Coordination of a Pyrazole Functionalized Acetylacetone to the Coinage Metal Cations: An Unexpected Packing Similarity and a Trinuclear Cu^II/Au^I Complex

Abstract

The heteroditopic molecule HacacMePz combines a Pearson hard acetylacetone donor site with a softer trimethylpyrazole and shows site selectivity towards the coinage metal cations. The coordination of the N donor function was achieved towards Ag^I and Au^I, leading to the salt [Ag(HacacMePz)₂]PF₆ (1) and the neutral complex [AuCl(HacacMePz)] (2). In either case, linear coordination about the coinage metal cation is observed. Interestingly, both complexes crystallize in space group $Pbca$ with similar cell parameters. The two solids do not qualify as isostructural, albeit being closely related in real and reciprocal space. To probe the ligand’s ability for the envisaged synthesis of bimetallic coordination polymers, the mixed-metal Cu^II/Au^I complex [Cu(acacMePzAuCl)₂] (3) was obtained. In this mixed-metal oligomer, the central Cu^II cation adopts a square planar coordination environment with two O,O^′-coordinated acacMePz⁻ ligands, whose softer N donor sites are saturated with a AuCl moiety.

1. Introduction

In recent years, the interest in organic–inorganic hybrid materials has flourished. Coordination polymers (CPs) are a subclass of the aforementioned, in which metal cations are bridged by ligands periodically in at least one dimension [1]. CPs are promising candidates for the increasing demand for tailored materials: the seemingly inexhaustible combinations of organic ligands with different metal species offer a huge playground for the modern physicist, chemist and material scientist. Gas sorption [2,3,4] and separation [5,6], catalysis [7,8,9] or luminescence [10,11,12] are only a few of the potentially useful properties of CPs. Often, these desirable characteristics directly arise from the metal species used, and thus it could be advantageous to incorporate more than one metal cation to combine their properties in a joint material. Our group utilizes heteroditopic ligands to enable a selectivity in this otherwise statistical [13] construction process (Figure 1) [14,15,16,17].

While the two binding sites of 4,4^′-bipyridine are equivalent and do not exhibit selectivity towards specific metals, each of the inequivalent donors of the functionalized acetylacetones in Figure 1 preferably binds to specific cations. This selectivity is based on different Pearson hardnesses [20,21] of the ligand’s donor sites and, vice versa, of the metal cations used. Recently, we reported the ditopic coordination chemistry of the pyrazole functionalized acetylacetone (HacacMePz, Figure 1 center) for the selective construction of bimetallic CPs. After deprotonation, the diketonate moiety acted as a chelating ligand towards small and notoriously hard Fe^III cations, whereas the N donor site selectively attached to larger cations carrying a lower charge, e.g., Hg^II or Ag^I [17]. In this study, we address cations with less pronounced differences in their Pearson hardness and focus on a single group in the periodic table, namely, the coinage metals. The elements in group 11 form continuous solid solutions [22]. We report here that the cations of the coinage metals selectively form coordination compounds with HacacMePz. Ag^I and Au^I afford N-coordinated building blocks suitable for further crosslinking, and ditopic coordination of the ligand leads to a bimetallic trinuclear Cu^II/Au^I complex.

2. Materials and Methods

All chemicals were purchased from common vendors and used without further purification. 3-(1,3,5-Trimethyl-1H-pyrazol-4-yl)acetylacetone [17] and [AuCl(THT)] [23] were synthesized using procedures published before. ATR FT IR spectra were obtained with a Shimadzu IRSpirit with a QATR-S ATR unit equipped with a diamond prism. The magnetic resonance spectrum was measured using a Bruker Avance II UltrashieldT11 plus 400 instrument ( 400 $\mathrm{M}$$\mathrm{Hz}$, referenced to SiMe₄).

