One-Pot Sequential Kinetic Profiling of a Highly Reactive Manganese

Article Cite This: J. Am. Chem. Soc. 2018, 140, 9244−9254

pubs.acs.org/JACS

One-Pot Sequential Kinetic Profiling of a Highly Reactive Manganese Catalyst for Ketone Hydroboration: Leveraging σ‑Bond Metathesis via Alkoxide Exchange Steps Vladislav Vasilenko, Clemens K. Blasius, and Lutz H. Gade* Anorganisch-Chemisches Institut, Universität Heidelberg, Im Neuenheimer Feld 270, 69120 Heidelberg, Germany

Downloaded via KAOHSIUNG MEDICAL UNIV on July 25, 2018 at 09:54:55 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: A comprehensive experimental and computational mechanistic study of the highly enantioselective hydroboration of ketones catalyzed by a manganese(II) alkyl boxmi pincer complex is reported. The catalyst operates at low catalyst loadings (down to 0.01 mol %) under very mild conditions (typically −40 °C) and facilitates the reduction of both aryl alkyl and dialkyl ketones with excellent selectivity (up to >95%ee). Catalyst activation pathways were investigated, demonstrating that a manganese(II) hydride and a manganese(II) alkoxide species are part of the catalytic cycle and can be generated via σ-bond metathesis of the alkyl precursor with the borane or by alcoholysis. Extensive kinetic experiments based on a “one-pot sequential kinetic profiling” approach under various conditions in combination with kinetic simulations reveal that two catalytic cycles are effective with this earth-abundant base metal catalyst: (i) a minor MnH/borane-mediated insertion cycle, in which the subsequent, product-releasing metathesis step is rate determining (km = 0.076 s−1), giving a background reaction, which is zeroth order in substrate concentrations, and (ii) a major MnOR/borane-based alkoxide exchange process, leveraging the high-barrier metathesis via the affiliation to an insertion step. The latter features non-integer reaction orders in both reagents due to a combination of an adduct formation step (ka = 2.12 M−1 s−1, k−a = 0.49 s−1) and a substrate insertion step of comparable rates (kai = 3.74 M−1 s−1). The kinetic findings are underpinned by high-level density functional theory calculations of the mechanism, control experiments, and kinetic isotope effect/Hammett/Eyring analysis in different concentration regimes. The study highlights the role of a rigorous mechanistic understanding of homogeneous catalytic processes in 3d metals for rational catalyst discovery and optimization.

■

INTRODUCTION Mechanistic studies of earth-abundant metal catalysts have been fueled by their increasing relevance in chemical synthesis over the past years.1−6 The variety of reactivity displayed by these readily available and frequently nontoxic 3d metal catalysts demonstrates that environmentally benign and inexpensive catalytic platforms can in many respects surpass known noble metal benchmark systems or give access to new synthetically relevant transformations.7−9 This is reflected by the high number of recently reported iron- or manganesemediated transformations which include oxidations,10−21 reductions,22−26 or even C−H activations.27−39 However, despite the tremendous growth of this field, there remains a lack of understanding with regard to the elementary processes which determine the efficacy and selectivity of the catalysts. This situation can, at least in part, be attributed to the mostly paramagnetic and often highly dynamic nature of the catalytic species as well as a number of energetically accessible spin states, making the experimental characterization and theoretical description of the catalytic pathways particularly cumbersome. Initiated by the pioneering work of Trovitch and co-workers a few years ago,40 the study of manganese © 2018 American Chemical Society

compounds for carbonyl activation and reduction has produced various intriguing catalysts.41−51 From a systematic point of view, these catalysts can be discriminated by their formal oxidation state (MnI vs MnII) and hypothesized mechanistic pathway (ligand-assisted outer-sphere vs innersphere, Figure 1 top). Notably, there appears to be little doubt about the nature of the catalytic cycle and associated intermediates for the diamagnetic MnI (transfer) hydrogenation catalyst A,52−54 which strongly resembles its antecedent FeII analog,25,55−58 both in structure and reactivity. In contrast, the mechanism by which Trovitch’s original (B) and modified (C) catalysts operate in the hydrosilylation and dihydrosilylation of ketones and carboxylates is significantly less explored, and preliminary initial rate kinetics and a set of control experiments have only just provided the basis for a proposal of the catalytic cycles involved.59,60 The application and mechanistic analysis of 3d metalmediated transformations has been a major concern of our Received: May 22, 2018 Published: June 26, 2018 9244

DOI: 10.1021/jacs.8b05340 J. Am. Chem. Soc. 2018, 140, 9244−9254

Article

Journal of the American Chemical Society

that the principal reaction pathway may change during the course of the conversion of the substrate. The identification of these factors requires a range of complementary analytical tools, in particular, a complete analysis of the reaction kinetics, combined with theoretical modeling. They are crucial for an insight-based “rational” optimization of a given catalyst which is presented in this work.

■

RESULTS AND DISCUSSION A principal aim of the study has been to uncover the mechanistic basis for the fast and highly selective hydroboration route as opposed to the slow direct insertion underlying catalytic hydrosilylations. It will first focus on the conversion of the manganese alkyl precatalyst to several active species, all of which play a role in the catalytic cycle. As indicated, a complete picture of the catalytic cycle could only be obtained with the aid of a comprehensive kinetic analysis based on high-quality concentration vs time profiles. The onepot kinetic profiling approach we have chosen in our study meets this prerequisite, and we will combine this method with the well-stablished reaction profile kinetic analysis (RPKA) which has been pioneered by Blackmond71 and was recently extended by Burés through his variable time normalization (VTN) analysis.72,73 The insights provided by this approach will then be supplemented by a mechanistic model obtained from density functional theory (DFT) calculations. This model in turn will allow the simulation of the critical steps of the catalytic cycle by means of a numerical solution of the associated differential rate equations (vide inf ra). Our findings suggest that the full kinetic profile can only be described consistently if a second productive cyclewith an alkoxideborane adduct as the hydrogen donoris included in the kinetic modeling, and, depending on the reaction conditions, this may actually constitute the dominating reaction pathway. We will conclude with a comparative overview of the factors that govern the reactivity of the manganese catalyst and the related iron complex. Catalyst Activation of the Manganese Alkyl Complex. The activation process was investigated in two ways: First, we studied the stoichiometric reaction of the substrates and other co-reagents with the precatalyst, leading to the formation of hypothesized catalytic intermediates (Scheme 1). In a second step, we investigated the activity of these intermediates in the catalytic cycle by monitoring the reaction progress of ketone hydroboration by low-temperature in situ react-IR techniques (Figure 3). Initially, we studied the activation of the precatalyst by its reaction with the fluorinated ketone substrate 4′-fluoroacetophenone, with the aim of monitoring the potential conversion of the alkyl species to an alkoxide by 19F NMR spectroscopy. Surprisingly, only the reversible coordination of the ketone was seen at rt, while the expected insertion product remained undetected even after longer periods of time (Scheme 1a). This reactivity strongly contrasts with that of many other known manganese alkyl compounds which are potent nucleophilic alkylating reagents.74−77 On the other hand, the addition of pinacol borane to a solution of the precatalyst yielded a clean and quantitative conversion to Me3SiCH2BPinthe σ-bond metathesis product of the reaction of the alkyl speciesand, presumably, a transient manganese hydride complex (Scheme 1b). In a similar fashion, the alcoholysis of 1 with 1-(4-fluorophenyl)-1ethanol produced a new 19F NMR resonance, corresponding to

