Formation of a C–C double bond from two aliphatic carbons. Multiple C–H activations in an iridium pincer complex

Iridium can mediate a reversible intramolecular coupling reaction involving up to four unactivated Csp3–H bonds, to give a carbon–carbon double bond.

1 H NMR (500 MHz, C 6 D 6 ): δ 8.39 (d, 3 J HH = 7.1 Hz, 1H, 12-H), 8.21 (d, 3  Aromatization of the cyclohexane-based pincer pigand. A Straus flask was charged with PCyP ligand (0.085 g, 0.212 mmol), [Ir(COD)Cl] 2 (0.071 g, 0.106 mmol) and 4 ml of toluene as well as 0.1 ml of tert-butylethylene inside a nitrogen atmosphere glovebox. The flask was sealed, fully immersed into an oil bath and heated at 220 ºC for 48 h. The resulting solution was filtered through a thin layer of Celite and evaporated. The residue was dissolved in hexane, partially evaporated, and kept in the freezer (-20 ºC) overnight. The mother liquor was decanted and the residue was washed with a small amount of cold hexane to give [2,6-( t Bu 2 PCH 2 ) 2 C 6 H 3 ]IrHCl (19) as a red powder (0.105 g, 80%). NMR spectra are consistent with the literature data 1 .
Comparison of the aromatization of the cyclohexane-based pincer ligand with and without hydrogen acceptor. Two Straus flasks were charged with (PCyP)IrHCl (1) (0.015 g, 0.024 mmol) and 5 ml of toluene inside a nitrogen atmosphere glovebox. To one of them, 0.1 ml of tert-butylethylene was added. The flasks were sealed, fully immersed into the same oil bath and heated at 220 ºC for specified times (note: the reaction is quite sensitive to the temperature gradients in the oil bath and to the heat flow, therefore the times may differ somewhat from the ones stated below). According to NMR monitoring, in the flask with tert-butylethylene, aromatization was complete within 48 h, while the flask without hydrogen acceptor contained ca. 55% 18, 25% 1, 10% of 19 and 10% of 21 (average from three experiments). Further heating, if the flask is not opened, does not result in any significant changes; if the flask is opened and purged with Ar several times the reaction slowly goes to completion. Therefore, for NMR monitoring of the acceptorless experiments, the flasks were not used further after being opened.
Synthesis of carbene complex 16 and mono-dehydrogenated isomeric complexes 17. Complex 1 (100 mg, 0.16 mmol) was placed into a Schlenk flask and heated in an oil bath for 9 h at 200 ºC while passing a slow stream of argon above the flask. Periodically, the flask was removed from the oil bath and the solid which sublimed above the oil level was returned to the bottom of the flask. While several solution and solid-state techniques for thermolysis of 1 were tried, this was found to produce the best yields of compounds 16 and 17. A mixture containing 43% of starting material, 5% of 16, 47% of 17 (28% of one isomer and 19% of another isomer), as well as 4% of 18 and 1% of 19 was formed, which was analyzed by NMR. Despite a number of tries, including solution and solid-state thermolysis of 1 under various conditions, we were unable to obtain complex 16 in a high yield and isolate it from the other products and so far 16 and Hydrogenation of 16, 17 and 18. A solution of complex 1 (0.015 g, 0.024 mmol) in 5 ml of toluene in a sealed Straus flask was heated for 2 h at 205 ºC, and a mixture of 1, 16, 17, and 18 was formed according to NMR. The solution was degassed, cooled to -196 ºC and the flask was refilled with H 2 . Upon heating the solution for 6 h at 155 ºC, a clean and quantitative conversion to 1 was observed.
Attempted hydrogenation of 19. A degassed solution of benzene-based complex 19 (0.015 g, 0.024 mmol) in 5 ml of toluene in a sealed Straus flask was cooled to -196 ºC and the flask was refilled with H 2 . After this, the flask was fully immersed into an oil bath and heated; at temperatures up to 215 ºC, no formation of hydrogenated products was observed.
Synthesis of diene complex 18 and C-C coupling product 21. A Straus flask was charged with PCyP ligand (0.150 g, 0.374 mmol), [Ir(COD)Cl] 2 (0.126 g, 0.188 mmol) and 6 ml of toluene as well as 0.1 ml of tert-butylethylene inside a nitrogen atmosphere glovebox. The flask was sealed, fully immersed into an oil bath and heated at 210 ºC for 32 h. According to NMR, at that moment the reaction mixture contained 75% of 18 and 12% of 21, and small amounts of 1 and 19 as well as traces of 16 and 17. The volatiles were removed in vacuum and the residue was chromatographed under inert atmosphere on alumina, which was preliminary heated for 2 h at ca. 150 ºC under vacuum. Dry and degassed benzene-hexane 1:1 mixture was used as eluent. As far as color changes are difficult to distinguish and target compounds rapidly decompose on TLC plates, NMR monitoring was required. Compounds 21 and 1 were eluted first, together with some amount of 18. These fractions were evaporated, the residue was dissolved in a minimum amount of hexane at 120 ºC into a Straus flask and the flask was allowed to reach ambient temperature. Relatively large, yellow crystals of 21 were formed which could be readily separated from the rest of solid material. Yield: 0.013 g, 6%. The middle fractions contained complex 18 in an up to 90% purity, while complex 16 eluted the last. Those middle fractions were evaporated and subjected to another chromatography on silica, which was preliminary heated for 2 h at ca. 150 ºC under vacuum. Dry and degassed benzene was used as eluent. Complex 19 was eluting first together with 18, while the next colored fractions contained pure 18. After evaporating the volatiles and drying in vacuum complex 18 was obtained as a red-orange powder. Yield: 0.036 g, 15%. Characterization of complex 18: S10 P( t Bu) 2 P( t Bu) 2 Ir Cl

