Hz, 1H, C₅H₂), 5.81 (m, 1H, C₅H₃), 5.86 (m,

1H, C₅H₃).

{C₅H₂(CHEt₂)₂}(SiMe₂)₂{C₅H₃((S)-CHMeCMe₃)}. THF (150 mL)

was vacuum transferred onto the dilithio salt Li₂[{C₅H₂(CHEt₂)₂}SiMe₂-

{C₅H₃((S)-CHMeCMe₃)}]‚DME (10.0 g, 19.5 mmol). At -78 °C,

SiMe₂Cl₂(1 equiv) was added by vacuum transfer, and the solution

was allowed to warm to room temperature overnight. All volatiles were

removed in vacuo. The ligand was extracted away from LiCl with

petroleum ether. After the ligand was dried under vacuum, it was

Kugel-Ro¨hr distilled (130 °C, <10^-3Torr) to give a pale yellow oil

(85%).

0.25 (s, 6H, SiMe₂), 1.25 (m, 10H, Cy CH₂), 1.65 (t, 6H, Cy CH₂),

1.90 (m, 4H, Cy CH₂), 2.27 (m, 1H, Cy CH), 2.50 (m, 1H, Cy CH),

3.45 (br s, 1H, C₅H₃), 5.87 (br s, 1H, C₅H₃), 6.28 (s, 1H, C₅H₃).

Li₂[(C₅H₂Cy₂)SiMe₂{C₅H₃((S)-CHMeCMe₃)}]‚DME. (C₅H₂Cy₂)-

SiMe₂{C₅H₃((S)-CHMeCMe₃)} was prepared analogously to the pro-

cedure used for {C₅H₃(CHEt₂)₂}SiMe₂{C₅H₄((S)-CHMeCMe₃)} to give

a dark yellow oil (98%). The dilithio salt of this oil was prepared by

deprotonation in the same manner as used for Li₂[{C₅H₃(CHEt₂)₂}-

SiMe₂{C₅H₄((S)-CHMeCMe₃)} (95%). ¹H NMR (300 MHz, THF-d₈):

δ 0.28 (s, 6H, Si(CH₃)₂), 0.84 (s, 9H, C(CH₃)₃), 1.18 (d, 3H, MN CH₃),

1.32 (m, 10H, Cy CH₂), 1.70 (m, 6H, Cy CH₂), 1.88 (m, 4H, Cy CH₂),

2.36 (m, 1H, Cy CH), 2.46 (qt, 1H, MN CH), 2.78 (m, 1H, Cy CH),

3.27 (s, 6H, DME CH₃), 3.42 (s, 4H, DME CH₂), 5.63 (m, 3H, C₅H₃),

5.82 (m, 2H, C₅H₂).

(C₅H₂Cy₂)(SiMe₂)₂{C₅H₃((S)-CHMeCMe₃)}. THF (50 mL) was

added by vacuum transfer to a swivel frit assembly containing the

dilithio salt Li₂[(C₅H₂Cy₂)SiMe₂{C₅H₃((S)-CHMeCMe₃)}]‚DME (3.32

g, 7.4 mmol). At -78 °C, SiMe₂Cl₂(1 mL, 8 mmol) was added by

vacuum transfer. The resulting slurry was stirred at -78 °C and allowed

to warm slowly to room temperature overnight. All volatiles were

removed in vacuo. Petroleum ether was added by vacuum transfer, and

the solution was filtered to give a yellow filtrate. All volatiles were

removed in vacuo, and the yellow oil obtained was pumped down under

high vacuum until a solid formed (73%). ¹H NMR (300 MHz, benzene-

d₆): δ 0.55 (s, 6H, Si(CH₃)₂), 0.62 (s, 6H, Si(CH₃)₂), 1.03 (s, 9H,

C(CH₃)₃), 1.28 (d, 3H, MN CH₃), 1.28-2.8 (m, 20H, Cy CH₂), 2.32

(t, 1H, Cy CH), 2.56 (m, 1H, MN CH), 2.82 (t, 1H, Cy CH), 3.5 (m,

1H, Cp), 6.46 (m, 2H, Cp), 7.02 (d, 1H, Cp).

