) UvA-DARE (Digital Academic Repository) Investigating the Active Species in a [(R-SN(H)S-R)CrCl3] Ethene Trimerization System: Mononuclear or Dinuclear?

Cr-catalyzed ethene trimerization is an industrially important process to produce 1-hexene. Despite its industrial relevance, the changing oxidation state and the structural rearrangements of the metal center during the catalytic cycle remain unclear. In this study, we have investigated the active species in a [(R-SN (H)S (cid:0) R)CrCl 3 ] (R = C 10 H 21 ) catalyzed ethene trimerization system using a combination of spectroscopic techniques (XAS, EPR and UV/VIS) and DFT calculations. Reaction of the octahedral Cr III complex with modified methylaluminoxane (MMAO) in absence of ethene gives rise to the formation of a square-planar Cr II complex. In the presence of ethene (1 bar), no coordination was observed, which we attribute to the endergonic nature of the coordination of the first ethene molecule. Employing an alkyne as a model for ethene coordination leads to the formation of a dinuclear cationic Cr III alkyne complex. DFT calculations show that a structurally related dinuclear cationic Cr III ethene complex could form under catalytic conditions. Comparing a mechanism proceeding via mononuclear cationic


Introduction
Several large chemical companies are faced with an increase in demand for the shorter (1-butene, 1-hexene and 1-octene) linear alpha olefins (LAOs) due to their application as a comonomer in the production of linear low-density polyethylene (LLDPE). [1,2]Traditionally, LAOs are produced by ethene oligomerization catalysts which generate a statistical product distribution, e. g. the nickel-based SHOP catalyst. [3]However, due to the growing market demand for shorter LAOs, several companies (e. g.Sasol and Chevron Phillips Chemical) have successfully commercialized chromium catalysts that are capable of selectively forming 1-hexene. [4,5]n early mechanistic proposal has suggested the involvement of metalacyclic intermediates to explain the observed product selectivity (Scheme 1) and this proposal has been validated via deuterium labeling experiments. [6,7]However, the oxidation state of chromium during the catalytic cycle is still subject to debate.10][11][12][13] None of these studies are however conclusive as no intermediates in the catalytic cycle have been directly detected.Various electron paramagnetic resonance (EPR) spectroscopy experiments have been performed in the presence of ethene in an attempt to observe some of these intermediates. [7,11,12,14]In none of these studies, novel resonances were observed that could be assigned to intermediates in the catalytic cycle.Possibly, these studies were hampered by (Xband) EPR-silence of (some of) the intermediates formed (e. g. dinuclear Cr I or mononuclear Cr II complexes). [15] technique that does allow for the detection of complexes that are unobservable by low-field EPR spectroscopy is X-ray Absorption Spectroscopy (XAS).XAS is a bulk technique that is sensitive to both the oxidation state and the coordination environment of the metal center.[16,17] This technique was applied by Brückner and coworkers in the study of the activation of a {Cr(PNP)} system and the formation of a neutral [(PNP)Cr II (Me) 2 ] complex was suggested.[12] Recently, we have developed a freeze-quench XAS technique which allows for the trapping of reactive complexes.[18][19][20] Freeze-quench XAS is required to study these ethene trimerization systems due to catalyst deactivation occurring on the same timescale as a high-quality EXAFS data acquisition times (several hours).The freeze-quench methodology allows for 'trapping' of reaction intermediates at different points in time, i. e. obtaining time-resolution, while freezing these intermediates allows long measurement times required for high quality data.We have previously applied this technique to study the reaction of a {Cr (SNS)} ethene trimerization system with excess AlMe 3 and demonstrated that the octahedral Cr III precursor is reduced to a Cr II square-planar complex (Scheme 2).[21] This study was performed in the absence of ethene.The presence of ethene could promote further reactivity to the metal center.
