ebook img

Regioselectivity in the Ene Reaction of Singlet Oxygen with Alkenes PDF

21 Pages·2000·0.33 MB·English
by  
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Regioselectivity in the Ene Reaction of Singlet Oxygen with Alkenes

TETRAHEDRON Pergamon Tetrahedron56(2000)1595–1615 Tetrahedron Report Number 519 Regioselectivity in the Ene Reaction of Singlet Oxygen with Alkenes Manolis Stratakis and Michael Orfanopoulos* DepartmentofChemistry,UniversityofCrete,Iraklion71409,Greece DedicatedtoProfessorC.S.Footeontheoccasionofhis65thbirthday Received4October1999 Contents 1. Introduction 1595 2. Regioselectivity ofthe Ene Reaction 1596 2.1. ‘cis Effect’ with trisubstituted alkenesand enol ethers 1596 2.2. syn Stereoselectivity with allylic alcohols 1597 2.3. Anti ‘cis effect’ selectivity 1597 2.4. Phenyl group directing synselectivity 1598 2.5. Regioselectivity with cis and transdisubstituted alkenes. The large group non-bonded effect 1598 2.6. Regioselectivity with geminal dimethyl and diethyl trisubstituted alkenes 1599 2.7. Geminal regioselectivity with respect to abulky allylic substituent1600 2.8. The role ofnon-bonded interactions between the cis alkyl substituentsduringthe formation of the new doublebond in the ene adduct 1603 2.9. Geminal selectivity with vinylsilanesand vinylstannanes 1603 2.10. Regioselectivity with allylic silanes, stannanesand germanes 1604 2.11. Regioselectivity with olefinsbearing an electron withdrawing group at the (cid:97)-position 1605 2.12. Regioselectivity with olefinsbearing an electron withdrawing group at the (cid:98)-position, and with homoallylic alcohols. The role of electronic repulsions 1605 2.13. Regioselectivity with twisted 1,3-dienes.Vinylic versusallylic hydrogen reactivity 1607 2.14. Regioselectivity in the autooxidation offullerene adducts 1607 2.15. Regioselectivity ofthe ene photooxygenation reaction within zeolites 1607 3. Solvent Effects in the Ene Regioselectivity 1608 4. Synthetic Transformations ofRegioselective and Stereoselective Ene Reactions 1609 5. Conclusions 16101.Introduction The chemistry of singlet molecular oxygen ( 1953,13 and it was revived after the pioneering work of Foote14 in the early 1960s. Ever since, amongst all singlet 1O ) has oxygen reactions, it has attracted major interest due to its 2 receivedremarkableattentionbychemistsnotonlybecause controversial reaction mechanism and fascinating regio- of its interesting mechanistic and synthetic aspects1–4 but chemistry and stereochemistry. also because of its large environmental5–7 and biomedical significance.8–10 Among the peculiar types of reaction involving 1O wi2th],oerngea,naincdsubstrates are: [4(cid:49) theenereactionproceedsthroughaconcertedorastepwise 2 [2(cid:49)2] additions to alkenes and oxidation of organic mechanism. The initially proposed synchronous pathway15 compounds of sulfur, phosphorus, etc. The ene reac- was challenged by biradical,16 zwitterionic,17 or perepox- tion 2,3,11,12 was originally discovered by Schenck in ide11 intermediates (Scheme 1). Kinetic isotope effects in the photooxygenation of tetrasubstituted,18 trisubstituted19 * Corresponding author. Tel.: (cid:49)30-81-393630; fax: (cid:49)30-81-393601; and cis alkenes20 supported the irreversible formation of e-mail:[email protected] an intermediate perepoxide, while for alkenes,21,22 a 0040–4020/00/$-seefrontmatter(cid:113)2000ElsevierScienceLtd.Allrightsreserved. PII: S0040-4020(99)00950-3 1596 M.Stratakis,M.Orfanopoulos/Tetrahedron56(2000)1595–1615 2,3-dihydro-(cid:103)-pyran and suggested the importance of an initial orbital interaction between the incoming oxygen and the enol ether. Stephenson47 suggested that an inter- action between the LUMO orbital of the oxygen and the HOMO orbital of the olefin stabilizes the transition state of perepoxide formation. This is particularly true for the case ofcis-2-butenewherethebondedorbitalsoftheallylic hydrogencontributetotheHOMOorbitaloftheolefin,thus generatingasystemsimilartothe(cid:99) stateofbutadiene.The 3 interaction of this system with the (cid:112)(cid:112) LUMO orbital of the Scheme1. oxygenisfavorable.Forthetransisomerthecontributionof the bonded orbitals ofthe allylic C–H bonds to the HOMO partial equilibration of the intermediate with the reactants orbital of