Brønsted Acid-catalyzed S N 2 Reaction in Bulk Solution.
Firstly, we calculated the SN2 reaction catalyzed by phosphoric acid in bulk solution. The S-enantiomer of the experimentally used 1-phenylethyl-2,2,2-trichloroacetimi-date (sub, Fig. 1) and the phosphoric acid (Cat) were chosen as the substrate and the catalyst in the calculations. The complexation of the trichloroacetimidate, methanol, and phosphoric acid leads to intermediate 1, which is slightly endergonic by 2.4 kcal/mol. In intermeidate 1, except the hydrogen bonding that between the methanol O-H bond and the catalyst’s oxygen atom, the O-H of the catalyst also forms a hydrogen bonding with the N atom of the trichloroacetimidate substrate. The methanol and the trichloroacetimidate group were located on opposite sites of the phenyl ring in 1. The phosphoric acid-assisted back-side nucleophilic substitution occurs via transition state 2-ts to produce intermediate 3, in which the chirality of the benzylic carbon has been inversed. This step has an activation free energy barrier of 7.9 kcal/mol and is exergonic by 11.7 kcal/mol with respect to sub. Finally, the release of the trichloroacetamide and regeneration of the phosphoric acid is exergonic by 4.7 kcal/mol to provide the ether product 8R, which has the opposite chirality compared to the starting reactant (sub).
Alternatively, the complexation of the three components could lead to intermediate 4, with the methanol and the trichloroacetimidate group locating on the same side of the benzylic carbon. Subsequently, the same-side nucleophilic attack takes place through 5-ts to generate 6, in which the chirality of the benzylic carbon has been retained. Similarly, the release of the trichloroacetamide and regeneration of acid catalyst lead to the production of the ether product 8S. This pathway has an activation free energy barrier of 9.2 kcal/mol, which is 1.3 kcal/mol higher than the pathway leading to 8R. Our calculations are consistent with experiment in which the chirality-inversed R-product was observed as the major product in bulk solution. We have also considered the other conformations of 1 and 4, in which the O-H of the catalyst forms hydrogen bondings with the O atom of the trichloroacetimidate (i.e., 1b and 4b, Fig. 1), however, calculations show that both are much more unfavorable in free energy and can be ruled out.
Distortion-interaction analysis[48] were performed on the selectivity-determining transition states. It can be seen that the interaction energy (ΔEint) between the substrate, methanol, and phosphoric acid is more favorable in 5-ts than in 2-ts. However, both the distortion of the substrate (ΔEdist-sub) and the distortion of the catalyst (ΔEdist-cat) is much larger in the former than the latter, which overcome the favorable interaction enengy and result in the unfavorable same-side attacking transition state 5-ts. Detailed geometric analysis shows that the larger distortion of the substrate in 5-ts comes from the more elongated C1-O1 bond (3.01 Å) compared to that in 2-ts (2.08 Å), which might be due to the more pronounced steric repulsion induced by the same-side methanol molecule with the leaving group. Similarly, the larger distortion of the phosphoric acid catalyst might be because of the elongated O3-H1 bond in 5-ts (1.63 Å) than in 2-ts (1.57 Å).
Thus, the much larger unfavorable distortion of the substrate and catalyst in the same-side nucleophilic substitution transition state than in the back-side nucleophilic substitution transition state, mainly account for the selectivity of the SN2 reaction between the 1-phenylethyl-2,2,2-trichloroacetimidate and methanol in bulk solution.
Supramolecular Cage-catalyzed S N 2 Reaction.
Simplified Supramolecular Cage. The experimentally-used supramolecular cage is a chiral cage which is formed by the self-assembly of S-type chiral L ligands and Ga3+ (Fig. 3a and 3b). The R-type chiral ligand was also used to create supramolecular cage with different chirality.
It should be noted that different chiral cages provided same reaction results, indicating that the chirality of the product is independent to the chirality of the cages and the ligands. Based on this, we have simplified the tert-butyl group on the terminal substituent of the ligand to methyl group (Fig. 3c, Cage) to reduce the computational cost.
Conformation Search of the Host-substrate Complex. Firstly, we have performed conformational search for the substrate-encapsulated host-guest complex (see SI for details). Our calculations show that the phenyl ring on the substrate tends to be situated near the apex of [Ga4L6]12-. This is also supported by the electrostatic potential (ESP) analysis and the non-covalent interaction (NCI) analysis. The ESP analysis shows that the electron-negative regions are located on the four apexes of the[Ga4L6]12- molecule, which is due to the presence of the phenolic oxygen atoms (Fig. 4a). The ESP analysis on the substrate indicates that the hydrogen atoms on the phenyl ring posses positive electron densities (Fig. 4b). Therefore, it would tend to interact with the [Ga4L6]12- cage at the electron-negative apex area. Importantly, such a conformation also brings stable π-π stacking interaction between the phenyl ring of substrate and the naphthalene ring of the ligand on the supramolecule, as shown by the NCI analysis (Fig. 4c, bottom left)
Reaction Pathway with One Methanol Molecule. After the determination of the most stable conformation of sub@C, the next step is the encapsulation of the the nucleophile, methanol, to proceed the nucleophilic substitution reaction. @@We have also calculated different conformations for the methanol encapsulated complex (i.e., sub@C with methanol) and the most stable one is given (1@C, Fig. 5) while all others are shown in Supporting Information (see SI for details).@@ Calculations show that the encapsulation of methanol is exergonic by 9.4 kcal/mol to form 1@C, in which the methanol forms hydrogen bonding interaction with the imine group of the substrate. Subsequently, the same-side nucleophlic substitution with the simutaneous proton transfer takes place via 2-ts@C to generate the chirality-rentention S-product 3@C. This step has an activation free energy barrier of 34.1 kcal/mol, which is dramatically high considering the reaction condition. It should be noted that the back-side nucleophilic substitution, which would lead to the chirality-reverse R-product, however, is unpractical because the proton transfer from methanol to imine during the nucleophilic attack is prohibited due to much longer distance (2-ts’@C). Overall, with one methanol molecule inside the [Ga4L6]12- cage, the same-side nucleophilic substitution reaction occurs with quite high barrier and is unfavorable.
