9 Reaction mechanisms Part (ii) Pericyclic reactions By IAN D. CUNNINGHAM Department of Chemistry University of Surrey Guildford UK GU2 5XH 1 Introduction Recent advances in our understanding of the mechanisms for pericyclic reactions have been summarised in reviews which cover quantum mechanical methods for modelling pericyclic reactions,1 metal-assisted cycloaddition reactions,2 radical cation cycloaddition reactions,3 frontier molecular orbital (FMO) theory4 and solvent e§ects on Claisen rearrangement reactions.5 2 Cycloaddition reactions Theoretical studies Pericyclic cycloaddition reactions with well established mechanisms are often used as benchmarks to test proposed advances in computational methodology.6 These studies provide justification for the routine use of computational modelling to rationalise the rates and stereochemical course of pericyclic reactions.For example a frontier molecular orbital analysis nicely rationalizes the syn regioselectivity observed for the reaction of penta-1,3-diene with juglone 1 (Scheme 1) by a consideration of secondary orbital interactions (SOI). The di§erences in the FMO coe¶cients for the diene determined by semi-empirical and ab initio calculations are large for non-bonding secondary sites (C2\C3) but minimal for the primary bonding sites. BF 3 coordinates to the Cl carbonyl of 1 and the observation that addition of this complex to penta-1,3- diene results in a reversal to anti regioselectivity can also be rationalised by frontier molecular orbital theory where calculation shows that there is a reversal in the magnitude of the LUMO coe¶cients at the secondary sites (C1[C4).Al(OR) 3 with bulky R groups (e.g. 2,6-diphenylphenyl) coordinates to the C4 carbonyl group but does not lead to the predicted syn regioselectivity. Steric e§ects are invoked to explain the failure to observe the predicted reaction products. Finally the results of an AM1/ab initio study suggest an extremely asynchronous approach to the transition state for the reaction catalysed by BH 3 so that the structure of this transition state approaches that expected for a zwitterionic reaction mechanism.7 The still controversial question of the explanation for the observation of endo selectivity in Diels–Alder reactions has been considered for the butadiene–cyclo- 273 O O OH O O OH O O OH syn 3 2 1 4 1 anti Scheme 1 Si Si Si Si Scheme 2 propene cycloaddition reaction using ab initio computational methods [HF CAS QCISD(T)].While it is clear that the endo selectivity is the result of a C–H · · · p interaction the authors argue whether this should be defined as a true secondary orbital e§ect (SOE).8 A related DFT study also identifies the C–H· · · p interaction as dominant in determining the endo selectivity for this reaction.9 The interplay between calculations and experiments is often fruitful. A prediction from density function theory of a novel and facile [4]2] cycloaddition of cyclohexa- 1,3-diene onto a Si(100)-2]1 surface (Scheme 2)10 has been supported by vibrational spectroscopic evidence for formation of a buta-1,3-diene adduct obtained in a separate study.11 Novel dienes and dienophiles There were many reports in the 1997 literature of cycloaddition reactions of novel dienes and dienophiles but relatively few of these provide significant new insights into the mechanisms of these reactions.However the reaction of alkynyldihaloboranes 2 (R\Bun Ph But Si(Pr*) 3 ; X\Cl Br) with 2-substituted dienes is of interest. Although some of the products of these reactions are consistent with a pericyclic [p44 ]p24] reaction mechanism,* the observation of anomalous regioselectivity solvent e§ects and reaction by-products is cited as evidence to support alternative reaction mechanisms including one that proceeds via the [4]3] adduct 3 (R@\Me But Ph).12 B R' X X R R BX2 2 – + 3 *The distinction of H.-W. Fru� hauf (ref. 2) between the mechanistic notation e.g. [p44 ]p24] and the topological or product-based notation e.g.[4]2] is used in this report. 