Noga
Eren
,
Renata
Svecova
,
Farzaneh
Fadaei-Tirani
,
Rosario
Scopelliti
and
Kay
Severin
*
Institut des Sciences et Ingénierie Chimiques, École Polytechnique Fédérale de Lausanne (EPFL), 1015 Lausanne, Switzerland. E-mail: kay.severin@epfl.ch
First published on 13th August 2025
Trinuclear Au3(Pz)3 complexes (Pz = pyrazolate) have been used extensively as components in molecular and polymeric nanostructures. However, the constitutional dynamic chemistry of Au3(Pz)3 complexes remains largely unexplored. We have investigated exchange reactions between Au3(Pz)3 complexes and pyrazole ligands in homogeneous solution. At room temperature and at low millimolar concentrations, several days were needed to establish the thermodynamic equilibrium. The slow exchange kinetics corroborate the inert character of Au3(Pz)3 complexes. When 3,5-diisopropylpyrazolato or 3,5-diphenylpyrazolato ligands were replaced by 3,5-bis(trifluoromethyl)pyrazolato ligands, a sigmoidal rate profile was observed. This observation led to the discovery that pyrazoles can act as potent (auto)catalysts for ligand exchange and ligand scrambling reactions with Au3(Pz)3 complexes. The kinetic studies were supplemented by crystallographic analyses of four heteroleptic Au3(Pz)2(Pz′) complexes.
Au3(Pz)3 complexes have been investigated in soft matter chemistry. Non-covalent assemblies of Au3(Pz)3 complexes were found to form liquid crystals,10 organogels,11 fibers,12 or multilayer vesicles (‘nano onions’).13 Some of these assemblies are luminescent, and stimuli-controlled emission changes have been explored.11
Molecularly defined nanostructures with cage-like architectures were obtained by combining AuI complexes with bridged pyrazole ligands,14 or by linking pre-formed Au3(Pz)3 complexes via metal–ligand interactions15 or dynamic covalent imine chemistry.3 The tetrahedral cage A (Fig. 1b), for example, was obtained by condensation of tris(2-aminoethyl)amine with a Au3(Pz)3 complex featuring pendant phenylaldehyde groups. This cage acts as a potent receptor for C60 and C70.3
![]() | ||
Fig. 1 The general structure of Au3(Pz)3 complexes (a) and graphic representations of a molecular cage (b) and polymeric networks (c) containing Au3(Pz)3 complexes. |
Au3(Pz)3 complexes have also been used for the construction of polymeric framework materials (Fig. 1c).16–18 Materials of this type were employed as catalysts for the carboxylation of alkynes with CO216 or, in combination with Au nanoparticles, as photocatalysts for H2 evolution.17
Metal–ligand exchange reactions are of key importance for the successful construction of metallosupramolecular structures. When the metal–ligand bonds are too inert, error correction processes are suppressed, resulting in the formation of side products. For example, [PtL4]2+ complexes (L = N-donor ligand) are more inert than the analogous [PdL4]2+ complexes, making the construction of [PtL4]2+-based metallosupramolecular structures significantly more challenging.19
Thus far, there is only limited knowledge about the constitutional dynamic chemistry of Au3(Pz)3 complexes. While synthesizing a Au3(Pz)3-containing coordination cage, we noticed that heteroleptic complexes of type Au3(Pz)2(Pz′) do not undergo fast ligand scrambling. The apparent inert character of the Au trimer is in contrast to what was reported for pyrazolate complexes of CuI and AgI, which can form rapid dynamic equilibria in solution.20,21
Below, we describe ligand exchange reactions of Au3(Pz)3 complexes. Surprisingly, exchange reactions between Au3(Pz)3 trimers and ‘free’ pyrazole ligands can display pronounced autocatalytic behavior. This finding led to the discovery that pyrazoles can act as catalysts for ligand exchange and ligand scrambling reactions with Au3(Pz)3 complexes. To supplement the kinetic studies, crystallographic analyses of four heteroleptic Au3(Pz)2(Pz′) complexes were performed.
First, we studied ligand exchange between Au3(PzPr)3 and H–PzCF3. The fluorinated pyrazole ligand H–PzCF3 was chosen because the formation of heteroleptic complexes can also be monitored by 19F NMR spectroscopy. A C2D2Cl4 solution containing Au3(PzPr)3 (4 mM) and H–PzCF3 (12 mM) was heated to 60 °C. After 24 h, the thermal equilibrium was reached. NMR spectroscopy indicated the presence of three main complexes: Au3(PzPr)3 (40%), Au3(PzPr)2(PzCF3) (51%), and Au3(PzPr)(PzCF3)2 (9%) (Scheme 1). The homotrimer Au3(PzCF3)3 was not detected in significant amounts (<3%). As expected, a similar equilibrium distribution was observed for the ‘inverse’ reaction between Au3(PzCF3)3 and H–PzPr (see the SI, Fig. S33 and S34). From these experiments, one can conclude that: (a) ligand exchange is possible but slow, and (b) excess ligand does not compromise the stability of Au3(Pz)3 trimers to a significant extent, and (c) PzPr-containing Au trimers are more stable than PzCF3-containing trimers.
