Theoretical Investigations on the Mechanism of Benzoin

Feb 9, 2011 - paths A and B. Our results are in nice agreement with the experimental result in a kinetic ... that the dimer could not catalyze the ben...
2 downloads 0 Views 4MB Size
ARTICLE pubs.acs.org/JPCA

Theoretical Investigations on the Mechanism of Benzoin Condensation Catalyzed by Pyrido[1,2-a]-2-ethyl[1,2,4]triazol3-ylidene Yunqing He†,‡ and Ying Xue*,†,§ †

College of Chemistry, Sichuan University, Chengdu 610064, People’s Republic of China Department of Chemistry and Engineering, Sichuan University of Arts and Science, Dazhou 635000, People’s Republic of China § State Key Laboratory of Biotherapy, Sichuan University, Chengdu 610041, People’s Republic of China ‡

bS Supporting Information ABSTRACT: A new annulated N-heterocyclic carbene (NHC), pyrido[1,2-a]-2-ethyl[1,2,4]triazol-3-ylidene, has been synthesized and its good catalytic activity for benzoin condensation has been experimentally determined by You and co-workers recently [Ma, Y. J.; Wei, S. P.; Lan, J. B.; Wang, J. Z.; Xie, R. G.; You, J. S. J. Org. Chem. 2008, 73, 8256]. In this work, the mechanism of the title reaction has been intensively studied computationally by employing the density functional theory (B3LYP) method in conjunction with 6-31þG(d) and 6-311þG(2d,p) basis sets. Our results indicate that path A (in which a sequence of intermolecular proton transfers between two carbene/ benzaldehyde coupling intermediates affords enamine) and path B (in which a t-BuOH assisted hydrogen transfer generates enamine) proposed on the basis of the Breslow mechanism are competitive for their similar barriers. In path A, the first intermolecular proton transfer between two N-heterocyclic carbene/benzaldehyde coupled intermediates to form tertiary alcohol and enolate anion is theoretically the rate-determining step with corresponding barrier (30.93 kcal/mol), while the t-BuOH assisted hydrogen transfer generating Breslow enamine is the rate-determining step with corresponding barrier (28.84 kcal/mol) in path B. The coupling of carbene and benzaldehyde, and the coupling of enamine and another benzaldehyde to form a C-C bond are partially rate-determining for their relatively significant barriers (24.06 and 26.95 kcal/mol, respectively), being the same in both paths A and B. Our results are in nice agreement with the experimental result in a kinetic investigation of thiazolium ion-catalyzed benzoin condensation performed by White and Leeper in 2001.

1. INTRODUCTION Recently, You and co-workers synthesized pyrido[1,2-a][1,2,4]triazol-3-ylidenes as a new family of stable annulated N-heterocyclic carbenes (NHC).1 They experimentally demonstrated that pyrido[1,2-a]-2-ethyl[1,2,4]triazol-3-ylidene (1 in Scheme 1, R = ethyl) can prompt the benzoin condensation (Scheme 1) with manifest catalytic activity with 93% yield when treating benzaldehyde in THF at 25 °C for 15 h, while its carbene dimer (2 in Scheme 1) cannot catalyze the benzoin condensation. The benzoin condensation is of much interest and has been intensively investigated for its important role in convenient carbon-carbon bond formation affording an important synthon R-hydroxycarbonyl group.1-24 The premier study may date back to 1832 when W€ohler and Liebig discovered the cyanidecatalyzed condensation of benzaldehyde to produce benzoin.2 In 1943, Ugai et al.3 ascertained thiazolium salts could also catalyze the benzoin condensation. A mechanism of an addition/ proton-transfer/condensation sequence was reported for cyanidecatalyzed benzoin condensation by Lapworth, in which a cyanohydrin intermediate was postulated as the active nucleophilic species.4,5 And in 1958, Breslow6,7 proposed a similar mechanism for a carbene compound thiazolin-2-ylidene-catalyzed benzoin r 2011 American Chemical Society

condensation on the basis of Lapworth’s studies, in which the active nucleophilic species is an enaminol-type Breslow intermediate similar to the cyanohydrin intermediate in Lapworth mechanism. Debate over the carbene-catalyzed benzoin condensation has led to other different mechanism models. Lemal et al.8 proposed a dimer-catalyzed mechanism (a variant of the Breslow mechanism) on the basis of the facility of carbene dimer formation, in which the carbene dimer acts as nucleophilic species in the first steps, and the corresponding Breslow enamine intermediate was given by the broken of the central C-C bond of the adduct. Lopez-Calahorra and co-workers9-12 developed a dimer mechanism without enamine-like intermediate appearing. The monomer- and dimer-catalyzed competing models have been intensively discussed; finally, the Breslow mechanism triumphed over the dimer ones.1,13-22 You and co-workers1 have gotten a tremendous chance to find some evidence of the mechanism of carbene-catalyzed benzoin condensation since they obtained a carbene (1 in Scheme 1) and its dimer (2 in Received: October 29, 2010 Revised: January 14, 2011 Published: February 9, 2011 1408

dx.doi.org/10.1021/jp110352e | J. Phys. Chem. A 2011, 115, 1408–1417

The Journal of Physical Chemistry A Scheme 1. Pyrido[1,2-a]-2-ethyl[1,2,4]triazol-3-ylideneCatalyzed Benzoin Condensation