The aim of the present study is to gain more insight into the oxidation state and the structure of the active species in the [(R-SN(H)SÀ R)CrCl 3 ] (R = C 10 H 21 ) ethene trimerization system.This was investigated by studying the activation of the Cr III precursor using a variety of activators (e. g.AlMe 3 and MMAO-12) in the absence and presence of a suitable substrate (ethene, other alkenes and alkynes).Spectroscopic techniques that were employed to investigate the oxidation state and structure of the formed intermediates include X-band EPR, Cr K-edge XAS and UV/VIS.The spectroscopic data was interpreted by performing DFTÀ D3 calculations.The obtained results suggest that a dinuclear complex could form under catalytic conditions.Therefore, we performed DFTÀ D3 calculations to compare a mechanism proceeding via mononuclear intermediates to a mechanism proceeding via dinuclear intermediates.

Effect of the Activator on Catalytic Performance
To assess the performance of the catalytic system, we performed catalytic experiments under one bar of ethene pressure in toluene (Scheme 3 and Table 1).The catalyst was an n-decyl substituted [(R-SN(H)SÀ R)CrCl 3 ] (1) complex. [22]As an activator we employed AlMe 3 , AlMe 3 and [Ph 3 C][Al{OC(CF 3 ) 3 } 4 ], and MMAO-12 (from now on denoted as MMAO) as activator. [23,24]MMAO was used as an industrially preferred activator.AlMe 3 , and AlMe 3 and [Ph 3 C][Al{OC(CF 3 ) 3 } 4 ] were used as more well-defined activators. [25]he choice of activator has a large influence on the performance of the catalyst.When 1 is activated with AlMe 3  ), the catalyst does become active at an elevated temperature of 50 °C (entry 3 and 4).When MMAO (400 eq.) is employed as an activator for 1, a significant increase in productivity of the catalyst is observed.The MMAO-activated catalyst is already active at room temperature (entry 5).
Increasing the reaction temperature to 50 °C further increases the activity (entry 6).In addition to formation of 1-hexene, formation of a relatively large quantity of polyethylene (PE) is observed.
In Ti IV ethene trimerization systems it is known that partially alkylated Ti IV species are responsible for polymer formation.In these Ti systems, the amount of PE produced can be reduced by pre-alkylation of the metal center. [26]In this {Cr(SNS)} system, we can significantly decrease the amount of PE (entry 7) formed by first pre-activating 1 with AlMe 3 (20 eq.) and subsequently introducing MMAO (400 eq.) into the reaction mixture.In that case a minimal amount of PE is observed while retaining high activity.In a patent publication it was also demonstrated that activation of a [(R-SN(H)SÀ R)CrCl 3 ] complex with a mixture of AlMe 3 and MAO can lead to reduced PE formation. [5,27]he obtained catalytic results show that the presence of Lewis acidic sites within the reaction mixture is important to generate an active catalyst.When no additional Lewis acidic sites are introduced (entry 1 and 2), no active catalyst is obtained.When Ph 3 C + cations or MMAO is introduced into the reaction mixture (entry 3-7), an active catalyst is obtained.
McGuinness et al. also investigated the effect of Lewis acids on the catalytic performance of the {Cr(SNS)} ethene trimerization system.They also observed a positive effect of Lewis acids on catalytic performance.In addition, they found {Cr II (SNS)} complexes to be active for selective ethene trimerization.Based on these results the authors have suggested a cationic Cr II /Cr IV redox couple to be operative in the {Cr(SNS)} system. [23]

Spectroscopic Investigation of the Activation Process in the Absence of a Substrate
The activation of 1 using AlMe 3 , AlMe 3 and [Ph 3 C][Al{OC(CF 3 ) 3 } 4 ] and MMAO was studied in toluene in the absence of ethene using a variety of spectroscopic techniques (X-band EPR, stopped-flow UV/VIS and Cr K-edge XAS).We had already investigated the activation of 1 with AlMe 3 and AlMe 3 and [Ph 3 C][Al{OC(CF 3 ) 3 } 4 ] in a previous study and observed the formation of a square-planar Cr II complex with a deprotonated amine functionality (Scheme 2). [21]In this study, we have reproduced these experiments.The results are reported in the Supporting Information (Section 2.2 and Section 2.3) and the obtained results are in line with our previous study.
