Details of the Mechanism of the Asymmetric Transfer Hydrogenation of

Dec 1, 2015 - The excellent ketone asymmetric transfer hydrogenation (ATH) systems using the precatalysts (S,S)-trans-[FeCl(CO)(PPh2CH2CH2NHCHPhCHPhNC...
0 downloads 7 Views 3MB Size
Research Article pubs.acs.org/acscatalysis

Details of the Mechanism of the Asymmetric Transfer Hydrogenation of Acetophenone Using the Amine(imine)diphosphine Iron Precatalyst: The Base Effect and The Enantiodetermining Step Weiwei Zuo,†,‡ Demyan E. Prokopchuk,† Alan J. Lough,† and Robert H. Morris*,† †

Department of Chemistry, University of Toronto, 80 Saint George Street, Toronto, Ontario M5S 3H6, Canada State Key Laboratory for Modification of Chemical Fibers and Polymer Materials, College of Materials Science and Engineering, Donghua University, Shanghai 201620, People’s Republic of China



S Supporting Information *

ABSTRACT: The excellent ketone asymmetric transfer hydrogenation (ATH) systems using the precatalysts (S,S)-trans-[FeCl(CO)(PPh2CH2CH2NHCHPhCHPhNCHCH2PAr2)]BPh4 (Ar = Ph (1), ptolyl (2)) have a fascinating dependence of activity on the base concentration, which is investigated here. The reaction of complex 1 or 2 with 1 equiv of the strong base potassium tert-butoxide in THF for 2−7 days produces the neutral amine(ene-amido) complexes [FeCl(CO)(PPh2CH2CH2NHCHPhCHPhNCHCHPAr2)] (8 and 9). These monodeprotonated complexes have been completely characterized by NMR, EA, and FT-IR spectroscopy and mass spectrometry, and the structure of 9 has been further confirmed by single-crystal X-ray diffraction to reveal a structure with the NH and FeCl bonds parallel and the proton and chloride ligands next to each other. The structures of 8 and 9 and their 1H NMR patterns are similar to those of the active catalyst for the ATH of ketones that is postulated to have NH and FeH bonds parallel with the protonic and hydridic hydrogens adjacent. Identical key nuclear Overhauser effect (NOE) correlations in both 8 and the hydrido complex provide further evidence for the postulated structure of the amine iron hydride intermediate. The catalyst system is not active for the transfer hydrogenation of acetophenone, unless greater than 2 equiv of base is added to activate the precatalyst. The addition of base (up to 8 equiv per iron) increases the reaction rate, while a further increase of the base concentration shows a reduction of activity. The loss of activity with less than 2 equiv of base results from a side reaction of the active amido(ene-amido) complexes with 2-propanol to form an inactive neutral bis(amido) iron complex, which was characterized by NMR spectroscopy. Structural evidence for this was provided by the X-ray crystal structure determination of an analogous bis(amino) iron(II) complex, generated from the reaction of the amine(ene-amido) iron complex with methanol in C6D6 in the presence of BF4−. The presence of excess base prevents this side reaction, thereby favoring the reaction which forms the active amine iron hydride species that is in the catalytic cycle. The structure of the transition state for the reaction of the amine hydrido iron catalyst with acetophenone has been successfully modeled using density functional theory (DFT) calculations. The (R) configuration of the product 1-phenylethanol is induced by the position of the phenyl groups on the catalyst and a π−π stabilizing interaction between the aryl ring on the ketone and the ene-amido moiety on the ligand. KEYWORDS: asymmetric transfer hydrogenation, iron catalyst, ketone reduction, mechanism, amido ligand



INTRODUCTION Developing and understanding the mechanism of homogeneous hydrogenation catalysts based on iron is an area of much interest as part of the effort to find green, sustainable chemical processes.1−35 We are developing iron-based catalysts for the asymmetric hydrogenation (AH)4,36−38 and asymmetric transfer hydrogenation (ATH)4,12,38−44 of prochiral ketones and imines to valuable enantioenriched alcohols and amines, respectively. The iron complex (S,S)-trans-[FeCl(CO)(PPh2CH2CH2NHCHPhCHPhNCHCH2PAr2)]BPh4 (1, Ar = Ph; Scheme 1) is a particularly effective precatalyst for the ATH of aryl ketones in isopropanol.38,42,43 It is activated by the reaction of an excess of the strong base, usually potassium tertbutoxide, to give the amido(ene-amido) complexes 3 and 4 and © XXXX American Chemical Society

amine hydride 5, which are proposed to be in the catalytic cycle. The activation involves the removal of a proton on the CH2 group α to the imine carbon by 2 equiv of base. We observed during the investigation of this catalytic mechanism that the reaction with 8 equiv of base per complex 1 in 2propanol gave superior rates rather than the 2 equiv expected on the basis of the proposed activation step. This is in contrast to the mechanism of the AH of ketones catalyzed by 1 and 2 equiv of base in THF;38 the addition of more base did not increase the reaction rate. The current study uncovers the reasons for these observations and provides further DFT Received: September 6, 2015 Revised: November 27, 2015

301

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis

Scheme 1. Proposed Mechanism for the Asymmetric Transfer Hydrogenation of Acetophenone Catalyzed by the Amine(imine)diphosphine Iron Carbonyl/Base System

of base in a very active ruthenium ketone TH catalyst system was to keep the coordinated alkoxide (C), regarded as essential to generate the hydrido complex E in the catalytic cycle, from being protonated to give the alcohol adduct D. In this report, we identify complexes obtained by the reaction of 1 with base and study the base dependence of the catalytic ATH reaction in more detail in order to understand the mechanism of action of the system using precatalyst 1. We also report the discovery of full-structure transition states for the ketone reduction step using DFT methods. Discovering the source of enantiodiscrimination in such bifunctional hydrogenations by DFT methods is of great interest in order to better design asymmetric hydrogenation catalysts.70−77 These results provide a clearer mechanistic picture of the individual steps of the catalytic cycle of this important ATH system.

evidence that the hydride 5 is responsible for the enantiodetermining hydride transfer step. Base has been found to be an important component in the asymmetric pressure dihydrogen45−55 and transfer hydrogenation29,56−65 reactions catalyzed by metal complexes. There have been reports of the acceleration of the rate of transfer hydrogenation of ketones when excess base is added to ruthenium catalyst systems. Various explanations have been offered for this effect. Aranyos et al.66 showed that the activating role of base in ruthenium triphenylphosphine chloro systems for the TH of ketones in 2-propanol was to create the ruthenium dihydride [RuH2(PPh3)3] responsible for ketone reduction, in other words the activation of the catalyst. Similarly, Noyori’s ATH catalysts [Ru(arene)Cl(TsNCHPhCHPhNH2)] only require base for the formation of the active amido complex.67 Our group elucidated the route from A via B to C or D (Chart 1) for the AH and ATH catalyst [RuH2(pyCMe2NH2)(binap)] and proposed that excess alkoxide might serve to labilize these equilibria68 on the basis of the suggestion by Hamilton and Bergens for an AH system45,46 and on classic base-catalyzed substitution at cobalt(III).69 Baratta and co-workers60 proposed that the role



RESULTS AND DISCUSSION Reactions of the Precatalysts 1 and 2 with Strong Base. Reactions of the precatalyst 1 in THF with base were investigated first to simplify the system from the complicating effect of the protic solvent 2-propanol that is used in ATH. Note that the NMR reactions involve complex and base concentrations of about 10−2 M, while ATH catalysis is conducted at catalyst concentrations of 10−5 M and base at 10−4 M. These catalyst solutions are very oxygen sensitive once they are activated; therefore, the sample preparation was done in a glovebox with very low oxygen levels. Reaction of 1 or 2 with 1 Equiv of Strong Base. The reaction of the cationic complex 1 or 2 with 1 equiv of KOtBu in THF at 25 °C for 5 min gives a mixture of isomers. Complex 1 produces isomers 6 and 7 with the probable structures shown in Scheme 2. Over a period of 1 week the isomers convert to the amine(ene-amido) iron(II) complex 8. The reaction of ptolyl derivative 2 with KOtBu produces similar isomers that eventually convert to the amine(ene-amido) iron(II) complex 9. On the basis of NMR data, DFT calculations, and an X-ray diffraction structure for 9, it has been established that complexes 8 and 9 have the N−H bond oriented parallel to the Fe−Cl bond, as shown in Scheme 2. This is in contrast with the starting complexes 1 and 2, which have the N−H group next to the Fe−CO group. There is not enough NMR information to make definitive structural assignments for the intermediate isomers 6 and 7, but Scheme 2 shows structures

Chart 1. Ruthenium Complexes That Are Activated by Base and Transfer Hydride (and Proton) Equivalents to Ketonesa

L = phosphine, amine, hydride, aryl or pyridyl donor; R′O− = 2-PrO− or other alkoxide; R = various aryl or alkyl substituents. a

302

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis

Scheme 2. Reaction of the Amine(imine)diphosphine iron(II) Carbonyl Complexes with 1 Equiv of Potassium tert-Butoxide in THFa

a

For the reaction of 1, after 5 min at 25 °C, the 6:7:8 ratio is 1:1.7:2.5, and for the reaction of 2 the 6:7 ratio is 1:2.

