1. Introduction
Propargylamines (PPAs) are nitrogenous compounds with unique chemical structures and extensive applications. They can be used as important intermediates to create a variety of bioactive molecules and are widely used in organic synthesis, pharmaceutical chemistry, and other fields [
1,
2,
3].
Multicomponent reactions are economical, selective, and energy-efficient [
4]. Transition metal catalysts exhibit strong reaction activity, high selectivity, and chemical stability. The synthesis of PPAs through a three-component coupling reaction of aldehyde, acetylene, and amines catalyzed by transition metals is an efficient and atomic-economic method [
5]. This type of reaction is also known as the A3 (alkyne, aldehyde, and amine) coupling reaction and has been extensively researched [
6].
In 1998, McNally [
7] and Dyatkin [
8] reported the preparation of PPAs via the reactions of aldehydes, acetylene, and secondary amines in the presence of a catalytic amount (10 mol %) of CuCl as the catalyst; the yield of the product was between 70% and 90% with different substituents. In 2002, Li et al. reported the reactions of phenylacetylene (PAE), aldehydes, and aromatic primary amines cocatalyzed by RuCl
3 and CuBr [
9]. When only CuBr was used as the catalyst, the reaction yield was extremely low; however, when RuCl
3 (3 mol %) and CuBr (30 mol %) were used as cocatalysts, the reaction yield was significantly improved from 30% to 90%.
In 2004, Shi et al. reported a rapid and microwave-assisted CuI (15 mol %) -catalyzed A3 coupling method, which involves microwave heating using water as a solvent; the yield of the product ranged from 41% to 93%, and this method has wide substrate applicability [
10]. Yadav et al. used an ionic liquid as a solvent and CuBr as a catalyst for the A3 coupling reaction. This reaction can be performed using aromatic primary amines as well as fatty primary amines [
11]. Subsequently, Eycken et al. used CuBr (20 mol %) as a catalyst to realize the A3 coupling reaction between an aliphatic primary amine substrate and terminal alkynes and aldehydes under microwave irradiation conditions, resulting in a significant shortening of the reaction time; the yield of the product ranged from 41% to 94% [
12].
In 2014, Khanna et al. found that when the pyridine ring in the primary amine substrate was replaced with a benzimidazole ring during the cocatalysis of Cu(I) and Ag(I), the intramolecular cyclization/oxidation of PPA intermediates occurred to generate pyrimidine benzimidazole compounds [
13]. Singh et al. used ortho-ester-substituted benzaldehyde substrates to react with terminal alkynes and primary amines via A3 coupling reactions using Cu(I) and chiral ligands as catalysts to generate chiral propargyl amines, followed by intramolecular internal amidation to generate optically active isoindoline ketones [
14].
Transition metal compounds are widely used as catalysts to promote organic reactions because of their rich variety and good catalytic activity. The three-component A3 coupling reaction employing Cu salt as the catalyst was first used to promote the synthesis of PPA and other products. Subsequently, many reports have been published on the use of various transition metal catalysts, such as Ag [
15,
16,
17,
18,
19], Au [
20,
21,
22], Cu [
23,
24,
25], In [
26,
27], Fe [
28,
29,
30], Ni [
31,
32], and Pd [
33], to promote this three-component reaction.
In 2014, Trose et al. synthesized pyridine-containing ligand silver(I) catalysts (3 mol %) and found that they could catalyze A3 coupling reactions under dielectric heating conditions, with the product yields ranging between 53% and 98% [
15]. In 2018, Cao et al. found that polyacrylonitrile fiber-supported Au catalysts (1 mol %) could be used to synthesize secondary PPAs with short retention times and efficient productivity, and they achieved a product yield up to 86% [
20]. In 2020, Nouruzi et al. prepared a covalent organic polymer Au catalyst (0.8 mol %) with a high activity for A3 coupling reactions and attained a product yield up to 90% [
34]. Li et al. synthesized 2,4-disubstituted cyclopentenones using the A3 coupling method, where the addition of acidic 2,2,2-trifluoroethanol (TFE) as the reaction solvent was important for ensuring successful reactions [
35]. In 2021, Xu et al. applied a microwave-assisted A3 coupling reaction to catalyze the reaction between two amines (formaldehyde and propionic acid) using Cu(I) (30 mol %) and attained a maximum product yield of 88%. Asymmetric 1, 4-diamino-2-butylene can be efficiently synthesized through a domino process [
36].
