Catalytic Activity of trans-Bis(pyridine)gold Complexes.

Gold catalysis has become one of the fastest growing fields in chemistry, providing new organic transformations and offering excellent chemoselectivities under mild reaction conditions. Methodological developments have been driven by wide applicability in the synthesis of complex structures, whereas the mechanistic understanding of Au(III)-mediated processes remains scanty and have become the Achilles' heel of methodology development. Herein, the systematic investigation of the reactivity of bis(pyridine)-ligated Au(III) complexes is presented, based on NMR spectroscopic, X-ray crystallographic, and DFT data. The electron density of pyridines modulates the catalytic activity of Au(III) complexes in propargyl ester cyclopropanation of styrene. To avoid strain induced by a ligand with a nonoptimal nitrogen-nitrogen distance, bidentate bis(pyridine)-Au(III) complexes convert into dimers. For the first time, bis(pyridine)Au(I) complexes are shown to be catalytically active, with their reactivity being modulated by strain.

Introduction

Despite humankind’s fascination with gold ever since ancient times, n class="Chemical">Au catalysis has lagged behind the chemistry of other transition metals, such as palladium, rhodium, and platinum. Recently, significant efforts have been put into the establishment and understanding of Au(I)-catalyzed processes,[1,2] whereas Au(III)-mediated reactions have received only minute interest.[3−5] They have sometimes been simplistically rationalized as Au(I) catalytic processes, with Au(III) being a precatalyst,[6] or as intermediates in a catalytic redox cycle starting with Au(I).[7−10] This lag is likely explained by the challenging lability of Au(III) complexes by the largely unexplored potential of their ligation and by the lack of experimental evidence necessary for understanding of the mechanism of Au(III)-mediated processes.[11] Ligated Au(III) complexes have so far primarily been developed for biological applications,[12] with only scarce examples of catalysts designed for organic reactions.[13−25] Such ligated complexes have been reported to exhibit higher stability and increased catalytic activity as compared to inorganic Au(III) salts. Current ligated Au(III) complexes are typically based on bidentate cis-chelating ligands and contain oxazolines,[15,23,26,27]N-heterocyclic carbenes,[24,28] aromatic groups,[25,29−31] or salen.[32]Trans-chelating ligands have so far barely been studied. Pyridine derivatives, a frequent motif in transition metal chemistry,[33,34] have also been explored for ligation of Au(III)[35−38] and have shown to yield an improved catalytic performance and stability[39] as compared to simpler Au(III) salts, e.g., K/NaAuCl4, AuCl3, and AuBr3, presumably due to activation and stabilization of the Au(III) center. Whereas some initial data exist, the catalytic activity of pyridine complexes has not yet been systematically investigated, and the mechanism of pyridine-ligated Au(III)-mediated reactions remains unexplored. Motivated by this knowledge gap, we studied bis(pyridine)Au(III) complexes, evaluating the influence of electronic and steric effects on their reactivity and investigating their coordination mode by using NMR spectroscopic, X-ray crystallographic, and computational (DFT) techniques.

Results and Discussion

Bis(pyridine)Au(III) chloride was pn class="Chemical">repared by addition of 2 equiv of pyridine to a methanol solution of KAuCl4, resulting in immediate precipitation of [(1-H)2-Au(III)]Cl (Figure and Table ). The single set of 1H, 13C, and 15N NMR signals observed for this product was compatible with that previously reported.[40] Complex [(1-H)2-Au(III)]Cl readily catalyzed the propargyl ester (5) cyclopropanation of styrene (6), yielding full conversion within 1 h (Table , entry 1). To evaluate the impact of electron density on catalytic activity, the 4-substituted complexes [(1-CF3)2-Au(III)]Cl, [(1-CH3)2-Au(III)]Cl, and [(1-OCH3)2-Au(III)]Cl were also prepared. The electron-poor [(1-CF3)2-Au(III)]Cl showed increased catalytic activity, giving full conversion within 30 min (Table , entry 2), most likely as a result of weaker coordination of the ligand. The electron rich-complexes [(1-CH3)2-Au(III)]Cl and [(1-OCH3)2-Au(III)]Cl were strongly deactivated (Table , entries 3 and 4). 15N NMR chemical shifts provided valuable information about metal coordination strength.[43] The 15N NMR coordination shifts (Δδ15Ncoord) of the [(1-R)2-Au(III)]Cl complexes (Table ) corroborate the electron density dependence of the reactivity. This is further supported by the increased stability (ΔGstab, Table ) of the more electron-rich complexes, as computed by DFT (for details, see the Supporting Information).

