Theoretical insight into the Au(I)-catalyzed hydration of halo-substituted propargyl acetate: dynamic water-assisted mechanism

Abstract

The hydration mechanism of halo-substituted propargyl acetate, catalyzed by a homogenous Au(I) complex, has been investigated with the aid of the density functional theory (DFT) method. Our results reveal that the hydration is initiated by the favoured 1,5-exo-dig cycloaddition in the anti manner, affording a desired regioselective Markovnikov product. We also verify that neither the pathway towards the anti-Markovnikov product triggered by 1,6-endo-dig cycloaddition, nor direct nucleophilic attack by water, would happen without the help of neighbouring carbonyl groups. The favoured pathway mainly includes three processes: nucleophilic attack after 1,5-exo-dig cycloaddition, protodeauration, and enol–keto tautomerization. It turns out that the third process (enol–keto tautomerization) is the rate-determining step. Additionally, different halo-substituents cannot change the reaction trend, but slightly affect the relative energies. Particularly, cluster-continuum solvent models were established for some proton-transfer steps to rationally simulate reaction processes and evaluate energy barriers. Our study suggests that the presence of an explicit water-bridge is crucial to promote the hydration reaction. Computational results provide theoretical support for experimental observations, and insight into the hydration.

---

1. Introduction

Hydration of alkynes, which provides access to various carbonyl derivatives, has been extensively studied for more than one century.^1–4 Because of the straightforward and atom-economic formation of high-value materials, such as ketones and aldehydes, the hydration of alkynes attracts important industrial interest. Moreover, alkynes can be considered as masked ketones in synthetic chemistry.⁴ Typical catalytic systems for this reaction usually consist of toxic mercury salts and acidic media to enhance the reactivity.^5,6 In order to efficiently promote the hydration in a green mode and extend the reaction scope, a series of organometallic catalysts, such as Pd,^7–9 Ru,¹⁰ Rh,^11,12 Pt,^13,14 and Au complexes^15–24 have been comprehensively utilized in syntheses. Among these catalysts, homogeneous Au(I) catalysts are powerful, since they are neither water nor air sensitive.^25–37 Besides, Au(I) catalysts can offer higher regioselective products than other catalysts. Corma²² and He¹⁸ have shown that a Au(I) complex, bearing the weakly coordinating bis-(trifluoromethanesolfonyl)imidate (NTf₂⁻) ligand, could catalyze the hydration of substituted alkynes at room temperature without acid additives. Nolan et al. demonstrated that a Au(I) nitrogen-heterocyclic carbene (NHC) cationic species such as [Au(IPr)]⁺, derived from [AuCl(IPr)] in the presence of AgSbF₆, is an excellent catalyst for the hydration of alkynes.¹⁷ This reaction could occur at very low catalyst loadings under acid-free conditions.

Additionally, some studies were aimed at controlling the regioselectivity with the assistance of neighboring groups. It was found that the progargyl carboxylates, as neighboring moieties in alkynes, would easily trigger regioselectivity.^19,38–40 Thus, rearrangements of propargylic esters have recently gained much experimental and theoretical attention. Generally, gold-catalyzed reactions of propargylic esters mainly undergo two types of skeleton rearrangement under different reaction conditions, i.e., 3,3-rearrangement and 1,2-acyloxy migration (cases (1) and (2) in Scheme 1). In contrast to the hydration of terminal alkynes, the hydration mechanism of propargylic esters have rarely been studied (case (3) in Scheme 1).

Scheme 1 Rearrangements and hydration of propargyl carboxylates.

Recently, Zhang et al. demonstrated that the presence of an electronegative halogen at the propargyl carboxylate terminus would not only enhance the polarization of alkynes, but also render regioselectivity at the alkyne's distal end⁴¹ (cases (1) and (2) in Scheme 1). Sahoo et al. reported hydration reactions of propargyl carboxylate and halo-substituted propargyl carboxylate (case (3) in Scheme 1) under Ph₃PAuCl/AgSbF₆ at room temperature.^38,39 After screening different reaction conditions, they found that the optimal reaction condition for propargyl carboxylate was 1% Ph₃PAuCl catalyst and 1% AgSbF₆ co-catalyst in 1,4-dioxane solvent and water (3 equiv.) at room temperature, whereas, for halo-substituted propargyl carboxylate, a mixture of solvent (1,4-dioxane/CH₃NO₂, 20 : 1) and water (3 equiv.) was required to improve the yield (Scheme 2).

Scheme 2 Experimentally reported reaction by Sahoo et al.³⁹

The corresponding proposed mechanism is summarized in Scheme 3. The reaction is initiated after the activation of the C≡C triple bond by cationic Au(I) catalyst, rendering a transient five-membered vinyl-Au species (II), via an intramolecular 1,5-exo-dig attack of the acetate carbonyl group at the alkyne. Then, the nucleophilic attack of H₂O at the carbocation leads to ring-opening and offers the species (III). The reaction finally produces the desired ketone after protodemetallation and isomerization. Obviously, in comparison to the direct addition of water at the alkyne, this mechanism is based on the assumption that the participation of the carboxyl moiety would accelerate the hydration.

Scheme 3 Plausible hydration mechanism supposed by Sahoo et al.^38,39

Actually, the above proposed mechanism is somehow rough, for describing the hydration. Some critical problems are hidden and need to be clear. For example, how is the competition between 1,5-exo-dig and 1,6-endo-dig cycloaddition? Does the reaction need only one water molecule or more water molecules to directly take part in the hydration? Which process is the rate-determining step? How is the hydrogen transfer in the reaction? These basic questions, which have caught long-standing attention, usually exist in hydrolysis or solvent-assisted rearrangement reactions.⁴² It is well-known that quantum mechanics (QM) methods can practically find the trajectory of reaction processes and provide electronic structure information. Therefore, these above questions could be answered by the detailed reaction mechanism at the molecular level.