Encapsulation of Additional Solvent Molecule and Conformational search on the Adduct complex. Since the reaction with one methanol in the cage is unfavorable due to high activation barrier, we have further considered that whether additional solvent molecule encapsulated in the cavity that can promote the nucleophilic substitution process. Previously, there are studies that determine the number of solvents in supramolecular cages. Gregori Ujaque[49] and coworkers used MD simulations to show that at maximum, two solvent molecules can be present inside the cavity of [Ga4L6]12− with the encapsulated guest [(Et3P)Au(I)(CH3)2]. The follow-up work by Gregori Ujaque[50] also used MD simulations proved [Ga4L6]12− encapsulated [R3PAu(MeOH)(CH3)2] can accommodate additional MeOH, and the reaction rate is regulated by the presence of solvent molecule. In our system, mixed solvent (methanol:water = 1:1) were used. Given that in experiment the alcohol product, which was resulted by the nucleophilic attack of water, was observed as a side product, therefore, we considered an additional water present inside the cavity. That is, the starting intermediate for the nucleophilic substitution reaction is a host-guests complex, in which the substrate, one methanol, and one water are encapsulated inside the cage. It should be noted that the performed ab initio molecular dynamics simulation also supports that the metallocage host can encapsulate one methanol and one water together with the substrate (see SI for more details).
Reaction Pathway with Methanol and Water Inside the Cage. From 1@C, the encapsulation of one water into the cage leads to 4@C, which is slightly endergonic by 2.0 kcal/mol. It should be noted that we have also performed conformational search on this host-guests complex and 4@C is obtained as the most stable structure (Fig. 6, see also SI for more details).
In 4@C, water molecule forms hydrogen bonding interaction with both the imine group of the substrate and the methanol molecule, and methanol is placed on the same side with the leaving group. Subsequently, the same-side nucleophilic substitution occurs through 5-ts@C to provide the chirality-retained product 6@C. This step has an activation free energy barrier of 29.1 kcal/mol and is exergonic by 10.8 kcal/mol with respect to 1@C. The encapsulation of water could also lead to 7@C, in which the methanol is positioned at back-side with the leaving group. The following classical back-side nucleophilic substitution takes place via 8-ts@C to generate the chirality-reverse product 9@C. This pathway has a free energy barrier of 34.0 kcal/mol and is exergonic by 5.3 kcal/mol. By comparison, the pathway that leads to chirality-reversed product has 4.9 kca/mol higher free energy barrier than that leads to chirality-retained product, which is consistent with experiment.
Origins of Selectivity Inside the Metallocage. By detailed analysis of the selectivity-determining transition states, it is found that although both TSs have some steric repulsion between the substrates and the metallocage (shown in red dashed lines in Fig. 7), there are more favorable hydrogen bonding interactions in 9-ts@C than in 12-ts@C (shown in green dashed lines in Fig. 7). In addition, in transition state 9-ts@C, the methanol group and the leaving group position at the same side, making it possible for the back side of the phenyl ring of the substrate to approach the naphthalene ring of the cage in a parallel way. Indeed, there is a π-π stacking interaction between the phenyl ring of substrate and the naphthalene ring of the supramolecule in 9-ts@C (3.10 Å, shown in blue dashed line in Fig. 7). On the other hand, in 12-ts@C, the methanol attacks the benzylic carbon from the opposite side of the leaving group. Therefore, both sides of the phenyl ring were occupied, preventing the approach of the phenyl ring to the naphthalene ring of the supramolecule and resulting in no π-π stacking interaction. Thus, the more favorable hydrogen bonding interaction and π-π stacking interaction that favor the same-side nucleophilic substitution transition state inside the cage account for the enantioselectivity.
The Side Pathway Leading to Alcohol with Water as the Nucleophile. The side pathway that leads to the alcohol product has also been studied. In this reaction, water plays the role of nucleophile instead of methanol. As can be seen in Fig. 8, from 1@C, the encapsulation of water and rearrangement would generate 10@C or 13@C, which then proceed through same-side or back-side nucleophilic substitution transition states 11-ts@C or 14-ts@C, respectively, to produce the chirality-retained or chirality-reversed product complexes 12@C or 15@C. In agreement with experiment, the chirality-retained pathway (i.e. via 11-ts@C) is 1.1 kcal/mol more favorable compared to the chirality-reversed pathway (i.e., via 14-ts@C). More importantly, the pathway for formation of alcohol (i.e., 1@C → 10@C → 11-ts@C → 12@C) has a free energy barrier of 29.8 kcal/mol, which is 0.7 kcal/mol unfavorable compared to the formation of ether product (i.e., 1@C → 4@C → 5-ts@C → 6@C, Fig. 7), which is consistent with experimental observation that the ether is the major product.