{(SiMe₂)₂[η⁵-C₅HCy₂][η⁵-C₅H₂((S)-CHMeCMe₃)]}ZrCl₂, (S)-7. To

a 50 mL flask were added (C₅H₂Cy₂)(SiMe₂)₂{C₅H₃((S)-CHMeCMe₃)}

(0.93 g, 1.9 mmol), Zr(NMe₂)₄(0.51 g, 1.9 mmol), and dry xylenes

(25 mL). The resulting yellow solution was refluxed under a strong Ar

purge to vent product HNMe₂. When the vented gas showed a neutral

reading on pH paper, all volatiles were removed in vacuo. The reaction

vessel was attached to a swivel frit assembly, and the residue was

dissolved in toluene (25 mL). To this solution was added TMSCl (2.2

equiv) by vacuum transfer. After the solution was stirred overnight at

room temperature, the volatiles were removed, and the residue was

washed with (TMS)₂O (61%). Crystals were obtained from CH₂Cl₂.

¹H NMR (300 MHz, benzene-d₆): δ 0.48 (s, 3H, Si(CH₃)₂), 0.50 (s,

3H, Si(CH₃)₂), 0.54 (s, 3H, Si(CH₃)₂), 0.58 (s, 3H, Si(CH₃)₂), 0.78 (s,

9H, C(CH₃)₃), 1.0-1.4 (m, 10H, Cy CH₂), 1.61 (d, 3H, MN CH₃), 1.6

(m, 6H, Cy CH₂), 1.73 (m, 4H, Cy CH₂), 2.5-2.7 (m, 2H, Cy CH),

2.87 (qt, 1H, MN CH), 6.42 (s, 1H, Cp), 6.64 (s, 1H, Cp), 6.75 (s, 1H,

Cp). Anal. Calcd for C₃₂H₅₀Si₂ZrCl₂: C, 58.20; H, 7.72. Found: C,

58.20; H, 7.93.

(R)-1-Cyclohexylethanol. rac-1-Cyclohexylethanol (50 mL, 390

mol) and (S)-ethyl thiooctanoate (73.4 mL, 390 mmol) were combined

in a 250 mL round-bottom flask. Two grams of the Novozyme-435

enzyme were added. With the flask venting properly in a fume hood

(ethanethiol evolution!), the temperature was increased to 43 °C with

constant stirring. The reaction was stopped short of 50% conversion

(5.5 h; % conversion by GC) followed by removal of the enzyme by

filtration. The product (R)-cyclohexylethyloctanoate (43 g) was isolated

from the reaction mixture by fractional distillation under reduced

{(SiMe₂)₂[η⁵-C₅H(CHEt₂)₂][η⁵-C₅H₂((S)-CHMeCMe₃)]}ZrCl₂, (S)-

6. In a swivel frit apparatus, the ligand {C₅H₂(CHEt₂)₂}(SiMe₂)₂{C₅H₃-

((S)-CHMeCMe₃)} (0.95 g, 2.0 mmol) was dissolved in 50 mL of ether.