274 I.D. Cunningham EtO N O N NO2 N –O O- O N O N NO2 O– O N –O OEt 5 + + 4 + + Scheme 3 NPh R O S R' R R' N O S Ph S N O R R' 7 6 Ph Scheme 4 The [4]2] cycloaddition of ethyl vinyl ether to the C––C–N––O part of 4,6-dinitrobenzofuran (DNBF) 4 has been studied (Scheme 3). The authors propose that the endo monoadduct 5 and the diadduct (not shown) form by consecutive Diels–Alder reactions in which there is inverse electron demand.13 The results of a theoretical density functional theory study of the related cycloaddition reaction between nitroethene and ethenol support a concerted pericyclic reaction mechanism but it was not possible to clearly identify the transition state of the reaction by a one-step mechanism in an AM1 study of the cycloaddition reaction between 4 and methyl vinyl ether.14 A detailed experimental study of the Diels–Alder reactions of the ambident dienophile 6 has shown that dienes with C- and Z-type substituents R and/or R@ show high C––S vs.C–– C chemoselectivity while those with X-type substituents (alkyl) show little chemoselectivity (Scheme 4). High regioselectivity as shown by 7 in Scheme 4 is observed for the cycloaddition reaction of the C–– S functional group and the stereoselectivity of the reaction generally favours formation of the endo product. The regioselectivity for the reaction of the thiocarbonyl group cannot be adequately modelled by FMO/AM1 theory because the coe¶cients for theLUMO C�S are similar at carbon and sulfur. However an AM1 analysis of transition state structures provides a better prediction for the regioselectivity and overall stereoselectivity for this reaction.These calculations model the experimental trend in chemoselectivities but they overestimate the di§erence between the activation barriers for the cycloaddition reactions of the C–– C and C–– S centres.15 The P–– S groups in Lawesson’s reagent 8 (Ar\4-MeOPh) show high 1,3- dipolarophilicity towards 9 and a value of k 2 \2.21dm3 mol~1 s~1 has been determined for the reaction in acetone at 37 °C. By comparison the P–– S group of 8 is 1500- and 1.4-fold more reactive as a dipolarophile respectively than representative C––C and C––S groups.16 275 Reaction mechanisms Part (ii) Pericyclic reactions Ar P S S N N N Ph N Ph Ph Ph 9 + – 8 2 Kinetics and mechanism of [4]2] and [2]2] cycloaddition reactions The qualitative categorization of pericyclic reactions as proceeding with normal or inverse electron demand is now well established.Sauer has published the results of a systematic study of the kinetics of the cycloaddition of polyhalogenated cyclopentadienes to triazolinedione maleimide styrene acrylate and other dienophiles. The dependence of the rate constant for these reactions on the Hammett substituent constant r is used to define the electron demand for the reaction. A surprising number of reactions give V-shaped Hammett plots in which the Hammett reaction constant q changes from positive for electron-withdrawing substituents to negative for electrondonating substituents. This result is indicative of neutral electron demand.17 N N N N R R N N N N R 12 11 10 The di¶culty in establishing a scale to define electron demand is illustrated by the results of a study of the cycloaddition of tetradehydrodianthracene 10 with its relatively high C–– C HOMO to the electron-deficient diene 11 (R\H Me CO 2 Me CF 3 ).The rate of this reaction decreases as R becomes more electron-withdrawing which is the opposite of the change expected for a reaction with inverse electron demand. It has been proposed instead that the rates for these reactions are sensitive to the steric bulk of R. This conclusion is supported by the results of PM3 calculations and by the observation of the expected inverse electron demand trend when the substituentsR separated from the reaction center in 12 (R\MeO H Cl).18 The question of how to distinguish ‘true’ concerted pericyclic and stepwise reaction mechanisms continues to tax the ingenuity of chemists particularly when the reactions involve novel reactants.Arenediazonium ions 13 with electron-withdrawing substituents at the phenyl ring undergo a [4]2] cycloaddition reaction with certain dienes such as (E)-penta-1,3-diene to form 14 and 15 and the aromatised analogs of these compounds (Scheme 5). A stepwise mechanism has been considered for this reaction involving initial electrophilic addition of the terminal nitrogen to the diene followed by cyclisation because of the observation that dienes such as 4-methylpenta-1,3-diene yield azo coupling products such as 16 derived from the intermediate allylic cation 17. Aconcerted mechanism is favoured for the formal [4]2] addition reaction because of 276 I.D.Cunningham N N Ar N N Ar N N Ar N N Ar OMe Ar N N MeOH 13 + 16 17 + 15 14 Scheme 5 O O Me Me R R Me R R Me CO2 18 + D Scheme 6 the failure to observe the products of trapping of an allylic cation intermediate by nucleophiles and because of the low reactivity of the Z isomer of penta-1,3-diene. The preference for formation of the regioisomer 14 as the product of reaction of highly electron-deficient 13 is attributed to the stabilization of a transition state with some cationic character at C4 (C6 of product).19 The concept of asynchronicity in pericyclic reactions is well established and there have been several studies of the borderline between concerted ansynchronous and stepwise reaction mechanisms. The thermal decarboxylation of b-lactone (oxetan-2- one) 18 is formally a [2]2] cycloreversion reaction (Scheme 6).The results of ab initio (RHF and MP2) calculations predict a concerted reaction mechanism through a highly asynchronous planar transition state with considerable biradical character. There is good agreement between the values of E!