the olefin is approximately half, thus making the wasproposed.Consistent withthe perepoxideintermediate interactionwiththeoxygenLUMOlessfavorable. is also the observation that the ene reaction proceeds as a highly suprafacial process, in which the conformational For a cycloalkene48 it was proposed that there is a correla- arrangement of the allylic hydrogen controls the stereo- tion between the orientation of the allylic hydrogen in the chemistry of the allylic hydroperoxide produced.23 groundstate anditsreactivity.Theallylichydrogensat the axialpositionaremorereactive,becausetheorbitaloverlap Trapping of the intermediate with sulfoxides,24 phos- betweenoxygenandallylichydrogensisoptimuminsucha phites,25 and sulfenates or sulfinate esters,26 in the photo- conformation. oxygenation of adamandylidenoadamantane, and theoreticalcalculations27–31aswell,supporttheperepoxide In 1981 Houk and coworkers,49 using STO-EG semi- asthemostpossibleintermediate.Manyauthorsconsideran empirical calculations, suggested that for acyclic alkenes intermediate exciplex32–34 instead of the polar perepoxide. differences in regioselectivity are related to the different Sincethegeometryoftheintermediateiswelldefinedfrom energiesrequiredforanallylichydrogentoadoptaperpen- isotope effects and stereochemical studies, if an inter- dicular arrangement to the olefinic plane. For example, in mediate exciplex is formed, its geometrical features must cis-2-butene the rotational barrier ofan allylic hydrogen to resemble thoseof aperepoxide. adopt a perpendicular alignment to the double bond plane from an eclipsed conformation was calculated to be In the photooxygenation of electron rich alkenes like 1.06kcal/mol, while for the trans isomer it is higher by enamines,35 enol ethers,36 enol esters37 and dienes,38–40 1kcal/mol. This assumption was in accordance with the thereisapartitioning betweenperepoxideand zwitterionic experimental observation that cis alkenes react about intermediates, affording mixtures of ene and cycloaddition 10–20 times faster than their transisomers.50 products, whose relative abundance depends on solvent polarity and temperature. Schuster and coworkers51,52 measured the activation para- meters(cid:68)H‡and(cid:68)S‡intheenereactionof1O withaseries 2 of alkenes and suggested that there is a type of ‘hydrogen 2. Regioselectivityof theEne Reaction bonding’ interaction in the rate-limiting step ofperepoxide formationbetweentheallylichydrogensandtheintermedi- 2.1.‘cis Effect’ with trisubstituted alkenes andenol ate. For example, in the reaction of cis-2-butene activation ethers entropy (cid:68)S‡ is less negative by 10e.u., than that of the correspondingtransisomer,whiletheactivation enthalpies In the reaction of 1O with trisubstituted alkenes41,42 and (cid:68)H‡ are very similar. The considerable difference in the 2 enol ethers,43,44 the more reactive side of the double bond activationentropieswasattributedtothefactthattransition isthemoresubstituted.Thissurprisingselectivityisreferred states in the cases of cis olefins require more organization to as the ‘cis effect’ (Table 1). Both experimental and because of the simultaneous interaction of the incoming theoretical chemists have offered explanations for the ‘cis oxygenwithtwoallylichydrogensasseenintheaccompa- effect’. Bartlett and Frimer45,46 studied the primary and nying scheme. On the other hand for trans alkenes, this secondary isotope effects in the photooxygenation of interaction occurs with only one allylic hydrogen. Table1.Numbersindicatepercentageofdoublebondformationintheene adducts M.Stratakis,M.Orfanopoulos/Tetrahedron56(2000)1595–1615 1597 Table2.syn/antiSelectivityinthephotooxygenationofallylicalcohols9-E crowded side to form PE, the oxygen interacts simul- I and9-Z taneously with the hydroxyl and one allylic hydrogen. Thisinteractionstabilizesthetransitionstate.Polarsolvents interact with the hydroxyl through hydrogen bonding and reduce its ability to interact with the oxygen. Thus, the activation energy of TS‡ increases significantly and leads I Substrate Solvent syn/antiSelectivity to a reversal of selectivity which is now controlled by steric factors. The observed stereoselectivity in methanol 9-E CCl4 75:25 (syn/anti(cid:136)33:67) is a typical example of anti ‘cis effect’ 9-E Benzene 73:27 behavior,similartothatobservedinthestericallyanalogous 9-E CHCN 41:59 9-E MeO3H 33:67 substrates 10-E and 10-Z(see accompanying Section 2.3). 