Figure 1

Pyridine-based ligands utilized for Au(I) and Au(III) ligation. Here, R may be H (1-H and 2-H), CF3 (1-CF3 and 2-CF3), CH3 (1-CH3 and 2-CH3), OCH3 (1-OCH3 and 2-OCH3), or asymmetric CH3/CF3 (2-CH3/CF3). Ligands 2–4[41,42] have previously not been utilized in Au-mediated catalysis.

Table 1

Reactivity of the [(1-R)2-Au(III)]+ and [(1-H)2-Au(I)]+ Complexes in Cyclopropanationa

entry	Au catalyst	reaction time	conv [%] (trans:cis)b
1	[(1-H)2-Au(III)]Cl	1 h	100 (25:75)
2	[(1-CF3)2-Au(III)]Cl	30 min	100 (36:64)
3	[(1-CH3)2-Au(III)]Cl	12 h	83 (17:83)
4	[(1-OCH3)2-Au(III)]Cl	12 h
5	[(1-H)2-Au(I)]BF4	24 h	<5 (<1:>99)

No conversion into product is observed without the presence of an Au catalyst.

As this ratio has previously been studied in detail,[11] it is not discussed here.

Table 2

Experimental δ15N and Δδ15Ncoord NMR Chemical Shifts in CD2Cl2, Calculated Au–N Bond Length, Calculated Changes of the Total Electron Population of the Pyridine Nitrogen Δn(N) upon Substitution: as Estimated by Natural Atomic Populations Analysis for Pyridine Ligands 1-R and the Calculated Stabilization Energies of Bis(pyridine) Au(I) and Au(III) Complexes of Pyridine Ligands 1-Ra

complex	δ 15Ncomplex	δ15Nligand	Δδ15Ncoord	Au–N bond length [Å]	103Δn(N)	ΔGstabb [kcal/mol]
[(1-H)2-Au(I)]BF4	–155.6	–67.0	–88.6	2.056
[(1-H)2-Au(III)]Cl	–154.2	–67.0	–87.2	2.046	0.0	0.0
[(1-CF3)2-Au(III)]Cl	–144.1	–58.0	–86.1	2.047	–17.0	10.4
[(1-CH3)2-Au(III)]Cl	–162.6	–71.6	–91.0	2.044	8.0	–1.9
[(1-OCH3)2-Au(III)]Cl	–178.2	–86.0	–92.2	2.042	27.0	–2.7

See the Supporting Information for computational details.

Calculated free energy change of the [(1-H)2-Au(III)]+ + 2 × 1-R → [(1-R)2-Au(III)]+ + 2 × 1-H isodesmic reaction.

No conversion into pn class="Chemical">roduct is observed without the presence of an Au catalyst.

As this ratio has pn class="Chemical">reviously been studied in detail,[11] it is not discussed here.

See the Supporting Infon class="Chemical">rmation for computational details.

Calculated free enen class="Chemical">rgy change of the [(1-H)2-Au(III)]+ + 2 × 1-R → [(1-R)2-Au(III)]+ + 2 × 1-H isodesmic reaction.

