Radical Mechanism of Ir ^III /Ni ^II -Metallaphotoredox-Catalyzed C(sp ³ )–H Functionalization Triggered by Proton-Coupled Electron Transfer: Theoretical Insight

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE1 Apr 2022Radical Mechanism of IrIII/NiII-Metallaphotoredox-Catalyzed C(sp3)–H Functionalization Triggered by Proton-Coupled Electron Transfer: Theoretical Insight Yu-Jiao Dong†, Bo Zhu†, Yun Geng, Zhi-Wen Zhao, Zhong-Min Su and Wei Guan Yu-Jiao Dong† Institute of Functional Material Chemistry, Faculty of Chemistry, Northeast Normal University, Changchun 130024 , Bo Zhu† Institute of Functional Material Chemistry, Faculty of Chemistry, Northeast Normal University, Changchun 130024 , Yun Geng Institute of Functional Material Chemistry, Faculty of Chemistry, Northeast Normal University, Changchun 130024 , Zhi-Wen Zhao Institute of Functional Material Chemistry, Faculty of Chemistry, Northeast Normal University, Changchun 130024 , Zhong-Min Su *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Institute of Functional Material Chemistry, Faculty of Chemistry, Northeast Normal University, Changchun 130024 College of Chemistry, Jilin University, Changchun 130012 and Wei Guan *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Institute of Functional Material Chemistry, Faculty of Chemistry, Northeast Normal University, Changchun 130024 https://doi.org/10.31635/ccschem.021.202100802 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Photoredox catalysis can be induced to activate organic substrates or to modulate the oxidation state of transition-metal catalysts via unique single-electron transfer processes, so as to achieve challenging C(sp3)–H functionalization under mild conditions. However, the specific reaction mechanism and relevant electron transfer process still need to be clarified. Here, a highly regioselective IrIII/NiII-metallaphotoredox-catalyzed hydroalkylation of asymmetrical internal alkyne with an ether α-hetero C(sp3)–H bond has been investigated by density functional theory (DFT) calculations. A novel radical mechanism was predicted to merge oxidative quenching (IrIII–*IrIII–IrIV–IrIII) and nickel catalytic cycles (NiII–NiIII–NiI–NiIII–NiII) for this C(sp3)–H functionalization to construct C(sp3)–C(sp2) bonds. It consists of seven major steps: the single-electron transfer involved in the photoredox cycle for generating active Ni(I)–chloride complexes, proton-coupled electron transfer process to provide α-carbon-centered tetrahydrofuran (THF) radicals, radical capture by Ni(II), reductive elimination to obtain 2-chlorotetrahydrofuran, alkyne oxidative hydrometallation, inner-sphere electron transfer, and σ-bond metathesis to yield the desired alkyne hydroalkylation product. Importantly, both the thermodynamic performance for redox potentials and the kinetic exploration for energy barriers and electron-transfer rates have also been evaluated for the corresponding electron transfer processes. In addition, the steric effects play a major role in determining the regioselectivity of alkyne oxidative hydrometallation. Download figure Download PowerPoint Introduction The functionalization of C–H bonds to construct C–C and C–X (X = N, O, S, etc.) bonds has been recognized as an active field in current organic research.1–3 It is not only due to the readily available starting materials to satisfy the requirement of organic transformations, but also the natural features of C–H bonds broadly existing in organic molecules, such as accessibility, activity, and selectivity.4 Over the past decades, activation of C–H bonds with a wide variety of directing groups has become an economically attractive synthetic strategy, which plays an important role in determining proximal C–H reactivity.5,6 Also, transition-metal catalysts, by Pd,7–9 Rh,10,11 Ru,12,13 Ir,14,15 and Ni16–18 complexes, are crucial for the functionalization of C–H bonds as a versatile and efficient strategy to construct the C–C and C–X bonds in organic synthesis.19–21 These reactions have been realized using the directing groups strategy,22,23 radical addition,24 the concerted metalation–deprotonation,25 and single-electron transfer (SET),26 respectively. It is still a significant challenge to realize regioselective remote C–H functionalization in the absence of directing groups.4 Therefore, it is highly desirable that a novel strategy