From a theoretical perspective, recent studies have shown that it would be better to describe the Au(I)-catalyzed mechanism for the hydration of alkynes by considering explicit solvent molecules in computations.^21,43–46 Moreover, our previous theoretical studies indicated that explicit solvent molecules could play an active role in reactions.^47,48 The micro-water effect, acting as a proton shuttle, could facilitate the proton-transfer. At this point, based on experimental observations and the proposed reaction mechanism, some explicit water molecules were appropriately taken into account in our computations. From published experimental results, the selected reaction in this work is shown in Scheme 2. Herein, to reduce computational cost and provide a general mechanism, a model system in which the catalysts [AuP(Ph)₃]⁺ and PhCH₂CO₂ were simplified to [AuP(Me)₃]⁺ and MeCO₂, was firstly used for computations. Note that we only considered the chloro-substituted propargyl acetate in this simplified model reaction to determine a favoured pathway. Then, based on the results of the model reaction, the hydration mechanisms of real systems for halo-substituted propargyl acetates were further studied to account for the substituent effect. As shown in Scheme 4, all possible channels for the model reaction are considered and compared in this work (see details in the Results section). Our computations should be sufficient to clarify the favourable hydration pathway catalyzed by the Au(I) complex.

Scheme 4 Schematic showing the possible pathways for the Au(I)-catalyzed hydration of 3-chloroprop-2-yn-1-yl acetate.

2. Computational details

All the relevant geometries were fully optimized by using the B3LYP density functional theory (DFT)^49–52 in conjunction with the self-consistent reaction field (SCRF) method. Numerous theoretical studies of Au-catalyzed reactions have been reported by using the B3LYP method, which confirm that this functional is suitable in this work.^53,54 The standard 6-31+G(d,p) basis set on the non-metal atoms and the Def2-TZVP basis set, including the effective core potential (ECP) on Au and I atoms, were used in the computations.^55–57 After geometry optimization, harmonic vibrational analyses were performed at the same level to confirm that each minimum has no imaginary frequency, or that each transition state (TS) has only one imaginary frequency. The minimum energy path (MEP) was also traced by using an intrinsic reaction coordinate (IRC)⁵⁸ method to ensure that each TS structure correctly links with two minima. Note that although the title reaction is carried out in organic solvent (a mixture of 1,4-dioxane and nitromethane), we assume that the hydration would practically take place at the microscopic interface of the organic solvent and water, or completely in a tiny dot of water. Therefore, the combination of explicit water molecules and the implicit IEF-PCM solvent model^59,60 (cluster-continuum solvent model)^61–73 was adopted to evaluate the solvent effect on the reaction in water (ε = 78). Moreover, since the free energy for each species was computed at 1 atm pressure, the entropy correction at higher pressure could be a simple way to model translational degrees of freedom in the solvent. Free energies have been evaluated at a pressure of 1354 atm to mimic a condensed phase.^44,74 Recently, some benchmark studies revealed that the double-hybrid functional is good for relative energies in gold-catalyzed reactions and that the dispersion correction is important for long-range interaction.^75,76 We therefore performed single-point energy calculations in water to refine relative energies by using the B2PLYP-D3 method.^77–80 In single-point energy calculations, except for Au and I atoms, 6-311++G(d,p) and 6-31+G(d,p) basis sets were used for non-metal atoms in simplified and real models, respectively. Finally, the corresponding total energies of each species were obtained by the sum of B2PLYP-D3 single-point electronic energy and thermal corrections of B3LYP at 25 °C and 1354 atm.

All computations were fulfilled with the Gaussian 09 program.⁸¹ All 3-D structures were generated by the CYL view program.⁸²

3. Results and discussion

3.1 Reaction pathways for the simplified model system

As mentioned above, Scheme 4 shows all the possible processes for the Au(I)-catalyzed nucleophilic addition of water towards 3-chloroprop-2-yn-1-yl acetate. It is worth noting that as shown in Fig. 1, two types (anti and gauche) of reactant complexes could exist in the formation of gold(I)-activated alkyne, due to the close binding energies (21.50 vs. 21.59 kcal mol⁻¹). That is, anti and syn attacking types would have the same probability of occurring (Scheme 4). Additionally, assisted by the neighboring γ-acetate group, 5-exo-dig and 6-endo-dig intramolecular cycloaddition would often competitively occur. After which, 1,2-acyloxy migration and 3,3-rearrangement processes might lead to variable reaction channels and thus be utilized in chemical synthesis. In all, three hydration pathways (paths A, B and C in Scheme 4) were considered and explored in this work. Path A (or path C) is associated with the indirect hydration process after intramolecular 1,5-exo-dig (or 1,6-endo-dig) attack, whereas path B corresponds to the direct addition of water on the alkyne. In order to clearly clarify the whole mechanism, we divide the hydration into three processes. Process 1 is to generate the enol intermediate, process 2 is associated with protodeauration via hydrogen-transfer, and process 3 corresponds to enol–keto tautomerization to furnish the α-acyloxy methyl ketone. Obviously, depending on different attacking positions by water in paths A and B, the oxygen in the formed ketone originates from different groups of substrates. Scheme 4 clearly demonstrates the oxygen (labelled with a red colour) migration in the hydration. One can see that although the product is the same, oxygen from water accumulates in the formed ketone moiety via the direct process (path B), whereas the indirect process (path A) leads to the exchange of oxygen of the acyloxy group with water. From our results (see details of other possible pathways in ESI†), the anti type of path A is found to be more favourable for the title reaction. In this sense, our computational results are consistent with the experimental ¹⁸O-labeled investigation.³⁹ Herein, we only focus on the process of path A in detail.

