Iron(II)-Catalyzed Hydrogenation of Acetophenone with a Chiral, Pyridine-Based PNP Pincer Ligand: Support for an Outer-Sphere Mechanism

Article abstract of DOI:10.1021/acs.organomet.7b00816

We report here the tridentate, P-stereogenic, C₂-symmetric PNP pincer ligand (S_P,S_P)-2,6-bis((cyclohexyl(methyl)phosphanyl)methyl)pyridine (1a) and its iron(II) complexes [FeBr₂(CO)(1a)] (2a), [FeHBr(CO)(1a)] (3a), and [FeH₂(CO)(1a)] (4a). In the presence of base, bromocarbonylhydride 3a catalyzes the hydrogenation of acetophenone to (S)-1-phenylethanol with 48% ee. The transition states of the enantiodetermining transfer of hydride from 3a to the carbonyl group of acetophenone were studied by density functional theory (DFT) with a full conformational analysis of the PNP ligand for the three different mechanistic models recently proposed for a related achiral catalyst. The DFT calculations show that the outer-sphere monohydride mechanism originally proposed by Milstein reproduces the experimentally observed sense of induction (S) and enantioselectivity, whereas the dihydride and inner-sphere pathways predict the formation of the R enantiomer.

Full text of DOI:10.1021/acs.organomet.7b00816

Article
Cite This: Organometallics XXXX, XXX, XXX−XXX
Iron(II)-Catalyzed Hydrogenation of Acetophenone with a Chiral,
Pyridine-Based PNP Pincer Ligand: Support for an Outer-Sphere
Mechanism
Raffael Huber, Alessandro Passera, and Antonio Mezzetti*
Department of Chemistry and Applied Biosciences, ETH Zurich, Zurich 8093, Switzerland
S
* Supporting Information
ABSTRACT: We report here the tridentate, P-stereogenic,
C2-symmetric PNP pincer ligand (SP,SP)-2,6-bis((cyclohexyl-
(methyl)phosphanyl)methyl)pyridine (1a) and its iron(II)
complexes [FeBr2(CO)(1a)] (2a), [FeHBr(CO)(1a)] (3a),
and [FeH2(CO)(1a)] (4a). In the presence of base,
bromocarbonylhydride 3a catalyzes the hydrogenation of
acetophenone to (S)-1-phenylethanol with 48% ee. The
transition states of the enantiodetermining transfer of hydride
from 3a to the carbonyl group of acetophenone were studied
by density functional theory (DFT) with a full conformational analysis of the PNP ligand for the three different mechanistic
models recently proposed for a related achiral catalyst. The DFT calculations show that the outer-sphere monohydride
mechanism originally proposed by Milstein reproduces the experimentally observed sense of induction (S) and enantioselectivity,
whereas the dihydride and inner-sphere pathways predict the formation of the R enantiomer.
In contrast to backbone-based chirality, stereogenic P
atomswhich are recognized as powerful units to impart
enantioselectivity in catalytic processes9are still virtually
unexplored with pincer ligands.10 An achiral scaffold that can be
readily modified to a chiral, P-stereogenic PNP pincer is 2,6-
dimethylenepyridine. Milstein has used achiral PNP pincer
ligands such as 1b (Chart 2) in the iron-catalyzed hydro-
INTRODUCTION
■
Since Casey’s and Guan’s seminal report of the first well-
defined iron catalyst for the hydrogenation of ketones,1 several
iron(II) catalysts have been developed for the asymmetric
hydrogenation of polar double bonds.2 Its enantioselective
version is one of the most important catalytic reactions in
industry,3 and two chiral ligand platforms play a major role with
iron(II). Tetradentate N2P2 ligands, either linear4 or macro-
cyclic,5,6 efficiently catalyze the asymmetric transfer hydro-
genation (ATH) of polar double bonds, whereas PNP pincers
are the ligands of choice in the asymmetric direct (H2)
hydrogenation (AH, Chart 1).7 Thus, complexes of the type
Chart 2. PNP Ligands and [FeBr2(CO)(PNP)] Complexes
Chart 1. Selection of Iron Catalysts for Asymmetric Direct
Hydrogenation
genation of ketones,11 esters,12 aldehydes,13 and amides.14 In
view of these remarkable results, we decided to prepare chiral,
C2-symmetric PNP ligands such as 1a in Chart 2 and to study
their Fe(II) complexes as catalysts for the asymmetric direct
hydrogenation of ketones.
During the development of this project, we realized that a
chiral PNP ligand based on the 2,6-dimethylenepyridine
backbone may help disentangle the controversy concerning
the mechanism of ketone hydrogenation with their iron(II)
complexes. In fact, three different mechanism types have been
[FeHBr(CO)(PNP)] (where PNP is a chiral pincer ligand with
an NP2 donor set) have found successful application in the
asymmetric direct hydrogenation of ketones.8 Morris prepared
chiral PNP pincers based on a disubstituted 2-phosphinoethy-
lamine,8a,c and Zirakzadeh introduced a ferrocenyl link between
the amine and a phosphine as further stereogenic element
(Chart 1).8b
Received: November 9, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.organomet.7b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
proposed for the direct hydrogenation of ketones with the
achiral bromocarbonylhydride [FeHBr(CO)(1b)] (3b) using
density functional theory (DFT) calculations (Chart 3).15−17
methylphosphine borane (7) was prepared by enantioselective
deprotonation of cyclohexyldimethylphosphine borane (6)
with (−)-sparteine and sec-butyllithium, followed by O2
oxidation (Scheme 2).20,21
Chart 3. Proposed Transition States for the Direct
Hydrogenation of Acetophenone with Catalyst 3b
Scheme 2. Synthesis of (R)-
Cyclohexyl(hydroxymethyl)methylphosphine Borane 7
The first suggestion,16 which has been recently revised,17 is the
outer-sphere mechanism D based on the dihydride complex
[FeH2(CO)(1b)] (4b) as catalytically active species. Alter-
natively, Milstein considered an inner-sphere mechanism
involving hydride transfer from a complex bearing a
deprotonated PNP ligand (I).11 This mechanism was later
discarded because of its unrealistically high calculated energetic
span18 (50.8 kcal mol−1), and an unprecedented outer-sphere
mechanism was put forward in which an iron(0) complex
transfers a benzylic H atom of the PNP ligand as hydride to
acetophenone (O).15 The much lower energetic span of O
(21.7/24.5 kcal mol−1) compared to that of I and the
observation that dihydride 4b is unreactive toward acetophe-
none make O the most likely mechanism.15
Crude hydroxymethylphosphine borane 7 was isolated as an
oil in 75% ee, and its enantiomeric purity was improved by
esterification with p-phenylbenzoyl chloride to (boranyl-
(cyclohexyl)(methyl)phosphanyl)methyl[1,1′-biphenyl]-4-car-
boxylate (8, Scheme 3).22 This highly crystalline ester was
Scheme 3. Synthesis of Enantiomerically Pure Secondary
Phosphine Borane 5
As the use of a chiral catalyst introduces additional
stereochemical information, we wondered whether the
enantioselectivity and sense of chiral induction observed
experimentally and those estimated by DFT calculations may
allow discriminating between mechanisms I, O, and D.
