Asymmetric Cyanation of Activated Oleﬁns with Ethyl Cyanoformate Catalyzed by Ti(IV)-Catalyst: A Theoretical Study

: The reaction mechanism and origin of asymmetric induction for conjugate addition of cyanide to the C = C bond of oleﬁn were investigated at the B3LYP-D3(BJ) / 6-31 + G** // B3LYP-D3(BJ) / 6-31G**(SMD, toluene) theoretical level. The release of HCN from the reaction of ethyl cyanoformate (CNCOOEt) and isopropanol (HOiPr) was catalyzed by cinchona alkaloid catalyst. The cyanation reaction of oleﬁn proceeded through a two-step mechanism, in which the C-C bond construction was followed by H-transfer to generate a cyanide adduct. For non-catalytic reaction, the activation barrier for the rate-determining C-H bond construction step was 34.2 kcal mol − 1 , via a four-membered transition state. The self-assembly Ti(IV)-catalyst from tetraisopropyl titanate, ( R )-3,3 (cid:48) -disubstituted biphenol, and cinchonidine accelerated the addition of cyanide to the C = C double bond by a dual activation process, in which titanium cation acted as a Lewis acid to activate the oleﬁn and HNC was orientated by hydrogen bonding. The steric repulsion between the 9-phenanthryl at the 3,3 (cid:48) -position in the biphenol ligand and the Ph group in oleﬁn raised the Pauli energy ( ∆ E (cid:44) Pauli ) of reacting fragments at the re -face attack transition state, leading to the predominant R -product.


Introduction
The asymmetric catalytic cyanation of C=X bond (X = O, N or C) provides an outstanding method to obtain various optically active nitriles [1][2][3][4][5][6]. Compared to the intensively studied cyanation of aldehydes [1,2,5], ketones [1,2,5] and imine (Strecker reaction) [3], reports on the conjugate addition of cyanide to the C=C bond are limited [4,6]. Since the products from cyanide addition to C=C double bonds in the α,β-unsaturated carbonyl compounds could conveniently convert to the enantioenriched intermediates with great synthesizing value and pharmaceutical importance (e.g., γ-aminobutyric acids), developing straightforward synthetic procedures and exploring the relevant reaction mechanisms are in high demand.
Jacobsen's group reported the first catalytic asymmetric cyanation of α,β-unsaturated imides, using Al(III) complex with chiral salen ligand as a catalyst. Based on the kinetic analyses, they proposed that the reaction involved a bimetallic, dual activation process. The salen-Al(III) complex-activated cyanide was delivered to the electrophile bound as an imidate complex for highly enantioenriched cyanide adducts [7]. In a heterobimetallic system with (salen)Al and (pybox)Er complexes (pybox = 2,6-bis(2-oxazolinyl) pyridine), two catalysts operated cooperatively complexes (pybox = 2,6-bis(2-oxazolinyl) pyridine), two catalysts operated cooperatively in the ratedetermining step, promoting the conjugate addition in a highly enantioselective manner [8]. The poly(norbornene)-supported (salen)AlCl catalyst could also realize this transformation. The proximity of catalytic sites in polymeric Al-catalyst facilitated the reaction to occur via a bimetallic pathway [9]. Shibasaki et al. developed chiral gadolinium complex catalysts for the cyanation of α,βunsaturated N-acylpyrroles [10,11] and enones [12]. Mechanistic studies suggested that the reaction is carried out through an intramolecular cyanide transfer from the gadolinium cyanide to the activated N-acylpyrrole substrate. The protic additive (e.g., HCN) efficiently facilitated both catalyst activity and enantioselectivity. Other metal complexes containing Sr(II) [13], Ru(II)/Li(I) [14,15], Mg(II) [16,17] and Li(I) [18] were also active catalysts for this kind of reaction. The bifunctional catalysis model was proposed to interpret the activation mode as well as the stereochemical outcome [15,16]. Besides, the reaction could be realized by organocatalysis [19,20] or a phase-transfer process [21,22]. The spectroscopic studies by Khan et al. verified that the N-oxide additive participated in the cyanation of nitroalkenes as a ligand and activator of trimethylsilylcyanide (TMSCN) [23]. Experiments and density functional theory (DFT) calculations by Minakata and co-workers revealed that the cyanation of the boron enolates generated from α,β-unsaturated ketones with ptoluenesulfonyl cyanide (TsCN) proceeded through a six-membered ring transition state (TS) [24]. Based on the NMR spectroscopy results, they proposed that imidazolium cations interacted with the substrate, facilitating the attack of cyanide ions generated by the activation of acetone cyanohydrin by the acetate counter ion in 1-butyl-3-methylimidazolium (BMIM)-based ionic-liquid-catalyzed conjugate cyanation of CF3-substituted alkylidenemalonates [25].
