Allyl thermodynamic control

Conjugated dienes undergo several reactions not observed for nonconjugated dienes. One is the 1,4-addition of electrophiles. When a conjugated diene is treated with an electrophile such as HCl, 1,2- and 1,4-addition products are formed. Both are formed from the same resonance-stabilized allylic carbocation intermediate and are produced in varying amounts depending on the reaction conditions. The L,2 adduct is usually formed faster and is said to be the product of kinetic control. The 1,4 adduct is usually more stable and is said to be the product of thermodynamic control. 

The silicon- and sulfur-substituted 9-allyl-9-borabicyclo[3.3.1]nonane 2 is similarly prepared via the hydroboration of l-phenylthio-l-trimethylsilyl-l,2-propadiene with 9-borabicy-clo[3.3.1]nonane36. The stereochemistry indicated for the allylborane is most likely the result of thermodynamic control, since this reagent should be unstable with respect to reversible 1,3-borotropic shifts. Products of the reactions of 2 and aldehydes are easily converted inlo 2-phenylthio-l,3-butadienes via acid- or base-catalyzed Peterson eliminations. 

It was recognized in early examples of nucleophilic addition to acceptor-substituted allenes that formation of the non-conjugated product 158 is a kinetically controlled reaction. On the other hand, the conjugated product 159 is the result of a thermodynamically controlled reaction [205, 215]. Apparently, after the attack of the nucleophile on the central carbon atom of the allene 155, the intermediate 156 is formed first. This has to execute a torsion of 90° to merge into the allylic carbanion 157. Whereas 156 can only yield the product 158 by proton transfer, the protonation of 157 leads to both 158 and 159. 

The latter results have been explained on the basis of the following reaction scheme. The 1,2-regioisomers derived from butadiene are obtained through a non-symmetrical iodonium ion intermediate. The subsequent nucleophilic attack on the allylic position gives, under kinetic control, 1,2-derivatives. Nevertheless, when poorer nucleophiles such as benzene or acetonitrile are employed, the conversion of the initially formed iodonium ion into the allylic cation has been suggested to give 1,4-products, under thermodynamic control. However, other alternatives like nucleophilic attack involving allylic participation have not been excluded for the formation of 1,4-derivatives. 

The influence of the classical anomeric effect and quasi-anomeric effect on the reactivity of various radicals has been probed. The isomer distribution for the deu-teriation of radical (48) was found to be selective whereas allylation was non-selective (Scheme 37). The results were explained by invoking a later transition state in the allylation, thus increasing the significance of thermodynamic control in the later reactions. Radical addition to a range of o -(arylsulfonyl)enones has been reported to give unexpected Pummerer rearrangement products (49) (Scheme 38).A mechanism has been postulated proceeding via the boron enolate followed by elimination of EtaBO anion. 

The different products arise from enthalpy differences in the second step, the reaction of Br and the allyl R. See Fig. 8-5. At -80°C the 1,2-adduct, the rate-controlled product, is favored because its formation has the lower A// . 1,2-Adduct formation can reverse to refurnish the intermediate allylic carbocation, R. At 40°C, R goes through the higher-energy transition state for formation of the more stable 1,4-adduct, the thermodynamic-controlled product. The 1,4-adduct accumulates because the addition, having a greater is more difficult to reverse than that for the 1,2-adduct. The 1,4-adduct has a lower enthalpy because it has more R groups on the C=C,... 

The diphenylallyl carbanion is conveniently formed by the abstraction of an allylic proton from the corresponding diphenyl propene. Under certain conditions it was found that the initial product formed from trans 1,3 diphenyl-2-methyl propene was the trans,trans anion which isomerised completely into the cis,trans conformation within a few minutes [3]. Clearly, the abstraction reaction proceeds under kinetic and-not thermodynamic control. 

In this connection, our finding that the allylic mercury intermediate represents the only isolated product following reaction in THF, CH2CI2, or HMPA is pertinent. The generation of this species appears to be rapid in all media. The ensuing step in which Hg(II) is reduced to Hg(O) likely serves as the rate-determining step. If so, the role of nonpolar benzene may be to stabilize the transition state and accelerate the overall rate. No allylic alcohol having an exocyclic double bond was formed, a feature that hints to the possible operation of thermodynamic control. 

We shall consider reactions catalysed by two different types of pro-catalyst the first (type A) employs Pd-allyl cations ([Pd(a]lyl)(PCy3)]+/Et3SiH or [Pd(allyl)(MeCN)2] + ), and the second (type B) employs Pd-alkyl or chloro complexes ([(phen)Pd(Me)(MeCN)]+, where phen = phenanthroline, and [(RCN)2PdCl2]). These two types of catalysts give very different products in the cyclo-isomerisation of typical 1,6-dienes such as the diallyl-malonates (10), Scheme 12.6. Since there is known to be a clear order of thermodynamic stability 11 < 12 <13, with a difference of ca. 3-4 kcal mol 1 between successive pairs, any isomerisation of products under the reaction conditions will tend towards production of 12 and 13 from 11 and 13 from 12. Clearly, when 11 is the major product (as with pro-catalysts of type A), it must be the kinetic product (see Chapter 2 for a discussion of kinetic and thermodynamic control of product distributions). However, when 12 is generated selectively, as it is with pro-catalysts of type B, there is the possibility that this is either generated by rapid (and selective) isomerisation of 11 or generated directly from 10. 

