An Anthracene-Thiolate-Ligated Ruthenium Complex: Computational Insights into Z-Stereoselective Cross Metathesis

Stereoselective control of the cross metathesis of olefins is a crucial aspect of synthetic procedures. In this study, we utilized density functional theory methods to calculate thermodynamic and kinetic descriptors to explore the stereoselectivity of cross metathesis between allylbenzene and 2-butene-1,4-diyl diacetate. A ruthenium-based complex, characterized primarily by an anthracene-9-thiolate ligand, was designed in silico to completely restrict the E conformation of olefins through a bottom-bound mechanism. Our investigation of the kinetics of all feasible propagation routes demonstrated that Z-stereoisomers of metathesis products can be synthesized with an energy cost of only 13 kcal/mol. As a result, we encourage further research into the synthetic strategies outlined in this work.


INTRODUCTION
Historically, one of the main challenges in cross metathesis (CM) was the control of the stereoselectivity of the reaction. 1or nonstereoselective catalysts, metathesis products typically consist of a mixture of E and Z isomers, creating operational difficulties due to the complicated and costly nature of component separation. 2 In 2009, Schrock et al. synthesized the first Z-selective olefin metathesis catalysts based on Mo 3,4 and W 5 complexes (see Figure 1).These catalysts were able to achieve yields above 50% with a Z isomer ratio greater than 98%.At that juncture, the field of alkylidene chemistry involving Mo and W had advanced substantially, resulting in the availability of numerous exceptionally active and tunable catalysts. 6In addition, over the past decade, significant progress has been made in the development of Z-selective Mo-and W-based catalysts for CM, which have provided new opportunities for the synthesis of Z-olefins with high selectivity and efficiency. 7These include, for example, a series of halogenated Mo-alkylidene complexes that were reported as exceptionally effective catalysts for olefin CM reactions, resulting in the production of acyclic 1,2-disubstituted Z-alkenyl halides with ratios Z:E of 93:7 or higher. 8Monoaryloxide chloride Mo-based catalysts also exhibited Z-selective CM reactions involving a series of olefins and Z-1,1,1,4,4,4-hexafluoro-2-butene in benzene at 22 °C. 9urthermore, in the same context of the Mo-catalyzed CM framework, notable advances in the synthesis of Z-trisubstituted alkenes have been recently achieved by Hoveyda et al. 10−12 In the case of Z-selective Ru-based catalysts, a series of functional-group-tolerant systems were reported by Grubbs et al. since 2011. 13,14−18 Nitratosubstituted catalysts were found to be more stable and efficient for CM. 19In 2013, Hoveyda et al. used a similar approach to synthesize a Z-selective catalyst, where the replacement of chlorides with a catechodithiolate ligand favored the formation of the Z isomer in ring-opening metathesis polymerization and ring-opening/cross-metathesis. 20In the same year, Jensen et al. achieved Z-selective CM by substituting one chloride of a Rubased catalyst complex with the bulky 2,4,6-triphenylbenzenethiolate (see Figure 1c). 21This substitution resulted in Z selectivity of up to 96% in the metathesis homocoupling of terminal olefins.Continued investigations in Jensen's group resulted in additional phosphine-based Z-selective complexes. 22n 2015, within Ru-based catalysts, Hoveyda et al. presented a Ru-disulfide complex to catalyze the synthesis of acyclic Z allylic alcohols via CM transformations, with yields of up to 80% and Z:E ratios as high as 98:2. 23In 2018, Mauduit et al. synthesized a variation of the adamantyl-NHC catalysts, which incorporated N-Dipp (N-2,6-diisopropylphenyl) and N-adamantyl substituents in the nonsaturated NHC ligand. 24This catalyst exhibited exceptional catalytic efficacy in both self-and CM reactions, even at a low catalyst loading, yielding internal olefins with remarkable conversion rates and high Z-selectivity exceeding 99%.Later in 2023, within the Ru-based framework, Mauduit's group synthesized a series of 16 Z-selective catalysts, and in the case of a catechodithiolate complex, it resulted in a remarkable performance in the Z-selective (>98%) asymmetric ringopening CM of exonorbornenes. 25Furthermore, the first Zstereoselective catechodithiolate-based ruthenium complexes containing cyclic(alkyl)(amino)carbene (CAAC) ligands were recently reported by Bertrand and co-workers (see Figure 1c). 26oderate to good yields and high Z-selectivity (>98%) were obtained in various metathesis transformations, including CM.
Grubbs et al. in collaboration with Houk's group provided a rationale for the design of Z-stereoselective catalysts by distinguishing between two possible metathesis mechanisms, bottom-bound 27 and side-bound 28 (see Figure 1d). 29They presented evidence to support the notion that the side-bound mechanism leads to improved d-orbital back-donation and van der Waals interactions between the catalyst and the Z isomer of olefin in transition states, and the steric compression of the olefin substituents is reduced.A possible explanation for the Zselectivity observed in catechodithiolate-based catalysts may be rationalized in terms of the side-bound mechanism as reported by Houk et al. 29 Sterically demanding ligands are known to stabilize the 14-electron Ru-activated catalyst 30 and prevent the formation of the E isomer. 31,32However, Z-selective catalysts that incorporate monodentate chelates (e.g., I − , (2,4,6-Ph 3 )Ph-S − , or Dipp-O − ) that allow bottom-bound pathways have also been developed.As a result, the question of whether the Zstereoselectivity arises entirely from the side-bound mechanism is a primary focus in the current study.
Recent developments in Z-stereoselective olefin metathesis have led to specific applications such as continuous-flow synthesis of natural products, including pheromones and macrocyclic odorant molecules. 33Inspired by the current state of Z-selective olefin metathesis, we aimed to investigate by means of quantum-chemical methods the development of an even more efficient catalyst for stereoselective CM reactions.To achieve this goal, we incorporated a polycyclic aromatic-carbon (anthracene) thiolate ligand to exclusively induce the formation of Z olefins.It may be argued that this framework replicates the design principles of Jensen's catalyst, 21 in which 2,4,6triphenylbenzenethiolate was incorporated as a ligand that induces the selective formation of Z olefins.Although similarities exist, ligands such as anthracenes or extended polycyclic rings cannot be spatially reoriented so that the formation of E olefins is prohibitive.Therefore, our current research is motivated by the observation that E olefins could still be detected in catalysts with improved Z-selectivity, which were synthesized through the incorporation of Z-inducing ligands such as thiolates, 22 sulfonates, and phosphates. 34ccordingly, we chose a model reaction from a set of six metathesis transformations proposed by Grubbs et al. to characterize newly synthesized catalysts, 31,35 specifically the CM of allylbenzene (A) and 2-butene-1,4-diyl diacetate (B) to produce 4-phenyl-2-buten-1-yl acetate (P) (see Figure 1e).The objective of our work was to conduct a thorough analysis of the thermodynamic and kinetic factors that determine the formation of the Z isomer of the products as well as to elucidate key strategies to encourage further experimental assessment.−40