2.1. Synthesis and Crystallization of [Ag(HacacMePz)₂]PF₆, 1

HacacMePz ($41.6$ $\mathrm{m}$$\mathrm{g}$, $0.2$ $\mathrm{mmol}$, $2.0$ $\mathrm{eq}.$) and AgPF₆ ($25.3$ $\mathrm{mg}$, $0.1$ $\mathrm{mmol}$, $1.0$ $\mathrm{eq}.$) were separately dissolved in ethyl acetate (2 $\mathrm{m}$$\mathrm{L}$ each). The two clear solutions were combined. The solution was stored at 4 °C, and colorless block-shaped crystals suitable for SCXRD formed overnight. Yield: 51.3 mg (77%). IR (ATR, $\tilde{\nu }$/cm⁻¹): 2965 (vw), 2925 (vw), 2359 (w), 2344 (w), 1601 (m), 1560 (m), 1488 (m), 1421 (m), 1389 (s), 1319 (m), 1301 (m), 1276 (m), 1136 (m), 994 (m), 914 (m), 843 (vs), 818 (vs), 715 (m), 668 (w), 637 (vw), 574 (w), 555 (vs), 495 (m), 475 (w). CHN: anal. calcd. for C₂₂H₃₂AgF₆N₄O₄P: C 39.5%, H 4.8%, N 8.4%; found: C 39.7%, H 4.8%, N 8.6%. PXRD performed on the bulk material allowed us to recognize the presence of a minor crystalline impurity (Figure S1).

2.2. Synthesis and Crystallization of [AuCl(HacacMePz)], 2

HacacMePz ($11.2$ $\mathrm{m}$$\mathrm{g}$, $0.06$ $\mathrm{mmol}$, $1.2$ $\mathrm{eq}.$) and [AuCl(THT)] ($16.0$ $\mathrm{m}$$\mathrm{g}$, $0.05$ $\mathrm{mmol}$, $1.0$ $\mathrm{eq}.$) were dissolved in deuterated chloroform (1 $\mathrm{m}$$\mathrm{L}$). Colorless block-shaped crystals suitable for SCXRD were obtained from CDCl₃ at 4 °C by slow gas diffusion of n-pentane into the solution. Yield: 16.2 mg (74%). ¹H NMR (CDCl₃, 400 MHz): $\delta$ 16.86 (s, 1H), 3.78 (s, 3H), 2.13 (s, 6H), 1.85 (s, 6H) ppm. IR (ATR, $\tilde{\nu }$/cm⁻¹): 2962 (w), 2920 (w), 1597 (s), 1560 (s), 1485 (s), 1414 (vs), 1389 (vs), 1365 (s), 1275 (s), 1220 (s), 992 (s), 917 (s), 714 (m), 670 (m), 535 (m), 474 (m). CHN: anal. calcd. for C₁₁H₁₆AuClN₂O₂: C 30.0%, H 3.7%, N 6.4%; found: C 33.0%, H 4.0%, N 7.0%. Phase purity was confirmed via PXRD (Figure S2).

2.3. Synthesis and Crystallization of [Cu(acacMePzAuCl)₂], 3

HacacMePz ($5.2$ $\mathrm{m}$$\mathrm{g}$, $0.025$ $\mathrm{mmol}$, $2.0$ $\mathrm{eq}.$), Cu(OAc)₂·H₂O ($2.7$ $\mathrm{m}$$\mathrm{g}$, $0.0125$ $\mathrm{mmol}$, $1.0$ $\mathrm{eq}.$) and [AuCl(THT)] ($7.7$ $\mathrm{m}$$\mathrm{g}$, $0.025$ $\mathrm{mmol}$, $2.0$ $\mathrm{eq}.$) were separately dissolved in acetonitrile (2 mL) each. The HacacMePz solution was combined with the Cu(OAc)₂ solution; then, the [AuCl(THT)] solution was added. The solution was stored at −25 °C. Eventually, after 11 day, light blue plate-shaped crystals suitable for SCXRD were obtained and isolated. Yield: $6.8$ $\mathrm{m}$$\mathrm{g}$ (56%). IR (ATR, $\tilde{\nu }$/cm⁻¹): 2998 (vw), 2930 (w), 2359 (w), 2248 (w), 1585 (vs), 1549 (m), 1416 (m), 1368 (vs), 1281 (s), 1036 (vs), 974 (m), 927 (m), 851 (w), 723 (m), 714 (w), 685 (m), 633 (m), 464 (vs). Microanalysis of the dried bulk: CHN: anal. calcd. for C₂₂H₃₀AuCl₂CuN₄O₄: C 28.0%, H 3.2%, N 5.9%; found: C 26.0%, H 3.6%, N 5.4%. PXRD confirms that the majority of the bulk material corresponds to the phase identified by SCXRD; however, a few reflections that cannot be attributed to 3 are visible in the diffractogram (Figure S2). Measurement under acetonitrile and perfluorinated oil was intended to reduce desolvation but did not improve the diffractogram.