Fig. 1 Two types (anti and gauche) of Au(I)-activated reactant complexes (bond distance in Å). Binding energies (kcal mol⁻¹) are reported.

3.1.1 Process 1 in path A. Fig. 2 and 3 depict relative energy profiles and some related optimized geometries. Note that the energy of the activated complex RC1 is taken as the reference state of the relative energies. The reaction is triggered by the coordination of the cationic catalyst [Au(PMe₃)]⁺ to the alkynyl of substrate RE. Influenced by the d → π* back donation, the bond length of C1–C2 in the activated complex RC1 is slightly elongated by 0.2 Å to 1.234 Å. The transition state TS1 for the 1,5-exo-dig cycloaddition is successfully optimized. In TS1, bond distances of C2–O2, C1–C2, and C1–Au are 2.124, 1.267 and 2.143 Å, respectively. Vibrational analysis verifies that the imaginary frequency is mainly associated with the C1–C2 bond stretching. Calculated energies indicate that it just requires 11.07 kcal mol⁻¹ to overcome the 1,5-exo-dig cycloaddition and generate a five-membered vinyl-gold intermediate IM1. C1 and C2 atoms change to sp² hybridization in IM1, rendering that the formal positive charge partially distributes on the C4 atom. In this way, it is favourable for water to undergo nucleophilic attack at the C4 atom. Note that the 1,5-exo-dig cycloaddition is a thermodynamically allowed process (ΔG = −1.16 kcal mol⁻¹).

Fig. 2 Relative Gibbs free energy (kcal mol⁻¹) profiles in the first process of path A.

Fig. 3 Some selected geometries (bond distance in Å) in the first process of path A. More optimized geometries are shown in the ESI.†

Next, we considered the nucleophilic addition of one water molecule. A bimolecular complex IM1-1W is located on the potential energy surface. The water molecule, stabilized by two weak hydrogen bonds, is almost right above C4 atom. Due to no external interaction on the water except the implicit solvent, we could not obtain the stepwise mechanism via an ion-pair intermediate; a concerted four-membered ring transition state TS2-1W is involved in this process. For TS2-1W, the bond distances of O(w)1–C4, H(w)1–O1, and H(w)1–O(w)1 are 1.505, 1.236, and 1.257 Å respectively. The bond angle of O2–H(w)–O(w) is 115.83°. Unfortunately, Fig. 2 indicates that the concerted addition/elimination mechanism requires about 6.47 kcal mol⁻¹ to afford a ring-opening intermediate IM3. The high energy barrier mainly results from the large ring-constraint of the proton-transfer. Ideally, it is optimal to shift a proton in a line between the acceptor and the donor.

Bulk water is formed by a hydrogen-bond network. Even several equivalents of water in organic solvent, a very small water dot (could be a water cluster) would dynamically occur. In this sense, more than one water molecule (i.e. water cluster) should be carefully taken into account in this hydration reaction. Generally, water can play three roles in a reaction: nucleophile, proton-shuttle, and micro-solvent. It is very interesting and crucial to understand how many water molecules would directly take part in an elementary step and efficiently help proton-transfer. Encouraged by our previous theoretical studies of hydrolysis mechanisms,^47,48 we tested two, three, four, and five explicit water molecules in computations, respectively.⁸³ Gratifyingly, to facilitate proton-transfer, we found that the backbone of a four-water chain would be necessary in the nucleophilic attack process.⁸⁴ As shown in Fig. 2, nucleophilic attack, assisted by a four-water cluster, significantly reduces the energy barrier. Interestingly, the nucleophilic addition of a water molecule across the C4–O1 bond divides into two steps. The first step is associated with a nucleophilic attack of oxygen of one water molecule on the carbocation (C4 atom). Concurrently, a proton moves from the attacking water molecule into a water-cluster. The second step is the cleavage of the C4–O1 bond with shifting a proton to the O1 atom. It shows that the water cluster can stabilize a proton and act as a shuttle to accept or donate a proton. Actually, the two-step mechanism is generally found and accepted in protic-solvent-catalyzed organic or biochemical reactions. In our computations, the transition state TS2-4W is constructed and optimized in the first step. The distances of C4–O(w)1 and O(w)1–H(w)1 in TS2-4W are 1.775 and 1.050 Å, respectively, indicating that nucleophilic attack precedes proton-transfer. The imaginary vibration is mainly associated with the mode of water attack (O(w)1–C4 bond stretch). Note that ΔG^‡ of the nucleophilic attack step dramatically decreases to 11.44 kcal mol⁻¹. In the following step, we optimized the transition state TS3-4W towards the C4–O1 bond cleavage and proton-transfer. The imaginary frequency vibration is almost related to the C4–O1 bond stretch, indicating that the C4–O1 bond cleavage and proton-transfer are asynchronous. Bond distances of C4–O1 and O1–H(w)4 are 1.907 and 1.064 Å, respectively, revealing that proton-transfer is complete prior to C4–O1 bond cleavage. Note that this elimination step needs 8.80 kcal mol⁻¹ to form an enol intermediate from IM2-4W.