Figure 1. Previous mechanistic studies on manganese-catalyzed reductions (top), our work in the field of enantioselective hydroboration, and a brief overview of the aspects covered in this manuscript (bottom left, and right).

group.61−66 Inter alia, we have been able to elucidate the elementary steps of the iron-catalyzed highly enantioselective hydrosilylation of ketones supported by the bis(oxazolinylmethylidene)isoindoline (“boxmi”) pincer ligand via a combined kinetic and computational approach.67−69 Although this boxmi iron-catalyst remains a benchmark system in the field with regard to both activity and selectivity, we have recently reported a related manganese hydroboration catalyst which substantially surpasses the iron system with respect to catalyst performance and substrate scope (Figure 1 bottom).70 In particular, our initial results for the manganese catalyst pointed toward the presence of two different, enantiodivergent hydroelementation pathways for this base metal catalyst, viz. a slow direct transfer in the hydrosilylation and a boraneassociated hydroboration process which is faster by several magnitudes (Figure 2). This proposal also accounted for the

Figure 2. Initial mechanistic proposal for the enantioselective manganese-catalyzed hydroboration (left) and hydrosilylation (right) of ketones, L = boxmi.

observation that the hydrosilylation provides the (R)enantiomer with an enantioselectivity which was independent of the silane employed. On the other hand, the activity and selectivity of the hydroboration was found to vary considerably depending on the borane, the stoichiometry, and the amount of coordinating additives, typically providing the (S)enantiomer with excellent enantiodiscrimination. A detailed study of the reaction mechanism has uncovered competing reaction pathways which operate depending on the reaction conditions and chosen stoichiometries, influencing the observed catalyst activity and selectivity, and displaying different kinetic characteristics. In fact, it has been found 9245


Article


With these data related to catalyst activation in hand, we performed the catalytic hydroboration of 4′-fluoroacetophenone under three different initial conditions, starting from (i) the precatalyst 1, (ii) the in situ generated manganese hydride, and (iii) the manganese alkoxide 2. Initial experiments showed that a 0.3 mol % catalyst loading and a temperature of −30 °C were required to collect kinetic data of sufficient time resolution and quality (Figure 3). The kinetic profile of the reduction starting from the precatalyst 1 featured a distinct, albeit short, induction period with a characteristic sigmoidal c−t profile, which is indicative of an activation process prior to the catalytic cycle. This finding rules out a role of the alkyl species as an active catalytic component and is in agreement with the observed reaction represented in Scheme 1a, in which a coordination of the ketone was the sole interaction between the (pre)catalyst and the substrate. In contrast, both the hydride and the alkoxide immediately initiated the catalytic conversion, suggesting that these complexes are either direct catalytic intermediates or are converted to such intermediates at rates that are faster than those of the underlying catalytic cycle. Kinetic Analysis and Initial Profiling of the Catalytic Cycle. Having identified the active catalytic species of the hydroboration, we investigated the rate law of the reaction by varying the concentrations of all components and monitoring their effect on the initial rate kinetics of the system. Changes in the catalyst loading led to an approximately linear alteration in the initial rate, suggesting a first-order dependence on the catalyst concentration. Similarly, variation of the ketone and the borane concentration produced first-order dependences for these reactants (see Supporting Information). However, given the quality of the raw kinetic data, the analysis also revealed a significant deviation of such an idealized rate equation. In addition, we found the late stage regime of the c−t profiles to deviate from the expected exponential behavior (vide infra), suggesting a change in reaction order during the course of the conversion. Since the excellent reproducibility and consistency of our experiments ruled out an inherent error associated with the measurement, we considered two underlying effects as possible reasons for our observations: (i) an overall non-integer reaction order due to a combination of forward/reverse and irreversible reactions with rate constants of a similar magnitude, or (ii) a complex reaction mechanism that is composed of more than one productive cycle with varying fractions of the dominating steps depending on the specific initial stoichiometry as well as overall conversion in the catalytic transformation. To discriminate between these possibilities a set of independent kinetic experiments featuring high intra- and interdata set precision was required. To this end, we chose a one-pot sequential approach due to the inherent advantages of this method. These include an excellent consistency between different substrate additions and the possibility to study the kinetics both under synthetically relevant and strongly deviating, that is, explicitly artificial conditions.78−81 Covering this range was particularly important to identify a potential interplay of competing pathways. In addition, this technique was also compatible with the variabletime normalization approach, thus allowing instant assessment of the reaction kinetics (vide infra). Importantly, such an analysis would be possible even in the case of reversible or irreversible deactivation processes, provided the associated kinetic relation was found.73,82

Scheme 1. Control Experiments To Identify Precatalyst Activation Pathways and Manganese Species That Are Part of the Active Catalytic Cyclea

a

Stoichiometric reactions were performed at rt in toluene.

Figure 3. Exemplary reaction profile of the catalytic hydroboration of 4′-fluoroacetophenone with HBPin at −30 °C in the presence of Ph boxmiMnCH2SiMe3 (1, red), the in situ generated PhboxmiMnH (blue), and PhboxmiMnOR (2, green). The profile was obtained from a time-resolved FT-IR spectrum of the reaction course and represents the extracted spectroscopic trace of the ketone concentration obtained by peak-height analysis.

a hypothesized manganese alkoxide 2 (Scheme 1c). Notably, addition of HBPin to this complex led to release of the product of the catalysis, which is also liberated during the stoichiometric reaction between the in situ generated hydride complex and the ketone (for details, see Supporting Information). Although we have been unable to crystallize and thus fully characterize the transient manganese hydride, we succeeded in isolating an alkoxide complex 3 (Ar = Ph) for characterization by mass spectrometry and elemental analysis. The structural resemblance between the boxmi iron and manganese complexes, however, suggests that 2 and 3 feature bonding metrics similar to the known iron alkoxide complex.67,70 9246


Article

Journal of the American Chemical Society A typical one-pot sequential reaction profile collected by time-resolved FT-IR spectroscopy is illustrated in Figure 4.

[P] = kobs

∫ [Ketone]α × [HBPin]β × [Mn]γ dt

According to this expression, a graphical representation of the product concentration vs the normalized time-axis should give a straight line with a slope of kobs, if the macroscopic reaction orders α, β, γ have been chosen correctly. Although this method has been used previously to establish reaction orders by achieving an overlay of several measurements, a single reaction profile should, in principle, be sufficient for this purpose. Whereas such an overlay is not possible in the case of a one-pot sequential profile, the normalized profile of each addition should form a straight line which is a natural continuation of the other data sets, thus allowing for a reliable calculation of the reaction orders by the least-squares method. We have applied this analysis to the profile obtained by the one-pot sequential addition method (Figure 5). Notably, assuming a reaction order of one for all components, an initial linear correlation is observed repeatedly, immediately after the addition of the ketone substrate for several turnovers (Figure 5 left). On the other hand, toward the end of the conversion, an increasingly positive curvature is detected. This observation corresponds to a reaction which is close to first order with respect to all components at the outset and features a noninteger, nearly (pseudo)-zeroth-order dependence for at least one reagent toward the end. The change in overall reaction order leads to a relative acceleration during the course of the reaction, and thus, completion is reached faster than expected for a standard integer order reaction. Accordingly, an excellent linear fit could be obtained when adjusting the individual reaction orders by the least-squares method (Figure 5 right), with gross macroscopic reaction orders of α = 0.58, β = 0.35, and γ = 1 for substrate, reducing agent, and catalyst, respectively. In an attempt to clarify the previously discussed trends, we have performed a comprehensive kinetic modeling, aiming to reproduce our experimental c−t data based on the simplified elementary steps of the initially proposed basic catalytic cycle illustrated in Figure 2. Simulations were carried out within Matlab R2017b software suite, deriving ordinary differential equations (ODEs) from the model reactions using massbalance principles.83 The left-hand side (LHS) of each ODE is the time derivative of a model quantity and the right-hand side (RHS) is defined using reaction fluxes that are derived from reaction rates and rate rules.