COMPUTATIONAL DETAILS
Density functional theory (DFT) calculations were performed using the quantum-chemistry code Jaguar version 11.2. 41ecke's three-parameter functional combined with the Lee− Yang−Parr correlation functional (B3LYP) was employed to optimize the geometries. 42,43During the geometry optimization process, the dispersion energy corrections developed by Grimme et al. (D3) were incorporated. 44The electronic configuration of the molecular systems was described using

The Journal of Physical Chemistry A
the Gaussian 6-31G** basis set, which includes polarization functions for the H, C, N, O, and S atoms.In the case of ruthenium, the small-core, quasi-relativistic effective core potential developed at Los Alamos National Laboratory (LA), 45 along with an associated double ζ plus polarization basis set, was used (standard LACVP** keyword in the Jaguar code).Analytical frequency calculations were performed for all localized stationary points at the B3LYP-D3/6-31G**∼LACVP** level of theory.Transition states and connecting local minima were located via linear-transit calculations by using the same DFT method.Electronic energies were obtained from the single-point M06-D3 calculations 46 that included a triple ζ basis set plus polarization and diffuse functions, 6-311G++(d,p) (for Ru: standard LACV3P++** keyword in Jaguar).The polarizable continuum model (PCM) 47,48 was employed to account for solvent effects, with dichloromethane as the solvent.Gibbs free energies were derived from the corrected electronic energies at the (PCM: CH 2 Cl 2 ) M06-D3/6-311G++**∼LACV3P++** // B3LYP-D3/6-31G**∼LACVP** level of theory.Corrections for zeropoint energy, thermal contributions to internal energy, and entropy term were calculated from vibrational frequencies at 298.15 K, assuming an ideal gas under standard conditions.Standard convergence criteria and a fine grid for DFT calculations were utilized in all cases.It can be argued that the M06 functional already incorporates medium-range dispersion, leading to overestimation of dispersion contributions due to double counting of these effects in M06-D3. 49However, the inclusion of D3 corrections to M06 was demonstrated to enhance the results for catalyzed olefin metathesis, particularly in treating weak interactions. 50Furthermore, the computational approach we chose has previously been benchmarked against experimental data. 51,52