Figure 1. 1H NMR chemical shifts of the amine(ene-amido) carbonyl iron hydrido and chloro compounds (5 and 8) as well as the two representative NOE interactions in these two analogues.

CHN distance contracts from 1.467(7) to 1.351(8) Å when x changes from 2 to 1, with the CH−N bond lengthening from 1.256(7) to 1.329(7) Å. The C1−N1 length of 1.467(6) Å and the C6−P2 distance of 1.758(6) Å are also clearly shorter than those in 2 (1.486(7) and 1.848(6) Å, respectively). All of these changes suggest that there is extensive delocalization of bonding in the ene-amido part of 9. The deprotonation also results in bond length changes in the amine side. For example, the Fe1−N2 and N2−C3 bond lengths of 2.035(5) and 1.458(7) Å have been shortened by 0.027 and 0.028 Å in comparison to those found in the mother complex, respectively. The metrical parameters of 8 from DFT calculations (see the Supporting Information) match with the corresponding parameters in 9. As proposed for 8, the N−H bond points in the same direction as the Fe−Cl bond in the structure of 9. The NOESY spectrum of complex 9 exhibits NOE contacts very similar to those shown for 8 in Figure 1. As can be seen in Figure 2, these contacts correspond to the short distances between H1a in the C(Ph)−H bond and H2n in the NH group and the short distance between the axial hydrogen of the CH2 group (H3a)

consistent with the available data, as discussed further in the Experimental Section. The carbonyl vibration in the IR spectrum of the neutral complex 8 is shifted to 1933 cm−1, in comparison to 1976 cm−1 for the cationic precatalyst complex 1. The 1H NMR pattern of 8 is quite similar to that of the amine(ene-amido) iron hydrido complex 5, which was identified previously and found to be in the catalytic cycle for ketone hydrogenation (Figure 1). Two representative NOE interactions observed in 8 are identical with those observed in the amine iron hydrido complex 5 (Figure 1). Thus, the two compounds are structurally closely related.42 An X-ray diffraction structure determination of a single crystal of 9 shows that the iron is coordinated in a distortedoctahedral fashion, with the nitrogen and phosphorus atoms of the P−N−NH−P ligand forming the equatorial plane with the chloride and carbonyl ligands located in axial positions (Figure 2). The ene-amido nitrogen−iron distance is 1.970(4) Å (N1− Fe1), shorter than the amine nitrogen−iron bond of 2.035(5) Å (N2−Fe1) but comparable to those found in our previous bis(ene-amido) complexes.39 On going from 2 to 9, the PCHx− 303

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis

reaction. The IR spectrum of 11 has a νCO absorption at 1869 cm−1. Products of the same composition can also be produced by directly reacting the precatalyst with base in 2-propanol, conditions closer to those used for transfer hydrogenation catalysis, where the amido(ene-amido) complex was thought to be generated in situ but not isolated. The reaction of 1 with exactly 2 equiv of base in 2-propanol at 25 °C for 3 min followed by the immediate removal of 2-propanol affords a 1:1 mixture of 10 and the hydride 5, on the basis of both the 1H and 31P{1H} NMR spectra. Reaction of 1 with More than 2 Equiv of Base. Slightly increasing the amount of base above 2 equiv with respect to 1 or 2 causes a significant increase in the ratio of the hydrido complex to the saturated bis(amido) complex. For comparison, the use of 2.1 equiv of base in the reaction of 2 results in the formation of mixture of the reduced species 11 and the hydride in a ratio of around 1:1, while the use of 2.5 equiv eliminates the formation of the reduced species. As we have reported previously,42 at base amounts of 2.5 equiv and above the product is exclusively the iron hydride 5 complex along with the amido(ene-amido) complex 3 (Scheme 4). Related Bis(amine) Iron Complex. The addition of excess methanol to a C6D6 solution of complex 8 in the presence of KBF4 leads to the formation of the new bis(amine) iron(II) complex 13 (Scheme 5), which crystallized from the reaction solution upon addition of CH2Cl2 (Figure 3). Presumably methanol is the reductant of the imine group. The spectroscopic properties of the BPh4 salt have already been reported.12 The solid-state structural features of 13 (Figure 3) are closely related to those described for its BPh4 analogue.12 In crystalline form, 13 is in a distorted-octahedral geometry with a P−NH− NH−P tetradentate ligand and trans-situated carbonyl and chloride ligands. The C5−C6 bond length of 1.508(5) Å is similar to that of C1−C2 (1.516(4) Å) and much longer than the corresponding values measured in 9, indicating a C−C single bond instead of double bond in 13. The distances for C5−N2 (1.469(4) Å) and N1−C2 (1.484(4) Å) are also indicative of C−N single bonds. Complex 13 has the two NH groups on the opposite sides of the plane. The crystal structure of 13 provides evidence for the structural assignments of 10 and 11. The metrical parameters of cation 13 have been confirmed by DFT calculations (see the Supporting Information). Hydrogenation Activity of the Amido(ene-amido) and Bis(amido) Complexes. The bis(amido) complexes 10 and 11 were tested and found to be inactive for both pressure H2 hydrogenation (up to 20 atm) and transfer hydrogenation catalysis, with temperatures in the range of 25−50 °C. There is

Figure 2. ORTEP plot of 9. Thermal ellipsoids are drawn at the 30% probability level. Some hydrogen atoms are omitted for clarity. Selected interatomic distances (Å) and angles (deg): Fe1−N1 1.970(4), Fe1−N2 2.035(5), Fe1−P1 2.260(2), Fe1−P2 2.266(2), N1−C5 1.329(7), C5−C6 1.351(8), N2−C3 1.458(7), C3−C4 1.540(7), N1−C1 1.467(6), N2−C2 1.491(7); P1−Fe1−P2 110.20(6).

and H2a of the other C(Ph)−H bond. The similarities among the NMR spectra of 9, 8, and 5, including the two identical key NOE correlations, provide further evidence for the postulated structure of the amine iron hydride catalyst 5 (Scheme 1). Reaction of 1 or 2 with 2 Equiv of Strong Base and 2Propanol. Complex 1 has been reported to react with 2 equiv of base in THF to produce the amido(ene-amido) complexes 3 and 4 shown in Scheme 1.42 These complexes react with 2propanol to afford red solutions, in either the presence or absence of excess base. We find now that, when there is no excess base present, an equimolar mixture of a new compound (10) and the known amine iron hydride complex 5 are produced (Scheme 3). Complex 10 was found by NMR experiments (see the Experimental Section) to be the neutral bis(amido) carbonyl iron(II) complex resulting from complete reduction of the backbone of the P−N−N−P ligand. The complex has a lowwavenumber carbonyl absorption at 1839 cm−1, 62 cm−1 lower than that of the amido(ene-amido) complex 3 at 1901 cm−1. Saturating the backbone of the ligand has a significant electronic effect, causing greater back-donation of electrons from Fe into the CO ligand π* orbitals. For the p-tolyl analogue the ratio between the reduced complex 11 and the corresponding hydride 12 is around 9:1, indicating that the substitution of the phenyl groups on the phosphine atom with the p-tolyl analogue favors the reduction

Scheme 3. Generation of the Intermediate Amido(ene-amido) Carbonyl Iron Complexes by Reaction with 2 Equiv of Base (Step 1) and Their Reaction with 2-Propanol (Step 2)

304

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis

Scheme 4. Generation of the Intermediate Amido(ene-amido) Carbonyl Iron Complexes by Reaction with Excess Base (Step 1) and Their Reaction with 2-Propanol Producing 5 and 3 in a 4:3 Ratio (Step 2)