As Au(III) has pn class="Chemical">reviously been suggested to show catalytic activity by being a precatalyst for Au(I)-mediated processes, we evaluated the catalytic activity of [(1-H)2-Au(I)]BF4 (Table , entry 5). Its low activity revealed that the oxidation state of Au(III) plays an important role in the catalysis of the propargyl ester cyclopropanation and that [(1-H)2-Au(III)]Cl is highly unlikely to act as a precatalyst for the corresponding Au(I) complex. Calculations predicted the preference of trans coordination of 1-R pyridine ligands in [(1-R)2-Au(III)]+complexes, with the cis geometries being 1.2–6.3 kcal/mol less stable (see the Supporting Information). This finding is corroborated by the X-ray crystallographic data of bis(pyridine)-type Au(III) complexes.[44] To evaluate whether a trans-coordinating ligand may lead to a catalytically active species, we synthesized the geometrically restricted 1,2-bis(4-R-pyridin-2-yl)ethynyl)benzene, 2 (Figure and Scheme ). This bidentate ligand, and its close structural analogues, have earlier been successfully utilized in palladium,[45−49] mercury,[50] copper,[51] silver, and halonium complexes,[41,52−56] making its application in Au(III) chemistry plausible. Coordination of KAuCl4 with 2-H readily took place, resulting in precipitation of [(2-H)-Au(III)]Cl upon mixing in methanol (Scheme ). Given its poor solubility in a wide range of solvents, we also prepared [(2-H)-Au(III)]BF4 from the corresponding Au(I) complex, [(2-H)Au(I)]BF4, by oxidation using dichloro(phenyl)-λ3-iodane (Scheme ).[57] Notably, the [(2-H)-Au(III)]BF4 complex could not be formed via an anion exchange with [(2-H)-Au(III)]Cl. The [(2-H)-Au(III)]Cl and [(2-H)-Au(III)]BF4 complexes both catalyzed the cyclopropanation reaction, giving full conversion within 5 h (Table , entries 1 and 2). Despite its somewhat lower reactivity as compared to the analogous bis(pyridine) complex [(1-H)2-Au(III)]Cl (Table , entry 1), this observation proves, for the first time, that Au(III) complexes of trans-chelating bidentate ligands are catalytically active.

Scheme 1

Formation of the Au(I) and Au(III) Complexes of Ligand 2-H and the Au(III) Complexes of Its 4-Substituted Analogues, 2-R, Where R Is H (2-H), CF3 (2-CF3), CH3 (2-CH3), OCH3 (2-OCH3), or Both CH3 and CF3 (2-CH3/CF3)

Table 3

Reactivity of the [(2-R)-Au(III)]+ and [(2-H)-Au(I)]+ Complexes in the Cyclopropanation Reaction

entry	Au catalyst	reaction time	conv [%] (trans:cis)
1	[(2-H)-Au(III)]BF4	5 h	100 (23:77)
2	[(2-H)-Au(III)]Cl	5 h	100 (21:79)
3	[(2-CF3)-Au(III)]Cl	30 min	100 (17:83)
4	[(2-CH3)-Au(III)]Cl	10 h	41 (24:76)
5	[(2-OCH3)-Au(III)]Cl	10 h	75 (18:82)
6	[(2-CH3/CF3)-Au(III)]Cl	10 h	8 (<1:>99)
7	[(2-H)-Au(I)]BF4	12 h	40 (26:74)

The lower n class="Chemical">reactivity of [(2-H)-Au(III)]Cl may be explained by the stronger coordination of Au(III) to 2-H as compared to pyridine (1-H), reflected by the 24 ppm smaller |Δδ15Ncoord| of [(1-H)2-Au(III)]Cl (87.2 ppm, Table ) as compared to [(2-H)-Au(III)]Cl (111.2 ppm, Table ). This different coordination strength is corroborated by the calculated shorter N–Au distance of [(2-H)-Au(III)]Cl, 2.037 Å (Table ) as compared to [(1-H)2-Au(III)]Cl, 2.046 Å (Table ). The 4-substituted analogues [(2-R)-Au(III)]Cl were prepared by following same the method described for [(2-H)-Au(III)]Cl and showed higher solubility in dichloromethane. The electron-poor [(2-CF3)-Au(III)]Cl and the asymmetric [(2-CH3/CF3)-Au(III)]Cl complex were less stable than [(2-H)-Au(III)]Cl as judged from their stability in a dichloromethane solution monitored by 1H NMR and confirmed by DFT computations (Table ). The reactivity order of complexes [(2-R)-Au(III)]Cl followed the same trend in the cyclopropanation model reaction (Table ) as the [(1-R)2-Au(III)]Cl complexes (Table ). This suggests that the geometrically restrained pyridine complexes react with the same mechanism as the unrestrained ones. Accordingly, the Δδ15Ncoord of [(2-R)-Au(III)]Cl (Table ) showed linear correlation to the σpara Hammett substituent constant (Figure ), indicating that the reactivity of pyridine complexes is systematically modulated by electron density. A similar correlation was observed for [(1-R)2-Au(III)]Cl complexes (Table ).