can remove basic additives and then broaden the scope of substrates under mild conditions. The visible-light photocatalysis,27–29 because of the natural advantages of the visible light, is attractive for developing important and accessible chemical conversions. Specifically, photoredox catalysis developed simultaneously by Nicewicz and MacMillan,30 Yoon et al.,31 and Stephenson et al.32 were applied to the direct asymmetric alkylation of aldehydes, efficient [2 + 2] enone cycloadditions, and SET instructional tin-free reductive dehalogenation reaction, respectively. In comparison with traditional C–H functionalization, the photocatalyst using its redox properties can directly produce reactive radical species from C–H bond substrates under extremely mild conditions.33 Furthermore, the synergistic catalysis of photocatalyst and transition-metal catalyst has been widely reported by Molander, Gutierrez, Doyle, et al.,34–39 and can be traced back to Pd/Ru cooperative catalytic system. It can convert some substrates to reactive radicals as nucleophilic coupling partners that are subsequently involved in the transition-metal-catalyzed cross-coupling reactions. This electron-transfer-mediated catalysis can break through the inherent limitations of the two-electron pattern and then as an innovative catalytic platform expand the scope of the building various C–C/C–X bonds. Based on the interest in theoretical mechanistic studies of iridium/nickel metallaphotoredox-catalyzed C–O, C–S, and C–N cross-couplings,40–42 we further questioned whether the previously proposed oxidation state modulation mechanism merging oxidative quenching and nickel catalytic cycles or an alternative radical mechanism is applicable to C–H functionalization to achieve C–C cross-couplings. Recently, a photoredox-mediated IrIII/NiII dual-catalyzed hydroalkylation of ether α-hetero C(sp3)–H bond has been achieved to construct C(sp3)–C(sp2) bonds, as shown in Scheme 1.43 Compared with transition-metal catalysis, this dual catalysis requires no basic additive and has wide substrate scope and an excellent regioselectivity of alkyne insertion. Based on the radical trapping experiments, a hypothetical radical mechanism consisting of reductive quenching and nickel catalytic cycles had been proposed to understand the dual IrIII/NiII-catalyzed C–H functionalization. Furthermore, a chlorine radical was speculated to be generated by the visible-light excitation of high-oxidation-state nickel(III) chloride, as shown in Scheme 2 (mechanism A). Next the chlorine radical may abstract an α-hydrogen atom of tetrahydrofuran (THF) to generate a key α-carbon-center radical. On the other hand, a unique photocatalytic strategy in which an excited photocatalyst is employed to directly activate substrates, such as R–H (R = C, Si, and S) bonds, by proton-coupled ET (PCET) process, has been proposed in previous studies.44–46 Thus, it is likely that the PCET process of photoexcited species *IrIII and THF may generate α-carbon-centered radical and the ground-state IrII species (mechanism B in Scheme 2). Subsequently, single-electron reduction of NiII catalyst by IrII generates active NiI and IrIII species. Another plausible mechanism is the oxidative quenching cycle (IrIII–*IrIII–IrIV–IrIII) where an active NiI species is generated via oxidative quenching of photoexcited *IrIII by NiII. Next, the oxidized IrIVCl with THF triggers a PCET process to afford α-carbon-centered THF radical (mechanism C in Scheme 2). Indeed, the generated chlorine radical as an electrophilic species can abstract hydrogen from the electron-rich α-C(sp3)–H bond of ethers via direct hydrogen atom transfer (HAT) using the polar effects.18,47,48 However, the chlorine photoelimination from a charge-transfer excited state of mononuclear nickel(III) complex49 may inevitably trigger a multistep prereaction process. Considering the successive HAT and other steps involved in the C–H functionalization reaction, a high energy barrier might need to be overcome to construct the desired C–C bond. Therefore, how the radical is generated in this radical mechanism of the C–H functionalization by such a dual catalysis still remains ambiguous but a quite crucial issue. In this regard, theoretical calculations complemented with experimental observations can predict enough short-lived reactive intermediates and transition