Redox Chemistry of Pt(II) Complex with Non-Innocent NHC Bis(phenolate) Pincer Ligand: Electrochemical, Spectroscopic and Computational Aspects

1. Introduction

Transition-metal complexes with redox-active ligands have been intensively studied over the past few decades due to their unique properties and intriguing chemical behavior [1,2,3,4,5,6,7,8]. They wildly applied in various fields of coordination chemistry including catalysis, organic synthesis and material science [1,9,10,11,12,13,14,15,16,17,18]. Recent attempts in controlling catalytic processes employing redox non-innocent ligands are mostly inspired by the metal-radical motifs in active sites of many metalloenzymes [5], such as galactose oxidase, which contains one copper ion and oxidizes a primary alcohol to the aldehyde in the presence of dioxygen [19,20]. The formation of a Cu(II)−phenoxyl radical species was proposed to be the key step in this transformation [21].

Indeed, electron-rich sterically hindered phenolates are among the most common non-innocent ligands that can assist in catalytic transformations by storing (in form of phenoxyl) and delivering charge during catalytic transformations [22,23,24,25,26,27]. However, bidentate ligand scaffold utilized in many catalytic reactions involving such phenolate ligands (or their aminophenol derivatives) often leads to unfavorable isomerization processes of the ligand field, precluding further transformations of a substrate [28,29]. In order to overcome the limitation the redox-active phenolates were merged with N-heterocyclic carbenes (NHCs), forming a pincer-type tridentate ligands [30,31]. NHCs are well-known as a powerful class of ligands in organometallic chemistry and homogeneous catalysis due to their unique steric and electronic properties. Moreover, the strongly σ-donating carbene makes the ligand less labile in its oxidized forms, while the strong trans effect provide hemilability of the auxiliary ligand, generating a vacant coordination site at the metal center which is a prerequisite for effective coordination and transformation of substrate molecules in the homogeneous catalysis conditions [32,33,34]. In this regard, the previously reported diphenolate imidazolyl and benzimidazolyl carbenes appeared promising for constructing transition metal complexes with the coordinated pro-radical ligand [35,36,37,38]. The oxidation of the above-mentioned species allowed to form an electronic structure in which a delocalized electronic hole is more likely located on the ligand. However, the precise determination of the metal oxidation state in transition-metal complexes is not always possible. In reality, metal ion and ligand are both affected by the oxidation and the resulting spin density is shared over both, allowing a multi-configurational state [9,25]. Therefore, a careful examination of metal complexes by various spectroscopic methods combined with computational studies is required for the better description of their electronic structure, which is essential for the design of catalytic cycles with participation of redox non-innocent ligands [39,40].

Over the last years, our research group has focused on the design, synthesis and catalytic application of pincer-type ligands and their complexes of group 10 metals [41,42,43,44,45,46,47]. The intrinsic diamagnetism of this metal ions in d⁸ configuration allows for fine assignment of the electronic structure of the oxidized species. At the same time, the natural content of spin-1/2 isotope for platinum (¹⁹⁵Pt, 33.775 %) gives the opportunity for careful examination of the electronic structure by EPR spectroscopy. Thus, herein we report on the synthesis of Pt(II) complex bearing tridentate diphenolate NHC ligand. The chemical and electrochemical generation and isolation of one-electron-oxidized species are also performed. Characterization of their electronic structures by combined UV-Vis/NIR- and EPR-spectroscopy, X-ray diffraction and DFT studies suggests the redox non-innocence of the ligand. The obtained results allowed to design an electrocatalytic process for 2-(methylamino)ethanol oxidation with participation of Pt(II)-phenoxyl radical complex.