RESULTS AND DISCUSSION
The primary aim of our current work is to demonstrate, via quantum-chemical calculations, the feasibility of designing Zstereoselective catalysts using the concept of the bottom-bound mechanism.To achieve this goal, we have proposed a Ru-based complex that contains an anionic ligand, anthracene-9-thiolate, which can induce the formation of the Z isomer of olefins through a bottom-bound mechanism.The NHC ligands are constituted of 2,4,6-trimethylphenyl (mesityl) and 2,4-dimethyl-6-(propylthiolate)phenyl groups, where the thiolate group is coordinated with Ru.In this regard, the alkylthiolate moiety was extended until the NHC plane reached a perpendicular alignment with the S−Ru−S axis, resulting in propyl as the alkyl chain.In the context of polycyclic aromatic hydrocarbons, Grela et al. previously synthesized Ru-based catalysts characterized by either a thiophene-based 53 or a 10-phenyl-9phenanthryl 54 ligand bonded to an unsymmetrical NHC moiety.In fact, in the case of Hoveyda−Grubbs complexes bearing N-(9-alkylfluorenyl)imidazol-2-ylidene ligands (alkyl: CH 3 , C 2 H 5 , or C 6 H 5 ), they conducted thorough investigations of olefin metathesis reactions, with a specific focus on achieving Zselectivity with Z/E ratios of up to 94/6. 55Furthermore, Rubased catalysts incorporating sulfurated chelates have also been reported by Lemcoff et al. 56−58 Therefore, we have formulated a possible synthesis route to the precatalyst in question (details in Figure S1 in the Supporting Information), which involves the preparation of thioether-imidazolium chlorides, 59 anilines, 60,61 and modifications to the Grubbs first-generation catalyst. 34,62.1.The Initiation Route.The catalytic cycle commences with the generation of an active catalyst, AC Ph , as a result of the dissociation of pyridine from pre.Two distinct conformations pre and pre′ are presented in Figure 2, which are differentiated by the orientation of benzylidene relative to the propyl-thiolate of NHC.The Ru-pyridine bond dissociation reported in Figure S2 in the Supporting Information reveals a flat potential (electronic) energy surface that reaches maxima of 16 and 21 kcal/mol for pre and pre′, respectively.Nonetheless, for the localized stationary points, the thermal contributions resulted in a Gibbs activation energy of <15 kcal/mol, as shown in Figure 2. As a result, activation of the catalyst is expected to take place under mild conditions at room temperature.The subsequent step involves olefin coordination (OC), and given our interest in synthesizing 4-phenyl-2-buten-1-yl acetate P through a productive metathesis reaction, there exist two possible