states, and thus disclose the rational reaction mechanism in a more advantageous approach.5,28,50–68 Here, we want to investigate the existing photoredox catalysis, clarify reasonable mechanism and propose the meaningful C–H activation modes for constructing C–C bond. More importantly, we will solve some problems as follows: (1) How do we understand the PCET process induced by the quenching of the excited photocatalyst? (2) What is the origin of regioselectivity in this asymmetrical hydroalkylation reaction? Scheme 1 | Photoredox-mediated IrIII/NiII dual-catalyzed asymmetrical hydroalkylation of ether α-hetero C(sp3)–H bond to achieve C–C cross-coupling. Download figure Download PowerPoint Scheme 2 | Three possible mechanisms regarding α-C(sp3)–H functionalization of THF. Download figure Download PowerPoint Methods All calculations were carried out with the Gaussian 09 package69 using density functional theory (DFT) functional (U)M06.70 Geometry optimizations, together with frequency analyses, were used to verify whether the stationary points are minima without imaginary frequencies or a transition state with only one imaginary frequency conducted in the gas phase to acquire the thermodynamic properties at 333.15 K and 1 atm. Intrinsic reaction coordinate (IRC)71,72 analyses were performed to identify the transition states to connect the correct reactants and products. Here, a mixed basis set approach was employed with LanL2DZ73 for Ir and Ni, 6-31++G(d,p)74 for the α-hydrogen atom of THF, and 6-31G(d)75 for the other main-group elements, respectively. The single-point energies of all stationary points were performed at the (U)M06/[6-311++G(d,p)/SDD76(Ir and Ni)] level. The solvent effect of THF in single-point energies was considered by the SMD77 solvation model. In addition, the translational entropy was corrected with the method developed by Whitesides et al.78 (see Supporting Information for computational details). In this work, all of the DFT calculations employ the default pruned numerical integration grid. To avoid numerical errors in both SCF and frequency calculations, two kinds of pruned numerical integration grids, fine (75, 302) and ultrafine (99, 590), have been used to evaluate the energy barrier of rate-determining step. The result shows that the integral grid has little influence on the energy barriers. In the present calculations, the photoredox-mediated IrIII/NiII dual-catalyzed C–C cross-coupling of THF and unsymmetrical alkynes (3,3-dimethylbut-1-yn-1-yl) benzene ( AL) were selected as the model reactions, where IrIII[dF(CF3)ppy]2(dtbbpy)PF6 was adopted as photocatalyst and abbreviated as Ir III (Scheme 1). Considering the computational cost, the realistic transition-metal catalyst NiII(dtbbpy)Cl2 was simplified to NiII(bpy)Cl2 ( Ni II). In addition, the absorption spectra of Ir III and Ni II were simulated by the time-dependent DFT (TDDFT) calculation at the SMD(THF)/M06/[6-31G(d)/LanL2DZ(Ir and Ni)] level to validate the computational rationality ( Supporting Information Figures S1 and S2). Note that, when the realistic ligand dtbbpy was employed instead of the simplified ligand bpy, the Gibbs activation energy of the rate-determining step does not differ very much, indicating the rationality of simplification ( Supporting Information Table S2). Considering the diversity of oxidation state and spin state of nickel, such as the planar singlet (SNiII) and tetrahedral triplet (TNiII) states of Ni(II) and the doublet (DNiI/III) and quartet (QNiI/III) states of Ni(I) and Ni(III), all species were discussed at the lowest energy spin state in the nickel catalytic cycles ( Supporting Information Tables S3 and S4 and Figure S4). In addition, to evaluate the effects of gas- and solvent-phase optimizations on the ion migration involved in the SET process, we have compared the Gibbs free energies of the key SET processes at two levels ( Supporting Information Figure S5). The calculated results show that the solvent effect of THF has little influence on their gas-phase optimized geometries. Results and Discussion The photomediated α-C(sp3)–H functionalization of THF Generation of α-carbon-centered radical Radical trapping experiments indicate that α-carbon-centered THF radicals participate in the reaction process.43 However, the generation pathway of α-carbon-centered radical needs to be clarified. As described