#5 determined by experiment and calculation for unsubstituted 18 (R\H) and fluoro-substituted 18 (R\F) which is 725-fold less reactive than the unsubstituted compound. The large 18 kJ mol~1 change in DG8 observed for a change in solvent from cyclohexane to acetonitrile is consistent with a change to a strongly zwitterionic transition state for a stepwise reaction mechanism in solution. The authors conclude that the high asynchronicity for the gas phase reaction provides a mechanism to circumvent the Woodward–Ho§man rules which forbid the ‘planar’ [p24 ]p24]-like transition state and they invoke a spectrum of transition states with structures that are strongly influenced by solvent.20 The authors of a second theoretical study prefer a concerted reaction mechanism through a polar asynchronous transition state for the gas phase reaction whose geometry corresponds to that expected for an asynchronous [p24 ]p24] reaction mechanism.Although this reaction is formally forbidden for 4-p electron systems the authors suggest that it can be explained using frontier molecular orbital theory.21 277 Reaction mechanisms Part (ii) Pericyclic reactions O N (CH2) n Ph Ph NO (CH2) n 20 + 19 k2 k–1 Scheme 7 Solvent e§ects The relative rates of cycloaddition of polyhalogenated cyclopentadienes to 4-phenyl- 1,2,4-triazoline-3,5-dione in solution were found to show a good correlation with the solvent parameters AN (acceptor number) DN (donor number) and SB (solvent basicity) but not with E5.The small e§ect of changing solvent on the reaction rate and the tendency of reactivity to increase with increasing AN are characteristic of a pericyclic cycloaddition reaction mechanism.22 Pericyclic reactions have been described as the ‘showpiece’ in the study of organic transformations in aqueous media.23 In an attempt to separate the contributions of hydrophobic and hydrogen bonding interactions to the large rate accelerations observed for pericyclic reactions in water Engberts has studied the cycloreversion of the naphthoquinone–cyclopentadiene adduct and has argued that the small value of DV8 observed for this reaction requires a reactant-like transition state and a minimal hydrophobic e§ect on the rate of the reaction.The observed 13 kJ mol~1 decrease in DG8 on going from hexane to water was therefore taken as a measure of the stabilization of this transition state by hydrogen bonding and it was concluded that the same stabilization will be observed in the reverse cycloaddition reaction. Application of this type of analysis to a related Diels–Alder reaction where both hydrophobic and hydrogen bonding interactions contribute to the rate acceleration in water allows calculation of a value of ca. 8 kJ mol~1 for transition state stabilization by a hydrophobic e§ect.24 A rather unusual rate-retarding e§ect of water has been observed for the heteroretro- Diels–Alder reaction of 19 (n\1).This e§ect has been quantified for the reaction of 20 (n\2 Scheme 7) by estimating the value of k ~1 for its cycloreversion reaction from the experimental values for k 2 and the overall equilibrium constant K\k ~1 /k 2 . The decrease in k ~1 observed for the reaction in water compared to organic solvents is attributed to the decrease in the hydrogen bonding interactions to 19 proposed for the change from reactant to transition state. This transition state is reactant-like but shows some resemblance to nitrosobenzene which an infrared spectroscopy study shows to form weaker hydrogen bonds than 19.23 The possibility that Diels–Alder reactions in supercritical CO 2 proceed with unusual regio- or stereoselectivity has attracted attention in recent years.However a report published in 1997 casts doubt on earlier findings of a reversal in the selectivity for the cycloaddition reaction of isoprene with methyl acrylate near the critical point of CO 2 . The authors suggest that incomplete sampling or an unknown complex phase behavior may have resulted in