New Mechanistic Insights on the Selectivity of Transition-Metal

Article pubs.acs.org/accounts

New Mechanistic Insights on the Selectivity of Transition-MetalCatalyzed Organic Reactions: The Role of Computational Chemistry Published as part of the Accounts of Chemical Research special issue “Computational Catalysis for Organic Synthesis”. Xinhao Zhang,*,† Lung Wa Chung,*,‡ and Yun-Dong Wu*,†,§ †

Lab of Computational Chemistry and Drug Design, Laboratory of Chemical Genomics, Peking University Shenzhen Graduate School, Shenzhen 518055, China ‡ Department of Chemistry, South University of Science and Technology of China, Shenzhen 518055, China § College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China CONSPECTUS: With new advances in theoretical methods and increased computational power, applications of computational chemistry are becoming practical and routine in many fields of chemistry. In organic chemistry, computational chemistry plays an indispensable role in elucidating reaction mechanisms and the origins of various selectivities, such as chemo-, regio-, and stereoselectivities. Consequently, mechanistic understanding improves synthesis and assists in the rational design of new catalysts. In this Account, we present some of our recent works to illustrate how computational chemistry provides new mechanistic insights for improvement of the selectivities of several organic reactions. These examples include not only explanations for the existing experimental observations, but also predictions which were subsequently verified experimentally. This Account consists of three sections discuss three different kinds of selectivities. The first section discusses the regio- and stereoselectivities of hydrosilylations of alkynes, mainly catalyzed by [Cp*Ru(MeCN)3]+ or [CpRu(MeCN)3]+. Calculations suggest a new mechanism that involves a key ruthenacyclopropene intermediate. This mechanism not only explains the unusual Markovnikov regio-selectivity and anti-addition stereoselectivity observed by Trost and co-workers, but also motivated further experimental investigations. New intriguing experimental observations and further theoretical studies led to an extension of the reaction mechanism. The second section includes three cases of meta-selective C−H activation of aryl compounds. In the case of Cu-catalyzed selective meta-C−H activation of aniline, a new mechanism that involves a Cu(III)-Ar-mediated Heck-like transition state, in which the Ar group acts as an electrophile, was proposed. This mechanism predicted a higher reactivity for more electron-deficient Ar groups, which was supported by experiments. For two template-mediated, meta-selective C−H bond activations catalyzed by Pd(II), different mechanisms were derived for the two templates. One involves a dimeric Pd−Pd or Pd−Ag active catalyst, and the other involves a monomeric Pd catalyst, in which a monoprotected amino acid coordinates in a bidentate fashion and serves as an internal base for C−H activation. The third section discusses a desymmetry strategy in asymmetric synthesis. The construction of rigid skeletons is critical for these catalysts to distinguish two prochiral groups. Overall, fruitful collaborations between computational and experimental chemists have provided new and comprehensive mechanistic understanding and insights into these useful reactions.

■

better reaction and/or catalyst.2 A notable example is the reaction mechanism of olefin metathesis involving a metalcarbene intermediate proposed by Chauvin, which motivated the development of the Schrock and Grubbs catalysts.3 Historically, structural information on experimentally obtained intermediates was used to explain the observed selectivity of reactions. However, the lack of understanding of the mechanistic details, e.g., the properties of transition states, sometimes led to misleading conclusions. Particularly, exper-

INTRODUCTION Discoveries of new reactions and catalysts have provided many new ways to synthesize various functional molecules covering different areas, including life sciences and material sciences. Sustainable development also requires environmentally benign chemical transformations. To produce the desired products with minimum waste of starting materials, atom economy emerges as an important concept.1 Moreover, selective synthesis of desired products has long been a central challenge for synthetic chemists. Understanding reaction mechanisms, especially the origin of the chemo-, regio- and stereoselectivity, is a critical issue for rational development, or even design, of a © XXXX American Chemical Society

Received: February 22, 2016

A

DOI: 10.1021/acs.accounts.6b00093 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

found in textbooks cannot rationalize these unusual selectivities. For the anti-addition anti-Markovnikov hydrosilylation of terminal alkynes, Crabtree11 proposed that a cis-Ir-vinyl intermediate undergoing isomerization via an iridacyclopropene to give a trans-Ir-vinyl intermediate accounts for both syn- and anti-addition products. Our DFT study proposed a new mechanism that explains the unprecedented regio- and stereoselectivity of the Ru(II)catalyzed intermolecular and intramolecular hydrosilylations of alkynes (Scheme 2).12 In contrast to the Chalk−Harrod and

imental observation of short-lived and rare transition states, coined by Polanyi and Zewail as the “Holy Grails” of chemistry,4 remains very challenging. As a representative case, characterization of major and minor intermediates by NMR was inconsistent with the observed preference of products: the enantioselectivity of the Rh(I)-catalyzed asymmetric hydrogenation cannot be simply understood by a conventional lockand-key concept.5 Calculations by Landis’ group revealed an anti-“lock-and-key” mechanism, because the more stable major catalyst−substrate intermediate has a much higher reaction barrier than the less stable minor catalyst−substrate diastereomer.6 Recently, Imamoto reinvestigated the system and proposed a new pathway.7 Computational chemistry now plays an increasingly important role in unraveling the origin of selectivity, improving the synthesis or design of new catalysts. Currently, the absolute accuracy of state-of-the-art computational methods is not adequate to describe completely a reaction process in silico. However, relative accuracy is believed to be more reliable because errors for similar reaction scenarios are likely to be canceled. For example, evaluation of entropy in solution is an inherent challenge, but the error of entropy can be largely eliminated when comparing transition states for a selectivitydetermining step. It is because that the step involves the same elementary reaction type and the same number of molecule(s). Moreover, selectivity is a bridge between computational chemistry and synthesis.8 Based on the Boltzmann distribution and transition state theory concepts, calculated relative barriers for different transition states in the selectivity-determining step can reflect product ratios which can be measured experimentally. Once critical factors of a reaction mechanism are determined, ligands may be optimized and new catalysts can be designed to improve the selectivity. In this Account, selected examples are presented to illustrate the role of computational chemistry in unraveling the origin of the selectivity of several metal-catalyzed organic reactions.