by the above strategies, three generation pathways of α-carbon-centered radical have been examined here (Figure 1). In mechanism A, although T Ni II may be oxidized by photoexcited * Ir III according to the redox potential comparisons ( Supporting Information Table S5), the reductive quenching process of * Ir III with T Ni II to afford Ir II and D Ni III species is predicated to be thermodynamically unfavorable with the Gibbs free energy change (ΔG°) of 13.5 kcal/mol. This reversal is speculated to be due to arduous PF6− migration between iridium and nickel to remain electrically neutral. Such inert photoredox-mediated excited-state SET process has been difficult enough to trigger the next step,31,79 and the photolysis of resultant oxidative product D Ni III and following HAT require ΔG° values of −4.5 and 7.9 kcal/mol, respectively, to deliver the α-carbon-centered THF radical. Thus, such high endothermic reactions induced by the above SET and HAT may not be the most favorable mechanism to produce the carbon-centered radical. Figure 1 | The Gibbs free energy change (ΔG°333.15 in kcal/mol) of selected key steps for generating α-carbon-centered radical in mechanisms A, B, and C. Spin densities for the main intermediates and transition states are given in grey italic font. Download figure Download PowerPoint In comparison with mechanism A, * Ir III could be reductively quenched by THF instead of T Ni II to provide an α-carbon-centered radical and Ir II in mechanism B. However, the whole PCET process is still highly endoergic at 32.0 kcal/mol, and mechanism B can be ruled out. It is noteworthy that the excited photocatalyst may simultaneously serve as a strong 1e-oxidant and -reductant, so the present photoredox catalysis could adopt an oxidative quenching mechanism, in addition to the above two reductive quenching mechanisms. In mechanism C, photoexcited * Ir III is oxidatively quenched by T Ni II via the SET process accompanied by the facile Cl− migration to generate the ground-state Ir IVCl and D Ni I with a ΔG° value of −5.6 kcal/mol, although the difference between E1/2red[IrIV/*IrIII] and E1/2red[NiII/NiI] is very small ( Supporting Information Table S5). Next, a stepwise PCET process between THF and Ir IVCl can release the desired α-carbon-centered THF radical (rather than β-carbon-centered THF radical, Supporting Information Figure S6) and regenerate Ir III to restart the photocatalytic cycle with ΔG° and ΔG°‡ values of −0.6 and 4.8 kcal/mol. Note that the dissociation of HCl from the THF•⋯HCl complex requires a ΔG° value of 4.1 kcal/mol and the HCl is more likely to interact with the α-C center rather than the O of THF• ( Supporting Information Figure S6c). In this stepwise PCET process, the proton transfer will occur prior to the ET through the redox potential comparisons ( Supporting Information Tables S5 and S6). Therefore, both ET steps are thermodynamically accessible for modulating the oxidation state of nickel catalyst and affording the α-carbon-centered radical. Kinetic evaluation for SET processes To further verify the rationality of the mechanism C, the kinetic exploration of ET processes involved in the above oxidative quenching cycle has been performed. The energy profiles of these processes are intuitive in Figure 2. Optimized structures for selected important stationary points in the energy profiles are shown in Supporting Information Scheme S1. First, the calculated absorption spectrum of Ir III is in good agreement with the experimental one and supports the robustness of our theoretical methods ( Supporting Information Figure S1). The strong absorption bands calculated at 292, 300, 341, and 372 nm mainly involve the charge-transfer transitions from the d orbital of iridium center to the ligands (denoted as S0 → SMLCT), localized π–π* transitions within the same phenylpyridine or bipyridine ligand, and the charge-transfer transitions from phenylpyridine ligands to bipyridine ligand ( Supporting Information Figure S3 and Table S1). Thereby, upon excitation by light, Ir III was initially excited to the Frank–Condon region mainly characterized as metal-to-ligand charge transfer (MLCT) and LCT states, as illustrated in Figure 2a. Subsequently, photoexcited * Ir III rapidly relaxes to the minimum state (S1) through internal conversion according to Kasha's rule. Then the intersystem crossing between