Pyridine Wingtip in [Pd(Py-tzNHC)₂]²⁺ Complex Is a Proton Shuttle in the Catalytic Hydroamination of Alkynes

Abstract

The cationic palladium(II) complex 1 of pyridyl-mesoionic carbene ligand catalyzes Markovnikov-selective intermolecular hydroamination between anilines and terminal alkynes into the corresponding imines. The reaction proceeds at room temperature, in the absence of additives, with exquisite selectivity and diverse functional group tolerance. The key intrinsic feature of the catalyst is the pyridine wingtip confined to the proximity of the alkynophilic metal active site, which mimics the function of enzyme-like architectures by assisting entropically favored proton transfers.

Ligands that provide additional functions beyond their traditional spectator roles are becoming increasingly important in transition-metal catalysis.¹ A metal–ligand cooperation concept has been realized by bifunctional ligands, mostly by establishing control of a catalytic process through weak catalyst–substrate interactions.² Examples where the pyridine fragment of such a bifunctional ligand assists the catalytic process by a proton transfer are emerging and have been demonstrated with ruthenium^3a,3b and iridium^3c catalysts. In palladium catalysis, proton shuttle has been rationalized by an SCS indenediide pincer complex.^3d

A palladium(II) complex [Pd(Py-tzNHC)₂]²⁺ 2BF₄^– (1, Figure 1)⁴ was sought ideal for the metal–ligand cooperation process. While pyridyl-1,2,3-triazol-5-ylidene (Py-tzNHC) strongly stabilizes the metal through the mesoionic triazolylidene (tz) NHC dent, the more labile pyridine wingtip could play a dual role. By coordination, it internally stabilizes the metal center throughout the preparation of the complex and, more importantly, in the resting state of the catalytic cycle. Upon dissociation, it should expose the Lewis acidic metal center to enable substrates to enter the reactive zone and initiate the catalytic cycle while acting as a local Lewis base functionality in the proximity of the metal, enabling a proton-transfer process.

Figure 1.

Complexes 1 (R = CH₃) and 1′ (R = H). X = BF₄.

To test this hypothesis and its potential utility, the intermolecular hydroamination of terminal acetylenes 2 with anilines 3 (Figure 2) was chosen as a model reaction, as it would likely benefit from such a combination of metal/base centers given the multiple proton transfers involved.

Figure 2.

Substrate scope of Pd-catalyzed intermolecular hydroamination. ^a Conversion and ^b isolated percent yields.

Numerous catalysts based on different metals have been developed to promote the intermolecular catalytic hydroamination of terminal acetylenes.⁵ Palladium catalysis is rare in hydroamination of terminal alkynes and mostly comprises the examples developed by the groups of Schmidt,⁶ Huynh,⁷ Cao,⁸ and Biffis.⁹ These catalysts operate at elevated temperatures (80–100 °C) and require additives (TfOH, AgOTf). The hydroamination is sometimes accompanied by undesired side reactions like cyclotrimerization of alkyne component, consequently requiring a large excess of this reagent.

To assess the utility of 1, phenylacetylene 2a and (2,6-dimethyl)aniline 3a were selected as model substrates. The test reaction was conducted in toluene at room temperature. The selection of this solvent was based on previously documented good performance, albeit the reported transformations took place at substantially higher temperatures (100 °C).^7,8 Monitoring the course of the reaction revealed slow formation of the desired imine 4a. This promising finding prompted us to screen through the reaction parameters including other solvents, substrates ratios, catalyst loading, and concentrations. This analysis revealed that 1 mol % loading of 1 efficiently promoted hydroamination of phenylacetylene 2a (1.2 equiv) with (2,6-dimethyl)aniline 3a (1.0 equiv, 0.5 M) in dichloromethane. Within 24 h at room temperature, the conversion exclusively to the Markovnikov product 4a was 83%. Besides 4a, unreacted aniline 3a (17%) and the excess amounts of acetylene 2a, without any side product, could be detected as judged by ¹H NMR analysis of the crude reaction mixture.

Using these optimal reaction conditions along with somewhat higher catalyst loadings of 2 mol % to ensure higher conversions, we examined the substrate scope of the reaction. As shown in Figure 2, a variety of functional groups is tolerated including hydroxyl (as phenol), methoxy (anisole), methyl, trifluoromethyl, ester, chlorine, bromine, and a C=C double bond in isopropenyl group. Sterically demanding anilines also react well. Excellent results were obtained with acetylene and aniline coupling partners having electron-releasing groups, whereas the presence of electron-deficient substituents slightly eroded the conversions (Figure 2). No isomeric anti-Markovnikov imine could be detected by ¹H NMR analyses of the crude reaction mixtures in any of these experiments.