In this study, we have also investigated the activation of 1 with MMAO (400 eq.).The results are reported in detail in the Supporting Information (Section 2.4).Bond distances for solutions of 1 and 1 activated with MMAO in toluene (frozen after 2 minutes) obtained via Cr K-edge EXAFS analysis are compared in Table 2. Similar coordination numbers and bond distances are observed when 1 is activated with either MMAO (Table 2), AlMe 3 , or AlMe 3 and [Ph 3 C][Al{OC(CF 3 ) 3 } 4 ] (Supporting Information Section 2.2 and 2.3).These results indicate that a squareplanar Cr II complex is also formed when 1 is activated with MMAO (Scheme 2).In addition to the formation of a squareplanar Cr II complex, the formation of a bis(η 6 -tolyl)Cr I complex is detected with X-band EPR spectroscopy (Figure S19 and Figure S20a).The concentration of the bis(η 6 -tolyl)Cr I complexes increases with time.After roughly an hour, the concentration is close to 10 % of the total chromium content (Figure S20b).The formation of these bis(η 6 -tolyl)Cr I complexes is a known deactivation pathway in the selective trimerization of ethene. [14,20]sing DFTÀ D3 calculations at the BP86/TZP level of theory we have studied the thermochemistry of the reduction of 1 via reaction with trialkylaluminum compounds (Supporting Information 3).Reduction was assumed to proceed via reaction of the complex with free AlMe 3 contained within MMAO, as AlMe 3 has a higher alkylation aptitude. [28,29]Our results are summarized in Scheme 4 and Table 2. Shown is the thermodynamically most favored structure taking account retention (model A) and deprotonation (model B) of the amine functionality.The calculated CrÀ N distance of model B is in close agreement with the experimentally observed CrÀ N distance, making it likely that the amine is deprotonated.
We also considered the interaction of Lewis acids (AlMe 3 , Ph 3 C + and MMAO) with lone pairs contained on the chloride and amide moiety (Scheme 4, model CÀ G).a] Model Coordination shell Experimental EXAFS data (1) [b] 1 CrÀ N described by Zurek and coworkers. [30]For all considered geometries, interaction of the complex with the Lewis acids was found to be highly exergonic.However, due to the close similarity of the calculated bond distances for the various models, we cannot conclusively assign a single model to our experimental XAS data (vide supra).

Spectroscopic Investigation of the Activation Process in the Presence of a Substrate
With an understanding of the structure of the metal complex after activation, we set out to study the reactivity of the metal complex towards C 2 H 4 using Cr K-edge XAS spectroscopy.Cr Kedge XAS experiments were done via activation of the metal center in the presence of C 2 H 4 (1 bar) and freezing these solutions after a set reaction time.The results are depicted in Figure 1.Initially, activation experiments were performed at room temperature in the presence of C 2 H 4 (1 bar) using AlMe 3 (40 eq.) and [Ph 3 C][Al{OC(CF 3 ) 3 } 4 ] (1.2 eq.), and MMAO (400 eq.) as activator.Under these conditions no changes are observed in the XANES (Figure 1a and c) and EXAFS region (Figure 1b and  d).Therefore, we performed experiments at À 50 °C and 50 °C.A temperature of À 50 °C was chosen to overcome a (potentially) entropically disfavored coordination of ethene to the metal center.A temperature of 50 °C was chosen, as the catalyst is more active at this temperature (Table 1).In most experiments, no differences are observed between the individual spectra.The XANES region for 1 activated with MMAO (400 eq.) at 50 °C does show minor differences.The XANES region matches closely with the XANES region of samples aged for 3 h and 12 h in the absence of ethene (Figure S41).Based on quantitative EPR measurements, relatively large quantities of bis(η 6 -tolyl)Cr I are expected after 3 h (~7 %) and 12 h (~18 %).We therefore attribute these differences in the Cr K-edge XANES region to an increase in concentration of bis(η 6 -tolyl)Cr I at elevated temperatures.