a singlet and a triplet electronic state occurs through the singlet–triplet crossing (STC) and the resultant TMLCT relaxes to its minimum state (T1). It is worth mentioning that there is competition between phosphorescence emission and SET from this T1 state. Comparing the rates of phosphorescent radiation (red line, kp = 2.35 × 104 s−1) and SET oxidative quenching (blue line, kSET = 2.22 × 1013 s−1) based on the value of reorganization energy 1, the SET process occurs much more quickly than the emission route; see Supporting Information for computational details of rate calculations. Based on this discovery, the ET processes are expected to play a crucial role in modulating the oxidation state of nickel and activating α-hetero C(sp3)–H bond of THF. Starting from the T1 state of Ir III, the SET oxidative quenching is instantaneously triggered by interaction with the ground state of T Ni II. Next, the excited-state charge-transfer * Ir III/ Ni II complex A (exciplex) is formed. The dimer configurations for * Ir III/ Ni II were initially screened out from the conformers with ≥3% probability in the Boltzmann distribution80 ( Supporting Information Scheme S2), and then the electron coupling values were calculated based on these conformers. To guarantee the ET from * Ir III to T Ni II, a relatively large electron coupling value of 0.058 eV was chosen to discuss the corresponding SET process, as shown in Figure 2b. A shorter π–π stacking distance between phenylpyridine and bipyridine ligands (3.47 Å) can be observed in such optimized dimer complex, which also effectively promotes the SET through stronger electronic coupling. Besides, the small reorganization energies with 0.13 and 0.20 eV could promote the rate of SET process. With structural relaxation and charge redistribution, the exciplex overcomes a very low energy barrier of 0.9 kcal/mol to afford Ir IVCl and D Ni I complexes (blue line in Figure 2a). Note that the energy transfer between T Ni II and * Ir III should not occur because of no overlap between the emission spectrum of Ir III and UV-absorption spectrum of T Ni II.41 Similarly, the PCET process between the resultant Ir IVCl and newcomer THF occurs with an energy barrier of 4.8 kcal/mol, the reorganization energies of 0.14 and 0.42 eV, and the electron-transfer rate of 2.64 × 1010 s−1, respectively. The isomers of THF/ Ir IVCl dimer complex have been also examined ( Supporting Information Scheme S2). Finally, the α-hetero C(sp3)–H bond of THF can be successfully cleaved to afford the desired α-carbon-centered THF radical, and the ground-state Ir III is regenerated. In addition, two nucleophilic substrates N,N′-dimethylacetamide (DMA) and dioxane employed in the experiments were chosen to confirm the feasibility of the mechanism C ( Supporting Information Figure S7). In brief, the ET processes involved in the oxidative quenching cycle have been supported by the kinetic exploration for energy barriers and electron-transfer rates. At present, mechanism C seems to be the only acceptable one. Figure 2 | (a) Energy profiles of ET processes in the oxidative quenching cycle. The kinetic calculating of reorganization energy λ1 and λ2 (eV), electronic free energy change ΔE (kcal/mol), Gibbs activation energy ΔG°⧧ (kcal/mol) and the charge-transfer rate k (s−1). The key characteristic transition points are schematically shown with their orbitals and key distances in Å. (b) The isomers of *Ir III/Ni II complex (exciplex) with the value of their electronic coupling Download figure Download PowerPoint The most favorable catalytic cycle The whole nickel catalytic cycle In addition to the iridium photoredox the possible nickel catalytic cycles be In of the previous studies alkyne three nickel catalytic cycles (Scheme have been to understand the present C–C cross-coupling reaction of THF and asymmetrical which were evaluated through DFT calculations in Figures and ( Supporting Information Table and Figure Scheme | Three proposed mechanisms regarding photoredox-mediated Ir III/Ni II dual-catalyzed hydroalkylation to achieve C–C cross-coupling. Download figure Download PowerPoint Figure | Energy profiles (ΔG°333.15 in kcal/mol) of the most favorable nickel catalytic cycle. bond distances and spin densities are given in and grey italic respectively. Download figure Download PowerPoint Figure | Energy profiles (ΔG°333.15 in