Platinum-Catalyzed Allylic C–H Alkylation with Malononitriles
Lian‐Feng Fan, Pei‐Pei Xie, Pu‐Sheng Wang, Xin Hong, Liu‐Zhu Gong
Abstract
Open AccessCCS ChemistryRESEARCH ARTICLE1 Apr 2022Platinum-Catalyzed Allylic C–H Alkylation with Malononitriles Lian-Feng Fan†, Pei-Pei Xie†, Pu-Sheng Wang, Xin Hong and Liu-Zhu Gong Lian-Feng Fan† Hefei National Laboratory for Physical Sciences at the Microscale, Department of Chemistry, University of Science and Technology of China, Hefei 230026 , Pei-Pei Xie† Department of Chemistry, Zhejiang University, Hangzhou 310027 , Pu-Sheng Wang *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] E-mail Address: [email protected] Hefei National Laboratory for Physical Sciences at the Microscale, Department of Chemistry, University of Science and Technology of China, Hefei 230026 , Xin Hong *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] E-mail Address: [email protected] Department of Chemistry, Zhejiang University, Hangzhou 310027 and Liu-Zhu Gong *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] E-mail Address: [email protected] Hefei National Laboratory for Physical Sciences at the Microscale, Department of Chemistry, University of Science and Technology of China, Hefei 230026 Center for Excellence in Molecular Synthesis of Chinese Academy of Sciences, Hefei 230026 https://doi.org/10.31635/ccschem.021.202100884 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail The direct functionalization of allylic C–H bonds is distinguished by enabling rapid assembly of structural complexity from simple molecules. Although Pd-catalyzed allylic C–H functionalization has been extensively studied, the discovery of new catalytic systems remains fairly underdeveloped. Here, we disclose a Pt-catalyzed allylic C–H alkylation of a wide range of α-alkenes by using phosphoramidites as ligands and malononitriles as alkylating reagents. Notably, the combination of chiral urea-catalyzed Michael addition and Pt-catalyzed allylic C–H alkylation can serve as an efficient protocol to access chiral tetrahydropyran with high levels of diastereo- and enantioselectivity. Mechanistic studies suggest that the Pt-catalyzed allylic C–H activation proceeds through a concerted proton and two-electron transfer process, which is analogous to transition state geometries of Pd catalysis. Download figure Download PowerPoint Introduction The direct functionalization of inert C(sp3)−H bonds, previously viewed as an entirely new perspective in comparison with the conventional logic of organic synthesis,1–4 now still represents an ultimate goal in synthetic chemistry, allowing for rapid assembly of densely functionalized molecules from simple molecules. Among the landmark advances with transition-metal catalysis,5,6 palladium (Pd)-catalyzed allylic C–H functionalization of readily available α-alkenes is distinguished by enabling the access to alkene-bearing structurally complex molecules with versatile bond-forming capacity and minimal manipulation of functional groups,7–9 thereby being regarded as an atom- and step-economic alternative to Tsuji–Trost allylation reactions by avoiding the use of preoxidized allylating reagents.10,11 In recent decades, our group12–24 and others25–28 have found that trivalent phosphorus ligands are capable of facilitating Pd-catalyzed allylic C–H cleavage through a concerted proton and two-electron transfer process (Scheme 1),22 thereby generating electrophilic π-allylpalladium intermediates that can be leveraged to forge chemical bonds with a wide range of nucleophiles.29 Upon using p-quinone as an oxidant, the 16-electron Pd(0) complex 1 bearing a phosphorus ligand, a quinone, and an α-alkene is most likely to be a key intermediate for allylic C–H activation.22 Although the Pd catalysis has been extensively studied, the feasibility of other transition-metal-based catalyst systems to drive the allylic C–H activation remains unclear and fairly underdeveloped.30–33 Scheme 1 | Allylic C–H cleavage through a concerted proton and two-electron transfer process. Download figure Download PowerPoint Since the