Scheme 1) in solutions at room temperature. They confirmed that the dimer could not catalyze the benzoin condensation. Till now, the Breslow mechanism is accepted as definitive.23 In 2001, White and Leeper24 performed experimentally the thorough kinetic study for 3-benzyl-5-(2-hydroxyethyl)-4methylthiazolium bromide catalyzed benzoin condensation in methanol buffered with Et3N/Et3NHþ-Cl- and verified that the three steps (the addition of the thiazolium to benzaldehyde, proton transer to form enamine intermediate, and the final condensation) of the Breslow mechanism are partially ratedetermining. Previous calculations12,20,25,26 to the carbenecatalyzed reactions usually presumed a direct 1,2 proton shift in second step to provide Breslow enamine intermediate (similar to the process presented in Scheme S1 of Supporting Information). However, this is a high barrier process of the enamine generation for the large strain associated with the three-membered transition state,27 being unreasonable and inconsistent with the kinetic investigation by White and Leeper.24 In 2008, Hawkes and Yates27 reported theoretically the mechanism of the Stetter reaction, where an intermolecular hydrogen transfer sequence between two N-teterocyclic carbene/benzaldehyde coupled intermediates was verified almost certainly to form enamine with relatively low energy barriers. And in 2009, Domingo et al.28 performed a mechanism study of the N-heterocyclic carbene-catalyzed ring-expansion of 4-formyl-β-lactams to succinimide derivatives at the B3LYP/6-31G (d,p) level, computationally determined that a proton-transfer process assisted by methanol affords the Breslow enamine intermediate with relative low barrier. On the other hand, Martí et al.12 performed a detailed calculation of the Breslow mechanism, Lemal-like mechanism, and the dimer mechanism suggested by LopezCalahorra and co-workers9-12 for the thiamin-catalyzed benzoin condensation using the AM1 semiempirical molecular orbital method in 1995, and figured out that the dimer mechanism with a biradical character and without enamine-like intermediate is reasonable for justifying that the benzoin condensation can carry out in nonpolar solvent. In our previous work about the cyanidecatalyzed benzoin condensation in aprotic solvents DMSO,29 the dominant pathway without protonic solvent assistance was determined at the B3LYP/6-31þG(d,p) level, in which the formation of the cyanohydrin intermediate proceeds through the stepwise intermolecular proton transfers assisted by the reactant benzaldehyde. So, it is necessary to have an intensive theoretical study to find some channel to decrease the barrier of Breslow enamine intermediate formation or some other possible mechanism to reduce the barriers of the overall reaction for the

ARTICLE

N-heterocyclic carbene-catalyzed benzoin condensation. In this study, a computational investigation of the mechanism for pyrido[1,2-a]-2-ethyl[1,2,4]triazol- 3-ylidene-catalyzed benzoin condensation is performed using the density functional theory (DFT) method. Since You et al.1 experimentally verified that the carbene dimer cannot catalyze the benzoin condensation, we do not consider any carbene-dimer catalyzed mechanism in this work. At first, the direct 1,2 proton-shift pathway to provide Breslow enamine intermediate through a three-membered transition state and the postulated pathway without enamine intermediate are examined and ruled out for their relatively high barriers (presented in Schemes S1 and S2 of the Supporting Information). So, only two possible pathways (paths A and B in Scheme 2) proposed on the basis of the Breslow mechanism are minutely investigated in this paper: a sequence of intermolecular proton transfers between two N-teterocyclic carbene/benzaldehyde coupled intermediates is proposed to form Breslow enamine in path A, while a proton-transfer process assisted by t-BuOH molecule (which is generated in reaction system) affords the Breslow enamine intermediate in path B.

2. COMPUTATIONAL DETAILS The structure optimizations for all stationary points, including reactant complex (RC), product complex (PC), transition state (TS), and intermediate (IM), were performed in the gas phase using the density functional theory B3LYP method30,31 in conjunction with 6-31þG(d) basis set considering the computational cost and the accuracy, thanks to the successful application of similar methods in analogous systems.26,32,33 Each stationary point was confirmed by the harmonic frequency analysis at the same calculational level as a true minimum with no imaginary frequency or a transition state with only one imaginary frequency. The frequency calculations without scaling also provided the thermodynamic quantities such as the zero-point vibrational energy, thermal correction, enthalpies, Gibbs free energies, and entropies at temperature of 298.15 K and pressure of 101325 Pa. Intrinsic reaction coordinate (IRC)34,35 calculations were performed at the same level of theory to verify the correct relationship between the transition states and corresponding the relevant reactants and products. Actually, the reaction takes place in solvent, so solvation effect was considered here. Directly using the geometries optimized in gas phase, the solvent effect of tetrahydrofuran (THF) was evaluated at B3LYP/6-31þG(d) and B3LYP/6-311þG(2d,p) levels using the polarizable continuum model (PCM).36,37 Energies from these PCM single point calculations were combined with the thermodynamic corrections in the gas phase to obtain the Gibbs free energies in THF. Each activation energy barrier results from the free energy difference between the transition state and its former species. Molecular orbital analysis, natural population analysis (NPA)38,39 were performed at B3LYP/6-31þG(d) level. All calculations were carried out using Gaussian 03 program.40 3. RESULTS AND DISCUSSIONS In this section, paths A and B depicted in Schemes 2 are minutely described from structural and energetic features. To present the pathways of the title reaction conveniently, the prefixes A, B, S1 and S2 are used to identify the stationary structures that pertain only to path A, path B, the direct 1,2 proton-shift related to a three-membered transition state to provide Breslow enamine intermediate (see Scheme S1, Supporting Information), 1409