Scheme 2. Wu−Trost Mechanism for the Ru(II)-Catalyzed Hydrosilylation of Alkynes and a New Mechanistic Pathway for Alkyne Insertion

modified Chalk−Harrod mechanisms, oxidative addition of a Si−H bond to form Ru(IV) hydride silyl complex (6) is unfavorable because of the unstable electron-deficient Ru(IV) species. Instead, a transition state involving oxidative addition concerted with hydrometalation (7) was found to directly give an uncommon ruthenacyclopropene intermediate (9). The transferring H in 7 can be regarded as a proton in this electronpoor system and should preferentially migrate to the electronrich carbon, partly determining the unusual Markovnikov regioselectivity (Scheme 3a). Another interesting feature for the formation of 9 is derived from rotation of the CC bond without an intermediate 8, after the H transfer process is mostly complete. Reaction path analysis revealed that rotation orientates the H away from the bulky silyl group and stereospecifically leads to 9. The presumed Ru(IV)-vinyl intermediate (8) is not formed because of the unfavorable 16-electron Ru(IV) center. Once 9 is formed, the stereochemistry is determined for the terminal alkynes, since the first step is the rate-determining step. The subsequent reductive silyl migration to the carbene via 10, as opposed to the classical reductive elimination, affords the uncommon product (11). Our computational findings agree well with all the experimental observations. The unique points in this Wu− Trost mechanism are the ruthenacyclopropene intermediate (9) and the subsequent silyl migration. Recently, the first crystal structure of a ruthenacyclopropene complex and the reversible silyl migration to a Ru-carbene were reported.13 The proposed ruthenacyclopropene intermediate and reductive silyl

■

REGIO-, STEREO-, AND CHEMOSELECTIVITY IN HYDROSILYLATION Hydrosilylation of alkynes is the most straightforward method to synthesize vinylsilanes. The anti-Markovnikov product is usually obtained using various catalysts. The classical Chalk− Harrod and modified Chalk−Harrod mechanisms were proposed to account for the observed stereoselectivity (synand anti-addition).9 Trost reported that unusual Markovnikov and anti-addition products preferentially were obtained from hydrosilylation of terminal alkynes by using [Cp*Ru(MeCN)3]+ or [CpRu(MeCN)3]+ as the catalyst (Scheme 1).10a Also, Trost and Ball observed unusual 6- and 7-endo-dig and anti-addition products from the intramolecular hydrosilylation reaction catalyzed by the same catalyst.10b Notably, the Chalk−Harrod and modified Chalk−Harrod mechanisms Scheme 1. Ru(II)-Catalyzed Hydrosilylation of Alkynes

B


Article

Accounts of Chemical Research Scheme 3a

Scheme 4a

a

(a) Ru(II)-catalyzed hydrosilylation of thioalkynes. (b) Our proposed mechanism for the hydrosilylation of thioalkynes with computed relative free energy (kcal/mol).

regioselectivity is favored to form a stable intermediate 13α, because the π-donating sulfenyl group stabilizes the electrophilic carbene in 13α. Distinct from our previous study, the ruthenacyclopropene intermediate is not the most stable intermediate. Several new isomeric intermediates and pathways were located (Scheme 4b). Intriguingly, the presence of the coordinating sulfenyl group triggers the formation of more stable sulfur-chelated σ-vinyl intermediates 14-anti and 15-syn. The isomerization of 13α to form 14-anti and 15-syn determines the E/Z-selectivity. Our calculations showed that the syn-addition pathway is more favorable than the antiaddition pathway, due to steric repulsion between the Cβ-R2 group and the bulky silyl group in the antiaddition pathway. Consequently, the excellent α-regio- and syn-stereoselective hydrosilylation of internal alkynes is achieved cooperatively through the Ru−S chelation and steric effects. Apart from the above-mentioned electronic influence on the regioselectivity, Sun also found a ligand-controlled regio- and stereodivergent hydrosilylation of internal alkynes through a striking steric factor. As shown in Scheme 5a, the hydrosilylation of silyl alkynes can be tuned to afford various vinyldisilanes with high but different regio- and stereoselectivity by using the catalysts 2 and 3. Catalyst 2 favors the β syn addition, whereas 3 prefers the α anti addition. Our

a

Favorable electronic effect of the alkynes in regioselectivity of the hydrometallation step for (a) electron-poor and (b) electron-rich metal systems. (c) Two possible energetic profiles for the Ru(II)catalyzed hydrosilylation of alkynes.

migration can also rationalize the failure of the hydrosilylation of olefins.10 The Wu−Trost mechanism also accounts for the unusual endo-selectivity in the intramolecular reaction,10b because the H−Si bond must be completely broken to give an unfavorable Ru(IV) hydride silyl intermediate in the exopathway.12 Furthermore, our calculations suggested the feasibility of the Wu−Trost mechanism (the new alkyne insertion pathway) for other hydrofunctionalizations.14 The Früstner and Murakami groups realized the unusual antiaddition hydrogenation, hydroboration, hydrogermylation, and hydrostannation of internal alkynes with the same or similar Ru catalyst.15 A striking difference between the Wu−Trost mechanism and the Chalk−Harrod mechanism, which involves a metal hydride, is the protonlike character of the transferring H and the electrophilic Cα in intermediate 9, which leads to the unique regioselectivity (Scheme 3). This underlying knowledge encouraged us to consider other substrates bearing different substituents. Since the first oxidative hydrometalation step is the rate- and regio-determining step for the terminal alkynes (black line in Scheme 3c), the regiochemistry should be influenced by an electronic factor (Scheme 3a). If the final reductive silyl migration is the rate-determining step (red line in Scheme 3c), the preceding isomerization step may occur and become reversible. Therefore, the latter steps can determine the selectivity. Sun recently observed a highly regio- and stereoselective hydrosilylation of electron-rich internal thioalkynes (Scheme 4a).16a Mechanistically, the presence of a sulfenyl substituent results in predicted α regioselectivity (Schemes 3a and 4a), but it gives an unusual syn-addition for 3. Our calculations indicated that the first oxidative hydrometalation step is still the rate- and regio-determining step. Therefore, as predicted, the α-

Scheme 5a

a

(a) Ru(II)-catalyzed hydrosilylation of silyl alkynes. (b) Our proposed mechanisms for the hydrosilylation of silyl alkynes.