In 2022, Cao synthesized a heterogeneous catalyst CuO–CeO
2 for the activation of terminal alkynes to promote cyclization/A3 coupling reactions. The results showed that the CuO–CeO
2 catalyst exhibited a good cycle performance and group compatibility in the A3 coupling reaction [
37]. Kumar et al. synthesized three-dimensional architectures of silver(I)-based coordination polymers that could be used as catalysts for A3 coupling reactions [
38]. Costabile et al. synthesized and tested A3-coupling reactions using silver and gold as catalysts (3 mol %). The maximum yields of the products were 88% for the silver catalyst and 90% for the gold catalyst. In addition, the catalytic behaviors of the two gold complexes with different substituents were compared using density functional theory (DFT) calculations. The results showed that both gold complexes had low transition state barriers for alkyne deprotonation [
39].
In 2017, Cao et al. prepared recoverable N-heterocyclic carbene silver (
CatAg) catalysts (0.5 mol %) to synthesize PPA;
CatAg has high activity and can efficiently perform A3 coupling reactions. Additionally, the triple bond and −N
3 group in PPA underwent a cycloaddition reaction to generate the final product (PR) with a maximum product yield of 92% [
40]. However, the specific reaction mechanism remains unclear, and to the best of our knowledge, no theoretical research has been conducted on the A3 coupling reactions catalyzed by
CatAg. Therefore, this study employed DFT calculations to investigate the catalytic mechanism of the A3 coupling reaction catalyzed by
CatAg. Detailed research on the experimental mechanism will help us to understand the catalytic mechanism and provide new strategies for the synthesis of propargylamines.
Scheme 1 shows a flowchart of
CatAg catalyzing these reactions [
40].
2. Results and Discussion
2.1. Addition Reaction between CatAg and PAE
Initially, we focused on the addition reaction between
CatAg and
PAE.
Figure 1 illustrates the path of the addition reaction, and
Figure 2 illustrates the molecular structures.
In the addition reaction, the C
2 atom of PAE bonds to the center metal, Ag, of the catalyst CatAg, forming the product Ag_PAE. Because of coordination in the structure of the product Ag_PAE, C
2–H
1 (1.068 Å) and C
2–C
3 (1.225 Å) bonds are longer than those (1.062 and 1.204 Å) in PAE; the Ag–C
1 (2.119 Å) bond is shorter than that (2.129 Å) in CatAg. The frontier molecular orbitals of product Ag_PAE also illustrate the apparent interaction between PAE and CatAg.
Figure 2B shows the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of Ag_PAE. The Ag
dx2−y2 orbital interacts with the
π orbitals of the triple bond in the HOMO. Meanwhile, the Ag
s orbitals appropriately combine with
π anti-bonding orbitals of the C
2–C
3 triple bond in the LUMO.
Table 1 lists the Wiberg bond indices (WBIs) and natural bond orbital (NBO) charges (natural charges) for the addition reaction between CatAg and PAE. WBIs of the C
2–H
1 (from 0.934 to 0.899) and C
2–C
3 (from 2.825 to 2.623) bonds in the product Ag_PAE gradually decreased compared with those in the reactant (CatAg + PAE), indicating the weakening of the C
2–H
1 and C
2–C
3 bonds. Furthermore, Q
NBO value of the C
2 atom in the product (−0.187 e) was higher than that in the reactant PAE (−0.201 e). Meanwhile, the WBIs of the Ag–C
1 (from 0.482 to 0.491) and Ag–C
2 (from 0.000 to 0.199) bonds in the product Ag_PAE increased compared with those in the reactant (CatAg + PAE), indicating the strengthening of the Ag–C
1 and Ag–C
2 bonds. The energy of the product Ag_PAE is slightly higher than that of the reactant (CatAg + PAE) by a value of 2.1 kcal/mol, indicating that the experiment is easy to perform.
2.2. Hydrogen Migration Reaction Involving Amine
After PAE bonds to CatAg, a hydrogen migration reaction involving the amine occurs through the reaction path presented in
Figure 3. The corresponding molecular geometries are shown in
Figure 4. The hydrogen migration reaction involving amines starts with TS1 and proceeds to the product (Ag_PAI + AmineH).