dx.doi.org/10.1021/jp110352e |J. Phys. Chem. A 2011, 115, 1408–1417

The Journal of Physical Chemistry A

ARTICLE

Scheme 2. Paths A and B of the Pyrido[1,2-a]-2-ethyl[1,2,4]triazol-3-ylidene-Catalyzed Benzoin Condensation

and the postulated pathway without enamine intermediate (see in Scheme S2, Supporting Information), respectively, while the prefix AB stands for the corresponding stationary points shared in paths A and B. For example, S1-TS2 denotes the threemembered transition state to provide Breslow enamine intermediate (see Scheme S1, Supporting Information), while AB-IM2 presents an intermediate shared in paths A and B. The stationary point shared in the common first step of two paths has no prefix. The numbering of the involved atoms in the following discussions is presented in corresponding figures of optimized structures. Because the sequence affording Breslow enamine in path A is performed between two carbene/benzaldehyde coupling intermediates, a prime (0 ) is used at the top right corner of an atomic numbering to identify the atom coming from one or another carbene/benzaldehyde intermediate. And because the carbene/benzaldehyde intermediate reacting with ButO- affording

t-BuOH is not the same one to react with t-BuOH generating enamine, another prime (00 ) is used to identify the atom coming from one or another carbene/benzaldehyde intermediate in path B. The relative free energies in THF at B3LYP/6-31þG(d) and B3LYP/6-311þG(2d,p) levels for all the species involved in paths A and B are presented in Table 1, and Table 2 shows the free energy barriers for the elementary steps of paths A and B in THF at B3LYP/6-31þG(d) and B3LYP/6-311þG(2d,p) levels. Some crucial NPA charges are exhibited in Table S1 of the Supporting Information. The optimized Cartesian coordinates of all the stationary points and the imaginary vibrational frequencies of all transition states along the potential energy surface (PES) of the processes shown in Scheme 2, Scheme S1, and Scheme S2 are also given in the Supporting Information. We use the free energies in THF at B3LYP/6-311þG(2d,p) level in the following discussions for convenience. 1410

dx.doi.org/10.1021/jp110352e |J. Phys. Chem. A 2011, 115, 1408–1417

The Journal of Physical Chemistry A

ARTICLE

Table 1. Relative Free Energies of the Species for Paths A and B in Tetrahydrofuran at B3LYP/6-31þG(d) and B3LYP/ 6-311þG(2d,p) Levels (Unit: kcal/mol, at 298.15 K) species

B3LYP/6-311þ

B3LYP/

G(2d,p)

6-31þG(d)

Table 2. Changes of the Free Energies for Paths A and B in Tetrahydrofuran at B3LYP/6-31þG(d) and B3LYP/ 6-311þG(2d,p) Levels (Unit: kcal/mol, at 298.15 K) steps

Common First Step (Relative to NHC þ R) NHC þ R

0.00

RC

11.52

11.26

TS1 IM1

24.06 19.40

23.48 18.77

NHC þ R f TS1

2IM1 þ R þ t-BuOH

0.00

0.00

A-IM1 þ R þ t-BuOH

15.53

15.72

A-TS2 þ R þ t-BuOH

30.93

31.58

A-IM2 þ R þ t-BuOH

9.93

12.41

A-TS3 þ R þ t-BuOH

19.17

23.13

13.35

18.82

-5.26 5.98

1.39 12.03

AB-TS4 þ AB-IM2 þ t-BuOH

21.69

26.86

AB-IM4 þ AB-IM2 þ t-BuOH

18.13

22.51

AB-TS5 þ AB-IM2 þ t-BuOH

20.26

24.82

AB-PC þ AB-IM2 þ t-BuOH

1.72

7.17 -

Step Generating t-BuOH (Relative to IM1 þ t-BuO ) IM1 þ t-BuO-

0.00

0.00

B-TS2

23.38

24.28

B-IM2

-15.69

-12.77

Path B (Relative to 2 IM1 þ t-BuOH þ R) 2IM1 þ t-BuOH þ R B-TS3 þ R þ IM1 AB-IM2 þ R þ t-BuOH þ IM1