Figure 4. One-pot sequential reaction profile of the hydroboration of 4′-fluoroacetophenone with HBPin catalyzed by Ph boxmiMnCH2SiMe3 (1) under standard conditions (−40 °C, [Mn]/M = 1.25 × 10−3, [HBPin]0/M = 1.14, [Ketone]0/M = 0.068, 0.109, 0.138, 0.166, 0.193, 0.235. The borane was added first to the mixture to ensure complete catalyst activation prior to the run. For clarity, only every fourth data point is shown (blue ketone, red HBPin, green product).

The addition of a large excess of HBPin to the precatalyst was chosen as the starting point for the sequential profiling. Having initially equilibrated this setup, we were able to obtain a range of well-behaved kinetic profiles upon each addition of a varying quantity of the ketone. Importantly, the excess of the reducing agent shifted from >15-fold to smaller than 2-fold throughout the experiment. Over the course of this reaction, the overall catalyst content was 0.2 mol %. Further investigations have shown that loadings as low as 0.01 mol % could be employed without a significant loss of catalyst activity and selectivity, highlighting the excellent performance of this catalytic system. Variable-time normalization analysis is a valuable tool to reliably determine the reaction order in each component. Within this framework, the time dimension of the c−t profile is transformed in such a way that the kinetic concentration effect of each component is removed from the profile.72,73 The elegance of this method is immediately evident from inspection of the integrated rate equation:

Figure 5. Variable time normalization analysis of the kinetic profiles shown in Figure 4 employing two sets of reaction orders α, β, γ. Different colors represent the consecutive additions of the ketone substrate after consumption of the previously added substrate portion. For clarity only every fourth data point is shown. 9247


Article

Journal of the American Chemical Society The system of ODEs is represented as ẋ = S × v. Here, ẋ is an M-by-1 vector containing the rates of change for the model quantities, S is an M-by-R stoichiometry matrix, and v is an Rby-1 flux vector. M equals the total number of species, and R equals the total number of reactions in the model. Within the deterministic framework, we used Matlab’s ode15s solver, specifically intended for stiff ODE problems, as the numerical method (for details see Experimental Section and Supporting Information).84 In our initial simulation we accounted for the following reaction steps: (i) an initial ketone insertion, (ii) a reversible coordination of the borane to a manganese alkoxide, and (iii) an irreversible metathesis step of the alkoxide-borane adduct with a concomitant release of the product. This basic catalytic cycle from Figure 2 is represented by reaction fluxes f1 to f4:

DFT Study of Mechanistic Alternatives. The identification of a second productive cycle as the reason for the characteristic features of the c−t profile initiated our search for alternative mechanistic pathways of the hydroboration. To this end, we employed unrestricted DFT (BP86 and PBE0 functionals) to study the reactivity of the system and kinetically feasible pathways (see Figure 7 and the Supporting Information for an overview of DFT results). Having ruled out radical transfer scenarios via the addition of radical traps such as 9,10-dihydroanthracene and triphenylmethane,70 we considered several variations of the insertion and metathesis steps (Figure 8 top). As expected, a relatively high activation barrier was found for the direct insertion of the ketone into the Mn−H bond (Mts1-R, ΔG⧧ = 19.0 kcal/mol) with the (R)enantiomer as the major product. Notably, the coordination of the ketone is endergonic, highlighting the poor Lewis acidity of the d5 high-spin manganese center and its low propensity to bind an additional fifth donor. The insertion barrier is reduced significantly in the borane-assisted process, with the hydrogen atom attached to the boron center playing the role of the hydrogen source (Bts1S, ΔG⧧ = 11.1 kcal/mol), producing the (S)-enantiomer as the major species. Here, the coordination of the borane is found to be exergonic, whereas the subsequent ketone coordination step is endergonic, albeit somewhat more favored compared to the direct insertion alternative. It should be noted that the rearranged product is not a minimum on the Gibbs energy hypersurface, leading to an instant release of the borane after the insertion process (Bint2). Interestingly, the isomeric species in which the ketone is coordinated to the borane and the hydrogen is transferred from the manganese center are much higher in energy and thus not part of a viable pathway for the hydroboration (R1ts1-S, R2 ts1-S, Figure 8 top right). Regardless of the insertion step, the formation of the alkoxide was found to be irreversible under the reaction conditions (ΔG = −23.2 kcal/mol). Starting from the alkoxide intermediate, we were able to identify two feasible scenarios for product release, one of them being the natural counterpart of the direct insertion, viz. direct metathesis. Following the low-barrier coordination of the borane (Bts2-S, ΔG⧧ = 4.3 kcal/mol), the direct metathesis step is found to be rate limiting, regenerating the manganese hydride with an activation barrier of ΔG⧧ = 22.4 kcal/mol (Bts3-S) and directly affording the product of the catalysis. Alternatively, an additional ketone can bind to the alkoxideborane adduct Bint3. The subsequent borane-assisted alkoxide exchange step features a low activation barrier (Ats1-S, ΔG⧧ = 9.6 kcal/mol) through the merger of a metathesis and an insertion process. In this case, the product is released and the manganese alkoxide species is regenerated. Strikingly, the borane-assisted alkoxide exchange also leads to the preferential formation of the (S)-enantiomer, which is in agreement with the experimentally observed enantiomer ratio. For a 3D graphical representation of selected enantiodiscriminating model steps see Figure 8 bottom. Due to the potential non-innocence of the boxmi framework, we evaluated the role of different spin states for the catalytic cycle. All relevant intermediates and transition states were found to feature a sextet ground state at the employed level of theory, with the excited quartet and doublet states having significantly higher energies (typically at least 5 kcal/ mol). This makes potential internal catalysis pathways via twostate reactivity very unlikely and is in contrast to many other

f1 = kbi[MnH‐HBPin][Ketone] f2 = ka[MnOR][HBPin] − k −a[MnOR‐HBPin]

f3 = km[MnOR‐HBPin] f4 = kh[MnH][HBPin] − k −h[MnH‐HBPin]

Although excellent agreement between measured and computed c−t profiles could be achieved for the early stage of the reaction, a significant discrepancy was evident for the end phase of the conversion (Figure 6, basic cycle). This

Figure 6. Global fitting of the experimental data (gray) to the basic productive catalytic cycle from Figure 2 (red) and to two productive cycles (blue). For clarity, only data points of the first injection are shown. The early and late stages of the reaction have been highlighted with green and light orange, respectively.

