Revisiting the Stability of endo / exo Diels-Alder Adducts between Cyclopentadiene and 1 , 4-benzoquinone

Neste trabalho é apresentada uma análise teórica detalhada da estabilidade relativa dos adutos endo/exo formados entre o ciclopentadieno (1) e a 1,4-benzoquinona (2). As coordenadas intrínsecas de reação (IRC) indicaram a presença de apenas um estado de transição para a reação, mostrando que se trata de um mecanismo concertado para ambos os adutos, endo 3 e exo 4. As energias dos adutos foram calculadas com um alto nível de teoria (CBS-Q) confirmando que o aduto endo é mais estável que o exo, o que está em desacordo com o que é observado para reações que usualmente seguem a regra de Alder. Uma análise estrutural eletrônica foi realizada através da metodologia NBO, a qual indicou que interações atrativas predominam sobre as interações estéricas repulsivas no aduto endo. Em resumo, para a reação de cicloadição estudada o aduto endo é o produto termodinâmico e cinético, o que pode ser confirmado também pelos dados experimentais mencionados neste trabalho.


Introduction
2][3][4][5][6][7][8] Historically, in 1906 Albrecht 9 published the reaction between the cyclopentadiene (1) and 1,4-benzoquinone (2) as a 1:1 adduct.However, Albrecht's considerations about the structure of the obtained adduct were inconsistent.After several studies, in 1928 Otto Diels and Kurt Alder, 10 established a correct structure for the mono-and bis-adducts formed by the reaction between these compounds through a [4+2] cycloaddition, in opposition to the reactional course suggested by Albrecht: an 1,4-addition of cyclopentadiene (1) into 1,4-benzoquinone (2).Since then, the Diels-Alder reaction has appeared in more than 32,000 papers involving synthetic and theoretical approaches. 11n general, Diels-Alder reactions are excellent reactional models for the transition states calculations due to the nature of their mechanism (usually concerted).Vol.6][17][18] Most of these studies attempt to explain some experimental results non-consistent with empirical rules or expected results concerning the selectivity.While experimental measurements can provide accurate rate constants for the different reaction pathways, high-level quantum chemical calculations are often necessary to explain the observed phenomena at an electronic level, to predict the substituent effects and also to evaluate steric interactions between the participating species.In this context, advances in the quantum-chemical calculations methodology (e.g., the development of new DFT functional methods) and the improvements of computational power provided the accuracy needed to quantitatively explain experimental observations.Moreover, convenient methods for the analysis of correlated wave functions, as natural bond orbital analysis (NBO), 19,20 can be used to evaluate the numerous stereoelectronic interactions, 21,22 including secondary orbital interactions (SOI). 23,24These interactions (SOI) were proposed by Woodward and Hoffmann 25 to rationalize the empirical endo principle of addition (Alder's rule) formulated by Alder and Stein, [26][27][28][29] but they had been object of some criticism. 3029][31][32][33][34] On the other hand, the endo product is usually less stable than exo.
While the stereoelectronic interactions, involved in the transition state stability, are widely studied in the literature, 14,15,35 the same attention has not been dedicated to understand the stability of the products.One of the most used examples of the Diels-Alder reaction, which is listed in several Organic Chemistry textbooks, [36][37][38] is that between cyclopentadiene (1) and maleic anhydride which, at room temperature, gives only the endo adduct that is then converted at 200 o C to the thermodynamically more stable exo adduct through a retro Diels-Alder reaction. 4According to several authors, [39][40][41][42][43][44][45][46][47][48][49][50][51][52] the Diels-Alder reaction between cyclopentadiene (1) and 1,4-benzoquinone (2) (Scheme 1), gives, in a similar way, only the kinetic endo adduct 3, although the exo adduct 4 has higher stability than the endo adduct 3 due to the steric repulsive interaction present in the endo form. Actualy, we observed from the 1 H NMR spectrum of the reactional mixture, 98% of compound 3 and 2% of compound 4 (Fig. 1).
Thus, in principle, the results for cycloaddition between compounds 1 and 2 follow the Alder's rule, however some surprising results were observed when an intrinsic reaction coordinate (IRC) and the NBO analysis of the endo/exo transition states and products were performed at high level of theory.These results are discussed in the present study.

Chemicals
All solvents and reagents were purchased from Merck, Acros or Aldrich.The cyclopentadiene (1) was freshly distilled and then used.p-Benzoquinone (2) was purified by sublimation in an appropriated apparatus (Aldrich).

Synthesis of endo
To a solution containing 541 mg (5.0 mmol) in 18 mL of dry methanol at -78 ˚C under nitrogen atmosphere, was added cyclopentadiene freshly distilled (344 mg, 5.2 mmol, in 4 mL of dry methanol) also cooled to -78˚C.Then, the reaction mixture was allowed to reach 0 ˚C (approx.1h).After that, the solvent was removed under reduced pressure and the product was crystallized by using hexane, yielding the yellow crystals (854 mg, 4.9 mmol, 98%).mp.63-65 ˚C.