0.00

0.00

28.84

29.04

-2.63

0.14

AB-IM3 þ t-BuOH þ IM1

8.62

10.78

AB-TS4 þ t-BuOH þ IM1

24.33

25.61

AB-IM4 þ t-BuOH þ IM1

20.77

21.27

AB-TS5 þ t-BuOH þ IM1 AB-PC þ t-BuOH þ IM1

22.90 4.36

23.58 5.93

3.1. Common First Step of the Carbene/Benzaldehyde Coupling Intermediate Formation. As illustrated in

Schemes 2, S1 and S2, the common first step shared in the processes is that the nucleophilic carbene (pyrido[1,2-a]-2ethyl[1,2,4]triazol-3-ylidene) attacks the carbonyl carbon atom in benzaldehyde, leading to a carbene/benzaldehyde coupling intermediate (IM1) including the tetrahedral carbon (C1). On the potential energy surface of this nuclephilic addition process, a possible transition state (TS1) is located. The optimized structures of the reactant (benzaldehyde), the nucleophilic N-heterocyclic carbene (NHC), and other stationary points in this step are shown in Figure 1, with some important bond lengths marked in angstrom. The distances of the C1-C2 bond in TS1 is 1.896 Å, suggesting the C1-C2 bond is forming in this transition state. The displacement vector of the imaginary vibrational frequency in TS1 for the addition step mainly corresponds to the addition of C2 in the carbene to C1 of the benzaldehyde. And in IM1, the C1-C2 bond (1.584 Å) has formed completely with simultaneous elongation of C1-O1, C1-C3, and C1-H1 bonds. The

24.06

23.48

IM1 f A-TS2

30.93

31.58

A-IM2 þ R f A-TS3 AB-IM2 þ R f AB-TS4

9.24 26.95

10.72 25.47

AB-IM4 þ R f AB-TS5

2.13

2.31

IM1 þ t-BuO- f B-TS2

23.38

24.28

IM1 þ t-BuOH f B-TS3

28.84

29.04

Path A

Path A (Relative to 2 IM1 þ R þ t-BuOH)

2AB-IM2 þ R þ t-BuOH AB-IM3 þ AB-IM2 þ t-BuOH

B3LYP/ 6-31þG(d)

IM1 Formation Step

0.00

A-IM3 þ R þ t-BuOH

B3LYP/ 6-311þG(2d,p)

Path B

2

sp hybridized C1 in benzaldehyde converts to a tetrahedral carbon (uneven sp3 hybridization) in IM1 with sp2.45 (O1), sp4.06(C2), sp2.61(C3), and sp3.22(H1) hybrid orbitals, respectively, taking part in the bond formation. NPA charge analysis shows the triazole ring is of rich negative charge but the carbene carbon atom is not negative (NPA charge on C2 is 0.12e in NHC) with a slight positive charge. When the carbene adds to the more positive C1 (NPA charge on C1 is 0.40e in R) of benzaldehyde, the negative charge of the triazol ring transfers partially to the oxygen atom of the carbonyl group of benzaldehyde, leading to the evident decreasing of positive charge on C1 (0.40e in R vs 0.09e in IM1) and the increasing of negative charge on O1 (-0.54e in R vs -0.87e in IM1) with simultaneous decreasing the negative charge on nitrogen atoms (-0.26e in NHC vs -0.19e in IM1 on N1, and -0.49e in RC vs -0.37e in IM1 on N2) and increasing positive charge on C2 (0.12e in NHC vs 0.46e in IM1). RC is a optimized reactant complex verified by IRC, but one can see from Table 1 that RC cannot exist steadily for its free energy is 11.52 kcal/mol higher than that of separating reactants (NHC þ R). So the free energy barrier of this first step should equal to the free energy difference between TS1 and the separating reactants (NHC þ R). As given in Table 2 and Figure 2, the free energy barrier of this common first step affording carbene/benzaldehyde coupling intermediate in tetrahydrofuran is 24.06 kcal/mol. 3.2. Path A. Considering that the regretful large strain associated to the three-membered transition state S1-TS2 (see Scheme S1 of the Supporting Information) results in the high barrier of the intramolecular step giving the Breslow enamine intermediate (AB-IM2), we deem that an intermolecular process probably decreases the barrier. Enlightened by prominent study of Hawkes and Yates,27 we proposed this pathway (path A presented in Scheme 2) based on the Breslow mechanism in which a similar intermolecular sequence between two N-heterocyclic carbene/benzaldehyde coupled intermediates affording enamine involved for the title reaction. As shown in Figures 3-5, the other four transition states (A-TS2, A-TS3, AB-TS4, and AB-TS5) were located along the potential energy surface of path A with the exception of TS1 presented supra. The optimized geometries and some important bond lengths of the stationary points in this pathway are exhibited in Figures 3 and 4. First, two sequent intermolecular steps of hydrogen-shift (H1 transfers from C1 to O10 with successive transfer of H10 from C10 to O1, 1411

dx.doi.org/10.1021/jp110352e |J. Phys. Chem. A 2011, 115, 1408–1417

The Journal of Physical Chemistry A

ARTICLE

Figure 1. Optimized structures of benzaldehyde and the stationary points in the step of carbene/benzaldehyde reactant complex formation common in all the pathways (bond length in Å; the atomic numberings marked in R, NHC, and RC).