In the reaction pathway, the H1 atom of Ag_PAE first coordinated to the nitrogen atom (N1) of the Amine, forming TS1. The bond length of C2–H1 (1.380 Å) and C2–C3 (1.234 Å) bonds in TS1 are longer than those in Ag_PAE (1.068 and 1.225 Å, respectively). TS1 has a unique imaginary frequency of –994.9 cm−1, which shows the vibration characteristics of hydrogen migration between C2 and N1. The frontier orbitals of TS1 also show their respective hydrogen migration reaction characteristics.
After the formation of TS1, the C2–H1 bond breaks to form the product (Ag_PAI + AmineH). The hydrogen atom H1 migrates from Ag_PAE to AmineH. Compared with those of the TS1 configuration, the C2–C3 (1.222 Å) and N1–H1 (1.032 Å) bonds of the product (Ag_PAI + AmineH) are shortened (1.234 and 1.277 Å) by 0.012 and 0.245 Å, respectively; the N1–H2 bond (1.021 Å) of AmineH is longer than that (1.015 Å) in the TS1 configuration by 0.006 Å.
Table 2 lists WBIs and NBO charges for the hydrogen migration reaction involving the amine. WBIs of the C
2–H
1 bond (from 0.899 to 0.000) in the product (Ag_PAI + AmineH) gradually decrease compared with those in the reactant (Ag_PAE + Amine), indicating the breakage of the C
2–H
1 bond. In contrast, WBIs of the N
1–H
1 bond (from 0.000 to 0.739) in the product (Ag_PAI + AmineH) increase compared with those in the reactant (Ag_PAE + Amine), indicating the N
1–H
1 bond is formed. Additionally, Q
NBO values of H
1 atom increase from 0.174 e in (Ag_PAE + Amine) to 0.458 e in (Ag_PAI + AmineH). Relative to the reactant (Ag_PAE + Amine), the energy barrier is 13.0 kcal/mol, and the reaction is exergonic, releasing 3.7 kcal/mol, revealing that it is feasible. The results are similar to those of a previous study in which alkyne deprotonation had a low transition state barrier when using gold catalysts [
39].
2.3. Amine–Aldehyde Condensation Reaction
The amine–aldehyde condensation reaction generates a water molecule and an imine ion [
41,
42]. The amine–aldehyde condensation reaction can procced via two pathways: with (Path one) or without (Path two) another amine catalyst.
Figure 5 and
Figure 6 illustrate the reaction path of the amine–aldehyde condensation reaction and the molecular structures, respectively.
In the amine–aldehyde condensation reaction, one hydrogen atom (H1) of AmineH first coordinates to the oxygen (O1) atom of the aldehyde (Ald) to form the intermediate IM1. Compared with the reactant (AmineH + Ald), because of coordination, C4–O1 (1.222 Å) and N1–H1 (1.044 Å) bonds in IM1 are lengthened by 0.018 and 0.012 Å, respectively. Subsequently, TS2 is formed, and the H1 atom migrates from N1 to O1. TS2 has a unique imaginary frequency of –138.3 cm−1, which shows the vibration characteristics of the H1 atom migration between N1 and O1. The N1–H1 (2.546 Å) and C4–O1 (1.272 Å) bonds are longer than those of IM1 (1.044 and 1.222 Å, respectively), whereas the O1–H1 bond (0.978 Å) is shorter than that (1.723 Å) in IM1. Relative to that of the reactant (AmineH + Ald), the energy potential barrier of TS2 is 27.9 kcal/mol. Compared with those in TS2, the C4–N1 (1.524 Å) and O1–H1 (0.963 Å) bonds in IM2 are shorter; however, the C4–O1 (1.403 Å) bond is longer.