Therefore, we prepared the P-stereogenic PNP pincer ligand
1a and its mono- and dihydride complexes and tested them as
catalysts in the asymmetric hydrogenation of ketones. Parallel
to this, we studied the transition states of the enantiodetermin-
ing hydride transfer to the carbonyl group of acetophenone by
DFT with the three mechanistic models recently proposed for
the related achiral catalyst 2b.15−17 The results of this
combined experimental and DFT study are reported below.
recrystallized from ethyl acetate/hexane until it was optically
pure, which typically required two crystallizations steps. After
hydrolysis and oxidation−decarboxylation,21 the secondary
phosphine borane 5 was obtained with 99.6% ee in 44% yield
over three steps.
The secondary phosphine borane 5 was deprotonated with
nBuLi, and addition of 2,6-bis(chloromethyl)pyridine gave the
borane-protected pincer ligand 9a in high yield (90%) (Scheme
4).
[FeBr2(CO)(PNP)]. Phosphine borane 9a was deprotected
with HBF4·OEt2 in dichloromethane (Scheme 5).23,24 Treat-
ment of the free ligand 1a with FeBr2 under a CO atmosphere
(1.1 atm) gave the deep blue dibromocarbonyl complex
[FeBr2(CO)(1a)] (2a), which was precipitated with pentane,
filtered off in air, and purified by washing with water, ethanol,
RESULTS AND DISCUSSION
■
Synthesis of the Pincer Ligand. As a chiral analogue of
1b, we prepared the pincer ligand (SP,SP)-2,6-bis((cyclohexyl-
(methyl)phosphanyl)methyl)pyridine (1a) by alkylation of 2,6-
bis(chloromethyl)pyridine with (S)-cyclohexylmethyl-
phosphine borane 5 (Scheme 1).
Scheme 1. Synthons for Pincer Ligand 1a
Scheme 4. Preparation of Borane-Protected PNP Pincer
Ligand 9a
The key synthon for 1a is (S)-cyclohexylmethylphosphine
borane (5, Scheme 1). Enantiomerically pure cyclohexylmethyl-
phosphine borane 5 has been previously prepared by separation
of the diastereomers of bornylthio(cyclohexyl)-
methylphosphine borane by preparative HPLC.19 In our
stereoselective approach, (R)-cyclohexyl(hydroxylmethyl)-
B
DOI: 10.1021/acs.organomet.7b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Scheme 5. Preparation of Dibromocarbonyl Complex 2a
Scheme 6. Preparation of Bromocarbonylhydride 3a
determined by 31P−1H correlation spectroscopy). We for-
mulate this species as [FeHBr(CO)(1a)] (3a), in which H is
trans to Br (Scheme 6) in analogy to Milstein’s [FeHBr(CO)-
(1b)] (3b).11 A minor 31P NMR AX system (δ = 72.9 and 65.0,
2JP,P′ = 159.9 Hz, 15%) that correlates to a 1H NMR signal at δ
−22.0 (dd, 2JP,H = 55.7, 49.6 Hz) is assigned to the isomer of 3a
with hydride trans to the pyridine nitrogen, as a hydride ligand
diethyl ether, and pentane. The 31P{1H} NMR spectrum of 2a
shows a singlet at δ 59.6, indicating a C2-symmetric complex
and hence mutually trans-bromido ligands. A trace amount of
its cis-isomer is also present (δ 52.2 and 42.9, 2JP,P′ = 203.3 Hz,
<2%).
Complex 2a is perfectly stable toward air and moisture both
in solution and in the solid state. No decomposition was
observed after storing solid samples thereof in air at room
temperature for several months. In solution, 2a is stable for at
least 2 weeks without precautions to exclude air or moisture.
The IR spectrum shows a CO band at 1945 cm−1 (2a), which
indicates that the π-basicity of the iron(II) center in 2a is very
similar to that in the achiral [FeBr2(CO)(1b)] analogue 2b
(νCO = 1944 cm−1).11 X-ray quality crystals of 2a were obtained
by slow diffusion of pentane into a THF solution. Complex 2a
displays a slightly distorted octahedral geometry with mutually
trans-bromides (Figure 1).
1
trans to CO would give a H NMR signal at a much higher
frequency.25 Finally, a 31P NMR singlet at δ 82.9 (12%)
probably corresponds to a dicarbonyl Fe(0) complex, which has
already been observed as a decomposition product in the case
of 3b.11 The presence of the latter complex as an impurity may
explain the poor elemental analysis of 3a. Attempts to further
purify the iron hydride 3a failed as it slowly decomposes in
solution. Upon prolonged (>1 month) storage of a C6D6
solution of 3a, some decomposition products in the form of
a brown precipitate are observed. The signals of both hydride
complexes have disappeared, and the 31P NMR spectrum shows
only the signals of the putative dicarbonyl complex (at δ 83.4)
and of dibromocarbonyl 2a.