Scheme 5. Formation of 13 from an Attempt To Crystallize 8 in a Mixed Solvent of C6D6, MeOH, and CH2Cl2 without First Removing the KBF4

proposed to generate the amido(ene-amido) complex 3, which, when there is a low concentration of base, is protonated by 2propanol at the carbon α to phosphorus to give the cationic intermediate 14. A slow hydride transfer from the alkoxide to the imine, also supported by DFT calculations below, produces the saturated complex 10, which is known to be catalytically inactive. The presence of excess base will suppress the formation of 14 and thus allow 3 to undergo rearrangement, which occurs on the order of seconds to minutes to give catalytically active 4.42 An alternative route to 4 (as in Scheme 2) is shown with dashed lines; however, the rearrangement to 8 takes hours in THF. DFT calculations as discussed below indicate that complex 10 and its products upon reaction with alcohol are more stable and represent a deactivation pathway for the catalyst. Reaction Profiles as a Function of Base and Ketone Concentrations. A systematic variation of base concentration was conducted in an attempt to investigate the role of base in catalysis. Plots of 1-phenylethanol concentration versus time at various amounts of base with other conditions being kept the same are shown in Figure 4. There is a dramatic dependence of the reaction rate on the base concentrations. No acetophenone conversion was observed when less than 2 equiv of base was added. When the base:catalyst ratio is slightly greater than 2, for example 2.2−2.5, there is some activity, but the activity is relatively low and the maximum conversion of acetophenone was not achieved in 10 min. This is in contrast to our standard reaction profile, where 80% conversion of acetophenone is easily achieved within 2 min. Further increasing the concentration of base with a base:catalyst ratio in the range 2:1−8:1 causes an obvious increase in the reaction rates. At higher concentrations of base, for example 12 equiv, the reaction shows a decrease in activity. It should be mentioned that, in the catalytic reactions using 2−3 equiv of base, the addition of the base to the mixed solution of precatalyst and acetophenone in 2-propanol leads to a red color of the resulting solution; this is the color of the reduced species 10 and 11. However, when greater than 5 equiv of base per iron complex was added, the color of the mixture was light yellow. These observations indicate that, during catalysis, the use of less base leads to the formation of red inactive species while in the

Figure 3. ORTEP plot of the cation of 13. Thermal ellipsoids are drawn at the 30% probability level. Some hydrogen atoms, solvent molecules, and counterions are omitted for clarity. Selected interatomic distances (Å) and angles (deg): Fe1−N1 2.053(3), Fe1−N2 2.046(3), Fe1−P1 2.266(1), Fe1−P2 2.270(1), N1−C2 1.484(4), C1−C2 1.516(4), N2−C5 1.469(4), C5−C6 1.508(5); P1− Fe1−P2 108.58(4).

no stoichiometric reaction of 10 or 11 with either 2-propanol or dihydrogen gas under various conditions. On the other hand, the amido(ene-amido) analogue, as we have reported previously, was found to react quickly with 2-propanol and slowly with dihydrogen gas at 25 °C to produce the amine iron hydrido complexes as key intermediates in the catalytic cycles.38,42 It is interesting to find that the only structural difference between the amido(ene-amido) complex 4 and the bis(amido) complex 10 or 11 is the presence of the CC double bond in 4 vs the C−C single bonds in 10 or 11. Dependence of Base Concentration on Catalysis and the Proposed Mechanism. It is clear from the results shown in Schemes 3 and 4 why more than 2 equiv of base per catalyst precursor is required to obtain catalytic activity. We propose the bifurcated mechanism shown in Scheme 6 to explain this critical base dependence. The first equivalent of base generates the chloro complex 6 (the kinetic product shown in Scheme 2). The second equivalent of base, when it is 2-propoxide, is 305

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis Scheme 6. Concentration of Base Determining How Much Inactive 10 and Active 4 Is Formed

Figure 4. Reaction profiles of base concentration dependence experiments. Conditions: [acetophenone] = 0.412 M; [1] = 6.73 × 10−5 M; [2propanol] = 12.4 M; 24 °C.

Simulation of the Reaction Profiles. The reaction profiles were fit to curves generated by the numerical integration of the equations of Scheme 7. The k1 and k‑1 values are effective rate constants, since they contain the concentration of the catalyst (and concentration of base; see below). A typical simulation is shown in Figure 5. This resulted in the effective rate constant k1 and equilibrium constant K1, as given in Table 1 and Figure 6. Proposed Explanation for the Base Dependence of Figure 6. Figure 6 shows that there are three regimes. At concentrations of base with base:catalyst ratios less than 2 there is very low activity, consistent with the catalyst being converted

presence of more base the formation of such inactive species is suppressed and the color is pale yellow, the color of an alkoxide complex or amido(ene-amido) complex 4. This is consistent with the observations made for the stoichiometric reactions. The acetophenone concentration dependence of the rate of catalysis was also studied (Figure S14 in the Supporting Information). In the initial 20 s, when the influence of 1phenylethanol on reaction kinetics is not significant, the reaction rates at different initial acetophenone concentrations are similar, in the range of 0.309−0.618 M. This indicates that, similar to the case of high-pressure H2 hydrogenation reactions, the rate of catalytic transfer hydrogenation is also independent of the concentration of acetophenone. 306

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis Scheme 7. Simplified ATH Equilibrium Reaction and Equations To Describe the Kinetics at a Constant Catalyst and Base Concentration

Figure 6. Effective rate constant k1 of Table 1 for the hydrogenation of acetophenone catalyzed by 1 as a function of the base to catalyst ratio with the initial [1] = 6.7 × 105 M.

other explanations. The deprotonation of the hydrido complex 5 to give an unreactive Fe(0) complex is unlikely, because the pKaTHF value of 5 is estimated to be greater than 3078 and the strongest base that can exist in 2-propanol is the pKa of the alcohol, approximately 18. In contrast, the base concentration has no effect on the rate of pressure hydrogenation of acetophenone catalyzed by 1 once it is activated with 2 equiv of KOtBu in THF.38 In this case, dihydrogen is not acidic enough to protonate the CC bond of the ene-amido group in catalyst 4 and cause its reduction to the inactive compound 10 and there is no 2-propoxide as a hydride source.38 Noyori’s group found that the activity of the ruthenium system [RuH(BH4)((S)-tolbinap)((S,S)-dpen)] for the pressure hydrogenation of acetophenone in 2-propanol went through a maximum rate as the base concentration was increased,50 in a fashion similar to that in our study. They found that a 10 mM concentration of base was optimum regardless of the base to catalyst ratio (which was actually approximately 20:1) and was independent of the nature of the base, either potassium alkoxide or neutral phosphazene base. In our case the optimum concentration is 0.5 mM of base with 0.07 mM iron catalyst. Chen and co-workers,49 Bergens and co-workers47 and Dub, Gordon, and co-workers48 have suggested that the increase in rate for the ruthenium system with base concentration is caused by the formation of a more reactive anionic ruthenium amido hydrido complex, with a potassium sitting on the amido group in the case when potassium alkoxide is used as the base. For our iron system the analogous mechanism would involve the deprotonation of the amine group of the hydride 5. This would have also been observed in the pressure hydrogenation study using 5 in THF with variable amounts of base if it were occurring, but it was not. In addition, DFT calculations to be described next are fully consistent with the amine-hydride complex 5 carrying out the enantiodetermining reduction of acetophenone. Exploratory DFT calculations on a potassium salt of an anionic iron complex did not lead to suitable structures or mechanism. Calculated Activity and Enantioselectivity Using DFT. Past theoretical studies40 on the catalytic cycle shown in Scheme 1 used simplified structures to reduce the computer time required for the DFT calculations. The turnover-limiting step, the reaction of an amido(ene-amido) model complex of 4 with 2-PrOH to give the hydride model complex of 5, was calculated to have a barrier of 12−13 kcal/mol. In contrast, the barrier for the same turnover-limiting step for the reaction of

Figure 5. Fit of the numerical solution to the equations of Scheme 7 for the ATH of acetophenone (0.412 M) in 2-propanol catalyzed by 1 (6.7 × 10−5 M) and KOtBu (2.2 × 10−4 M) at 25 °C.

Table 1. Relative Rate Constants k1 for the ATH of Acetophenone Catalyzed by 1 as a Function of the Base/ Catalyst Ratio KOtBu/1a [base]:[catalyst]

k1 (s−1)

K1

0 2.2 3 5 8 12

0.0 0.002 0.011 0.021 0.035 0.027

4.5 4.5 4.5 4.5 4.5

a Conditions: [acetophenone] = 0.412 M; [1] = 6.73 × 10−5 M; [2propanol] = 12.4 M; 24 °C.

into inactive iron compound 10, as shown in Scheme 6. Above 2 equiv of base per catalyst the rate increases with the base concentration until about 8 equiv of base per iron. The amount of 3 in Scheme 6 will be determined by the side equilibrium to form 14. Increasing the base concentration will thus feed more 3 and more active iron complexes 4 and 5 into the system. Excess base may also labilize more stable product alkoxide complexes that are present, as presented below. High concentrations of base (e.g., 12 equiv) slow catalysis. We speculate that the catalyst is unstable in the presence of high concentrations of alkoxides and enolates but that there may be 307

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis

Figure 7. Calculated ground states 4, 5, 10, 15, and 16 and transition states TSR (ν = 668i cm−1), TSs (ν = 671i cm−1), and TS4,10 (ν = 1133i cm−1). Free energies (Go) are given in kcal/mol relative to that of 4 plus the energies of all relevant small molecules.