erroneous experimental results.25 Catalysis There were many reports in 1997 of Lewis acid-catalysed Diels–Alder reactions which 278 I.D. Cunningham Al O O Cl O MeO O MeO OH HO Ph Ph 21 22 Mg2+ O O O O MeO MeO OMe OMe H Ar H Ar OH OH HO HO 23 24 Ar = 4-chlorophenyl involve the lowering of theLUMOof the dieneophile. An interesting example involves autoinduction in the cycloaddition of methyl acrylate to cyclopentadiene in the presence of (S)-VAPOL 21 and Et 2 AlCl. It is proposed that the enantioselectivity for formation of the product complex (e.g.22) is higher than when the reaction proceeds in the presence of (S)-VAPOL alone. The further increase in enantioselectivity to[99% observed upon addition of bulky achiral ligands such as di-tert-butyl 2,2-dimethylmalonate suggests that the enhancement of enantioselectivity (e.g. in 22) is due mainly to an increase in the steric bulk around the chiral catalyst.26 Catalysis of Diels–Alder reactions by solid materials is of current interest. Convincing evidence has been obtained that catalysis and stereoselectivity in the cycloaddition reaction of acrolein and cyclohexa-1,3-diene occurs within the cavities of solid 23. Non-catalytic cycloaddition of methyl acrylate to cyclohexa-1,3-diene is also facilitated by solid 23 and an X-ray crystallographic analysis has identified the presence of reactants within the cavities of this solid.The enhanced reactivity of these bound substrates was attributed to a proximity e§ect even though it was noted that the reagents embedded in the solid are not aligned towards the classical Diels–Alder transition state. The lack of catalysis of the reaction of methyl acrylate compared to acrolein was attributed to the stronger binding of the cycloadduct of the former within the cavities of 23.27 Studies of the rate enhancement of cycloaddition reactions caused by alkali and alkali earth metal ions continue. Evidence has been presented in a kinetic study of the MgClO 4 -catalysed Diels–Alder reaction of cyclopentadiene and 2-(4-chlorobenzylidene) malonic acid dimethyl ester that the rate-limiting step for the reaction involves addition of the diene to a dienophile–Mg2` complex 24.The value of DS8[[167 J mol~1K~1 is typical of an uncatalysed Diels–Alder reaction and it was therefore suggested that the enhanced reactivity of 24 is due mainly to a tendency of the metal ion to lower DH8.28 Cumulenes The question of whether cycloadditions of allenes to alkenes proceed by biradical or concerted pericyclic [2]2] reaction mechanisms continues to be debated. The ob- 279 Reaction mechanisms Part (ii) Pericyclic reactions O N O H H R NR O Ph D H H NR O • Ph D 26 + 25 LUMOC1'-C2' 2' 1' 4 LUMO HOMOC4-C1' O O Scheme 8 served stereoselectivity for the cycloaddition reaction of 25 with a range of alkenes and alkynes provides evidence for the concerted reaction mechanism (Scheme 8).Two frontier molecular orbital-based schemes of which 26 appears to be the more credible are discussed and the observed high reactivity of 25 for cycloaddition is attributed to a lowering of the energy of the p* C1{–C2{ LUMOby an interaction with the rather distant p* N–R (R\-SO 2 Tol or -COPh) orbital. The e§ect of the increase in the energy of the pC4–C1{ HOMO by the adjacent nitrogen does not seem to have been considered.29 1,3-Dipolar cycloaddition reactions Reactive ylides 27 (RA\H Ph) have been generated by laser flash photolysis reactions of 28 in acetonitrile and values of 106–1010dm3 mol~1 s~1 have been determined for the second-order rate constants for the reaction of 27 with dipolarophiles 29 (R R@\H CO 2 Me CN Scheme 9). The values of log k for the reaction of 27 (RA\H) with acrylonitrile show a linear dependence on both the values of p0 for the aryl ring substituents and of *E(LUMO$*10-!301)*-% [HOMO:-*$%) determined by AM1 calculations.30 Asymmetric cycloaddition reactions Asymmetric cycloaddition reactions have a profound significance in synthetic organic chemistry which has led inevitably to an interest in their mechanism. The mechanism for the Sharpless alkene asymmetric dihydroxylation reaction has long been the subject of contentious debate with the controversy centering around whether the key-step for the reaction is a [3]2] or a [2]2] cycloaddition reaction (Scheme 10). A comparison of the 13C and the 2H kinetic isotope e§ects determined by experiment with the calculated isotope e§ects for [3]2] and [2]2] reactions favours the [3]2] mechanism via a symmetrical transition state.This is o§ered to counter the recent evidence supporting the [2]2] reaction