C


Article


hindrance from the bulky phosphine ligand. Consequently, it results in the anti-Markovnikov product. Second, hydrosilylation of ketone and nitrile undergoes an ionic SN2-Si outer-sphere mechanism (23) instead of the inner-sphere mechanisms (21 or 22). The basic phosphine ligand promotes the Si−H activation via back-donation, and this facilitates the attack of the lone pair from acetone or acetonitrile on the activated silyl group. In addition to the Ru-catalyzed hydrosilylation (Scheme 6c), we studied the reaction mechanism of hydrosilylation of acetone and acetonitrile mediated by neutral hydrido(hydrosilylene)tungsten complexes.20a Two different substrates were found to proceed through two different mechanisms (an uncommon metal hydride migration mechanism and a silyl migration mechanism, respectively). Moreover, our DFT calculations showed that hydrosilylation of carbonyl compounds catalyzed by high-valent Re(V)-dioxo complexes should initiate with the [2 + 2] addition of the Si−H bond across a ReO bond.20b Overall, our systematic studies on the effects of the metal, ligand and substrate on the mechanism of hydrosilylation have provided a new and comprehensive mechanistic understanding and insights to this useful H−Si functionalization.

computation revealed that these reactions proceed by a similar mechanism described above (Scheme 2), i.e., oxidative hydrometalation, isomerization, and metallocyclopropeneform reductive silyl migration. However, the presence of the bulky silyl groups in the alkynes leads to two significant effects (Scheme 3c). First, due to the steric repulsion between the substrate and the catalyst, the ruthenacyclopropene is not of sufficiently low in energy to drive the reaction via a single transition state. More importantly, the final reductive silyl migration step, rather than the first step, becomes ratedetermining. Scheme 5b shows two reductive silyl migration transition states which lead to the observed products. In the case of the Cp* ligand, the repulsion between the substrate and the Cp* ring becomes critical. Therefore, a less bulky Cβ-H preferentially rotates toward Cp* in 18-TS, which leads to an α anti product. While using the Cp ligand, the repulsion between the substrate and the Cp ring is not very severe. Consequently, the bulky silyl group from the silane tends to migrate to the less bulky Cβ, with the H atom placed on the same side of the metallocyclopropene plane (19-TS). These computational results explained the observed regio- and stereodivergence based on synergetic steric effects. Similar results were obtained by using a tBu alkyne substrate with diminished electronic effect but similar bulkiness. Nikonov reported that hydrosilylation of acetone and acetonitrile can be achieved by the catalyst [CpRu(MeCN)2PR3]+ 20, in which an acetonitrile ligand in 2 is replaced by a phosphine ligand (Scheme 6a).17,18 Nevertheless,

■

META-SELECTIVITY OF C−H FUNCTIONALIZATION C−H bond functionalization is one of the most challenging areas in organic chemistry. Tremendous developments have been reported in the past decade. Due to the ubiquity of C−H bonds in substrates, selectivity becomes a critical issue to be addressed. Utilization of directing groups is the most common strategy to achieve the desired selectivity. In the case of arene C−H functionalization, various directing groups have been installed to produce the ortho-selective C−H functionalization reactions. In contrast, meta-selective C−H functionalization is rare. Several remarkable examples were reported recently.21 A meta-selective Cu-catalyzed arylation of anilides and α-aryl carbonyl compounds using diaryliodonium salt as the aryl source (Scheme 7) was reported by Gaunt.22 Interestingly, Pd-

Scheme 6a

Scheme 7. Regioselective C−H Arylation of Anilide Catalyzed by Pd(OAc)2 and Cu(OTf)2

catalyzed arylation of anilides under similar conditions gives the ortho product.23 We carried out computational and experimental studies to reveal the origin of the meta-selectivity.24 Three mechanisms were considered (Scheme 8), all involving a Cu(III)-aryl species generated from the reaction of the Cu complex and diaryliodonium salt. The first mechanism is the dearomatizing oxy-cupration proposed by Gaunt and Phipps.22a The Cu(III)-aryl complex attacks the meta-position concerted with the addition of the carbonyl group at the ortho-position (24). However, this mechanism can be ruled out because the calculated barrier for the oxy-cupration step is over 50 kcal/mol. The second mechanism is the concerted metalation-deprotonation (CMD) (25). This mechanism was commonly proposed to account for ortho-selective metal-catalyzed C−H functionalization. How-

a

(a) Hydrosilylation of acetone and acetonitrile catalyzed by a Ru(II)complex 20. (b) Three possible mechanistic pathways for the hydrosilylation. (c) Summary of mechanism and regio- and chemoselectivity of hydrosilylation of alkynes, acetone, and acetonitrile catalyzed by 2 and 20.

the catalyst 2 cannot catalyze the hydrosilylation of acetone and acetonitrile. Interestingly, the catalyst 20 also catalyzes hydrosilylation of terminal alkynes, but it gives the antiMarkovnikov product (instead of the Markovnikov product with 2). We conducted a comparative computational study to address the fascinating ligand effects on these reactions.19 Our computation led to two key findings. First, hydrosilylation of alkyne initiates with a silyl migration (22), rather than an H migration (21) (Scheme 6b), to the less bulky Cβ due to steric D