2. Results and Discussion

The diphenolate NHC precursor L^H3Cl was prepared according to our previous report [48]. The synthesis of its platinum complex was performed using modified procedure suggested by Mauro, Dagorne, Bellemin-Laponnaz and co-authors for the related diphenolate imidazolyl and benzimidazolyl carbene complexes [30]. However, it should be noted, that in our case providing the synthesis at 100 °C (as reported by the authors) lead unfavorable deposition of black precipitate (more likely metallic platinum). Therefore, the tetrahydropyrimidin-1-ium based pro-ligand was treated with Pt(COD)Cl₂ and an excess of potassium carbonate in pyridine at 50 °C for 20 h (Scheme 1). The isolation of the complex by a filtration of the CH₂Cl₂ solution of crude material through a silica gel plug with further evaporation of the solvent afforded the platinum complex Pt(L)Py in 81% yield as an orange solid. The formation of the corresponding diphenolate NHC complex was confirmed by the presence of a carbene signal at δ_C 159.2 ppm in the ¹³C{¹H} NMR spectrum, while ¹H NMR spectroscopy confirmed the absence of any residual tetrahydropyrimidinium and phenol moieties. The proposed formulation was also confirmed by MALDI mass spectrometry and elemental analysis.

The UV-Vis spectrum of Pt(L)Py in CH₂Cl₂ (Figure 1) shows a weak absorption band at ca. 350 nm (ε = 6000 M^–1⋅cm^–1). According to time-dependent DFT calculations (Figure S4 and Table S1), the observed absorption originates from the electronic transition composed of a phenolate-to-NHC intraligand excitation mixed with a metal-to-metal transition (HOMO → LUMO+2). In should also be noted that no NIR transitions were observed for Pt(L)Py complex (see Supplementary Materials for details).

The electrochemical properties of Pt(L)Py were investigated by cyclic voltammetry (CV) and differential pulse voltammetry (DPV) in CH₂Cl₂ containing 0.1 M tetra-n-butylammonium tetrafluoroborate as supporting electrolyte. CV curve of complex Pt(L)Py exhibits two reversible oxidation peaks at E_1/2¹ = 0.25 V and E_1/2² = 0.80 V vs the ferrocenium/ferrocene external standard (Figure 2). The number of electrons involved was confirmed by the analysis of the shape of the DPV curve. Thus, the values of 1.060 and 1.004 were obtained for the first and second oxidation, respectively. The two oxidation waves can be attributed to ligand-based oxidation processes leading to a mono- and presumably a bis-phenoxyl radical species [49,50]. No reduction peak observed when scanning towards the cathodic values of potentials in CH₂Cl₂. It is interesting to note, that the reversibility of the second oxidation peak disappears when polar coordinating solvent is used (DMF, CH₃CN). At the same time, DMF enabled determination of the electroreduction process for the complex at –2.80 V, which was ascribed to the Pt^II → Pt⁰ conversion (see Supplementary Materials for details).

The relatively large separation between the halfwave potentials for both processes (ΔE_1/2 = 0.55 V) indicates that oxidation of one phenolate moiety in the complex has a significant influence on the second one. Thus, in order to evaluate the nature of the species formed in the electrooxidation process, the “chemical” oxidation of Pt(L)Py with AgBF₄ as oxidizing agent (E_1/2 = 0.65 V vs Fc⁺/Fc [51]) in dry CH₂Cl₂ at room temperature under an inert atmosphere of nitrogen was performed (Scheme 2). A color change of the solution from orange to dark green, as well as the precipitation of Ag⁰ at the bottom of the flask verified the successful oxidation of Pt(L)Py. The product [Pt(L)Py][BF₄] was isolated in 77 % yield as an air-stable dark green solid. The proposed composition was confirmed by MALDI mass spectrometry by the characteristic ions with m/z 792.4 (positive mode) and m/z 87.1 (negative mode) corresponding to the [Pt(L)Py]⁺ and [BF₄]ˉ species respectively. The elemental analysis is also in agreement with the expected product.

X-ray-quality dark green needle-like crystals were grown by diffusion of pentane into a saturated solution of the product in dichloromethane at ca. 0 °C, and X-ray diffraction studies allowed the structure determination of [Pt(L)Py][BF₄]. The complex crystallizes as CH₂Cl₂ adduct in the $P\overline{1}$ (2) triclinic space group, with two molecules per unit cell. As depicted in Figure 3, a distorted-square-planar environment around the metal is observed. Thus, the pyridine ligand occupies the coordination site trans to the pincer carbene carbon, while two oxygen atoms complete the metal coordination sphere standing trans to each other. The Pt−C_carbene distance (1.954(2) Å) lies within the range of related structures [30,52,53,54,55,56]. The {OCO}Pt chelate in the complex is significantly distorted from planarity (O(1)−N(1)−N(2)−O(2) = –14.63(8)°). The Pt−N bond length (2.094(2) Å) is consistent with a trans influence of the carbene ligand (see ref [57] for example of [Pt(Py)₄]²⁺ complex, where the average Pt–N bond length is 2.024 Å). Selected bond distances and angles are listed in the figure description.