n-BuLi (2.78 mL, 4.45 mmol) was added by syringe to the flask. The

solution was allowed to stir for 24 h at room temperature under Ar. In

a separate flask, ZrCl₄(541 mg, 2.3 mmol) was dissolved in ether, and

the solution was cannula transferred to the solution containing the

ligand. The resulting mixture was allowed to stir at room temperature

under Ar for 24 h. The solution was filtered, and the solvent was

removed in vacuo. Hexamethyldisiloxane (40 mL) was added by

vacuum transfer to precipitate a white solid, which was collected by

filtration, washed, and dried under vacuum. The product was recrystal-

1

lized from cold toluene. H NMR (300 MHz, benzene-d₆): δ 0.52,

0.54, 0.59, 0.62 (s, 4H, Si(CH₃)₂), 0.58 (td, J ) 1.7, 7.2 Hz, 6H,

CH₂CH₃), 0.81 (s, 9H, C(CH₃)₃), 1.02 (m, 6H, CH₂CH₃), 1.37 (m, 2H,

CH₂), 1.56 (m, 2H, CH₂), 1.64 (d, J ) 7.2 Hz, MN CH₃), 1.69 (m, 2H,

CH₂), 2.38 (m, 2H, CH₂), 2.68 (m, 2H, CH₂), 2.90 (q, J ) 7.2 Hz, 1H,

MN CH), 6.45 (s, 1H, C₅H₁), 6.67 (d, J ) 1.8 Hz, 1H, C₅H₂), 6.79 (d,

J ) 1.8 Hz, 1H, C₅H₂). Anal. Calcd for C₃₀H₅₀Si₂ZrCl₂: C, 57.28; H,

8.01. Found: C, 56.57; H, 8.40.

6-Cyclohexylfulvene. The product was prepared from reaction of

cyclopentadiene and cyclohexanone, using the same method as

described for 6,6-diethylfulvene (85%). ¹H NMR (300 MHz, benzene-

d₆): δ 1.29 (m, 2H, CH₂), 1.41 (m, 1H, CH₂), 2.34 (t, 4H, CH₂), 6.62

(m, 4H, C₅H₄).

Cyclohexylcyclopentadiene Isomers, “CyCpH”. The product was

prepared from 6-cyclohexylfulvene in a manner similar to that of (3-

1

pentyl)CpH (91%). H NMR (300 MHz, benzene-d₆): δ 1.2 (m, 5H,

CH₂), 1.64 (m, 3H, CH₂), 1.80 (m, 1H, CH₂), 1.92 (m, 1H, CH₂), 2.22

(m, 1H, CH), 2.73, 2.80, 5.95, 6.18, 6.24, 6.35, 6.5, 6.55 (m, 5H, C₅H₅).

3,6-Dicyclohexylfulvene. The product was prepared from CyCpH

and cyclohexanone using the same method described for (3-pentyl)-

6,6-diethylfulvene. The crude oil was Kugel-Ro¨hr distilled (130 °C,

<10^-3Torr) to yield a yellow oil (70%). ¹H NMR (300 MHz, benzene-

d₆): δ 1.1-2.4 (m, 21H, CH₂), 6.29 (m, 1H, C₅H₃), 6.55 (dd, J ) 1.6,

5.3, 1H, C₅H₃), 6.63 (dd, J ) 1.6, 5.3, 1H, C₅H₃).

1, 3-Dicyclohexylcyclopentadiene Isomers, “Cy₂CpH”. The prepa-

ration of this compound was analogous to that described for (3-

1

pentyl)₂CpH to yield a pale yellow oil (94%). H NMR (300 MHz,

9

8228 J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004

Kinetic Resolution of Chiral R-Olefins

A R T I C L E S

pressure. The (R)-cyclohexylethyloctanoate was hydrolyzed in a 1 M

solution of NaOH in MeOH (400 mL) for 3 d. The product alcohol

was then extracted with ether. This required the addition of water and

brine to get two layers. The combined ether layers were washed with

water and brine and then dried over MgSO₄. Evaporation of the ether

gave 18 g of pure alcohol, ee_R> 96%, by enantioselective GC of the

trifluoroacetyl derivative.

stirred for 15 h and then neutralized by addition of CH₃COOH. Water

and ether were added, and the layers were separated. The aqueous layer

was extracted with ether, and the combined ether fractions were washed

with water and brine. The solution was dried over MgSO₄and filtered,

and the solvent was removed to give the product as a light brown liquid

1

(10 g, 90%). H NMR (CDCl₃): δ ) 0.84-1.00 (br m, 5H), 1.02-

1.79 (br m, 6H), 1.13 (d, 3H), 2.14 (s, 6H), 2.36 (q, 1H), 6.11 (m, 1H),

6.39 (dd, 1H), 6.49 (dd, 1H).

(R)-1-Cyclohexylethylmethanesulfonate. (R)-1-Cyclohexylethanol

(5 g, 39 mmol) was dissolved in dry CH₂Cl₂(200 mL). Triethylamine

(8.2 mL, 59 mmol) was added, and the resulting solution was cooled

to 0 °C. Methanesulfonyl chloride was added dropwise over 5-10 min.

The solution was stirred further for 1 h at 0 °C, and then was washed

with water, HCl (10%), NaHCO₃, and brine. The solution was dried

over MgSO₄, and the solvent was removed in vacuo to give a light

C₅H₄(CHMe₂)((S)-1-cyclohexylethyl). In a three neck flask fitted

with an addition funnel and an efficient reflux condenser, LiAlH₄(2.1

g, 54 mmol) was suspended in ether (350 mL). A solution of (S)-1-

cyclohexyl-6,6-dimethylethylfulvene (9.8 g, 45.1 mmol) in ether (100

mL) was added dropwise over 30 min. Stirring for 15 h gave a yellow

mixture that was carefully quenched by dropwise addition of water (5

mL), followed by 7.5 mL of 15% NaOH(aq). This caused a granular

white solid to form. Once H₂evolution had ceased, a further 250 mL

of water was added. The ether layer was then separated and washed

with water. The yellow solution was dried over MgSO₄and filtered,

and the solvent was evaporated to give a light yellow oil (9.3 g, 95%).