![]() | ||
Scheme 1 Reaction between Au3(PzPr)3 (4 mM) and H–PzCF3 (12 mM) in either C2D2Cl4 or toluene-d8 and the product distribution after equilibration. |
A similar ligand exchange reaction was performed using toluene-d8 instead of C2D2Cl4 as the solvent. The reaction was slower, and 3 days were required to reach the equilibrium. The spectra of the equilibrated mixture revealed the presence of two trinuclear complexes: Au3(PzPr)3 (61%), Au3(PzPr)2(PzCF3) (39%). These results show that the nature of the solvent can influence the equilibrium distribution substantially. A possible factor for the altered distribution is a solvent effect on the relative stability of H–PzPr and H–PzCF3. Pyrazoles are known to self-aggregate via hydrogen bonding,23 and the stability of these aggregates is expected to depend on the nature of the solvent and on the substituents on the pyrazole.
An unexpected finding was the sigmoidal rate profile for the ligand exchange reactions in both C2D2Cl4 and in toluene-d8 (Fig. 3, traces in blue). To further investigate this unusual kinetic behavior, we performed a kinetic study in C2D2Cl4 at room temperature.
1H NMR spectroscopic monitoring of the reaction at RT confirmed the presence of a pronounced induction period (see the SI, Fig. S39). For the first 200 h, the concentration of H–PzCF3 diminished by only 4%. Subsequently, a rapid decrease of the H–PzCF3 concentration was observed until its concentration approached the equilibrium distribution.
Sigmoidal rate profiles are a typical characteristic of autocatalytic reactions.24 The presence of an autocatalytic system can be corroborated by enhanced reaction kinetics in the presence of externally added products. The reaction between Au3(PzPr)3 and H–PzCF3 produces two new Au trimers along with H–PzPr. The latter appeared to be the most likely candidate for a potential catalyst. Therefore, we repeated the reaction between Au3(PzPr)3 and H–PzCF3 in the presence of different amounts of H–PzPr (16 or 27 mol% with respect to the Au-trimer), both in toluene-d8 and in C2D2Cl4. Significant rate enhancements were observed in the presence of H–PzPr (Fig. 3). As expected, the presence of H–PzPr also shifted the equilibrium distribution of the complexes, with more remaining H–PzCF3 in the presence of H–PzPr.
The results described above show that H–PzPr can act as a catalyst, but not H–PzCF3. The latter pyrazole ligand is significantly less basic than H–PzPr: the calculated pKa of protonated (H–PzCF3–H)+ is 0.17, whereas a value of 3.70 was calculated for (H–PzPr–H)+.25 On the other hand, the neutral ligand H–PzCF3 (pKa = 9.57) is more acidic than H–PzPr (pKa = 15.08). Assuming a correlation between ligand basicity and ligand donor strength, H–PzCF3 is expected to be a worse ligand than H–PzPr, both in the neutral and in the deprotonated form.
A plausible explanation for the autocatalytic behavior in the reaction between Au3(PzPr)3 and H–PzCF3 is that ligand exchange proceeds via an associative mechanism, with initial formation of an adduct between Au3(PzPr)3 and H–PzPr. The adduct then undergoes fast ligand exchange with H–PzCF3 to give Au3(PzPr)2(PzCF3) and the regenerated catalysts H–PzPr.
Attempts to detect an adduct between Au3(PzPr)3 and H–PzPr spectroscopically were not successful. The 1H NMR spectrum of an equimolar mixture of Au3(PzPr)3 (4 mM) and H–PzPr (12 mM) in toluene-d8 after equilibration for 24 h showed neither signals of a new complex nor significant shifts when compared to the spectra of the individual compounds (see the SI, Fig. S32). The proposed intermediate is therefore not formed in larger amounts.
Analogous exchange reactions were then performed with Au3(PzPh)3 and three equivalents of H–PzCF3 (C2D2Cl4, RT, Scheme 2). The 19F NMR spectrum of the equilibrated reaction mixture indicated the formation of two main new complexes, which can be assigned as the heteroleptic complexes Au3(PzPh)2(PzCF3) (45%) and Au3(PzPh)(PzCF3)2 (26%). As in the case of Au3(PzPr)3, the rate profile displayed an induction period. However, the reaction was overall faster, with equilibrium being established after ∼10 hours. The faster ligand exchange for reactions involving Au3(PzPh)3 instead of Au3(PzPr)3 likely reflects a lower relative stability of the former trimer. A pronounced rate enhancement was observed when the reaction was performed in the presence of 30 mol% of H–PzPh, indicating again an autocatalytic behavior (Fig. 4).