divergence could not be eliminated even in the case of noninteger reaction orders, demonstrating that the proposed catalytic cycle did not account for the full mechanistic complexity posed by this manganese catalyst. Therefore, we explored several simple mechanistic effects, such as a reversible catalyst inhibition by the substrate or the reducing agent, as the source for these deviations (for details see Supporting Information).79 Importantly, while such additions to the mechanistic picture led to a slight improvement of the overall fitting quality, they failed to consistently reflect the characteristic features of the reaction profile. A systematic study of the effects of the addition of more complex mechanistic steps finally led us to believe that a second productive cycle (vide infra) is the source of the observed late stage behavior (Figure 6, full simulation). 9248


Article


Figure 7. Gibbs free enthalpy profile of the manganese-catalyzed hydroboration of ketones obtained from unrestricted DFT calculations calculated for 233 K and 1 atm. All energies are relative to the separated starting materials and the manganese hydride (pre)catalyst and calculated in kcal mol−1 (BP86/def2TZVP//PBE0/def2QZVPP/SMD(toluene)). Different pathways are highlighted in green, red, and blue. Unless stated otherwise, all energies refer to the sextet spin state and the (S)-alkoxide attached to Mn. For details see Experimental Section and the Supporting Information.

insertion/alkoxide release process (Figure 9 top). Interestingly, the cycles intersect at the alkoxide intermediate. Whereas the former requires the formation of a hydride and an alkoxide intermediate, the latter also proceeds in the absence of a manganese hydride. Kinetic Modeling of Borane-Assisted Insertion and Metathesis Pathways. Having established a more detailed mechanistic model of the underlying processes of the manganese-catalyzed hydroboration, we employed this information as the basis for an improved kinetic analysis. Accounting for both cycles in the kinetic simulation via the addition of an additional flux f5 = kai[MnOR-HBPin][Ketone] considerably improved the quality of the fit (see Figure 6, full simulation and Figure 9 bottom), finally reflecting the characteristic features of the c−t profile over the whole course of the reaction. The kinetic parameters obtained from the fitting process indicate that the profile can be deconvoluted into a (major) borane-assisted alkoxide exchange path and the (minor) borane-assisted insertion/metathesis cycle. The former is the dominating pathway under synthetic conditions and dictates in the early and midstage of the reaction, whereas the latter plays a crucial role for its late-stage characteristics. The distinct separation of the dominant time regimes of the two pathways stems from their disparate kinetic properties: The alkoxide-

Figure 8. Overview of insertion and metathesis models (top) and 3D graphical representation of the favored enantiodiscriminating transition states Bts1-S and Ats1-S (bottom left and right, respectively). Energetically favored transition states are highlighted in green.

3d metal-mediated transformations, in which such processes have been attributed a key role.85−88 In summary, the DFT calculations revealed two feasible catalytic cycles, that is, (i) a borane-mediated insertion/direct metathesis pathway and (ii) a borane-assisted, combined 9249


Article


Figure 10. Simulated kinetic profiles: Estimated fraction of product formed via the alkoxide exchange (blue) and direct metathesis (red) pathways over the course of the reaction and corresponding reaction rates (dotted lines) under various conditions.

Figure 9. Proposed catalytic cycle employed as the simulation model to extract kinetic parameters (top) and global fitting of the experimental data to this model (bottom).

37%) due to its pseudo zeroth-order kinetics. Whereas the product provenance is almost independent of the borane concentration, it is highly sensitive to (relative) changes in the ketone and catalyst concentrations. Providing high concentrations of the ketone strongly favors the alkoxide exchange cycle, while lower loadings trigger product formation preferentially via the direct metathesis pathway (Figure 10b,c). For the extreme case of a highly diluted catalytic reaction ([Ketone]0 and [HBPin]0 < 0.02 M), this provides the product almost exclusively via the direct metathesis pathway. Complementary Evidence for the Presence of Two Productive Pathways. The sensitivity of the catalytic cycles toward concentration changes should, in principle, allow a semi-independent study of the features of both pathways by choosing the appropriate reaction conditions. An additional characterization of the catalyst by complementary procedures such as Hammett correlations, kinetic isotope effect (KIE) measurements, and Eyring plots of the enantiomeric ratio was therefore considered to provide additional and complementary evidence for the presence of two productive catalytic cycles. The H/D KIE resulting from the employment of HBPin vs DBPin was measured in order to evaluate the contribution of the X−H bond scission to the transition state of the ratelimiting step. Under typical catalytic conditions (−40 °C, high concentration regime) a direct competition experiment furnished a KIE of kH/kD = 1.2 ± 0.05. The small influence of H/D exchange is in good agreement with the computed value of 1.22 for the rate-limiting alkoxide-exchange step obtained from normal coordinate analysis of the DFT reactant and the transition-state structures (Aint1 → Ats1) by the Bigeleisen−Mayer approach.89−92 Interestingly, a significantly larger effect (kH/kD = 1.6 ± 0.05) was detected when the experiment was carried out at higher dilution, pointing toward a more pronounced contribution of the X−H bond cleavage in the formation of the rate-determining transition state. The increased KIE obtained upon dilution of the reaction matches the expected mechanistic switch from an alkoxide exchange to

exchange mechanism features an overall non-integer reaction order due to rate constants of similar magnitudes for the addition (ka = 2.12 ± 0.12 M−1 s−1, k−a = 0.49 ± 0.05 s−1) and insertion (kai = 3.74 ± 0.06 M−1 s−1) steps. In the early and midstages of the catalysis this cycle is fast, but owing to its dependence on the substrate and the borane concentrations, it significantly loses momentum over the course of the reaction. On the other hand, the slow direct metathesis step (km = 0.076 ± 0.004 s−1) is clearly rate-limiting for the borane-assisted insertion/metathesis mechanism, producing the expected pseudo zeroth-order kinetic profile with a linear consumption of the starting materials regardless of the composition of the reaction mixture. This behavior is also evident from inspection of the reaction fluxes f R used for the full kinetic simulation (* denotes the flux of the pseudozeroth-order rate-limiting direct metathesis reaction): f1 = kbi[MnH‐HBPin][Ketone] f2 = ka[MnOR][HBPin] − k −a[MnOR‐HBPin] f3 = km[MnOR‐HBPin]

*

f4 = kh[MnH][HBPin] − k −h[MnH‐HBPin]

f5 = kai[MnOR‐HBPin][Ketone]

It is worthwhile noting that the fraction of the product formed through either one of the catalytic cycles is thus controlled by the absolute concentration of the reagents and the ketone to catalyst ratio (Figure 10). Under the standard catalytic conditions for the kinetic measurement (Figure 10a), the rate crossing point of both mechanisms is reached at approximately 79% overall conversion, with 57 and 22 percentage points formed through alkoxide exchange and direct metathesis steps, respectively. Upon reaction completion, the ratio of the product provenance shifts slightly in favor of the direct metathesis pathway (63% vs 9250


Article


concentration regimes (Figure 12). Despite the small enantiomeric excess range used for the study (84.5−96.5%

a direct metathesis pathway, the latter featuring a large singlestep KIE of 3.27 according to the computed structures (Bint4 → Bts3). It should be pointed out that the difference between the experimentally determined KIE and the computed value arises from the multistep character of the reaction and can be attributed to a shift of the catalyst resting states upon dilution with an estimated upper bound for the KIE of kH/kD = 1.9−2.6 (for a product distribution simulation, see Supporting Information). As a result, adduct formation between MnOR and HBPin and ketone coordination processes, rather than direct metathesis of the alkoxide borane adduct, becomes increasingly rate-determining under such dilute conditions. This effect partially reduces the large single-step KIE expected for the direct metathesis reaction. The rate laws discussed in the previous section suggest that the nature of the ketone should be important for both mechanistic pathways and could provide valuable insights into the electronic structure of the transition states from the perspective of the ketone. An analysis of the relative reaction rates of various para-substituted acetophenone derivatives (H, F, Cl, Br, Me, OMe, CO2Me) obtained in a competitive 13C NMR experiment under the typical catalytic conditions gave a Hammett parameter ρ = 0.52 ± 0.02 for the highconcentration regime (Figure 11).