9-Z CCl 72:28 4 9-Z CH3CN 40:60 2.3. Anti ‘cis effect’ selectivity The proposal that there is a stabilizing effect by the simul- Itcan beconcluded that mostofthe proposedexplanations taneousinteractionof1O withtwoallylichydrogensonthe 2 are consonant with the existence of an interaction between same side of the olefin during the formation of the perep- theincoming oxygenandtwoallylic hydrogensthathighly oxide, suffices to explain the rare cases where anti ‘cis stabilizesthe transition state ofperepoxide formation. effect’ selectivity has been observed in the photooxygena- tionoftheseriesofacyclictrisubstitutedalkenes59shownin Table 3. 2.2. synStereoselectivity with allylic alcohols Recently,Adamandcoworkers,inordertoexplainthehigh For these substrates, the less substituted side of the double diastereoselectivity (d.e. (cid:44)90%) observed in the photo- bond is the more reactive. In the case of alkenes 10-E and oxygenation of chiral allylic alcohols53,54 and amines,55,56 10-Z,forexample,formationoftheperepoxideonthemore proposed that singlet oxygen forms an exciplex with the substituted side of the olefin is not favorable because of steric repulsions between the oxygen and the two alkyl allylic alcohol, in which the oxygen coordinates to the groups. Also, during the formation of the intermediate, hydroxyl. This coordination, in combination with the 1,3- oxygen is interacting with only one allylic hydrogen on allylicstrain,provides,innon-polarsolvents,highselectiv- ityforthreoallylichydroperoxides.Thiseffectwasfoundto both sides of the double bond. Thus, the stabilizing inter- control the syn/anti stereoselectivity57,58 in the photo- actionofoxygenwithtwoallylichydrogensinthesameside oxygenation of allylic alcohols 9-E and 9-Z. In non-polar of the double bond is absent in these substrates. This is solvents (CCl , benzene) the syn methyl group is more evidentinScheme3wheretheintermediatesinthereaction reactive (‘cis4effect’ type selectivity), while in polar of 1O2with 10 and 4are found. solvents,suchasacetonitrileormethanol,whicharecapable of hydrogen bonding with the hydroxyl, regioselectivity is Ingeneral,theanti‘ciseffect’selectivityisrelated:(a)tothe degree of crowdedness on the more substituted side of the almost reversed (Table 2). olefin;(b)tothenon-bondedinteractionsduringtheforma- tion of the new double bond; and (c) most importantly, to Examination of the possible transition states in Scheme 2 helps to understand the observed stereoselectivity. For thelackofsimultaneousinteractionoftheincomingoxygen with twoallylic hydrogens. allylic alcohols 9-E and 9-Z, singlet oxygen is capable of interactingwithonlyoneallylichydrogenoneither sideof theolefin.Thelackofstereoselectivity,whethertheantior In the literature there are a few examples of cyclic olefins whichexhibitanti‘ciseffect’selectivity.Typicalexamples the syn methyl is labeled, indicates that perepoxide forma- tionoccursintheratelimitingstep.InTS‡(appliedto9-E), are substrates 1360 and 1461 in Table 3. Examination using I wheretheelectrophileapproachestheolefinfromthemore molecularmodelsrevealssevereconformationalrestrictions for the tertiary allylic hydrogens on the more substituted side of the double bond in adopting the necessary orienta- tion either for abstraction or for ‘positive’ interaction with the incoming oxygen. Table3.Numbersindicatepercentageofdoublebondformationintheene adducts Scheme2. 1598 M.Stratakis,M.Orfanopoulos/Tetrahedron56(2000)1595–1615 Table 4. syn/anti Stereoselectivity in the ene reaction of 1O with (cid:98),(cid:98)- 2 dimethylstyrene Solvent syn/antiSelectivitya CCl 56:44 4 CHCN 71:29 3 MeOH 82:18 asyn(cid:136)H-abstraction,anti(cid:136)D-abstraction. Scheme3. 2.4.Phenylgroupdirecting synselectivity thetransitionstatefortheformationoftheperepoxide,justas thephenylgroupdoesinthecaseofsubstrate15. Surprisingly, in the reaction of 1O with (cid:98),(cid:98)-dimethyl- 2 styrene (15),62 the ene products which are formed, apart from dioxetane and diendoperoxides,63,64 exhibit an unexpected solvent-dependent syn selectivity, although the intermediate in the more substituted side of the olefin can interactwithonlyoneallylichydrogen.Thusthesynmethyl groupof15is56%reactiveinCCl ,whileinthemorepolar 4 methanol the reactivity becomes 82%(Table 4). 