Furthermore, in order to estimate the effect of micro-solvent, we added one more water molecule into the above four-water assisted process. The corresponding optimized structures are also shown in Fig. 3. In this way, the added water molecule might stabilize the charge-separated transition-state in the first step. Compared with the four-water assisted steps as stated above, it turns out that the additional one explicit solvent has little effect on geometries. However, it causes the free energy barriers to go down by about 4.0 kcal mol⁻¹ (ΔG^‡ = 7.23 kcal mol⁻¹ from IM1-5W to TS2-5W). For the overall addition/elimination, the C4–O1 bond cleavage would be slower than the C4–O(w)1 bond formation.

In all, with the help of a four-water chain, the proton-transfer is efficient in this nucleophilic attack process. Importantly, the two-step mechanism (addition and elimination) dramatically reduces the energy barrier.

3.1.2 Process 2: protodeauration. After the formation of intermediate IM3, it is likely to undergo the protodeauration via proton-transfer. The corresponding relative free energy profiles and optimized geometries are shown in Fig. 4 and 5. Herein, we firstly tried one explicit water molecule as a proton-shuttle to facilitate proton-transfer. IM4-1W is generated by the interaction of one explicit water molecule with IM3. The corresponding transition state TS4-1W indicates that the vibration of the imaginary frequency (122.61 cm⁻¹) is unexpectedly associated with the wagging mode of H₃O⁺ in the process of proton-transfer from O(w)1 to O1. We found that the energy barrier of TS4-1W seems high (ΔG^‡ = 12.76 kcal mol⁻¹ from IM4-1W). Similarly, it is also attributed to the ring-constraint of proton-transfer. We then examined the addition of several water molecules, which can build a water-bridge to help to transfer the proton. We found that three or four-water chains would be acceptable for the water-assisted protodeauration. Note that compared with IM3 and IM4-1W, the optimized structures IM4-3W and IM4-4W reveal that H(w)1 has already completely dissociated into a water-cluster without an energy barrier, i.e., the proton attaches at O(w)3 in IM4-3W and IM4-4W, and the cationic structure IM3 should easily liberate an instantaneous proton in water. In transition states TS4-3W and TS4-4W, the imaginary frequency mainly corresponds to the double proton-transfer (i.e., abstracting a proton from the hydronium and simultaneously transferring a proton to the terminal carbon). In this process, the enol structure is produced and the binding mode of the gold(I) catalyst changes from Au–C σ bond to d–π coordination. Free energy barriers of the protodeauration were found to be 5.29 and 7.69 kcal mol⁻¹ for TS4-3W and TS4-4W, respectively. After checking the geometrical parameters, we think that the difference in energy barriers would be explained by the hydrogen-bond interaction of O1 and Hw2. For TS4-3W, the angle of C2–O1–Hw2 is 120.5 and the distance of O1–Hw2 is 1.723 Å. On the other hand, for TS4-4W, the angle of C2–O1–Hw2 is 126.2 and the distance of O1–Hw2 is 1.950 Å. It shows that the orientation of the hydrogen-bond between water and O1 in TS4-4W deviates more from the ideal sp³-type, rendering weaker interaction and higher energy barrier; i.e., the optimal water-bridge structure determines how many water molecules could be accommodated by proton-transfer. We therefore suggest TS4-3W would be the favoured transition state in this water-assisted protodeauration. In addition, the relative energy dramatically decreases from IM3 to IM5 by 25.50 kcal mol⁻¹ and the formed enol intermediate, IM5, is significantly thermodynamically favourable in the process of protodeauration.

Fig. 4 Relative Gibbs free energy (kcal mol⁻¹) profiles in the second process of path A.

Fig. 5 Optimized geometries (bond distance in Å) in the second process of path A.