The ¹H NMR spectrum is complex, consistent with a mixture of double

bond isomers.

Li[C₅H₃(CHMe₂)((S)-1-cyclohexylethyl)]. In a large swivel frit

assembly, C₅H₄(CHMe₂)((S)-1-cyclohexylethyl) (6.6 g, 30 mmol) was

dissolved in pentane (130 mL), and the solution was cooled to -78

°C. To this was added 1.6 M n-BuLi in hexanes (21 mL, 33 mmol) via

syringe. The solution was warmed to room temperature and stirred for

24 h, during which time the color changed from orange to very pale

yellow. Addition of a small excess of DME (1.3 mL) and further

pentane (50 mL), followed by stirring at low temperature (>1 h), gave

a white precipitate. The mixture was filtered while cold to give an off-

1

yellow oil (95%). H NMR (CDCl₃): δ ) 0.9-1.3 (br m), 1.35 (d,

3H), 1.44-1.84 (br m), 2.96 (s, 3H), 4.55 (q, 1H).

(S)-1-Cyclohexylethylcyclopentadiene. Lithium cyclopentadienide

(6.9 g, 96 mmol) was weighed into a 200 mL Schlenk tube in a nitrogen

glovebox. On a Schlenk line, dry THF was added (60 mL) to give an

amber solution. (R)-1-Cyclohexylethylmethanesulfonate (15.3 g, 74

mmol) was added via cannula as a solution in dry THF (30 mL),

changing the solution color to orange. N,N,N′,N′-Tetramethylethylene-

diamine (29.1 mL, 193 mmol) was then added via syringe. The solution

was heated to reflux under an argon atmosphere for 12 h and then

stirred at room temperature for a further 12 h to effect reaction

completion. During this time, the solution became cloudy and brown

in color. Once the reaction was complete, excess LiCp was quenched

by the slow addition of water. The solution was then treated to washings

with water, 1 M KHSO₄, saturated NaHCO₃, and then brine. The

solution was dried over MgSO₄and filtered, and the solvent was

removed in vacuo to give a dark brown oil. The product could be

purified by Kugel-Ro¨hr distillation at 65-80 °C under high vacuum

1

white powder which was dried under vacuum (6.1 g, 90%). H NMR

(THF-d₈): δ ) 0.86-1.37 (br m, 6H), 1.14 (m, 9H), 1.53-1.81 (br

m, 5H), 2.37 (q, 1H), 2.75 (q, 1H), 5.31 (m, 1H), 5.35 (m, 2H).

{C₅H₃(CHMe₂)((S)-1-cyclohexylethyl)}SiMe₂Cl. In a swivel frit

assembly, an excess of SiMe₂Cl₂(3.4 mL, 28 mmol) was added to a

cold (-78 °C) solution of Li[C₅H₃(CHMe₂)((S)-1-cyclohexylethyl)] (4.2

g, 19 mmol) in THF (80 mL) via syringe. The initially cloudy mixture

was very slowly warmed to room temperature (>5 h) and was stirred

for a total of 24 h to give a clear yellow solution. The solvent was

evaporated, and the product was extracted into petroleum ether followed

by filtration to remove LiCl. Removal of the petroleum ether solvent

1

to give a light yellow oil (70%). The H NMR spectrum is complex,

consistent with a mixture of double bond isomers.

Li[(S)-1-cyclohexylethylcyclopentadienide]. To a solution of (S)-

1-cyclohexylethylcyclopentadiene (1.2 g, 6.63 mmol) in cold diethyl

ether (20 mL, -78 °C) was added 1.6 M n-BuLi (5 mL, 7.3 mmol) in

hexanes. The mixture was allowed to warm to room temperature, and

a white precipitate formed over the next 1 h. The solvent was evaporated

and replaced with petroleum ether. The white product (1.1 g, 92%)

was isolated after filtration and further washing with petroleum ether.