Table 4

Experimental δ15N and Δδ15Ncoord NMR Chemical Shifts CD2Cl2, Calculated Au–N Bond Length, and Calculated Changes of the Total Electron Population of the Pyridine Nitrogen Δn(N) upon Substitution: as Estimated by Natural Atomic Populations Analysis for Ligands 2-R

complex	δ 15Ncomplex	δ 15Nligand	Δδ 15Ncoord	Au–N bond length [Å]	ΔGstaba [kcal/mol]
[(2-H)-Au(III)]Clb	–179.8b	–63.6b	–111.2b	2.037	0.0
[(2-H)-Au(III)]BF4c	–178.9c	–75.0c	–103.9c	2.037	0.0
[(2-CF3)-Au(III)]Cl	–142.6	–50.7	–91.9	2.037	7.3
[(2-CH3)-Au(III)]Cl	–186.5	–69.2	–117.3	2.035	–3.1
[(2-OCH3)-Au(III)]Cl	–202.4	–83.5	–118.9	2.034	–4.8
[(2-CH3/CF3)-Au(III)]Cl	-159.5	–53.8	–105.7	2.042	2.0
	–189.5	–71.8	–117.7	2.030
[(2-H)-Au(I)]BF4	–150.1	–64.5	–85.6	2.064	–d
[(2-H)-Au(I)]BF4c	–151.3c	–75.0c	–76.3c	2.064	–d

Energies are given as stabilization energies relative to [Au(III)(2-H)]Cl.

Acquired in DMSO-d6.

Acquired in methanol-d4.

Not relevant to compare to the Au(III) complex due to lack of the two coordinating chlorides.

Figure 2

Correlation of the Hammett substituent constant and the Δδ15Ncoord shift of the [(2-R)-Au(III)]Cl complexes. R2 = 0.99.

Not relevant to compan class="Chemical">re to the Au(III) complex due to lack of the two coordinating chlorides.

This experimental finding is suppon class="Chemical">rted by the variation of the net atomic charge of the pyridine nitrogen atom upon para substitution, as obtained by natural atomic population (NAP) analysis,[58] being well-correlated to the Δδ15Ncoord of pyridine complexes (Table ; for details see the Supporting Information).

Whereas n class="Chemical">[(2-H)-Au(III)]Cl is less reactive (100% in 5 h) than [(1-H)2-Au(III)]Cl (100% in <1 h), [(2-H)-Au(I)]BF4 shows higher reactivity (40% in 12 h) than its unrestrained analogue [(1-H)2-Au(I)]BF4 (<5% in 2 h). The Δδ15Ncoord of [(1-H)2-Au(I)]BF4 (−88.6 ppm, Table ) is comparable yet somewhat larger than that of [(2-H)-Au(I)]BF4 (−85.6 ppm, Table ), and the Au(I)–N bond length of the conformationally adjustable 1-H complex is somewhat shorter (2.056 Å) than that of the geometrically restricted complex of 2-H (2.064 Å). This suggests that a weak geometrical restraint that enforces weaker than optimal coordination of pyridine promotes the reactivity of Au(I). This is significant as pyridine is widely accepted to deactivate Au(I), and thus the reactivity of [(2-H)-Au(I)]BF4 in cyclopropanation is unprecedented. This chelating ligand design could pave the way for the development of new types of catalytically active Au(I) complexes in organic reactions.

The DFT calculated Au–N bond lengths of the complexes n class="Chemical">[(1-H)2-Au(III)]Cl (2.046 Å) and [(2-H)-Au(III)]Cl (2.037 Å) are shorter than the corresponding bond lengths of the analogous Au(I) complexes (Tables and 4), in agreement with Au(III) being smaller and more electrophilic than Au(I). Overall, the reactivity of both Au(I) and Au(III) complexes appears to be a function of coordination strength. Weaker coordination results in higher catalytically activities and is associated with longer Au–N bond lengths and smaller |Δδ15Ncoord|.