kcal/mol) of the two possible and bond distances and spin densities are given in and grey italic respectively. Download figure Download PowerPoint As illustrated in Figure the α-carbon-centered electrophilic THF radical be rapidly by the triplet Ni(II) complex 1 to generate a doublet species 2 with a ΔG° value of kcal/mol ( Supporting Information Figure the bond via reductive elimination occurs from 2 to afford the through the with a small ΔG°⧧ value of kcal/mol. The of with the experimental Next, the oxidative between and occurs via the to yield the alkyne In the distance is to from in the and bonds are to and respectively, indicating the bond is cleaved and the and bonds are The ΔG°⧧ and ΔG° values of this oxidative step are and kcal/mol, respectively. Note that oxidative is more favorable than the alternative stepwise mechanism consisting of the oxidative addition of Ni(I) to HCl and the alkyne to bond ( Supporting Information Figure Based on the hydrogen of the alkyne hydroalkylation product is expected to be from both the THF C(sp3)–H bonds and the following proton with in the reaction In the present calculations, HCl is generated with the of α-carbon-centered THF radical. It can be expected that in THF•⋯HCl complex or intermediates HCl and could with to the following hydrogen in the addition to In the reaction, a mixed complex is initially with a ΔG° value of kcal/mol, in which and NiI are by a the dimer a radical migration from to Ni(II) occurs via the to afford the and regenerate an Ni(II) catalyst Such reaction can be also as inner-sphere The ΔG°⧧ and ΔG° values of this inner-sphere ET are and kcal/mol to respectively. In the σ-bond spin between the singlet and triplet energy profiles effectively the activation barrier and the ground-state from the singlet state to the triplet state via a minimum energy crossing In the favorable the σ-bond metathesis between and bonds occurs via the and the triplet to afford the alkyne hydroalkylation product and regenerate the other triplet Ni(II) catalyst The singlet transition state of σ-bond metathesis has also been examined to be unfavorable than the present triplet one ( Supporting Information Figure In the of the and bonds are to and from in and in the and distances are and respectively, indicating the bond and the bond occur The ΔG°⧧ and ΔG° values of the rate-determining σ-bond metathesis process are and kcal/mol, respectively, which can be overcome under the experimental In addition, by the favorable SET from photoexcited * Ir III to T Ni II D Ni I (Figure an alternative oxidative quenching of * Ir III by has also been evaluated to be unfavorable to provide Ir IVCl and species. 1 ( T Ni simultaneously with is much and to the *Ir III than by the energy barrier and redox potentials ( Supporting Information Figures and although the oxidative addition of to species and reductive elimination from nickel(III) require energy barriers. As the of the above mechanism the direct radical capture by D Ni I and the alkyne with D Ni I could to possible inner-sphere 2 and as shown in Figure ( Supporting Information Figure Note that the following results are discussed based on the Gibbs energy calculated on the triplet energy because the singlet profiles than the triplet ( Supporting Information Figure Starting from the active D Ni the α-carbon-centered THF radical could be by D Ni I to generate a relatively Ni(II) species or coordinate with to from a Ni(I) with the ΔG° values of and kcal/mol, respectively. In is the bond of through a to afford an The ΔG°⧧ and ΔG° values of this alkyne are and kcal/mol, respectively. In the electrophilic THF radical to the of could occur via the to afford ( Supporting Information Figure This alkyne step requires the ΔG°⧧ and ΔG° values of and kcal/mol, respectively. At a σ-bond metathesis between and HCl could yield the alkyne hydroalkylation product and regenerate the complex However, the regioselectivity of alkyne in was examined to be to the experimental results (Scheme 2 and are favorable than 1 because than 1 based on the Gibbs energy calculated on the potential energy Scheme | The of selected alkyne step. The Gibbs free energies are in kcal/mol. experimental values were from Download figure Download PowerPoint of alkyne oxidative As shown in Scheme 1, the hydroalkylation reaction of alkynes excellent regioselectivity of

Topics & Concepts