¹H NMR (THF-d₈): δ ) 0.84-1.38 (br m, 6H), 1.15 (d, 3H), 1.42-

1.80 (br m, 5H), 2.43 (q, 1H), 5.5 (m, 4H). ¹³C NMR (THF-d₈): δ )

18.5 (CH₃), 26.1 (CH₂), 26.2 (CH₂), 26.2 (CH₂), 29.8 (CH₂), 30.5 (CH₂),

39.5 (CH), 45.1 (CH), 99.7 (Ar), 100.4 (Ar), 123.0 (Ar).

1

gave a dark yellow oil (4.3 g). The H NMR spectrum is complex,

consistent with a mixture of double bond isomers.

{C₅H₃(CHMe₂)((S)-1-cyclohexylethyl)}SiMe₂(C₅H₅). A large swivel

frit assembly was charged with {C₅H₃(CHMe₂)((S)-1-cyclohexylethyl)}-

SiMe₂Cl (6.7 g, 22 mmol), and THF (100 mL) was added by vacuum

transfer. To this was added a solution of lithium cyclopentadienide (1.9

g, 26 mmol) in THF (50 mL) via cannula. The reaction was stirred for

24 h to give an amber solution. The solvent was evaporated to give a

thick yellow paste. The product was extracted with petroleum ether,

and the solution was filtered to remove LiCl. Removal of petroleum

ether gave an orange oil (7.2 g).

(rac/meso)-Iron Bis(1-cyclohexylethylcyclopentadienide). The iron

complexes were prepared as described above for iron bis(methylneo-

pentylcyclopentadienide). The ¹H NMR spectrum was less informative

in this case, as the rac and meso forms showed essentially the same

set of resonances. Instead, ¹³C NMR data were used to assess

enantiopurity. (S,S)-Iron bis(1-cyclohexylethylcyclopentadienide) was

synthesized from lithium (S)-1-cyclohexylethylcyclopentadienide. ¹³

C

NMR [(S,S), CDCl₃]: δ ) 16.65 (CH₃), 26.72 (CH₂), 26.81 (CH₂),

26.86 (CH₂), 29.76 (CH₂), 30.56 (CH₂), 38.80 (CH), 45.36 (CH), 66.29

(Cp), 67.33 (Cp), 67.61 (Cp), 69.59 (Cp), 93.90 (Cp). A 1:1 mixture

of rac and meso forms was made from lithium (S/R)-1-cyclohexyleth-

ylcyclopentadienide. ¹³C NMR (rac/meso, CDCl₃): δ ) 16.6 (CH₃),

16.7 (CH₃), 26.7 (coincident CH₂’s), 26.8 (coincident CH₂’s), 26.9

(coincident CH₂’s), 29.6 (CH₂), 29.8 (CH₂), 30.58 (CH₂), 30.57 (CH₂),

38.66 (CH), 38.78 (CH), 45.34 (coincident CH’s), 66.38 (CpH), 66.86

(Cp), 67.27 (Cp), 67.42 (coincident Cp’s), 67.70 (Cp), 69.18 (Cp), 69.68

(Cp), 94.05 (Cp), 94.16 (Cp).

Li₂[{C₅H₂(CHMe₂)((S)-1-cyclohexylethyl)}SiMe₂(C₅H₄)]. The dil-

ithio salt was prepared from {C₅H₃(CHMe₂)((S)-1-cyclohexylethyl}-

SiMe₂(C₅H₅) (7.2 g, 21 mmol) and 1.6 M n-BuLi (28 mL), in a manner

similar to that of Li[C₅H₃(CHMe₂)((S)-1-cyclohexylethyl)] (see above)

1

(9.2 g, 80%). The H NMR (THF-d₈) spectrum is consistent with the

presence of two linkage isomers: δ ) 0.34 (m, 12H, SiCH₃’s), 0.9-

1.46 (br m, 12H, Cy), 1.14 (m, 18H, CH₃’s), 1.50-1.96 (br m, 10H,

Cy), 2.39 (q, 1H, CH), 2.77 (q, 1H, CH), 2.84 (q, 1H, CH), 3.21 (q,

1H, CH), 5.63 (m, 4H, Cp), 5.81 (m, 4H, Cp), 5.97 (m, 4H, Cp).