discovery of the Shilov catalyst system,34 platinum (Pt) complexes have played a significant role in the fundamental understanding of inert C–H bond activation,35 and therefore, a large number of structurally diverse Pt complexes have been synthesized for the discovery of potential catalysts.36 A 16-electron Pt(0) complex 3, first prepared and identified by Chetcuti et al. in 1981,37 is coordinatively unsaturated but isolable, and thus holds the potential for additional ligand coordination and exchange. In terms of geometrical pattern, the structure of 3 is very similar to Pd(0) complex 1, a precursor for allylic C–H activation. As such, we envisioned that a ligand exchange of ethylene in 3 with α-alkene might also occur to generate an active precursor that can undergo allylic C–H cleavage via a concerted proton and two-electron transfer process, leading to a π-allylplatinum intermediate that might be able to couple with a suitable nucleophile resembling those of the Pd catalysis (Scheme 1). Herein, we introduce an unprecedented Pt-catalyzed allylic C–H alkylation of a wide range of α-alkenes occurring via a concerted proton and two-electron transfer process, enabling rapid assembly of alkene-bearing structurally complex compounds from simple starting materials. Experimental Methods General procedure for allylic C−H alkylation with allyl ethers To a flame-dried and N2-purged Schlenk tube (10 mL) were added malononitrile (0.2 mmol), phosphoramidite L2 (6 mol %), thymoquinone (1.1 equiv), and a stirring bar. The Schlenk tube was then evacuated and filled with N2. This cycle was repeated three times and followed by the addition of Pt(cod)Me2 (5 mol %) in the glovebox; then toluene (0.4 mL) and allyl ether (0.24 mmol) were added. The mixture was stirred at 110 °C for 21 h. Then the solvent was removed under vacuum, and the crude residue was purified by column chromatography on silica gel to provide the products. The detailed experimental methods are available in the Supporting Information. Results and Discussion Computational studies of concerted proton and two-electron transfer process with Pd and Pt catalysis Initially, a series of density functional theory (DFT) calculations were conducted to evaluate the feasibility of Pt-catalyzed allylic C–H activation through a concerted proton and two-electron transfer process. A set of precursors and C–H activation transition states ( TS-2) for both Pd and Pt catalysis were computed (Scheme 2). By employing 1-butene as a model substrate, a very similar geometry of the precursors and transition states for Pd and Pt catalysis was revealed. The root-mean-square deviations (RMSDs) of overlaid precursors and transition states of Pd and Pt catalysis are only 0.0745 and 0.0544 Å, respectively. The high similarity suggests that the Pt catalysis can undergo a similar C–H activation mode to the Pd catalysis, and this deduction can be further corroborated by the comparable activation barrier (19.6 kcal/mol for Pd catalysis and 19.9 kcal/mol for Pt catalysis). In addition, for both Pd and Pt catalysis, the calculated activation energy is correlated to the calculated heterolytic bond dissociation energy (HBDE, H as cation) of the allylic C–H bond (see Supporting Information Tables S4–S9). These results indicate that the manipulation of Pt-catalyzed allylic C−H functionalization is truly feasible. Scheme 2 | Calculated activation energy of allylic C–H cleavage with Pd and Pt catalysis.a Bonds of Pd complex are red and Pt complex are blue. aB3LYP-D3(BJ)/6-311+G(d,p)-SDD-SMD(toluene)//B3LYP-D3(BJ)/6-31G(d)-LANL2DZ. Download figure Download PowerPoint Pt-catalyzed allylic C−H alkylation of α-alkenes with malononitriles In consideration of the relatively high activation energy of allyl ether, the initial experimental investigation was focused on a model reaction of malononitrile 4 and allyl benzyl ether 5 using phosphoramidite L2 as ligand and thymoquinone (TMQ) as oxidant at 110 °C (Table 1). The examination of several commercially available Pt precursors showed that Pt(II) species bearing coordinated anions (entries 1 