2.5. Regioselectivity withcis andtransdisubstituted alkenes. The largegroupnon-bondedeffect Thisstereoselectivityisunexpected,consideringthatinthe more substituted side of the olefin oxygen is capable of In the reaction of singlet oxygen with non-symmetrical cis interacting with only one allylic hydrogen. Thus, anti ‘cis disubstituted alkenes, an unexpected regioselectivity was effect’ selectivity would be expected. found.65 The allylic hydrogens next to the large alkyl sub- stituentaremorereactivethanthosenexttothesmallgroup. Intermolecular isotope effects revealed that in the photo- Some examplesare shownin Table 5. oxygenation of 15 formation of the perepoxide is the rate determining step.62 Although the incoming oxygen is In substrate 19 where L(cid:136)t-butyl and s(cid:136)methyl, the double oriented towards the more substituted side of the olefin bond is formed next to the bulkier t-butyl group 75% and (TS‡), it interacts with only one allylic hydrogen; never- I only25%nexttothemethyl.ByretainingtheL(cid:136)t-butyland theless this transition state affords the major product increasingthesizeofthealkylgroupontheothersideofthe (Scheme 4). Thus, the stabilization must arise from the olefin from methyl to phenyl (substrates 20 and 21) a interaction of the perepoxide with the phenyl group. In decrease in the percentage of hydrogen abstraction from TS‡, which leads to perepoxide PE, the benzylic carbon is I I the side of the t-butyl is observed. For substrate 21 where partially behaving as a carbocation which is stabilized by thecompetingsubstituentsareat-butylgroupandaphenyl resonance with the phenyl group. Interaction of the nega- group,thereisasmallpreferencefordoublebondformation tively charged oxygen of perepoxide with the phenyl ring, next tothe bulkier t-butyl.Thisdemonstrates that conjuga- whichhaslostpartofitselectronicdensity,resultsinsignifi- tion of the phenyl group with the forming double bond is cant stabilization. By increasing the polarity of the solvent notsufficientlydevelopedinthetransitionstatetoovercome thisstabilizationincreasesbecausethecationiccharacterat the benzylic carbon becomes more profound, thus making TS‡ more dominant due to solvation. Transition state TS‡ I II which leads to the anti ene product (D-abstraction) is less favored, being stabilized by only a single interaction between the oxygen and the one allylic hydrogen. The same rationale can also be used to explain the ‘cis effect’behaviorwhichwasobservedearlierinthephotoox- ygenation of trisubstituted enol ethers43,44 (Section 2.1). Oxygen prefers to form the intermediate on the more crowded side, where it interacts with the electronegative methoxysubstituent. In the transition state for the formation oftheperepoxidethecarbonnexttomethoxyhassignificant carbocationiccharacterandisstabilizedbythemethoxygroup through resonance. An interaction of the negatively charged oxygen of the perepoxide and the methoxy group stabilizes Scheme4. M.Stratakis,M.Orfanopoulos/Tetrahedron56(2000)1595–1615 1599 Table5.Numbersindicatepercentageofdoublebondformationintheeneadducts the control of regioselectivity by steric effects (see also The photooxygenations of the trans alkenes proceed substrate 23). In the extreme case where the large sub- approximately20timesmoreslowlythanthoseofthecorre- stituent is the triphenylmethyl group (substrate 25) the spondingcisisomers.49Thefewexampleswhichhavebeen reaction is highly regioselective. The regioselectivity trend studied,66 are summarized in Table 6. The regioselectivity is altered when there is only one allylic hydrogen on the does not differ at all from that of the corresponding cis morebulkysideofthedoublebond(e.g.substrate26)dueto isomers found in Table 5. Considering that the formation theunfavorableconformationofthemethinehydrogenatom of perepoxide in the photooxygenation of trans alkenes is to adopt a perpendicular geometry to the former double reversible,22the1,3-non-bondedinteractionsintheproduct bond, since considerable 1,3-allylic strain is developing. forming transition states between the oxygen and the alkyl In addition, there are steric repulsions between the cis substituentsappeartocontroltheregioselectivityinasimi- alkylgroupsontheformingdoublebondduringthemethine lar fashion to thoseshownin Scheme 5. hydrogen atom abstraction step (two alkyl groups adopt a cis geometryin the double bond ofthe adduct). 