Strong Ligand Stabilization Based on π‐Extension in a Series of Ruthenium Terpyridine Water Oxidation Catalysts

Abstract The substitution behavior of the monodentate Cl ligand of a series of ruthenium(II) terpyridine complexes (terpyridine (tpy)=2,2′:6′,2′′‐terpyridine) has been investigated. 1H NMR kinetic experiments of the dissociation of the chloro ligand in D2O for the complexes [Ru(tpy)(bpy)Cl]Cl (1, bpy=2,2’‐bipyridine) and [Ru(tpy)(dppz)Cl]Cl (2, dppz=dipyrido[3,2‐a:2′,3′‐c]phenazine) as well as the binuclear complex [Ru(bpy)2(tpphz)Ru(tpy)Cl]Cl3 (3  b, tpphz=tetrapyrido[3,2‐a:2′,3′‐c:3′′,2′′‐h:2′′′,3′′′‐j]phenazine) were conducted, showing increased stability of the chloride ligand for compounds 2 and 3 due to the extended π‐system. Compounds 1–5 (4=[Ru(tbbpy)2(tpphz)Ru(tpy)Cl](PF6)3, 5=[Ru(bpy)2(tpphz)Ru(tpy)(C3H8OS)/(H2O)](PF6)3, tbbpy=4,4′‐di‐tert‐butyl‐2,2′‐bipyridine) are tested for their ability to run water oxidation catalysis (WOC) using cerium(IV) as sacrificial oxidant. The WOC experiments suggest that the stability of monodentate (chloride) ligand strongly correlates to catalytic performance, which follows the trend 1>2>5≥3>4. This is also substantiated by quantum chemical calculations, which indicate a stronger binding for the chloride ligand based on the extended π‐systems in compounds 2 and 3. Additionally, a theoretical model of the mechanism of the oxygen evolution of compounds 1 and 2 is presented; this suggests no differences in the elementary steps of the catalytic cycle within the bpy to the dppz complex, thus suggesting that differences in the catalytic performance are indeed based on ligand stability. Due to the presence of a photosensitizer and a catalytic unit, binuclear complexes 3 and 4 were tested for photocatalytic water oxidation. The bridging ligand architecture, however, inhibits the effective electron‐transfer cascade that would allow photocatalysis to run efficiently. The findings of this study can elucidate critical factors in catalyst design.

Mass spectrometry (MS): MS analysis was performed on a Bruker solariX (2010) Hybrid 7 TFT-ICR for MALDI and with a Bruker Ultraflex III MALDI TOF/TOF for MALDI/TOF measurements. Mass data were converted using Compass DataAnalysisViewer V5.0.
Photochemical water oxidation catalysis: [2] all studies were carried out in de-aereated solvents under inert atmosphere. All catalytic experiments were tempered by a custom air cooling setup (25 °C, Fig S1). A screw cap, hermetically sealed vial (diameter: 12.75 ± 0.25 mm, length: 99.00 ± 0.50 mm) equipped with two sensor spots (see "Oxygen detection" below) was used as reaction vessel. The mixtures were stirred during catalysis by a 3D printed custom made stirring bar.
Oxygen detection: [2] Oxygen concentrations were determined using a FireStingO2 optical oxygen meter (Pyroscience, Germany) using oxygen sensitive optical sensor spots (OXSP5, with optical isolation). The spots were glued to the inner glass vessel wall of a screw-capped vial (transparent silicone glue, SPGLUE). The sensor spots are stable between pH 1 -14, stable in the presence of strong oxidants (the manufacturer suggests cleaning in 3 % aqueous H2O2), are not affected by the solvents used (as shown by repeated reproducible catalytic measurements) and are autoclavable at 120 oC. For further information, see the manufacturer information: https://www.pyro-science.com/contactless-fiber-optic-oxygensensor-spots.html. O2 concentration was measured in μmol/L (solution) and mbar (gas-phase). Both spots were calibrated by two-point calibration: gas-phase calibration was performed against ambient air and Ar-atmosphere. Liquid-phase calibration was performed using a de-oxygenated reaction solution. Solution TONs were calculated based on the detected concentration; gas-phase TONs were calculated via the ideal gas equation.
Calculation of equilibrium constants: Equilibrium constants (Ke) were calculated using equations 1 and 2: Where m is the slope of the linear plot of logarithm of the difference in the integrals of H a Cl and H a OD (see Figure 1, main manuscript) derived from the 1 H-NMR spectra for compounds 1, 2 and 3b vs. time.  [Ru(tpy)Cl3]: This compound was synthesized according to a literature-known procedure. [3] RuCl3 x H2O (0.56 g, 2.1 mmol) and 2,2':6',2''-terpyridine (0.50 g, 2.1 mmol) were added to absolute ethanol (10 mL) and refluxed for 3 h. After evaporation of the solvent in vacuum, the solid was washed well with water and diethyl ether. The product could be isolated as brown solid (80%, 0.74 g) and was used directly for the next steps.
[  3 1 v/v) and the mixture was refluxed for 24 h. A color change from red to dark-red/brown could be observed. The solution was allowed to cool down to room temperature and the solvent was evaporated under reduced pressure. Upon addition of an excess of aqueous NH4PF6, the precipitated product could be filtered off. The crude product was further purified by slow diffusion of ether to an acetonitrile solution to yield a dark red solid (85%, 82 mg).  Figure S1: 1

H-NMR spectrum of [Ru(tbbpy)2(tpphz)Ru(tpy)Cl](PF6)3 in acetonitrile-d3 at 25 °C.
S9 Figure S5: 1 [4] Complex 3 Ligand exchange R 0.0 (0.0) AR 0.6 (4.9) TS 26.0 (30.3) AP 15.0 (18.0) P 5.6 (9.9)  Table S4. Ru-Cl bond dissociation energy scans for complexes 1, 2 and 3, starting from the equilibrium geometry of each complex. In the case of complex 3, due to the considerably increased size, the relative energy was only evaluated at the limit of 5.0 Å. Figure S19 Plotted values of the Ru-Cl bond dissociation energy scans for complexes 1,2 and 3, from the data depicted in Table S4.   Table S7 Relevant geometrical changes across the ligand exchange step, with corresponding reactants (R), associated reactants (AR), transition states (TS), associated products (AP) and product (P) steps of complex 1, 2 and 3. The structure of the complexes has been represented in a more simplified manner.