To gauge the direct involvement of the pyridine wingtip into the hydroamination process, we performed the same reaction as described above by using [Pd(ImNHC)] complex A and [Pd(Ph-tzNHC)] complex B. In the latter, the ligand strongly resembles Py-tzNHC in 1 with the pyridine wingtip being replaced by phenyl (Figure 3). In contrast to 1, the pyridine and NHC parts are not bound together in A and B, allowing the complete dissociation of pyridine from the metal upon engaging in a reaction. No product formation could be observed in the reaction of 2a and 3a after 24 h in the presence of these complexes, neither in dichloromethane at room temperature nor in toluene at 100 °C. This striking difference between the activities of 1 and complexes A and B implies that the pyridine wingtip of the iPEPPSI ligand indeed plays an explicit role in the overall function of catalyst.

Figure 3.

No hydroamination occurs with complex A or B.

To expose the role of the pyridine wingtip in this catalytic process we scrutinized the mechanism of hydroamination to generate imine product 4b using solution-state density functional theory (DFT) simulations carried out using DFT as implemented in ORCA 4.0.1.2, employing the method PBE0-D3/def2-SV(P)/RIJCOSX/Grid4,GridX4 for geometry optimizations and PBE0-D3/def2-TZVPP/Grid5,GridX5 for single-point calculations (Supporting Information).¹⁰ Solvation energies in dichloromethane as solvent were computed with the latter computational method, using the SMD implicit solvation model. In our in silico investigation we used slightly truncated model of the complex with 4-tolyl replaced by phenyl group (1′, Figure 1). The methyl substituents, relatively far from the metal center and without significant electronic effects, do not affect the general catalytic properties of the [Pd(Py-tzNHC)] system and overall energy barriers. Additional details of computational investigations are provided in the Supporting Information. The energy profile of the most likely pathway for the catalytic formation of the imine (4b) from phenylacetylene (2a) and aniline (3b) together with a schematic description of the operative mechanism is illustrated in Figures 4 and 5. In addition, Figure 5 highlights five chemically well-identifiable subprocesses, starting with C–N bond formation and terminating with imine dissociation. Competing processes and less likely alternative mechanisms for some of these subprocesses are also discussed in detail in the Supporting Information, together with other pathways that were ruled out due to being unreasonably high in energy.

Figure 4.

Proposed reaction pathway.

Figure 5.

Computational energy trajectory.

As expected, and previously proposed for the catalytic hydroamination reactions,⁵ the operative pathway begins with acetylene coordination to the catalyst (1′) forming the adduct 5 (Figures 4 and 5). In order to accommodate the acetylene substrate in the first coordination sphere of a stable square-planar arrangement, one of the pyridine side arms decoordinates from the Pd(II) center in 5, allowing tight binding of acetylene to the metal. Although unbound, as being tethered to the mesoionic NHC ring, this pyridine remains in the proximity (<5.0 Å) of the metal forming together a multicenter reactive zone comprising a Lewis base and an acid center. This replacement of the pyridine by the incoming acetylene takes place through an associative interchange process traversing the tbp transition state TS⁵ with solution-state Gibbs free energy of 11.7 kcal mol^–1. The η² binding of acetylene to palladium via one of its π bonds activates its carbon atoms toward nucleophilic attack facilitating direct carbon–nitrogen bond formation with the amine substrate. The tell-tale structure of the corresponding transition state (TS⁶), displayed in Figures 4 and 5, conforms to the expected electronic transitions, namely, binding to the metal through an electron rich carbon meanwhile accepting the lone-pair of the amine at the electron deficient carbon center. This C–N bond formation process is associated with an activation barrier of about 23 kcal mol^–1, which is surmountable under standard conditions. While the direct product of this step, complex 6, is not particularly stable (8.5 kcal mol^–1), a subsequent intramolecular proton transfer stabilizes the system.

This proton transfer takes place between the secondary ammonium fragment formed upon C–N bond formation and the liberated pyridine side arm leading to intermediate 7 with an amine substrate and a pyridinium group. A thermodynamically rather stable enamine complex, 8, is formed upon passing of this pyridinium proton to the carbanion center of the substrate. Both of these proton transfer events are elementary processes associated with a low activation barrier of ∼5–7 kcal mol^–1 to the preceding intermediates and, accordingly, are facile transformations under the experimental conditions. The structure of TS^6/7 (Figures 4 and 5) is evidence that the pyridine side arm has an ideal distance and arrangement to engage in structurally unconfined proton-transfer processes.

