Catalytic Hydrogenation of Carbon Dioxide Using Ir(III)-Pincer Complexes Ryo Tanaka, Makoto Yamashita, and Kyoko Nozaki* Department of Chemistry and Biotechnology, Graduate School of Engineering, The UniVersity of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan Received May 2, 2009; E-mail: nozaki@chembio.t.u-tokyo.ac.jp Formic acid is an important chemical product used as a preservative and an insecticide as well as for tanning leathers. In addition, formic acid plays a major role in synthetic chemistry as an acid, a reductant, and a carbon source. In industrial processes, formic acid is mainly synthesized by carbonylation of alcohols using carbon monoxide followed by hydrolysis, which requires a high pressure of toxic carbon monoxide gas. 1 Therefore, development of an alternative route to formic acid or its salt is highly desired. Hydrogenation of carbon dioxide to generate formic acid is a promising candidate because carbon dioxide is abundant, inexpen- sive, and a less toxic C1 source. Hydrogenation of carbon dioxide has been widely investigated using transition-metal complexes, 2 especially with rhodium, 3 ruthenium, 4-6 and iridium 7,8 catalysts. Noyori and co-workers carried out the reaction in supercritical carbon dioxide using a Ru(II) catalyst and obtained ammonium formate, alkyl formate, and formamide in high yields. 4 For example, the highest turnover frequency (TOF) (95 000 h -1 ) has been achieved using RuCl(OAc)(PMe 3 ) 4 , which is soluble in supercritical CO 2 . 5 Himeda reported a highly active cationic Cp*Ir(III) catalyst containing phenanthroline derivatives as ligands. The maximum turnover number (TON) of the hydrogenation reached 222 000 (the highest TON reported to date) in a basic aqueous solution. 8 They stated that electron-donating ligands accelerated the reaction. Pincer ligands are known as multidentate ligands that strongly bind to a metal center to prevent dissociation of the ligand from the metal. 9 The idea that an alkylphosphine-based pincer ligand would be an efficient electron donor prompted us to use a transition- metal-pincer complex as a catalyst for hydrogenation of CO 2 . Although insertion of CO 2 into PCP-ligated rhodium dihydrogen complexes such as 1 and 2 to form formato complexes were intensively studied, 10 catalytic hydrogenation of CO 2 with these complexes has never been reported. Herein we describe the synthesis of new PNP-Ir(III) hydride complexes and their application to the catalytic hydrogenation of CO 2 in aqueous base. The PNP-Ir(III) trihydride complex 5 bearing isopropyl groups on the phosphorus atoms shows higher TON and TOF than reported for any other catalyst to date. Deprotonation of the PNP ligand is proposed to play an essential role in the catalytic cycle. PNP-ligated chloroiridium(III) dihydride complexes 4a and 4b were synthesized by the reaction of an Ir(I) source with an excess amount of alkyl-substituted PNP pincer ligands 3a and 3b, respectively, under H 2 pressure (Scheme 1). The structures were determined by NMR spectroscopy. In the 1 H NMR spectra of 4a and 4b, each of the two hydrides is coupled with the two magnetically equivalent phosphorus atoms and with the other hydride. Thus, two hydrides are suggested to occupy positions cis and trans to the nitrogen atom, forming an octahedral structure. PNP-ligated iridium(III) trihydride complex 5 was synthesized from 4b in 77% yield by addition of an excess amount of NaH. X-ray crystallographic analysis of 5 revealed a meridional coordination mode of the PNP pincer ligand, with three peaks in the Fourier map that could be assigned as hydrides to form the octahedral geometry. The IR spectrum of 5 showed a trans iridium dihydride stretch at 1678 cm -1 , which is a much lower wavenumber than that previously reported for mer - IrH 3 (PPh 3 ) 3 (1745 cm -1 ), 11 indicating that the Ir -H bonds in 5 are weakened by the electron-rich PNP pincer ligand. Structural optimization of 5 using B3LYP//6-31G(d)/LANL2DZ calculations reproduced the crystal structure as well as the IR spectrum. The hydrogenation of carbon dioxide in aqueous KOH was investigated using Ir(III)-PNP pincer complex 4a, 4b, or 5 or t Bu-PCP-ligated iridium(III) complex 6a 9a or 6b 9c as the catalyst (Table 1). At a reaction temperature of 200 °C, potassium formate was formed even in the absence of catalyst 12 (entry 1), and thus, background corrections to the catalyst TONs based on the data from entry 1 were made. PNP complex 4a showed higher activity than PCP complex 6a (entries 2, 3). Changing the t Bu groups on the phosphorus atoms in 4a to i Pr groups in 4b afforded higher activity, probably because of the higher solubility of 4b (entries 3, 4). Addition of THF as a cosolvent was necessary because of the poor solubility of 4b in water (entry 5). Iridium trihydride complex 5 showed much higher catalytic activity than its chloroiridium dihydride analogue 4b (entries 4, 7). PCP-ligated iridium dihydride complex 6b also showed higher activity than its chloroiridium hydride analogue 6a (entries 2, 6). These results indicate that chloride may suppress the catalytic reaction. Using 5 increased the TON to 150 000 h -1 at 200 °C (entry 7), 13 which is 1.5 times as high as the best TOF value in the literature. 5 At 120 °C and lower catalyst loading (100 ppb) (entry 9), the TON of 3 500 000 was 15 times higher and the TOF almost Scheme 1. Synthesis of PNP-Ligated Chloroiridium(III) Dihydride Complexes 4a and 4b and PNP-Ligated Iridium(III) Trihydride Complex 5 and ORTEP Drawing of 5 (50% Thermal Ellipsoids, Hydrogen Atoms except Hydrido Ligands Have Been Omitted for Clarity) Published on Web 09/23/2009 10.1021/ja903574e CCC: $40.75  2009 American Chemical Society 14168 9 J. AM. CHEM. SOC. 2009, 131, 14168–14169