3.1.3 Process 3: enol–keto tautomerization. As mentioned above, the final process of the gold(I)-catalyzed hydration should be the tautomerization from enol to ketone. The enol–keto tautomerization in water has been extensively investigated theoretically.^85,86 Undoubtedly, the water-assisted mechanism is generally accepted for proton-transfer. To undergo enol–keto tautomerization from IM5, two points should be clear at first: (1) whether the catalyst will displace to the substrate (alkynes) or not; (2) hydrogen should change the orientation, via O–H bond rotation, to facilitate the formation of a water-bridge. The result indicates that the binding energy of IM5 is 22.79 kcal mol⁻¹, which is almost equal to that of RC1. Therefore, the displacement of catalyst to bind with RC1 would be slow and the catalyst would keep on binding with enol in IM5. Next, the calculated ΔG^‡ is 8.72 kcal mol⁻¹ in the rotation from IM5 to IM6 (see Fig. 6 and 7), indicating that the rotation would be feasible, although a strong intramolecular hydrogen-bond exists in IM5. Unexpectedly, unlike IM6, IRC computations of TS6-nW during enol–keto tautomerization confirmed another type of intermediate (IM7), where the catalyst interacts with the chlorine atom rather than the C=C double bond. We think that once the hydrogen atom starts transferring to the terminal carbon atom (enol–keto tautomerization), C1 would gradually change to sp³ hybridization and the Au(I) catalyst would move from the C1=C2 double bond to the chlorine atom. It indicates that as the calculated binding energy of IM7 is reduced to 13.58 kcal mol⁻¹ for the weak Au–chlorine interaction, the catalyst might be liberated from IM7 and go into the next catalytic cycle. We also explored the subsequent enol–keto tautomerization in the presence of the gold catalyst. As shown in Fig. 6 and 8, we tested one, two, three, and four-water-assisted tautomerization. All types of tautomerization were found to be a concerted and asynchronous step, via only one transition state. Due to the ring-constraint, one or two-water-bridges could not efficiently help proton-transfer. The corresponding free energy barriers are 31.66 and 18.91 kcal mol⁻¹ for TS6-1W and TS6-2W, respectively, in the tautomerization. Meanwhile, we found that three or four water molecules, ready for proton-transfer, assembled well in precursors IM7-3W and IM7-4W. Then, transition states TS6-3W and TS6-4W were optimized and the corresponding ΔG^‡ values were 15.97 and 13.49 kcal mol⁻¹. The vibrational analysis indicates that imaginary frequencies (458.87 cm⁻¹ for TS6-3W and 454.61 cm⁻¹ for TS6-4W) are almost related to the double-proton-transfer towards terminal C–H bond formation. H(w)1–O(w)2 becomes a normal bond (0.998 and 0.994 Å in TS6-3W and TS6-4W), revealing that the O1–H(w)1 bond has completely broken and that H(w)1 has already rapidly transferred to the water cluster. Vibrational analysis, IRC results, and geometrical parameters exhibit that proton-transfer is asynchronous and that the liberation of a proton from the O-terminal would be faster than its addition to C-terminal in enol–keto tautomerization. Furthermore, it could also be regarded as the formation of an instantaneous hydroxonium in advance, after the dissociation of the O1–H(w)1 bond. In addition, we also added two more water molecules to evaluate the micro-solvent effect (see IM7-6W and TS6-6W in Fig. 8). Compared with TS6-4W, the corresponding geometrical framework of TS6-6W has a little change, but its ΔG^‡ changes to 15.32 kcal mol⁻¹. That is to say, if using insufficient and inappropriate explicit micro-solvent molecules in computations, the results might be worse than that using an implicit solvent model. This is why we just emphasize the backbone of the water-bridge for proton-transfer under the average solvent effect by using the implicit solvent model in this work. As depicted in Fig. 8, due to a small water-bridge connecting O1 and C1 in TS6-3W, the orientation of Hw1 has a large change from IM7-3W to TS6-3W and deviates extremely from the lone-pair of O1, whereas the position of the Hw1 atom in TS6-4W changes little and still points to the lone-pair of O1. In view of the free energy barrier and geometrical parameters, the four-water assisted process would be likely for this enol–keto tautomerization. Finally, after the displacement of the gold catalyst, the product is furnished during the title reaction.

Fig. 6 Relative Gibbs free energy (kcal mol⁻¹) profiles in the third process of path A.

Fig. 7 Optimized geometries (bond distance in Å) during the O–H rotation in the third process of path A.

Fig. 8 Some selected geometries (bond distance in Å) for enol–keto tautomerization in the third process of path A. More optimized geometries are shown in the ESI.†

For comparison, the enol–keto tautomerization was also investigated without the effect of the gold catalyst. The corresponding results are collected in Fig. 9 and the ESI.† As stated above, we just considered three, four, and six-water assisted cases. We found these geometrical parameters are similar to those in the presence of the catalyst. The corresponding ΔG^‡ are 18.58, 18.75, and 19.06 kcal mol⁻¹ for n-TS6-3W, n-TS6-4W, and n-TS6-6W, respectively, higher than those in the presence of gold catalyst. Therefore, we deem that the Au(I) catalyst would be innocent in the enol–keto tautomerization. Gold-catalyst, binding with a chlorine atom, could slightly weaken the C1-chlorine bond and activate the C1 atom to accept a proton more easily.

Fig. 9 Relative Gibbs free energy (kcal mol⁻¹) profiles in the third process of path A without gold catalyst.

3.1.4 Overall mechanistic scenario. Except for the above-mentioned description of path A, the results of the other possible pathways are shown in the ESI.† According to our results, we can draw a conclusion that the regioselectivity of the title hydration reaction would rely on the first step in each pathway. As shown in Table 1, the lowest free energy barrier (11.07 kcal mol⁻¹) is associated with the anti type of the 1,5-exo-dig cycloaddition in path A; i.e., the intramolecular 1,5-exo-dig cycloaddition, benefiting from the strong back-donation of the gold catalyst, is more favourable for the title hydration reaction. The water-assisted protodeauration and the enol–keto tautomerization occur in sequence. The enol–keto tautomerization in the final process would be the rate-determining step. Therefore, the desired Markovnikov product was obtained in the experiment. Importantly, the origin of the O atom of the formed aldehyde would come from the neighbouring carbonyl group, rather than water, which is in agreement with the experimental ¹⁸O-labeled observation.³⁹ Note that, although the discrete water molecules were built for each process and the backbones of the water-bridge were considered, we deem that the geometrical framework should not be deformed if more water molecules are used for further investigation, such as QM/MM computations.