In 2010, Feng's group developed a modular catalyst generated in situ from cinchona alkaloid, tetraisopropyl titanate (Ti(OiPr)4) and achiral 3,3′-disubstituted biphenol, achieving the efficient asymmetric cyanation of N-p-toluenesulfonyl aldimines and ketimines, as well as ketones and aldehydes [26,27]. Interestingly, this self-assembled catalyst system also exhibited excellent performance in the asymmetric cyanation of C=C bonds, using ethyl cyanoformate (CNCOOEt) as a cyanide source [28]. The enantioenriched cyanide adducts could be obtained with high yield (97%) and enantiomeric excess (ee) (up to 94%). In the reaction, the axial chirality of the biphenol ligand was induced in the formation of the complex to achieve asymmetric activation [29]. Based on previous work, they proposed that the catalytic species was (R)-biphenol, and chiral cinchonidine ligand coordinated to the Ti(IV) center simultaneously.
In this work, we employed DFT calculations to understand the mechanism for the asymmetric cyanation of activated olefin in detail (Scheme 1). The key factors controlling the enantioselectivity could be revealed to expedite the rational design of new Ti(IV)-complex catalysts. Scheme 1. Asymmetric cyanation of activated olefin catalyzed by Ti(IV)-complex.

Computational Details
DFT calculations were performed using the Gaussian 09 package [30] at the B3LYP-D3(BJ)/6-31G**(SMD, toluene) theoretical level. Geometries were optimized in toluene solvent and characterized by calculating the harmonic vibrational frequencies. The self-consistent reaction field

Release of HCN or HNC Species from CNCOOEt
The previous experimental and theoretical investigations suggested that HCN was the real cyanation reagent in the cyanation of imines with TMSCN catalyzed by Ti(IV)-complex [29]. Moreover, isopropyl alcohol (HOiPr) had a positive effect on the release of HCN from TMSCN, consequently accelerating the reaction [26,27,29,44]. Based on these results, we first studied the formation of HCN or HNC species from the reaction between HOiPr and the cyanide source, ethyl cyanoformate (CNCOOEt). Four possible pathways were considered, in which HCN was formed along paths 1 and 3, while its isomer HNC was afforded along paths 2 and 4. As shown in Scheme 2, when CNCOOEt approached HOiPr, the H atom of HOiPr transferred to the C and N atoms of CNCOOEt, yielding HCN and HNC species, respectively. The potential energy surfaces for these pathways are shown in Figure 1. The activation energies were 54.9 and 56.5 kcal mol −1 via 1-TS1 and 2-TS1, respectively. Besides, the reaction could also occur via other stepwise mechanisms along paths 3 and 4, in which the H atom transferred initially to the O atom of the carbonyl group in CNCOOEt having a more negative charge (−0.548 at the O atom vs. −0.501 at the CN moiety) via a four-membered ring transition state TS1. Then, this H atom transferred to the C and N atoms of the CN group through transition states 3-TS2 and 4-TS2, producing HCN (path 3) and HNC (path 4), respectively. The calculations indicated that the ∆G associated with the generation of HCN along path 3 and HNC along path 4 were 39.6 and 40.1 kcal mol −1 , respectively, which were lower than those along paths 1 and 2.  For comparison, the formation of HNC assisted by cinchona alkaloid (L1) was studied ( Figures  1 and 2). In the initial complex (C-I-COM), the cinchona alkaloid activated HOiPr by the N atom of the tertiary amine ring and OH group simultaneously with (O)H···O and N···H distances of 1.858 and 1.834 Å , respectively. These hydrogen bonds could be verified by atoms-in-molecules (AIM) analysis, with the positive Laplacian (∇ 2 ρ) on (3, −1) bond critical points (BCPs) ( Figure S1). In the next step, the O atom of HOiPr approached the C atom of CNCOOEt, accompanied with H transfer from the O atom to N atom via C-I-TS1. Then, HCN was formed by breaking the C-C bond via C-I-TS2, with a ΔG of 20.0 kcal mol −1 . From the viewpoint of energy, cinchona alkaloid could promote the transformation of CNCOOEt to HNC by organocatalysis.   