To generate the imine product (4b) from the enamine substrate in 8, a formal imine–enamine tautomerism needs to take place along the reaction coordinate. Our simulations imply that such a tautomerism is most effective if the enamine substrate binds to the metal through its amine (N) functionality rather than via its C=C (C) π bond. The requisite C/N coordination switch of the enamine substrate may easily occur in a single step, through TS^8/9. The tautomerization process begins from complex 9 with a proton transfer, then again, to the pyridine wingtip forming another transient intermediate with pyridinium functionality (10). Traversing TS^10/11, shown in Figures 4 and 5, the proton of the pyridinium moiety is transferred to the terminal CH₂ group forming the imine–palladium complex 11. The two-step process from 9 to 11 is virtually identical to a conventional base-catalyzed imine–enamine tautomerism, but the role of the base is played by the pyridine wingtip of the Py-tzNHC ligand.

Finally, the extrusion of the imine product 4b and the simultaneous regeneration of catalyst 1′ occurs traversing a tbp transition state, TS^4b+1′, characterizing a classical associative interchange ligand substitution of the imine to the pyridine side arm. The overall reaction is exergonic by about −23 kcal mol^–1.

Evident from the mechanisms of enamine to imine tautomerization and first proton-transfer steps is the explicit role of pyridine in the discovered catalytic reactivity of 1. Namely, the moderate basicity of pyridine opens low energy reaction channels for the former proton transfer processes through stable pyridinium-containing intermediates, 7 and 10, with solution-state relative stability of about 2 kcal mol^–1 to free reactants.

An experimental mechanistic investigation comprising kinetics and KIE studies provides support for the key features of our mechanistic proposal; a method of initial rates for the model reaction between phenylacetylene 2a and (2,6-dimethyl)aniline 3a revealed orders of 0.5 in acetylene, 0.3 in aniline, and 1.4 in complex 1. In addition, the use of deuterated (2,6-dimethyl)aniline (PhND₂, 3a-d₂) leads to a primary KIE of 2.1, whereas the use of PhC≡CD (2a-d) results in the absence of primary KIE (1.1). These observations are intuitive to interpret by an early pre-equilibrium (1′ + 2a = 5) and a late rate-determining proton-transfer process. Indeed, the imine–enamine tautomerism, i.e., going from 8 to 11, is associated with an apparent, rate-determining barrier of 25 kcal mol^–1 centered by TS^10/11 and representing proton transfer between the pyridinium group and the terminal CH₂ group, giving rise to the observed primary KIE when using 3a-d₂. The structure of this rate determining TS accounts for the modest primary kinetic effect when deuterium is transferred—the effect is moderate in comparison to typical KIEs of 5–6 because the structure and bonding of 8 and TS^10/11 is significantly different. While the approximate first-order kinetics in the catalyst agrees with a metal-mediated tautomerization process, the dependence of the rate also on the concentrations of acetylene (0.5) and aniline (0.3) is best explained by the slightly uphill pre-equilibrium process between catalyst 1′ and acetylene (2a) to form adduct 5 and subsequent C–N bond formation with aniline. Accordingly, these experimental observations are in conceptual agreement with the key features of the mechanism established computationally.

In conclusion, we have demonstrated that beyond mere stabilization and activation of the (pre)catalyst’s metal center, the pyridyl-mesoionic carbene ligand in palladium complex [Pd(Py-tzNHC)₂]²⁺ 2BF₄^– efficiently assists the metal along the reaction trajectory in enzyme-like proton transfer events. In transition-metal catalysis, the examples in which the hemilabiltiy of ligands is related to their catalytic properties are still scarce, and to the best of our knowledge, this is the first demonstration of pyridine-assisted proton shuttle in palladium catalysis. We believe this case study is another step toward the development of novel catalytic systems that are based on these principles.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.orglett.0c00203.

• Experimental procedures and analytical data for the compounds described, computational details, copies of NMR spectra (PDF)

• Cartesian coordinates (PDF)

Author Contributions

^# M.V. and M.M. contributed equally.

The Slovenian Research Agency (Research Core Funding Grant P1-0230, Project J1-8147, Project J1-9166, and Young Researcher Grant to M.V. including support for his 6-months stay at the Universidad Técnico Federico Santa María), the Ministry of Education, Science and Sport, Republic of Slovenia (PhD Scholarship to M.M.), and the Department of Chemistry of UTFSM are gratefully acknowledged for financial support.

The authors declare no competing financial interest.