{C₅H₂(CHMe₂)((S)-1-cyclohexylethyl)}(SiMe₂)₂(C₅H₄), “(S)-1-cy-

clohexylethylThp”, “(S)-ceThp”. In a swivel frit assembly, 9.2 g of

Li₂[{C₅H₂(CHMe₂)((S)-1-cyclohexylethyl}SiMe₂(C₅H₄)] was reacted

with SiCl₂Me₂(1.2 equiv) in THF (110 mL). The reaction was stirred

for 15 h to give an amber solution. Removal of solvent gave a beige

C₅H₃CMe₂((S)-1-Cyclohexylethyl), “6,6-dimethyl,(S)-1-cyclohexyl-

ethylfulvene”. To a solution of (S)-1-cyclohexylethylcyclopentadiene

(9 g, 52 mmol) and acetone (5.7 mL, 78 mmol) in MeOH (350 mL)

was added pyrrolidine (6.9 mL, 82 mmol). The brown solution was

9

J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004 8229

A R T I C L E S

Baar et al.

Scheme 11

paste, from which the product was extracted with petroleum ether.

Filtration and evaporation of petroleum ether gave the product as a

thick orange oil (5.1 g). The ¹H NMR spectrum is complex, consistent

with a mixture of double bond isomers.

Li₂[{C₅H(CHMe₂)((S)-1-cyclohexylethyl)}(SiMe₂)₂(C₅H₃)], “Li₂-

[(S)-ceThp]”. The protio form of the ligand, (S)-ceThp (0.98 g, 2.5

mmol), was dissolved in ether (50 mL), and the solution was cooled to

-78 °C. To this was added 1.6 M n-BuLi in hexanes (3.3 mL, 5.2

mmol) via syringe. After 24 h, the solvent was removed and replaced

with petroleum ether. The off-white solid product was recovered by

filtration and washed with petroleum ether (2 × 5 mL) (1 g, 95%). ¹H

NMR (THF-d₈): δ ) 0.22 (s, 3H), 0.24 (s, 3H), 0.28 (s, 3H), 0.29 (s,

3H), 0.84-1.82 (br m, 10H, Cy), 2.01 (br d, 1H, Cy), 2.69 (q, 1H,

CH), 3.14 (q, 1H, CH), 3.74 (m, 1H), 5.89 (s, 1H), 6.07 (br t, 1H),

6.15 (d, 2H). ¹³C NMR (THF-d₈): δ ) 5.26 (SiCH₃), 5.75 (SiCH₃),

5.77 (SiCH₃), 5.87 (SiCH₃), 22.59 (CH₃), 27.27 (CH₃), 27.77 (CH₃),

28.10 (CH₂), 28.22 (CH₂), 28.34 (CH₂), 30.45 (CH₂), 31.66 (CH₂), 33.79

(CH), 41.52 (CH), 46.64 (CH), 102.45, 110.07, 112.90, 113.01, 113.96,

115.26, 120.37, 120.47, 137.76, 139.78.

(S)-ceThpZrCl₂, 8a. Zirconium tetrachloride (0.56 g, 2.5 mmol) and

Li₂[(S)-ceThp] were loaded into a swivel frit assembly in the glovebox.

Toluene (25 mL) was added by vacuum transfer. After being stirred

for 20 h, the mixture was filtered (very slowly) to remove LiCl.