For comparison, the formation of HNC assisted by cinchona alkaloid (L1) was studied ( Figures  1 and 2). In the initial complex (C-I-COM), the cinchona alkaloid activated HOiPr by the N atom of the tertiary amine ring and OH group simultaneously with (O)H···O and N···H distances of 1.858 and 1.834 Å , respectively. These hydrogen bonds could be verified by atoms-in-molecules (AIM) analysis, with the positive Laplacian (∇ 2 ρ) on (3, −1) bond critical points (BCPs) ( Figure S1). In the next step, the O atom of HOiPr approached the C atom of CNCOOEt, accompanied with H transfer from the O atom to N atom via C-I-TS1. Then, HCN was formed by breaking the C-C bond via C-I-TS2, with a ΔG of 20.0 kcal mol −1 . From the viewpoint of energy, cinchona alkaloid could promote the transformation of CNCOOEt to HNC by organocatalysis. For comparison, the formation of HNC assisted by cinchona alkaloid (L1) was studied (Figures 1  and 2). In the initial complex (C-I-COM), the cinchona alkaloid activated HOiPr by the N atom of the tertiary amine ring and OH group simultaneously with (O)H···O and N···H distances of 1.858 and 1.834 Å, respectively. These hydrogen bonds could be verified by atoms-in-molecules (AIM) analysis, with the positive Laplacian (∇ 2 ρ) on (3, −1) bond critical points (BCPs) ( Figure S1). In the next step, the O atom of HOiPr approached the C atom of CNCOOEt, accompanied with H transfer from the O atom to N atom via C-I-TS1. Then, HNC was formed by breaking the C-C bond via C-I-TS2, with a ∆G of 20.0 kcal mol −1 . From the viewpoint of energy, cinchona alkaloid could promote the transformation of CNCOOEt to HNC by organocatalysis.

Noncatalytic Reaction
The isomerization between HCN and HNC can occur quickly in the presence of cinchona alkaloid [29], establishing rapid HCN-HNC equilibrium. Then, HNC can act as an active cyanide species to construct a C-C bond by interacting with an olefin in the cyanation reaction. Based on these results, we first studied the mechanism of the noncatalytic cyanation of olefins with HNC ( Figure 3). The reaction occurred through a two-step process: C-C bond formation followed by C-H bond construction. In the initial step, HNC interacted with the COOEt moiety through hydrogen bonding,

Noncatalytic Reaction
The isomerization between HCN and HNC can occur quickly in the presence of cinchona alkaloid [29], establishing rapid HCN-HNC equilibrium. Then, HNC can act as an active cyanide species to construct a C-C bond by interacting with an olefin in the cyanation reaction. Based on these results, we first studied the mechanism of the noncatalytic cyanation of olefins with HNC ( Figure 3). The reaction occurred through a two-step process: C-C bond formation followed by C-H bond construction. In the initial step, HNC interacted with the COOEt moiety through hydrogen bonding, forming a molecular complex b-IM1. Then, the CN group attacked the olefin to form a C-C bond via a seven-membered ring transition state b-TS1. In the final step, the product P-S was formed by shifting an H atom to a C atom (tautomerization of enol to keto), with a ∆G of 34.2 kcal mol −1 . This H-transfer step was predicted to be the rate-determining step (RDS) in the background reaction. We also located the TS involving HCN as a proton donor (b-TS1-1, Figure S2). The relative Gibbs energy of b-TS1-1 was higher than that of b-TS2 (the highest point in the energy profile) by 8.4 kcal mol −1 . These results indicated that the noncatalytic reaction was difficult to achieve owing to a higher barrier.
