Stannanes racemization

Stannane 37 lost 50% of its optical activity when allowed to stand as a 0.2 M solution in benzene for 17 days. Addition of A1BN to the solution at 80 °C caused complete racem-ization after 30 min. With added hydroquinone, the benzene solution at 80 °C showed no decrease in rotation after 2 h. It was thus concluded that racemization proceeds by homolysis. 

Racemization of stannane 37 was also observed in donor solvents as shown in Table 2. Although stable for short periods in phenethylamine and acetonitrile, stannane 37 is appreciably racemized in DMSO, HMPA and especially MeOH. Pentacoordinated tin species are likely intermediates. Interestingly, exposure to HMPA leads to significant formation... 

The tin-lithium exchange reactions are thought to proceed with retention of stereochemistry. However, as the stannanes employed in this study were racemic, there is no evidence in support of this pathway. 

Racemic diquinane enone rac-6 was prepared by Piers and Orellana starting from cyclopentenone (Scheme 6) [11]. After the preparation of the heterocuprate from stannane 20, conjugate addition to cyclopentenone in the presence of BF3 Et20 provided carbonyl compound 21. It was expected that conversion of 21 by intramolecular alkylation and subsequent hydrogenation should provide the desired endo-substituted diquinane rac-13. While other hydrogenation methods proved to be rather unselective, reduction in the presence of Wilkinson s catalyst finally resulted in the formation of rac-13 with good facial diastereoselectivity [11]. 

Comparison of the configuration of the stannane with the prodncts of reaction reveals that primary alkyl halides that are not benzyhc or a to a carbonyl react with inversion at the lithium-bearing carbon atom. In the piperidine series, the best data are for the 3-phenylpropyl compound, which was shown to be >99 1 er. In the pyrrolidine series, the er of the analogous compound indicates 21-22% retention and 78-79% inversion of configuration. Activated alkyl halides such as benzyl bromide and teri-butyl bromoacetate afford racemic adducts. In both the pyrrolidine and piperidine series, most carbonyl electrophiles (i.e. carbon dioxide, dimethyl carbonate, methyl chloroformate, pivaloyl chloride, cyclohexanone, acetone and benzaldehyde) react with virtually complete retention of configuration at the lithium-bearing carbon atom. The only exceptions are benzophenone, which affords racemic adduct, and pivaloyl chloride, which shows some inversion. The inversion observed with pivaloyl chloride may be due to partial racemization of the ketone product during work-up. 

The formation of rings with more than seven atoms has unfavorable rates because the addition step is often too slow to allow it to compete successfully with other pathways open to the radical intermediate. In stannane based chemistry for example, premature hydrogen abstraction from the organotin hydride is difficult to avoid. However, Baylis-Hillman adducts 111 derived from enantiopure 1-alkenyl (or alkynyl)-4-azetidinone-2-carbaldehydes are used for the stereoselective and divergent preparation of highly functionalized bicycles 112 and 113 fused to medium-sized heterocycles (Scheme 38) [80, 81]. The Baylis-Hillman reaction using nonracemic protected a-amino aldehydes has been attempted with limited success due to partial racemization of the chiral aldehyde by DABCO after... 

New methods for the preparation of germanes and stannanes reported since 1995 are dealt with in Section n. In Section III, radical chain chemistry involving trialkyltin hydrides is examined. In particular, the synthetic utility of tributyltin hydride will be reviewed, as well as that of other stannanes. Recent advances in the area of asymmetric radical chemistry involving chiral non-racemic stannanes are also included. Section IV details a limited number of examples of non-radical stannane chemistry, while Section V covers recent advances in germane and plumbane chemistry. While we have restricted ourselves largely to the literature since the beginning of 1996, some salient features of earlier work are included when relevant to the discussion. 

Enantioselective free-radical chemistry has benefited through the development of chiral, non-racemic stannanes. Despite having been prepared on a limited number of occasions prior to 1996, the use of chiral stannanes in enantioselective free-radical chemistry was only reported on one occasion31. The field effectively lay dormant until Nanni and Curran reported the preparation and uses of (6,)-4,5-dihydro-4-methyl-3//-dinaphtho[2,l-c l/,2/-ejstannepin (11) prepared from (S)-2, 2/-bis(bromomethyl)-l,l,-binaphthyl (equation 11)32 and a few years later Curran and Gualtieri reported the preparation of C2-symmetric... 