and 2) were ineffective at driving the conversion of allyl benzyl ether, while Pt(cod)Me2 (entry 3), a potential Pt(0) precursor via carbon–carbon reductive elimination of ethane,38–40 afforded the desired product 6 in 63% yield. Then a series of structurally different monodentate phosphorus ligands were evaluated. Interestingly, PPh3 (entry 4), a common ligand in Pd-catalyzed allylic C–H functionalization,17,25 proved to be invalid with Pt catalysis. Regarding H8-BINOL-derived (BINOL = [1,1′]-Binaphth-2-ol) phosphoramidite ligands, the presence of electron-deficient aryl substituents at the 3,3′-positions (entry 5) was essential to maintain the reaction performance. The presence of relatively electron-rich aryl substituents or the absence of aryl substituents significantly eroded the catalytic efficiency (entries 6 and 7). In addition, the alteration of ligand backbone from H8-BINOL to BINOL also resulted in much diminished yield (entry 8). Fine-tuning of the p-quinone oxidants showed that TMQ with relatively large steric hindrance was superior for enhancing the reaction performance (entries 9 and 10). Furthermore, higher concentration was found to be beneficial to provide the desired product 6 (entries 11 and 12), and a 76% isolated yield was afforded at 0.5 M concentration (entry 12). Table 1 | Optimization of Reaction Conditionsa Entry [Pt] L [BQ] Yieldb 1 Pt(acac)2 L2 TMQ <5 2 Pt(cod)Cl2 L2 TMQ <5 3 Pt(cod)Me2 L2 TMQ 63 4 Pt(cod)Me2 PPh3 TMQ <5 5 Pt(cod)Me2 L3 TMQ 63 6 Pt(cod)Me2 L4 TMQ 10 7 Pt(cod)Me2 L5 TMQ 11 8 Pt(cod)Me2 L6 TMQ 25 9 Pt(cod)Me2 L2 2,5-DMBQ 42 10 Pt(cod)Me2 L2 2,6-DMBQ 50 11c Pt(cod)Me2 L2 TMQ 69 12d Pt(cod)Me2 L2 TMQ 79 (76e) Note: DMBQ, dimethyl-1,4-benzoquinone. aReaction conditions: 4 (0.1 mmol), 5 (0.12 mmol), [Pt] (0.005 mmol), L (0.006 mmol), [BQ] (0.11 mmol), toluene (0.5 mL), 110 °C, under nitrogen. b/l < 20:1. bDetermined by 1H NMR analysis of the crude products based on 1,3,5-triacetylbenzene as the internal standard. cWith 0.3 mL toluene. dWith 0.2 mL toluene. eIsolated yield. Under the optimized reaction conditions, the generality of malononitriles was first explored (Scheme 3). A wide array of monosubstituted malononitriles participated in this protocol to furnish the corresponding allylation products in moderate to good yields. Benzyl and substituted benzyl groups bearing either electron-donating or electron-withdrawing substituents at the para-, meta-, or ortho-position ( 7– 15) were well tolerated. It was worth noting that medicinally relevant heterocycles such as furan ( 16) and thiophene ( 17) could undergo the allylation efficiently with good results. In addition, long-chain alkyl ( 18), alkyl ether ( 19), and allyl groups ( 20 and 21) were also compatible with the reaction. Scheme 3 | Substrate scope of α-alkenes.aa Condition A: 4 (0.2 mmol), alkene (0.24 mmol), Pt(cod)Me2 (0.01 mmol), L2 (0.012 mmol), TMQ (0.22 mmol), toluene (0.4 mL), 110 °C, under nitrogen, isolated yield. Condition B: 4 (0.2 mmol), alkene (0.24 mmol), Pt(cod)Me2 (0.01 mmol), L3 (0.012 mmol), 2,5-DMBQ (0.22 mmol), toluene (1.0 mL), 110 °C, under nitrogen, isolated yield. For the majority of monosubstituted malononitriles, condition A was applied. b2,5-DMBQ was used. Download figure Download PowerPoint Then the substrate scope with respect to α-alkenes was examined (Scheme 3). A variety of allyl ethers with diverse steric properties were evaluated, furnishing the corresponding branched products in good yields and with excellent regioselectivities. Allyl benzyl ethers with various substituents like methoxy, ester, trifluoromethyl, and bromo groups, irrespective of their positions ( 22– 25), were nicely tolerated. Notably, five-membered heterocyclic groups ( 26 and 27) were compatible with this oxidative condition. In addition, allyl alkyl ethers containing various functional handles on the alkyl moiety, such as long-chain alkyl, chloride, bromide, hydroxyl, acetoxy, silyl ether, cyclic ether, ketal, amide, and carbamate groups ( 28– 37), were also viable to furnish the desired products. Unaltered reaction efficiency, albeit poor