Catalysts 2020, 10, x FOR PEER REVIEW 6 of 14 forming a molecular complex b-IM1. Then, the CN group attacked the olefin to form a C-C bond via a seven-membered ring transition state b-TS1. In the final step, the product P-S was formed by shifting an H atom to a C atom (tautomerization of enol to keto), with a ΔG ≠ of 34.2 kcal mol −1 . This H-transfer step was predicted to be the rate-determining step (RDS) in the background reaction. We also located the TS involving HCN as a proton donor (b-TS1-1, Figure S2). The relative Gibbs energy of b-TS1-1 was higher than that of b-TS2 (the highest point in the energy profile) by 8.4 kcal mol −1 .
These results indicated that the noncatalytic reaction was difficult to achieve owing to a higher barrier.

Catalytic Reaction
A previous study indicated that aldimine and HOiPr could coordinate to the Ti(IV), forming a reactive hexacoordinated Ti(IV)-complex in the Strecker reaction catalyzed by Ti(IV)-complex catalyst with cinchona alkaloid and achiral 3,3′-disubstituted 2,2′-biphenol ligands [29]. Based on experimental observation [28], a cinchonidine(L1)/Ti(IV)/(R)-biphenol (L2a) catalyst was employed as an active species in the reaction. Considering that there were two O-donors in the olefin substrate, a bidentate model was first studied in the present work, in which two carbonyl groups of olefin coordinated to the Ti(IV) center simultaneously, forming three possible hexacoordinated Ti(IV)complexes (d-I-d-III). For comparison, the monocoordinated models (m-I-m-VI) were also investigated (Scheme 3). Nine low-energy Ti(IV)-complexes in mono-and bidentate models were located to allow the favorable si-face attack pathway observed in the experiment (see Table 1 and Figure S3). Table 1 shows that a bidentate Ti(IV)-complex (d-I-COM-si) had the lowest Gibbs free energy among nine models. In other words, this complex was the only starting species available in the reaction mixture.

Catalytic Reaction
A previous study indicated that aldimine and HOiPr could coordinate to the Ti(IV), forming a reactive hexacoordinated Ti(IV)-complex in the Strecker reaction catalyzed by Ti(IV)-complex catalyst with cinchona alkaloid and achiral 3,3 -disubstituted 2,2 -biphenol ligands [29]. Based on experimental observation [28], a cinchonidine(L1)/Ti(IV)/(R)-biphenol (L2a) catalyst was employed as an active species in the reaction. Considering that there were two O-donors in the olefin substrate, a bidentate model was first studied in the present work, in which two carbonyl groups of olefin coordinated to the Ti(IV) center simultaneously, forming three possible hexacoordinated Ti(IV)-complexes (d-I-d-III). For comparison, the monocoordinated models (m-I-m-VI) were also investigated (Scheme 3). Nine low-energy Ti(IV)-complexes in mono-and bidentate models were located to allow the favorable si-face attack pathway observed in the experiment (see Table 1 and Figure S3). Table 1 shows that a bidentate Ti(IV)-complex (d-I-COM-si) had the lowest Gibbs free energy among nine models. In other words, this complex was the only starting species available in the reaction mixture.
Catalysts 2020, 10, x FOR PEER REVIEW 6 of 14 forming a molecular complex b-IM1. Then, the CN group attacked the olefin to form a C-C bond via a seven-membered ring transition state b-TS1. In the final step, the product P-S was formed by shifting an H atom to a C atom (tautomerization of enol to keto), with a ΔG ≠ of 34.2 kcal mol −1 . This H-transfer step was predicted to be the rate-determining step (RDS) in the background reaction. We also located the TS involving HCN as a proton donor (b-TS1-1, Figure S2). The relative Gibbs energy of b-TS1-1 was higher than that of b-TS2 (the highest point in the energy profile) by 8.4 kcal mol −1 .
These results indicated that the noncatalytic reaction was difficult to achieve owing to a higher barrier.