After the formation of the intermediate IM2, the reaction can proceed via two pathways: with (Path one) or without (Path two) another amine catalyst. In Path one, with another amine catalyst, the H2 hydrogen atom of IM2 coordinates to the N2 nitrogen atom of another amine, forming TS3. TS3 has a unique imaginary frequency of –865.2 cm−1, which shows the vibration characteristics of the H2 atom migration between N1 and N2. The N1–H2 bond (1.300 Å) in TS3 is longer than that of IM2 (1.023 Å), whereas the C4–N1 bond length decreases from 1.524 Å in IM2 to 1.490 Å in TS3. Thereafter, the formation of IM3 occurs. The N1–H2 bond (1.740 Å) in the IM3 configuration is longer than that (1.300 Å) in the TS3, whereas the N2–H2 bond (1.075 Å) is shorter. The H2 atom then migrates from N1 to N2. After the formation of IM3, TS4 is formed. TS4 has a unique imaginary frequency of –610.8 cm−1, which shows the vibration characteristics of the H2 atom migration between N2 and O1. The N2–H2 (1.202 Å) and C4–O1 (1.954 Å) bonds increase in length. Thereafter, the N2–H2 bond is further lengthened to 1.805 Å, and another intermediate IM4 is formed. The H2 atom migrates from N2 to O1. Afterward, the N2–H2 and C4–O1 bonds break, leading to the formation of the product (Imine + H2O).
Table 3 lists WBI and Q
NBO values for the amine–aldehyde condensation reaction. As the reaction proceeds from the reactant (AmineH + Ald) to form the product (Imine + H
2O), N
1–H
1 and C
4–O
1 WBIs decrease from 0.739 to 0.009 and from 1.886 to 0.002, respectively, implying the breakage of these bonds. However, WBIs of the O
1–H
1 (from 0.071 to 0.813) and O
1–H
2 (from 0.004 to 0.813) bonds gradually increase, indicating the formation of free H
2O. Relative to the reactant (AmineH + Ald), the energy barrier of Path one is 30.7 kcal/mol and the energy of the product (Imine + H
2O) is similar.
Path two for the amine–aldehyde condensation reaction starts from IM2. In contrast to Path one, which requires another amine catalyst, Path two directly yields the same product through a four-membered cyclic hydrogen migration transition state (TS3A). TS3A has a unique imaginary frequency of –1431.2 cm−1, which shows the vibration characteristics of the H2 atom migration between N1 and O1. Relative to the reactant (AmineH + Ald), the energy barrier of Path two was 49.9 kcal/mol. This value is very high compared to that of Path one (30.7 kcal/mol), and, therefore, Path two is not the dominant pathway.
2.4. Imine Reaction with Ag_PAI
Imine reacted with Ag_PAI to generate the A3 coupling reaction product PPA.
Figure 7 and
Figure 8 illustrate the reaction path of Imine with Ag_PAI and the molecular structures, respectively.
The reaction first forms a carbon–carbon (C
2–C
4) bonding transition state (TS5). The Ag–C
2 (2.074 Å), C
2–C
3 (1.234 Å), and C
4–N
1 (1.334 Å) bonds in the TS5 configuration are longer than those (2.028, 1.222, and 1.283 Å, respectively) in the reactant (Ag_PAI + Imine). TS5 has a unique imaginary frequency of –134.8 cm
−1, which shows the vibration characteristics of carbon(C
2)–carbon(C
4) bonding. The frontier molecular orbitals of TS5 showed their respective carbon–carbon bonding reaction characteristics (
Figure 8B). The C
2 px orbital appropriately combined with the C
4 py orbital of the LUMO. The formation of IM5 subsequently occurred. The C
2–C
4 bond length decreased from 2.151 Å in TS5 to 1.501 Å in IM5. Thereafter, the Ag–C
2 and Ag–C
3 bonds break, releasing the catalyst components CatAg and PPA. The C
2–C
4 (1.470 Å) and C
2–C
3 (1.207 Å) bonds in the PPA configuration are shorter than those in IM5, whereas the C
4–N
1 (1.478 Å) bond is longer.
Table 4 lists WBI and Q
NBO values for the imine ion and the silver phenylacetylide (Ag_PAI) reaction. As the reaction proceeds, from the reactant (Ag_PAI + Imine) to form the product (CatAg + PPA)
, the Ag–C
2 bond gradually elongates until its breakage (WBI decreases from 0.627 to 0.000). The C
4–N
1 bond gradually elongates until it changes from a double bond to a single bond (WBI decreases from 1.669 to 0.946), and the C
2–C
4 bond shortens until a stable bond is formed (WBI increases from 0.000 to 1.034). Relative to the reactant (Ag_PAI + Imine), the energy barrier is 25.7 kcal/mol and the energy of the product (CatAg + PPA) is slightly higher (3.8 kcal/mol).
