Evidence for a terminal Pt(iv)-oxo complex exhibiting diverse reactivity

Abstract

Terminal oxo complexes of transition metals have critical roles in various biological and chemical processes^1,2. For example, the catalytic oxidation of organic molecules^3,4, some oxidative enzymatic transformations^5,6,7, and the activation of dioxygen on metal surfaces⁸ are all thought to involve oxo complexes. Moreover, they are believed to be key intermediates in the photocatalytic oxidation of water to give molecular oxygen, a topic of intensive global research aimed at artificial photosynthesis and water splitting^{9,10,11,12,13}. The terminal oxo ligand is a strong π-electron donor, so it readily forms stable complexes with high-valent early transition metals. As the d orbitals are filled up with valence electrons, the terminal oxo ligand becomes destabilized². Here we present evidence for a dⁿ (n > 5) terminal oxo complex that is not stabilized by an electron withdrawing ligand framework. This d⁶ Pt(iv) complex exhibits reactivity as an inter- and intramolecular oxygen donor and as an electrophile. In addition, it undergoes a water activation process leading to a terminal dihydroxo complex, which may be relevant to the mechanism of catalytic reactions such as water oxidation.

Main

To date, the very few terminal oxo complexes having more than five valence d electrons are all stabilized by powerful electron accepting ligands: namely NaRe^(I)(O)(PhCCPh)₂, in which extensive back bonding to the alkyne ligands was noted¹⁴, and complexes stabilized by the encapsulating, highly effective π-accepting polytungstate ligand, including Pt^(IV) (ref. 15), Pd^(IV) (ref. 16) and Au (ref. 17). A d⁵ Fe(iii) oxo complex stabilized by a protective hydrogen bond cavity has been reported¹⁸. Isolation of a terminal d⁶ metal-oxo complex not bearing strong electron withdrawing groups is desirable, as it may provide insight into the reactivity potential of the metal-oxo bond, in addition to information regarding spectroscopic and structural aspects. Particularly desirable are terminal oxo complexes of platinum, as this metal and its compounds are widely used by industry in oxygen-based technologies. Towards this goal, we have used a pincer complex of platinum.

Pincer complexes, involving a meridional, tridentate ligand framework, are of considerable current interest, owing to the structural versatility of such systems, leading to a variety of bond activation and catalytic processes¹⁹. Reaction of the cationic pincer PCN Pt complex 1a (PCN = C₆H₃[CH₂P(t-Bu)₂](CH₂)₂N(CH₃)₂; ref. 20) with a freshly prepared acetone solution of dioxirane²¹ resulted in immediate formation of a new complex, formulated as the terminal oxo complex 2 (Fig. 1a). The ³¹P{¹H} NMR spectrum of 2 shows a singlet at 64.27 p.p.m. with Pt satellites (coupling constant J_Pt-P = 2,686 Hz) and the ¹⁹⁵Pt{¹H} NMR spectrum exhibits a signal at -645.88 p.p.m. (broad doublet, J_P-Pt = 2,686 Hz), definitely indicating a Pt(iv) oxidation state (for comparison, the chemical shift of the starting Pt(ii) complex 1a is -4,015 p.p.m.). (Here {¹H} indicates proton-decoupled.) In the ¹³C{¹H} NMR spectrum, the ipso carbon gives rise to a doublet at 147.25 p.p.m. (J_P-C = 5 Hz), and in the ¹⁹F{¹H} NMR spectrum, the outer-sphere BF₄ anion gives rise to a singlet at -149.87 p.p.m. The NMR data exclude the possibility of the corresponding (PCN)Pt(OH) complex, which has a completely different NMR pattern²⁰. Electrospray mass spectrometry analysis is in line with the structure of 2, indicating two signals at mass/charge ratios m/z⁺ 518.07(M+1) and m/z^- 87.29 (BF₄). The infrared spectrum in acetone solution shows an absorption at 783 cm^-1, which is in the range expected for Pt = O vibration (Supplementary Fig. 4)²².

Figure 1: Preparation and DFT structure of complex 2.

a, Preparation of complex 2 by reaction of the Pt(ii) complex 1a with dioxirane. b, DFT structure of complex 2 optimized at the PBE0/pc1 level of theory (hydrogen atoms are omitted for clarity).