Table 1 Relative free energy (kcal mol⁻¹) for the first step of each pathway

	ΔG^‡
a Relative to RC1.b Relative to RC2.
1,5-exo-Dig in path A	11.07^a
Direct attack (1W) in path B	14.61^a
Direct attack (2W) in path B	13.65^a
1,6-endo-Dig in path C	13.13^a
1,5-exo-Dig (syn) in path A	19.22^b
Direct attack (1W, syn) in path B	25.79^b

---

3.2 Comparative study on real systems

To match with the experimental observation and account for the substituent effect, we considered the real system (Scheme 2). Based on the above results of the simplified model reaction, herein, we only explored the favoured pathways in real systems. All relevant optimized structures are shown in Fig. S10–S18 in the ESI,† and corresponding relative Gibbs free energies are depicted in Fig. 10–12. Note that the structures of real systems are labelled by a prefix X (Cl, Br, or I) to differ from those of the simplified model. We found that the structural parameters of central backbones in real systems are very close to those in the simplified model. Compared with the chloro-substituted substrates between simplified and real systems, relative Gibbs free energies have a little change, but the trend and general mechanism are the same. Furthermore, some clues to the halo-substituent effect appear in some elementary steps. For instance, ΔG^‡ is 7.97, 9.38, and 11.83 kcal mol⁻¹ for Cl-TS1, Br-TS1, and I-TS1 in the 1,5-exo-dig cycloaddition, respectively. It turns out that the trend of ΔG^‡ should be I > Br > Cl, consistent with the decrease in electronegativity. Stronger electronegativity at the C1 atom induces more electron deficiency at the C2 atom, enhancing the reactivity of intramolecular nucleophilic attack at the C2 atom. Additionally, ΔG^‡ is 7.98, 8.73, and 11.61 kcal mol⁻¹ for Cl-TS5, Br-TS5, and I-TS5, respectively, in the process of O–H bond rotation, whereas, ΔG^‡ is 15.93, 12.66, and 9.39 kcal mol⁻¹ for Cl-TS6-4W, Br-TS6-4W, and I-TS6-4W, respectively, in the enol–keto tautomerization. Note that we also deem that the gold catalyst should be innocent in the enol–keto tautomerization, because the interaction energies (E_in) are 19.04, 23.07, and 27.65 kcal mol⁻¹ for Cl-IM7, Br-IM7, and I-IM7, respectively, by using the Boys counterpoise (CP) correction method. The bond distance of X–Au is 2.502, 2.573, and 2.685 Å in Cl-IM7, Br-IM7, and I-IM7, respectively, revealing that the interaction would not be neglected between Au catalyst and halo-substituted enol intermediates. Unfortunately, the attempt to build a relationship between the computed relative energies and experimental yield of products failed. Furthermore, other possible pathways (paths B and C for direct hydration and 1,6-endo-dig cycloaddition) were also examined and it was found that both of them would be unfavourable in real systems.

Fig. 10 The relative Gibbs free energy (kcal mol⁻¹) profile for the first process of the favoured pathway in halo-substituted real systems (values in parentheses, Cl; values in plain text, Br; underlined values, I).

Fig. 11 Relative Gibbs free energy (kcal mol⁻¹) profile for the second process of the favoured pathway in halo-substituted real systems (values in parentheses, Cl; values in plain text, Br; underlined values, I).

Fig. 12 Relative Gibbs free energy (kcal mol⁻¹) profile for the third process of the favoured pathway in halo-substituted real systems (values in parentheses, Cl; values in plain text, Br; underlined values, I).

4. Conclusions

The whole Au(I)-catalyzed hydration of terminal halo-substituted propargyl acetate was theoretically investigated by the density functional theory (DFT) method in this work. Three main processes, nucleophilic attack after the 1,5-exo-dig cycloaddition, protodeauration, and enol–keto tautomerization, were involved and explored. Our findings are summarized as follows.

(1) Two types (anti and gauche) of Au(I)-activated alkyne could be formed in the initial stage. The anti-type pathway is more favourable for the reaction.

(2) In comparing the 1,6-endo-dig cycloaddition (path C) and direct nucleophilic attack by water (path B), the 1,5-exo-dig cycloaddition (path A) is more favourable to start the reaction.

(3) After 1,5-exo-dig cycloaddition, the reaction requires a four-water-bridge to efficiently help proton transfer, reduce the energy barrier, and accelerate the reaction rate in nucleophilic attack. In the presence of explicit water molecules, the first process of path A divides into two steps (addition and elimination).

(4) In the process of protodeauration, the three-water assisted mechanism would be more efficient.

(5) In the enol–keto tautomerization, the O–H bond rotation would occur first. Importantly, the gold-catalyst would be innocent in accelerating the reaction rate. Benefiting from a four-water bridge, the enol–keto tautomerization just needs 13.49 kcal mol⁻¹ via the transition state TS6-4W, in the simplified chloro-substituted reaction.

(6) It is found that the third process (enol–keto tautomerization) is the rate-determining step in the overall hydration.

(7) A comparative study on real systems indicates that the halo-substituent has little effect on structures and a little effect on relative energies. However, the general reaction trend of real systems is the same as the simplified reaction model.

To sum up, a water cluster plays an important role in the hydration and it is very crucial to consider the appropriate water molecules for a water-bridge in computations. Our computational results are consistent with experimental observations and provide a deep insight into the hydration process. Moreover, our results could afford a general framework for the nucleophilic addition, protodeauration, and enol–keto tautomerization with the assistance of water.

Acknowledgements

This work is financially supported by the National Natural Science Foundation of China (No. 21573172) and the National Key Basic Research Program of China (No. 2012CB720904). L. Jin thanks the Postdoctoral Science Foundation of China (No. 2015M582635). Y. Wu thanks the Fundamental Research Funds for the Central University (No. xjj2015064).