The CO stretch of 3a (1902 cm−1) is red-shifted with respect
to dibromocarbonyl 2a (1945 cm−1), as expected for the more
electron-rich hydride complex, in which π-backbonding to CO
is increased. The CO band of 3a also displays a shoulder at
1850 cm−1, which might be assigned to the νFe−H stretching or,
alternatively, to the putative dicarbonyl impurity. Hydride 3a is
stable when stored as solid under argon at −20 °C in the
absence of light but is highly air-sensitive in solution. Exposure
of CD2Cl2 solutions to air leads to an immediate color change
from green to yellow, and a brown, unidentified precipitate is
formed.
Figure 1. ORTEP drawing of 2a (ellipsoids at 30% probability).
Selected bond lengths (Å) and (torsion) angles (deg): Fe−P(1),
2.2313(8); Fe−P(2), 2.2414(8); Fe−N(1), 2.042(2); Fe−C(9),
1.744(3); Fe−Br(1), 2.4533(5); Br(1)−Fe−P(1), 86.68(2); Br(1)−
Fe−P(2), 91.67(3); P(1)−Fe−N(1)−C(2), 14.3(2); P(2)−Fe−
N(1)−C(3), 16.2(2).
[FeH2(CO)(1a)]. Dihydride 4a was prepared by treating
dibromocarbonyl 2a with NaHBEt3 (2.1 equiv) in THF
(Scheme 7). The addition of the hydride source was
Scheme 7. Preparation of Carbonyldihydride 4a
The most striking difference between the chiral 2a and
achiral [FeBr2(CO)(1b)] (2b) are the Fe−P bond lengths,
which are considerably shorter in 2a (2.2313(8), 2.2414(8) Å)
than in 2b (2.3020(14), 2.2799(14) Å).11 We attribute this to
the lower steric bulk of cyclohexylmethylphosphine in
comparison with that of diisopropylphosphine. In 2b, the
repulsion between the isopropyl substituent and bromide leads
to an obtuse Br(1)−Fe−P(2) angle of 94.35(4)°. This
distortion is less pronounced in 2a (91.67(3)°).
accompanied by an immediate color change from deep blue
to brownish-yellow. After evaporation of the volatiles, the
complex was extracted with diethyl ether to give a yellow
solution, which was filtered and evaporated to dryness.
The resulting brown-yellow product contained a mixture of
products. The major hydride species (93% of total intensity)
shows a characteristic signal at δ −7.65 (t, 2JP,H = 41.5 Hz) and
a broad 31P NMR triplet at δ 86.4 (2JP,H = 41.5 Hz), as
confirmed by 31P−1H correlation spectroscopy, and is
formulated as trans-[FeH2(CO)(1a)] (4a) based on the C2-
symmetry of the complex. Accordingly, trans-[FeH2(CO)(1b)]
[FeHBr(CO)(PNP)]. Bromohydridocarbonyl complex 3a was
obtained by treatment of dibromocarbonyl 2a with NaBHEt3
(1 equiv) in THF at room temperature for 4 h (Scheme 6).
Evaporation of the solvent and washing with pentane gave a
brown solid that was dried, stored under argon at −20 °C, and
characterized by NMR and IR spectroscopy. The 31P{1H}
NMR spectrum of the reaction solution of 2a shows a mixture
of products. The major one is a hydride complex featuring a 1H
2
NMR signal at δ −20.95 (dd, JP,H = 56.6, 50.9 Hz) and a 31P
2
NMR AX system at δ 76.1 and 64.9 (d, JP,P′ = 139.0 Hz) (as
C
DOI: 10.1021/acs.organomet.7b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
(4b) has a hydride chemical shift of δ −7.27.15 A minor,
a
Table 1. Optimization of the Hydrogenation of 10
1
unidentified, C1-symmetric hydride (7%) gives an apparent H
b
b
entry cat.
S/C/B
solvent p(H2) (bar) conv. (%) ee (%)
NMR triplet at δ −9.91 (2JP,H = 49.6 Hz) and an AX system in
the 31P{1H} NMR spectrum at δ 80.0 and 69.6 (2JP,P′ = 139.2
Hz). Attempts to obtain a pure product failed, as 4a is soluble
even in alkanes and slowly decomposes in solution.
1
2
3
4
5
6
7
8
9
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
4a
4a
100/1/5
100/1/5
100/1/5
200/1/5
1000/1/5
100/1/5
100/1/5
100/1/5
100/1/5
100/1/5
100/1/5
100/1/0
100/1/5
EtOH
EtOH
EtOH
EtOH
EtOH
EtOH
MeOH
nPrOH
nBuOH
THF
5
20
50
50
50
50
50
50
50
50
50
50
50
6
nd
48
48
48
49
46
nd
48
48
65
quant.
22
Addition of a stoichiometric amount of acetophenone to a
C6D6 solution of this mixture did not lead to a change in the
15
c
1
quant.
4
hydride region of the H NMR spectrum, with the species at δ
= −7.65 still present as the major signal after 24 h. In the
31P{1H} NMR spectrum, trace amounts (<3%) of a new AX
54
2
53
system at δ = 84.2 and 74.9 (d, JP,P′ = 30.4 Hz) appeared.
31P−1H correlation spectroscopy shows that this species is not a
hydride complex. After hydrolysis, GC-MS analysis revealed
that trace amounts (<2%) of phenylethanol were formed, which
may originate from an impurity in the dihydride complex,
though, because the major trans-dihydride does not react. We
conclude that dihydride 4a is not a reactive intermediate in
catalysisas previously observed by Milstein for its achiral
analogue [FeH2(CO)(1b)] (4b).15
10
11
12
13
0
d
e
EtOH
EtOH
EtOH
quant.