PowerPoint slide

To exclude a dimeric [(PCN)Pt(μ-O)]₂ structure, PGSE (pulsed gradient spin echo) NMR studies were performed in acetone-d₆. Diffusion coefficients (D) and hydrodynamic radius (r) of complex 2 and, for comparison, complex 3 (see below) were determined, giving D = 1.078 × 10^-9 m² s^-1, r = 6.1 Å for 2, and D = 1.186 ×10^-9 m² s^-1, r = 5.6 Å for 3. Both the absolute D and r values of complex 2 and their good similarity to that of the unambiguously monomeric 3 clearly indicate the monomeric structure of 2 (ref. 23), as also strongly indicated by density functional theory (DFT) calculations, which show that it is impossible to obtain a Pt(iv) μ-oxo dimer, or a μ-hydroxo-dimer, starting from complex 2 because of severe steric constraints. The radius obtained by rotating the DFT structure of 2 (see below) about its geometric centre is 6.06 Å, in correlation with the experimental hydrodynamic radius. The diverse reactivity of complex 2 is commensurate with its monomeric structure (see later).

DFT studies performed on complex 2 show that the reaction 1a → 2 is thermodynamically favourable (ΔG⁰₂₉₈ = –11.1 kcal mol^-1 in acetone solution). According to the calculations, the optimized structure of 2 adopts a distorted square planar geometry with the oxygen atom being 35.3° out of the C–Pt–P plane (Fig. 1b). The calculated Pt–O bond length (1.811 Å) is unambiguously shorter than typical Pt(iv)–OH single bonds, which are in the range 1.943–2.079 Å (based on the Cambridge Structural Database), but it is longer than that reported¹⁵ for K₇Na₂[O = Pt(H₂O)(PW₉O₃₄)₂], 1.720 Å. Analysis of the covalent part of the Pt–O bonding in complex 2 shows that the low stability of electron-rich terminal oxo complexes can be attributed to occupation of strongly antibonding Pt–O π* molecular orbitals. The non-planar geometry of the complex decreases the overlap between metal-ligand orbitals, reducing this unfavourable occupation. The phenyl group serves as an electron donor, and no backdonation to it takes place because of the high energy of the vacant phenyl orbitals (Supplementary Fig. 2 presents a simplified molecular orbital diagram), as supported also by a comparison of natural bond orbital charges in 2 and 1 (Supplementary Fig. 3). According to the DFT calculation, acetone coordination is unfavourable (see Supplementary Information).

The local atomic environment and charge state of Pt in complex 2 were investigated by two methods of X-ray absorption spectroscopy, namely, extended X-ray absorption fine structure (EXAFS) and X-ray absorption near edge structure (XANES).

XANES data (Fig. 2a) of 2 and other compounds used for calibration (complexes 1a (solv = H₂O), 1c and 4 (Fig. 3) as well as Pt foil) confirmed the nominal charge states of Pt in all the complexes. The observed differences in the d-hole density which is proportional to the absorption peak area are consistent with the difference in the formal valences of Pt in complexes 2 and 4 (4+), 1a and 1c (2+) and Pt foil (0).

Figure 2: XANES spectra of complexes and Fourier transform magnitudes of EXAFS data.

a, XANES spectra for complex 2 and reference complexes 1a (solv = H₂O), 1c, 2 and 4. b, Fourier transform (FT) magnitudes, uncorrected for photoelectron phase shifts, of k-weighted EXAFS data for complex 2 and individual contributions Pt = O, Pt–OH and Pt–OH₂. Inset, the ∼0.1 Å shift to lower distances of the Pt = O peak relative to Pt–OH (the latter peak is scaled up by a factor of 2 for clarity).

PowerPoint slide

Figure 3: Reactivity of complex 2.

Oxygen transfer to H₂ and to PPh₃ with reduction to the Pt(ii) complex 1a; oxygen transfer to CO forming complex 1b and CO₂; intramolecular oxo transfer to form complex 3; nucleophilic attack by a hydride, generating the Pt(ii) hydroxo complex 1c; and water activation to form the dihydroxo Pt(iv) complex 4.

PowerPoint slide

Analysis of EXAFS data in the four complexes was performed self-consistently by imposing constraints dictated by our X-ray diffraction and DFT analyses: for example, we assumed that the lengths of Pt–P, Pt–N, Pt–C, Pt = O, Pt–OH and Pt–OH₂ bonds are unique and do not significantly depend on a particular complex. By using such a ‘fingerprinting’ approach, we obtained partial contributions χ_Pt–OH, and χ_{Pt = O} to EXAFS data of these complexes. These three types of platinum–oxygen bond were found to have very different strengths: Pt = O is the stiffest, Pt–OH₂ the weakest and Pt–OH the intermediate in strength between the two former bonds, as intuitively expected (Fig. 2b). Our results demonstrate that the Pt = O peak maximum is shifted to the left relative to the Pt–OH peak maximum (Fig. 2b, inset) by approximately 0.1 Å, in agreement with our DFT calculations described above, which indicate that the Pt = O in 2 differs from the Pt–OH in complex 4 (see below) by 0.14 Å (at the PBE/sdd level). FEFF6 calculations of the theoretical EXAFS signal for the DFT model of complex 2 are in qualitative agreement with the DFT calculations.