Article

Accounts of Chemical Research Scheme 8. Possible Mechanisms of Cu-Catalyzed meta-C−H Arylation of Anilide

Scheme 9. Template-Directed meta-Selective C−H Olefination of (a) Toluene and (b) Hydrocinnamic Acid Derivatives

ever, this mechanism has a higher calculated barrier than the third mechanism, namely a Heck-like mechanism (26), in which the Cu(III) coordinates with the carbonyl group and the Cu-Aryl bond adds to Cortho and Cmeta, respectively. This Heck-like mechanism explains several experimental observations:22a (1) The substrates with R1tBu gave the best yield. This is because a bulky R1 group pushes copper coordination to its trans position, facilitating the electrophilic attack to the ortho carbon. (2) meta-Arylation product was observed for the substrate with a meta-OMe R2 group, suggesting a competition between CMD and Heck-like mechanisms. (3) The reactivity is significantly reduced when the aryl group (Ph in Scheme 7) bares an electron-donating group, because it directly involves in the Heck-like transition state as an electrophile. Himo reported a similar four-center reductive elimination mechanism to explain the C2 selectivity in a copper-catalyzed amidation of indoles.25 In connection with the report of Gaunt and coworkers, the reaction shown in Scheme 7 may be achieved with a copperfree condition.22b Our kinetic study clearly showed that either Cu(I) or Cu(II) significantly increases the reactivity compared to a copper-free condition.24 More recently, Gaunt and coworkers also reported excellent enantioselectivity for the Cu(II)-catalyzed arylation reaction of Ar2IOTf with amides,26 indicating the direct involvement of the copper catalyst in the enantio-determining step. Notably, some recent coupling reactions involving diaryliodonium salt were reported to proceed without metal.27 These require further mechanistic study. Another pioneering progress of meta-selective C−H functionalization was reported by Yu and coworkers28 In order to direct the Pd to the vicinity of a meta-C−H bond, long but relatively rigid templates were designed (Scheme 9). Substrates installed with easily removable nitrile-containing templates undergo remote meta-C−H functionalization. This strategy potentially affords a general protocol for meta-selective C−H functionalization. We carried out mechanistic studies to elucidate the origin of the meta-selectivity, aiming to provide structural basis for further rational design.29 For the olefination of 27, the originally proposed monomeric model 31 cannot reproduce the experimental observation (Scheme 10). The ortho-C-H activation was calculated to be much more favorable than the meta-C-H activation. Distortion analysis revealed that the meta- or para-transition states bearing an 11-membered or 12-membered ring are highly destabilized by ring strain. We therefore proposed a dimeric model (32, MPd), in which the nitrile binds to one Pd and the arene C− H bond is activated by the other Pd (Scheme 10). This dimeric transition state was found to be lower in energy than the monomeric transition states. A systematic conformational search of transition states was conducted to ensure that a

Scheme 10. Commonly Proposed Monomeric PalladiumAcetate Model 31 and Novel Dimeric Model 32

reasonable distribution could be obtained. The transition state for the meta-C−H activation was found to be the lowest in free energy. Moreover, compared to the ortho-C−H bond activation, more conformations of transition states for the meta-C−H activation, within 2 kcal/mol from the lowestenergy conformer, were located. The Boltzmann distribution of these TS conformers derives a meta/ortho-products ratio close to that found experimentally. This result indicated that there may exist many low-energy conformers for a large system and the contribution of these conformers should not be ignored. Encouraged by the indispensable role of the Ag salt in the experiment and Pd−Ag heterodimer complex reported in the literature,30 we then further tested the possibility of replacing one Pd by Ag (32, MAg). Such a heterodimeric TS model also favors the meta-C-H activation with an even lower barrier. Similar Pd−Ag models have been proposed by Schaefer and coworkers in C−H amination.31 In the C−H functionalization of 29, mono-N-protected amino acid (MPAA) ligands were found to be essential. In the presence of a MPAA ligand, the nitrile template of 29 also overrides the electronic preference for the ortho C−H to achieve meta-selectivity. We then carried out a combined mass spectrometric/computational study to understand the role of MPAA.32 Mass spectrometry disclosed that MPAA binds to the metal center in a bidentate mode, stabilizing the monomer. In contrast to the Pd(OAc)2, the dimeric Pd complexes with an MPAA bridge were calculated to be unstable. Therefore, it is unlikely to form a dimer as the previous case. Given that the four coordination sites of Pd(II) were saturated by MPAA, the nitrile directing group and the C−H bond, there is no room at the Pd center to coordinate to the acetate which is the conventional proton acceptor in a CMD process. Consequently, we proposed a new model (33) for the C−H E


Article


conformational search to obtain low-energy conformers and to estimate the contribution from these conformers.

activation with the Pd/MPAA couple (Scheme 11). The carbonyl of the N-acyl group on the dianionic MPAA ligand

■

STEREOSELECTIVITY IN DESYMMETRIC REACTIONS We next applied this Pd/MPAA model (33) to account for the stereoselectivity of the asymmetric synthesis (Scheme 12).35 Indeed, this model reproduced the experimentally observed enantioselectivity of various prochiral substrates. More importantly, based on the transition structures, a general relay mechanism was proposed to explain the enantioselectivity induced by the Pd/MPAA catalyst (Scheme 13). Steric

Scheme 11. Possible Mechanisms for Pd/MPAA Catalyzed meta-C−H Activation Reaction