In an attempt to evaluate whether an asymmetric coordination of the nitrogen ligands with a trans-bis(pyridine)Au(III) complex could be advantageous for catalytic activity, we synthesized the asymmetrically substituted complex [(2-CH3/CF3)-Au(III)]Cl. The different bond strengths of the electron-poor and the electron-rich pyridines were confirmed by the two Au–N bond lengths, as calculated by DFT, 2.042 and 2.030 Å, respectively (Table ), with these bond lengths being longer and shorter than those of the corresponding symmetric complexes [(2-CH3)-Au(III)]Cl (2.035 Å) and [(2-CF3)-Au(III)]Cl (2.037 Å). The different strength of the Au–N bonds in the asymmetric complex is corroborated by different |Δδ15Ncoord| shifts of the two nitrogens (Table ). The |Δδ15Ncoord| of the electron-rich pyridine of [(2-CH3/CF3)-Au(III)]Cl (117.7 ppm) is comparable to the |Δδ15Ncoord| of the corresponding symmetric complex [(2-CH3)-Au(III)]Cl (117.3 ppm), whereas the |Δδ15Ncoord| of the CF3 substituted pyridine (105.7) is larger than its corresponding symmetric complex [(2-CF3)-Au(III)]Cl (91.9 ppm) (Table ). This suggests that dissociation of the weaker coordinative Au–N bond, which is the one to the 4-trifluoromethylpyridine, is less favored than dissociation of the corresponding bond of [(2-CF3)-Au(III)]Cl. Accordingly, [(2-CH3/CF3)-Au(III)]Cl shows lower catalytic activity in cyclopropanation (8% in 10 h) than [(2-CF3)-Au(III)]Cl (100% in 30 min, Table ). The Au–N bond dissociation may thus play a key role in the mechanism of Au(III)-mediated cyclopropanation.

To gain insights into the mechanism, we computed the Gibbs free enen class="Chemical">rgy profile for the envisioned intermediates and the related transition state of the key reaction between [(1-H)2-Au(III)]Cl and propargyl ester (5) as shown in Figure (see computational details in the Supporting Information). The reaction necessitates coordination of the substrate, which presumes decoordination of one of the four Lewis bases of the tetracoordinate Au(III) species. DFT calculations predict the decoordination of a pyridine to be vastly more favored (ΔΔG > 20 kcal/mol) over that of a chloride. The most favored structure of the proposed reaction intermediate, formed upon the exchange of a pyridine with propargyl ester 5, has chlorides in a cis arrangement (intermediate A, Figure ). This intermediate is predicted to be fairly unstable (ΔG = 23.2 kcal/mol) with respect to the [(1-R)2-Au(III)]+ + 5 reactant state; however, it is extremely reactive and easily initiates the 1,2-acyloxy migration process after rotation of the ester to form the transition state (TS). This computationally identified transition state (TS, Figure ) is at 24.4 kcal/mol relative to the reactants and leads to the energetically low-lying cyclic intermediate B (for details, see the Supporting Information). Importantly, the energy barrier represented by this transition state, TS, varies upon the electronic nature of the pyridine 4-substituent. The energy order of 1-CF3 < 1-H < 1-CH3 < 1-OCH3 is in qualitative agreement with the experimentally observed reactivity order of the [(1-R)2-Au(III)] complexes (Table ).

Figure 3

(a) Free energy data predicted for reactive intermediate A, transition state TS initiating the 1,2-acyloxy migration process, and subsequent cyclic intermediate B identified computationally for the reaction between [(1-H)2-Au(III)]+ and propargyl ester 5. Relative stabilities (in kcal/mol) are shown in blue with respect to the [(1-H)2-Au(III)]+ + 5 reactant state. (b) Optimized structures of A, TS, and B. Energy data obtained for the analogous reactions with 1-R = 1-CF3, 1-CH3, and 1-OCH3 are shown before the labels. Details for DFT calculation are given in the Supporting Information.

(a) Free enen class="Chemical">rgy data predicted for reactive intermediate A, transition state TS initiating the 1,2-acyloxy migration process, and subsequent cyclic intermediate B identified computationally for the reaction between [(1-H)2-Au(III)]+ and propargyl ester 5. Relative stabilities (in kcal/mol) are shown in blue with respect to the [(1-H)2-Au(III)]+ + 5 reactant state. (b) Optimized structures of A, TS, and B. Energy data obtained for the analogous reactions with 1-R = 1-CF3, 1-CH3, and 1-OCH3 are shown before the labels. Details for DFT calculation are given in the Supporting Information.

