Pure & Appl. Chern., Vol. 70, No. 5, pp. 1077-1082, 1998.

Printed in Great Britain. Q 1998 IUPAC

Kinetic vs. thermodynamic control in hydrozirconation reactions

Peter Wipf,* Hidenori Takahashi, and Nian Zhuang Department of Chemistry, University of Pittsburgh, Pittsburgh, PA 15260, USA
Abstract - Most synthetic applications of organo(chloro)zirconocenes involve the use of terminal zirconocenes derived from hydrozirconation of monosubstituted or symmetrical triple or double bonds. In our studies to effect stereoselective hydrozirconations of internal alkynes and alkenes, we have explored the alkoxy-directed hydrozirconation of homoallylic and bis-homoallylic alcohols. The discovery of a new method to prepare zirconocene hydrides in situ from readily available, air-stable zirconocene dihydride has allowed us to demonstrate for the first time that kinetic hydrozirconation products of internal, non-functionalized alkenes can be formed in preference over isomerized, thermodynamic products.

Hydrozirconation of alkynes and alkenes generates synthetically useful alkenyl- and alkylzirconocene reagents (ref. 1). In addition to stereocontrolled transmetalation reactions of organozirconocenes, we have applied zirconium complexes for cationic cascade reactions (ref. 2). O r most u recent synthetic protocols include the use of alkylzirconocenes in the asymmetric Cu(1)-assisted conjugate addition (eq. 1, ref. 3), the asymmetric synthesis of allylic alcohols via Zr+Zn transmetalation (eq. 2, ref. 4), and oxidative couplings (eq. 3, ref. 5 ) .
1. CpnZr(H)CI, THF

Ph TBSO d N 75% 94% de K

0
, Ph

0
(l)

TBSO1

2. CuBrSMe., ( a x m o'-"~ ~ j '^ i 0 2 h, 21 "C BF3'0Et.j (1 .O equiv) p h d N K o 2 \ / Ph'

1. CpnZr(H)CI,CHzCI2, rt.
2. Me2Zn, -65 O C

eN\
4. PhCHO, -20 'C, 8 h

80%; 95%ee

S
7

,
O

1. CpzZr(H)CI,CHZCIZ, rt. ~ D B 2. Cu(salen)n (1 equiv) rt., 18 h

O 64%

O
8

(3)

Almost exclusively, these processes involve the use of terminal zirconocenes derived from monosubstituted or symmetrical triple or double bonds. A key problem remains the stereoselective hydrozirconation of internal, unsymmetrical alkynes and alkenes. Since isomerization of organozirconocenes is generally faster than hydrozirconation, the thermodynamically most stable, e.g. least sterically crowded isomer is rapidly formed. For 2-octene, for example, this results in the exclusive formation of 1-metalated compound (ref. 6). The use of alkylzirconocenes could be considerably expanded if the thermodynamically favored isomerization to the terminus of an alkyl chain could be suppressed.
1077

1078

P. WIPF e t al.

cationic Zr-chemistry (ref. 1): cascade reactions aldehyde additions epoxide rearrangements Diels-Alder reactions acetal, THF & orthoester formations orthogonal orthoesterprotective group chemistry chiral orthoester

j M-Z, M = Cu, Zn

transmetalationchemistry (ref. 1):

- conjugate additions to enones - acylations - additions to aldehydes - asymmetric C,C-bond formation via stoichiometric auxiliaries or chiral catalysts polyene synthesis: alternatives to Stille couplings & Wiftlg reactions

II

il
t

The problem of rapid chain-migrations of secondary or tertiary organometallics is not unique to zirconocenes. Organoboranes isomerize at temperatures from 70 to 200 "C (ref. 7), and, in the presence of transition metal catalysts, hydroalumination of internal alkenes provides only terminal organoalanes at 70 T (ref. 8). The temperature cutoff for zirconocene migration is not known, and no clear mechanistic understanding has emerged. Since, in general, only starting atkene and terminal zirconocene are observed in NMR, the initial hydrozirconation with the sparingly soluble Cp2Zr(H)C1 must be considerably faster than subsequent p-elimination and further hydrozirconation. It is unlikely that zirconocene hydrochloride diffuses away from the alkene into solution, even though the x-complexes 9 and 1 1 offer very little stabilization due to a lack of back-bonding from &-zirconium to the alkene. However, the isolation of a mixture of alkenes in the hydrozirconation of ( 3 - 5 decene with sub-stoichiometric amounts of zirconocene hydrochloride as well as the isotope composition observed in decane after deuteriozirconation provide evidence that a participation by pathway b cannot be excluded (ref. 9).