2.6. Regioselectivitywith geminal dimethyl anddiethyl trisubstitutedalkenes The regioselectivity was explained65 by examining the possible transition states leading to the allylic hydro- The regioselectivity trend in the photooxygenation of peroxides (Scheme 5). In transition state TS‡, which leads geminal dimethyl and diethyl trisubstituted olefins is similar II to the minor product, the repulsive 1,3-non-bonded inter- tothatobservedincisortransdisubstitutedalkenes,withthe actions between the oxygen atom and the large group L, allylic hydrogens next to the bulky alkyl substituent being which exist in the intermediate perepoxide, still remain morereactive.Theresults66aresummarizedinTable7. (the large group non-bonded effect). This transition state is expected to be higher in energy than TS‡ where these Inarepresentativeexample(substrate34),whereL(cid:136)t-butyl I repulsiveinteractionsarereleased.Thenon-bondedinterac- and R(cid:136)H, the methylene hydrogens are statistically 6–7 tionsdevelopinginTS‡ aresimilartothoseappearinginthe times more reactive than the hydrogen atoms of the two II axial conformers ofmonosubstituted cyclohexanes. geminal methyl groups. Starting from 2-methyl-2-pentene Scheme5. 1600 M.Stratakis,M.Orfanopoulos/Tetrahedron56(2000)1595–1615 Table6.Numbersindicatepercentageofdoublebondformationintheene adducts where R(cid:136)H and L(cid:136)H and increasing the size of the L substituent, after ethyl substitution, a significant increase in the percentage of the anti-Markovnikov product is Scheme6. observed. When the L substituent is the very bulky tri- phenylmethyl group (substrate 38) the reaction becomes 2.7. Geminal regioselectivity with respect toa bulky highlyregioselective(92:8). allylic substituent Thetransitionstatesinthehydrogenabstractionstep,asseen In1988ClennanandChen71reportedthatreplacementofan earlier in Scheme 5, help to explain the observed change in allylic hydrogen in tetramethylethylene with a series of regioselectivity.Inatransitionstatewherethereisastrength- functional groups (sulfides, sulfoxides, sulfones, cyano, eningoftheC–Obondonthetertiarycarbon,areleaseofthe halides etc.) affords substrates which, upon photooxygena- 1,3-non-bondedinteractionsbetweentheoxygenatomandthe tion,giveeneproductswithasurprisinggeminalselectivity L group with respect to the intermediate perepoxide occurs. with respect to the allylic functionality (Table8). Therefore,thistransitionstateisexpectedtobelowerinenergy thanatransitionstatewherenon-bondedinteractionsexist. Three possible explanations were proposed to account for the observed regioselectivity: When the geminal dimethyl group is replaced by geminal diethylfunctionality(substrates39and40),ahigherdegree (a) Electronic repulsions between the lone pairs of the ofregioselectivityisobserved.Thetransitionstatesleading heteroatoms and the negatively charged oxygen of the totheminorproductareevenlessfavored,notonlybecause perepoxide. Such a repulsive interaction would favor ofthe1,3-non-bondedinteractionsbetweenthegroupLand the intermediate on the other side of the double bond the oxygen, but also because of the repulsions developing withrespecttothefunctionality,whichgiveseneproduct whentheethylgroup(Scheme6)adoptsacisgeometrywith with geminal hydrogen abstraction. amethylgroupinthenewlyformingdoublebond(seealso (b)Anchimericassistancebytheallylicsubstituentlead- discussion in Section 2.8). We would predict that if the ing to regioselective opening of the perepoxide inter- geminal dimethyl group is replaced by a geminal di-n- mediate by an S2 mechanism. However, this could N propyl or a longer di-n-alkyl group, the reaction would be explain the regioselectivity only in those cases where regiospecific ((cid:46)97%selectivity). the allylic substituent issulfur. For a series of geminal dimethyl trisubstituted alkenes the same trend in regioselectivity was recognized earlier by Thomas69 and Rautenstrauch.70 By examining molecular models and assuming that the reaction is concerted, Thomas and coworkers