complex 16 is significantly lower in energy (−8.5 kcal/mol). The pro-R and pro-S EDS are represented by TSR and TSS, respectively, and formation of the R product is lower in energy by 1.5 kcal/mol, also in agreement with the predominant formation of (R)-1-phenylethanol during catalysis (88% R). We have summarized our results of the comparison between observed and calculated activity/enantioselectivity in Table 2. Using conventional transition state theory,79 we calculated the TOF for each proposed transition state and compared it with

the saturated bis(amido) complex with 2-PrOH was 15−16 kcal/mol, significantly higher. This accounted for the lack of activity of the saturated species.12 Calculating the thermodynamic properties of the full catalyst system 4 by incorporating phenyl substituents on the phosphines and enantiopure diamine now allows us to obtain more accurate hydride transfer free energy values for the proposed catalyst cycle in Scheme 1. Furthermore, the enantiodetermining step (EDS) can be evaluated by comparing the difference in energies between the transition states that lead to the formation of R or S alcohols and the catalyst resting state(s). The bottom of Figure 7 shows three key ground state structures and two transition states proposed to be involved in catalysis that were calculated along with their free energies, all relative to 4, in addition to relevant small molecules under catalytic conditions (1 atm, 301 K, 2-propanol solvent continuum). The general trends are in agreement with previous calculations on a model system;40 FeH−NH complex 5 and 2propoxide complex 15 have similar free energies (−5.8 and −4.5 kcal/mol, respectively), while the product alkoxide

Table 2. Comparison of Experimental42 and Calculated (from DFT) Turnover Frequencies (s−1) at 25 °C and Enantioselectivity (ee, %) Parameters during the ATH of Acetophenone Using Amido(ene-amido) Catalyst 4

308

parameter

value

exptl TOF calcd TOF exptl ee calcd ee

119 s−1 1 to 102 s−1 88% 8 × 101 % DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis

Figure 8. (top) Three-dimensional space-filling (van der Waals) representations of the calculated EDS TSR. (middle) Three-dimensional cappedstick representations of the calculated EDS TSR with selected bond distances (Å). (bottom) Three-dimensional capped-stick representations of the calculated EDS TSS with selected bond distances (Å).

the observed value of 119 s−1, which corresponds to ΔGo = 14.8 kcal/mol. The calculations reveal that the most likely resting states during catalysis are the 2-propoxide complex 15 and the hydride 5 and that the maximum TOF is between approximately 102 and 1 s−1 (ΔGo = 16 ± 1 kcal/mol) to generate (R)-1-phenylethanol. Since 2-propanol is the reductant and reaction solvent, the catalyst will predominantly be in the alkoxide form during initial catalytic turnovers. As the product alcohol concentration increases, the more stable

alkoxide 16 will dominate in solution, reducing catalytic activity. The free energy difference between the pro-S and pro-R EDS is calculated to produce (R)-1-phenylethanol with an ee of (8 × 101)%, consistent with the observed ee of 88%. The top of Figure 7 also expands on our previous bis(amido) calculations with a model system.40 After reaction of 2-propanol and proton transfer to the ligand, there is a high imine reduction TS barrier TS4,10 (ΔGo = 22.6 kcal/mol) to generate bis(amido) complex 10. This is somewhat consistent with the 309

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis

of the hydrogenation of ketones but has not been characterized crystallographically. The results are consistent with the cycle of Scheme 1, with the hydride transfer from 2-propanol to the amido(ene-amido) complex 4 to form the hydrido complex 5 being the turnover-limiting step under the reaction conditions that we have examined. A transition state calculated using DFT methods for the hydrogenation of acetophenone explains the favored production of (R)-1-phenylethanol when using the (S,S)-amine(ene-amido) hydrido complex 5. There is an unusual base dependence of the ATH of acetophenone catalyzed by 1, where there is no activity until more than 2 equiv of base per iron is added and then maximum activity takes place at 8 equiv per iron. Below 2 equiv of base, an amido(ene-amido) intermediate is proposed to undergo protonation and hydride addition to give a bis(amido) complex that does not have ATH catalytic activity. Thus, the ene-amido moiety in the active catalyst structure is critically important. Conditions with greater than 2 equiv of base promote the formation of the active species by preventing the protonation and slow reduction of the alkene group in the ene(amido) complex; excess base might also labilize alkoxide resting states of the catalyst. The amount of active species increases in proportion to the base concentration until a maximum rate is reached.

observed slow conversion from 1 to 13 in the presence of methanol, as shown in Scheme 5, and the rapid formation of 10 from 4 by reaction with 2-propanol and 2 equiv of base (Scheme 3) suggests that additional 2-propanol molecules increase the rate of ligand reduction. Nevertheless, 10 is in fact more stable than amido(ene-amido) complex 4 (ΔGo = −5.6 kcal/mol) in the absence of 2-propanol. The lack of any observed catalytic activity of 10, even at elevated temperatures, is surprising considering that we reported previously that the energy barriers for an ATH cycle analogous to that shown in Scheme 1 but with a simplified version of 10 were only a few kilocalories per mole higher than for those using 4.40 Catalyst deactivation might occur via side reactions of 10 under catalytic conditions. This might involve ligand dissociation due to the pair of basic anionic, charge localized iron-amido moieties weakening the Fe−P bonds of the saturated alkylphosphine arms in the trans positions or amide protonation as in complex 13. However, this has been difficult to prove because of the low concentrations of catalyst used. We argue that the origins of enantioselectivity in the ATH of acetophenone using catalyst 4 arise from combined CH−π and π−π stabilizing interactions among the aryl ring of the ketone, an aryl phosphine ring, and the ene-amido moiety on the PNNP ligand. Figure 8 (top) shows space-filling models of the incoming ketone during the EDS for TSR, positioning the substrate aromatic ring over top of the ene-amido moiety. One phenyl ligand attached to phosphorus on the ene-amido side is also rotated nearly orthogonal to acetophenone to accommodate the incoming ketone. The capped-stick representations in Figure 8 (middle) show that the centroid of the aryl ring on acetophenone is positioned directly over top of the ene-amido carbon adjacent to nitrogen, with a π−π distance of 3.539 Å (measured from the ligand carbon to the centroid of the aromatic ring on acetophenone). Previous theoretical studies investigating aryl−aryl π interactions suggests that the calculated centroid−(ene-amido) distance in TSR is within range to stabilize the system, likely contributing to the predominant formation of (R)-1-phenylethanol during catalysis.80−82 Furthermore, a C−H group on the aromatic ring of acetophenone is interacting with the nearest arylphosphine ring, resulting in a CH−π interaction with a distance of 2.696 Å (measured from hydrogen to the centroid of the phenylphosphine ring). This type of CH−π interaction has been shown to stabilize the EDS in [RuH{(R,R)-XCH(Ph)CH(Ph)NH2}(η6-arene)] (X = NTs, O) complexes for the ATH of ketones.75 In transition state TSS, there are two separate CH−π interactions; however, there is no π−π stabilizing interaction and the structure is visibly more crowded, with the aryl group of the substrate interacting with the aryl groups of the phosphine (Figure 8, bottom). Our study of the ATH of other ketones catalyzed by solutions of 1 revealed that the highest ee values and rates were obtained for aryl ketones.42 Thus, the features of the transition states for acetophenone revealed here may apply generally to these other substrates.



EXPERIMENTAL SECTION General Considerations. All procedures and manipulations involving air-sensitive materials were performed under an argon or nitrogen atmosphere using Schlenk techniques or a glovebox with N2 or argon. Solvents were degassed and dried using standard procedures prior to all manipulations and reactions. Deuterated solvents were purchased from Cambridge Isotope Laboratories, Inc., and distilled and dried over activated molecular sieves. Complexes 1 and 2 were prepared as described previously.42,43 All of the other reagents used in the procedures were purchased from commercial sources and utilized without further purification. NMR spectra were recorded at ambient temperature and pressure using Varian Gemini 600, 400, and 500 MHz spectrometers (1H, 600, 500, and 400 MHz; 13C{1H}, 150, 125, and 100 MHz; 31P{1H}, 242, 202, and 161 MHz). The 31P NMR spectra were referenced to 85% H3PO4 (0 ppm). Elemental analyses were performed using a PerkinElmer 2400 CHN elemental analyzer at the Department of Chemistry at the University of Toronto. The electrospray ionization mass spectrometry (ESI-MS) data were collected on an AB/Sciex QStar mass spectrometer with an ESI source. Single-crystal X-ray diffraction data were collected using a Nonius Kappa-CCD diffractometer with Mo Kα radiation (λ = 0.71073 Å). The structures were solved and refined using SHELXTL V6.1. Synthesis of (S,S)-trans-[FeCl(CO)(PPh 2CH 2CH 2NCHPhCHPhNCHCHPPh2)] (8). In an argon glovebox, a ground powder of 1 (75 mg, 0.089 mmol) and KOtBu (10 mg, 0.089 mmol) were mixed in a vial (20 mL) that was charged with a stirring bar. THF (10 mL) was added, and the solution turned red immediately. The solution was further stirred at 25 °C for around 5 min until all the yellow powder in the bottom of the vial had disappeared. The solution was kept undisturbed in the argon glovebox for 1 week, and THF was removed under vacuum to yield a red powder. Benzene (10 mL) was added, and the resultant solution was filtered through a syringe filter PTFE membrane (pore size 0.2 μm), followed by filtration through a pad of Celite. Benzene