(CpnZr(H)CI)

slow

Zr(lV)

8 (18e')
hydrometal. 7
9

hydrometal. C 2 :R p ZCI 12

"\

fast patha

At least for functionalized alkenes, there is encouraging literature evidence for the existence of internal zirconocene products (ref. 1). Styrenes give both secondary benzyl alcohols and primary alcohols upon hydrozirconatiodoxidation (ref. lo), and heteroatoms present in oxazoline 17 and homoallylic alcohol 19 can stabilize internal zirconocenes through complexation (refs. 11,12).
1. CpnZrCldRed-Al (2 equiv)'THF' 25'C,4h; 2 . 0 2 , 2 5 "C
~

15: 1 : 3

0 1998 IUPAC, Pure a n d Applied Chemistry 70, 1077-1082

&

O^;;^""

&

Hydrozirconation reactions
64% D
28% D

1079

1. CpzZrHCl (2equiv), dloxane, 40 OC, 20 h;

I
8% D

2.DzO

-K O
19

CpzZrHCI, THF, r.t.

10-40%

In our own work, we have observed that hydrozirconation of silyl ether 21 followed by copper(1)assisted acylation with benzoyl chloride provided the trans-cyclohexyl ketone 22 as the major isomer (ref. 13). Presumably, axial attack anti to the bulky silyl ether is favored due to steric hindrance, and subsequent isomerizations might be disfavored by the inductive effect of the oxygen substituent. We were interested to determine if related neighboring group participations were effective in suppressing zirconocene isomerization in acyclic alcohols.
TBDPSo

21

6
6

1. CpzZrHCI, 2. CH~CHZCOCI, CuBrSMez THF'400C

+,,r ;
4
~

OTBDPS

OTBDPS

l% 52%

!
CH3CHzCOCI CuBrSMeZ
~ ~

Hydrozirconatioddeuteration of homoallylic ether 27 provided terminally labeled 28 as the sole detectable isomer. The deprotected homoallylic alcohol 29 exerted a more interesting neighboring group effect. The ratio of terminal to internal deuterides was highly depended on solvent and concentration (ref. 14). In THF, a 10 : 1 ratio of terminal deuteride 32 and secondary deuteride 33, possibly derived from the biszirconocenes 30 and 31,was isolated from a standard 0.2 M reaction mixture. At a higher concentration, the relative amount of terminal to internal zirconocene was decreased to 3.5 : 1. A similar ratio was obtained in CH2C12 at room temperature. Lowering the reaction temperature is likely to decrease the solubility of Schwartz reagent and thus its concentration in the reaction mixture, and at 0 "C the ratio of 32 to 33 increased to 5.8 : 1. The ratio observed after 25 min in CH2C12 was largely unchanged after 18 h, and since toluene provided a >25 : 1 ratio of products after the same reaction time, we conclude that the kinetic barrier toward isomerization is unusually high due to the coordinating alkoxide ligand. No combination of solvent, reaction time, and temperature provided the internal deuteride 33 as the major product.
OTlPS Ph
27
1. Cp Zr(H)CI (2 equiv) sofvent, rt 1. Cp2Zr(H)Ci THF, rt

OTIPS PhA
20

2.D20

OZrCpzCl
+

Ph
29

PhLZrCpEl 31

solvent THF THF CHzCIz CHzCIz CHzCIz toluene

temp. rt rt rt rt 0% rt

time 25mln 25 min 25 min 18 h 25 min 18 h

conc. 0.2M 0.5M 0.2M 0.2M 0.2 M 0.2 M

32 : 33 10: 1 3.5:1

2.Dz0

1
+
Ph
33

L Ph
32

3.6:1 3.2:1 5.8:1

>25: 1

0 1998 IUPAC, Pure and Applied Chemistry70, 1077-1082

1080

P. WlPF e t a / .

Since the in siru prepared zirconocene alkoxide failed to shift the regioselectivity of the intermolecular hydrozirconation of the homoallylic alcohol in favor of the secondary zirconocene, we investigated the use of zirconocene dihydride to induce an intramolecularmetalation pathway. In CH2C12 and THF we obtained indeed a 1 : 1 mixture of deuterides 32 and 33, presumably formed via the cyclic zirconocenes 34 and 35,respectively. Longer reaction times in THF and toluene provided ca. 2 : 1 mixtures.