References

1. L. Hintermann and A. Labonne, Synthesis, 2007, 1121–1150.
2. F. Alonso, I. P. Beletskaya and M. Yus, Chem. Rev., 2004, 104, 3079–3160.
3. C. Bruneau and P. H. Dixneuf, Chem. Commun., 1997, 507–512.
4. M. B. Smith and J. March, March's Advanced Organic Chemistry: Reaction, Mechanism and Structures, Wiley, Hoboken, NJ, 6th edn, 2007.
5. R. J. Thomas, K. N. Campbell and G. F. Hennion, J. Am. Chem. Soc., 1938, 60, 718–720.
6. S. Uemura, H. Miyoshi and M. Okano, J. Chem. Soc., Perkin Trans. 1, 1980, 1098–1103.
7. M. Tokunaga and Y. Wakatsuki, Angew. Chem., Int. Ed., 1998, 37, 2867–2869.
8. T. Suzuki, M. Tokunaga and Y. Wakatsuki, Org. Lett., 2001, 3, 735–737.
9. M. Tokunaga, T. Suzuki, N. Koga and Y. Wakatsuki, J. Am. Chem. Soc., 2001, 123, 11917–11924.
10. T. Khan, S. B. Halligudi and S. Shukla, J. Mol. Catal., 1990, 58, 299–305.
11. B. R. James and G. L. Rempel, J. Am. Chem. Soc., 1969, 91, 863–865.
12. J. Blum, H. Huminer and H. Alper, J. Mol. Catal., 1992, 75, 53–160.
13. W. Hiscox and P. W. Jennings, Organometallics, 1990, 9, 1997–1999.
14. L. W. Francisco, D. A. Moreno and J. D. Atwood, Organometallics, 2001, 20, 4237–4245.
15. Y. Fukuda and K. Utimoto, J. Org. Chem., 1991, 56, 3729–3731.
16. E. Mizushima, K. Sato, T. Hayashi and M. Tanaka, Angew. Chem., Int. Ed., 2002, 41, 4563–4565.
17. N. Marion, R. S. Ramón and S. P. Nolan, J. Am. Chem. Soc., 2008, 131, 448–449.
18. L. Xie, Y. D. Wu, W. Yi, L. Zhu, J. Xiang and W. He, J. Org. Chem., 2013, 78, 9190–9195.
19. W. Wang, B. Xu and G. B. Hammond, J. Org. Chem., 2009, 74, 1640–1643.
20. S. S. Lathe, L. A. Jones, F. Mohr and M. Laguna, Organometallics, 2007, 26, 952–957.
21. J. Cordón, G. Jiménez-Osés, J. M. López-de-Luzuriaga, M. Monge, M. E. Olmos and D. Pascual, Organometallics, 2014, 33, 3823–3830.
22. A. Leyra and A. Corma, J. Org. Chem., 2009, 74, 2067–2074.
23. N. Mézailles, L. Ricard and F. Gagosz, Org. Lett., 2005, 7, 4133–4136.
24. R. Casado, M. Contel, M. Laguna, P. Romero and S. Sanz, J. Am. Chem. Soc., 2003, 125, 11925–11935.
25. R. Dorel and A. M. Echavarren, Chem. Rev., 2015, 115, 9028–9072.
26. B. Ranieri, I. Escofet and A. M. Echavarren, Org. Biomol. Chem., 2015, 13, 7103–7118.
27. D. Qian and J. Zhang, Chem. Soc. Rev., 2015, 44, 677–698.
28. J. Xie, C. Pan, A. Abdukader and C. Zhu, Chem. Soc. Rev., 2014, 43, 5245–5256.
29. A. Hashmi and K. Stephen, Acc. Chem. Res., 2014, 47, 864–876.
30. L. Zhang, Acc. Chem. Res., 2014, 47, 877–888.
31. Y. M. Wang, A. D. Lachner and F. D. Toste, Acc. Chem. Res., 2014, 47, 889–901.
32. C. Obradors and A. M. Echaavarren, Acc. Chem. Res., 2014, 47, 902–912.
33. D. H. Zhang, X. Y. Tang and M. Shi, Acc. Chem. Res., 2014, 47, 913–924.
34. A. Fuerstner, Acc. Chem. Res., 2014, 47, 925–938.
35. B. Alcaide and P. Almendros, Acc. Chem. Res., 2014, 47, 939–952.
36. L. Fensterbank and M. Malacria, Acc. Chem. Res., 2014, 47, 953–965.
37. H. K. Yeom and S. Shin, Acc. Chem. Res., 2014, 47, 966–977.
38. N. Ghosh, S. Nayak and A. K. Sahoo, J. Org. Chem., 2011, 76, 500–511.
39. N. Ghosh, S. Nayak, B. Prabagar and A. K. Sahoo, J. Org. Chem., 2014, 79, 2453–2462.
40. T. R. Pradhan, K. L. Mendhekar and D. K. Mohapatra, J. Org. Chem., 2015, 80, 5517–5531.
41. Y. Wang, B. Lu and L. Zhang, Chem. Commun., 2010, 46, 9179–9181.
42. R. B. Sunoj and M. Anand, Phys. Chem. Chem. Phys., 2012, 14, 12715–12736.
43. C. M. Krauter, A. S. K. Hashmi and M. Pernpointner, ChemCatChem, 2010, 2, 1226–1230.
44. G. Mazzone, N. Russo and E. Sicilia, Organometallics, 2012, 31, 3074–3080.
45. M. Lein, M. Rudolph, S. K. Hashmi and P. Schwerdtfeger, Organometallics, 2010, 29, 2206–2210.
46. G. Mazzone, N. Russo and E. Sicilia, J. Chem. Theory Comput., 2010, 6, 2782–2789.
47. Y. Wu, L. Jin, Y. Xue, D. Q. Xie, C. K. Kim, Y. Guo and G. S. Yan, J. Comput. Chem., 2008, 29, 1222–1232.
48. Y. Wu, Y. Xue, D. Q. Xie, C. K. Kim and G. S. Yan, J. Phys. Chem. B, 2007, 111, 2357–2364.
49. A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