7
47
nd
42
99
a
Standard procedure: catalyst and KOtBu (amounts given as S/C/B)
were dissolved in ethanol (3 mL); acetophenone (50 mg, 48.5 μL,
0.416 mmol) was added, and the mixture was pressurized under H2 for
16 h at 20 °C. Determined by GC on a β-dex column. The hydride
complex was generated in situ by dissolving the corresponding
dibromocarbonyl complex in THF (0.3 mL) and adding NaBHEt3 (1
b
c
When the isomeric mixture of dihydride 4a was dissolved in
ethanol-d6, a red solution was obtained, whose 1H NMR
spectrum showed no hydride resonance. Upon addition of
KOtBu (1 equiv), the color of the solution changed
immediately to yellow, but no hydride signal was visible by
1H NMR spectroscopy. Therefore, in both cases, putative
hydride-containing species are either absent (due to decom-
position) or, less probably, paramagnetic.
Asymmetric Hydrogenation of Acetophenone. The
reaction conditions in the direct hydrogenation of acetophe-
none (10) to 1-phenylethanol (11) were optimized for catalyst
3a. Starting from similar reaction conditions (EtOH, 20 °C, 5
bar H2) as used for the achiral analogue [FeHBr(CO)(1b)]
(3b) (Scheme 8),11 quantitative acetophenone conversion was
d
e
equiv). PMe3 (30 mol % vs catalyst) was added. No base was added.
of 3a is currently underway in our laboratory and will be
published in due course.
As Yang16 and Hopmann17 have suggested that dihydride 4b
is the active species in ketone hydrogenation, we tested its
chiral analogue 4a as catalyst. No base was added to check
whether 4a is active per se (the function of the base is to
activate the bromocarbonylhydrido complex 3a). With other
conditions remaining the same, the conversion of acetophe-
none was only 7% (entry 12). Addition of base restored
catalytic activity and gave slightly lower enantioselectivity
(entry 13).
Mechanistic Background. As mentioned in the Introduc-
tion, the mechanism of the direct hydrogenation of
acetophenone with the achiral precatalyst [FeHBr(CO(1b)]
(3b) in the presence of base has been thoroughly studied by
DFT calculations.15−17 As acetophenone is prostereogenic, the
use of a chiral pincer ligand introduces additional information
concerning the enantiodetermining step of the catalytic cycle,
that is, the hydride transfer.26−28 This may allow one to
discriminate between the D, I, and O mechanisms (Scheme 9)
suggested for 3b, provided that they lead to significantly
different enantioselectivity and/or different sense of induction.
To clarify whether such a discrimination based on the absolute
configuration of the product is possible, we performed DFT
calculations on the enantiodetermining step of the hydro-
genation of acetophenone with complex 3a according to the
three different mechanisms (Scheme 9). Before exposing our
results, we summarize the main features of mechanisms D, I,
and O for the achiral catalyst 3b.
Scheme 8. Asymmetric Hydrogenation of 10 with 3a
achieved at 50 bar H2 (Table 1, entry 3). Catalyst loadings
lower than 1 mol % (entries 4 and 5) led to lower conversion,
in contrast to Milstein’s achiral analogue [FeHBr(CO)(1b)]
(3b), which operates at catalyst loadings as low as 0.05 mol
%.11 As observed with 3b,11 ethanol is the solvent of choice for
n
n
the reaction. Catalyst 3a is less active in PrOH and BuOH
(with no effect on the enantioselectivity, entries 8 and 9),
poorly active in MeOH (entry 7), and inactive in THF (entry
10).
Generating the hydride complex 3a in situ from 2a gave
results almost identical to those of the isolated complex (entry
3 vs 6), which is more convenient in view of the easy handling
of the dibromocarbonyl precursors. Under optimized con-
ditions, the isolated hydride 3a gave quantitative conversion
and 48% ee under 50 bar of H2 pressure (entry 3). Addition of
PMe3 as a poisoning agent for Fe nanoparticles4c to a catalytic
run with 3a did not alter the reaction outcome, which indicates
that the catalyst remains homogeneous under reaction
conditions (entry 11). An investigation of the substrate scope
Mechanism D was initially suggested by Yang16 and is based
on the dihydride complex [FeH2(CO)(1b)] (4b) (Scheme 1),
which transfers a hydride to a noncoordinated acetophenone
molecule. Hopmann has recently proposed a revised version
that differs for the regeneration of the catalyst but not for the
hydride transfer.17 However, Milstein has shown that this
dihydride does not react with acetophenone in a stoichiometric
fashion.15 As these findings strongly disfavor the involvement of
dihydride in the catalytic cycle, Milstein considered alternative
mechanisms. The classical inner-sphere Schrock−Osborne
D
DOI: 10.1021/acs.organomet.7b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
C(2)−C(1) and P(2)−Fe−N(1)−C(3)−C(4) assume an
envelope conformation with the P atom in the endo position
and equatorial cyclohexyl groups. The conformational inversion
of the ring swaps the cyclohexyl and methyl P-substituents
between equatorial and axial positions and tilts the pyridine ring
with respect to the P(1)−Fe−N(1) and P(2)−Fe−N(1)
planes. Additionally, a five-membered chelate ring can also
assume a planar conformation, which is energetically disfavored
for cyclopentane.38 The conformation of the five-membered
chelate rings in the iron-bound PNP pincer ligand is described
by the torsion angles P(1)−Fe−N(1)−C(2) (θ1) and P(2)−
Fe−N(1)−C(3) (θ2) (Figure 2).