2.4. Structure Determinations

For 1, intensity data were collected on a STOE STADIVARI goniometer (STOE & Cie GmbH, Darmstadt, Germany; location of the instrument: Institute of Inorganic Chemistry, RWTH Aachen University) equipped with a DECTRIS Pilatus 200K area detector, GeniX 3D HF Mo microsource ($\lambda$ = $0.71073$ Å, multilayer optics). Temperature was maintained with an Oxford Cryostream 800 instrument (Oxfordshire, UK). Data were collected and integrated with the X-Area program package [24]. Scaling and multi-scan absorption correction were applied with LANA [25].

For 2 and 3, intensity data were collected on a Bruker D8 goniometer (Bruker AXS, Madison, USA; location of the instrument: Institute of Inorganic Chemistry, RWTH Aachen University) equipped with an APEX CCD area detector, Incoatec microsource with Mo-K_α radiation ($\lambda$ = $0.71073$ Å, multilayer optics). Temperature was maintained with an Oxford Cryostream 700 instrument (Oxfordshire, UK). Data were collected with SMART [26] and were integrated with SAINT+ [27]. Scaling and multi-scan absorption correction were applied with SADABS [28].

The structures were solved by intrinsic phasing with SHELXT-2014/5 [29]; full-matrix least squares refinements against $F^2$ were carried out with SHELXL-2019/2 [30]. Non-H atoms were refined anisotropically. If possible, protic H atoms were located from difference Fourier synthesis maps, and their positional coordinates were refined freely. Other H atoms were placed in idealized positions and refined as riding. Isotropic displacement parameters for H atoms were constrained to multiples of their parent atoms, namely, $U_{\mathrm{iso}}$(H) = $1.5U_{\mathrm{eq}}$(C,N,O) for methyl/protic hydrogen and $U_{\mathrm{iso}}$(H) = $1.2U_{\mathrm{eq}}$(C) for other H atoms.

2.5. Powder Diffraction and Rietveld Refinement

Powder diffractograms were recorded on flat samples at room temperature using a STOE STADI-P diffractometer (STOE & Cie GmbH, Darmstadt, Germany; location of the instrument: Institute of Inorganic Chemistry, RWTH Aachen University) with Guinier geometry (Cu-K_α1, $\lambda$ = $1.54059$ Å, Johann germanium monochromator, STOE image plate detector IP-PSD, 0.005° step width in $2\theta$). Rietveld refinements for 1 and 2 were carried out with the program GSAS-II [31] using a pseudo-Voigt profile function. Residual values for 1: w$R_{\mathrm{P}}$ = 0.2680, w$R_{\mathrm{B}}$ = 0.3980 for 496 Bragg reflections (3835 data points) and 5 refined parameters. Residual values for 2: w$R_{\mathrm{P}}$ = 0.1337, w$R_{\mathrm{B}}$ = 0.1764 for 465 Bragg reflections (3835 data points) and 12 refined parameters.

3. Results

A classic candidate for the selective coordination of the rather Pearson soft pyrazole N site of HacacMePz is Ag^I. In a preceding article, we reported the AgNO₃ complex [Ag(NO₃)(HacacMePz)₂], in which the nitrate anion is still coordinating and prevents a strictly linear arrangement of the pyrazole moieties [17]. Here, we disclose the complex of HacacMePz with AgPF₆. The PF₆⁻ anion is less coordinating and more likely to form a bis-pyrazolyl silver(I) cationic complex than coordinating anions such as NO₃⁻ [32]. The compound [Ag(HacacMePz)₂]PF₆ (1) crystallizes in the orthorhombic space group $Pbca$ with $Z=4$ (Figure 2).