Scheme 13. Proposed Chirality Relay Model for Pd/MPAACatalyzed Asymmetric C−H Activation

serves as a base to deprotonate the C−H bond. The barriers for this new model were calculated to be lower than the alternative model (34) which required an external acetate. More importantly, the meta-selective TS was found to be the most stable one in our new model. Further examination of the most stable ortho, meta, and para transition structures revealed the origin of the meta-selectivity. In the meta-TS, the arene plane is perpendicular to the coordination plane of Pd and thus the optimum orbital overlap between the π orbital of arene and the d orbital of the Pd can be achieved. In the subsequent ionmobility mass spectrometric study, we captured the critical [Pd(MPAA)(substrate)] complex at different stages. The observation that the C−H bond can be activated in the absence of an external base in the gas phase supported the internal base model.33 The direct involvement of dianionic MPAA as the proton acceptor in C−H activation opens a new window for ligand design.34 Besides obtaining two novel models to account for the unique meta-regioselectivity, another lesson from these two studies is the necessity of performing conformational search. When the size of a model increases, the conformational space increases simultaneously. Taking these two cases as examples, the largest-ring system involves 16-membered rings. In this kind of conformational space, there may be many conformers with comparable relative energies. With the development of computation hardware and new algorithms, computational chemists can simulate a much larger system than two decades ago. Therefore, it is necessary to conduct a systematic

repulsion of the amino acid side-chain (R1) pushes the Nprotecting group below the Pd coordination plane. As the internal base for deprotonation, the carbonyl of the Nprotecting group directs the C−H bond downward. The sp2 C−H bond orientation also confines the prochiral carbon at the ortho position of the aryl ring downward and the aryl ring upward. Together with the coordination of the directing group, the substrate forms a rigid skeleton with the catalyst. The RL group (the large substituent) should take the sterically less crowded position. Desymmetrization also provides an indirect strategy for building all-carbon quaternary stereocenters. Scheme 14a presents an example of an enantioselective Cu-catalyzed Narylation.36 The steric congestion of the prochiral carbon in 41 makes the direct manipulation at the C center rather challenging. With the assistance of BINOL-derived ligand (43), selective N-arylation differentiates two aryl groups and thus results in a chiral product 42. A Cu(I)/Cu(III) catalytic

Scheme 12. Pd/MPAA-Catalyzed Desymmetric C−H Activation and Functionalization Reactions

F


Article

Accounts of Chemical Research Scheme 14a

Scheme 15a

a

(a) Enantioselective formation of an all-carbon quaternary stereocenter through copper-catalyzed asymmetric N-arylation reaction. (b) Proposed model for the enantio-determining step of the reaction. a

(a) Pd-catalyzed asymmetric aryl C−O bond formation with SDP(O) ligand. (b) Origin of the enantioselectivity.

cycle involving oxidative addition of aryl halide and reductive elimination of aryl and nucleophile was commonly proposed.37 Oxidative addition of C−I bond to the Cu center is expected to distinguish the two aryl groups and to determine the enantioselectivity. The transition states for the oxidative addition step, 44 and 45, are shown in Scheme 14b. Distinct from traditional BINOL-metal complexes, Cu(I) binds to the BINOL ligand with only one dative bond and leaves one coordination site to coordinate with the amine in the substrate. To fix the substrate into the chiral environment, an additional hydrogen bond between BINOL and the amine is formed. Due to the steric repulsion between the R group (−CN) and the 9anthracenyl group, 45 was calculated to be 1.9 kcal/mol higher in free energy than 44. When R is a hydroxyl group, the opposite enantioselectivity was observed. This phenomenon can be easily understood based on our model. Because of the stronger O−H--O hydrogen bond, BINOL prefers to bind with the hydroxyl group (R) rather than the amine. Accordingly, the chirality is switched. The finding that the chiral ligand interacts with the substrate directly to form a compact complex encourages further development of better ligands. Another desymmetric reaction we have studied is a metalcatalyzed aryl C-heteroatom coupling reaction (Scheme 15a). Unlike the previous case, the differentiation here between two hydroxyl groups becomes the enantioselectivity-determining step. The 1,1′-spirobiindane backbone (SDP(O)) ligand (49) coordinates to Pd to create an asymmetric chiral environment.38 One of the hydroxyl groups binds with Pd to form a seven-membered ring (50). Twelve conformers of 50 were located. The most stable conformers which lead to the R and S products are 50-R and 50-S (Scheme 15b). Both 50-R and 50S adopt a boat conformation. From the Newman projection through the axis of Cprochiral−O bond, the anti and gauche conformations clearly account for the preference of 50-S over 50-R. Interestingly, the short C−O bond length enhances such a preference.

C−H bond functionalization reactions are mainly discussed here, we believe that these mechanisms may also apply to other reactions. On the other hand, we have witnessed rapid development of various computational methods (e.g., hybrid QM/MM method, fragmentation methods, efficient highly accurate methods and free energy simulations methods with enhanced sampling methods) and increased computational power. Meanwhile, to attain mild, highly efficient and selective reactions, some of organic reactions have recently been developed by using complex bi(multi)-functional catalysts (e.g., 32 in Scheme 10) or combining with complex reaction environment (e.g., protein, DNA, MOF and COF). Therefore, mechanistic studies on complex organic reactions at the atomic level become much more challenging experimentally. We foresee that computational organic chemistry combined with multiscale/approach simulations (e.g., MM, QM/MM and QM methods; geometry optimization, molecular dynamics, Monte Carlo, and genetic algorithm approaches) and “Big Data” (application of cheminformatics to available experimental and computational data) should play an even more important role in unraveling mechanisms at atomic and dynamic details and in the rational design of new and better reactions/catalysts.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest. Biographies

■

Xinhao Zhang received his B.Sc. from University of Science and Technology of China in 2002 and Ph.D. from the Hong Kong University of Science & Technology (HKUST) (with Y.-D. Wu) in 2007. He then worked as a Humboldt Fellow at TU Berlin with H. Schwarz. He joined Peking University Shenzhen Graduate School as an associate Professor in 2011. His research focuses on utilizing mass spectrometry and computational chemistry to understand reaction mechanism and aggregation processes.