Figure 2. Free energy profile of the common first step affording the carbene/benzaldehyde coupling intermediate (IM1) in THF at B3LYP/ 6-311þG(2d,p) level (unit: kcal/mol, free energies relative to NHC þ R).

through corresponding transition states A-TS2 and A-TS3, respectively) between the two molecules of the carbene/benzaldehyde intermediates (IM1) result in two molecules of Breslow enamine (A-IM3, which includes two molecules of enamine AB-IM2). And then, an addition of C1 (or C10 ) of Breslow enamine (AB-IM2) to C4 of another benzaldehyde molecule is performed with simultaneous transferring of the hydrogen atom H10 (or H1) of hydroxyl in AB-IM2 to the oxygen (O2) of carbonyl in another benzaldehyde molecule via a transition state (AB-TS4) (we choose one Breslow enamine molecule generated

in the intermolecular sequence in the following discussions for convenience). Finally, the elimination of the catalyst (pyrido[1,2a]-2-ethyl[1,2,4]triazol-3-ylidene) is performed via a transition state (AB-TS5) resulting in benzoin. As shown in Figure 3: first, H1 of one carbene/benzaldehyde complex molecule transfers from C1 to the carbonyl oxygen (O10 ) of another carbene/benzaldehyde complex molecule via a transition state (A-TS2), generating a tertiary alcohol molecule and a enolate anion molecule; second, the hydrogen (H10 ) on the tertiary carbon (C10 ) of the tertiary alcohol transfers to O1 of the enolate anion via a transition state (A-TS3) forming two molecules of enamine. In A-TS2, the distances of C1-H1 and O10 -H1 are 1.339 and 1.281 Å, respectively, suggesting H1 is transferring from C1 to O10 . The displacement vector of the imaginary vibrational frequency of A-TS2 mainly corresponds to the transfer of H1 from C1 to O10 . After the barrier of A-TS2 is overcome, the intermediate A-IM2 is obtained. In A-IM2, the distances of C1-H1 and O10 -H1 are 4.459 and 1.021 Å, respectively, indicating H1 has transferred to O1 completely. A-IM2 actually includes a molecule of tertiary alcohol and a molecule of enolate anion, in which the hybridization of C1 is sp2 and that of C10 is sp3. There is a hydrogen bond (O10 H1 3 3 3 N1, the distances O10 -H1 and H1 3 3 3 N1 are 1.021 and 1.686 Å) in A-IM2. In A-TS3, the distances of C10 -H10 1412

dx.doi.org/10.1021/jp110352e |J. Phys. Chem. A 2011, 115, 1408–1417

The Journal of Physical Chemistry A

ARTICLE

Figure 3. Optimized structures of the stationary points of the intermolecular proton transfer sequence between two carbene/benzaldehyde coupled intermediates affording Breslow enamine intermediate in path A (bond length in Å).

and O1-H10 are 1.399 and 1.231 Å, indicating the hydrogen (H10 ) is transferring to O1. The displacement vector of the imaginary vibrational frequency of A-TS3 mainly corresponds to the transferring of H10 from C10 of the tertiary alcohol to O1 of the carbonyl in enolate anion. After the barrier of A-TS3 is overcome, the intermediate (A-IM3) including two molecules of Breslow enamine (AB-IM2) is obtained. The hybridizations of C10 and C1 convert from sp3 in IM1 to sp2 in A-IM3. As presented in Table 1, the complex A-IM1 containing two carbene/ benzaldehyde coupling molecules cannot exist stably for its higher free energy than the sum of two single IM1 molecules by 15.53 kcal/mol, so we deem this intermolecular sequence begins from two separating IM1 molecules. As illustrated in Table 2, the free energy barrier of the first intermolecular proton transfer to form a molecule of tertiary

alcohol and a molecule of enolate anion is 30.93 kcal/mol, and that of the second intermolecular proton transfer to produce the final two molecules of Breslow enamine is 9.24 kcal/mol, both evidently lower than that of the direct intramolecular 1,2-proton transfer to form the Breslow enamine (45.23 kcal/mol, see Scheme S1, Supporting Information). Symmetry allowing and charge distributions result in the relatively low barriers of the sequence. The slight positive charges on C1 (0.09e in IM1) and the fruitful negative charge on O10 (i.e., -0.87e on O1 in IM1) implies O10 can capture H1 from C1 easily; and the electron-rich O1 (-0.83e in A-IM2) can capture H10 from the electrondeficient C10 (0.04e on C1 in A-IM2) facilely. After the formation of the Breslow enamine, the carbonyl oxygen atom (O2) of another benzaldehyde approaches the hydrogen (H1 or H10 ) of the hydroxyl in AB-IM2. For 1413

dx.doi.org/10.1021/jp110352e |J. Phys. Chem. A 2011, 115, 1408–1417

The Journal of Physical Chemistry A

ARTICLE

Figure 4. Optimized structures of the stationary points after Breslow enamine intermediate formation in path A (bond length in Å).

Figure 5. Free energy profiles of paths A and B in THF at B3LYP/6-311þG(2d,p) level (unit: kcal/mol, energies relative to 2IM1 þ t-BuOH þ R).

convenience, we choose one Breslow enamine molecule (which includs C1, O1 and H10 ) generated in the intermolecular sequence in the following discussions, while the orther molecule (which includs C10 , O10 and H1) is similar to it. As presented in Figure 4, the hydrogen (H10 ) of the hydroxyl in enamine shifts

from O1 to oxygen atom (O2) in another aldehyde molecule with simultaneous approaching of the C1 atom to C4 via a fivemembered ring transition state (AB-TS4), which is in fact an addition-like transition state from the structure. In AB-TS4, the distances of C1-C4, C4-O2, O2-H10 , H10 -O1, and O1-C1 1414

dx.doi.org/10.1021/jp110352e |J. Phys. Chem. A 2011, 115, 1408–1417

The Journal of Physical Chemistry A

ARTICLE

Figure 6. Optimized structures of stationary points in path B (bond length in Å).