Scheme 9. Enantiodetermining Step in Mechanisms D, I,
and O
mechanism I29−31 in which the CO double bond of the
coordinated ketone inserts into the Fe−H bond of mono-
hydride 13 gave an unrealistically large energetic span (50.8
kcal mol−1).15 Therefore, an outer-sphere mechanism O was
suggested that involves the Fe(0) complex 15, from which a
benzylic H atom of the PNP ligand is transferred as hydride to
the carbonyl group of acetophenone. A hydrogen bond
involving the coordinated ethanol molecule directs the
incoming substrate and activates its carbonyl group toward
nucleophilic attack,11,15 which is reminiscent of the well-
established bifunctional mechanism.31,32 Based on the reason-
able calculated energy span for O (21.7/24.5 kcal mol−1) and
on the lack of reactivity of dihydride [FeH2(CO)(1b)] toward
acetophenone, Milstein has concluded that O is the most
probable mechanism.15
DFT Studies. The transition state (TS) structures and their
relative free energies for mechanisms D, I, and O with the chiral
pincer complex 3a (Scheme 9) were studied by DFT. The TS
structures previously optimized by Hopmann17 for mechanism
D and by Milstein15 for mechanisms I and O were modified by
introducing the cyclohexyl and methyl substituents on the P
atoms. The phosphine fragments were taken from the X-ray
structure of 2a. Then, the structures were optimized for each
TS (see Conformational Issues below and Supporting
Information). The TSs and minima were calculated with the
hybrid density functional B3LYP33 with Grimme’s D3BJ
empirical correction for dispersion.34 The SDD basis set with
the associated Stuttgart-Dresden pseudopotential was used for
Fe35 and Dunning’s basis set36 with double-ζ for all other
atoms. The solvent was described by the implicit polarizable
continuum model.37 Stationary points were characterized by
vibrational analysis (only real frequencies for minima, one
imaginary frequency for transition states), and intrinsic reaction
coordinate (IRC) calculations were carried out on the
transition states in order to confirm their correct identification.
Computed harmonic frequencies were used to calculate the
thermal contribution to Gibbs free energy at 298 K.
Figure 2. Conformation of the five-membered chelate ring (with
Newman projection and dihedral angle θ).
The conformation of the chelate ring determines (a) whether
the P-substituents are equatorial or axial in the envelope
conformation, or inclinal in the planar one,39 and (b) the tilt of
the pyridine ring with respect to the P(1)−Fe−N(1) and
P(2)−Fe−N(1) planes. In the dibromocarbonyl derivative 2a
(Figure 1), the chelate rings assume envelope conformations
(θ1 = 16°, θ2 = 14°) with the P atom in the endo position.
Positive θ values in the range of 12 to 27° give an envelope
conformation in which the cyclohexyl group is equatorial. For
negative θ values (−12 to −27°), the cyclohexyl group is axial.
Dihedral angles θ close to 0° correspond to an inclinal
cyclohexyl group. Largely different θ1 and θ2 angles tend to
distort the P(1)−Fe−P(2) angle away from the ideal value of
180°. Therefore, the conformation of the chelate rings in the
TSs D, I, and O (Scheme 9) was optimized by scanning θ1 = θ2
between −35 and 35° at 10° intervals (for a total of eight
conformations). The resulting TSs are discussed below for the
different mechanisms.
Dihydride Mechanism D. Mechanism D involves hydride
transfer from dihydride complex 4a to the noncoordinated
acetophenone (Scheme 9).16,17 The corresponding TSs (Figure
3) were modeled and optimized as described below.
The enantioface exposed to the hydride in the TS determines
the absolute configuration of the alcohol, which is known from
experiment. Overall, the structure of the transition states
depends on (a) the enantioface of acetophenone, (b) the
orientation of the substrate, and (c) the conformation of the
pincer ligand. Initially, we studied the orientation of the
substrate with the CO bond of acetophenone parallel to the
plane of the complex with a planar conformation of the pincer
ligand.40 The P(1)−Fe−CO angle was scanned at 60°
Conformational Issues. The X-ray structure of 2a (Figure
1) shows that the five-membered rings P(1)−Fe−N(1)−
E
DOI: 10.1021/acs.organomet.7b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 3. Transition states for the hydride transfer in mechanism D,
torsion angles θ1 and θ2, and relative Gibbs free energies (in kcal
mol−1). Absolute configuration of 1-phenylethanol in brackets.
intervals (6 TSs) for the R and S enantiofaces of the
acetophenone. For each enantioface, the resulting starting
structures either did not converge or gave a single TS in which
the oxygen atom of acetophenone is involved in two C−H···O
interactions41 with one of the two benzylic H atoms of the
ligand and with the H atom on the ipso carbon of the adjacent
methyl or cyclohexyl groups on the phosphine, in agreement
with Hopmann’s TSs.17 In D(1), the O atom of the carbonyl
group is frozen between these two H atoms (O···H distances
are 2.22 and 2.20 Å). Hence, for each enantioface, only one
orientation of the substrate is possible (P(1)−Fe−CO =
−17° in D1, −155° in D2, Figure 3).
The C−H···O interaction between a benzylic H atom and
the CO group forces the chelate rings to assume planar (D(1),
D(2), D(4)) or partially flattened conformations (D(3) and
D(5), Figure 3), which are less stable than the envelope. In
particular, the PNP pincer in D(3) (θ1 = 24°, θ2 = 14°) and
D(5) (θ1 = −10°, θ2 = −21°) exhibits a highly distorted overall
conformation.
In the low-energy TSs D(1) and D(2), the phenyl ring of the
acetophenone points toward the pyridine backbone. A closer
analysis reveals that the phenyl group of the acetophenone
forms a C−H···π interaction with the benzylic H atom (2.56 Å
in D(1), 2.58 Å in D(2)). The energetically lowest TS D(1)
predicts the formation of (R)-1-phenylethanol, which is against
experiment and strongly disfavors this mechanism. The next TS
lies 0.9 kcal mol−1 higher in energy and would give the S
enantiomer.
Inner-Sphere Mechanism I. For the achiral PNP pincer
complex 3b, Milstein has suggested that mechanism I involves a
hydride ketone complex analogous to 13 (Scheme 9 and Figure
4). Intermediate 13, which features a deprotonated benzylic
position and dearomatized pyridine ring,11,42 evolves toward
the insertion of the CO double bond into the Fe−H one
Figure 4. Transition states for the hydride transfer in mechanism I.