CONCLUSION The deprotonation of the iron carbonyl precatalyst complex 1 with 1 equiv of strong base allows the isolation of the welldefined amine(ene-amido) iron chloride complex 8, which shows a similar 1H NMR pattern and, more importantly, key NOE interactions identical with those of the amine(ene-amido) hydrido complex 5. This strengthens our NMR-based assignment of the structure of 5, which is found in the catalytic cycle 310

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis

218.5 (dd, JCP = 31.3 Hz, JCP = 30.3 Hz, CO). 31P{1H} NMR (202 MHz; C6D6): δ 55.7 (d, JPP = 31.2 Hz), 55.9 (d). Anal. Calcd for C45H43ClFeN2OP2: C, 69.20; H, 5.55; N, 3.59. Found: C, 69.05; H, 5.48; N, 3.57. HRMS (ESI-TOF, THF)” m/z calculated for [(C45H43ClFeN2OP2) + H]+ 781.1967; found, 781.1968. FT-IR (KBr, cm−1): 1931 s. Observation of Intermediate Isomeric Iron Complexes Shown in Scheme 2. The reaction of the complex 1 with 1 equiv of KOtBu in THF at 25 °C for 5 min gives a mixture of the three isomers 6−8. The 31P{1H} NMR spectrum of the reaction mixture in THF using a D2O insert for locking and shimming displayed three sets of doublets at δ 57.6 and 58.4 (2JPP = 33.0 Hz), δ 62.6 and 74.3 (2JPP = 50.3 Hz), and δ 55.4 and 57.2 (2JPP = 30.9 Hz). The last set of resonances is assigned to the amine(ene-amido) iron(II) complex 8. Over a period of 7 days, the other two isomers 6 and 7 are slowly converted to 8 in THF (see Figures S1 and S2 in the Supporting Information). The isomers 6 and 7 have mutually cis phosphorus nuclei; two possible structures that are sensible intermediates to 8 are shown in Scheme 2. The diastereomer with an N−H group anti-parallel to the Fe−Cl (6) is shown at the top of Scheme 2. This is proposed to be a kinetic product of deprotonation of the CH group α to the phosphorus atom of 1 and would be expected to have spectra very similar to those of 8. Thus, its 31P resonances are tentatively assigned as 57.6 and 58.4 ppm (2JPP = 33.0 Hz). The formation of the middle diastereomer 7 containing a cis-α P−N−NH−P ligand folded at the amine nitrogen could be an intermediate to 8, and its different structure would explain its 31P NMR spectrum, which is distinct from that of 8. The folding of the ligand P−NR− NR−P or P−NR−P at the NR to give fac ligand structures in metal complexes is a common occurrence: e.g. for R = H28,83−92 and R = alkyl.93−98 The reaction of the p-tolyl derivative 2 with 1 equiv of KOtBu under similar conditions (Scheme 2) initially generates only two isomers, both of which are then converted to complex 9 within 2 days (see Figures S3 and S4 in the Supporting Information). The p-tolyl analogue 9 has 1H, 31P{1H} and 2-D NMR spectra similar to those of 8, indicating that it has a structure similar to that of 8 (Figure S7 in the Supporting Information). The detailed structures of the two intermediate isomers are difficult to determine by NMR spectroscopy, due to the overlap of multiple resonances in the 1H NMR spectrum. They have similar νCO absorptions at 1931 cm−1 in the IR (KBr pellet) spectrum, with other peaks almost identical with those of 9 (see Figures S8 and S9 in the Supporting Information), indicating that these two diastereomers have structures consistent with those shown in Scheme 2. Reaction of 1 or 2 with 2 Equiv of Strong Base and Then 2-Propanol. Complex 1 has been reported to react with 2 equiv of base in THF to produce the amido(ene-amido) complexes 3 and 4 shown in Scheme 1.42 The amido(eneamido) isomers react with 2-propanol to afford red solutions, in either the presence or absence of extra base. This can be shown by first mixing the precatalyst complex 1 or 2 with base in THF with stirring for 5 min, followed by removing the THF under reduced vacuum to afford the amido(ene-amido) product along with the inorganic salts. 2-Propanol can then be added to the dark powder with vigorous stirring at 25 °C for 1 min, after which the 2-propanol is removed immediately under vacuum. When there is no excess of base present, controlled by adding 1.9 equiv of base to 1 to ensure that all the added base is consumed in the first step, 1H and 31P{1H} NMR spectra of the

was removed under vacuum to afford a red powder. Crystalline product was obtained by slow diffusion of diethyl ether into the solution of 8 in a mixture of benzene and hexane (volume ratio 1:1). Yield: 60 mg, 90%. 1H NMR (600 MHz, C6D6): δ 2.13− 2.16 (m, 1H CH2NH and 1H CH2P), 2.34 (brs, 1H, CH2P), 2.57 (m, 1H, CH2NH), 3.81 (dd, JHH = 10.6 Hz, JHH = 10.8 Hz, 1H, CH(Ph)NH), 4.14 (dd, JHH = 4.3 Hz, CHP), 5.02 (d, JHH = 10.6 Hz, 1H, CH(Ph)N), 5.32 (dd, JHH = 10.6 Hz, JHH = 10.8 Hz, NH), 6.36−8.15 (m, ArH, overlapping with signals of the NMR solvent). 7.60 (NCH, determined by 1H−13H HSQC spectrum). 13C{1H} NMR (150 MHz; C6D6): δ 34.1 (d, JCP = 22.0 Hz, PCH2), 49.5 (d, JCP = 4.8 Hz, NHCH2), 71.8 (dd, 1JCP = 54.9 Hz, 3JCP = 5.6 Hz, CHP), 75.5 (s, CH(Ph)), 79.0 (s, CH(Ph)), 127.4 (d, JCP = 9.0 Hz, ArC), 127.6−128.4 (m, ArC, overlapping with signals of the NMR solvent), 128.5−128.6 (m, ArC), 129.9 (d, JCP = 6.0 Hz, ArC), 132.7−132.9 (m, ArC), 134.5 (d, JCP = 9.0 Hz, ArC), 135.0 (d, JCP = 9.0 Hz, ArC), 137.1−138.0 (m, ArC), 137.9 (s, ArC), 140.9 (d, JCP = 43.5 Hz, ArC), 141.7 (s, ArC), 165.1 (d, JCP = 17.7 Hz, CHN), 218.8 (dd, CO). 31P{1H} NMR (242 MHz; THF): δ 55.4, 57.2, JPP = 30.9 Hz. Anal. Calcd for C43H39ClFeN2OP2: C, 68.58; H, 5.22; N, 3.72. Found: C, 68.55; H, 5.17; N, 3.66. HRMS (ESI-TOF, THF): m/z calculated for [(C43H39ClFeN2OP2) + H]+, 753.1654; found, 753.1658. FT-IR (KBr, cm−1): 1933 s, 3054 w. The 1H NMR spectrum of 8 in C6D6 exhibits one doublet of doublets at δ 5.32 for the NH proton, one doublet of doublets at δ 4.14 for the CH proton that is adjacent to phosphorus atom in the ene-amido group, and one doublet at δ 5.02 and one doublet of doublets at δ 3.81 for the protons of the CH(Ph)CH(Ph) moiety, respectively (Figure S5 in the Supporting Information). In the 2-D NOESY spectrum, a strong NOE interaction between the NH proton at δ 5.32 and the proton of the NCH(Ph) at 5.02 is observed (Figure 1 and Figure S6 in the Supporting Information). In addition, the CH(Ph)NH proton has an NOE interaction with the NHCHaxial hydrogen. These details together with the crystal structure of the p-tolyl analogue 9 are consistent with a structure with the hydrogen of the N−H bond close to the chlorine of the Fe−Cl bond. Synthesis of (S,S)-trans-[FeCl(CO)(PPh2CH2CH2NCHPhCHPhNCHCHPtol2)] (9). A procedure similar to that for 8 was used to synthesize 9, except that the THF solution after the deprotonation step was only kept still for 2 days. Single crystals suitable for X-ray analysis were obtained by slow diffusion of diethyl ether into the solution of 9 in a mixture of benzene and hexane (volume ratio 1:1). Yield: 90%. 1H NMR (500 MHz, C6D6): δ 1.99 (s, 3H, CH3), 2.02 (s, 3H, CH3), 2.16 (m, 1H, CH2NH, and 1H, CH2P), 2.37 (m, 1H, CH2P), 2.62 (m, 1H, CH2NH), 3.84 (dd, JHH = 10.5 Hz, JHH = 11.4 Hz, 1H, CH(Ph)NH), 4.17 (dd, JHH = 4.9 Hz, CHP), 5.04 (d, JHH = 10.5 Hz, 1H, CH(Ph)N), 5.38 (dd, JHH = 10.4 Hz, JHH = 11.4 Hz, NH), 6.84−7.85 (m, ArH, overlapping with signals of the NMR solvent). 13C{1H} NMR (125 MHz; C6D6): δ 20.8 (d, JCP = 1.0 Hz, CH3), 20.9 (d, JCP = 1.2 Hz, CH3), 33.8 (dd, JCP = 20.4 Hz, JCP = 2.9 Hz, PCH2), 49.2 (d, JCP = 4.8 Hz, NHCH2), 71.5 (dd, 1JCP = 52.0 Hz, 3JCP = 8.8 Hz, CHP), 75.2 (s, NCH(Ph)), 78.5 (s, NHCH(Ph)), 127.2 (s, ArC), 127.9 (s, ArC), 128.1−128.3 (m, ArC), 129.5 (d, JCP = 1.6 Hz, ArC), 132.3 (d, JCP = 9.5 Hz, ArC), 132.5 (d, JCP = 9.3 Hz, ArC), 134.2 (d, JCP = 8.9 Hz, ArC), 134.8 (d, JCP = 8.4 Hz, ArC), 137.6 (s, ArC), 137.7 (d, JCP = 2.7 Hz, ArC), 137.8 (d, JCP = 2.2 Hz, ArC), 141.4 (s, ArC), 164.5 (d, JCP = 17.0 Hz, CHN), 311