diastereoselectivity, was obtained with allyl alkyl ethers containing tertiary stereocenters ( 38– 40). Although this Pt-catalyzed allylic C−H alkylation reaction still occurred with other α-alkenes, the E/Z- and regioselectivity were susceptible to the type of α-alkenes (Scheme 3). In comparison with allyl ethers, inactivated α-alkenes, without an alkoxy group to stabilize the positive charge of the corresponding π-allylplatinum intermediate,41 provided densely functionalized linear products in moderate to high yields with excellent regioselectivity and moderate E/Z-selectivity. Notably, the minor generation of the thermodynamically unstable E-configuration product might be attributed to a rapid σ/η3/σ interconversion of the π-allyl-Pt complex compared with the rate of nucleophile attack on this intermediate.42 A wide range of functional groups such as substituted phenyl and heterocyclic groups, acetoxy, silyl ether, terminal epoxide, amide, sulfamide, and chloride ( 41– 52) were all nicely tolerated. Encouragingly, a chiral polyoxygenated compound bearing an α-alkene handle ( 53) was also amenable to the direct functionalization of allylic C−H bonds, highlighting the synthetic significance of this protocol toward late-stage functionalization. For allylbenzene, the linear product ( 54) was delivered in excellent yield and E/Z-selectivity, but with much diminished regioselectivity. Notably, the Pt-catalyzed allylic C–H alkylation reaction of malononitrile 4 and allyl ether 5 was feasible for reaction scale-up, providing the allylation product 6 with maintained reaction performance (Scheme 4). Furthermore, the allylation products could undergo versatile late-stage modification to provide structurally diverse alkene-functionalized derivatives (Scheme 4). For instance, the product 6 was readily converted into aldehyde 55 through a Rh-catalyzed hydroformylation. A ring-closing metathesis of 1,6-diene 21 with Grubbs-II catalyst smoothly provided cyclopentene derivative 56. Additionally, the terminal alkene derivative 6 could also be transformed into internal alkene 57 by a Pd-catalyzed Heck coupling reaction. Scheme 4 | Preparative-scale reaction and versatile transformations. Condition A: Rh(acac)(CO)2 (4 mol %), BiPhePhos (4.8 mol %), 1∶1 H2/CO (2 bar), toluene, 50 °C 25 h. Condition B: Grubbs-II (5 mol %), CH2Cl2, 40 °C, 39 h. Condition C: Pd(OAc)2 (10 mol %), PPh3 (20 mol %), Et3N (1.5 equiv), 1-iodo-3-methoxybenzene (2 equiv), dimethylformamide (DMF), 90 °C, 24 h. Download figure Download PowerPoint With these results in hand, an enantioselective allylic C–H alkylation with the use of chiral ligand is highly desirable to provide chiral allylation products. Unfortunately, the attempt to accomplish the enantioselective reaction of malononitrile 4 and allyl ether 5 by screening a broad range of chiral phosphoramidite ligands43–45 provided the allylation product 6 with unsatisfactory (≤45%) enantiomeric excess (ee) (Scheme 5). Alternatively, inspired by our long interest in organo/metal combined catalysis,46–49 a sequential protocol consisting of a chiral urea-catalyzed asymmetric Michael addition,50–54 and the Pt-catalyzed allylic C–H alkylation furnished a variety of chiral tetrahydropyrans ( 59– 63) in moderate yields and with high levels of diastereo- and enantioselectivities (Scheme 5). Scheme 5 | Exploration of asymmetric transformation. aSee Supporting Information Table S1 for reaction development and optimization. bPt(cod)2 and TMQ were used. Download figure Download PowerPoint Studies toward the reaction mechanism and origins of selectivities After showcasing the practical applicability, we turned our attention to the reaction mechanism. Under the optimal reaction conditions, a series of kinetic studies were conducted (Scheme 6). Interestingly, the progress curve of the reaction initiated with Pt(cod)2 ( Supporting Information Figure S1), a highly efficient precursor for the generation of Pt(0) species,55 was completely different from the curve of the reaction that was initiated with Pt(cod)Me2. The reaction starting with Pt(cod)Me2 