Figure 12. Illustration of the temperature dependence of the enantioselectivity for the manganese-catalyzed hydroboration of 4′fluoroacetophenone in the high- and the low-concentration regimes (red and blue, top) and Eyring-type analysis for both selectivity determining steps obtained for the whole temperature range.

ee), we could establish two distinct linear relationships indicative of two different selectivity determining steps in the temperature range of interest; a higher enantioselectivity being obtained for the more dilute sample (84.5%ee vs 89%ee at rt and 93.5%ee vs 96.5%ee at −40 °C). Graphical analysis of the Eyring plot gave difference values of the activation parameters for a re- vs si-face addition for both steps, with a slightly larger difference in free enthalpies of activation ΔΔG⧧sel,233 K found for the borane-mediated process (see insets in Figure 12). As expected from the DFT calculations, both mechanisms favor the si-face attack over a re-face addition and are characterized by a small activation entropy difference ΔΔS⧧sel typical of two structurally closely related diastereotopic transition states. Comparison to the Related Iron Catalyst. The close structural resemblance of the boxmi iron and manganese catalysts as well as their high activity and contrasting selectivity fosters the search for the principles that govern these observations. From a kinetic point of view, it is the markedly different electrophilicity of these base metal catalysts that facilitates the direct insertion pathway for iron and essentially prevents it for manganese. Notably, despite the different electronic properties, the following direct metathesis step was found to be rate-limiting for both metals.68 On the other hand, the ability of the manganese hydride catalyst to accomplish an efficient hydrogen transfer in the presence of a suitable borane highlights the intrinsically high nucleophilicity of neutral boxmi manganese hydride complexes which contrasts with the observed properties of the previously reported iron congeners.

Figure 11. Hammett correlations for the manganese-catalyzed hydroboration of para-substituted acetophenone derivatives under concentrated conditions.

The overall positive, but small value reflects the buildup of negative charge in the alkoxide exchange step which, however, is the result of an overall partial compensation of the buildup of negative charge in the ketone and of positive charge in the alkoxy substituent. Notably, a distinct observation of the Hammett parameter for the direct metathesis mechanism is impeded by the small variation of the fraction of product formed by this pathway for a reaction with an excess of ketone, regardless of the concentration regime employed (4% vs 32%, cf. Supporting Information). Although the slightly reduced substitution constant of ρ = 0.49 ± 0.02 can be interpreted in terms of the larger fraction of product formed by the direct metathesis pathway, the small change in the value and the relatively large error associated with it preclude a definitive assignment in this case. The complex mechanistic picture uncovered during our studies precluded a standard Eyring analysis of the temperature dependence of initial rates. However, we have been able to perform an Eyring-type analysis of the temperature-dependence of the enantiomer ratio of the hydroboration in both

■

CONCLUSION The unprecedented reactivity of the (boxmi)MnII alkyl complexes as precatalysts for the hydroboration of ketones and their unusual enantioselectivity trends encountered in the presence of varying reducing agents led us to the mechanistic investigation of activation and conversion processes of this 9251


Article


pentane, producing the alkoxido complex in excellent yield (86%, 90.5 mg, 149 μmol). EA: calcd C: 71.05%, H: 5.13%, N: 6.90%; found: C: 71.15%, H: 5.35%, N: 6.57%. Kinetic Experiments. A custom-made, flame-dried reaction flask was equipped with a magnetic stirring bar and an IR probe of an iC-IR spectrometer (Mettler Toledo ReactIR 15 with a SiComp probe head, connected with DST-AgX-fiber optics, 9.5 mm diameter). The flask was purged three times and cooled to the set temperature (typically 40 °C) using a cryostat. A solution of the catalyst in 4 mL abs. toluene was added, and after an initial thermal equilibration a background of 1024 scans was collected over the full scan range (3000 cm−1 to 700 cm−1) at 4 cm−1 intervals. The measurement was then started, collecting 30 scans every 15 s. After the addition of 1 mL neat HBPin in one portion, the flask was thermally re-equilibrated, and the final collection stage was initiated by the addition of the first ketone portion. Peak height analysis in the range from 1717 to 1653 cm−1 was used to assess the ketone concentration. Kinetic Analysis. Data fitting of experimental reaction profiles and product distribution simulations were performed with the Simbiology program as implemented in the Matlab R2017b software suite.83 Ode15s with absolute tolerance scaling was employed as the solver with constrained nonlinear least-squares problems (lsqnonlin) as the estimation method.84 The sequential addition of the ketone substrate was included as dosing information. The kinetic model and a list of the final fitting and simulation settings are printed in the Supporting Information. Computational Details. All calculations were performed with the software suite Gaussian 09 Revision D.01 at the DFT level of theory.99 All geometry optimizations were carried out with Ahlrichs’ def2 basis set family,100,101 employing def2TZVP basis functions for Mn, N, B and def2SVP for the remaining elements along with the BP86 functional, and no structural simplifications have been applied.102,103 Tight convergence criteria and an ultrafine integration grid have been used throughout all calculations. Single point energies of the optimized structures were calculated with PBE0/def2QZVPP as the computational tool, accounting for solvation effects of toluene with the SMD model.104−106 Dispersion effects were included with Grimme’s recent dispersion correction GD3 as implemented in Gaussian.107,108 The presence of a single imaginary frequency for transition states and their absence in the case of stationary points was confirmed by frequency calculations. Visualizations were created with CYLview software.109 Additional information on experimental and computational methods can be found in the Supporting Information.

intriguing catalytic platform. With the aid of a combination of stoichiometric and kinetic experiments we could show that manganese alkyls can enter the catalytic cycle as hydride and alkoxide species after metathesis with HBPin or alcoholysis with ROH, respectively. The kinetic study of initial rates of this transformation pointed toward a complex interplay of the effects of the reagent concentrations on the reaction rate, and we hence performed a complete kinetic profile analysis based on a one-pot multiple dose approach to extract the kinetic parameters of this system. The characteristic late-stage kinetics observed therein could only be explained by incorporating two productive cycles in the analysis, and this finding was supported by Hammett correlations, KIE measurements, and Eyring analysis as well as the DFT modeling of the catalytic pathways. The mechanistic proposal which rests upon these results consists of a main pathway in which a metathesis step is accompanied by a simultaneous ketone insertion (Figure 9 top left), yielding an overall alkoxide exchange process, and a side reaction made up of a borane-assisted insertion followed by a direct metathesis pathway (Figure 9 top right). Our study also highlights the mechanistic complexity encountered in the study of 3d metal-mediated transformations in general and of manganese-catalyzed reductions in particular. This should open new perspectives in this burgeoning field of research, in which in depth mechanistic investigations have been frequently heavily impeded by the intricacy of the catalytic reaction networks involved.