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis

material. 1H NMR (600 MHz, C6D6): δ 1.90 (s, 3H, CH3), 1.90 (s, 3H, CH3), 2.71 (m, 2H, CH2P), 3.03 (m, 2H, CH2P), 3.55− 3.61 (m, 4H, 2H, CH2N, and 2H, CH(Ph)), 4.05 (m, 2H, CH2N), 6.67−8.13 (m, ArH). 13C{1H} NMR (150 MHz; D2O insert in THF): δ 20.7 (s, CH3), 20.8 (s, CH3), 37.7−37.9 (m, PCH2), 52.3 (m, NCH2), 69.3 (m, NCH(Ph)), 125.8 (s, ArC), 127.3 (s, ArC), 127.7−128.5 (m, ArC), 128.8 (s, ArC), 129.0 (s, ArC), 132.2 (d, JCP = 9.3 Hz, ArC), 132.3 (d, JCP = 9.4 Hz, ArC), 132.8 (d, JCP = 11.9 Hz, ArC), 138.8 (m, ArC), 139.0 (m, ArC), 139.3 (s, ArC), 149.2 (m, ArC), 219.5 (m, CO). 31P{1H} NMR (242 MHz; C6D6): δ 82.7 (d, JPP = 45.9 Hz), 84.2 (d). FT-IR (KBr, cm−1): 1869 s. The 1H NMR spectrum of 11 displayed four sets of broad multiplet resonances at δ 2.71, 3.03, 3.57, and 4.05 for the ethylene backbone moieties. The presence of the CH(Ph) protons was confirmed by 1H−13C HSQC and HMBC spectroscopy (see Figures S12 and S13 in the Supporting Information). The 31P{1H} NMR (C6D6) spectrum of 11 displayed two phosphorus doublet resonances at δ 82.7 and 84.2 with JPP = 45.9 Hz, due to the inequivalent phosphorus atoms. This is in contrast with the spectrum of 10, for which only one broad resonance is observed at 25 °C. Observation of Hydride (S,S)-trans-[FeH(CO)(PPh2CH2CH2NHCHPhCHPhNCHCHPtol2)] (12). The procedure for synthesis of 12 is similar to that for 11 except that 2.5 equiv (or more) of base was employed. 1H NMR (600 MHz, C6D6): δ −2.25 (dd, 2JHP = 71.0 and 71.8 Hz), 1.98 (brs, 3H, CH3), 2.02 (brs, 3H, CH3), 3.81 (dd, JHH = 10.8 Hz, JHH = 11.2 Hz, 1H, CH(Ph)NH), 4.01 (m, 1H, NH), 4.36 (dd, JHP = 2.0 Hz, JHH = 4.4 Hz, 1H, PCH), 4.53 (d, JHH = 10.8 Hz, CH(Ph)NCH); remaining resonances were obscured by those of other species. Crystallization of 13. In an argon glovebox, a finely ground powder of 1 (75 mg, 0.089 mmol) and KOtBu (10 mg, 0.089 mmol) were mixed in a vial (20 mL) that was charged with a stirring bar. THF (10 mL) was added, and the solution turned red immediately. The solution was further stirred at 25 °C for around 5 min until all the yellow powder in the bottom of the vial disappeared. The solution was kept still in the argon glovebox for 1 week, THF was removed under vacuum, and C6D6 was added. The resultant red mixture was transferred to a J. Young tube with a pipet. One drop of methanol and one drop of CH2Cl2 (thought to help stabilize the crystal lattice) were also added, and the tube was sealed. The solution was kept still in the argon glovebox for 1 week, and light red crystals formed. Catalysis. The same procedure as we reported before was used for catalysis.42 Potassium tert-butoxide was dried under vacuum at 80 °C for 2 h prior to use in order to remove potential traces of 2-methyl-2-propanol.42 Calculations. DFT calculations were performed using Gaussian 09 (rev. D.01)99 at the M06L100/TZVP101/TZVPfit level of theory. This level of theory was chosen for being suitably fast and accurate on systems of this scale (90−110 atoms),102 and optimization cycles were several times faster in comparison with our previously reported conditions.40 Either normal (opt) or tight (opt = tight) convergence criteria were met using a pruned (99,590) integration grid (grid = ultrafine). Optimizations were performed in an 2-propanol solvent continuum (1 atm, 298 K) using the integral equation formalism polarizable continuum model (IEF-PCM)103,104 with radii and nonelectrostatic terms from the SMD solvation model (scrf = smd).105 Full vibrational and thermochemical analyses were performed on optimized structures to obtain

reaction mixture in C6D6 reveal the formation of an equimolar mixture of compounds 10 and 5 (Scheme 3). Observation of (S,S)-[Fe(CO)(PPh 2CH 2CH 2 NCHPhCHPhNCH2CH2PPh2)] (10). In an argon glovebox, a finely ground powder of 1 (39 mg, 0.047 mmol) and KOtBu (10 mg, 0.089 mmol, 1.9 equiv relative to iron) were mixed in a vial (20 mL) that was charged with a stirring bar. THF (10 mL) was added, and the solution turned blue immediately. Note that complex 1 and KOtBu should be extremely pure. Recently we have updated the synthetic protocols of complexes 1 and 2 by washing the crude product with water to completely remove the inorganic salts followed by recrystallization in hot methanol to afford pure products in a larger scale synthesis.43 The potassium tert-butoxide was dried under vacuum at 80 °C for 2 h prior to use in order to remove the potential 2-methyl-2propanol. The reaction mixture was stirred at 25 °C for 5 min until all the yellow powder in the bottom of the vial disappeared. THF was removed under vacuum to yield a dark powder. 2-Propanol (2 mL) was added to the powder, and the resultant solution was stirred vigorously at 25 °C for 1 min. The color changed immediately from black to red. The 2propanol was immediately removed under vacuum using the small antechamber of the glovebox. The obtained red powder was dried under vacuum overnight and later extracted with C6D6 for NMR analysis. The IR sample was prepared inside the glovebox using a KBr pellet. 1H NMR (600 MHz, C6D6): δ 2.68 (m, 2H, CH2P), 3.02 (m, 2H, CH2P), 3.56−3.57 (m, 2H, CH2N and 2H, CH(Ph)), 4.04 (m, 2H, CH2N), 6.79−7.71 (m, ArH). 31P{1H} NMR (242 MHz; C6D6): δ brs 84.5. 13C{1H} NMR (150 MHz; C6D6; determined by 1H−13C HSQC); δ 37.4−37.5 (PCH2), 52.4 (NCH2), 69.2 (NCH(Ph)). FT-IR (KBr, cm−1): 1839 s. In the 1H NMR spectrum of 10, the P−N−N−P ligand backbone shows four sets of broad resonances at δ 2.68, 3.02, 3.57, and 4.04, which correlate in the 1H−13C heteronuclear single quantum coherence (HSQC) spectrum with the two secondary carbon resonances at δ 37.5 and 52.4, identified by a distortionless enhancement of polarization transfer (DEPT) NMR experiment. In addition, the signals of the two CH(Ph) protons are assigned by their correlations at (δ 3.57, 69.2) in the 1H−13C HSQC spectrum and are found to overlap with those of the CH2 groups in the δ 3.57 region of the 1H NMR spectrum. Complex 10 is fluxional in solution and produces broad proton resonances at 298 K (Figure S10). The 31P{1H} NMR spectrum also displays a broad singlet at δ 84.5 for the two phosphorus atoms. When the NMR sample solution containing 5 and 10 was kept overnight at 25 °C, the signals for the hydride resonance of 5 disappeared and resonances due to the chloro complex 8 grew in (KCl was still present). The corresponding 31P{1H} NMR (C6D6) spectrum showed one major compound 10 along with 8 as two doublets at δ 55.4 and 57.2 (JPP = 30.9 Hz) (Figure S10 in the Supporting Information). This indicates that the hydrido complex 5 is slowly converted to the chloro analogue 8 in the NMR sample tube in the presence of a chloride and a proton source. It is presumed that this conversion proceeds via the reaction of the hydride of 5 with a weak acid, most likely residual 2-propanol, to form the ion-paired dihydrogen 2-propoxide complex.51 The release of dihydrogen and chloride coordination would lead to the formation of 8. Observation of (S,S)-[Fe(CO)(PPh 2CH 2CH 2 NCHPhCHPhNCH2CH2Ptol2)] (11). A procedure similar to that for 10 was employed except using complex 2 as the starting 312