The Journal of Physical Chemistry A
sequences of the cycle: allylbenzene A association followed by propagation with 2-butene-1,4-diyl diacetate B, as illustrated in Figure 1e, or vice versa.Nevertheless, AC Ph is expected to exhibit a preference for reacting with allylbenzene initially, owing to the greater reactivity of less substituted alkenes such as olefin A, 63 as confirmed in our earlier work. 64As a result, we focus only on the initiation route through A.
Our analysis and energy comparisons were conducted starting from the olefin coordination intermediate OC A .We identified eight distinct configurations for the coordination of A with the active catalyst AC Ph (or AC′ Ph in which benzylidene is trans to propyl-thiolate), as depicted in Figure 2. The phenyl group in A can acquire either a contracted or an expanded conformation.Similarly to our previous research findings, it has been demonstrated that the destabilization of OC intermediates is correlated with the destabilization of transition states. 65herefore, we investigated the initiation phase exclusively for the most stable conformers.Consequently, we suggest the criterion of ΔG rel < 3 kcal/mol to select the intermediates that undergo a kinetically favored initiation phase, which in this study are all OC A intermediates characterized by the expanded configuration of phenyl in A. This thermodynamic stabilization can be attributed to a more active alkene orientation, as described by Straub. 66he activated complex pyr/AC Ph or pyr/AC′ Ph is isoenergetic with respect to OC A since we did not observe significant energy variations for the exchange of pyridine with olefin A (ΔG = 0.4 and −1.2 kcal/mol respectively; see Figure 3).This fact indicates that catalyst activation is susceptible to reversibility when the concentration of A is insufficient.The 2,2-cycloaddition from OC A via transition state TS1 A produces metallacyclobutane MCB A .The rupture of MCB A through TS2 A leads to the 2,2-cycloreversion intermediate CR A , followed by the release of the propagating activated catalyst AC A (or AC′ A in which alkylidene is trans to propyl-thiolate) and the corresponding olefin.Species OC A1 and OC A5 are associated with productive metathesis so that styrene (S 0 ) is generated through the initiation route.On the other hand, OC A3 and OC A7 are related to nonproductive metathesis, so that substrate S 1 (1propene-1,3-diphenyl) is instead generated.The Gibbs free energy profiles, ΔG, for the initiation phase via OC A1 and OC A5 (productive metathesis) are depicted in Figure 3, along with the reaction energies and barriers for the 2,2-cycloaddition steps (ΔG 1 and ΔG 1 ‡ ) and 2,2-cycloreversion steps (ΔG 2 and ΔG 2 ‡ ).
−69 That is, the third energy barrier reported in Figure 3 is calculated as ΔG 3 ‡ = G(TS2 A ) − G(OC A ) for each pathway.The ΔG 3 ‡ value for path A1 is greater than the ΔG 1 ‡ and ΔG 2 ‡ values, indicating that the critical states of this reaction are OC A1 and TS2 A1 .In the case of path A5, ΔG 2 ‡ is slightly greater than ΔG 3 ‡ ; therefore, the crucial states of this pathway are MCB A5 and TS2 A5 describing the 2,2cycloreversion step.As indicated in Figure 3, our findings suggest that the catalyst can be promptly activated toward productive metathesis through pathways A1 and A5, which have overall costs of 8.3 and 6.1 kcal/mol, respectively.The generation of the active catalysts, AC′ A and AC A , through the release of styrene via paths A1 and A5, respectively, led to favorable driving forces of 6.5 and 4.5 kcal/mol, which were computed using the most stable isolated fragments: G(AC′ A ) or G(AC A ) + G(styrene) − G(OC A1 ).In the cases of AC′ A and AC A , we optimized eight conformers and chose the most thermodynamically stable structures, as illustrated in Figure S3 in the Supporting Information.In contrast, nonproductive routes A3 and A7 are comparatively less kinetically competitive due to limiting initiation barriers higher than 10 kcal/mol.Furthermore, the formation of substrate S 1 is endergonic, unlike the exergonic initiation through the release of styrene (see Figure S4 in the Supporting Information for a detailed description of the corresponding energy profiles).
As a final point, we employed the activation-strain model strategy 70−73 to assess the relative energy of the system in the 2,2-cycloaddition transition states, considering the structural strain and intermolecular interactions involved.The strain energy quantifies the amount of energy required to alter the geometry of both olefin A and active catalyst AC Ph (or AC′ Ph ) to form the structure of TS1 A , while the interaction energy assesses the extent of interactions between the distorted fragments when they form TS1 A .Similarly to our previous work, 64 the overall cost ΔG 3 ‡ for the initiation route is linearly correlated (R 2 = 0.77) with the level of structural strain and, to some extent, the molecular interactions (R 2 = 0.53, see inset in Figure 2).Consequently, faster reaction rates are associated with a lower The Journal of Physical Chemistry A degree of structural strain in the reacting entities as well as to moderate molecular interactions that prevent the formation of highly stabilized metallacyclobutanes, thereby promoting the course of the reaction.