Palladium(0)-catalyzed allylation of nucleophiles (the Tsuji-Trost reaction) is a versatile synthetic method that has gained immense popularity in recent years. Rarely applied to ambident nucleophilic aromatic heterocycles before 1991, the Tsuji-Trost reaction has been extensively used in the chemistry of these compounds since 1991. Two factors have played decisive roles in this increased interest in the Pd(0)-catalyzed allylation of such heterocyclic rings one is that, unlike other alkylation procedures, the Pd(0)-catalyzed allylation can sometimes give the product of thermodynamic control when applied to ambident nucleophiles and the second is that the Tsuji-Trost allylation has become one of the standard methods for synthesizing carbanucleosides, which are important antiviral compounds (93MI1, 93MI2). Of course, the double bond of an allylic system can be modified in different directions, thus adding versatility to the Tsuji-Trost reaction. 

We include in Sections I,A and I,B some general features of the Tsuji-Trost reaction with comments on kinetic versus thermodynamic control in allylations and in alkylations in general. Then we review in Sections II, III, and IV all cases known to the authors of the application of the Tsuji-Trost reaction to ambident nucleophilic aromatic heterocycles. This leaves out of the review the allylation of such heterocyclic ambident nucleophiles as 2-piperidone and the like. By aromatic, we mean any heterocycle for which a tautomeric or mesomeric formula can be written that is aromatic in the normal structural sense of having 4n + 2n- electrons cyclically conjugated. 

This review covers the Pd(0)-catalyzed allylations of aromatic ambident heterocyclic compounds, including all rings for which an aromatic tautomeric or resonance form can be written. Cases of C vs. O, C vs. N, N vs. O, and S vs. N allylation are discussed from all available viewpoints regioselectivity, kinetic vs. thermodynamic control, mechanism, stereo-... 

Direct allylation of rhodanine 49 (Scheme 13) under Pd(0)-catalysis with cinnamyl ethyl carbonate affords the /V-allylated compound 50. However, allylation with cinnamyl bromide and a base is not regioselective, producing a mixture of 50 and sulfide 51. Sulfide 51 isomerizes to 50 under palladium catalysis (N > S), thus indicating that Pd(0)-catalyzed allylation of 49 is thermodynamically controlled (93T1465). 

In one of the first papers on the subject, Billups et al. (80SC147) reported that the Pd(0)-catalyzed allylation of indole 96 with allyl acetate gave N-allyl- (97) and 3-allylindole (98) plus the diallylation product 99 (Scheme 21). They also showed that the yV-allyl isomer 97 rearranged under Pd(0) catalysis to the C-3 isomer 98, thus indicating that the formation of 98 was thermodynamically controlled (C > N). The work of Billups also includes the use of allyl alcohol instead of allyl acetate in the Tsuji-Trost reaction. 

The Pd(0)-catalyzed allylations of 120, benzothiazolethione 123, and benz-imidazolethione 125 (Scheme 26) have been studied by the current authors (93T1465). In all cases, the sulfides 121 (R = Ph), 124, and 126 were formed, either as major or as exclusive products. Attempts to isomerize sulfides into A-cinnamyl derivatives were made. Slow conversions were observed for sulfides 121 (R = Ph) and 124 under forcing conditions, thus indicating that sulfides are the products of kinetic control and the /V-cinnamyl derivatives are the products of the thermodynamic control (N > S), which is not easily attainable. Of course, the reaction of 125 with two equivalents of allylating agent gives the S./V-dicinnamyl derivative 127. 

The palladium-allylation of ambident aromatic heterocycles is covered by Professor Moreno-Mafias and Dr. Pleixats (Barcelona, Spain) in the second chapter of this volume. The preference for carbon versus oxygen, nitrogen, and sulfur allylation is discussed from the diverse viewpoints of regioselectivity, kinetic versus thermodynamic control, mechanisms, stereochemistry, and synthetic targets in the first general survey of this topic. 

At the higher temperature, the reaction becomes reversible and is under thermodynamic control. This means there is enough energy available for either product to reform the allylic carbocation by an SN1 type ionization and then form the other product. As a result the products are in equilibrium. At equilibrium, the relative amount of the products is controlled only by the difference in energy between them (AG). In this case, the 1,4-addition product (called the thermodynamic product) is more stable than the 1,2-addition product, so more of it is present in the equilibrium mixture. This same equilibrium mixture of products (15% of the 1,2-addition product and 85% of the 1,4-addition product) is produced when the low-temperature reaction product mixture (80% of the 1,2-addition product and 20% of the 1,4-addition product) is heated to 45°C. 

---