As we observed weaken class="Chemical">r coordination to increase the catalytic activity, the Au(III) complex of 3 (Figure ) was investigated because this ligand was expected to enforce a longer nitrogen–nitrogen distance (Scheme , bottom left) and thereby expectably a weaker coordination to Au(III). Following the synthetic route described for the generation of [(1-H)2-Au(III)]Cl, a mixture of two complexes possessing comparable Δδ15Ncoord, that is, −81.4 ppm [3-Au(III)2] and −80.8 ppm [3-Au(III)]2Cl2, was formed. The complex [3-Au(III)2] was isolated by crystallization and identified by single-crystal X-ray crystallography. Instead of the expected 1:1 ligand to gold stoichiometry, the structure features two Au(III) ions for every ligand (Scheme ). This complex gave full conversion in the cyclopropanation reaction within 10 h, thus showing a decreased reaction rate as compared to the Au(III) complexes of 1-H (Table ) and 2-H (Table ). The lower reactivity of [3-Au(III)2] (10 mol % Au(III), full conversion in 10 h) is in agreement with a decreased reactivity of Au(III) complexes that do not allow easy decoordination of a nitrogenous ligand.[59] Because of the lack of a suitable single crystal for the second ligated complex of 3, DFT calculations were performed to identify plausible structures. The monomeric [3-Au(III)]Cl (Scheme b), possessing a bis-coordinated Au(III) center analogous to [2-Au(III)]Cl, is highly strained and unstable as compared to dimeric structures that have two N–Au–N bridges (Scheme ). Out of the possible dimeric geometries, [3-Au(III)]2Cl2 was found to be the most stable (Supporting Information), presumably due to the advantageous π-stacking of its phenanthrenes.

Scheme 2

Conditions for Coordination of Au(III) by 3 and (a) the Complex Identified by X-ray Analysis, [3-Au(III)2] Complex, and (b) Computationally Identified Monomeric and Dimeric Forms of Complex [3-Au(III)]2Cl2

The free energy of dimerization (ΔG) is shown above the arrow. Hydrogen atoms are omitted for clarity. Details for DFT calculation are given in the Supporting Information.

The free enen class="Chemical">rgy of dimerization (ΔG) is shown above the arrow. Hydrogen atoms are omitted for clarity. Details for DFT calculation are given in the Supporting Information.

To evaluate whether horizontal twisting of the coordination plane of the N–Au–N interaction may modulate the catalytic activity of Au(III), we mixed the bidentate ligand 4 with AuCl3 (Scheme ). Crystals suitable for X-ray analysis were grown by slow diffusion of n-pentane into a dichloromethane solution. Formation of a dimeric helix, [4-Au(I)]2(AuCl4)2, associated with partial reduction of Au(III) to Au(I) was revealed by single-crystal X-ray crystallography. Hence, Au(I) was observed to coordinate to ligand 4 and AuCl4– to act as a counterion. Upon coordination of Au(I) to ligand 4, starting from chloro(dimethyl sulfide)Au(I) and AgBF4, the analogous [4-Au(I)]2(BF4)2 complex was obtained (Scheme ). The Δδ15Ncoord of −83.2 for [4-Au(I)]2(BF4)2 was slightly smaller than the coordination shift of [(1-H)2-Au(I)]BF4 and [(2-H)-Au(I)]BF4, confirming the weaker coordination of Au(I) to 4 as compared to 1-H and 2-H (Tables and 4). Formation of the dimeric helical [4-Au(I)]2 geometry allows close to linear N–Au–N interactions (176.6°) and optimal Au(I)–N bond lengths (2.01 Å) and thus avoids horizontal twisting that would have been enforced in a [4-Au(III)]Cl single-helix geometry (Scheme a).

Scheme 3

Coordination Condition for Formation of (a) [4-Au(I)]2(AuCl4)2 by Addition of the Ligand 4 Dissolved in Dichloromethane to AuCl3 Dissolved in Methanol and (b) [4-Au(I)]2(BF4)2 Starting from Chloro(dimethyl sulfide)gold(I) and AgBF4 in Dichloromethane

Crystal structures of complexes formed with two different counterions are depicted at the bottom of the scheme.

Crystal stn class="Chemical">ructures of complexes formed with two different counterions are depicted at the bottom of the scheme.