2.5. Cycloaddition Reaction
The triple bond and N
3 group in PPA can undergo cycloaddition reactions [
43] to generate the PR.
Figure 9 and
Figure 10 illustrate the path of the cycloaddition reaction and the molecular structures, respectively.
The CatAg first re-coordinates with the C
2–C
3 bond of PPA and forms the intermediate IM6. Because of the coordination, the C
2–C
3 bond (1.227 Å) in IM6 is longer than that (1.207 Å) in PPA by 0.020 Å. Immediately after the formation of IM6, the five-member cycloaddition occurs and forms TS6. TS6 has a unique imaginary frequency of –423.7 cm
−1, which shows the vibration characteristics of a cycloaddition reaction. The frontier orbitals of TS6 exhibit their respective cycloaddition reaction characteristics (
Figure 10B).
Figure 10B shows the HOMO and LUMO orbitals of TS6. The C
2–C
3 bond
π orbital interacts with the
π orbital of the −N
3 group in HOMO. Meanwhile, the C
2–C
3 bond
π orbital combines with the
π anti-bonding orbital of the −N
3 group in LUMO. In the TS6 configuration, the C
2–N
3 (2.091 Å) and C
3–N
5 (2.243 Å) bonds are significantly shorter than those in IM6 (3.107 and 3.675 Å). Next, TS6 releases CatAg, thereby forming PR. Significantly, the C
2–N
3 (1.359 Å) and C
3–N
5 (1.370 Å) bonds in the PR are further shortened.
Table 5 lists WBI and Q
NBO values for the cycloaddition reaction. In the pathway of cycloaddition reaction, when the reaction proceeds from the reactant (PPA) to the PR
, the C
2–C
3 bond gradually elongates until it changes from a triple bond to a double bond (WBI decreases from 2.716 to 1.420), and the C
2–N
3 (WBI increases from 0.002 to 1.206) and C
3–N
5 (WBI increases from 0.001 to 1.310) bonds shorten until stable bonds are formed. Compared with the reactant (PPA), the energy barrier is 25.4 kcal/mol, and the reaction is exergonic, releasing 64.7 kcal/mol. Therefore, the reaction is easy to conduct, which is similar to the experimental results [
40].
To further study these reactions, conceptual density functional theory (CDFT) calculations were performed (
Table 6). The global reactivity index (GRI) values reveal the nucleophilicity and electrophilicity of the molecules. The global nucleophilicity (
NNu) values of PAE, PPA, and Ag_PAI are 2.825, 3.216, and 4.215, respectively, which indicate the nucleophilicities of these molecules. Meanwhile, the global electrophilicity (
ω) values of CatAg and Imine are 3.795 and 4.172, respectively, which indicate that they are electrophilic.
Table 6 and
Figure 11 present the Fukui function data for certain molecules. The data illustrate that the Fukui function values (
f+) of Ag (0.693) and C
4 (0.212) atoms are the most significant parameters for CatAg and Imine; therefore, Ag (CatAg) and C
4 (Imine) atoms have higher reactivity. At the same time, the Fukui function values (
f−) of C
2 in PAE (0.197), PPA (0.064), and Ag_PAI (0.172) are the highest. This indicates that the C
2 atoms have higher reactivity and can easily react with Ag (CatAg) and C
4 (Imine) atoms to complete the reactions. This result is similar to the experimental results [
40].
In order to better understand the intrinsic characteristics of the reaction and to understand the influence of the concentration of the catalyst (
CatAg) on the reaction, the concvar program [
44] was used. This program simulates the change in the concentration of various substances over the reaction time of the study and provides kinetic data based on the calculated reaction profile. Note that in order to facilitate the study, the reaction pathway was appropriately simplified, and some low-barrier transition states and insignificant intermediates in the reaction path were ignored. The initial concentrations of each of the three reactants (PAE, Amine, and Ald) for this simulation were set to 1.0 M and the catalyst concentration was set to 0.02 M at 333.15 K for 500,000 s. The concentration changes in each substance throughout the simulation process are shown in
Figure 12.