■

EXPERIMENTAL SECTION

Unless stated otherwise, all manipulations have been carried out under exclusion of air and moisture using standard Schlenk and glovebox techniques. Argon 5.0 purchased from Messer Group GmbH was dried over Granusic phosphorus pentoxide granulate before use, and solvents were dried over activated alumina columns using M. Braun SPS 800 or deuterated solvents according to standard literature-known methods and stored in glass ampules under an inert gas atmosphere.93 NMR experiments were performed on Bruker Avance (400 MHz, 600 MHz) spectrometers. Chemical shifts (δ) are reported in parts per million (ppm) and are referenced to residual solvent signals or carbon resonances.94,95 BF3·OEt2 (11B), CCl3F (19F), and SiMe4 (29Si) were used as external standards. Mass spectra were acquired on a JEOL JMS-700 magnetic sector (LIFDI) spectrometer at the mass spectrometry facility of the Institute of Organic Chemistry at the University of Heidelberg. Elemental analyses were carried out on an Elementar vario MICRO cube in the Microanalysis Laboratory of the Heidelberg Chemistry Department. HPLC and GC traces were collected on an Agilent 1200 Series chromatograph employing a chiral Daicel OJ-H, column. Mn(CH2SiMe3)2,96,97 boxmi ligands,61 and DBPin98 were prepared following literature procedures. Liquid substrates and reagents were degassed by three successive freeze−pump−thaw cycles and used without further purification. Manganese salts were purchased with a trace metal purity of 99.99% or higher. Preparation of the Manganese Alkyl Complex 1.70 A mixture of Mn(CH2SiMe3)2 (113 mg, 493 μmol, 1.0 equiv) and the boxmi ligand (214 mg, 493 μmol, 1.0 equiv) was dissolved in 10 mL toluene and stirred at rt for 1 h. The solvent was removed under reduced pressure furnishing the desired precatalyst Phboxmi-MnCH2SiMe3 (1) as a dark-brown solid in quantitative yield. EA: calcd C: 66.88%, H: 5.79%, N: 7.31%; found: C: 67.21%, H: 6.02%, N: 7.61%. Preparation of the Manganese Alkoxide Complex 3. To a solution of the alkyl complex 1 (100 mg, 173 μmol, 1.0 equiv) in 5 mL toluene was added an equimolar amount of (S)-1-phenyl-1ethanol (21.3 mg, 173 μmol, 1.0 equiv). The mixture was stirred for 1 h, the solvent was evaporated, and the residue was washed with

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.8b05340. Description of methods, selected NMR spectra, product distribution simulations and other additional information. Cartesian coordinates of the stationary and transition states of computed structures (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*[email protected] ORCID

Lutz H. Gade: 0000-0002-7107-8169 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We acknowledge the award of a predoctoral fellowship to V.V. from the Landesgraduiertenförderung (LGF Funding Program of the state of Baden-Württemberg) and generous funding by the University of Heidelberg as well as the Deutsche 9252


Article


(31) Ilies, L.; Nakamura, E. Iron-Catalyzed C−H Bond Activation. In C-H Bond Activation and Catalytic Functionalization II; Springer: Switzerland, 2015; pp 1−18. (32) Casitas, A.; Krause, H.; Lutz, S.; Goddard, R.; Bill, E.; Fürstner, A. Organometallics 2018, 37 (5), 729−739. (33) Sun, X.; Li, J.; Huang, X.; Sun, C. Curr. Inorg. Chem. 2012, 2 (1), 64−85. (34) Cera, G.; Ackermann, L. Top. Curr. Chem. 2016, 374 (5), 57. (35) Yahaya, N. P.; Appleby, K. M.; Teh, M.; Wagner, C.; Troschke, E.; Bray, J. T. W.; Duckett, S. B.; Hammarback, L. A.; Ward, J. S.; Milani, J.; et al. Angew. Chem., Int. Ed. 2016, 55 (40), 12455−12459. (36) Cai, S.-H.; Ye, L.; Wang, D.-X.; Wang, Y.-Q.; Lai, L.-J.; Zhu, C.; Feng, C.; Loh, T.-P. Chem. Commun. 2017, 53 (62), 8731−8734. (37) Yu, X.; Tang, J.; Jin, X.; Yamamoto, Y.; Bao, M. Asian J. Org. Chem. 2018, 7 (3), 550−553. (38) Wang, H.; Lorion, M. M.; Ackermann, L. Angew. Chem., Int. Ed. 2017, 56 (22), 6339−6342. (39) Chen, S.-Y.; Li, Q.; Wang, H. J. Org. Chem. 2017, 82 (20), 11173−11181. (40) Mukhopadhyay, T. K.; Flores, M.; Groy, T. L.; Trovitch, R. J. J. Am. Chem. Soc. 2014, 136 (3), 882−885. (41) Elangovan, S.; Topf, C.; Fischer, S.; Jiao, H.; Spannenberg, A.; Baumann, W.; Ludwig, R.; Junge, K.; Beller, M. J. Am. Chem. Soc. 2016, 138 (28), 8809−8814. (42) Elangovan, S.; Garbe, M.; Jiao, H.; Spannenberg, A.; Junge, K.; Beller, M. Angew. Chem., Int. Ed. 2016, 55 (49), 15364−15368. (43) Kallmeier, F.; Irrgang, T.; Dietel, T.; Kempe, R. Angew. Chem., Int. Ed. 2016, 55 (39), 11806−11809. (44) Perez, M.; Elangovan, S.; Spannenberg, A.; Junge, K.; Beller, M. ChemSusChem 2017, 10 (1), 83−86. (45) Elangovan, S.; Neumann, J.; Sortais, J.-B.; Junge, K.; Darcel, C.; Beller, M. Nat. Commun. 2016, 7, 12641. (46) Andérez-Fernández, M.; Vogt, L. K.; Fischer, S.; Zhou, W.; Jiao, H.; Garbe, M.; Elangovan, S.; Junge, K.; Junge, H.; Ludwig, R.; et al. Angew. Chem., Int. Ed. 2017, 56 (2), 559−562. (47) Peña-López, M.; Piehl, P.; Elangovan, S.; Neumann, H.; Beller, M. Angew. Chem., Int. Ed. 2016, 55 (48), 14967−14971. (48) Kelly, C. M.; McDonald, R.; Sydora, O. L.; Stradiotto, M.; Turculet, L. Angew. Chem., Int. Ed. 2017, 56 (50), 15901−15904. (49) Maji, B.; Barman, M. Synthesis 2017, 49 (15), 3377−3393. (50) Mastalir, M.; Glatz, M.; Pittenauer, E.; Allmaier, G.; Kirchner, K. J. Am. Chem. Soc. 2016, 138 (48), 15543−15546. (51) Widegren, M. B.; Harkness, G. J.; Slawin, A. M. Z.; Cordes, D. B.; Clarke, M. L. Angew. Chem., Int. Ed. 2017, 56 (21), 5825−5828. (52) Nguyen, D. H.; Trivelli, X.; Capet, F.; Paul, J.-F.; Dumeignil, F.; Gauvin, R. M. ACS Catal. 2017, 7 (3), 2022−2032. (53) van Putten, R.; Uslamin, E. A.; Garbe, M.; Liu, C.; Gonzalezde-Castro, A.; Lutz, M.; Junge, K.; Hensen, E. J. M.; Beller, M.; Lefort, L.; et al. Angew. Chem., Int. Ed. 2017, 56 (26), 7531−7534. (54) Garbe, M.; Junge, K.; Walker, S.; Wei, Z.; Jiao, H.; Spannenberg, A.; Bachmann, S.; Scalone, M.; Beller, M. Angew. Chem., Int. Ed. 2017, 56 (37), 11237−11241. (55) Bornschein, C.; Werkmeister, S.; Wendt, B.; Jiao, H.; Alberico, E.; Baumann, W.; Junge, H.; Junge, K.; Beller, M. Nat. Commun. 2014, 5, 4111. (56) Kallmeier, F.; Kempe, R. Angew. Chem., Int. Ed. 2018, 57 (1), 46−60. (57) Alberico, E.; Sponholz, P.; Cordes, C.; Nielsen, M.; Drexler, H.J.; Baumann, W.; Junge, H.; Beller, M. Angew. Chem., Int. Ed. 2013, 52 (52), 14162−14166. (58) Lange, S.; Elangovan, S.; Cordes, C.; Spannenberg, A.; Jiao, H.; Junge, H.; Bachmann, S.; Scalone, M.; Topf, C.; Junge, K.; et al. Catal. Sci. Technol. 2016, 6 (13), 4768−4772. (59) Mukhopadhyay, T. K.; Ghosh, C.; Flores, M.; Groy, T. L.; Trovitch, R. J. Organometallics 2017, 36 (18), 3477−3483. (60) Trovitch, R. J. Acc. Chem. Res. 2017, 50 (11), 2842−2852. (61) Deng, Q.-H.; Wadepohl, H.; Gade, L. H. Chem. - Eur. J. 2011, 17 (52), 14922−14928.