mechanism which is also discussed in this paper.31 280 I.D. Cunningham N Ar R¢¢ H N R R R' R' R'' H Ar R2C CR'2 Ar C N C H R'' 27 + – 29 hn 28 Scheme 9 Os O O O O L Os O O O L O Os O O O L L = amine ligand + [2 + 2] [3 + 2] O Scheme 10 The results of mixed density function–molecular mechanics calculations also favour a [3]2] cycloaddition reaction. However it is proposed that the dihydroxylation reaction of styrene proceeds with the initial formation of a complex that is stabilised by a p–p interaction between the ligand L and substrate.32 The reactive species in the Ti–TADDOLate-catalysed (TADDOL\a,a,a@,a@-tetraaryl- 1,3-dioxolane-4,5-dimethanol) asymmetric cycloaddition reaction has been the subject of speculation.Jorgenson has modelled the cycloaddition reaction of 3-[(E)- but-2@-enoyl]-1,3-oxazolidin-2-one to benzylidinephenylamine N-oxide (R\Pr*O) and proposed the structure 30 for the transition state for an exo-approach of reactants. The decrease in exo selectivity with increasing bulk of the ligand X (Cl Br OTf) is cited as evidence that these cycloaddition reactions are subject to axial ligand control.33 Cl Cl Cl Cl L S M Ti O O N O R R R X C H N Ph Ph O steric interaction 30 31 Cl Cl The ‘Inside-Alkoxy’ model for stereoselectivity has been proposed to hold for the inverse electron demand cycloaddition of alkenes to hexachlorocyclopentadiene. The diastereoselectivities observed by experiment and from the results of ab initio calculations are modelled by the transition state 31 where the largest group L is anti to the incoming diene and the medium-sized alkoxy groupMis positioned ‘inside’ in order to minimise the electrostatic and steric repulsion with the chlorine atoms of hexachlorocyclopentadiene.34 281 Reaction mechanisms Part (ii) Pericyclic reactions N O R R' N O R R' N C O R' R 2 32 R = H Me CF3 R' = H Me CF3 F H Scheme 11 3 Electrocyclic reactions Spin-coupled theory has been shown to be a useful computational approach to model electrocyclic reactions. This form of modern valence bond theory yields orbitals which are highly localised and therefore more compatible with the traditional view of organic reactivity. Application of spin-coupled theory to model the ring-opening reaction of cyclobutene to form butadiene results in the prediction that most of the changes in the geometry observed along the intrinsic reaction coordinate occur after the reaction transition state and that this trend is most pronounced for the high-energy disrotatory process.35 The mechanism for ring-opening of the cyclobutene radical cation has received relatively little experimental attention.Wiest has reported a theoretical study using UHF MP2 QCISD(T)/QCISD and DFT methods. The results of these calculations mostly favour a concerted reaction mechanism with the lone electron localised on one methylene group and the positive charge distributed across the other three carbon atoms. Steric factors are invoked to rationalise the calculated preference for conrotatory ring opening.36 The above report highlights concerns that the observation of a stereoselective reaction may not be a suitable criterion for a concerted pericyclic reaction mechanism in cases where steric e§ects favour a stereoselective stepwise reaction.Dolbier et al. have examined the cyclisation reaction of isocyanate 32 (Scheme 11) in an attempt to determine the significance of torquoselectivity in a 6p system where the steric e§ects on product formation are minimal (Scheme 11). Values of E!#5 and DS8 are typically 105 kJ mol~1 and [54 J mol~1K~1 respectively and are consistent with a 6p pericyclic electrocyclisation reaction mechanism. However the results of calculations at the MP2/6-31G*//RHF/6-31G* level predict a transition state geometry in which the terminal p-orbital on C2 is twisted and oriented for overlap with a C–– Orather than an N––C p-orbital and the overall movement of electrons corresponds to a pseudopericyclic process.However the variation in the relative rate constants for the cyclisation of the Z and E isomers of 32 is still interpreted within the framework of the torquoselectivity for formation of a product-like transition state.37 The cyclisation of 33 has been studied in superacidic media (Scheme 12). The linear dependence of the rate constant on the acidity function H 0 and the lack of a plateau observed at values ofH 0 above the pK! for monoprotonated 33 (ca.