These observations lead to two hypotheses: i) either the first ethene coordination event is endergonic in nature and coordination of ethene cannot be observed in the experiment as performed or ii) a minority species is responsible for catalysis and the majority species is incapable of coordinating ethene.The first hypothesis is in line with a recent kinetic study of a {Cr (PN)} ethene tetramerization system, where it is demonstrated that the first and/or the second ethene coordination event is reversible. [31]The second hypothesis has been proposed by Bercaw and coworkers in a {Cr(PNP)} ethene trimerization system. [32]o investigate the first hypothesis, we resorted to different substrates.As discussed earlier, catalytic tests hint at a cationic mechanism being operative in this ethene trimerization system due to the observed positive effect of Lewis acids.Three free coordination sites are required on the metal center to successfully complete the catalytic cycle for a tridentate ligand (Scheme 1).We therefore envisioned an initiation pathway, Scheme 4. Calculated structures that were used to compare to the experimental bond distances.Gibbs free energies (ΔG) and electronic energies (ΔE) are reported in kcal mol À 1 for the high-spin states of the depicted complexes.Frequency calculations were performed at 298.15 KDFT-D3 calculations were performed at the BP86/TZP level of theory in the gas phase.Full calculations are found in the Supporting Information (Section 3).Selected structural parameters are.
where the halide is abstracted from the metal center and a cationic, electron-poor metal center is generated (Scheme 5).To this cationic metal center, ethene can subsequently coordinate.Generation of a cationic metal center in solvents with low dielectric constants (e. g. toluene) is expected to be disfavored. [33]To stabilize such a cationic complex, we turned our attention to more electron-rich substrates.
Initial tests were done with 1-octene and 2,3-dimethyl-2butene. Upon addition of these olefins to a MMAO-activated solution (400 eq.) of 1, no changes in the UV/VIS spectra were observed after 2 h (Figure S42).In a previous study, Bercaw and coworkers have demonstrated that ethene trimerization systems can also be employed as alkyne cyclotrimerization catalysts. [7]Upon activation of 1 with MMAO (400 eq.) or with AlMe 3 (40 eq.) and introducing 3-hexyne (60 eq.), an immediate color change from green to orange is observed.Heating the solution to 70 °C and letting the mixture react for 1 h allows for the formation of hexaethylbenzene, as is observed by 1 H NMR spectroscopy (Figure S40).The proposed mechanism of chromium-catalyzed cyclotrimerization (Scheme 6) shows similarities to the trimerization of ethene and alkynes might therefore be used as a model for the reactivity of ethene. [34,35]For our spectroscopic investigation, we employed 1-phenyl-2-trimethylsilylacetylene as a substrate.This substrate was employed as its steric bulk might hamper coordination of a second alkyne to the metal center, allowing us to study the first coordination event in detail.
Upon reaction of 1 with MMAO (400 eq.) and addition of 1phenyl-2-trimethylsilylacetylene (~100 eq.), the color of the reaction mixture gradually changes from green to purple over the course of an hour (Figure S42).Interestingly, upon reaction of 1 with AlMe 3 (40 eq.) and addition of 1-phenyl-2-trimethylsilylacetylene (~100 eq.), no color change is observed.X-band EPR was applied to study the oxidation state of the metal center upon introduction of the alkyne.The experiment was performed by mixing 1 with MMAO (400 eq.) in toluene.Five minutes after activation, 1-phenyl-2-trimethylsilylacetylene (100 eq.) was added.After 2 h, an aliquot of the solution was taken and measured at 20 K.The resulting EPR spectrum is shown in Figure 2. The spectrum is composed of overlapping resonances, which can be ascribed to resonances from a Cr I and a Cr III complex (vide infra).The EPR spectrum was simulated using two components (S = 1/2 and S = 3/2).For spin counting, double integration was performed on the simulated spectra of the two components. [36][39] Here, g eff represents the effective g-value, S represents the spin and c represents the concentration of the complex.For the Cr I complex normalization was achieved by dividing the double integral by 1.985 (g eff ) and 0.75 (S = 1/2).For the Cr III complex normalization was achieved by dividing the double integral by 3.317 (g eff ) and 3.75 (S = 3/2).The corresponding normalized double integrals were compared to the double integral of a TEMPO solution with a known concentration.The double integral determined for TEMPO was also normalized by dividing by 2.008 (g eff ) and 0.75 (S = 1/2).The obtained results are discussed below.