Catalytic Reaction
A previous study indicated that aldimine and HOiPr could coordinate to the Ti(IV), forming a reactive hexacoordinated Ti(IV)-complex in the Strecker reaction catalyzed by Ti(IV)-complex catalyst with cinchona alkaloid and achiral 3,3′-disubstituted 2,2′-biphenol ligands [29]. Based on experimental observation [28], a cinchonidine(L1)/Ti(IV)/(R)-biphenol (L2a) catalyst was employed as an active species in the reaction. Considering that there were two O-donors in the olefin substrate, a bidentate model was first studied in the present work, in which two carbonyl groups of olefin coordinated to the Ti(IV) center simultaneously, forming three possible hexacoordinated Ti(IV)complexes (d-I-d-III). For comparison, the monocoordinated models (m-I-m-VI) were also investigated (Scheme 3). Nine low-energy Ti(IV)-complexes in mono-and bidentate models were located to allow the favorable si-face attack pathway observed in the experiment (see Table 1 and Figure S3). Table 1 shows that a bidentate Ti(IV)-complex (d-I-COM-si) had the lowest Gibbs free energy among nine models. In other words, this complex was the only starting species available in the reaction mixture.  The reaction mechanisms of the cyanation of olefins in the presence of Ti(IV)-complex catalyst could be very similar for the nine coordination models, although the reaction might begin with different hexacoordinated Ti(IV)-complexes: C-C bond construction followed by H-shift. The potential energy surfaces for the si-face and re-face attack to produce Rand S-configuration enantiomers from d-I-COM-si and d-I-COM-re are shown in Figure 4. Like the noncatalytic reaction, the catalytic process occurred via a stepwise mechanism. Firstly, HNC was coordinated to the tertiary amine of cinchona alkaloid through hydrogen bonding, with an H···N distance of 1.612 Å . Then, the C-C bond was constructed by the attack of the CN group to olefin, with ΔG ≠ of 5.1 kcal mol −1 via the TS d-I-TS1-si. Finally, the catalytic cycle was finished when the H atom was transferred from the N to C atom, via the TS d-I-TS2-si. A protonic reagent (e.g., HOiPr) could accelerate proton shift by hydrogen-bonding [29], and the H-shift barrier for HOiPr via the TS d-I-TS2-si-HOiPr was decreased by 3.4 kcal mol −1 ( Figure S4). The activation barrier in the chiralcontrolling step (C-C bond formation step) in the catalysis was lower than that for the noncatalytic process by 14.2 kcal mol −1 . We also optimized the TSs in the chiral-controlling C-C bond construction along with the si-face attack in the d-II and d-III models ( Figure S5). As expected, the activation free energies via d-II-TS1-si (ΔG ≠ = 10.4 kcal mol −1 ) and d-III-TS1-si (ΔG ≠ = 7.9 kcal mol −1 ) were higher than that via d-I-TS1-si (ΔG ≠ = 5.1 kcal mol −1 ) in the d-I model. As shown in Figure 5, the ΔG of d-I-TS1-si was lower than that of d-I-TS1-re in a chiral-controlling step by 2.3 kcal mol −1 at 273 K, indicating the product with R-configuration was predominant. The theoretical enantioselectivity (ee %) was 96% using the Curtin-Hammett principle, which was close to that obtained experimentally (93% ee) [28]. Like the noncatalytic reaction, the catalytic process occurred via a stepwise mechanism. Firstly, HNC was coordinated to the tertiary amine of cinchona alkaloid through hydrogen bonding, with an H···N distance of 1.612 Å. Then, the C-C bond was constructed by the attack of the CN group to olefin, with ∆G of 5.1 kcal mol −1 via the TS d-I-TS1-si. Finally, the catalytic cycle was finished when the H atom was transferred from the N to C atom, via the TS d-I-TS2-si. A protonic reagent (e.g., HOiPr) could accelerate proton shift by hydrogen-bonding [29], and the H-shift barrier for HOiPr via the TS d-I-TS2-si-HOiPr was decreased by 3.4 kcal mol −1 ( Figure S4). The activation barrier in the chiral-controlling step (C-C bond formation step) in the catalysis was lower than that for the noncatalytic process by 14.2 kcal mol −1 . We also optimized the TSs in the chiral-controlling C-C bond construction along with the si-face attack in the d-II and d-III models ( Figure S5). As expected, the activation free energies via d-II-TS1-si (∆G = 10.4 kcal mol −1 ) and d-III-TS1-si (∆G = 7.9 kcal mol −1 ) were higher than that via d-I-TS1-si (∆G = 5.1 kcal mol −1 ) in the d-I model. As shown in Figure 5, the ∆G of d-I-TS1-si was lower than that of d-I-TS1-re in a chiral-controlling step by 2.3 kcal mol −1 at 273 K, indicating the product with R-configuration was predominant. The theoretical enantioselectivity (ee We also studied the catalytic mechanism starting from the monodentate Ti(IV)-complex m-I-COM-si for comparison, and the corresponding C-C bond construction TS (m-I-TS1-si) in the chiralcontrolling step was located ( Figure 5). Compared with d-I-COM-si, the olefin substrate in m-I-COMsi was slightly weakened, with a large Wiberg bond index of 1.664 for the C1=C2 bond. Accordingly, m-I-TS1-si was less stable than d-I-TS1-si by 11.6 kcal mol −1 . Thus, olefin tended to participate in the cyanation reaction in a bidentate fashion.