CONCLUDING REMARKS Our systematic computational mechanistic investigations not only successfully offer deep understandings and insights into the unusual selectivity of some recent organic reactions, but also motivate development of new reactions and/or better catalysts. Although our proposed mechanisms for the Si−H and G


Article


dig hydrosilylation catalyzed by ruthenium: evidence for a new mechanistic pathway. J. Am. Chem. Soc. 2003, 125, 30−31. (11) Tanke, R. S.; Crabtree, R. H. Unusual activity and selectivity in alkyne hydrosilylation with an iridium catalyst stabilized by an oxygendonor ligand. J. Am. Chem. Soc. 1990, 112, 7984−7989. (12) Chung, L. W.; Wu, Y.-D.; Trost, B. M.; Ball, Z. T. A Theoretical Study on the Mechanism, Regiochemistry, and Stereochemistry of Hydrosilylation Catalyzed by Cationic Ruthenium Complexes. J. Am. Chem. Soc. 2003, 125, 11578−11582. (13) Bernal, M. J.; Torres, O.; Martín, M.; Sola, E. Reversible Insertion of Carbenes into Ruthenium−Silicon Bonds. J. Am. Chem. Soc. 2013, 135, 19008−19015. (14) Chung, L. W. Ph.D. Dissertation, HKUST, Clear Water Bay, Hong Kong, 2006. (15) (a) Rummelt, S. M.; Fürstner, A. Ruthenium-Catalyzed transSelective Hydrostannation of Alkynes. Angew. Chem., Int. Ed. 2014, 53, 3626−3630. (b) Sundararaju, B.; Fürstner, A. A trans-Selective Hydroboration of Internal Alkynes. Angew. Chem., Int. Ed. 2013, 52, 14050−14054. (c) Radkowski, K.; Sundararaju, B.; Fürstner, A. A Functional-Group-Tolerant Catalytic trans Hydrogenation of Alkynes. Angew. Chem., Int. Ed. 2013, 52, 355−360. (d) Matsuda, T.; Kadowaki, S.; Yamaguchi, Y.; Murakami, M. Ruthenium-Catalyzed trans-Hydrogermylation of Alkynes: Formation of 2,5-Disubstituted Germoles through Double trans-Hydrogermylation of 1,3-Diynes. Org. Lett. 2010, 12, 1056−1058. (16) (a) Ding, S.; Song, L.-J.; Wang, Y.; Zhang, X.; Chung, L. W.; Wu, Y.-D.; Sun, J. Highly Regio- and Stereoselective Hydrosilylation of Internal Thioalkynes under Mild Conditions. Angew. Chem., Int. Ed. 2015, 54, 5632−5635. (b) Ding, S.; Song, L.-J.; Chung, L. W.; Zhang, X.; Sun, J.; Wu, Y.-D. Ligand-Controlled Remarkable Regio- and Stereodivergence in Intermolecular Hydrosilylation of Internal Alkynes: Experimental and Theoretical Studies. J. Am. Chem. Soc. 2013, 135, 13835−13842. (17) Gutsulyak, D. V.; Nikonov, G. I. Chemoselective Catalytic Hydrosilylation of Nitriles. Angew. Chem., Int. Ed. 2010, 49, 7553− 7556. (18) Gutsulyak, D. V.; Vyboishchikov, S. F.; Nikonov, G. I. Cationic Silane σ-Complexes of Ruthenium with Relevance to Catalysis. J. Am. Chem. Soc. 2010, 132, 5950−5951. (19) Yang, Y.-F.; Chung, L. W.; Zhang, X.; Houk, K. N.; Wu, Y.-D. Ligand-Controlled Reactivity, Selectivity, and Mechanism of Cationic Ruthenium-Catalyzed Hydrosilylations of Alkynes, Ketones, and Nitriles: A Theoretical Study. J. Org. Chem. 2014, 79, 8856−8864. (20) (a) Zhang, X.-H.; Chung, L. W.; Lin, Z.; Wu, Y.-D. A DFT Study on the Mechanism of Hydrosilylation of Unsaturated Compounds with Neutral Hydrido(hydrosilylene)tungsten Complex. J. Org. Chem. 2008, 73, 820−829. (b) Chung, L. W.; Lee, H. G.; Lin, Z.; Wu, Y.-D. Computational Study on the Reaction Mechanism of Hydrosilylation of Carbonyls Catalyzed by High-Valent Rhenium(V)Di-Oxo Complexes. J. Org. Chem. 2006, 71, 6000−6009. (21) (a) Wang, X.-C.; Gong, W.; Fang, L.-Z.; Zhu, R.-Y.; Li, S.; Engle, K. M.; Yu, J.-Q. Ligand-enabled meta-C-H activation using a transient mediator. Nature 2015, 519, 334−338. (b) Kuninobu, Y.; Ida, H.; Nishi, M.; Kanai, M. A meta-selective C−H borylation directed by a secondary interaction between ligand and substrate. Nat. Chem. 2015, 7, 712−717. (c) Dong, Z.; Wang, J.; Dong, G. Simple Amine-Directed Meta-Selective C−H Arylation via Pd/Norbornene Catalysis. J. Am. Chem. Soc. 2015, 137, 5887−5890. (22) (a) Phipps, R. J.; Gaunt, M. J. A Meta-Selective CopperCatalyzed C−H Bond Arylation. Science 2009, 323, 1593−1597. (b) Duong, H. A.; Gilligan, R. E.; Cooke, M. L.; Phipps, R. J.; Gaunt, M. J. Copper(II)-Catalyzed meta-Selective Direct Arylation of α-Aryl Carbonyl Compounds. Angew. Chem., Int. Ed. 2011, 50, 463−466. (23) Daugulis, O.; Zaitsev, V. G. Anilide ortho-Arylation by Using CH Activation Methodology. Angew. Chem., Int. Ed. 2005, 44, 4046− 4048. (24) Chen, B.; Hou, X.-L.; Li, Y.-X.; Wu, Y.-D. Mechanistic Understanding of the Unexpected Meta Selectivity in Copper-