The product stereochemistry is given in brackets. Relative Gibbs free
energies are expressed in kcal mol−1. Deprotonation sites and θ1 (left)
and θ2 (right) values are given in the Newman projections.
according to a classical Schrock−Osborn mechanism.31 The
structure of the resulting transition states is determined by (a)
the enantioface of acetophenone exposed to the catalyst, (b)
the position of the deprotonated benzylic carbon, and (c) the
conformation of the chelate rings (Figure 4).
The enantioface exposed in the TS determines the absolute
configuration of the resulting alcohol, which is the experimental
parameter. The relative orientation of the substrate is
constrained by the coordination of the O atom to the metal
center and by the interaction with the hydride. Concerning (b),
either −CH2− benzylic unit of the pyridine backbone can be
deprotonated, as indicated by the Newman projections in
Figure 4. As for point (c), the conformation of the chelate ring
determines whether the cyclohexyl and methyl substituents on
phosphorus are either axial or equatorial. Eight different
conformations were scanned as explained above. The
F
DOI: 10.1021/acs.organomet.7b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
optimization of these overall 32 structures converged to the
eight TSs shown in Figure 4.
The four energetically lower optimized transition states
I(1)−I(4) have envelope conformations with axial cyclohexyl
substituents (the dihedral angles θ1 and θ2 are negative in the
range between −23 and −17°). Transition state I(5) features
an envelope conformation for the P(1)−Fe−N(1)−C(1)−
C(2) ring (θ1 = 12°), whereas the P(2)−Fe−N(1)−C(3)−
C(4) ring is flattened with θ2 close to zero, and the cyclohexyl
group is inclinal. In I(6), both chelate rings are planar (both θ1
and θ2 close to 0°) with inclinal cyclohexyl substituents. The
last two transition states (I7−I8) show a positive value of the
two dihedral angles between 13 and 24°, typical of the envelope
conformation with equatorial cyclohexyl substituents.
Rather surprisingly, the four lowest-lying transition states
I(1)−I(4) have at least one axial cyclohexyl group. A closer
analysis reveals that the phenyl group of the acetophenone
forms a C−H···π interaction with the H atom on the ipso
carbon of the adjacent P-Cy group (2.34 Å in I(1), 2.32 Å in
I(2)). The P-Me group is involved in analogous but weaker
interactions in I(3) and I(4) (2.57 and 2.50 Å, respectively).
Eventually, the two lowest in energy transition states of the
inner-sphere mechanism (I(1) and I(2)) predict the formation
of (R)-1-phenylethanol, in contrast with the experimental
results (48% ee (S)).
Outer-Sphere Mechanism O. In the enantiodetermining
step of Milstein’s unprecedented outer-sphere mechanism O,15
a benzylic H atom of the ligand is transferred as hydride from
the five-coordinate iron(0) complex [Fe(CO)(EtOH)(1a)]
(15) to acetophenone (Scheme 9). Complex 15 exists as
conformers 15a and 15b. Negative θ1 and θ2 torsion angles are
indicative of axial cyclohexyl substituents such as in 15b, O(1),
and O(3). Positive θ torsion angles imply axial P-Me groups
such as in 15a, O(2), and O(4) (Figure 5). The approach of
acetophenone to 15a/15b is directed by the incipient CO···
Fe interaction and by the CO···HOEt hydrogen bond to
coordinated ethanol. Two parameters determine the structure
of the possible TSs, that is, (a) the prostereogenic face of
acetophenone and (b) the benzylic position involved in the
hydride transfer. The latter correlates to the conformation of
the chelate rings: The benzylic H atom on the left points
toward acetophenone only for θ > 0° (that is, for equatorial
cyclohexyl groups, Figure 5 right).
Figure 5. Iron(0) complexes 15a and 15b and transition states O(1)−
O(4) for the hydride transfer in mechanism O with product
stereochemistry in brackets. Newman projections give θ1 (left) and
θ2 (right). Relative Gibbs free energies are expressed in kcal mol−1.
bears equatorial cyclohexyl substituents, is more stable than
15b by 0.4 kcal mol−1 (Figure 6).
The conformation of structures involving hydride transfer
from the left benzylic methylene was optimized by scanning θ
between 35 and 5° (see above) with both acetophenone
enantiofaces. This optimization converged to the TSs (O(2)
and O(4) in Figure 5). The same procedure was repeated with
θ < 0° for transfer from the benzylic position on the right,
which gave O(1) and O(3). As acetophenone approaches 15a/
15b, the conformation of the chelate ring changes to allow the
attack of the benzylic H atom as hydride onto the carbonyl C
atom, as indicated by the small, but significant, increase in |θ|
values (Figure 5).
An intriguing result is that the most stable TS O(1) features
axial cyclohexyl groups and hence the less stable conformation.
To find out how the conformation of 15a/15b correlates with
that of O(1)−O(4), we performed an IRC analysis. The
calculated reaction coordinates were found to connect O(2)
and O(4) to 15a, which has the same conformation (equatorial
cyclohexyl groups) as the starting species (Figure 5, right).
Analogously, O(1) and O(3) are linked to 15b, both bearing
axial cyclohexyl groups (Figure 5, left). Conformer 15a, which
Figure 6. Computed reaction pathway for the hydride transfer in
mechanism O. Relative Gibbs free energies are expressed in kcal
mol−1. The Gibbs free energies of 16 are not in scale.
G
DOI: 10.1021/acs.organomet.7b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
No significant conformational changes were detected
between 15a/15b and the corresponding transition states.