Firstly, a resonance with axial symmetry is observed (g x,y = 1.977, g z = 2.002) and was quantified to consist of ~19 % of the total chromium content.The symmetry and g-values allow us to assign these resonances to the formation of bis(η 6 -tolyl)Cr I complexes.An additional resonance with rhombic symmetry is observed (g x = 4.043, g y = 3.914, g z = 1.995).These resonances are expected for Cr III complexes with a large zero-field splitting. [40]The concentration of this Cr III complex was found to consist of ~68 % of the total chromium content.A relative accuracy of � 30 % is expected for quantitative EPR measurements. [11,38]Within experimental error of the measurement, it cannot be concluded whether solely these two complexes are present in solution or whether other EPR-silent complexes are also present.
Cr K-edge XAS experiments were performed to gain further insight into the structure of the formed complexes.The Cr Kedge XANES data is reported in the Supporting Information (Figure S45) and the Cr K-edge EXAFS analysis is discussed Scheme 6. Mechanism for the cyclotrimerization of alkynes.The aromatic compound is formed either via a bicyclic or a metallocycloheptatriene intermediate.
Figure 2. Experimental X-band EPR spectrum for the activation of 1 with MMAO (400 eq.), followed by the addition of 1-phenyl-2-trimethylsilylacetylene (100 eq.).The reaction was performed in toluene and was frozen after 2 hours.Also shown is the simulated spectrum and the two components used for the simulated spectrum.For the Cr I complex (component 1), a fit was obtained using g x,y = 1.977 and g z = 2.001 and by applying Gaussian broadening (50 MHz).For the Cr III complex (component 2), a fit was obtained using g x = 4.043, g y = 3.914 and g z = 1.995 and by applying Gaussian broadening (221, 234 and 469 MHz respectively).
below.The obtained parameters are presented in Table 3 and an example of the data quality is presented in Figure 3.
Care had to be taken in the EXAFS analysis, as X-band EPR spectroscopy shows the presence of (at least) a Cr I complex and a Cr III within the solution.We were capable of successfully fitting data by assuming that the EXAFS spectrum was composed of contributions from a five-coordinate Cr III alkyne complex and a bis(η 6 -tolyl)Cr I complex.For the Cr III alkyne complex, the best fitting results were obtained when we assumed the complex to contain 2 atoms in a CrÀ C/CrÀ N shell close to the metal center (~2.10 Å) and 3 atoms in a CrÀ Cl/ClÀ S shell (~2.40 Å).A third CrÀ C shell (~3.10 Å) containing four atoms was required to fit the data.Carbon atom contained within the ligand backbone are expected to scatter at this distance from the metal center.The bis(η 6 -tolyl)Cr I complex is expected to contribute a CrÀ C shell containing 12 atoms at a distance of 2.13 Å from the metal center. [41]This shell overlaps with the CrÀ C/CrÀ N shell from the Cr III alkyne complex.In our fitting model we introduced a parameter (f 1 ) which describes the contribution (0-100 %) of the bis(η 6 -tolyl)Cr I complex to the Cr K-edge EXAFS spectrum.Via equations given in Table 3, the coordination number of the coordination shells was related to the contribution and this parameter was optimized.The amount of bis(η 6 -tolyl)Cr I complex was estimated to be 8 % � 11 %.Within experimental error this value agrees with the value determined by EPR spectroscopy.We considered three possibilities for the structure of the reaction product.Alkynes can act as a neutral donor or can give rise to a two-electron oxidation of the metal center. [35]Either i) the alkyne directly reacts with the metal center to form a Cr II / Cr IV alkyne complex or ii) upon coordination of the alkyne a disproportionation reaction occurs and a Cr I and a Cr III complex are formed or iii) upon coordination of the alkyne a dinuclear Cr III complex is formed.