solvent THF THF CHzCl2 toluene

temp. rt rt

rt rt

time 4h 18 h 4h 18 h

conc. 0.2M 0.2 M 0.2M 0.2 M

32 : 33 i:1 2.4: 1 1:1 2 :1

Hydrozirconation of the potassium alkoxide did not provide any further improvements in the formation of the internal deuteride; the ratio of 32 to 33 ranged from 3 : 1 in THF to 8 : 1 in CH2C12. Generation of the lithium alkoxide by deprotonation of 29 with BuLi at -78 C, followed by addition of zirconocene hydrochloride and warming to room temperature, provided for the first time the internal deuteride as the major isomer. Comparable results were obtained in the hydrozirconation of the tertiary homoallylic alcohol 36.Use of 2 equivalents of zirconocene hydrochloride provided the terminal deuteride 37 in >25 : 1 selectivity. In contrast, this regioselectivity was shifted in favor of the internal deuteride 38 with the in situ prepared lithium alkoxide.
1. KH; CpnZr(H)CI solvent, rt PL h 29 2.D20 32

33

THF: 3 : 1 CHzCIz: 8 : 1 toluene: 5 : 1

1. BuLi; CpzZr(H)CI THF, -78 C to rt 29 2. D20

32

33

1 : 2

?H

1. Cp2Zr(H)CI(2 equiv) THF, rt

PH ?

1. BuLi; Cp2Zr(H)CI THF. -78 O to rt C 36 2. D20

37

38

1 : 2

The modest levels of regioselectivity for the internal secondary zirconocene obtained in the directed hydrozirconation of homoallylic alcohols 29 and 36 could not be extended to the disubstituted alkene 39 and the bis-homoallylic alcohols 41,43and 45.Only terminal deuterides were obtained under a range of reaction conditions.
1. BuLi; CpzZr(H)CI THF, -78 C to rt

Ph 39 OTlPS Ph 41

2.D20

PhL

40
OTlPS PhL 42 D

k 2.D20

1, CpZZr(H)CI THF, rt 2. D20

1. BuLi; CpnZr(H)CI THF, -78 C to rt Ph

43

PhL 44

0 1998 IUPAC, Pure and Applied Chemistry 70, 1077-1082

Hydrozirconation reactions
1. BuLi; Cp2Zr(H)CI THF, -78 O to tt C ph&
45

1081

2.D2O

46

O r early attempts to control the regioselectivity of the hydrozirconation of internal, u unfunctionalized alkenes were unsuccessful. Hydrozirconation-acylation of cyclopentene 47 led to a mixture of five isomeric ketones (ref. 13). The isolation of the methylene ketone 54 implies that at the reaction temperature necessary to effect hydrozirconation of the trisubstituted alkene even the tertiary zirconocene 5 0 is formed transiently.

47

48

51

2. PhCOCl CuBrSMe2_ 85%

&,s~ph

dph
0
+

&Ph

54

52 (52+53)154
D

53

12 : 1

A decrease in the temperature necessary for hydrozirconation would likely inhibit any undesired thermodynamic equilibration and allow the isolation of kinetic products. However, the insolubility of zirconocene hydrochloride in organic solvents effectively limits the lower temperature range for hydrozirconation of alkynes and terminal alkenes at 0 c to room temperature. Internal alkenes are less reactive and require higher temperatures such as refluxing THF for reasonable reaction rates. Since chloride is an excellent bridging ligand and thus facilitates zirconocene oligomerization, we investigated next the rate of hydrozirconation of alkenes with derivatives Cp2Zr(H)X, where X=Br, I, and OTf. These studies were greatly facilitated by our discovery that treatment of the readily available, highly polymeric and air-stable Cp2ZrH2 with halogen sources such as CH2Br2, MEM-I, and others, converts the dihydride in situ into the desired monohydrides, which are much more effective hydrozirconating agents than Cp2ZrH2 (ref. 15). In spite of the heterogeneous nature of this reaction, only stoichiometric amounts of halogen sources are required. Trapping of the terminal zirconocene derived from 1-octene with Br2 or I2 led to high yields of 1halooctanes. Methylene bromide and TMS-triflate were among the most effective reagents in this process. In the absence of an electrophilic partial quenching of the zirconocene dihydride, no monohalogenated product was isolated.