![]() | ||
Scheme 2 Reaction between Au3(PzPh)3 (4 mM) and H–PzCF3 (12 mM) and the product distribution after equilibration. |
Next, we studied ligand exchange reactions between two trimeric complexes, Au3(PzPr)3 and Au3(PzPh)3. Ligand scrambling was expected to give four complexes, the heteroleptic complexes Au3(PzPh)2(PzPr) and Au3(PzPh)(PzPr)2 along with the homotrimers Au3(PzPr)3 and Au3(PzPh)3 (Scheme 3). For monitoring the reaction, we used in situ1H NMR spectroscopy. The three complexes containing PzPh ligands show well-resolved signals for the aromatic C–H atoms of the heterocycle at 6.98–7.04 ppm.
As expected, ligand scrambling in C2D2Cl4 at RT was very slow. After 208 h, only 6.5% of Au3(PzPh)3 had converted into new complexes (Scheme 3 and SI, Fig. S54). When 30 mol% of the ‘free’ pyrazole ligand H–PzPh was added to the reaction mixture, a pronounced rate enhancement was observed (32% conversion). A similar rate enhancement was noted for reactions in the presence of H–PzPr (33% conversion). An inspection of the 1H NMR spectra showed that H–PzPr got incorporated into the Au trimers, with liberation of H–PzPh (see the SI, Fig. S50). The latter pyrazole then takes over the role as the catalyst. Attempts to use other N-donors such as pyridine or 1-isopropylimidazole as catalysts led to partial decomposition of the gold trimers and to the formation of Au nanoparticles.
The targeted synthesis of Au3(Pz)2(Pz′) complexes from the corresponding precursors was found to be difficult because mixtures of complexes were typically obtained. An exception was the heteroleptic trimer Au3(PzPh)2(PzCF3), which could be isolated on a preparative scale in low yield (20%) by combining AuCl(SMe2) with a mixture of H–PzPh and H–PzCF3 in methanol in the presence of NEt3 (for details, see the SI, section 2.2). Crystals of heteroleptic complexes containing PzMe were obtained by combining toluene solutions of the homoleptic complexes Au3(PzPh)3 or Au3(PzCF3)3 with a toluene solution of H–PzMe.
The molecular structures of Au3(PzPh)2(PzCF3), Au3(PzMe)2(PzPh), Au3(PzMe)2(PzCF3), and Au3(PzCF3)2(PzMe), as determined by single-crystal XRD, are depicted in Fig. 5. Key structural parameters for the four complexes are summarized in the SI, Tables S1–S4.
The presence of two different ligands does not result in a desymmetrization of the central nine-membered (Au–N–N)3 ring system. The Au(I) centers in the four complexes all show a nearly perfect linear coordination geometry (αN–Au–N: 174.98–179.88°), and the Au–N bond distances are all within a narrow range (1.964–2.040 Å).
All four heteroleptic Au3(Pz)2(Pz′) complexes display intermolecular Au⋯Au contacts in the solid state (Fig. 5, graphics on the right side). Crystalline Au3(PzPh)2(PzCF3) features isolated pairs of co-planar trimers with Au⋯Au distances of 3.3065(7) and 3.3415(7) Å. The central (Au–N–N)3 rings are arranged in an eclipsed fashion,26 with the PzCF3 ligand of one trimer facing a PzPh ligand of an adjacent trimer. The close Au⋯Au contacts observed for Au3(PzPh)2(PzCF3) are in contrast to the structure of the homoleptic complex Au3(PzPh)3, which does not show intermolecular Au⋯Au interactions in the solid state.27
The three other heteroleptic complexes Au3(PzMe)2(PzPh), Au3(PzMe)2(PzCF3), and Au3(PzCF3)2(PzMe) all show a columnar arrangement of co-planar Au trimers. The individual complexes are linked via two close Au⋯Au contacts.
Our result should be considered when using Au3(Pz)3 complexes for the construction of supramolecular architectures. On one hand, error correction processes may be compromised by the inert character of Au3(Pz)3 complexes. On the other hand, it would be possible to facilitate thermal equilibration processes by using pyrazoles as catalysts.
CCDC 2453718, 2453719, 2453717 and 2440472 contains the supplementary crystallographic data for this paper.28a–d
This journal is © The Royal Society of Chemistry 2025 |