Removal of solvent gave an oily solid. Addition of petroleum ether

(15 mL) gave a powdered product. The mixture was cooled, and the

product was filtered off and washed with cold petroleum ether (2 × 5

mL). Drying under vacuum gave a white powder (0.24 g). The ¹H NMR

(benzene-d₆) spectrum showed that a single diastereomer was iso-

lated: δ ) 0.38 (s, 3H, SiCH₃), 0.44 (s, 3H, SiCH₃), 0.52 (s, 3H,

SiCH₃), 0.54 (s, 3H, SiCH₃), 0.9-1.88 (br m, 10H, Cy), 0.88 (d, 3H,

CH₃), 0.91 (d, 3H, CH₃), 1.39 (d, 3H, CH₃), 2.29 (br t, 1H, Cy), 2.78

(q, 1H, CH), 2.89 (sep, 1H, CH), 6.31 (t, 1H), 6.51 (s, 2H), 6.70 (m,

2H). ¹³C NMR (benzene-d₆): δ ) -1.31 (SiCH₃), -1.14 (SiCH₃), 3.60

(coincident SiCH₃’s), 21.08 (CH₃), 21.24 (CH₃), 26.96 (CH₃), 27.74

(CH₂), 27.75 (CH₂), 28.95 (CH₂), 29.14 (CH₂), 29.78 (CH₂), 34.30 (CH),

40.16 (CH), 109.73, 110.36, 114.67, 115.21, 116.29, 138.12, 138.20,

161.97, 164.59, 165.01. Anal. Calcd for C₂₄H₃₇Si₂ZrCl₂: C, 53.92; H,

6.88. Found: C, 53.84; H, 6.98.

glovebox, under an atmosphere of nitrogen. A Teflon needle valve was

fitted to the Schlenk flask, and it was degassed on the high-vacuum

line. Olefin (1.8 mL) was vacuum transferred from LiAlH₄or CaH₂

into a measuring cylinder and then into the flask containing the MAO/

tetradecane mixture. The olefin was stirred in the MAO suspension

for 1 h to ensure the removal of all traces of moisture. Prior to catalyst

injection, an aliquot was taken via the sidearm of the Schlenk flask

(fitted with a septum) and was immediately quenched with n-butanol.

The flask was placed under an atmosphere of argon, and the catalyst

solution was added via syringe (0.50 mL of a 3.5 × 10^-3M solution

in toluene is typical). The solution usually changed from colorless to

light yellow upon catalyst addition. When the desired conversion was

reached, a final aliquot was taken, and the contents of the flask were

immediately frozen at 77 K and degassed under high vacuum. The

unreacted olefin was vacuum transferred into a receiving flask, and

1

the purity was checked by H NMR. The aliquots were used for the

GC determination of conversion, with tetradecane serving as the internal

standard for integration. The polymer could be isolated by first

quenching the residue with methanol (10 mL) followed by a 1 M

solution of HCl in methanol (10 mL). The methanol was removed in

vacuo, and the tetradecane was then removed by Kugel-Ro¨hr distil-

lation. The polymer was suspended in methanol, filtered, washed, and

dried in vacuo. To more easily isolate the polymer, the polymerization

could be carried out in toluene. The solvent was easily removed,

precluding the need for the Kugel-Ro¨hr distillation.

Enantioassay of Recovered Olefins. Optical purity was determined

by enantioassay of the methyl ester derivative on a GC column with a

chiral stationary phase. A typical procedure for the derivatization

follows. Olefin (200 mg) was added to a 25 mL flask containing CCl₄

(4 mL), CH₃CN (4 mL), H₂O (6 mL), and NaIO₄(4 equiv). RuCl₃‚

3H₂O (10 mg) was added, and the biphasic reaction mixture was stirred

vigorously for >12 h at room temperature, after which time H₂O (5

mL) and CH₂Cl₂(5 mL) were added. The organic layer was separated

(S)-ceThpZrCl₂, 8b. To access the dichloride precatalyst, complex

9b (0.6 g, 1.1 mmol) was treated with excess TMSCl (0.84, 6.6 mmol)

in toluene (20 mL). After 3 h, all volatiles were removed in vacuo to

give the product as an oily yellow solid which solidified over time

under high vaucuum. Unfortunately, the extreme solubility of 8b

precluded further purification. The ¹H NMR spectrum (benzene-d₆)

showed only minor impurities (<5%) and that a single diastereomer

8b was dominant (ca. 10:1 ratio): δ ) 0.36 (s, 3H, SiCH₃), 0.44 (s,

3H, SiCH₃), 0.48 (s, 3H, SiCH₃), 0.64 (s, 3H, SiCH₃), 0.90 (d, 3H,

CH₃), 1.26 (d, 3H, CH₃), 1.34 (d, 3H, CH₃), 2.78 (m, 1H), 2.87 (sep,

1H), 6.3 (t, 1H), 6.42 (s, 1H), 6.72 (m, 2H).

-

and washed with Na₂S₂O₃(aq) (to reduce any I₂and I₃present) and

then brine. The solution was dried over MgSO₄and filtered, and the

solvent was removed in vacuo to give a light yellow oil. The methyl

esters were prepared by boiling a BF₃/MeOH (15%) solution (5 mL)

of the carboxylic acid (∼100 mg) for 5-10 min. The methyl ester was

extracted into petroleum ether (15 mL) which was dried prior to

injection on the GC column with a chiral stationary phase.