Oscar Lung Wa Chung received his B.Sc. (first Hon.) degree in 2000 and Ph.D. degree under the supervision of Y.-D. Wu at the HKUST in 2006. He then worked as a postdoctoral fellow with K. Morokuma at Fukui Institute for Fundamental Chemistry, Kyoto University. In 2013, he started his tenure-track faculty position at South University of Science and Technology of China. He is very honored to follow four of the best computational chemistry families. His interests are in simulations of complex chemical, biochemical, and biophysical processes. Yun-Dong Wu received his B.Sc. from Lanzhou University in 1981 and Ph.D. from University of Pittsburgh in 1986. He had a long association with K. N. Houk, both as a graduate student and as a research associate. In 1992, he joined the HKUST. In 2010, he moved to Peking University. His research group is interested in understanding the mechanisms of catalysis, molecular designs with peptides, modeling of protein folding, and protein/protein interactions.

■

ACKNOWLEDGMENTS We are grateful to all students and research associates who contributed to these studies. Financial support was provided by the National Natural Science Foundation of China (21133002, 21302006, 21232001, and 21473086), the MOST of China (2013CB911501), and the Shenzhen Science and Technology Innovation Committee (KQTD201103 and KQTD20150717103157174).

■

REFERENCES

(1) Trost, B. M. The atom economy–a search for synthetic efficiency. Science 1991, 254, 1471−1477. (2) (a) Houk, K. N.; Cheong, P. H.-Y. Computational prediction of small-molecule catalysts. Nature 2008, 455, 309−31. (b) Houk, K. N.; Paddon-Row, M.; Rondan, N.; Wu, Y.-D.; Brown, F.; Spellmeyer, D.; Metz, J.; Li, Y.; Loncharich, R. Theory and modeling of stereoselective organic reactions. Science 1986, 231, 1108−1117. (3) Chauvin, Y. Olefin Metathesis: The Early Days (Nobel Lecture). Angew. Chem., Int. Ed. 2006, 45, 3740−3747. (4) (a) Baraban, J. H.; Changala, P. B.; Mellau, G. C.; Stanton, J. F.; Merer, A. J.; Field, R. W. Spectroscopic characterization of isomerization transition states. Science 2015, 350, 1338−1342. (b) Polanyi, J. C.; Zewail, A. H. Direct observation of the transition state. Acc. Chem. Res. 1995, 28, 119−132. (5) (a) Halpern, J. Mechanism and Stereoselectivity of Asymmetric Hydrogenation. Science 1982, 217, 401−407. (b) Landis, C. R.; Halpern, J. Asymmetric hydrogenation of methyl (Z)-.alpha.acetamidocinnamate catalyzed by [1,2-bis(phenyl-o-anisoyl)phosphino)ethane]rhodium(I): kinetics, mechanism and origin of enantioselection. J. Am. Chem. Soc. 1987, 109, 1746−1754. (6) Landis, C. R.; Feldgus, S. A Simple Model for the Origin of Enantioselection and the Anti “Lock-and-Key” Motif in Asymmetric Hydrogenation of Enamides as Catalyzed by Chiral Diphosphine Complexes of Rh(I). Angew. Chem., Int. Ed. 2000, 39, 2863−2866. (7) Gridnev, I. D.; Imamoto, T. Challenging the Major/Minor Concept in Rh-Catalyzed Asymmetric Hydrogenation. ACS Catal. 2015, 5, 2911−2915. (8) Cheng, G.-J.; Zhang, X.; Chung, L. W.; Xu, L.; Wu, Y.-D. Computational Organic Chemistry: Bridging Theory and Experiment in Establishing the Mechanisms of Chemical Reactions. J. Am. Chem. Soc. 2015, 137, 1706−1725. (9) Chalk, A. J.; Harrod, J. F. Homogeneous Catalysis. II. The Mechanism of the Hydrosilation of Olefins Catalyzed by Group VIII Metal Complexes. J. Am. Chem. Soc. 1965, 87, 16−21. (10) (a) Trost, B. M.; Ball, Z. T. Markovnikov Alkyne Hydrosilylation Catalyzed by Ruthenium Complexes. J. Am. Chem. Soc. 2001, 123, 12726−12727. (b) Trost, B. M.; Ball, Z. T. Intramolecular endoH