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis

(16) Moulin, S.; Dentel, H.; Pagnoux-Ozherelyeva, A.; Gaillard, S.; Poater, A.; Cavallo, L.; Lohier, J. F.; Renaud, J.-L. Chem. - Eur. J. 2013, 19, 17881−17890. (17) Wienhöfer, G.; Baseda-Krüger, M.; Ziebart, C.; Westerhaus, F. A.; Baumann, W.; Jackstell, R.; Junge, K.; Beller, M. Chem. Commun. 2013, 49, 9089−9091. (18) Wienhofer, G.; Westerhaus, F. A.; Junge, K.; Ludwig, R.; Beller, M. Chem. - Eur. J. 2013, 19, 7701−7707. (19) Fong, H.; Moret, M. E.; Lee, Y.; Peters, J. C. Organometallics 2013, 32, 3053−3062. (20) Li, Y.; Yu, S.; Wu, X.; Xiao, J.; Shen, W.; Dong, Z.; Gao, J. J. Am. Chem. Soc. 2014, 136, 4031−4039. (21) Sues, P. E.; Demmans, K. Z.; Morris, R. H. Dalton Trans. 2014, 43, 7650−7667. (22) Chakraborty, S.; Dai, H.; Bhattacharya, P.; Fairweather, N. T.; Gibson, M. S.; Krause, J. A.; Guan, H. J. Am. Chem. Soc. 2014, 136, 7869−7872. (23) Morris, R. H. Acc. Chem. Res. 2015, 48, 1494−1502. (24) Chirik, P. J. Acc. Chem. Res. 2015, 48, 1687−1695. (25) Zell, T.; Milstein, D. Acc. Chem. Res. 2015, 48, 1979−1994. (26) Chakraborty, S.; Bhattacharya, P.; Dai, H.; Guan, H. Acc. Chem. Res. 2015, 48, 1995−2003. (27) Mérel, D. S.; Do, M. L. T.; Gaillard, S.; Dupau, P.; Renaud, J.-L. Coord. Chem. Rev. 2015, 288, 50−68. (28) Bigler, R.; Huber, R.; Mezzetti, A. Angew. Chem., Int. Ed. 2015, 54, 5171−5174. (29) Hopewell, J. P.; Martins, J. E. D.; Johnson, T. C.; Godfrey, J.; Wills, M. Org. Biomol. Chem. 2012, 10, 134−145. (30) Lu, L. Q.; Li, Y.; Junge, K.; Beller, M. J. Am. Chem. Soc. 2015, 137, 2763−2768. (31) Misal Castro, L. C.; Li, H. Q.; Sortais, J. B.; Darcel, C. Green Chem. 2015, 17, 2283−2303. (32) Bauer, I.; Knölker, H. J. Chem. Rev. 2015, 115, 3170−3387. (33) Bertini, F.; Mellone, I.; Ienco, A.; Peruzzini, M.; Gonsalvi, L. ACS Catal. 2015, 5, 1254−1265. (34) Li, H. X.; Hall, M. B. ACS Catal. 2015, 5, 1895−1913. (35) Yang, X. ACS Catal. 2012, 2, 964−970. (36) Lagaditis, P. O.; Sues, P. E.; Sonnenberg, J. F.; Wan, K. Y.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2014, 136, 1367−1380. (37) Sonnenberg, J. F.; Lough, A. J.; Morris, R. H. Organometallics 2014, 33, 6452−6465. (38) Zuo, W.; Tauer, S.; Prokopchuk, D. E.; Morris, R. H. Organometallics 2014, 33, 5791−5801. (39) Lagaditis, P. O.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2011, 133, 9662−9665. (40) Prokopchuk, D. E.; Morris, R. H. Organometallics 2012, 31, 7375−7385. (41) Mikhailine, A. A.; Maishan, M. I.; Morris, R. H. Org. Lett. 2012, 14, 4638−4641. (42) Zuo, W.; Lough, A. J.; Li, Y. F.; Morris, R. H. Science 2013, 342, 1080−1083. (43) Zuo, W.; Morris, R. H. Nat. Protoc. 2015, 10, 241−257. (44) Smith, S. A. M.; Morris, R. H. Synthesis 2015, 47, 1775−1779. (45) Hamilton, R. J.; Bergens, S. H. J. Am. Chem. Soc. 2006, 128, 13700−13701. (46) Hamilton, R. J.; Bergens, S. H. J. Am. Chem. Soc. 2008, 130, 11979−11987. (47) John, J. M.; Takebayashi, S.; Dabral, N.; Miskolzie, M.; Bergens, S. H. J. Am. Chem. Soc. 2013, 135, 8578−8584. (48) Dub, P. A.; Henson, N. J.; Martin, R. L.; Gordon, J. C. J. Am. Chem. Soc. 2014, 136, 3505−3521. (49) Hartmann, R.; Chen, P. Angew. Chem., Int. Ed. 2001, 40, 3581− 3585. (50) Sandoval, C. A.; Ohkuma, T.; Muñiz, K.; Noyori, R. J. Am. Chem. Soc. 2003, 125, 13490−13503. (51) Abdur-Rashid, K.; Clapham, S. E.; Hadzovic, A.; Harvey, J. N.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2002, 124, 15104−15118. (52) Abdur-Rashid, K.; Faatz, M.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2001, 123, 7473−7474.

solvent-corrected free energies (G°) and enthalpies (H°). The freqchk utility was used to obtain thermochemistry data under catalytic conditions (1 atm, 301 K). Optimized ground states were found to have 0 imaginary frequencies while transition states had 1 imaginary frequency. Three-dimensional representations were generated by Mercury 3.6.106 Calculations were performed in part by the facilities of the Shared Hierarchical Academic Research Computing Network (SHARCNET) and Compute/Calcul Canada.107



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscatal.5b01979. NMR and IR spectra, reaction conversion profiles, DFT calculated structures and metrical parameters, and DFT energies and Cartesian coordinates (PDF) Crystallographic data for 9 (CIF) Crystallographic data for 13 (CIF)



AUTHOR INFORMATION

Corresponding Author

*E-mail for R.H.M.: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors wish to acknowledge the Canadian Foundation of Innovation, project number 19119, and the Ontario Research Fund for funding of the Centre for Spectroscopic Investigation of Complex Organic Molecules and Polymers. R.H.M. thanks the Natural Sciences and Engineering Research Council (NSERC) for a Discovery grant, and D.E.P. thanks the NSERC for scholarship funding. W.Z. thanks the “The Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning”.



REFERENCES

(1) Daida, E. J.; Peters, J. C. Inorg. Chem. 2004, 43, 7474−7485. (2) Bart, S. C.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2004, 126, 13794−13807. (3) Enthaler, S.; Erre, G.; Tse, M. K.; Junge, K.; Beller, M. Tetrahedron Lett. 2006, 47, 8095−8099. (4) Sui-Seng, C.; Freutel, F.; Lough, A. J.; Morris, R. H. Angew. Chem., Int. Ed. 2008, 47, 940−943. (5) Enthaler, S.; Junge, K.; Beller, M. Angew. Chem., Int. Ed. 2008, 47, 3317−3321. (6) Casey, C. P.; Guan, H. J. Am. Chem. Soc. 2009, 131, 2499−2507. (7) Langer, R.; Diskin-Posner, Y.; Leitus, G.; Shimon, L. J. W.; BenDavid, Y.; Milstein, D. Angew. Chem., Int. Ed. 2011, 50, 9948−9952. (8) Morris, R. H. Chem. Soc. Rev. 2009, 38, 2282−2291. (9) Junge, K.; Schroder, K.; Beller, M. Chem. Commun. 2011, 47, 4849−4859. (10) Langer, R.; Leitus, G.; Ben-David, Y.; Milstein, D. Angew. Chem., Int. Ed. 2011, 50, 2120−2124. (11) Darwish, M.; Wills, M. Catal. Sci. Technol. 2012, 2, 243−255. (12) Mikhailine, A. A.; Maishan, M. I.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2012, 134, 12266−12280. (13) Ziebart, C.; Federsel, C.; Anbarasan, P.; Jackstell, R.; Baumann, W.; Spannenberg, A.; Beller, M. J. Am. Chem. Soc. 2012, 134, 20701− 20704. (14) Fleischer, S.; Zhou, S. L.; Werkmeister, S.; Junge, K.; Beller, M. Chem. - Eur. J. 2013, 19, 4997−5003. (15) Gopalaiah, K. Chem. Rev. 2013, 113, 3248−3296. 313