The first hypothesis can be excluded due to the observation of a large quantity of Cr III in the X-band EPR spectrum.The second and third hypothesis are harder to distinguish.However, if a disproportionation reaction were to occur, the amount of Cr III is expected not to exceed 50 %.In the performed spin counting experiments, 68 % of a Cr III complex is observed, deeming disproportionation unlikely.We therefore propose that the observed changes in the Cr K-edge EXAFS and X-band EPR data are due to the formation of a dinuclear Cr III alkyne complex.Unfortunately, the CrÀ Cr contribution of the dimer cannot be observed directly in EXAFS due to the limited EXAFS data quality and expected long CrÀ Cr distance (> 3.5 Å, vide infra).
To obtain further insights into the structure of the formed Cr III alkyne complex we performed DFTÀ D3 calculations at the BP86/TZP level of theory for various plausible geometries (Scheme 7 and Table 4).Optimization of a neutral dinuclear Cr III alkyne complex was unsuccessful.For this geometry, dissociation of one of the sulfide donors was observed.The monocationic (model I) and dicationic (model J) show very similar bond distances.They however differ in the coordination number of the CrÀ Cl/CrÀ S shell.Comparing the expected coordination numbers for the CrÀ Cl/CrÀ S shell of the monocationic (3) and dicationic (2) complex, formation of a monocationic is deemed likely based on agreement with the Cr K-edge EXAFS results.
[b] In these equations, f 1 is a parameter that describes the amount of bis(η 6 -tolyl) Cr I complex relative to a Cr III alkyne complex.
[d] The coordination number of the CrÀ Cl/CrÀ S shell is calculated via the equation (3-3 f 1 ).
[e] The coordination number of the CrÀ C shell is calculated via the equation (4-4•f 1 ).
Figure 3. Cr K-edge experiments for the activation of 1 with MMAO (400 eq.), followed by the addition of 1-phenyl-2-trimethylsilylacetylene (50 eq.).The reaction was performed in toluene and was frozen after 2 hours.The reaction was performed in toluene.Depicted is a) the EXAFS data and b) the corresponding Fourier transform of k 2 -weighted Cr K-edge EXAFS data.The final concentration in Cr was 3.61 mM.
1.2 kcal mol À 1 ) is almost neutral in energy.Close agreement between calculated and experimental Cr-C/Cr-N (experimental: 2.09(2) Å, calculated: 2.06(2) Å) and CrÀ Cl/CrÀ S (experimental: 2.42(1) Å, calculated: 2.457 Å) bond distance is found for model L, suggesting that this structure most closely resembles the structure of the formed Cr III alkyne complex.It should be stated that the presence of bis(η 6 -tolyl)Cr I will give rise to a slight overestimation of the CrÀ C/CrÀ N bond distance.

DFT Calculations on the Mechanism of Ethene Trimerization
EPR spectroscopy and Cr K-edge XAS experiments show the oxidation of a square-planar Cr II complex towards a dinuclear cationic Cr III alkyne complex in the presence of 1-phenyl-2trimethylsilylacetylene.The formation of this complex can be interpreted as being the result of dinuclear oxidative addition of the alkyne to a cationic Cr II complex and a neutral Cr II complex (Scheme 8).Based on these results, we were interested in whether a similar oxidation of Cr II to Cr III might occur in the presence of ethene.To address this question, we compared the thermochemistry of dimerization of a neutral Cr II complex and a cationic Cr II complex to form a dinuclear Cr III complex.These DFTÀ D3 calculations were performed at the BP86/TZP level of theory with implicit solvent corrections (COSMO) for toluene. [42]s a bridging moiety, ethene and 1-phenyl-2-trimethylsilylacetylene were employed (Scheme 8).For 1-phenyl-2trimethylsilylacetylene it is indeed predicted that upon coordination of 1-phenyl-2-trimethylsilylacetylene oxidation from Cr II to Cr III is feasible (ΔG = À 21.8 kcal mol À 1 ), in line with experimental observation.For ethene, an oxidation of Cr II to Cr III to form a dinuclear cationic Cr III complex is also expected to be favored (ΔG = À 10.8 kcal mol À 1 ), suggesting similar reactivity could take place with ethene.