proposed that the difference in regioselectivitycouldbeduetothedifferentconformations and stericrepulsionsin the transition states. Table7.Numbersindicatepercentageofdoublebondformationintheeneadducts.aRef.67,bRef.68 M.Stratakis,M.Orfanopoulos/Tetrahedron56(2000)1595–1615 1601 Table8.Numbersindicatepercentageofdoublebondformationintheeneadducts (c) Different barriers to rotation of the methyl groups of tothemajoreneproduct.TransitionstateTS‡ isnotfavored II the substrate. A similar explanation has been reported due to the retention of non-bonded interactions. For the earlier by Houk and coworkers49 in order to explain the same reason, TS‡ from perepoxide II is also not favored. IV syn preference of singlet oxygen ene reactions with Bothofthesetwotransitionstatesleadtotheminorproduct. trisubstituted alkenes. Finally, in TS‡ although there is relief of 1,3-non-bonded III interactionsbetweentheoxygenatomandthet-butylgroup, In order to examine the factors affecting geminal regio- two alkyl substituents (t-butyl and methyl) adopt a cis selectivity, the groups of Orfanopoulos72 and Clennan73 conformation in the newly forming double bond, which is independentlysynthesizedaseriesofalkylsubstitutedtetra- highly unfavorable, and no product derived from this methylethylenes at the allylic position (Table 9). Several pathway isformed. 1-neopentyl substituted cycloalkenes74 were also studied. The photooxidation products revealed that the general Geminalselectivityisalsoobservedincaseswherethebulky scheme of regioselectivity is exactly the same as in the substituent is a vinylic substituent.66,72,74 The examples are case where the allylic substituents are heteroatoms (Table shown in Table 10. The explanation of the geminal selec- 8),thusindicatingthatthereisnosignificantcontributionof tivity observed for these substrates is exactly the same as electronic effects (explanations a and b), for the observed previously proposed in the case of the alkenes bearing a regioselectivity.ItisobviousthatasthesizeoftheLsubsti- large allylic substituent (transition states of Scheme 7). tuent increases the degree of geminal selectivity is higher. For example, in substrate 56 the non-bonded interactions Thesameprinciplealsoappliesinthecasesof1-neopentyl betweenthet-butylgroupandtheoxygenatomintheinter- substituted cycloalkenes with a surprisingly high degree of mediateweakentheC–Obondnexttothebulkysubstituent regioselectivity. The geminal selectivity varies from 71% intheproductformingtransitionstate.Insubstrates58,the for 52, to (cid:46)97%for 51. presenceofvariouselectronicsubstituentsattheparaposi- tion66,76 does not alter the geminal regioselectivity in any Examination of the possible product forming transition way, despite the fact that the photooxygenation of the states (Scheme 7) in the reaction of singlet oxygen with p-MeO derivative proceeds much faster than that of the 2,3,5,5-tetramethyl-2-hexene (49) helps to understand the p-CF substituted styrene. This finding is expected, con- 3 geminal selectivity. Each of the two perepoxides I and II, sideringthatformationoftheperepoxideistheratelimiting canleadthroughtransitionstatesTS‡ toproducts.InTS‡, step.11,22 For these styrene-type substrates, although the I–IV I relief of 1,3-repulsions with respect to the intermediate active stereosize of the phenyl group is smaller than the perepoxide I is observed. Indeed, this transition state leads t-butyl group in 56, the geminal selectivity is higher. This Table9.Numbersindicatepercentageofdoublebondformationintheeneadducts 1602 M.Stratakis,M.Orfanopoulos/Tetrahedron56(2000)1595–1615 Scheme7. maybeattributedtothefactthattheformingdoublebondis barrier to rotation (0.85kcal/mol) with respect to the other in conjugation with the phenyl group. Furthermore, in two methyls. substrate61,the1,3-repulsions(seealsoScheme7)between the oxygen atom and the allylic t-butyl group and those with the vinylic t-butyl are expected to be similar. Thus, a moderate regioselectivity of55:45 isobserved. Forcycloalkenes,thevariationinthepercentageofgeminal regioselectivity is probably due to the combination of both effects: (a) the different conformational arrangements of theallylichydrogensintheringsystemsasproposedearlier by Rautenstrauch et al.48 and (b) the 1,3-non-bonded interactions