|Home | About | Journals | Submit | Contact Us | Français|
A series of Cp*Ir catalysts are the most active known by over an order of magnitude for water oxidation with Ce(IV). DFT calculations support a Cp*Ir=O complex as active species.
Few homogeneous water oxidizing catalysts are known. Ones that contain Ru, Mn or Ir have been reported,2 but high activity and stability are often lacking. A few are only active with the 2e− primary oxidants, such as oxone or OCl−. Ones active with Ce(IV) are more significant, being more relevant to the 4 × 1e− process of photosynthesis. We now describe several robust water oxidation catalysts using Ce(IV) as the primary oxidant that are the most active yet reported, achieving an initial rate of 54 t.o./min., and are active beyond seven hours.
Bernhard et al.2i have recently reported a robust homogeneous Ir catalyst, 3 in Fig. 1, that maintains activity with Ce(IV) for one week. Having confirmed this result, we moved to complexes with much more strongly donating ligand sets to look for improved activity, and now report 1a–b and 2 as highly active catalyst precursors for water oxidation, easily accessible from commercially available [Cp*IrCl2]2.3 The crystal structure of 2, shown in Fig. 2, confirms the expected atom connectivity.
Complex 1a dissolves in water (50 mM), but only on addition of excess Ce(IV) in the form of CAN (cerium ammonium nitrate), and gas bubbles are observed within one minute. For oxygen detection by the Clark electrode, bubble formation has to be avoided, so rate data for the reaction was obtained under the dilute conditions noted in Figure 3 by adding catalyst (10 μL, 38nmol) dissolved in a MeCN/H2O mixture to the Ce(IV) solution (7 mL, 78 mM). This shows initial oxygen evolution rates for 1a, 2 and 3 in water with CAN and no lag time is observed upon injection of the catalyst. Complex 1a (54 t.o./min) and 2 (17 t.o./min) are significantly more active than the known 3 (3 t.o./min), and more active by at least an order of magnitude than those shown in Table 1. Because reaction is so fast, measurements using the Clark electrode are limited to 1 min., dilatometry was used for measurements over longer periods (see SI). No catalyst deactivation is observed after 7.5 hours.4 The rate over seven hours is lower (0.1 s−1) than the initial rate, consistent with most catalysts reported in Table 1.
Cyclic voltammetry (CV) of the chloro complex 1a in MeCN shows three irreversible oxidation waves in contrast to triflato complex 1b which only shows two peaks. This, together with the much greater rate seen for 1a versus 1b (see SI) suggests that chloride oxidation may accelerate O2 evolution via oxidation to OCl−,5 or Cl2.6 This is consistent with previous reports of an accelerating effect of Cl− ion.7 However, we see 11.9 t.o./min from the triflate complex 1b, so Cl− is not required for activity. The [Cp*IrCl2]2 dimer also evolves O2, albeit at a much slower rate (8 t.o./min). The formation of IrO2 or other heterogeneous catalyst is unlikely: the catalyst can be fully recovered from the reaction mixture (NMR), no lag phase is seen, no dark deposit is formed, evolution is reproducible and the reaction is 1st order in 1b. Furthermore, the reported reaction rate catalyzed by IrO2 is very slow (Table 1, Entry 8). Other homogeneity tests are reported in the SI. Water was confirmed as the source of O by the detection of 36O2 in 18O labeling experiments.
Assuming the chelate remains bidentate, 1–2 only have one labile site. If so, this limits the mechanistic possibilities and the steric bulk present may inhibit μ-oxo dimer formation or other bimolecular deactivation processes.
Moving from the pyridine complex 1a to the pyrimidine complex 2 causes a rate decrease. Pyrimidines are less donor than pyridines and at low pH they can also protonate at the distal N. The need for strong donor ligands is consistent with a requirement for easy attainment of a high oxidation state intermediate, plausibly an IrV oxo. A 2e− oxidation from IrIII is also consistent with the presence of two oxidation waves in the CV for 1b in MeCN (see SI).
The minimum number of metal atoms required for water oxidation is debated, but several reports2c,2i,2k,2l suggest that a single metal is sufficient. The reaction is first order with respect to 1b in the 1–80 μM range (see SI). This makes [(Cp*)Ir(O)(ppy)]+ a plausible intermediate. DFT calculations8 show that the electronic structure of this pseudo-octahedral (t2g)4 complex can be understood by the formal combination of the [(Cp*)Ir(ppy)]3+ and O2− fragments (Figure 4).9 The electrons on the metal are localized in the non-bonding dx2−y2 orbital, which does not interact with O p orbitals. The Ir d orbitals oriented towards N, dN, and C, dC, make dπ/pπ interactions with the O p orbitals, yielding two bonding π(Ir=O) orbitals, πN and πC, and two anti-bonding π*(Ir=O) orbitals, π*N and π*C. In the singlet state, π*N is doubly occupied and π*C is empty, whereas in the triplet both orbitals are half occupied. The presence of low-lying π*(Ir=O) orbitals, which have significant O contribution, may promote the formation of an O-O bond by nucleophilic attack to the oxo group, which is the key step in water oxidation.10,11 The singlet state is only 3.7 kcal mol−1 above the triplet, due to the similar energies of the π*N and π*C orbitals. The accessibility of the singlet state may also facilitate the O-O bond formation, because the addition of the nucleophile will yield a diamagnetic IrIII octahedral complex. The electronic structure of [(Cp*)Ir(O)(ppy)]+ is consistent with this complex being the active species of the catalytic system. A thorough experimental and theoretical investigation is now in progress.
R.H.C, G.W.B. and J.F.H. acknowledge NSF (CHE-0614403) and NIH (GM32715) for funding. D.B. thanks Sanofi-Aventis for a postdoctoral fellowship. We thank J. N. Harvey for helpful discussions.
Supporting Information: Synthesis and characterization of compounds 1a–b and 2, and crystal parameters. Experimental details for oxygen evolution and kinetic measurements, CV and homogeneity tests. Optimized geometries and energies of the singlet and triplet states of [(Cp*)Ir(O)(ppy)]+. Frontier molecular orbitals of the singlet. This material is available free of charge via the Internet at http://pubs.acs.org.