Scheme 7. Plausible reaction pathway via which the activated square-planar Cr II complex can react to form a dinuclear Cr III alkyne complex.
a] Model Coordination shell Experimental data (1 + MMAO-12 (400 eq.) + 1-phenyl-2-trimethylsilylacetylene (50 eq.) (2 h)) [b] 2 CrÀ C/CrÀ N [a] Geometry optimizations were performed in the absence of the anion.We considered the high-spin state in an antiferromagnetic (singlet) and ferromagnetic (heptet) configuration.The singlet and heptet spin state were found to be in close proximity to one another (~3 kcal mol À 1 ).The singlet spin state was found to be slightly more favorable and is reported in this table .[b] Given is the experimental (CrÀ X) (Å) distance and the expected coordination numbers for the Cr III alkyne complex, as is obtained from Cr K-edge EXAFS measurements.Samples were measured as frozen toluene solutions (100 K).Full EXAFS data analysis results are provided in the ESI.
spin state.Overall, the reaction is predicted to proceed with an activation energy of ΔG � = 30.7 kcal mol À 1 , corresponding to the energy difference between the lowest lying intermediate A1 and the rate-determining transition state TSA2.This model is capable of correctly predicting the selectivity for 1-hexene.The formation of 1-butene from chromacyclopentane A4 has a significantly increased barrier (TSA4, ΔΔG � = 39.9 kcal mol À 1 ) compared to the migratory insertion of ethene to form the chromacycloheptane A6 (ΔΔG � = 22.1 kcal mol À 1 ).In addition, attempts to optimize a structure with a fourth ethene molecule coordinated to chromacycloheptane A6 failed.
During geometry optimization the ethene molecule liberates from the metal center.This was also observed Yang et al.They ascribe this observation to the steric repulsion caused by the ligand and by the chromacyclopheptane intermediate, hampering the formation of 1-octene. [8]n interesting finding of these calculations is that coordination of the first ethene molecule is predicted to be only mildly endergonic (ΔG = 2.8 kcal mol À 1 ) at 298 K. Performing spectroscopic experiments at elevated ethene pressures and/or at lower temperatures could thus provide opportunities to stabilize these ethene-coordinated intermediates.Model employed for DFTÀ D3 calculations of the mechanism for ethene trimerization via mononuclear cationic Cr II /Cr IV intermediates.Geometry optimization were performed at the BP86/TZP level of theory in the gas phase and solvent corrections were applied by performing single-point calculations using an implicit solvent model for toluene (COSMO).The singlet (not depicted), triplet (red) and quintet (blue) spin state were taken into account for chromium.For the dimer, the quartet (olive) spin state is depicted.Gibbs free energies and electronic energies (italics) are reported for the various intermediates of the catalytic cycle.Frequency calculations were performed at 298.15 K. Methyl substituents on the ligand were used to reduce computational cost.The model AlOMe) 10 .AlMe 3 was used for MMAO, as described in reference 30.
Full Papers 890 Some differences are observed between the calculations performed in the present study and the study performed by Yang et al.Firstly, the activation energy in the present study (ΔG � = 30.7 kcal mol À 1 ) is much lower compared to the activation energy (ΔG � = 55.9 kcal mol À 1 ) found by Yang et al. for a monocationic Cr II /Cr IV redox couple. [45]In addition, the activation energy found in this study also is lower compared to the neutral Cr I /Cr III model where the amine functionality was retained (ΔG � = 40.1 kcal mol À 1 ). [45]The DFT calculations performed in the present study thus provide a model in agreement with experimental findings (inclusion of activator and deprotonation of the amine functionality) and provides a physically more realistic activation energy for a reaction already proceeding at room temperature.It should however be stated that the calculated activation energy in the present study is still high for a reaction already proceeding at room temperature.
We also performed calculations for a mechanism proceeding via dinuclear cationic Cr II /Cr III intermediates.Our obtained results are described in detail in the Supporting Information (Section 5.5).The calculated mechanism is incapable of explaining the selectivity for 1-hexene.Very similar barriers are observed for the transition state responsible for metallacycle growth and the transition state responsible for product formation.Likely, such a dinuclear intermediate will give rise to a non-selective ethene oligo-or polymerization catalyst.