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314

Research Article

ACS Catalysis (53) Abdur-Rashid, K.; Lough, A. J.; Morris, R. H. Organometallics 2000, 19, 2655−2657. (54) Clapham, S. E.; Iuliis, M. Z.-D.; Mack, K.; Prokopchuk, D. E.; Morris, R. H. Can. J. Chem. 2014, 92, 731−738. (55) Clapham, S. E.; Morris, R. H. Organometallics 2005, 24, 479− 481. (56) Kuzma, M.; Vaclavik, J.; Novak, P.; Prech, J.; Januscak, J.; Cerveny, J.; Pechacek, J.; Sot, P.; Vilhanova, B.; Matousek, V.; Goncharova, II; Urbanova, M.; Kacer, P. Dalton Trans. 2013, 42, 5174−5182. (57) Carrion, M. C.; Sepulveda, F.; Jalon, F. A.; Manzano, B. R.; Rodriguez, A. M. Organometallics 2009, 28, 3822−3833. (58) Sandoval, C. A.; Li, Y. H.; Ding, K. L.; Noyori, R. Chem. - Asian J. 2008, 3, 1801−1810. (59) Noyori, R.; Yamakawa, M.; Hashiguchi, S. J. Org. Chem. 2001, 66, 7931−7944. (60) Baratta, W.; Siega, K. Pierluigi Rigo. Chem. - Eur. J. 2007, 13, 7479−7486. (61) Vastila, P.; Zaitsev, A. B.; Wettergren, J.; Privalov, T.; Adolfsson, H. Chem. - Eur. J. 2006, 12, 3218−3225. (62) Zhang, H.; Yang, C. B.; Li, Y. Y.; Donga, Z. R.; Gao, J. X.; Nakamura, H.; Murata, K.; Ikariya, T. Chem. Commun. 2003, 51, 142− 143. (63) Yamakawa, M.; Ito, H.; Noyori, R. J. Am. Chem. Soc. 2000, 122, 1466−1478. (64) Guo, R.; Elpelt, C.; Chen, X.; Song, D.; Morris, R. H. Org. Lett. 2005, 7, 1757−1759. (65) Chelucci, G.; Baldino, S.; Baratta, W. Acc. Chem. Res. 2015, 48, 363−379. (66) Aranyos, A.; Csjernyik, G.; Szabo, K. J.; Bäckvall, J. E. Chem. Commun. 1999, 351−352. (67) Haack, K. J.; Hashiguchi, S.; Fujii, A.; Ikariya, T.; Noyori, R. Angew. Chem., Int. Ed. Engl. 1997, 36, 285−288. (68) Hadzovic, A.; Song, D.; MacLaughlin, C. M.; Morris, R. H. Organometallics 2007, 26, 5987−5999. (69) Basolo, F.; Pearson, R. G. Mechanisms of Inorganic Reactions; Wiley: New York, 1967. (70) Reetz, M. T.; Meiswinkel, A.; Mehler, G.; Angermund, K.; Graf, M.; Thiel, W.; Mynott, R.; Blackmond, D. G. J. Am. Chem. Soc. 2005, 127, 10305−10313. (71) Di Tommaso, D.; French, S. A.; Zanotti-Gerosa, A.; Hancock, F.; Palin, E. J.; Catlow, C. R. A. Inorg. Chem. 2008, 47, 2674−2687. (72) Gridnev, I. D.; Imamoto, T.; Hoge, G.; Kouchi, M.; Takahashi, H. J. Am. Chem. Soc. 2008, 130, 2560−2572. (73) Maeda, S.; Ohno, K. J. Am. Chem. Soc. 2008, 130, 17228−17229. (74) Wang, T.; Zhuo, L.-G.; Li, Z.; Chen, F.; Ding, Z.; He, Y.; Fan, Q.-H.; Xiang, J.; Yu, Z.-X.; Chan, A. S. C. J. Am. Chem. Soc. 2011, 133, 9878−9891. (75) Dub, P. A.; Ikariya, T. J. Am. Chem. Soc. 2013, 135, 2604−2619. (76) Faza, O. N.; Fernandez, I.; Lopez, C. S. Chem. Commun. 2013, 49, 4277−4279. (77) Feng, R.; Xiao, A.; Zhang, X.; Tang, Y.; Lei, M. Dalton Trans. 2013, 42, 2130−2145. (78) Morris, R. H. J. Am. Chem. Soc. 2014, 136, 1948−1959. (79) McQuarrie, D. A.; Simon, J. D. Physical Chemistry: A Molecular Approach; University Science Books: South Orange, NJ, 1997. (80) Sinnokrot, M. O.; Sherrill, C. D. J. Phys. Chem. A 2003, 107, 8377−8379. (81) Mignon, P.; Loverix, S.; De Proft, F.; Geerlings, P. J. Phys. Chem. A 2004, 108, 6038−6044. (82) Arnstein, S. A.; Sherrill, C. D. Phys. Chem. Chem. Phys. 2008, 10, 2646−2655. (83) Bigler, R.; Otth, E.; Mezzetti, A. Organometallics 2014, 33, 4086−4099. (84) Friedrich, A.; Drees, M.; Kass, M.; Herdtweck, E.; Schneider, S. Inorg. Chem. 2010, 49, 5482−5494. (85) Sgro, M. J.; Stephan, D. W. Dalton Trans. 2013, 42, 10460− 10472.

(86) Prokopchuk, D. E.; Lough, A. J.; Morris, R. H. Dalton Trans. 2011, 40, 10603−10608. (87) Atoh, M.; Kashiwabara, K.; Fujita, J. Bull. Chem. Soc. Jpn. 1985, 58, 3492−3499. (88) Fillaut, J.-L.; de los Rios, I.; Masi, D.; Romerosa, A.; Zanobini, F.; Peruzzini, M. Eur. J. Inorg. Chem. 2002, 2002, 935−942. (89) Atoh, M.; Sorensen, H. O.; Andersen, P. Acta Chem. Scand. 1997, 51, 1169−1177. (90) Chin, R. M.; Dubois, R. H.; Helberg, L. E.; Sabat, M.; Bartucz, T.; Morris, R. H.; Harman, W. D. Inorg. Chem. 1997, 36, 3553−3558. (91) Rozenel, S. S.; Padilla, R.; Camp, C.; Arnold, J. Chem. Commun. 2014, 50, 2612−2614. (92) Rozenel, S. S.; Arnold, J. Inorg. Chem. 2012, 51, 9730−9739. (93) Sgro, M. J.; Dahcheh, F.; Stephan, D. W. Organometallics 2014, 33, 578−586. (94) Mealli, C.; Sabat, M.; Zanobini, F.; Ciampolini, M.; Nardi, N. J. Chem. Soc., Dalton Trans. 1985, 479−485. (95) Bianchini, C.; Masi, D.; Meli, A.; Peruzzini, M.; Sabat, M.; Zanobini, F. Organometallics 1986, 5, 2557−2559. (96) Rossin, A.; Gutsul, E. I.; Belkova, N. V.; Epstein, L. M.; Gonsalvi, L.; Lledos, A.; Lyssenko, K. A.; Peruzzini, M.; Shubina, E. S.; Zanobini, F. Inorg. Chem. 2010, 49, 4343−4354. (97) MacBeth, C. E.; Harkins, S. B.; Peters, J. C. Can. J. Chem. 2005, 83, 332−340. (98) Bianchini, C.; Farnetti, E.; Graziani, M.; Nardin, G.; Vacca, A.; Zanobini, F. J. Am. Chem. Soc. 1990, 112, 9190−9197. (99) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ã .; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09 Revision D.01; Gaussian, Inc., Wallingford, CT, 2009. (100) Zhao, Y.; Truhlar, D. G. J. Chem. Phys. 2006, 125, 194101. (101) Schäfer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100, 5829−5835. (102) Gusev, D. G. Organometallics 2013, 32, 4239−4243. (103) Tomasi, J.; Mennucci, B.; Cancès, E. J. Mol. Struct.: THEOCHEM 1999, 464, 211−226. (104) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105, 2999−3094. (105) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B 2009, 113, 6378−6396. (106) Macrae, C. F.; Bruno, I. J.; Chisholm, J. A.; Edgington, P. R.; McCabe, P.; Pidcock, E.; Rodriguez-Monge, L.; Taylor, R.; van de Streek, J.; Wood, P. A. J. Appl. Crystallogr. 2008, 41, 466−470. (107) www.sharcnet.ca, accessed Nov 26, 2015.

314

DOI: 10.1021/acscatal.5b01979 ACS Catal. 2016, 6, 301−314