An EPR analysis of a solution of [Pt(L)Py][BF₄] in CH₂Cl₂ at 298 K shows a signal at g = 2.070 with signal coupling coming from the ¹⁹⁵Pt having nuclear spin of ¹/₂ and natural content of 33.775%. (Figure 4a). The g value is significantly larger than that expected for free phenoxyl radicals (typically 2.005) [58], but smaller than the average g values usually reported for paramagnetic platinum complexes. Variable-temperature EPR study (Figure 4b) revealed the g-anisotropy. Thus, the frozen solution EPR spectrum of [Pt(L)Py][BF₄] in CH₂Cl₂ displays a highly anisotropic S = ½ signal at g₁ = 1.943, g₂ = 2.123, and g₃ ≈ 2.149 (g_iso ≈ 2.071), that together with large g value suggests a considerable contribution of the metal d orbital to the SOMO [12,59]. Taken together, these data provide strong evidence that the distribution of the unpaired electron in [Pt(L)Py][BF₄] is not uniform, supporting the assumption that the radical is at least partially localized over the ligand.

The Mulliken spin-density distribution in [Pt(L)Py]⁺ calculated at the DFT-optimized structure (Figure 5) exhibits a predominant phenoxyl-radical character. Although the spin density is shared between the metal center and the ligand, its major contribution is delocalized within the ligand. It should also be noted that the nitrogen atoms of the carbene linker hold a small spin population, in contrast to the diphenolate imidazolyl based nickel analogue, which does not feature any significant spin density [35].

The UV-Vis and NIR spectra of [Pt(L)Py][BF₄] in CH₂Cl₂ are shown in Figure 6. In visible region the complex exhibits an absorption band at ca. 466 nm, which is red-shifted compared to its neutral analogue. In addition, [Pt(L)Py][BF₄] shows a remarkable intense NIR band at 1242 nm (ε = 7500 M^–1⋅cm^–1) with a narrow bandwidth at half height Δν _1/2 of 1712 cm^-1 (see Supplementary Materials for details).

According to the Marcus–Hush relationship:

$$\mathsf{\Delta }{\mathsf{\nu }}_{\mathrm{HTL}}=\sqrt{16\mathrm{l}\mathrm{n}2RT{\mathsf{\nu }}_{max}}$$

(1)

and using the ν _max = 8050 cm^-1 (1242 nm) the minimum bandwidth predicted in the high temperature limit (HTL) is 4308 cm^-1, which differs significantly from the experimental value. This observation is consistent with the inability of Marcus–Hush theory to predict IVCT band shape at the class II/III borderline [59]. Moreover, a sharp and intense NIR transition, where Δν _1/2 ≤ 2000 cm⁻¹, ε ≥ 5000 M⁻¹ cm⁻¹, is an indicator that the electron hole is fully delocalized over the ligand scaffold [60]. Thus, on the basis of the high intensity and narrow bandwidth, the lowest energy band can be assigned to the intraligand HOMO (donor) to SOMO (radical) electronic transition. In order to validate this hypothesis, absorption properties of [Pt(L)Py]⁺ were studied with time-dependent DFT (Figure S4 and Table S1). The calculated energy of the NIR band with a large oscillator strength is 1162 nm (f_osc = 0.23), which matches well with the experimental data (1242 nm). The principal excitation is indeed the β-HOMO → β-LUMO transition (97%), which is predominately π-π* intraligand in nature. The calculated electronic transition of a higher energy (λ_calc = 429 nm) corresponds to the experimental band observed at 466 nm. The principal excitations that contribute to the band are α-HOMO → α-LUMO (58%) and α-HOMO → α-LUMO+1 (32%) transitions.