(S)-ceThpZr(NMe₂)₂, 9b. A Schlenk tube fitted with a reflux

condenser was charged with the protio ligand, (S)-ceThp (0.42 g, 1

mmol), and Zr(NMe₂)₄(0.28 g, 1 mmol). THF (25 mL) was then added

by vacuum transfer. The solution was heated to 120 °C to ensure gentle

refluxing and was stirred for 24 h, while open to a mercury bubbler.

Next, a strong argon purge was passed over the flask for a further 24

h, after which time vented gas gave only slightly basic readings. The

solvent was removed to give an orange oil which solidified under high

vacuum. Further purification was not possible due to the extreme

solubility of 9b. The ¹H NMR spectrum (benzene-d₆) showed the

presence of mainly diastereomer 9b (ca. 10:1): δ ) 0.52 (s, 3H, SiCH₃),

0.60 (s, 3H, SiCH₃), 0.69 (s, 3H, SiCH₃), 0.80 (s, 3H, SiCH₃), 1.03 (d,

3H, CH₃), 1.23 (d, 3H, CH₃), 1.29 (d, 3H, CH₃), 2.83 (s, 6H, NMe₂),

2.88 (s, 6H, NMe₂), 6.19 (t, 1H), 6.23 (s, 1H), 6.51 (m, 2H).

Polymerization Procedures. A typical polymerization procedure

was carried out as follows. MAO (250 mg) and tetradecane (1.5 mL,

distilled from Na) were weighed into a 10 mL Schlenk tube in the

(S)-3-Methyl-1-pentene. The procedure followed that outlined in

Scheme 11. ^L-Isoleucine was converted to the chloroacid (81%)

according to literature precedents (refs 22a,b), modified slightly by

adding urea (2.5 equiv) after a 1 d reaction time to remove soluble

N₂O₄gas. The chloroacid was reduced to the alcohol (66%) with LiAlH₄

and quenched very carefully with water and 15% NaOH(aq) to give a

granular white precipitate. The alcohol was then transformed to the

iodide following the procedure in ref 22c. DMF was used as the solvent

to increase the yield (72%). In the final step, (S)-3-methyl-1-pentene

was obtained by dehydrohalogenation, following the procedure in ref

22d (56%).

Mandelic Ester Derivatization of Recovered Olefins. The car-

boxylic acid was first prepared from the desired olefin as outlined above.

The acid (63 µL, 0.50 mmol), (S)-methylmandelate (83 mg, 0.50 mmol),

and DMAP (10 mg, 0.08 mmol) were combined in a dry 25 mL flask.

The flask was flushed with argon and cooled to 0 °C, and dicyclo-

hexylcarbodiimide (0.82 mL of a 0.672 M solution in CH₂Cl₂, 0.55

9

8230 J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004

Kinetic Resolution of Chiral R-Olefins

A R T I C L E S

mmol) was added. After the mixture was stirred at 0 °C for 1 h, the

ice bath was removed, and the reaction mixture was stirred for 12 h at

room temperature. The resulting slurry was filtered to remove the

majority of the dicyclohexyl urea and was diluted with diethyl ether

(10 mL). After washings with 1 N KHSO₄, saturated NaHCO₃, and

brine, the organic layer was dried (MgSO₄) and filtered, and the solvent

was removed in vacuo. The oily material was then dissolved in CDCl₃,

Marsh for his assistance in solving the structure of (S)-2 and

Professor Malcolm Green for suggesting the use of thiolate

ligands. C.R.B. and C.J.L. would like to thank NSERC (Canada)

for postdoctoral fellowships.

Supporting Information Available: X-ray data for (S)-2, (S)-

5, (S)-6, and (S)-7 (PDF and CIF). This material is available

free of charge via the Internet at http://pubs.acs.org.

1

and H and ¹³C NMR spectra were taken.

Acknowledgment. This work was been supported by the NSF

(Grant No. CHE-0131180). We would like to thank Richard E.

JA040021J

9

J. AM. CHEM. SOC. VOL. 126, NO. 26, 2004 8231

Article Doi

Kinetic resolution of chiral α-olefins using optically active ansa-zirconocene polymerization catalysts

DOI: 10.1021/ja040021j

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja040021j

Full text of DOI:10.1021/ja040021j

Products guided by the article

R&D Labs maybe for 732985-01-2

Relevant to this article

Hot Product