Article

Accounts of Chemical Research Catalyzed Anilide C−H Bond Arylation. J. Am. Chem. Soc. 2011, 133, 7668−7671. (25) Santoro, S.; Liao, R.-Z.; Himo, F. Theoretical Study of Mechanism and Selectivity of Copper- Catalyzed C−H Bond Amidation of Indoles. J. Org. Chem. 2011, 76, 9246−9252. (26) Cahard, E.; Male, H. P. J.; Tissot, M.; Gaunt, M. J. Enantioselective and Regiodivergent Copper-Catalyzed Electrophilic Arylation of Allylic Amides with Diaryliodonium Salts. J. Am. Chem. Soc. 2015, 137, 7986−7989. (27) (a) Sokolovs, I.; Lubriks, D.; Suna, E. Copper-Catalyzed Intermolecular C−H Amination of (Hetero)arenes via Transient Unsymmetrical λ3-Iodanes. J. Am. Chem. Soc. 2014, 136, 6920−6928. (b) Ciana, C.-L.; Phipps, R. J.; Brandt, J. R.; Meyer, F.-M.; Gaunt, M. J. A Highly Para-Selective Copper(II)-Catalyzed Direct Arylation of Aniline and Phenol Derivatives. Angew. Chem., Int. Ed. 2011, 50, 458− 462. (c) Kita, Y.; Morimoto, K.; Ito, M.; Ogawa, C.; Goto, A.; Dohi, T. Metal-Free Oxidative Cross-Coupling of Unfunctionalized Aromatic Compounds. J. Am. Chem. Soc. 2009, 131, 1668−1669. (28) Leow, D.; Li, G.; Mei, T.-S.; Yu, J.-Q. Activation of remote metaC-H bonds assisted by an end-on template. Nature 2012, 486, 518− 522. (29) Yang, Y.-F.; Cheng, G.-J.; Liu, P.; Leow, D.; Sun, T.-Y.; Chen, P.; Zhang, X.; Yu, J.-Q.; Wu, Y.-D.; Houk, K. N. Palladium-Catalyzed Meta-Selective C−H Bond Activation with a Nitrile-Containing Template: Computational Study on Mechanism and Origins of Selectivity. J. Am. Chem. Soc. 2014, 136, 344−355. (30) Neo, Y. C.; Vittal, J. J.; Hor, T. S. A. Pd(O2CCF2CF3-O)2(dppf) and PdAg(μ-O2CCF2CF3)2(O2CCF2CF3-O) (dppf) [dppf = 1,1′bis(diphenylphosphino)ferrocene]: A stable Pd(II) phosphine carboxylate as a precursor to heterobimetallic carboxylates. J. Chem. Soc., Dalton Trans. 2002, 337−342. (31) Anand, M.; Sunoj, R. B.; Schaefer, H. F. Palladium−Silver Cooperativity in an Aryl Amination Reaction through C−H Functionalization. ACS Catal. 2016, 6, 696−708. (32) Cheng, G.-J.; Yang, Y.-F.; Liu, P.; Chen, P.; Sun, T.-Y.; Li, G.; Zhang, X.; Houk, K. N.; Yu, J.-Q.; Wu, Y.-D. Role of N-Acyl Amino Acid Ligands in Pd(II)-Catalyzed Remote C−H Activation of Tethered Arenes. J. Am. Chem. Soc. 2014, 136, 894−897. (33) Cheng, G.-J.; Chen, P.; Sun, T.-Y.; Zhang, X.; Yu, J.-Q.; Wu, Y.D. A Combined IM-MS/DFT Study on [Pd(MPAA)]-Catalyzed Enantioselective C-H Activation: Relay of Chirality through a Rigid Framework. Chem. - Eur. J. 2015, 21, 11180−11188. (34) Xiao, K.-J.; Lin, D. W.; Miura, M.; Zhu, R.-Y.; Gong, W.; Wasa, M.; Yu, J.-Q. Palladium(II)-Catalyzed Enantioselective C(sp3)−H Activation Using a Chiral Hydroxamic Acid Ligand. J. Am. Chem. Soc. 2014, 136, 8138−8142. (35) (a) Shi, B.-F.; Maugel, N.; Zhang, Y.-H.; Yu, J.-Q. PdII-Catalyzed Enantioselective Activation of C(sp2)-H and C(sp3)-H Bonds Using Monoprotected Amino Acids as Chiral Ligands. Angew. Chem., Int. Ed. 2008, 47, 4882−4886. (b) Shi, B.-F.; Zhang, Y.-H.; Lam, J. K.; Wang, D.-H.; Yu, J.-Q. Pd(II)-Catalyzed Enantioselective C−H Olefination of Diphenylacetic Acids. J. Am. Chem. Soc. 2010, 132, 460−461. (c) Chu, L.; Wang, X.-C.; Moore, C. E.; Rheingold, A. L.; Yu, J.-Q. PdCatalyzed Enantioselective C−H Iodination: Asymmetric Synthesis of Chiral Diarylmethylamines. J. Am. Chem. Soc. 2013, 135, 16344− 16347. (d) Cheng, X.-F.; Li, Y.; Su, Y.-M.; Yin, F.; Wang, J.-Y.; Sheng, J.; Vora, H. U.; Wang, X.-S.; Yu, J.-Q. Pd(II)-Catalyzed Enantioselective C−H Activation/C−O Bond Formation: Synthesis of Chiral Benzofuranones. J. Am. Chem. Soc. 2013, 135, 1236−1239. (36) Zhou, F.; Cheng, G.-J.; Yang, W.; Long, Y.; Zhang, S.; Wu, Y.D.; Zhang, X.; Cai, Q. Enantioselective Formation of Cyano-Bearing All-Carbon Quaternary Stereocenters: Desymmetrization by CopperCatalyzed N-Arylation. Angew. Chem., Int. Ed. 2014, 53, 9555−9559. (37) (a) Jones, G. O.; Liu, P.; Houk, K. N.; Buchwald, S. L. Computational Explorations of Mechanisms and Ligand-Directed Selectivities of Copper-Catalyzed Ullmann-Type Reactions. J. Am. Chem. Soc. 2010, 132, 6205−6213. (b) Yu, H.-Z.; Jiang, Y.-Y.; Fu, Y.; Liu, L. Alternative Mechanistic Explanation for Ligand-Dependent Selectivities in Copper-Catalyzed N- and O-Arylation Reactions. J. Am.

Chem. Soc. 2010, 132, 18078−18091. (c) Zhang, H.; Yao, B.; Zhao, L.; Wang, D.-X.; Xu, B.-Q.; Wang, M.-X. Direct Synthesis of High-Valent Aryl−Cu(II) and Aryl−Cu(III) Compounds: Mechanistic Insight into Arene C−H Bond Metalation. J. Am. Chem. Soc. 2014, 136, 6326− 6332. (38) Shi, J.; Wang, T.; Huang, Y.; Zhang, X.; Wu, Y.-D.; Cai, Q. PdCatalyzed Asymmetric Intramolecular Aryl C−O Bond Formation with SDP(O) Ligand: Enantioselective Synthesis of (2,3Dihydrobenzo[b][1,4]dioxin-2-yl)methanols. Org. Lett. 2015, 17, 840−843.

I


New Mechanistic Insights on the Selectivity of Transition-Metal

Recommend Documents