Computational Details
All calculations were performed with Gaussian 03 program. 53Full geometry optimization for adducts were performed applying HF, B3LYP, 54,55 MP2 full 56,57 (full specifies that all electrons are included in a correlation calculation) and CBS-Q 58 methods and Dunning's correlation consistent basis set 59,60 were used.The CBS-Q method was used to solve the major source of error in most ab initio calculations of molecular energies, which is due to truncation of the one-electron basis set, and the mean absolute deviation in electronic energy, using CBS-Q method, is less than 1 kcal mol -1 . 58he transition states calculations (TS) and NBO analysis 61 were performed at the B3LYP/cc-pVTZ level of theory, while Intrinsic Reaction Coordinate (IRC) calculation was performed at the B3LYP/6-31g (d,p) level.All stationary points were characterized as minima or transition structures by calculating the harmonic vibrational frequencies (ZPE), expected for MP2 full calculations.
The IRC calculations (Fig. 2) for both reaction pathways were performed at the B3LYP/6-31g (d,p) level and it was observed that there is only one transition state for both adducts, which suggests that for this reaction the mechanism is concerted.It can also be observed, that the transition state energy to form the endo adduct is smaller than for exo adduct.
Theoretical calculations at high level of theory (Table 1) were performed for adducts endo and exo to get more accurate energies values to check if the endo adduct is in fact the thermodynamic product.It can be seen from Table 1 that in all levels of theory applied in the present study, the endo adduct is the more stable than exo.
Structure stability is usually explained by stereoeletronic interactions [attractive (delocalization interaction) or repulsive (steric interaction)]. 63,64The NBO analyses can be invoked to explain the unexpected stability observed for endo adduct and quantify the delocalization interaction.There is a simple way to determine which interaction (attractive or repulsive) is more pronounced in any structure.First the NBO analysis of a full wave function was performed.Then, only the attractive delocalization interactions are deleted (NBOdel) and the energy is recalculated without these delocalization interactions, which means that only steric interactions will be present.The energy change values (Table 2) can be used to determine the amount of attractive or repulsive interactions in endo and exo adducts.This energy value is the difference between a full molecular electronic energy and the molecular electronic energy calculated deleting the delocalization interactions, and the larger is this energy change the large attractive delocalization interaction the structure has.
When these energy changes for endo and exo are computed (Table 2) at the HF/cc-pVTZ and B3LYP/ cc-pVTZ level, deleting the delocalization interactions (NBOdel), the exo adduct becomes more stable than endo by 4.7 and 3.3 kcal mol -1 , respectively.The data from Table 2 corroborate that in the endo adduct the steric repulsive interactions are larger than in exo, however, the attractive delocalization interaction in endo is higher than in exo and compensates the steric repulsion.This is the reason why the endo 3 is also the thermodynamic product and not the exo 4, as it is reported in the current literature.
The most important delocalization orbital interactions from NBO analysis involving bonding and antibonding orbital, responsible for the stabilization of endo adduct, are listed in Table 3; similar interactions for exo adduct were also collected.The interactions listed at Table 3, are those only for the molecular fragment that differentiate endo from the exo adduct.As can be seen from Table 3, the attractive interactions (delocalization orbital interactions) are higher for endo than exo adduct, and probably are responsible for extra stabilization acquired for endo in comparison to exo.Probably, other interactions involving other parts of molecule system should also increase the stabilization of endo.
The results discussed in the present paper can be used for a correct explanation of the experimental results obtained by Yates and Switlak 48 for thermal stereoisomerization of adduct endo 3.In their study, 48 the authors reflux a solution of adduct endo in toluene under argon for 36 h and the products of this thermal isomerization consisted of 1:9 mixture of adduct exo and endo (50%), one hydroquinone derivative (1,4-dihydro-1,4-methanonaphthalene-5,8-diol) and 1,4-benzoquinone.The authors suggest that the small amount of exo adduct is due to the position equilibrium was not established under that conditions.However, the reason for small amount of exo adduct obtained in the thermo stereoisomeration reaction performed by Yates and Switlak 48 is simply due to the fact that the endo adduct is the thermodynamic and also the kinetic product.

Conclusions
The study described in the present work was performed at high level of theory using sophisticated theoretical methods and shows some evidences that the oldest Diels-Alder reaction follow Alder rule.However, the adduct endo 3 is not only the kinetic product, but also the thermodynamic product, which is in agreement with   experimental data.Also, this study demonstrates that some stabilizing stereoeletronic interactions are present in the endo product similar to observed in the endo transition state.
The most important point that we want to emphasize is that the stability acquired by any structure will depend of an energy balance between attractive and repulsive interactions present in that structure.
Total SCF energy -Deletion energy.b DE = energy change for endo -energy change for exo.

Table 1 .
Calculated energies (kcal mol -1 ) for adducts endo and exo at different level of theory

Table 2 .
Energies obtained from NBO deletion calculation for endo and exo adducts at the HF and B3LYP method using cc-pVTZ basis set

Table 3 .
Delocalization orbital interaction energies (kcal mol -1 ) from NBO analysis for endo and exo adduct, calculated at the HF/cc-pVTZ level