[6) indicates that the reaction proceeds via the O,O-diprotonated intermediate 34. The results of 282 I.D. Cunningham O R¢ R O R¢ R R R¢ O H H + + 34 33 R = H CF3 Me R' = H Ph 4-CF3Ph 4-MePh Me Et Scheme 12 HB HA HAHB 35 repulsion rotation 1 2 Scheme 13 deuterium exchange studies exclude a Friedel–Crafts-type reaction of the protonated alkene and the relatively small e§ect of substituents R@ on the rate of the reaction is consistent with a 4p electrocyclisation reaction.The phenylallyl carbocation that is generated by the initial protonation of the carbonyl group is apparently unreactive towards electrocyclization and additional activation by a second protonation of the substrate to give a strongly activating –H 2 O` substituent is required for observation of a cyclisation reaction.38 An ab initio study of substituent e§ects on a prototype 4p pentadienyl cation electrocyclisation reaction has appeared.39 4 Sigmatropic rearrangements [1,n] Shifts Several computational studies have been reported which attempt to provide a rationalisation for novel experimental findings.The rearrangement of vinylpropene to cyclopentene (Scheme 13) shows some characteristics of both stepwise and concerted reaction mechanisms and the formation of products of all possible reaction modes si (suprafacial–inversion) ar (antarafacial–retention) ai and sr is observed with the si product predominating as expected for a Woodward–Ho§man allowed [p24 ]r2!] process. There is an additional ring opening–closing reaction which results in the geometric isomerisation of the ring. DFT calculations for reaction by the preferred si mode show a concerted mechanism for the rearrangement reaction but with an extensive plateau region running through the diradical-like transition state 35 and with the C2 methylene group twisted towards the final inverted configuration.The energies of the transition states leading to the other reaction products are within 3–7 kJ mol~1 of the transition state for the si reaction. The twisting of the C2 methyl- 283 Reaction mechanisms Part (ii) Pericyclic reactions H Ar N O O R N R O O Ar 36 E + Scheme 14 CO2Me CO2Me CO2Me AlCl3 38 37 + + Scheme 15 ene group is attributed to a repulsive interaction between the C––C p-orbital and the r-orbital of the cleaving C–C bond.40 Similar results are obtained for a CASSCF calculation and a transition state for the ring isomerisation reaction is also reported.41 Not all experimental findings are so easily rationalised by theory. Several computational studies on [1,3] migrations of Si in allylsilanes and siloxyacetylenes favour a sr reaction mode in apparent contradiction to experimental results.42 An experimental study of a [1,5]-intermolecular migration between maleimides and arylpropenes (an ene reaction) has appeared (Scheme 14).The concerted pericyclic reaction mechanism is supported by the observation of a very small aryl substituent e§ect on reactivity. The exclusive formation of the E product 36 is proposed to reflect a reaction through an endo transition state which is stabilised by a secondary orbital interaction between the ene HOMO and the enophile LUMO.43 Jenner has reported that the AlCl 3 - or ZrCl 4 -catalysed reaction of cyclohexene with methyl propiolate yields 37 from a [2]2] cycloaddition reaction and the product 38 of an ene reaction (Scheme 15).The chemoselectivity is constant with changing pressure and this is consistent with similar values of DS8 for both reaction pathways.44 Houk et al. have published the results of a computational study of the ene reaction of cyclopropene with ethene propene and cyclopropene using ab initio (RHF and MP) and DFT methods including CASSCF. The endo transition state for the dimerisation of cyclopropene by an ene reaction is predicted to be stabilised by the secondary orbital interaction between C–H and C––C. In some cases a degree of diradical character is suggested for this transition state and the reaction pathway has been illustrated by use of a More O’Ferrall–Jencks diagram.45 The temperature dependent broadening of the NMR peaks from the spectrum of 7-phenylsulfanylcyclohepta-1,3,5-triene has been attributed to migration of the PhS group by a series of [1,7] shifts.46 [n,n] Shifts The allyl thiocyanate to allyl isothiocyanate rearrangement follows clean first-order kinetics and there is no rate acceleration observed upon addition of thiocyanate ion.These results are consistent with a pericyclic [3,3] reaction mechanism. The results of 284 I.D. Cunningham N NH+ + 39 N N N N 40 –H+ [3,3] Scheme 16 ab initio calculations are in agreement with this conclusion and predict that there is some separation of charge in the reaction transition state. This conclusion cannot simply