Complex 2, which has no significant electron accepting framework that can stabilize the terminal oxo ligand, exhibits diverse reactivity (Fig. 3). Initially, we explored its oxygen transfer reactivity. Reaction of 2 with PPh₃ at room temperature resulted in quantitative formation of OPPh₃ with parallel quantitative formation of complex 1a. Oxygen transfer activity was also observed upon reaction with CO. When 2 was treated with 4 equivalents of CO gas, it was quantitatively transformed into the cationic carbonyl complex 1b with concomitant formation of CO₂ (83% yield) as determined by gas chromatography. Complex 2 also reacts with hydrogen. Stirring 2 under 5 atm of H₂ led to deoxygenation, forming complex 1a in 52% yield, with complex 3 (see later) being a competitive product. Mechanistic studies of H₂ oxidation by MnO₄^- and RuO₄ have been reported²⁴.

Complex 2 can also be attacked by nucleophiles. Addition of KH to a THF solution of 2 resulted in formation of the reported²⁰ Pt(ii) hydroxo complex 1c, probably by hydride attack at the metal centre, followed by its migration to the oxo ligand (Fig. 3). Complex 1c constitutes 78% of the organometallic products (see Supplementary Information).

Complex 2 is only moderately thermally stable. After 7–10 h at room temperature, in the absence of an external oxygen acceptor, it undergoes intramolecular oxygen transfer, resulting in insertion of the oxo ligand into the Pt–P bond to give the new complex 3 (Fig. 3). Complex 3 was fully characterized by multinuclear NMR spectroscopy. The ³¹P{¹H} NMR spectrum of 3 exhibits a singlet at 76.57 p.p.m. with a very small Pt–P coupling constant (40 Hz) due to oxygen incorporation. The ¹⁹⁵Pt NMR spectrum exhibits a singlet at -2,842.87 p.p.m. Colourless needles of 3 suitable for a single-crystal X-ray diffraction study were obtained from an acetone/pyridine solution by slow vapour diffusion of ether at room temperature (Fig. 4a). Complex 3 (solv = pyridine) exhibits a slightly distorted square planar structure. The oxygen atom is bound to the Pt and P atoms, and the Pt–O length (2.031 Å) indicates a single-bond character, while the P–O bond length of 1.529 Å is short enough to be considered a double bond (based on the Cambridge Structural Database). Results of DFT calculations on the structure of complex 3 are in good agreement with the X-ray data. The geometries of complexes 2, 3 and 4 were optimized at the same level of theory. The observed good correlation between theoretical and X-ray single-crystal data for 3 and 4 (Supplementary Tables 2 and 3) lends credence to the calculated geometry of complex 2.

Figure 4: ORTEP views of molecules of 3 and 4 with the thermal ellipsoids at 50% probability.

a, 3; b, 4. Hydrogen atoms were omitted for clarity (OH hydrogen atoms of complex 4 were not located).

PowerPoint slide

Terminal oxo intermediates are implicated as key intermediates in the artificial catalytic water oxidation to O₂, the mechanism of which is a subject of numerous theoretical and experimental studies^25,26,27,28, although the nature of the step leading to the generation of dioxygen is not clear^{9,10,11,12,13}. In this context, it is desirable to know how a late transition metal terminal oxo complex reacts with water. Interestingly, when excess water was added to 2, the new dihydroxo Pt(iv) complex 4 was immediately formed (Fig. 3). Complex 4 was fully characterized by multinuclear NMR and mass spectroscopy (see Supplementary Information). Colourless crystals of 4 suitable for X-ray diffraction analysis were obtained from a CH₂Cl₂ solution at -30 °C (Fig. 4b). Complex 4 has a classical octahedral structure, the Pt atom being coordinated to the PCN ligand and to three oxygen atoms. The Pt–O bond lengths are 1.997(5), 2.017(5) and 2.171(5) Å for atoms O1, O2 and O3, respectively. Lacking the Pt = O functionality, and being significantly more stable than 2, complex 4 does not exhibit oxygen transfer reactivity and remains unchanged for weeks upon addition of PPh₃.

We are unaware of prior direct observations of water activation by a terminal oxo complex, and trapping of the product of such activation. It is noteworthy that oxidation of Pt(ii) to Pt(iv) by O₂ in water to generate a Pt(iv) dihydroxo complex is proposed as a key step in a hypothetical catalytic cycle for alkane oxidation by O₂ to give alcohols^29,30. Our results suggest that such a process might involve O–H activation of water by a Pt(iv) oxo intermediate. DFT calculations show that in our system such activation is very favourable. Upon addition of two water molecules to the apical positions of the oxo complex 2, geometry optimization gave the dihydroxo structure 4. The calculated free energy change of this reaction is ΔG^o₂₉₈ = -19.8 (-20.6) kcal mol^-1 in acetone (water) solution.