The cationic [AgPz₂]⁺ moieties are linear for reasons of symmetry, with the Ag^I cation occupying the inversion center on Wyckoff position $4b$. The PF₆⁻ anion is located on the other inversion center on Wyckoff position $4a$. The closest contact between the PF₆⁻ and the adjacent bis-pyrazolyl complex amounts to $2.9444\left(15\right)$ Å. Interestingly, no Ag^I complex of unsubstituted trimethylpyrazole has been reported to this date.

The reaction of HacacMePz with the precursor [AuCl(THT)] leads to the linear complex [AuCl(HacacMePz)] (2), which crystallizes in the orthorhombic space group $Pbca$ with $Z=8$ (Figure 3).

As expected, the Cl–Au–N angle is close to linear with 178.15(12)°. The enol hydrogen is clearly localized, with a C–O single bond ( $1.303\left(6\right)$ Å) and a C=O double bond ( $1.265\left(6\right)$ Å). While there are many reports of aurophilic interactions in the crystal structures of AuCl complexes [34,35,36,37,38], 2 does not exhibit this behavior. The steric bulk of the methyl substituents might impede a closer approach; alternatively, a weak aurophilic contact might be overcome by a more suitable packing.

Copper(II) acetate is sufficiently basic to deprotonate the acetylacetone site of the Au^I building block 2. If concentration and temperature are suitably chosen, single crystals of the compound [Cu(acacMePzAuCl)₂]·2MeCN (3·2MeCN) are obtained (Figure 4). It crystallizes as a discrete complex in the monoclinic space group $P{2}_1/c$ with $Z=2$.

The central Cu1 ion is located on the inversion center with Wyckoff letter $2a$ and adopts square planar coordination with two chelating acetylacetonate moieties of HacacMePz. Both N donor sites of the acacMePz⁻ ligand coordinate a AuCl moiety in a linear fashion, with an angle of 177.12(17)°. Interestingly, the $\omega$ angle deviates from orthogonality and amounts to the rather unfavorable value of 71.7(4)° [17]. Two symmetry-equivalent acetonitrile molecules come close to the central Cu1 ion, albeit not in the regular linear N donor fashion. The N and C atom of each nitrile group are located at a distance of about $3.7$ Å with respect to Cu1 to engage in a very long side-on coordination. This is larger than the sum of their van der Waals radii, but has an effect on the C≡N stretching vibrations observed via IR spectroscopy. The latter undergoes a minor red shift from 2253 cm⁻¹ to 2248 cm⁻¹, indicating a slightly weakened C≡N bond as a result of the weak interaction with the Cu^II cation. No further unexpectedly short intermolecular contacts were encountered.

4. Discussion

4.1. Surprising Packing Similarity of 1 and 2

A comparison of the powder diffractograms of the gold(I) complex 2 with that for the AgPF₆ complex 1 reveals surprising similarities (Figure 5).

Both structures crystallize in the same space group $Pbca$ with similar lattice parameters, although with a and b being anti-parallel to each other in the two structures. They also exhibit a different number of molecules per unit cell: the gold(I) complex crystallizes with $Z=8$ and the molecule in general position, whereas the silver(I) complex exhibits $Z=4$ with the Ag^I ion on a center of inversion. The Au^I ion in 2 is shifted in b direction for about b/4. A close inspection of the two structures shows they can indeed be superimposed to show a similar packing motif if the unit cells are slightly shifted (Figure 6).