Forschungsgemeinschaft (DFG-Ga488/9-2). The computational work presented herein was supported by the bwHPC initiative and the bwHPC-C5 project provided through the associated compute services of the JUSTUS HPC facility located at the University of Ulm; bwHPC and bwHPC-C5 are funded by the Ministerium für Wissenschaft, Forschung und Kunst, and the Deutsche Forschungsgemeinschaft. The authors thank Dr. Torsten Roth for helpful suggestions.

■

REFERENCES

(1) Woods, B. P.; Orlandi, M.; Huang, C.-Y.; Sigman, M. S.; Doyle, A. G. J. Am. Chem. Soc. 2017, 139 (16), 5688−5691. (2) Mikhailine, A. A.; Maishan, M. I.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2012, 134 (29), 12266−12280. (3) Casey, C. P.; Guan, H. J. Am. Chem. Soc. 2009, 131 (7), 2499− 2507. (4) Park, Y.; Kim, Y.; Chang, S. Chem. Rev. 2017, 117 (13), 9247− 9301. (5) Jackson, E. P.; Malik, H. A.; Sormunen, G. J.; Baxter, R. D.; Liu, P.; Wang, H.; Shareef, A.-R.; Montgomery, J. Acc. Chem. Res. 2015, 48 (6), 1736−1745. (6) Weix, D. J. Acc. Chem. Res. 2015, 48 (6), 1767−1775. (7) Enthaler, S.; Junge, K.; Beller, M. Angew. Chem., Int. Ed. 2008, 47 (18), 3317−3321. (8) Fürstner, A. ACS Cent. Sci. 2016, 2 (11), 778−789. (9) Egorova, K. S.; Ananikov, V. P. Organometallics 2017, 36 (21), 4071−4090. (10) Rose, E.; Andrioletti, B.; Zrig, S.; Quelquejeu-Ethève, M. Chem. Soc. Rev. 2005, 34 (7), 573−583. (11) Gelalcha, F. G.; Bitterlich, B.; Anilkumar, G.; Tse, M. K.; Beller, M. Angew. Chem., Int. Ed. 2007, 46 (38), 7293−7296. (12) Chen, J.; Yoon, H.; Lee, Y.-M.; Seo, M. S.; Sarangi, R.; Fukuzumi, S.; Nam, W. Chem. Sci. 2015, 6 (6), 3624−3632. (13) Shen, D.; Saracini, C.; Lee, Y.-M.; Sun, W.; Fukuzumi, S.; Nam, W. J. Am. Chem. Soc. 2016, 138 (49), 15857−15860. (14) Niwa, T.; Nakada, M. J. Am. Chem. Soc. 2012, 134 (33), 13538−13541. (15) Cussó, O.; Garcia-Bosch, I.; Ribas, X.; Lloret-Fillol, J.; Costas, M. J. Am. Chem. Soc. 2013, 135 (39), 14871−14878. (16) Cussó, O.; Ribas, X.; Costas, M. Chem. Commun. 2015, 51 (76), 14285−14298. (17) Cussó, O.; Ribas, X.; Lloret-Fillol, J.; Costas, M. Angew. Chem., Int. Ed. 2015, 54 (9), 2729−2733. (18) Zhang, W.; Loebach, J. L.; Wilson, S. R.; Jacobsen, E. N. J. Am. Chem. Soc. 1990, 112, 2801−2803. (19) Katsuki, T. Coord. Chem. Rev. 1995, 140, 189−214. (20) Liao, S.; List, B. Angew. Chem., Int. Ed. 2010, 49 (3), 628−631. (21) Paradine, S. M.; Griffin, J. R.; Zhao, J.; Petronico, A. L.; Miller, S. M.; Christina White, M. Nat. Chem. 2015, 7 (12), 987−994. (22) Casey, C. P.; Guan, H. J. Am. Chem. Soc. 2007, 129 (18), 5816−5817. (23) Bauer, G.; Kirchner, K. A. Angew. Chem., Int. Ed. 2011, 50 (26), 5798−5800. (24) Zuo, W.; Lough, A. J.; Li, Y. F.; Morris, R. H. Science 2013, 342 (6162), 1080−1083. (25) Werkmeister, S.; Junge, K.; Wendt, B.; Alberico, E.; Jiao, H.; Baumann, W.; Junge, H.; Gallou, F.; Beller, M. Angew. Chem., Int. Ed. 2014, 53 (33), 8722−8726. (26) Bigler, R.; Huber, R.; Mezzetti, A. Angew. Chem., Int. Ed. 2015, 54 (17), 5171−5174. (27) Loup, J.; Zell, D.; Oliveira, J. C. A.; Keil, H.; Stalke, D.; Ackermann, L. Angew. Chem., Int. Ed. 2017, 56 (45), 14197−14201. (28) Su, M.; Li, C.; Ma, J. J. Chin. Chem. Soc. 2016, 63 (10), 828− 840. (29) Liu, W.; Ackermann, L. ACS Catal. 2016, 6 (6), 3743−3752. (30) Shang, R.; Ilies, L.; Nakamura, E. Chem. Rev. 2017, 117 (13), 9086−9139. 9253