between the large groupsand the oxygen. Inordertoshedsomelightonthisregioselectivityproblem, Clennanandcoworkers,73basedonthetheoreticalmodelof Orfanopoulos and coworkers calculated the rotational Houk (MM2 calculations),49 found that for some alkyl barriers for the allylic methyls, with the HF-STO-3G substituted tetramethylethylenes (TME-R) there is a sur- method, in a series of trisubstituted alkenes and compared prising correlation between the rotational barriers and themtotheobservedexperimental regioselectivity.75Some their reactivity towards 1O . This model predicts that the results are shown in Table 11. For alkenes 10-E and 10-Z, 2 lower the barrier to rotation the higher the hydrogen thesynmethylgroupshavelowerrotationalbarriersthanthe abstraction from the perepoxide. Some numerical values corresponding anti ones by 0.6kcal/mol. This is in the in kcal/mol are depicted below for substrates 54 and 49. opposite direction to the proposed theoretical model. In 2,3,5,5-tetramethyl-2-hexene (49) for example, the However, for alkene 4 there is a correlation between rota- more reactive methyl group (geminal, 78%) has the lowest tional barriers and ene reactivity. Alkenes 11 and 55 also Table10.Numbersindicatepercentageofdoublebondformationintheeneadducts M.Stratakis,M.Orfanopoulos/Tetrahedron56(2000)1595–1615 1603 Table 11. (a) Percentage of double bond formation in the ene adducts. ethyls, during the formation of the minor allylic hydroper- (b)Barrierstorotationofthemethylgroups(kcal/mol) oxide. In the photooxygenation of 62 repulsions exist between an ethyl and a methyl group, whereas in 63 a b between an ethyl and a propyl. By increasing the size of 76 1.63 thealkylgrouptophenylorisopropyl(substrates64and65 24 1.11 respectively),a reverse regioselectivity trend isfound. For these specific cases, there are two equally competing 74 1.63 factorsintheproductformingtransitionstateswhichaffect 26 1.11 the regioselectivity (Scheme 8): (a) the 1,3-non-bonded interactions between the oxygen atom and the phenyl or isopropyl substituents in the transition state for hydrogen 14 1.64 86 0.40 abstraction from the methyl group, and (b) repulsions between the cis alkyls, due to the formation of the new double bond by methylene hydrogen atom abstraction. 66 2.91 Thestabilizingcontributionoftheconjugationofthedouble 34 0.91 bond with the phenyl group (alkene 64) does not affect regioselectivityaspreviouslynoticed inthephotooxygena- tion of 22(Table5). 64 1.45 36 1.22 Similarargumentscanexplainthechangeinregioselectivity ofsubstrates67and66whencomparedto56and16,respec- tively.In56forexample,replacementofthegeminalmeth- yl group with an ethyl group causes a significant drop in demonstrate impressively that there is no correlation geminal selectivity. It is instructive to note in general that between ene reactivity and rotational barriers. Again in the reactivity of the methylene hydrogens is significantly these examples the more reactive trans methyl group is decreased when a geminal methyl group is present, as theonewhichshowsthehigherrotationalbarrierincontrast clearlydemonstratedbytheexamplespresentedinTable13. tothe predictions ofthe proposedtheoretical model. These results indicate that there is not always a correlation For instance, from comparison of substrates 4 and 69, it is between the reactivity of the methyl groups and their rota- obviousthatthereactivityoftheantimethylgroupincreases tional barriers.More importantly, accordingto the Curtin– substantially when a methyl group is replaced by an ethyl Hammett principle, the rotational barriers are irrelevant to leading to anti ‘cis effect’ selectivity.59 the product distribution since their values are too small (0.5–2kcal/mol) compared to the activation energies of 2.9. Geminalselectivity with vinylsilanes and the reactions(6–13kcal/mol). vinylstannanes 2.8.Theroleofnon-bondedinteractionsbetweenthecis The reaction of 1O with vinyl silanes78–81 and stannanes82 2 alkyl substituentsduringthe formation of thenew hasbeen studied extensively. Apartfrom small amountsof double bondin theene adduct (cid:97),(cid:98)-unsaturatedketonesproducedbydecomposition ofthe labile (cid:97)-hydroperoxy allylic silanes and stannanes, the ene Inthesecondstepoftheenereaction,duringtheformation reaction