Origin of Stereoselectivity
The hexacoordinated Ti(IV)-complexes d-I-COM-si and d-I-COM-re had a pocket-like chiral cavity, with the dihedral angles formed by the 9-phenanthryl groups at the 3,3′-position in L2a of 93.7° and 87.3°, respectively. The topographic steric maps [45][46][47] of d-I-COM-si and d-I-COM-re are shown in Scheme 4, which characterizes the surface that the ligand L2a offers to the olefin substrate. The percentage of buried volume (%VBur) quantified the first coordination sphere around the Ti center occupied by L2a ligand. For d-I-COM-re, %VBur was 39.7, which was larger than that of d-I-COM-si (35.8). Importantly, the 9-phenanthryl groups provided stronger hindrance in the north-western and south-eastern quadrants (yellow colored area). Consequently, the unfavorable steric repulsion between the Ph group of olefin and 9-phenanthryl group of ligand L2a (in the south-eastern quadrant) became more significant in d-I-COM-re as well as in the corresponding C-C bond formation TS.
Then, we further analyzed the structures of d-I-TS1-si and d-I-TS1-re ( Figure 6). The Ph group of the olefin moiety in d-I-TS1-re was in proximity to the neighboring 3-substituted group in the biphenol ligand, with a Ph···Ph distance of about 2.5 Å . Accordingly, the steric repulsion raised the Pauli energy (∆E ≠ Pauli) of the reacting fragments at d-I-TS1-re (139.0 vs. 130.4 kcal mol −1 ) ( Table 2). Consequently, the ΔG of d-I-TS1-re was higher than that of d-I-TS1-si (7.4 vs. 5.1 kcal mol −1 ). This steric repulsion was also visualized by noncovalent interaction analysis using Multiwfn software [48], in which the larger yellow area between the Ph of olefin and phenanthryl group of L2a ( Figure S6) in d-I-TS1-re was observed. In contrast, this unfavorable steric interaction was avoided efficiently in d-I-TS1-si because the Ph groups were located far away. Consequently, two reacting fragments could interact easily, with a more stabilizing ΔE ≠ int (−5.6 kcal mol −1 ) and low reaction barrier of 5.1 kcal mol −1 in the C-C bond formation step. We also studied the catalytic mechanism starting from the monodentate Ti(IV)-complex m-I-COM-si for comparison, and the corresponding C-C bond construction TS (m-I-TS1-si) in the chiral-controlling step was located ( Figure 5). Compared with d-I-COM-si, the olefin substrate in m-I-COM-si was slightly weakened, with a large Wiberg bond index of 1.664 for the C1=C2 bond. Accordingly, m-I-TS1-si was less stable than d-I-TS1-si by 11.6 kcal mol −1 . Thus, olefin tended to participate in the cyanation reaction in a bidentate fashion.