Spectroscopic studies performed in the presence of ethene have allowed for the detection of a neutral square-planar Cr II complex.No ethene is coordinated to this complex.This complex likely serves as the resting state during catalysis (Figure 4).DFT calculations indicate that coordination of the first ethene molecule to this square-planar complex is endergonic.Abstraction of the bound chloride and formation of a cationic Cr II complex is required to proceed through the catalytic cycle.DFTÀ D3 calculations show that a dinuclear ethene Cr III complex could subsequently be formed under catalytic conditions (Scheme 8), which is similar in structure to the spectroscopically observed Cr III alkyne complex.Calculations of a dinuclear Cr II /Cr III calculations however show that such a Cr III ethene complex is not selective for the formation of 1-hexene and likely yields a distribution of LAOs.This dinuclear intermediate thus likely acts as an off-cycle intermediate.
Only the mechanism proceeding via mononuclear intermediates is capable of correctly predicting the observed product selectivity.Hence, this dinuclear complex most likely has to dissociate in order to generate the catalytically active mononuclear cationic Cr II complex.This mononuclear Cr II complex subsequently can enter the catalytic cycle (Figure 4), to produce 1-hexene in the experimentally observed selectivity.

Conclusions
In this study, the activation of [(R-SN(H)SÀ R)CrCl 3 ] with MMAO was investigated in the presence and absence of substrates (alkenes and alkynes) via the use of spectroscopic techniques (XAS, UV/VIS and EPR) and DFT calculations.In the absence of ethene, the octahedral Cr III precursor is reduced to a square-planar Cr II complex and the amine functionality is deprotonated.Upon introduction of ethene into the reaction mixture, no coordination of ethene is observed under the employed experimental conditions (1 bar C 2 H 4 ).Likely this neutral Cr II complex serves as the resting state during catalysis.Using an alkyne as a model for ethene coordination, oxidation from Cr II to Cr III is observed.This oxidation is attributed to the formation of a dinuclear cationic Cr III alkyne complex.DFT calculations suggest a/similar dinuclear Cr III ethene complex could form under catalytic conditions.Via DFT calculations we have compared a mechanism proceeding via cationic mononuclear Cr II /Cr IV intermediates to a mechanism proceeding via cationic dinuclear Cr II /Cr III intermediates.Only the mononuclear Cr II /Cr IV mechanism is capable of correctly predicting the observed product selectivity.Future studies will focus on stabilizing the proposed cationic Cr ethene complexes.We aim to achieve this by performing the activation in the presence of ethene at elevated ethene pressures and/or lower temperatures.

Experimental Section
The experimental details are reported in the Supporting Information.

Figure 1 .Scheme 5 .
Figure 1.Cr K-edge XAS experiments performed in the presence of ethene (1 bar) with various activators.a) XANES and b) EXAFS region of activation experiments with MMAO (400 eq.) in the presence and absence of ethene at various temperatures.c) XANES and b) EXAFS region of activation experiments with AlMe 3 (40 eq.) and [Ph 3 C][Al{OC(CF 3 ) 3 } 4 ] (1.2 eq.) in the presence and absence of ethene at various temperatures.The samples were measured as a frozen solution in toluene in fluorescence mode.The legend is only shown in Figure 1a and 1c.A similar coloring scheme is used in Figure 1b and d.

Figure 4 .
Figure 4. Model employed for DFTÀ D3 calculations of the mechanism for ethene trimerization via mononuclear cationic Cr II /Cr IV intermediates.Geometry optimization were performed at the BP86/TZP level of theory in the gas phase and solvent corrections were applied by performing single-point calculations using an implicit solvent model for toluene (COSMO).The singlet (not depicted), triplet (red) and quintet (blue) spin state were taken into account for chromium.For the dimer, the quartet (olive) spin state is depicted.Gibbs free energies and electronic energies (italics) are reported for the various intermediates of the catalytic cycle.Frequency calculations were performed at 298.15 K. Methyl substituents on the ligand were used to reduce computational cost.The model AlOMe) 10 .AlMe 3 was used for MMAO, as described in reference 30.

Table 1 .
Overview of the performance of 1 in the presence of various activators under one atmosphere of ethene in toluene.[a]