Crystallography.
Intensity data were collected at 293 K with an Oxford Diffraction Xcalibur 3 system using ω-scans and Mo-Kα (λ = 0.71073 Å). CCD data were extracted and integrated using Crysalis RED 2 . The structures were solved using direct methods and refined by full-matrix least-squares calculations on F 2 using SHELXTL 5.1 3 . Non-H atoms were refined with anisotropic displacement parameters. Hydrogen atoms were constrained to parent sites, using a riding model.

Summary of Additional Computational Results
All Density Functonal Theory (DFT) calculations were carried out with Jaguar 7.6 program package by Schrödinger LLC 4 . For geometry optimization, solvation energy and frequency calculations, Becke´s three-parameter hybrid functional and the LYP correlation functional (B3LYP) 5,6 was used with LACVP** level core potential and basis set, while single point energy corrections were performed with the M06 7 , M06-L 8 or B3LYP-D3 9 functional using the LACV3P**++ basis set which, as suggested by Martin 10 , was augmented with two f-polarization functions on Ir. Frequency calculations at the same level were performed on the optimized geometries to verify that the geometries correspond to minima (no imaginary frequency) o transition states (one imaginary frequency) and to provide the thermochemical data at different temperatures (298 K, 393 K and 413 K), which include entropy contributions. Single-point solvation energies were also calculated using the Poisson-Boltzmann reactive field implemented in Jaguar 7.6 (PBF) 11 with standard parameters for benzene. The Gibbs free energies were defined as following equation: G = E(B3PLYP-D3/LACV3P**++2f on Ir) + Gsolv + ZPE + H 298 -TS 298 + 1.9 [concentration correction to the free energy of solvation from M(g) → M(aq) to atm(g) → M(aq)].