Origin of Stereoselectivity
The hexacoordinated Ti(IV)-complexes d-I-COM-si and d-I-COM-re had a pocket-like chiral cavity, with the dihedral angles formed by the 9-phenanthryl groups at the 3,3 -position in L2a of 93.7 • and 87.3 • , respectively. The topographic steric maps [45][46][47] of d-I-COM-si and d-I-COM-re are shown in Scheme 4, which characterizes the surface that the ligand L2a offers to the olefin substrate. The percentage of buried volume (%V Bur ) quantified the first coordination sphere around the Ti center occupied by L2a ligand. For d-I-COM-re, %V Bur was 39.7, which was larger than that of d-I-COM-si (35.8). Importantly, the 9-phenanthryl groups provided stronger hindrance in the north-western and south-eastern quadrants (yellow colored area). Consequently, the unfavorable steric repulsion between the Ph group of olefin and 9-phenanthryl group of ligand L2a (in the south-eastern quadrant) became more significant in d-I-COM-re as well as in the corresponding C-C bond formation TS.
Then, we further analyzed the structures of d-I-TS1-si and d-I-TS1-re ( Figure 6). The Ph group of the olefin moiety in d-I-TS1-re was in proximity to the neighboring 3-substituted group in the biphenol ligand, with a Ph···Ph distance of about 2.5 Å. Accordingly, the steric repulsion raised the Pauli energy (∆E Pauli ) of the reacting fragments at d-I-TS1-re (139.0 vs. 130.4 kcal mol −1 ) ( Table 2). Consequently, the ∆G of d-I-TS1-re was higher than that of d-I-TS1-si (7.4 vs. 5.1 kcal mol −1 ). This steric repulsion was also visualized by noncovalent interaction analysis using Multiwfn software [48], in which the larger yellow area between the Ph of olefin and phenanthryl group of L2a ( Figure S6) in d-I-TS1-re was observed. In contrast, this unfavorable steric interaction was avoided efficiently in d-I-TS1-si because the Ph groups were located far away. Consequently, two reacting fragments could interact easily, with a more stabilizing ∆E int (−5.6 kcal mol −1 ) and low reaction barrier of 5.1 kcal mol −1 in the C-C bond formation step.   The influence of the 3,3 -substitute of biphenol on the stereoselectivity was further studied. When the 9-phenanthryl group in L2a was replaced by a 3,5-dimethyl phenyl group (L2b), an opening chiral pocket in d-I-COM-si-L2b was observed, with the dihedral angle of the two substituent groups at the 3,3 -position of 109.6 • . The relative energy of the two competing TSs (d-I-TS1-si-L2b and d-I-TS1-re-L2b) in the chiral-controlling step (i.e., C-C bond construction step) was comparable (7.3 vs. 7.2 kcal mol −1 ), affording racemic products (Figure 7). These results indicated that the bulky substituent at the 3,3 -position of the biphenol ligand was essential for asymmetric induction in the Ti(IV)-complex-catalyzed cyanation of activated olefins.  The influence of the 3,3′-substitute of biphenol on the stereoselectivity was further studied. When the 9-phenanthryl group in L2a was replaced by a 3,5-dimethyl phenyl group (L2b), an opening chiral pocket in d-I-COM-si-L2b was observed, with the dihedral angle of the two substituent groups at the 3,3′-position of 109.6°. The relative energy of the two competing TSs (d-I-TS1-si-L2b and d-I-TS1-re-L2b) in the chiral-controlling step (i.e., C-C bond construction step) was comparable (7.3 vs. 7.2 kcal mol −1 ), affording racemic products (Figure 7). These results indicated that the bulky substituent at the 3,3′-position of the biphenol ligand was essential for asymmetric induction in the Ti(IV)-complex-catalyzed cyanation of activated olefins. Optimized geometries of TSs in the C-C bond formation step (d-I-TS1-si-L2b and d-I-TS1-re-L2b) in the cyanation reaction of olefins catalyzed by Ti(IV)-complex with L2b along two competing pathways for R-and S-configuration products, respectively, associated with the relative Gibbs free energies. The intermolecular distance is in Angstroms (Å). The color definitions of atoms are red = oxygen, blue = nitrogen, gray = carbon and white = hydrogen.