On the other hand, a-alkoxyorganolithiums are not configurationally stable on a macroscopic timescale when they are secondary and allylic or benzylic. For example, despite the known (see section 5.2.1) stereospecificity of the tin-lithium and lithium-tin exchanges of similar compounds, tin-lithium exchange of 110 with rc-BuLi/TMEDA at -78 °C gives an organolithium 111 which has completely racemised after 10 min stannylation returns racemic stannane 112.55 Similarly, 111 racemises rapidly at -70 °C in pentane/cyclohexane in the... 

The second is the non-heterosubstituted lithiated anilide 297. This compound presents a remarkable example because it is configurationally unstable in the presence of TMEDA, but configurationally stable in the presence of (-)-sparteine or its achiral analogue di-n-butylbispidine 299.138 Treatment of the enantiomerically enriched stannane 296 with s-BuLi returns, after 1 h, racemic material in the presence of no diamine additive, or in the presence of TMEDA, indicating a configurationally unstable intermediate on this timescale. However,... 

Further confirmation that the deprotonation determines the ee of the product came from the result shown below. When the intermediate 387 is formed as a racemate, either by deprotonation with s-BuLi-TMEDA or by tin-lithium exchange from a racemic stannane, the products 390 of electrophilic quench are racemic too, even if the electrophile is added in the presence of (-)-sparteine. 

The lithiation and silylation of the amide 71 is enantioselective for a very different reason. When the intermediate organolithium is made from the racemic stannane 72, it still gives the product 74 in good enantioselectivity provided the electrophile reacts in the presence of (-)-sparteine.72 The reaction must therefore be an enantioselective substitution. Furthermore, reaction of the deuterated analogue 75 gives a result which is not consistent with asymmetric deprotonation yield, deuterium incorporation and product ee are all high. 

Additions of Achiral and Racemic Oxygenated Allylic Stannanes to Aldehydes... 

Several racemic a- and /3-oxygenated aldehydes were examined under chelation-con-trolled conditions in which MgBra OEt2 served as the chelating Lewis acid. Reaction of a c -y-OTBS allylic stannane with a-benzyloxybutyraldehyde was highly selective for the syn, syn adduct (Eq. 31). /3-Oxygenated butyraldehydes were somewhat less selective. In these additions, the anti, syn adducts predominated by 4 1 (Table 33). 

Initial efforts in this area involved the addition of BuaSnLi to fratw-crotonaldehyde and conversion of the racemic hydroxy stannane adduct to diastereomeric (-)-menthyloxy-methyl ethers by reaction with (-)-menthyloxymethyl chloride (Eq. 32) [52]. These dia-stereomers could be separated by careful chromatography. They formed diastereomeric anti, (Z) adducts with aldehydes upon heating to 130 °C. The results parallel those seen for the racemic OMOM allylic stannanes (Table 25). Formation of the (Z) double bond in these adducts is attributed to steric interactions between the allylic OR substituent and the adjacent stannane butyl groups in a chair-like transition state as pictured in Eq. (9). The excellent stereoselectivity of these additions is suggestive of a highly ordered transition state. 

Oxygenated (E)-allylic stannanes can also be formed by 1,4-addition of a BuaSn cyanocuprate to enals and in situ trapping of the enolate with TBSCl (Eq. 39) [59]. Tlie corresponding (Z) isomers are not produced in these reactions. Thus far the method has only been applied to the synthesis of racemic stannanes a suitable chiral catalyst for the cuprate addition has not been found. 

The MgBr2-promoted additions are strongly substrate-controlled. As a result it is possible to effect kinetic resolution of racemic y-oxygenated allylic stannanes thereby circumventing the need to employ enantioenriched stannane. The degree of enantio discrimination is somewhat dependent upon the y-oxygen substituent as illustrated by the additions to a threonine-derived aldehyde given in Eq. (46) [66]. 

Further evidence for the racemization premise was obtained from experiments employing (R)-a-methyl-/3-ODPS propanal (Eq, 85) [93]. Addition of the allenylin-dium chloride derived from an enantioenriched (P)-allenyl stannane yielded a 60 40 mixture of anti, anti and anti, syn adducts, not unlike that obtained when racemic alle-nylstannane was used to generate the transient allenylindium chloride. When the (5) aldehyde was employed for this addition a 40 60 mixture of anti, anti and anti, syn adducts was formed. Thus it can be concluded that substrate control (Felkin-Ahn or chelation) is, at best, only modest in these reactions, and that the rate of racemization is only slightly less than the rate of addition. The use of -benzyloxy-a-methyl propa-... 

---