Having all this results in hand, we decided to test the platinum complex Pt(L)Py as a mediator in the process of electrocatalytic oxidation of amines. It is well known, that the electrochemical oxidation of amines is an important alternative to conventional chemical approaches, providing critical pathways for the synthesis and modification of a wide range of chemically valuable compounds, including pharmaceuticals and agrochemicals [61]. Details on the mechanistic aspects of the electrochemical oxidation of aliphatic amines can be found elsewhere [61]. However, the direct electrooxidation processes often lead to a passivation of the working electrode surface by the formation of a polymer film, which can sharply decrease current efficiency [62]. Indirect electrooxidation using an electron transfer mediator is an effective technique to avoid this problem [63]. Besides the minimization of the electrode surface fouling effect, the application of redox mediators allows to decrease the overpotential needed for the direct oxidation of amines [61]. Pt(L)Py, which is prone to electrochemical oxidation to [Pt(L)Py]⁺, meets the requirements usually addressed to a redox mediator (or catalyst): it has a reversible redox couple, while both the oxidized and reduced forms are relatively stable in the solution. Thus, 2-(methylamino)ethanol (MEA) was chosen as a model substrate for the catalytic tests. Figure 7 shows the CV curves, obtained for 1 mM solution of Pt(L)Py with different concentrations of MEA (0.00 M, 0.12 M and 0.24M) in the range of 0.00 to 0.45 V vs Fc⁺/Fc in CH₂Cl₂ in the presence of n-Bu₄NBF₄ (0.1 M). It should be noted, that MEA is electrochemically inactive in this region (see black curve in Figure 7). Reaching a concentration of MEA up to 0.12 M (see blue curve in Figure 7) leads to a disappearance of reversibility of [Pt(L)Py]⁺_/Pt(L)Py couple with simultaneous increase in oxidation peak current. The peak oxidation current continues to increase until the substrate concentration reaches 0.24 M. At MEA concentrations greater than 0.24 M, saturation behavior is observed as the current becomes independent of the concentration of MEA. The efficiency of electrocatalytic MEA oxidation can be estimated via the ratio of the maximum catalytic current (i_cat) to the peak current (i_p) in the presence and absence of amine, respectively. Thus, the i_cat/i_p value of 1.9 was obtained for Pt(L)Py as a mediator of this process.

These results are indicative of the electrochemical catalysis of secondary amine (MEA) oxidation via the active Pt(II)-phenoxyl radical complex oxidant. Further experiments related to reactivity and catalytic activity of platinum complexes [Pt(L)Py]⁺ and Pt(L)Py as well as other group 10 metals with diphenolate NHC ligand L are currently under progress.

3. Materials and Methods

General Considerations

All reactions were performed using standard Schlenk procedures under a dry nitrogen atmosphere. Organic solvents (pentane, CH₂Cl₂, pyridine, DMF, CH₃CN, THF) were purified and degassed by standard procedures. CDCl₃ was degassed by three freeze–pump–thaw cycles and kept over 3Å molecular sieves before use. Ligand L^H3Cl was obtained by the previously described procedure [48]. All other reagents (AgBF₄ (Aldrich, 98%), n-Bu₄NBF₄ (Acros Organics, 98%), ferrocene (Alfa Aesar, 99%), Pt(COD)Cl₂ (Sigma-Aldrich, 99%)) were used without further purification.

NMR spectra were recorded on a high-resolution BRUKER AVANCE-400 (Germany—Switzerland) spectrometer at frequencies of 400.17 MHz (¹H) and 100.62 MHz (¹³С). ¹H and ¹³C{¹H} chemical shifts are reported in parts per million (ppm) downfield of tetramethylsilane and were calibrated against the residual resonance of the deuterated solvent. UV–Vis spectra were recorded on SPECORD 50 PLUS Analytik Jena (Spain) spectrophotometer in 10 mm closed quartz cuvette at 298 K. NIR absorption spectra were recorded by a Bruker Vertex 70 spectrometer (Germany) in 10 mm closed quartz cuvette at 298 K. The electron paramagnetic resonance (EPR) measurements were performed with a Bruker Elexsys E-500 spectrometer utilizing a 100-kHz field modulation and X-band microwaves (Germany). MALDI-TOF experiments were performed using an Ultraflex III TOF/TOF (Bruker Daltonics, Germany) mass spectrometer equipped with a Nd:YAG laser. The spectra were measured in both positive and negative ion linear modes. para-Nitroaniline (PNA) was used as a matrix. The FlexControl software (Bruker Daltonik GmbH, Version 3.0) was used for instrument control and data acquiring. Data processing was performed by FlexAnalysis software (Bruker Daltonik GmbH, Version 3.0). Elemental analysis was performed on a high-temperature Elementar vario MACRO cube (Germany) analyzer.