1
readily available, airstable, polymeric

561

1. 1-Octene, halogen source THF, 0 C ICpnZrHn) 2. X2 (X = Br, I), 0 C

* x58

halogen source CH2Br2: CH212: MEM-Br: MEM-I: Mel:

12:

MeOTf: TMSOTf

yield 90% (rt) . . 70% 82% (tt) 75% 42% 49% 63% 93%

Application of these new preparations for modified Schwartz reagents to (Q-2-octene provided, for the first time, the kinetic, internal trapping products. In the presence of 0.5 equiv of CH2Br2, trapping of a mixture of 1.0 equiv of Cp2ZrH2 and 0.83 equiv of 2-octene in THF with Br2 provided only traces of terminal bromide in addition to the major compound, dibromide 60, which was formed from unreacted 2octene. A slightly better ratio was obtained with 1 equiv of CH212, however, the use of 0.5 equiv of TMStriflate resulted in the formation of the internal trapping product 2-bromooctane as the major isomer. With (E)4-octene, in siru prepared zirconocene hydrotriflate led to a 1 : 1.8 ratio of 1-bromo- to 4-bromooctane.

0 1998 IUPAC, Pure and Applied Chemistry70, 1077-1082

1082

P. WlPF eta/.

Interestingly, independently prepared Cp2Zr(H)OTf (ref. 16) did not provide any kinetically controlled internal trapping product, and no 2- or 3-brominated octanes were detected in any reaction mixture.
1.2-Octene,CH2Br2 THF, 0-20 "C
ICpzZrHz)

2. Br2,O "C 1 . 2-0ctene, CH212 THF, 0-20 "C

D

B 59

1 : 23 (CpzZrHd 59
60

2. Biz, 0 "C 1.2-0ctene,TMSOTf THF, 0-20 "C

D

1 : 15

ICpzZrHd

59

2. Brz, 0 "C

1 : 1.2

ICP2.w2J

1. 4-Octene,TMSOTf THF, 0-20 "C

59

2. Br2, 0 O C
1 : 1.8

T Br 62

In conclusion, lithium &oxide-directed hydrozirconation or in situ prepared zirconocene hydrotriflate allow for the formation of internal, secondary zirconocenes in preference over the primary isomers obtained by migration along the carbon chain. With unfunctionalized internal alkenes, the thermodynamically preferred isomerization to the terminal position can be partially suppressed at 0 "C by the use of zirconocene dihydride and TMS-triflate, and electrophilic trapping of kinetically formed internal zirconocenes has been achieved. Further studies will be directed at low-temperature (-78 "C) hydrozirconation protocols that should allow for the exclusive formation of kinetic hydrozirconation products.

ACKNOWLEDGMENT: This work was supported by grants from the National Science Foundation and the National Institutes of Health.

1. P. Wipf and H. Jahn. Tetrahedron 52, 12853 (1996). 2. P. Wipf, W. Xu, H. Takahashi, H. Jahn and P. D. G. Coish. Pure Appl. Chem. 69, 639 (1997). 3. P. Wipf and H. Takahashi. J. Chem. SOC. Chem. Commun. 2675 (1996). 4. P. Wipf and S. Ribe, manuscript in preparation. 5. P. Wipf and W. Xu, manuscript in preparation. 6. D. W. Hart and J. Schwartz. J. Am. Chem. SOC.96, 8115 (1974). 7. (a) H. C. Brown and G. Zweifel, J. Am. Chem. SOC. 1967, 89, 561. (b) H. C. Brown. Hydroboration, pp.136-149. W. A. Benjamin, New York (1962). 8. E. C. Ashby and S. A. Noding. J. Org. Chem. 44, 4364 (1979). 9. U. Annby, S. Karlsson, S. Gronowitz, A. Hallberg, J. Alvhall and R. Svenson. Acra Chem. Scand. 47, 425 (1993). 10. T. Gibson. Organometallics 6, 918 (1987). 11. J. Alvhall, S. Gronowitz and A. Hallberg. Chem. Scr. 25, 393 (1985). 12. H. Takaya, M. Yamakawa and K. Mashima, K. J. Chem. SOC.Chem. Commun. 1283 (1983). 13. P. Wipf and W. Xu. Synlett 718 (1992) 14. In this and all subsequent deuteration experiments, the internal deuteride 33 was isolated as a 1 : 1 to 1 : 1 0 mixture of diastereomers. No attempts to assign the relative stereochemistry of the major isomers were made. 15. Precedence for the use of Cp2ZrH2 as a precursor for zirconocene monohydrides was found in the conversion of Cp2ZrH2 in CH2C12, reported to form Cp2ZrHC1 which is slowly transformed further into Cp2ZrC12: P. C. Wailes and H. Weigold. J. Organomet. Chem. 24, 405 (1970). 16. G. A. Luinstra, U. Rief and M. H. Prosenc. Organometallics 14, 1551 (1995). In the solid state, this compound has been shown to be a hydride-bridged dimer.

0 1998 IUPAC, Pure and Applied Chemistry70,1077-1082