had a pronounced induction period (<30 min), and the reaction rate was much slower than that of the reaction initiated with Pt(cod)2. These results suggest that Pt(cod)Me2 should also serve as an active Pt(0) catalyst precursor, but the transformation of Pt(cod)Me2 to active Pt(0) complex would be slow. In addition, a significant primary kinetic isotope effect (KIE, kH/kD = 2.5) was determined for the Pt(cod)2-catalyzed allylic C–H alkylation of allyl benzyl ether 5 and deuterated allyl benzyl ether 5- d 2 (Scheme 6; Supporting Information Figure S2), suggesting that the allylic C−H cleavage was involved in the rate-limiting step. Scheme 6 | Kinetic studies. Download figure Download PowerPoint Furthermore, to clarify the activation pathway of Pt(cod)Me2, a series of mechanistic studies were conducted. When the reaction of malononitrile 4, Pt(cod)Me2, and L2 was conducted in toluene at 110 °C for 6 h, CH4 rather than C2H6, was detected as the sole gaseous product in gas chromatography (GC) analysis (Scheme 7a; Supporting Information Figure S3), implying that a direct carbon–carbon reductive elimination of ethane to give Pt(0) species was unfeasible; instead, the malononitrile 4 served as a Brønsted acid to protonate the Pt-methyl bond. Intermediates ( L2)PtIIMe 64, Pt(II) 65, Pt(0) 66, and Pt(0) 67 were detected by high-resolution mass spectrometry (HRMS) analysis of the stoichiometric reaction mixture of malononitrile 4, allyl ether 5, Pt(cod)Me2, L2, and 2,5-DMBQ (DMBQ = dimethyl-1,4-benzoquinone) in toluene at 110 °C for 1.5 h (Scheme 7b; Supporting Information Figure S4). Notably, Pt(0) 67 actually was the key intermediate capable of undergoing allylic C–H cleavage through the concerted proton and two-electron transfer process. Based on these experimental evidences, we proposed that the Pt(II) 68, generated from Pt(cod)Me2 and L2, was initially protonated with malononitrile 4 to provide Pt(II) 65. Then Pt(II) 65 underwent β -H elimination to furnish methylenemalononitrile 69 and Pt(II) 70; meanwhile, an active Pt(0) 71 was furnished through reductive elimination of CH4 (Scheme 7c). Scheme 7 | (a–c) Mechanistic studies. Download figure Download PowerPoint To further elucidate the reaction mechanism at the molecular level, we next studied the catalytic cycle with DFT calculations (Scheme 8; for computational details, see Supporting Information Table S10). The free-energy profile of the operating catalytic cycle with the model phosphoramidite ligand L9 is shown in Scheme 8. The substrate-coordinated Pt(0) complex 72 first isomerizes to the less stable conformer 73, and a subsequent concerted proton and two-electron transfer process occurs via TS-3. This process requires a barrier of 25.9 kcal/mol and is endergonic by 17.7 kcal/mol to generate the allylplatinum(II) species 74. Subsequent coordination of malononitrile 4 leads to the intermediate 75, and the deprotonation of malononitrile through TS-4 generates 76. Intermediate 76 isomerizes to 77 with the nitrogen coordinated to Pt, which allows the C–C reductive elimination through TS-5. The product-coordinated Pt(0) complex 78 eventually liberates the product 6 and regenerates the active catalytic species 72. Based on the DFT-computed free-energy profile, the on-cycle resting state is the substrate-coordinated complex 72. The overall barrier of the catalytic cycle is 25.9 kcal/mol compared with the resting state 72, and the rate-determining step is the concerted proton and two-electron transfer process via TS-3, which is consistent with the experimental kinetic studies (Scheme 6). Scheme 8 | DFT-computed free-energy changes of the most favorable pathway. Download figure Download PowerPoint Conclusion A phosphoramidite-Pt complex-catalyzed allylic C–H alkylation reaction has been developed using malononitrile as alkylating reagents and thymoquinone as oxidant. This protocol can tolerate a wide range of α-alkenes and malononitriles and thus enables rapid assembly of alkene-bearing structurally complex molecules from simple starting materials. In addition, an asymmetric variant has also been