Conclusions
DFT calculations on the reaction mechanism of asymmetric cyanation of activated olefincatalyzed Ti(IV)-complexes revealed the following results: (i). Cinchona alkaloid facilitated the reaction between HOiPr and ethyl cyanoformate (CNCOOEt) to release the reacting species HCN (or HNC) by organocatalysis with ∆G ≠ of 26.0 kcal mol −1 . (ii). The cyanation reaction of olefin proceeded via a two-step mechanism, in which the C-C bond construction was followed by H-transfer to generate a cyanide adduct. For noncatalytic reaction, the ∆G ≠ for the rate-determining C-H bond construction step was up to 34.2 kcal mol −1 , through a four-membered TS. In the catalytic reaction, the olefin coordinated to the self-assembly cinchonidine/Ti(IV)/(R)-3,3′-disubstituted biphenol catalyst in the bidentate model, forming a highly reactive hexacoordinated Ti(IV)-complex. The HNC activated by the quinuclidine tertiary amine moiety of cinchonidine ligand performed a nucleophilic attack towards the activated C=C bond of olefin, generating a cyanide adduct. The catalytic reaction required about 19.9 kcal mol −1 lower energy barrier compared to the noncatalytic reaction. (iii). EDA showed that the steric repulsion between the bulky group (e.g., 9-phenanthryl substituent) at the 3-position in the biphenol ligand and the phenyl group in olefin raised the Pauli energy

Conclusions
DFT calculations on the reaction mechanism of asymmetric cyanation of activated olefin-catalyzed Ti(IV)-complexes revealed the following results: (i). Cinchona alkaloid facilitated the reaction between HOiPr and ethyl cyanoformate (CNCOOEt) to release the reacting species HCN (or HNC) by organocatalysis with free energy barrier of 26.0 kcal mol −1 . (ii). The cyanation reaction of olefin proceeded via a two-step mechanism, in which the C-C bond construction was followed by H-transfer to generate a cyanide adduct. For noncatalytic reaction, the ∆G for the rate-determining C-H bond construction step was up to 34.2 kcal mol −1 , through a four-membered TS. In the catalytic reaction, the olefin coordinated to the self-assembly cinchonidine/Ti(IV)/(R)-3,3 -disubstituted biphenol catalyst in the bidentate model, forming a highly reactive hexacoordinated Ti(IV)-complex. The HNC activated by the quinuclidine tertiary amine moiety of cinchonidine ligand performed a nucleophilic attack towards the activated C=C bond of olefin, generating a cyanide adduct. The catalytic reaction required about 19.9 kcal mol −1 lower energy barrier compared to the noncatalytic reaction.
(iii). EDA showed that the steric repulsion between the bulky group (e.g., 9-phenanthryl substituent) at the 3-position in the biphenol ligand and the phenyl group in olefin raised the Pauli energy (∆E Pauli ) of the reacting fragments at the re-face attack TS, leading to the predominant R-product through the si-face attack, as observed in the experiment.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4344/10/9/1079/s1, Figure S1: Laplacian (∇ 2 ρ) and electronic density (ρ, in parentheses) of selected bond critical points (BCP) for molecular complex C-I-COM were obtained by AIM analysis, using Multiwfn software. Figure S2: Optimized geometry of transition state b-TS1-1. The intermolecular distance is in Angstroms (Å). The color definitions of atoms are Red = oxygen, blue = nitrogen, gray = carbon, and white = hydrogen. Figure S3: Optimized geometries of low-energy hexacoordinated Ti(IV)-complexes formed by coordinating olefin to metal center in (a) bidentate (b) monodentate fashion along si-face attack pathway. The intermolecular distance is in Angstroms (Å). The color definitions of atoms are Red = oxygen, blue = nitrogen, gray = carbon, and white = hydrogen. Figure S4: Optimized geometries of hexacoordinated Ti(IV)-complexes and H-shift transition state in the presence of HOiPr along si-face attack pathway. The intermolecular distance is in Angstroms (Å). The color definitions of atoms are Red = oxygen, blue = nitrogen, gray = carbon, and white = hydrogen. Figure S5: Optimized geometries of two competing transition states in C-C bond formation step in d-II and d-III models as well as their relative Gibbs free energies (in kcal mol −1 ). The intermolecular distance is in Angstroms (Å). The color definitions of atoms are Red = oxygen, blue = nitrogen, gray = carbon, and white = hydrogen. Figure S6: Visualization of the main noncovalent interaction described by contour plots of the reduced density gradient isosurfaces (density cutoff of 0.7 au) for transition states d-I-TS1-si and d-I-TS1-re. The surface color code is blue for strongly attractive, green for weakly attractive, and red for strongly repulsive interactions.