X-ray Structure Determination

X-ray crystallography data. X-ray diffraction analysis of the structure [Pt(L)Py][BF₄] was performed on a Bruker D8 QUEST (Germany) automatic three-circle diffractometer with a PHOTON III two-dimensional detector and an IμS DIAMOND microfocus X-ray tube (λ [Mo Kα] = 0.71073 Å) at cooling conditions. Data collection and processing of diffraction data were performed using the APEX3 software package. The structure was solved by the direct method using the SHELXT program [64] and refined by the full-matrix least squares method over F² using the SHELXL program [65]. All the calculations were performed in the WinGX software package [66], the calculation of the geometry of the molecule and the intermolecular interactions in the crystal was carried out using the PLATON program [67] and the drawings of the molecules were performed using the ORTEP−3 [66] and MERCURY [68] programs. The non-hydrogen atoms were refined in the anisotropic approximation. The hydrogen atoms were placed in geometrically calculated positions and included in the refinement in the “riding” model. One of the solvate molecules of methylene chloride is disordered by the center of symmetry over two positions. The crystallographic data of structure were deposited at the Cambridge Crystallographic Data Center and the registration number and the most important characteristics are given in Table 1.

Cyclic voltammetry (CV) and differential pulse voltammetry (DPV)

In CV and DPV experiments the concentration of the complex was 5∙mM in CH₂Cl₂ as a solvent with (n-Bu₄N)BF₄ (0.1 M) as a supporting electrolyte. The volume of the working solution was 5 mL. All experiments were performed under an inert nitrogen atmosphere in a three-channel electrochemical cell equipped with a working electrode, an auxiliary electrode, and a reference electrode. A glassy carbon (GC) electrode (working surface 3.14 mm²) served as a working electrode. A Pt wire with the diameter of 1 mm was used as an auxiliary electrode. Ag/AgNO₃ (0.01 M solution in CH₃CN) was used as a reference electrode (Е⁰(Fc⁺/Fc) = +0.20 V). Values were recalculated vs the ferrocenium/ferrocene external standard. The GC electrode was polished with 0.05 µm aluminium oxide polishing paper, and the electrode surface was cleaned before each experiment. Curves were recorded at a constant potential scan rate of 100 mV∙s^–1 using a Е2Р potentiostat from BASi Epsilon (USA). The device comprises a measuring unit and a PC DellOptiplex 320 with the Epsilon-EC-USB-V200 software.

Experimental Procedures and Product Characterization

Synthesis of Pt(L)Py.