Article

Journal of the American Chemical Society (62) Deng, Q.-H.; Wadepohl, H.; Gade, L. H. J. Am. Chem. Soc. 2012, 134 (6), 2946−2949. (63) Deng, Q.-H.; Wadepohl, H.; Gade, L. H. J. Am. Chem. Soc. 2012, 134 (26), 10769−10772. (64) Deng, Q.-H.; Bleith, T.; Wadepohl, H.; Gade, L. H. J. Am. Chem. Soc. 2013, 135 (14), 5356−5359. (65) Deng, Q.-H.; Rettenmeier, C.; Wadepohl, H.; Gade, L. H. Chem. - Eur. J. 2014, 20 (1), 93−97. (66) Bleith, T.; Deng, Q.-H.; Wadepohl, H.; Gade, L. H. Angew. Chem., Int. Ed. 2016, 55 (27), 7852−7856. (67) Bleith, T.; Wadepohl, H.; Gade, L. H. J. Am. Chem. Soc. 2015, 137 (7), 2456−2459. (68) Bleith, T.; Gade, L. H. J. Am. Chem. Soc. 2016, 138 (14), 4972− 4983. (69) Bleith, T.; Gade, L. H. J. Am. Chem. Soc. 2016, 138 (18), 6074− 6074. (70) Vasilenko, V.; Blasius, C. K.; Wadepohl, H.; Gade, L. H. Angew. Chem., Int. Ed. 2017, 56 (29), 8393−8397. (71) Blackmond, D. G. Angew. Chem., Int. Ed. 2005, 44 (28), 4302− 4320. (72) Burés, J. Angew. Chem., Int. Ed. 2016, 55 (6), 2028−2031. (73) Burés, J. Angew. Chem., Int. Ed. 2016, 55 (52), 16084−16087. (74) Cahiez, G.; Duplais, C.; Buendia, J. Chem. Rev. 2009, 109 (3), 1434−1476. (75) Cahiez, G.; Figadere, B. Tetrahedron Lett. 1986, 27 (37), 4445− 4448. (76) Cahiez, G.; Moyeux, A.; Buendia, J.; Duplais, C. J. Am. Chem. Soc. 2007, 129 (45), 13788−13789. (77) Cahiez, G.; Alexakis, A.; Normant, J. F. Synth. Commun. 1979, 9 (7), 639−645. (78) Strieter, E. R.; Blackmond, D. G.; Buchwald, S. L. J. Am. Chem. Soc. 2003, 125 (46), 13978−13980. (79) Singh, U. K.; Strieter, E. R.; Blackmond, D. G.; Buchwald, S. L. J. Am. Chem. Soc. 2002, 124 (47), 14104−14114. (80) Rosner, T.; Sears, P.; Nugent, W.; Blackmond, D. Org. Lett. 2000, 2 (16), 2511−2513. (81) Das, U. K.; Higman, C. S.; Gabidullin, B.; Hein, J. E.; Baker, R. T. ACS Catal. 2018, 8 (2), 1076−1081. (82) Burés, J. Top. Catal. 2017, 60 (8), 631−633. (83) Matlab R2017b software suite and SimBiology; MathWorks: Natick, MA, 2017. (84) Shampine, L. F.; Reichelt, M. W. SIAM J. Sci. Comput. 1997, 18 (1), 1−22. (85) Hu, L.; Chen, H. J. Am. Chem. Soc. 2017, 139 (44), 15564− 15567. (86) Schröder, D.; Shaik, S.; Schwarz, H. Acc. Chem. Res. 2000, 33 (3), 139−145. (87) Sun, Y.; Tang, H.; Chen, K.; Hu, L.; Yao, J.; Shaik, S.; Chen, H. J. Am. Chem. Soc. 2016, 138 (11), 3715−3730. (88) Holland, P. L. Acc. Chem. Res. 2015, 48 (6), 1696−1702. (89) Bigeleisen, J.; Mayer, M. G. J. Chem. Phys. 1947, 15 (5), 261− 267. (90) Brown, S. B.; Dewar, M. J. S.; Ford, G. P.; Nelson, D. J.; Rzepa, H. S. J. Am. Chem. Soc. 1978, 100 (25), 7832−7836. (91) Dewar, M. J. S.; Olivella, S.; Rzepa, H. S. J. Am. Chem. Soc. 1978, 100 (18), 5650−5659. (92) Bell, R. P. Trans. Faraday Soc. 1959, 55 (1), 1−4. (93) Armarego, W. L. F.; Chai, C. L. L. Purification of Laboratory Chemicals, 6th ed.; Elsevier/Butterworth-Heinemann: Amsterdam; Boston, 2009. (94) Fulmer, G. R.; Miller, A. J. M.; Sherden, N. H.; Gottlieb, H. E.; Nudelman, A.; Stoltz, B. M.; Bercaw, J. E.; Goldberg, K. I. Organometallics 2010, 29 (9), 2176−2179. (95) Gottlieb, H. E.; Kotlyar, V.; Nudelman, A. J. Org. Chem. 1997, 62 (21), 7512−7515. (96) Andersen, B. R. A.; Carmona-Guzman, E.; Gibson, J. F.; Wilkinson, G. J. Chem. Soc., Dalton Trans. 1976, 21, 2204−2211.

(97) Alberola, A.; Blair, V. L.; Carrella, L. M.; Clegg, W.; Kennedy, A. R.; Klett, J.; Mulvey, R. E.; Newton, S.; Rentschler, E.; Russo, L. Organometallics 2009, 28 (7), 2112−2118. (98) Braunschweig, H.; Guethlein, F.; Mailänder, L.; Marder, T. B. Chem. - Eur. J. 2013, 19 (44), 14831−14835. (99) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Petersson, G. A.; Nakatsuji, H.; Li, X.; Caricato, M.; Marenich, A. V.; Bloino, J.; Janesko, B. G.; Gomperts, R.; Mennucci, B.; Hratchian, H. P.; Ortiz, J. V.; Izmaylov, A. F.; Sonnenberg, J. L.; Williams-Young, D.; Ding, F.; Lipparini, F.; Egidi, F.; Goings, J.; Peng, B.; Petrone, A.; Henderson, T.; Ranasinghe, D.; Zakrzewski, V. G.; Gao, J.; Rega, N.; Zheng, G.; Liang, W.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Throssell, K.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M. J.; Heyd, J. J.; Brothers, E. N.; Kudin, K. N.; Staroverov, V. N.; Keith, T. A.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A. P.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Millam, J. M.; Klene, M.; Adamo, C.; Cammi, R.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Farkas, O.; Foresman, J. B.; Fox, D. J. Gaussian 09 Revision D.01; Gaussian, Inc.: Wallingford, CT, 2013. (100) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7 (18), 3297−3305. (101) Weigend, F. Phys. Chem. Chem. Phys. 2006, 8 (9), 1057−1065. (102) Becke, A. D. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38 (6), 3098−3100. (103) Perdew, J. P. Phys. Rev. B: Condens. Matter Mater. Phys. 1986, 33 (12), 8822−8824. (104) Adamo, C.; Barone, V. J. Chem. Phys. 1999, 110 (13), 6158− 6170. (105) Ernzerhof, M.; Scuseria, G. E. J. Chem. Phys. 1999, 110 (11), 5029−5036. (106) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B 2009, 113 (18), 6378−6396. (107) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys. 2010, 132 (15), 154104. (108) Grimme, S.; Ehrlich, S.; Goerigk, L. J. Comput. Chem. 2011, 32 (7), 1456−1465. (109) Legault, C. Y. CYLview, 1.0b; University of Sherbrooke: Quebec, 2009.

9254


One-Pot Sequential Kinetic Profiling of a Highly Reactive Manganese

Recommend Documents