In the resulting overlay, the positions of Ag^I ions and PF₆⁻ anions formally coincide with Au^I ions in the alternative structure. Do our compounds 1 and 2 qualify as isostructural? No, because it would require both similar stoichiometries of the two crystalline substances and equivalent occupation of special positions. But what about isomorphism? According to the IUCr dictionary [39,40], isomorphous structures need to exhibit three characteristics:

Two crystals are said to be isomorphous if (a) both have the same space group and unit-cell dimensions and (b) the types and the positions of atoms in both are the same except for a replacement of one or more atoms in one structure with different types of atoms in the other (diadochy), such as heavy atoms, or the presence of one or more additional atoms in one of them (isomorphous addition). Isomorphous crystals can form solid solutions.

The first condition is obviously fulfilled, but the unit cell shift of b/4 required for efficient overlay maps a special position in 1 on a general position in 2. Such relationships are difficult if not impossible to classify. A quick search in the CSD [41] reveals that there are 67 other orthorhombic structures with similar lattice parameters. Quite obviously, this leads to similarities in the reciprocal space and similar line positions in the diffraction patterns. If heavy atoms occupy special or pseudo-special positions in these structures, the contribution of light atoms to reflection intensities may be minor, and the overall X-ray powder diffractograms may appear to be similar. As an example the simulated powder patterns of 1 and a structurally unrelated Pd^II complex (refcode DOXMEL [42]) are compared in Figure 7. DOXMEL has been chosen quite arbitrarily for comparison, with similar lattice parameters as 1 and the presence of a heavy atom as the only criterion.

Despite their apparent similarity in reciprocal space, the real space match between 1 and DOXMEL is poor: the inversion center corresponding to Wyckoff position $4a$ is occupied by the Pd^II centers in DOXMEL and in 1 by the PF₆⁻ anions, and the remaining light atoms cannot be superimposed at all. With respect to our much closer related structures 1 and 2, we therefore stick to the admittedly vague expression “packing similarity”.

4.2. Structural Features of 3

The coordination of the Au^I cations in 2 and 3 can be compared with chemically related species published earlier, namely, the gold complexes of trimethylpyrazol itself (XACQUQ) and an alternative substituted and potentially ditopic trimethylpyrazol (VAVMIS) (

Table 1. Comparison of the coordination environment about the Au^I in 2, 3 and two selected literature structures.

Compound	Au–Cl/Å	Au–N/Å	Cl–Au–N/°	$\mathit{\omega }$/°
2	2.2446 (14)	2.011 (4)	178.15 (12)	79.0 (3)
3	2.2458 (19)	2.010 (6)	177.12 (17)	71.7 (4)
XACQUQ [43]	2.253 (2)	2.011 (7)	177.3 (2)
VAVMIS [44]	2.2441 (17)	2.020 (4)	179.08 (11)

).

The compounds in this comparison are essentially linear with similar metal–ligand distances. Aurophilic interactions are encountered in none of the above-mentioned AuCl complexes. As stated before, this might be attributed to the steric bulk of the methyl groups. The overlay plots of 3 with 2 and XACQUQ [43] reveal how similar the conformations in these complexes are (Figure 8).

5. Conclusions

The packing similarities between the chemically quite different cationic Ag^I bis–ligand complex 1 and the neutral Au^I complex 2 are surprising yet difficult to classify. They are probably caused by the coincidence of several facts:

• Packing in each structure is dominated by the necessity to accommodate eight neutral N-coordinated HacacMePz ligands per unit cell; in either case, this involves the interaction of methyl H atoms with the pyrazole -system;
• Both metal cations prefer a linear coordination;
• The distance between the non-coordinating counter-anion PF₆⁻ and the Ag^I cation is compatible with the Au⋯Au separation between neighboring molecules.

We are not aware of an established scientific term for such similarities.

With respect to CP design, the gold(I) complex 2 may be used as metalloligand towards a harder Cu^II cation in predictable geometry.

The resulting mixed-metal complex 3 qualifies as an oligonuclear compound rather than a coordination polymer. It proves, however, that our ligand is sufficiently selective to distinguish even chemically related cations from the same group based on their different Pearson hardness. Therefore, 3 paves the way to future extended and well-ordered mixed-metal structures. The next step towards a “real” polymer will imply a substitution of the terminal chlorido ligands attached to Au^I.