The Propagation Route.
In relation to the process of productive metathesis, the active catalysts AC′ A and AC A are subsequently coordinated with compound B to form the propagating OC P species.However, the flexibility of the structure of olefin B results in the formation of several conformers of OC P (the selection of OC P species is reported in Figure S5 in the Supporting Information).Similarly to the preceding subsection, our analysis of the propagation phase focused exclusively on the most stable conformers, specifically, OC P2 and OC P3 .The corresponding reaction mechanisms are illustrated in Figure 4.
The Gibbs free energy profiles for the propagation phase through OC P2,3 are reported in relation to the sum of the energies of AC′ A and B. To obtain the energy of B, we selected the most stable structure from a group of three cis conformers (see Figure S3 in the Supporting Information for details).For example, the formation of OC P2 resulted in no energy change (ΔG = 0.1 kcal/mol) and the 2,2-cycloaddition energy barrier is ΔG 1 ‡ = 10.9 kcal/mol.MCB P2 is destabilized since ΔG 1 = 9.0 kcal/mol, but the 2,2-cycloreversion proceeds with an energy barrier of only ΔG 2 ‡ = 4.1 kcal/mol under an exergonic process of ΔG 2 = −3.4kcal/mol.Olefin decoordination from CR P2 resulted in a favorable driving force of 4.2 kcal/mol, giving rise to the acetate propagating alkylidene (AC P ) and the metathesis product P.The OC P2 and TS2 P2 stationary points are the ratelimiting states in this part of the pathway, resulting in an overall cost of ΔG 3 ‡ = 13.1 kcal/mol, so that the reaction can be carried out under laboratory conditions.In our prior communication, 64 we reported equivalent energy barriers for a second-generation Grubbs catalyst, so that a similar catalytic activity is expected for the complex in question.Path P3 is kinetically less competitive but accessible since we calculated ΔG 3 ‡ = 22.1 kcal/mol as the limiting barrier.Furthermore, if the phenyl-alkylidene is located in the opposite direction relative to AC′ A (that is, AC A in which the phenyl-alkylidene is cis to propyl-thiolate), the propagating structure OC P6 , which is analogous to OC P2 , was destabilized by 6 kcal/mol compared to OC P2 .Nonetheless, the overall energy required for the formation of P via path P6 was calculated to be only 9 kcal/mol, as illustrated in Figure S6 and discussed in the Supporting Information.Consequently, the catalyst in question is versatile, which means that metathesis products can be generated regardless of the orientation of the phenyl-alkylidene within the activated catalyst.Our calculations indicate that Zstereoselective catalysts can be effectively designed by considering the bottom-bound mechanism, making our study a reference point for future synthetic procedures as an alternative to strategies based on the side-bound mechanism.

CONCLUSIONS
We performed predictive DFT calculations to assess the catalytic activity of a Ru-based complex containing an anthracene-9thiolate ligand.The catalyst was designed in silico to induce the formation of the Z isomer of metathesis products through the bottom-bound mechanism.Additionally, the anthracene-9thiolate ligand impedes the formation of the E isomer.Through an in-depth analysis of the thermodynamics and kinetics of the formulated mechanisms, we explored the potential applications of the Ru complex in the CM of allylbenzene and 2-butene-1,4diyl diacetate.Our results indicate that this new catalyst should facilitate efficient catalyst activation, as the energy barrier for the initiation route was <15 kcal/mol essentially attributed to the Ru-pyridine bond dissociation.Furthermore, we demonstrated that the overall energy cost for olefin metathesis through the propagation phase, which leads to the synthesis of the Z isomer of the metathesis product, 4-phenyl-2-buten-1-yl acetate, was only 13 kcal/mol.Our findings hold significant promise for the future exploration of this new catalyst candidate through experimental studies, and we hope that they will serve as a useful guide for further computational investigations aimed at developing new Z-stereoselective catalysts.In this regard, we additionally proposed a laboratory procedure for the synthesis of this novel catalyst.
Discussion of alternative initiation and propagation phases; detailed analysis of molecular structures and energies of the investigated chemical species and their structural isomers, along with their respective DFT absolute energies (PDF) The Journal of Physical Chemistry A

Figure 1 .
Figure 1.Selected Z-stereoselective catalysts: (a) early Mo-and W-based complexes, (b) set of catalysts designed by Grubbs, and (c) Ru-based complexes ligated to thiolates.(d) Comparison between the side-bound and bottom-bound conformations as a strategy in the design of Z-selective catalysts.(e) Cross metathesis studied in this work.

Figure 2 .
Figure 2. (Top) Gibbs free energy changes for catalyst activation through the dissociation of pyridine.(Bottom) Conformers of the activated catalyst coordinated to allylbenzene (OC A ) and Gibbs free energy comparisons relative to OC A1 .Inset: The activation-strain model for the 2,2-cycloaddtion transition state (TS1 A ).All energies are given in kcal/mol.

Figure 3 .
Figure 3. Gibbs free energy profiles for the initiation phase for the most stable OC A species associated with productive metathesis.Illustration and energies (kcal/mol) are relative to OC A1 .The equivalent schematization regarding the benzylidene of AC′ Ph localized in the opposite direction corresponds to path A5.ΔG 3 ‡ = G(TS2) − G(OC).

Figure 4 .
Figure 4. Gibbs free energy profiles (in kcal/mol) for the propagation phase through 2-butene-1,4-diyl diacetate as the substrate (B) and the catalyst activated via the metathesis of allylbenzene (AC′ A ).