Figure 12 shows that the three reactants (PAE, Amine, and Ald) are consumed at approximately the same rate; hence, it is reasonable to consider that the starting concentrations of the three reactants are nearly equal. This is consistent with the experimental results. At the beginning of the reaction, the concentration of Imine increases first and then decreases gradually as the reaction progresses and the product is formed. The concentration of PPA remains low because PPA reacts easily and forms PR. The concentrations of the product (PR) and H
2O increase rapidly at the beginning and then more slowly as the reaction progresses.
The yield of the product (PR) varies with the catalyst (CatAg) concentration at a constant reaction time of 500,000 s, as shown in
Figure 13. The yield of the product (PR) increases with increasing catalyst (CatAg) concentration. In particular, the product (PR) yield increases rapidly when the catalyst (CatAg) concentration increases from 0.00 M to 0.02 M. After 0.02 M, the product (PR) yield increases slowly with increasing catalyst (CatAg) concentration.
3. Computational Methods
All computations were performed using the DFT method employing the Gaussian 16 program [
45]. All structures were optimized and characterized by frequency analysis calculations to be minima (without imaginary frequencies) or transition states (TSs) (having unique imaginary frequencies) at the
B3LYP-
D3(BJ)/BS level [
46,
47,
48,
49]. BS, as a basis set, unites the SDD [
50] for Ag and 6-311G (d,p) for nonmetal atoms. A pseudo-potential base set was adopted for Ag atoms. The intrinsic reaction coordinate (IRC) [
51] approach was used to identify the transition states connecting the reactants with the products. The thermodynamic corrections (the correction factor is 0.9670) were calculated based on a rigid-rotor harmonic oscillator (RRHO) model using the Shermo program under 333.15 K and 1 atm [
52]. To improve the energetic results, single-point energy calculations for the optimized structures were performed at the M06-2X/def2-TZVP level [
53,
54] using the SMD [
55] solvent effects model. Free energies obtained using M06-2X (SMD, solvent = acetonitrile)/def2TZVP were considered throughout this study; the relative enthalpies were also provided for reference. The WBI and NBO charges [
56] were calculated at the
B3LYP-
D3(BJ)/BS level. Houk and coworkers [
57] recommended the M06-2X//B3LYP combination to study transition metal-catalyzed reactions. The protocol has been successfully applied to study the mechanisms of transition metal-catalyzed reactions [
58,
59,
60,
61,
62,
63,
64,
65,
66].
Frontier orbitals and certain CDFT data, such as the GRI [
67,
68,
69,
70] and Fukui functions [
71,
72], were obtained using the Multiwfn program 3.8 [
73]. The diagrams for the frontier orbitals and isosurface of the Fukui function were plotted using the VMD program 1.9.3 [
74]. The change in the concentration of various substances over the reaction time was simulated using the concvar program [
44], and the reaction rate constants were calculated based on the transition state theory [
75]. Additional computational details and detailed data of Fukui functions values for some molecules are given in
Table S1. Energies, vibrational frequencies and Cartesian coordinates of all the optimized structures are given in
Table S2.
4. Conclusions
The A3 coupling reactions catalyzed by CatAg were calculated using DFT. First, an addition reaction between CatAg and PAE forms Ag_PAE. Subsequently, one hydrogen atom of the Ag_PAE migrates to the nitrogen atom of the Amine. Thereafter, the amine–aldehyde condensation reaction generates a water molecule and an Imine. The amine–aldehyde condensation reaction can proceed through two pathways: with (Path one) or without (Path two) another amine catalyst. Path one has a lower reaction barrier than Path two. The Imine reacts with Ag_PAI to generate the A3 coupling reaction product, PPA. Moreover, the cycloaddition reaction between the triple bond and −N3 group in PPA generates the PR, and the CatAg is regenerated. The entire reaction is strongly exothermic, and, therefore, it is easy to perform, which is consistent with our experimental results. Additionally, CDFT data analyses confirmed this reaction mechanism. The detailed mechanistic insights of these reactions presented in this paper advance the understanding of the reactions of aldehydes, alkynes, and amines catalyzed by CatAg and will be beneficial for developing a new synthesis strategy for similar functional compounds.