exhibits a high degree of geminal regioselectivity oftheadductdoublebond,iftwoalkylsubstituentsadopta with respect to the trimethylsilyl or stannyl groups cis orientation the transition state leading to the adduct is (Table14). highlyunfavorable.Thiscanbedemonstratedbyexamining theregioselectivityintheseriesofsubstrates77ofTable12. Although steric factors (bulky vinyl trimethylsilyl sub- stitutent) may be responsible for the geminal selectivity For olefin 62, the methylenic hydrogens are more reactive (for the carbon analogues see Section 2.7), the selectivity thanthehydrogensofthemethyls(ratio70:30).Byincreas- was attributed to electronic factors. For the silanes, it was ingthesizeofthealkylgroup(substrate63),themethylenic proposed that antibonding interactions of the C–Si (cid:115) bond hydrogensbecomemuchlessreactive((cid:44)5%).Thisremark- and the lone pair of the non-terminal oxygen of the perep- able change in regioselectivity can be attributed to higher oxideareresponsibleforthehighdegreeofregioselectivity repulsionsbetweenthepropylgroupswhencomparedtothe (Scheme9).Duetothehigherenergyofthe(cid:115)orbitalofthe Table12.Numbersindicatepercentageofdoublebondformationintheeneadducts 1604 M.Stratakis,M.Orfanopoulos/Tetrahedron56(2000)1595–1615 Scheme9. Scheme8. Table13.Numbersindicatepercentageofdoublebondformationintheeneadducts C–Snbond,theinteractionwiththelonepairsoftheoxygen To explain the regioselectivity trend in the photooxygena- is expected to be higher in stannanes. Yet, the geminal tion of allylic silanes, it was proposed that an interaction regioselectivities observed in stannanes are somewhat between the negatively charged oxygen of the perepoxide lower than those of silanes. However, since the C–Sn and the silicon controls the abstraction of allylic hydrogen bond is longer than the C–Si bond and the non-bonded atoms (Scheme 10). interactions between oxygen and the vinylic substituent are lower in stannanes,the degree of geminal selectivity is The formation of cis ene adducts87 was attributed either to consequently higher for silanesthan for stannanes. the intermediacy of a zwitterionic intermediate, or to a non-concerted pathway. A perepoxide intermediate was considered unlikely. 2.10.Regioselectivitywithallylic silanes,stannanesand germanes Photooxygenation of allylstannanes affords, apart from the ene derived allylic hydroperoxides, stannyldioxolanes.88–90 Allylsilanes83–85 exhibit a different regioselectivity pattern Thedioxolaneswereproposedtoderivethroughringopen- whencomparedtotheircarbonanalogues(Section2.6).The ning of the intermediate perepoxide with 1,2-migration of major ene product arises from hydrogen abstraction from the stannyl group to form azwitterion. theothersideofthedoublebond,withrespecttothesilicon substituent.Also,inadditiontotrans,significantamountsof ene productsareformed withcisgeometryaroundthe new doublebond.Itisinterestingtocomparetheregioselectivity in the photooxygenation of alkyl substituted alkenes with allylicsilanes.ThesizeoftheC–Sibondis1.89A˚,almost 25%longerthantheC–Cbond.AlthoughtheMe Sigroup 3 is more bulky than the t-butyl, the 1,3-non-bonded inter- actionsmustbe farlessimportant, since the methyl groups lie away from the oxygen of the perepoxide. For example, replacementofthet-butylwiththetrimethylsilylgroupfor The regioselectivity of the ene pathway is surprising, with substrates 34 and 18 versus 74 and 75 (Table 15) leads to the trialkyltin group directing allylic hydrogen abstraction very different regioselectivity, as well as stereoselectivity next to it, in contrast to the regioselectivity observed in (mixtures of E and Z allylic hydroperoxides). The few allylic silanes,91 asseen in substrates78 and 79 ofScheme studies of photooxygenation of some cyclic allyl- germanes,86 indicate a similar regioselectivity trend to the silicon analogues, however, the directing effect of Table15.Numbersindicatepercentageofdoublebondformationintheene adducts germanium islesspronounced. Table14.Numbersindicatepercentageofdoublebondformationintheene adducts

Description:
photooxygenation ene reaction, it was interesting to assess the effect of the hydroxyl at the more remote homoallylic position. The regioselectivity in the reaction
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.