are 1.833, 1.334, 1.320, 1.129, and 1.378 Å, respectively, evidently indicating that the C1-C4 bond is forming and H10 is transferring from O1 to O2. The displacement vectors of the imaginary vibrational frequency of AB-TS4 mainly correspond to the transfer of H10 from O1 to O2 and the approaching of C1 atom to C4. After the barrier of AB-TS4 is overcome, an intermediate (AB-IM4) is generated. In AB-IM4, the distances of C1-C4 and O2-H10 are 1.594 and 0.988 Å, revealing the C1-C4 bond has formed and the O2-H10 bond is completely formed, and the C1-O1 bond length is shortened (1.409 Å in AB-IM2 vs 1.338 Å in AB-IM4) with simultaneous elongation of the C4-O2 bond (1.426 Å in AB-IM4 vs 1.218 Å in R (corresponds to C1-O1 in Fingure 1)) and the C1-C2 bond (1.373 Å in AB-IM2 vs 1.609 Å in AB-IM4). The NPA charges on C1 and C4 are 0.24e and 0.05e in AB-IM4, suggesting the positive charge on C1 in benzaldehyde has been partially restored and the positive C4

in benzaldehyde converts to a nearly neutral one. The hybridization of C4 converts from sp2 in benzaldehyde to sp3 in AB-IM4, while that of C1 converts from sp2 in AB-IM2 to uneven sp3 in AB-IM4. At last, the pyrido[1,2-a]-2-ethyl[1,2,4]triazol-3-ylidene is eliminated from AB-IM4 via the corresponding transition state (AB-TS5). In AB-TS5, the distance of C1-C2 is 1.929 Å, suggesting that pyrido[1,2-a]-2-ethyl[1,2,4]triazol-3-ylidene is eliminating in this transition state. The displacement vector of the imaginary vibrational frequency of AB-TS5 mainly corresponds to the breaking of the C1-C2 bond. Consequently, the catalyst (i.e., pyrido[1,2-a]-2-ethyl[1,2,4]triazol-3-ylidene) of the title reaction is regenerated with simultaneous bond length shortening of C1-C4 (1.594 Å in AB-IM4 vs 1.543 Å in AB-PC), C1-O1 (1.338 Å in AB-IM4 vs 1.231 Å in AB-PC), and C1-C3 (1.553 Å in AB-IM4 vs 1.487 Å in AB-PC), generating benzoin. 1415

dx.doi.org/10.1021/jp110352e |J. Phys. Chem. A 2011, 115, 1408–1417

The Journal of Physical Chemistry A

ARTICLE

Scheme 3. Step Affording t-BuOH for Path B

The free energy of the complex (AB-IM3) verified by IRC is higher than that of AB-IM2 þ R by 11.24 kcal/mol, as shown in Table 1, suggesting AB-IM3 cannot exist stably. So the free energy barrier of the addition step forming a C1-C4 bond with simultaneous transferring of H10 from O1 to O2 equals the free energy difference (26.95 kcal/mol) between AB-TS4 and ABIM2 þ R. This is a relatively low free energy barrier, possibly owing to the baring hydroxyl in HOMO (see Figure S1, Supporting Information) of AB-IM2, the attraction to H10 from carbonyl oxygen of another benzaldehyde molecule and the current of the approaching of the slight positive C1 (0.14e in AB-IM2) to the more positive carbonyl carbon atom (0.40e in R). The free energy barrier of the elimination of pyrido[1,2-a]-2ethyl[1,2,4]triazol-3-ylidene to produce benzoin is just 2.13 kcal/ mol, for the C1-C2 bond is a discomplete covalence bond (its bond length is 1.609 Å) in AB-IM4 and the large interaction of C1(0.24e) and O1 (0.90e). The energy profile of path A beginning from 2IM1 þ R is presented in Figure 5. Considering the overall potential surface, the free energy barriers corresponding to TS1, A-TS2, A-TS3, AB-TS4, and AB-TS5 are 24.06, 30.93, 9.24, 26.95, and 2.13 kcal/ mol, respectively, so the step of the first intermolecular proton transfer between two N-heterocyclic carbene/benzaldehyde coupled intermediates to form A-IM2 (a molecule of tertiary alcohol and a molecule of enolate anion) ready to create enamine is the rate-determining step, while the coupling of carbene and benzaldehyde, the coupling of enamine and another benzaldehyde to form C-C bond accompanied by the hydrogen shift from the hydroxyl of the enamine to the carbonyl oxygen of another benzaldehyde are partially rate-determining for their significant barriers. 3.2. Path B. Experimentally, the reaction was carried out in presence of t-BuOK 2%.1 So we attempted to find the possible t-BuOK/t-BuOH pair catalyzed mechanism to provide the Breslow enamine. As presented in Scheme 3 and Figure 6, H100 of one carbene/benzaldehyde coupling complex molecule transfers from C100 to the carbonyl oxygen (O3) of t-BuO- via a transition state (B-TS2), affording t-BuOH. In B-TS2, the distances of C100 -H100 and O3-H100 are 1.276 and 1.432 Å, respectively, suggesting H100 of IM1 is transferring to O3 of t-BuO-. The displacement vector of the imaginary vibrational frequency of B-TS2 mainly corresponds to the transfer of H100 from C100 to O3. After the barrier of B-TS2 is overcome, a tert-butanol molecule is generated. In B-IM2, with the O3-H100 (0.995 Å) single bond formation, the C5-O3 bond length is elongated from 1.344 Å in t-BuO- to 1.425 Å in B-IM2, suggesting the formation of t-BuOH. The hybridization of C100 converts from sp3 in IM1 to sp2 in B-IM2. O3 with much negative NPA charge (-0.96e) in t-BuO- can capture H100 (its NPA charge is 0.18e in IM1) from C100 (its NPA charge is 0.09e in IM1) easily, which is verified by the relative low free energy barrier (23.38 kcal/mol, as presented in Table 2) of this step generating t-BuOH. While the