Therefore, we conclude that the inversion of the chelate rings
occurs between the starting species 15a and 15b. We were
unable to locate a transition state for their interconversion,
which suggests that the energy hypersurface for the chelate ring
flip is flat without large barriers. Accordingly, the pseudor-
otation of the five-membered ring in cyclopentane through the
half-chair conformation has a time scale on the order of
picoseconds or sub-picoseconds at room temperature.43 For
these reasons, we assume that the equilibrium between 15a and
15b is fast, and that they react under Curtin−Hammett
conditions to give the 1-phenylethoxide complexes 16a/16a′
and 16b/16b′ (Scheme 9). This implies that the ratio of the
products depends on the energy difference between the
transition states.44,45
The above conclusion concerning the conformational
equilibria is corroborated by a closer analysis of the four TSs.
In O(1), the benzylic C−H bond to the hydrogen atom that is
being transferred as hydride is very weakly perturbed (1.23 Å),
whereas the distance to the carbonyl C atom is as long as 1.68
Å. Both observations suggest an early transition state, in
agreement with the Hammond−Leffler postulate for exergonic
reactions such as hydride transfer.46 Also, there is only little
structural reorganization between 15 and O(1)−O(4), as
indicated by the minor changes in |θ| (Figure 5), which further
supports the retention of conformation along the reaction
coordinate.
Perusal of TSs O(1)−O(4) in Figure 5 suggests that their
energy depends on the steric interactions of the phenyl ring of
acetophenone (a) with the pincer backbone and (b) with the
substituents on phosphorus, which depends on the con-
formation of the chelate rings. In the energetically low-lying
TSs O(1) and O(2), acetophenone directs the methyl group
above the pyridine and phenyl in TSs O(3) and O(4). For the
P-substituents, acetophenone places the phenyl group toward
the small P-Me group in O(1) and toward the large P-Cy group
in O(2), which is in qualitative agreement with O(1) being
more stable. However, Figure 5 shows that there is no clear and
easily predictable trend in the interactions between acetophe-
none the P-substituents in the case of O(3) and, in particular,
of O(4), whose high energy possibly mainly derives from the
conformational strain (θ = 27°; see below).
Figure 7. Complexes before (16b) and after (17b) the proton transfer
in mechanism O.
located on the energy hypersurface, the proton is equidistant
from the oxygen atoms of ethoxide (1.22 Å) and 1-
phenylethanolate (1.20 Å). IRC analysis indicates that the
reaction is without barrier (Figure S24). As the energy of
TS(16b−17b) is much lower than that of O(1), we conclude
that only the preceding hydride transfer affects the
enantioselectivity. In 17b, the proton is located on the oxygen
atom of 1-phenylethanol (1.01 Å) and is involved in a hydrogen
bond with ethoxide (the O···H distance is 1.64 Å) (see
Supporting Information). Hence, at difference with the achiral
catalyst 3b, for which concerted H−/H+ transfer occurs, the
mechanism for 3a is stepwise with consecutive hydride and
proton transfers.15
Conformational Flexibility. Our DFT study highlights the
pivotal role played by the changes of the ligand conformation in
the different intermediates. Surprisingly, the conformation with
axial cyclohexyl groups in the Fe(0) intermediate 15a/15b is
just 0.4 kcal mol−1 less stable than with equatorial cyclohexyls
(Figure 6). To gain further insight into the flexibility of the
PNP ligand 1a, we applied the activation strain model (ASM)
to transition states O(1)−O(4).47 The breakdown of the
electronic energy ΔΔE⧧ indicates that the interaction of the
tot
catalyst with acetophenone in the transition state (ΔΔE⧧int) is
less important than the distortion of the two reactants during
their approach (ΔΔE⧧strain) (Table 2). Compared to O(1), a
a b
,
Table 2. ASM Analysis of Transition States O(1)−O(4)
O(1)[S]
O(2)[R]
O(3)[R]
O(4)[S]
ΔΔE⧧
ΔΔE⧧
ΔΔE⧧
ΔΔE⧧
ΔΔE⧧
|Δθ|
0
0
0.7
−1.4
2.1
1.6
−3.0
4.6
3.6
−1.4
5.0
tot
int
H−/H+ Transfer. Each TS O(1)−O(4) shows a strong
hydrogen bond between the H atom of the coordinated ethanol
O−H and the oxygen atom of acetophenone. In O(1), the C
O···HOEt distance is 1.50 Å. It has been shown previously15
that this hydrogen bond directs and activates the incoming
acetophenone toward hydride attack. In the case of mechanism
O, we have studied the hydride transfer step by IRC profile
analysis, which shows that TSs O(1)−O(4) evolve to four
diastereomeric intermediates (16a, 16a′, 16b, and 16b′; see
Supporting Information) that contain coordinated 1-phenyl-
ethoxide involved in a strong hydrogen bond to the iron-bound
ethanol. In 16b, which derives from the lowest-energy TS
O(1), the hydrogen atom is still closer to ethoxide (1.02 Å)
than to the oxygen atom of 1-phenylethanolate (1.58 Å), which
implies that the transfers of hydride and proton are not
concerted.
0
strain
0
−1.0
3.1
2.1
−0.2
5.2
strain/sub
strain/cat
0
2.5
2°
9°
5°
11°
a
b
Relative electronic energies are expressed in kcal mol−1. For each
TS, |Δθ| = |θ(O) − θ(15)| refers to the conformation of the chelate
ring involved in hydride transfer (see Figure 5).
significant deformation of acetophenone is observed only for
O(3). Thus, the major contribution to the strain energy
ΔΔE⧧
is the distortion of the catalyst, which is mainly due
strain
to the conformational change of the chelate ring involved in
hydride transfer (Figure 5) and can be expressed as the increase
of |θ| in the corresponding chelate (|Δθ|, Table 2). The data in
Table 2 show that TSs O(1) and O(3), which relate to the less
stable conformer 15b bearing axial cyclohexyl groups, undergo
less conformational rearrangement than O(2) and O(4), which
connect to the more stable 15a. Accordingly, O(1) and O(3)
Although the possibility that H+ transfer might be kinetically
relevant appears remote, we studied the proton transfer step
from ethanol to 1-phenylethanolate, which converts 16b to 17b
(Figure 7). In the corresponding TS(16b−17b), which was
H
DOI: 10.1021/acs.organomet.7b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
have the lowest ΔΔE⧧strain/cat values, which implies that 15b has
the best conformation for the approach of acetophenone.