established by a sequential process consisting of the chiral urea-catalyzed asymmetric Michael addition and the Pt-catalyzed allylic C–H alkylation, delivering a variety of chiral tetrahydropyrans with high levels of diastereo- and enantioselectivities. Mechanistic studies suggest that the allylic C–H cleavage events proceed via a concerted proton and two-electron transfer process, and the precursors and transition states of allylic C–H activation with either Pd or Pt catalysis are very similar in terms of geometry and activation energies. More importantly, the success of this protocol, to a certain degree, indicates that the concerted proton and two-electron transfer process may be a general activation mode for transition-metal-catalyzed allylic C–H activation, highlighting its potential in the future development of other transition-metal catalysts, especially non-noble metal-based catalyst systems. Supporting Information Supporting Information is available and includes characterization data, detailed experiment procedures, and computational details. Conflict of Interest There is no conflict of interest to report. Acknowledgments The financial support from the National Nature Science Foundation of China (NSFC) (nos. 21831007 and 21672197 for L.-Z.G. and nos. 21702182 and 21873081 for X.H.), the Chinese Academy of Sciences (grant no. XDB20020000) and Youth Innovation Promotion Association CAS, Fundamental Research Funds for the Central Universities (no. 2020XZZX002-02 for X. H.), and the State Key Laboratory of Clean Energy Utilization (no. ZJUCEU2020007 for X.H.) is gratefully acknowledged. Calculations were performed on the high-performance computing system at the department of chemistry, Zhejiang University. References 1. Gutekunst W. R.; Baran P. S.C-H Functionalization Logic in Total Synthesis.Chem. Soc. Rev.2011, 40, 1976–1991. Google in Synthesis Soc. Rev.2011, 40, Google for and Google the and of 5, Google Wang by Google by Google Allylic Functionalization Google 8. Allylic C–H The of Google Wang R.; Allylic Functionalization via Google Allylic Google in Google Wang P. Gong for of with Google X. Gong with by Google Gong a with under the of and Google X. Wang P. Allylic Google X. Wang P. Gong Allylic Alkylation of with Google Wang P. Gong General Allylic Alkylation of Google Wang P. Gong Allylic Alkylation of with by of and Google Wang Wang P. Gong Synthesis of via Allylic of with Google Wang Wang Gong in Allylic C−H Google Wang Wang Gong Allylic C–H Alkylation of with Google P. Wang P. Wang Hong Gong and in the Allylic Alkylation of Google Wang Wang P. Gong Allylic Alkylation of Allyl with Google Wang Wang P. Gong Allylic Alkylation of and China Google Alkylation of by Google and Allylic Alkylation through Google Allylic through Google of for and Allylic of Google Wang P. Gong Allylic and and Google and Allylic Google Allylic Functionalization with a Google C–H Bonds of Google X. of Allylic and Google Shilov in of of Google Google of by Pt Google Chetcuti of and with of Soc. Google into the of the and of and Google of on the of A of from and Google R.; of Upon to the Google of and for the of Google of Allylic Google in Google in Google W. and in Google Gong Reaction of and Google X. Gong with of and Google Gong Google Wang Gong in of with Google Wang Wang to Google to Michael of to Google Wang Michael of to by a A for the Synthesis of a Google of to by Google Michael of to Google of from and to Under Google Information Chinese proton and two-electron financial support from the National Nature Science Foundation of China (NSFC) (nos. 21831007 and 21672197 for L.-Z.G. and nos. 21702182 and 21873081 for X.H.), the Chinese Academy of Sciences (grant no. XDB20020000) and Youth Innovation Promotion Association CAS, Fundamental Research Funds for the Central Universities (no. 2020XZZX002-02 for X. H.), and the State Key Laboratory of Clean Energy Utilization (no. ZJUCEU2020007 for X.H.) is gratefully acknowledged. Calculations were performed on the high-performance computing system at the department of chemistry, Zhejiang University. times