direct transfer of H100 atom from t-BuOH to O100 atom in B-IM2 to get enamine (AB-IM2) is verified to be very difficult, which is reflected by that the energy of B-IM2 is lower than the sum of AB-IM2 and t-BuO- by 50.15 kcal/mol in gas phase. As presented in Figure 6, the generated small quantity of tBuOH instead of t-BuO- may catalyze the hydrogen shift from C1 to O1 in IM1 affording enamine intermediate (AB-IM2) (as presented in Scheme 2, thanks to the prominent study of Domingo et al.28). The potential energy surface of this step of t-BuOH assisted proton shift to give enamine in the gas phase reveals one transition state (B-TS3) (see Figure 6). B-TS3 has a five-membered ring related to the hydrogen (H1) transferring from C1 of another carbene/benzaldehyde complex molecule to oxygen (O3) of t-BuOH with simutaneous hydrogen (H100 ) transferring from oxygen (O3) of t-BuOH to O1 of IM1. The distances of C1-H1, H1-O3, O3-H100 , O1-H100 , and C1O1 are 1.256, 1.414, 1.878, 0.993, and 1.427 Å in B-TS3, respectively, suggesting H1 is transferring from C1 to O3, and the O1-H100 single bond is forming with simultaneous O3H100 bond breaking. The displacement vectors of the imaginary vibrational frequency of B-TS3 mainly correspond to the transfer of H1 from C1 of IM1 to O3 of t-BuOH and that of H100 from tBuOH to O1 of IM1. After the barrier of the t-BuOH assisted transition state (B-TS3) is overcome, the Breslow enamine intermediate (AB-IM2) is generated with simultaneous regenerating of t-BuOH. The hybridization of C1 converts from sp3 in IM1 to sp2 in AB-IM2. After the generation of the Breslow enamine intermediate (AB-IM2), the successive condensation step and the elimination of pyrido[1,2-a]-2-ethyl[1,2,4]triazol-3ylidene are same as those of path A. The free energy profile of path B from 2IM1 þ t-BuOH þ R is presented in Figure 5. As illustrated in Table 2, the free energy barriers of TS1 (giving the carbene/benzaldehyde coupling intermediate (IM1)), B-TS3 (related to the t-BuOH assisted Breslow enamine formation), AB-TS4, and AB-TS5 are 24.06, 28.84, 26.95, and 2.13 kcal/mol, respectively, so the step of the t-BuOH assisted Breslow enamine forming is theoretically the rate-determining in path B, while the coupling of carbene and benzaldehyde, and the coupling of enamine and another benzaldehyde to form CC bond are partially rate-determining for their significant barriers. The generation of the catalyst t-BuOH ready to assist the Breslow enamine forming is also significant for path B for its barrier (23.38 kcal/mol).

4. CONCLUSIONS In this paper, the mechanism of the pyrido[1,2-a]-2-ethyl[1,2,4]triazol-3-ylidene catalyzed benzoin condensation was investigated at B3LYP/6-31þG(d) and B3LYP/6-311þG(2d,p) levels. Two pathways (paths A and B) proposed on the basis of the Breslow mechanism were intensively studied. In path A, the Breslow enamine is generated via two sequent intermolecular 1416

dx.doi.org/10.1021/jp110352e |J. Phys. Chem. A 2011, 115, 1408–1417

The Journal of Physical Chemistry A proton transfers between two N-heterocyclic carbene/benzaldehyde coupled intermediates, the first intermolecular proton transfer to form tertiary alcohol and enolate anion is theoretically the rate-determining with corresponding barrier (30.93 kcal/ mol), while the t-BuOH assisted hydrogen transfer generating Breslow enamine is the rate-determining step with corresponding barrier (28.84 kcal/mol) in path B. The coupling of carbene and benzaldehyde and the coupling of enamine and another benzaldehyde to form C-C bond are partially rate-determining for their relatively significant barriers (24.06 kcal/mol and 26.95 kcal/mol, respectively) in both paths A and B, while the free energy barrier (23.38 kcal/mol) of t-BuOH (catalyst for the generation of Breslow enamine) generation is also significant for path B. So paths A and B are competitive for their similar barriers, and the results of these two pathways are in nice agreement with the experimental result in a kinetic investigation of thiazolium ion-catalyzed benzoin condensation performed by White and Leeper24 in 2001.