account for the small rate retardation observed for reactions in solvents of increasing polarity and these solvent e§ects are rationalised in terms of opposing e§ects of solvent polarity and hydrogen bonding on the reaction rate.47 The interconversion of the products 39 and 40 of the cycloaddition reaction shown in Scheme 16 has been shown to occur by a [3,3] rearrangement rather than by cycloreversion.It is not clear whether this rearrangement proceeds in one or two steps.48 The benzidine rearrangement reaction of 42 is proposed to proceed by a [9,9] sigmatropic shift through the transition state 41. The reaction is second-order in the concentration of acid with an observed third-order rate constant of 0.13dm6 mol~2 s~1. Only indirect evidence is cited for the proposed concerted pericyclic reaction mechanism and the authors admit that they cannot rule out the possibility that the reaction occurs by two consecutive [5,5] shifts. However the observation that the well documented [5,5] shift observed for the reaction of the less hindered hydrazobenzene is slower (0.024dm6 mol~2 s~1) than the reaction of 42 and the failure to observe any other products of rearrangement reactions is consistent with a reaction by a [9,9] sigmatropic shift.49 A study of the kinetics of the O toNallyl migration of 1-aryl-5-allyloxytetrazoles 43 inDMSOand in 1,1,2,2-tetrachloroethane has been reported (Scheme 17).A concerted mechanism for this reaction has been proposed on the basis of the lack of evidence (CINDP EPR) for a biradical intermediate. While the values of E!#5 are roughly independent of the aryl ring substituent and the alkene substituent R the values of *H8 and *S8 show significant compensating variations. These variations are consistent with a dipolar transition state in which there is a partial positive charge on the allyl fragment of substrate and a partial negative charge on the tetrazole fragment.However there is no evidence for the formation of ions in a polar LiClO 4 –ether medium.50 The variability of substituent e§ects on Claisen and related [3,3] rearrangements is made apparent in the above paper. Houk has formulated a theory for the e§ect of substitution of O-donor groups based upon RHF DFT and CASSCF calculations. A 285 Reaction mechanisms Part (ii) Pericyclic reactions O N N O H H N O N O HH HH N N N N Ar O R N N N N Ar R [3,3] 43 41 2+ 42 O Scheme 17 Ar Ar Ar Ar Ar Ar 45 46 44 + • Scheme 18 modification of Marcus theory has been used to model the relative contribution of the intrinsic reaction barrier and the thermodynamic driving force to the observed activation barrier.The results are interpreted using frontier molecular orbital theory and the reactant transition state and product orbitals obtained from RHF/3-61G* calculations. 51 The di¶culties in distinguishing stepwise and concerted reaction mechanisms are readily apparent for the rearrangement reactions of radical cations. However this question appears to have been resolved for the photo-induced electron-transfer (PET) Cope rearrangement of 44 to form 45 (Ar\4-MeOC 6 H 4 -) where the radical cations 44·` and 45·` and a distonic reaction intermediate 46·` have been detected following pulse radiolysis of the substrate (Scheme 18). It is not yet clear whether this stepwise mechanism is strongly favoured by the methoxyphenyl group.52 [n,m] Shifts where nDm Most of the literature reports of these rearrangement reactions concern their use in chemical syntheses.However a novel [3,4] shift has been described for the conversion of allene oxides such as 47 to the ketone 48 (Scheme 19). The proposed shift in the 6p 286 I.D. Cunningham CHO O O– CHO O CHO 47 + [3,4] 49 48 Scheme 19 O O C O O O O C O O CO 51 50 + D Scheme 20 electron system 49 is allowed by the Woodward–Ho§man rules but it has not yet been demonstrated whether this reaction proceeds by a ‘true’ pericyclic reaction mechanism. 53 5 Miscellaneous Birney et al. have published the results of an ab initio computational study of thermal chelotropic decarbonylation reaction mechanisms which have been described as pseudopericyclic.These calculations predict a transition state 51 for decarbonylation of 50 which is consistent with the ‘in-plane’ departure of CO and the orbital topology shown in Scheme 20 in which the lack of formal cyclic orbital overlap characterises the reaction as pseudopericyclic.54 References 1 O. Wiest D. C. Montiel and K. N. Houk J. Phys. Chem. 1997 101 8378. 2 H.-W. Fru� hauf Chem. Rev. 1997 97 523. 3 M. Schmittel C. Wo� hrle and I. Bohn Acta Chem. Scand. 1997 51 151. 