50. A. D. Becke, J. Chem. Phys., 1996, 104, 1040–1046.
51. C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785–789.
52. R. G. Parr and W. Yang, Density-Functional Theory of Atoms and Molecules, Oxford University, Oxford, 1989.
53. F. Q. Shi, X. Li, Y. Xia, L. Zhang and Z. X. Yu, J. Am. Chem. Soc., 2007, 129, 15503–15512.
54. G. Kovács, G. Ujaque and A. Lledós, J. Am. Chem. Soc., 2008, 130, 853–864.
55. Def2-TZVP basis set is taken from EMSL (http://bse.pnl.gov/bse/portal) and shown in ESI.†.
56. F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005, 7, 3297–3305.
57. D. Andrae, U. Haeussermann, M. Dolg, H. Stoll and H. Preuss, Theor. Chim. Acta, 1990, 77, 123–141.
58. K. Fukui, Acc. Chem. Res., 1981, 14, 363–368.
59. V. Barone and M. Cossi, J. Phys. Chem. A, 1998, 102, 1995–2001.
60. M. Cossi, N. Rega, G. Scalmani and V. Barone, J. Comput. Chem., 2003, 24, 669–681.
61. R. BabaAhmadi, P. Ghanbari, N. A. Rajabi, A. S. K. Hashmi, B. F. Yates and A. Ariafard, Organometallics, 2015, 34, 3186–3195.
62. P. L. Silva, C. M. Silva, L. Guimarães and J. R. Pliego Jr, Theor. Chem. Acc., 2015, 134, 1591.
63. D. R. Ramos, R. Castillo, M. L. Canle, M. V. García, J. Andrés and J. A. Santaballa, Org. Biomol. Chem., 2009, 7, 1807–1814.
64. T. T. Mihaylov, T. N. Parac-Vogt and K. Pierloot, Org. Biomol. Chem., 2014, 1395–1404.
65. L. L. Han, S. J. Li and D. C. Fang, Phys. Chem. Chem. Phys., 2016, 18, 6182–6190.
66. C. H. Lim, A. M. Holder and C. B. Musgrave, J. Am. Chem. Soc., 2013, 135, 142–154.
67. J. M. Fox, O. Dmitrenko, L. Liao and R. D. Bach, J. Org. Chem., 2004, 69, 7317–7328.
68. R. Gómez-Bombarlli, E. Calle and J. Casado, J. Org. Chem., 2013, 78, 6868–6879.
69. R. Gómez-Bombarlli, E. Calle and J. Casado, J. Org. Chem., 2013, 78, 6880–6889.
70. V. S. Nguyen, T. M. Orlando, J. Leszczynski and M. T. Nguyen, J. Phys. Chem. A, 2013, 117, 2543–2555.
71. X. M. Sun, X. G. Wei, X. P. Wu, N. B. Wong and W. K. Li, J. Phys. Chem. A, 2010, 114, 595–602.
72. M. T. Nguyen, G. Rasoet and L. G. Vanquickenborne, J. Phys. Chem. A, 1997, 101, 7379–7388.
73. B. J. Wang and Z. X. Cao, J. Phys. Chem. A, 2010, 114, 12918–12927.
74. R. L. Martin, P. J. Hay and L. R. Pratt, J. Phys. Chem. A, 1998, 102, 3565–3573.
75. R. Kang, H. Chen, S. Shaik and J. Yao, J. Chem. Theory Comput., 2011, 7, 4002–4011.
76. G. Ciancaleoni, S. Rampino, D. Zuccaccia, F. Tarantelli, P. Belanzoni and L. Belpassi, J. Chem. Theory Comput., 2014, 10, 1021–1034.
77. S. Grimme, J. Chem. Phys., 2006, 124, 34108.
78. G. Ciancaleoni, S. Rampino, D. Zuccaccia, F. Tarantelli, P. Belanzoni and L. Belpassi, J. Chem. Theory Comput., 2014, 10, 1021–1034.
79. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104.
80. S. Grimme, S. Ehrlich and L. Goerigk, J. Comput. Chem., 2011, 32, 1456–1465.
81. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09, Revision C.02, Gaussian, Inc., Wallingford, CT, 2009.
82. C. Y. Legault, CYLview, 1.0b, Université de Sherbrooke, 2009, http://www.cylview.org.
83. D. J. Wales, J. P. K. Doye, A. Dullweber, M. P. Hodges, F. Y. Naumkin, F. Calvo, J. Hernández-Rojas and T. F. Middleton, Initial guess structures of reference states (H₂O)_n, n = 2–6, are taken from the Cambridge Cluster Database, http://www.wales.ch.cam.ac.uk/CCD.html.
84. For the nulceophilic attack and elimination processes assisted by five-water cluster backbone, it is found that the water-bridge has a little deformation due to the ring strain and flexibility. Moreover, the corresponding ΔG^‡ in each elementary step is higher than that of four-water assisted mechanism, i.e. TS2-4W and TS3-4W. More details are shown in Fig. S7 and 8 in ESI.†.
85. T. Kaweetirawatt, T. Yamaguchi, T. Higashiyama, M. Sumimoto and K. Hori, J. Phys. Org. Chem., 2012, 25, 1097–1104 , and cited therein.
86. N. Capron, B. Casir, N. Sisourat, M. N. Piancastelli, M. Simon and S. Carniato, Phys. Chem. Chem. Phys., 2015, 17, 19991–19996.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra13897g

---