’ ASSOCIATED CONTENT

bS

Supporting Information. Table of NPA charges. Schemes of reaction pathways. HOMO of AB-IM2. Optimized Cartesian coordinates and geometrical structures of all stationary points along the potential energy profiles. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Tel: þ86 28 85418330.

’ ACKNOWLEDGMENT This project has been supported by the National Natural Science Foundation of China (Grant Nos. 20773089 and 20835003), National Basic Research Program of China (973 Program) (Grant No. 2011CB201202), and the Scientific Research Foundation of the Education Department of Sichuan Province (Grants 09ZC065 and 08ZC020). ’ REFERENCES (1) Ma, Y. J.; Wei, S. P.; Lan, J. B.; Wang, J. Z.; Xie, R. G.; You, J. S. J. Org. Chem. 2008, 73, 8256. (2) W€ohler, F.; Liebig, J. Ann. Pharm. 1832, 3, 249. (3) Ugai, T.; Tanaka, S.; Dokawa, S. J. Pharm. Soc. Jpn. 1943, 63, 296. Chem. Abstr. 1951, 45, 5148). (4) Lapworth, A. J. Chem. Soc. 1903, 83, 995. (5) Lapworth, A. J. Chem. Soc. 1904, 85, 1206. (6) Breslow, R. Chem. Ind. (London) 1957, 893. (7) Breslow, R. J. Am. Chem. Soc. 1958, 80, 3719. (8) Lemal, D. M.; Lovald, R. A.; Kawano, K. I. J. Am. Chem. Soc. 1964, 86, 2518. (9) Castells, J.; Lopez-Calahorra, F.; Domingo, L. J. Org. Chem. 1988, 53, 4433. (10) Castells, J.; Domingo, L.; Lopez-Calahorra, F.; Martí, J. Tetrahedron Lett. 1993, 34, 517. (11) Martí, J.; Castells, J.; Lopez-Calahorra, F. Tetrahedron Lett. 1993, 34, 521. (12) Martí, J.; Lopez-Calahorra, F.; Bofill, J. M. J. Mol. Stru. (Theochem) 1995, 339, 179. (13) Diederich, F.; Lutter, H. D. J. Am. Chem. Soc. 1989, 111, 8439. (14) Lopez-Calahorra, F.; Rubires, R. Tetrahedron. 1995, 51, 9713. (15) Breslow, R.; Kool, E. Tetrahedron, Lett. 1988, 29, 1635.

ARTICLE

(16) Yano, Y.; Tamura, Y.; Tagaki, W. Bull. Chem. Soc. Jpn. 1980, 53, 740. (17) Breslow, R.; Kim, R. Tetrahedron, Lett. 1994, 35, 699. (18) van den Berg, H. J.; Challa, G.; Pandit, U. K. J. Mol. Catal. 1989, 51, 1. (19) Lopez-Calahorra, F.; Castro, E.; Ochoa, A.; Martí, J. Tetrahedron Lett. 1996, 37, 5019. (20) Lopez-Calahorra, F.; Castells, J.; Doming, L.; Martí, J.; Bofill, J. M. Heterocycles. 1994, 37, 1579. (21) Chen, Y.-T.; Barlctta, G. L.; Haghjoo, K.; Cheng, J. T.; Jordan, F. J. Org. Chem. 1994, 59, 7714. (22) Breslow, R.; Schmuck, C. Tetrahedron Lett. 1996, 37, 8241. (23) Linghu, X.; Bausch, C. C.; Jeffrey, S. H. J. Am. Chem. Soc. 2005, 1833. (24) White, M. J.; Leeper, F. J. J. Org. Chem. 2001, 66, 5124. (25) Goldfuss, B.; Schumacher, M. J. Mol. Model. 2006, 12, 591. (26) Gronert, S. Org. Lett. 2007, 9, 3065. (27) Hawkes, K. J.; Yates, B. F. Eur. J. Org. Chem. 2008, 5563. (28) Domingo, L. R.; Aurell, M. J.; Arno, M. Tetrahedron 2009, 3432. (29) He, Y. Q.; Xue, Y. J. Phys. Chem. A 2010, 114, 9222. (30) Beck, A. D. J. Chem. Phys. 1993, 98, 5648. (31) Parr, R. G.; Yang, W. Density Functional Theory of Atoms and Molecules; Oxford University Press: Oxford, 1989. (32) Yamabe, S.; Yamazaki, S. Org. Biomol. Chem. 2009, 7, 951. (33) Dudding, T.; Houk, K. N. Proc. Natl. Acad. Sci. 2004, 101, 5770. (34) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1989, 90, 2154. (35) Gonzalez, C.; Schlegel, H. B. J. Phys. Chem. 1990, 94, 5523. (36) Miertus, S.; Tomasi, J. Chem. Phys. 1982, 65, 239. (37) Cossi, M.; Barone, V.; Cammi, R.; Tomasi, J. Chem. Phys. Lett. 1996, 255, 327. (38) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985, 83, 735. (39) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. Rev. 1988, 88, 899. (40) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniel, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V. ;Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T. ; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head.Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 03, Reversion D.01; Gaussian, Inc.: Pittsburgh, PA, 2005.

1417

dx.doi.org/10.1021/jp110352e |J. Phys. Chem. A 2011, 115, 1408–1417