Ligand 4 has been repon class="Chemical">rted to form a monomeric helix when coordinating iodine(I) [4-I(I)]BF4.[42] The difference in topology of the [4-Au(I)]2(BF4)2 dimeric helix from [4-I(I)]BF4 (Scheme ) is corroborated by DFT calculations, which show a thermodynamic preference for formation of a dimeric, double-helical structure for [4-Au(I)]2(BF4)2 over a monomeric helical complex. This is most plausibly explained by stronger π-stacking interactions in the double helical [4-Au(I)]2(BF4)2 geometry due to the smaller van der Waals radius of Au(I) as compared to I(I), which favors the monomeric helical structure (for details, see the Supporting Information).

Scheme 4

Computational (DFT) Investigation of the Relative Stability of Monomeric and Dimeric Helical Structures of 4 in Complexes with Au(I) and I(I) Reveal Opposite Preferences, Corroborating the Experimental Observations

Thus, Au(I) prefers to form a dimeric, whereas I(I) a monomeric helix. Energy values shown in parentheses refer to the free energy changes for the interconversion of the monomeric and dimeric geometries. Distances between the Au(I) and I(I) centers are given in Å. Details for the DFT calculation are given in the Supporting Information.

Thus, Au(I) pn class="Chemical">refers to form a dimeric, whereas I(I) a monomeric helix. Energy values shown in parentheses refer to the free energy changes for the interconversion of the monomeric and dimeric geometries. Distances between the Au(I) and I(I) centers are given in Å. Details for the DFT calculation are given in the Supporting Information.

The [4-Au(I)]2(AuCl4)2 complex gave full conven class="Chemical">rsion within 1 h as catalyst for the cyclopropanation model reaction, whereas barely 19% conversion was obtained within 24 h for the analogous [4-Au(I)]2(BF4)2. This reveals the AuCl4– counterion of the former complex is primarily responsible for the catalytic activity, whereas the Au(I) within the complex has a low catalytic activity. This is in excellent agreement with the low catalytic activity of [(1-H)2-Au(I)]BF4 (Table ). The reactivity order [(1-H)2-Au(I)]BF4 < [4-Au(I)]2(BF4)2 < [(2-H)-Au(I)]BF4 suggests that the catalytic activity of bis(pyridine)Au(I) complexes is modulated by the geometric strain of the N–Au–N three-center bond.

Conclusion

Bis(pyridine)Au(III) complexes an class="Chemical">re applicable as catalysts as shown by using the cyclopropanation of styrene (6) with propargyl ester (5) as a model reaction. By systematic NMR spectroscopic, X-ray crystallographic, and computational assessment of a series of structurally related complexes, we demonstrated that the electron density of the pyridine ligands has a pivotal influence on the catalytic activity of Au(III). Accordingly, the energy requirement of the formation of a key intermediate and the subsequent transition state is predicted by DFT to depend on the electron density of the nitrogen of the pyridine ligand. The catalytic activity of geometrically restrained trans-bis(pyridine)Au(III) complexes was investigated here for the first time. These complexes also show electron-density-dependent reactivity, which is somewhat lower than that of the Au(III) complexes of free pyridines. DFT calculations suggest that these types of complexes convert into a cis-dichlorido(pyridine)Au(III) intermediate upon decoordination of a pyridyl functionality. Geometrical restriction of bis(pyridine)-type ligands enforced increased strain in the N–Au–N bond, resulting in the formation of a variety of dimeric structures, including among others an unusual double-helical supramolecular assembly. Whereas cyclopropanations initiated with bis(pyridine)Au(III) complexes are shown to be Au(III)-mediated reactions, the lower yet significant catalytic activity of bis(pyridine)Au(I) complexes is demonstrated here for the very first time. The reactivity appears to be modulated by strain and thus is expected to be tunable.

Despite a quickly growing inten class="Chemical">rest in Au(III) catalysis, its understanding lags behind that of transition metal catalysis in general. So far, the mechanism of most Au(III)-catalyzed reactions is not well understood. Furthermore, systematic, combined spectroscopic and theoretical investigations of the mechanism of Au(III)-mediated reactions are rare. This work is expected to contribute to providing the basis of an improved understanding of Au(III) chemistry and the development of new, robust synthetic methodologies.