Choice of functional: M06, M06-L versus B3LYP-D3
Since DFT studies for iridium complex with PCP pincer ligands with different density functionals have been reported 12 , we carried out comparable calculations using these three functionals (M06, M06-L and B3LYP-D3). In general all the calculations follow the send trend and agree with the experimental data, however the agreement with the experimental observations is a bit larger with the density functional B3LYP-D3. Therefore, we believe that the results from B3LYP-3/ LACV3P**++ method are more reliable.
Hydrogenation of 4. DFT calculations on the reaction of complex 4 with dihydrogen confirmed the experimental observation that complexes 5 and 6 are in equilibrium depending on the temperature and hydrogen-pressure used in the reaction ( Figure 2). The formation of complex 6 occurs, through H 2  addition to complex 4 forming the Ir(III) complex 5. After this, IrH insertion into the double bond of the pincer ligand leads to the intermediate 8. This Ir(III) complex undergoes an oxidative addition to form the experimentally observed Ir(V) complex 6. The activation energy barrier for this conversion is calculated to be 20.5 kcal/mol via TS8-5.The calculations showed that complex 6 is more stable than intermediate 8 (under dihydrogen atmosphere). Depending on the choice of functional to calculate the electronic energy, (a) M06, (b) M06-L, or (c) B3LYP-D3, they all favor formation of complex 6 vs complex 8, however the preference for 6 over 8 is a bit larger with B3LYP-D3 in agreement with the experimental observations. The equilibrium between 5 and 6 appears to be most accurate with B3LYP-D3 followed by M06 and M06-L ( Figure S1).
. H2   Different mechanistic scenarios were considered for the formation of complex 7 at the experimental conditions (5 atm of H 2 , 140 ºC) ( Figure S2). The green and black paths of Figure S2 shows how complex 5, via intermediate 8 undergoes an α-alkyl elimination to form carbene 9. The activation barrier is calculated at 31.4 kcal/mol via transition state TS8-9 (See Figure S3 for a more detailed picture of transition states). Complex 9 proceeds through an α-insertion to form intermediate 10, which under hydrogen atmosphere conditions undergoes a H 2 addition that finally forms the more stable Ir(V) complex 11. The transition state TS9-10 is calculated at 25.8 kcal/mol on the free energy surface. This step is followed by CH reductive elimination from IrC sp3 and Ir-H, which leads to formation of a new C sp3 H bond in the pincer ligand, a phosphorusbound tert-butyl group. The free energy barrier via transition state TS11-12 is calculated to be 11.8 kcal/mol. Finally the process finishes with an oxidative addition of one dihydrogen molecule leading to Ir(V)H 4 complex 7. The overall activation free energy corresponds to the CC bond cleavage step (31.4 kcal/mol). Note that the results that are reported here have been calculated at 1 atm of H 2 , and the experimental were carried out at 5 atm. The effect of the pressure can affect the total values around ≈1 kcal. Other mechanistic proposals are evaluated in Figure S2 for the formation of complex 7: (a) through complex 22, and (b) through complex 13. (A) Carbene 9 can undergo a reductive elimination forming the CH bond and generating Ir complex 22. This transition state has a free energy barrier of 41.0 kcal/mol. By comparing with the green αinsertion pathway, it is clear that this path is essentially impossible under the experimental conditions, and the reaction favorably goes via transition state TS9-10. (B) Through complex 13 (figure 2b), there are two possible alternative pathways. From S14 complex 8, a reductive elimination step can occur to generate intermediate 13 with a plausible free energy barrier for the transition state (23.7 kcal/mol), followed by a CC oxidative addition that leads to complex 10. The free energy of the transition state TS13-10 is 53.6 kcal/mol. Alternatively, from complex 13, an oxidative addition of a dihydrogen molecule to Ir can occur to form complex 14 which undergoes CC oxidative addition generating complex 11. The activation energy barrier for this pathway is even higher at 62.9 kcal/mol via TS14-11. It can clearly be concluded, similarly to the statement above, that the proposal the green α-insertion pathway is the most plausible pathway due to the significantly lower free energy barriers, (TS9-8 : 31.4 kcal/mol) vs (TS13-10 : 53.6 kcal/mol) and (TS14-11 : 62.9 kcal/mol).    Carbon-carbon bond formation. As mentioned, it was found experimentally that the formation of complex 7 is reversible, proceeding in one direction or other depending on the reaction conditions. In the presence of a hydrogen atmosphere, complex 7 is obtained quantitatively, and in the presence of a hydrogen acceptor, such as tertbutylethylene, the forward reaction occurs leading to the olefin complex 15 ( Figure S7). Complex 12 proceeds through a CH oxidative addition process forming Ir(V) complex 11. The free energy barrier via transition state TS12-11 is calculated to be 25.6 kcal/mol. This step is the only one that is feasible for both conditions (under H 2 and in the presence of tert-butylethylene). In the reaction with tert-butylethylene, complex 11 goes through reductive elimination followed by α-elimination forming iridium complex carbene 9. The activation barrier is calculated at 22.6 kcal/mol via transition state TS10-9. Complex 9 via CC coupling transition state TS9-8 (28.0 kcal/mol) leads to the formation of intermediate 8, which proceeds through β-elimination and finally reductive elimination to generate the more stable complex 15. The activation barrier is calculated at 17.0 kcal/mol via transition state TS8-5. In the reaction with tert-butylethylene, the overall activation energy barrier corresponds to the CC bond formation step (TS9-8) that is 28.0 kcal/mol. The forward reaction is not possible under hydrogen pressure due to the higher energy barriers. Under hydrogen pressure, the formation of complex 5 needs to cross over a free energy barrier of 45.7 kcal/mol (TS9-8), that makes this CC coupling step unfavorable under these conditions.              NBO studies of carbenes: With the aim of evaluating the nature of the iridium carbene complexes, NBO calculations 13 were made. The results are given in Table S1. The NBO analysis data exhibit typical scheme for the Fischer complexes; 16 and 22 have a Ir-C carbene  and  bond, while 9 has only  bond. The Ir-C carbene bond of complexes 16 and 22 are clearly polarized towards the carbon end (only 40% and 42%, respectively, are at the iridium end) while the  bond is polarized towards the metal moiety (65% and 64%). The NBO bonding pattern suggest that the Ir-C  bond is even more polarized towards the metal moiety because the optimal Lewis structure has a iridium lone-pair d() orbital rather than a  bond. The calculated hybridization shows that the  bond has mainly d character at the iridium end (65%d and 63%d), while the  bond at iridium is purely d(Ir) (97% and 94%). Note that the and electrons in Schrock complexes are more polarized towards the carbon end. In the case of the carbene 9, the calculated polarization of the Ir-C carbene bonds are more equally distributed (57%). The NBO results for the complexes show that the formal charge are between -0.04 and +0.06. This is gratifying, because typically the metal of the Schrock complexes is more positively charged and has a lower 5d population than the Fischer complexes. 14 As a summary, all the complexes showed very different NBO results from Schrock complexes, and several similarities with Fischer complexes.