A mixture of 1,3-bis(3,5-di-tert-butyl-2-hydroxyphenyl)-5,5-dimethyl-3,4,5,6-tetrahydropyrimidin-1-ium chlorideL^H3Cl (100 mg, 0.179 mmol), Pt(COD)Cl₂ (85 mg, 0.227 mmol), and potassium carbonate (30 equiv, 744 mg, 5.4 mmol) was suspended in pyridine in a Schlenk flask. The mixture was sonicated for 50 min and stirred at 50 °C for 20 h under a nitrogen atmosphere. The resulting suspension was concentrated under reduced pressure and then dissolved in dichloromethane; this solution was filtered through a Celite plug and concentrated under reduced pressure, affording complex Pt(L)Py (115 mg, 81%, orange solid). ¹H NMR (CDCl₃, 400.17 MHz, 300 K.): 1.12 (s, 18H, C(CH₃)₃), 1.27 (s, 6H, C(CH₃)₂), 1.30 (s, 18H, C(CH₃)₃), 3.73 (s, 4H, CH₂), 6.92 (d, ⁴J_HH = 2.4 Hz, 2H, CH_Phenoxy), 7.01 (d, ⁴J_HH = 2.4 Hz, 2H, CH_Phenoxy), 7.41 (td, ³J_HH = 6.3, ⁴J_HH = 1.4 Hz, 2H, CH_Py), 7.81 (tt, ³J_HH = 7.7, ⁴J_HH = 1.5 Hz, 1H, CH_Py), 8.82 (dd, ³J_HH = 4.8, ⁴J_HH = 1.5 Hz, 2H, NCH_Py). ¹³C{¹H}NMR (100.62 MHz, CDCl₃, 300 K) δ 24.72 (C(CH₃)₂), 28.40 (C(CH₃)₂), 29.76 (C(CH₃)₃), 31.69 (C(CH₃)₃), 34.30 (C(CH₃)₃), 35.29 (C(CH₃)₃), 58.41 (NCH₂), 115.68 (CH_Ar), 119.52 (CH_Ar), 124.65 (CH_Py), 136.15 (C–O), 138.01 (C_Py), 138.16 (C_Ar), 151.67 (CH_Py), 159.37 (CH_carbene). MALDI-TOF MS m/z: calcd for C₃₉H₅₅N₃O₂Pt [M]⁺ 792.97, found 792.4. Anal. Calcd (%) for C₃₉H₅₅N₃O₂Pt: C, 56.37; H, 6.51; N, 5.12. Found: C, 56.59; H, 7.78; N, 5.05.

Synthesis of [Pt(L)Py][BF₄].

In a Schlenk flask, Pt(L)Py (33 mg, 0.041 mmol) was dissolved in dichloromethane (7.5 mL). Afterwards, the solution of AgBF₄ (7.4 mg, 0.037 mmol) in dichloromethane (7.5 mL) was added dropwise to Pt(L)Py and the mixture was stirred for 1 h at room temperature. After the completion of the reaction, the solution was filtered, concentrated under reduced pressure, and washed three times with pentane (5 mL). Complex [Pt(L)Py][BF₄] was isolated in 63% yield (20.5 mg) as a dark green solid. Crystals suitable for X-ray diffraction analysis were grown by diffusion of pentane into a saturated solution of the product in dichloromethane at 0 °C. MALDI-TOF MS m/z: calcd for C₃₉H₅₅BF₄N₃O₂Pt [M]⁺ 792.97, found 792.4. Anal. Calcd (%) for C₃₉H₅₅BF₄N₃O₂Pt: C, 50.63; H, 5.73; N, 4.83. Found: C, 50.19; H, 7.00; N, 4.98.

Quantum-chemical Calculations

Geometry optimizations were carried out using the PBE0/LANL2DZ computational procedure, which employs the PBE0 functional [69] in conjunction with the LANL2DZ basis set and associated ECP [70,71,72,73] for Pt. Absorption spectra were calculated with time-dependent DFT using the B3LYP functional [74,75] and a larger basis set (def2-TZVP [76,77,78]). All calculations were done with the ORCA package (version 4.0) [79].

4. Conclusions

In summary, we have reported the direct and good-yielding synthesis and characterization of a Pt(II) complex bearing tridentate diphenolate tetrahydropyrimidin-1-ium based NHC ligand. The electrochemical behavior of the complex, investigated by voltammetric techniques, revealed the reversible redox wave at 0.25 V vs Fc⁺/Fc couple, which corresponds to the oxidation of the noninnocent phenolate moiety with the formation of Pt(II)-phenoxyl radical complex. The latter was obtained preparatively via “chemical” oxidation of neutral analogue by means of AgBF₄ and its electronic structure was determined by combined UV-Vis/NIR- and EPR-spectroscopy, X-ray diffraction and DFT studies. The obtained g value and spin-density plot suggest the predominant phenoxyl-radical character of the complex. The intense NIR band supports the delocalization of the unpaired electron over the ligand. It was found that neutral platinum complex is active in electrocatalytic oxidation of secondary amine (MEA) via the formation of active Pt(II)-phenoxyl radical complex oxidant with the i_cat/i_p value of 1.9.