4 N.T. Anh and F. Maurel New J. Chem. 1997 21 861. 5 J. J. Gajewski Acc. Chem. Res. 1997 30 219. 6 Advances in theoretical methods are covered elsewhere in Annual Reports but a good example is by M. K. Diedrich F.-G. Kla� rner B. R. Beno K. N. Houk H. Senderowitz and W. C. Still J. Am. Chem. Soc. 1997 119 10 255.7 J. Motoyoshiya T. Kameda M. Asari M. Miyamoto S. Narita H. Aoyama and S. Hayashi J. Chem. Soc. Perkin Trans. 2 1997 1845. 8 M. Sodupe R. Rios V. Branchadell T. Nicholas A. Oliva and J. J. Dannenberg J. Am. Chem. Soc. 1997 119 4232. 287 Reaction mechanisms Part (ii) Pericyclic reactions 9 B. S. Jursic J. Org. Chem. 1997 62 3046. 10 R. Konecny and D. J. Doren J. Am. Chem. Soc. 1997 119 11 098. 11 A. V. Teplyakov M.J. Kong and S. F. Bent J. Am. Chem. Soc. 1997 119 11 100. 12 S.-W. Leung and D. A. Singleton J. Org. Chem. 1997 62 1955. 13 J.-C. Halle� D. Vichard M.-J. Pouet and F. Terrier J. Org. Chem. 1997 62 7178. 14 S. Pugnaud D. Masure J.-C. Halle� and P. Chaquin J. Org. Chem. 1997 62 8687. 15 Y. Tamaru H. Harayama H. Sakata H. Konishi K. Fugami M. Kimura S. Tanaka T. Okajima and Y.Fukazawa Liebigs Ann./Recuil 1997 907. 16 R. N. Butler E. C. McKenna and D. C. Grogan Chem. Commun. 1997 2149. 17 E. Eibler P. Ho� cht B. Prantl H. Roßmaier H. M. Schuhbauer H. Wiest and J. Sauer Liebigs Ann./Recuil 1997 2471. 18 J. Sauer J. Breu U. Holland R. Herges H. Neumann and S. Kammermeier,Liebigs Ann./Recuil 1997 1473. 19 M. Hartnagel K. Grimm and H. Mayr Liebigs Ann./Recuil 1997 71. 20 R. Ocampo W.R. Dolbier Jr. M. D. Bartberger and R. Parades J. Org. Chem. 1997 62 109. 21 I. Morao B. Lecea A. Arrieta and F. P. Cossý� o J. Am. Chem. Soc. 1997 119 816. 22 J. Sauer and H. M. Schuhbauer Liebigs Ann./Recuil 1997 1739. 23 J. W. Wijnen and J. B. F. N. Engberts Liebigs Ann./Recuil 1997 1085. 24 J. W. Wijnen and J. B. F. N. Engberts J. Org. Chem. 1997 62 2039. 25 A. R.Renslo R. D. Weinstein J. W. Tester and R. L. Danheiser J. Org. Chem. 1997 62 4530. 26 D. P. Heller D. R. Goldberg and W. D. Wul§ J. Am. Chem. Soc. 1997 119 10 551. 27 K. Endo T. Koike T. Sawaki O. Hayashida H. Masuda and Y. Aoyama J. Am. Chem. Soc. 1997 119 4117. 28 G. Desimoni G. Faita M. Mella M. Ricci and P.-P. Righetti Tetrahedron 1997 53 13 495. 29 M. Kimura Y. Horino Y. Wakamiya T. Okajima and Y. Tamaru J. Am. Chem. Soc. 1997 119 10 869. 30 E. Albrecht J. Mattay and S. Steenken J. Am. Chem. Soc. 1997 119 11 605. 31 A. J. DelMonte J. Haller K. N. Houk K. B. Sharpless D. A. Singleton T. Strassner and A. A. Thomas J. Am. Chem. Soc. 1997 119 9907. 32 G. Ujaque F. Maseras and A. Lledo� s J. Org. Chem. 1997 62 7892. 33 K. V. Gothelf and K. A. Jørgensen J. Chem. Soc. Perkin Trans.2 1997 111. 34 J. Haller S. Niwayama H.-Y. Duh and K. N. Houk J. Org. Chem. 1997 62 5728. 35 J. M. Olivia J. M. Gerratt P. B. Karadakov and D. L. Cooper J. Chem. Phys. 1997 107 8917. 36 O. Wiest J. Am. Chem. Soc. 1997 119 5713. 37 L. Luo M.D. Bartberger and W. R. Dolbier Jr. J. Am. Chem. Soc. 1997 119 12 366. 38 T. Suzuki T. Ohwada and K. Shudo J. Am. Chem. Soc. 1997 119 6774. 39 D. A. Smith and C. W. Ulmer II J. Org. Chem. 1997 62 5110. 40 K. N. Houk M. Nendal O. Wiest and J. W. Storer J. Am. Chem. Soc. 1997 119 10 545. 41 E. R. Davidson and J. J. Gajewski J. Am. Chem. Soc. 1997 119 10 543. 42 T. Yamabe K. Nakamura Y. Shiota K. Yoshizawa S. Kawauchi and M. Ishikawa J. Am. Chem. Soc. 1997 119 807; M. Takahashi and M. Kira J. Am. Chem. Soc. 1997 119 1948; M. Oblin F. Fotiadu M. Rajzmann and J.-M.Pons J. Chem. Soc. Perkin Trans. 2 1997 1621. 43 I. D. Cunningham A. Brownhill I. Hamerton and B. J. Howlin Tetrahedron 1997 53 13 473. 44 G. Jenner New J. Chem. 1997 21 1085. 45 Q. Deng B. E. Thomas IV K. N. Houk and P. Dowd J. Am. Chem. Soc. 1997 119 6902. 46 G. A. Dushenko I. E. Mikhailove A. Zschunke N. Hakam C. Mu� gge and V. I. Minkin Mendeleev. Commun. 1997 50. 47 M. Kotani Y. Shigetomi M. Imada M. O3 ki and M. Nagaoka Heteroat. Chem. 1997 8 35. 48 K. Beck P. Ho§man and S. Hu� nig Chem. Eur. J. 1997 3 1588. 49 K. H. Park and J. S. Kang J. Org. Chem. 1997 61 3794. 50 M.L. S. Cristiano and R. A. W. Johnstone J. Chem. Soc. Perkin Trans. 2 1997 489. 51 H. Y. Yoo and K. N. Houk J. Am. Chem. Soc. 1977 119 2877. 52 H. Ikeda A. Ishida T. Takasaki S. Tojo S. Takamuka and T. Miyashi J. Chem. Soc. Perkin Trans. 2 1997 849. 53 I. Erden F.-P. Xu and W.-G. Cao Angew. Chem. Int. Ed. Engl. 1997 36 1516. 54 D.M. Birney S. Ham and G. R. Unruh J. Am. Chem. Soc. 1997 119 4509; see also J. A. Ross R. P. Seiders and D. M. Lemal J. Am. Chem. Soc. 1976 98 4325 for a definition of ‘pseudopericyclic’.