Overall, the above DFT analysis suggests that the energy
difference between different chelate ring conformations is small.
In mechanism D, the most stable structures show a planar
chelate ring, whereas the envelope conformation with axial
cyclohexyl groups is the most stable for mechanism I. In
mechanism O, the most stable transition state has envelope
conformations with axial cyclohexyl groups, but the most stable
starting minimum has envelope conformations with equatorial
cyclohexyl groups, although the energies of the minima are very
close. We conclude that PNP ligand 1a is flexible and its
conformation can be easily swapped by weak intermolecular
interactions, such as the C−H···O bond in mechanism D and
the C−H···π interaction in mechanisms D and I. In the same
way, the steric hindrance significantly affects the relative
conformation stability, as seen in mechanism O. Therefore,
rigidifying the PNP backbone is the first priority in the attempt
to achieve higher enantioselectivity.
Comparison of Mechanisms D, I, and O. Overall, the
DFT study on the outer-sphere mechanism O indicates that the
lowest energy transition state O(1) gives the alcohol with the S
absolute configuration, which is in agreement with the
experimental observation. The TSs responsible for the minor
R enantiomer lie 0.6 and 0.9 kcal mol−1 higher than that leading
to the S product. The Boltzmann distribution of the four
transition states leads to 26% ee, which is in fair agreement with
the experimentally observed enantioselectivity of 48% ee. The
reaction profile for the reaction segment connecting the Fe(0)
ethanol complex 15 to the Fe(II) 1-phenylethanol species 17
confirms that the hydride transfer step is kinetically competent.
In contrast, the inner-sphere mechanism I predicts the wrong
sense of induction, as the two lowest energy transition states
lead to the R enantiomer instead of the experimentally observed
S one.
this result is that it is based on the experimentally observed
sense of induction and is hence complementary to Milstein’s
considerations based on the calculated energetic span of the
catalytic reaction.15 Finally, we note that the discrimination of
different mechanisms based on sense of induction and
enantioselectivity is still a rare approach48 and should be
added to the toolbox of mechanistic investigation.
CONCLUSION
■
The C2-symmetric enantiopure Fe-PNP dibromocarbonyl
complex described herein directly transfers the highly
successful, achiral pyridine-based Fe-PNP pincer chemistry to
a closely related chiral analogue. The slightly altered reactivity
and the structural differences in the solid state of the chiral
analogue highlight the importance of the substituents at
phosphorus in such systems.
Even more importantly, the DFT study on P-stereogenic 3a
discloses the pivotal role played by the conformation of the
pincer ligand in the mechanism of enantioselection. The
combination of stereogenic P donors with the 2,6-dimethyle-
nepyridine backbone gives highly flexible complexes, which
requires a thorough analysis and investigation of the conforma-
tional aspects. In particular, conformations in which the largest
substituents occupy axial positions are easily stabilized by
secondary interactions, which hampers simple, qualitative
predictions of the relative stability of the transition states. We
are currently investigating such aspects with related PNP
ligands.
The P-stereogenic pincer 1a gave a valuable contribution to
the mechanistic debate concerning the catalytic hydrogenation
of ketones with 2,6-dimethylenepyridine-based PNP pincer
ligands. Indeed, Milstein’s mechanism involving hydride
transfer from the benzylic carbon of the ligand was the only
one to reproduce the experimental sense of induction and
enantioselectivity.
Also dihydride mechanism D predicts the wrong R
enantiomer. Additionally, Milstein observed that dihydride
[FeH2(CO)(1b)] (4b) does not react with acetophenone.15
The chiral analogue [FeH2(CO)(1a)] (4a) gives, at best, traces
of 1-phenylethanol in the presence of acetophenone (1 equiv)
after 24 h, which strongly disfavors its involvement in catalysis.
We explain the residual activity (Table 1, entry 12) with the
presence of unidentified catalytically active impurities, as
suggested by the NMR spectra of 4a. Upon addition of base
(5 equiv), dihydride 4a catalyzes the hydrogenation of
acetophenone with similar conversion and enantioselectivity
as 3a. However, this does not lend support to dihydride
mechanism D. In fact, the addition of base most probably
triggers the deprotonation of the benzylic position of the PNP
ligand. The resulting negatively charged and hence electron-rich
dihydride might transfer hydride to acetophenone. The
resulting five-coordinate hydridocarbonyl complex might
undergo reductive elimination of H+, reprotonation of the
ligand, and EtOH coordination to form 15, which is the
catalytically active species of mechanism O (Scheme 9).
As a word of caution, though, we notice that our DFT studies
do not categorically exclude mechanism D and, in particular,
mechanism I, as the energy differences involved are small as
compared to the accuracy of DFT methods. Still, the combined
experimental and computational study on chiral catalyst 3a
supports Milstein’s conclusion with its achiral analogue 3b that
the outer-sphere mechanism (O) is more likely than the
dihydride (D) and inner-sphere one (I).15 The importance of
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.organo-
met.7b00816.
Experimental procedures and characterization, X-ray
crystallographic data for 2a, and computational details
(PDF)
Interactive structure (XYZ)
Accession Codes
CCDC 1584795 contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
via www.ccdc.cam.ac.uk/data_request/cif, or by emailing data_
request@ccdc.cam.ac.uk, or by contacting The Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge CB2
1EZ, UK; fax: +44 1223 336033.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: mezzetti@inorg.chem.ethz.ch.
ORCID
Antonio Mezzetti: 0000-0002-1824-7760
Notes
The authors declare no competing financial interest.
I
DOI: 10.1021/acs.organomet.7b00816
Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
127, 11934−11935. With 7, the benzoyl ester was a colorless oil at
room temperature.
(23) McKinstry, L.; Livinghouse, T. Tetrahedron 1995, 51, 7655−
7666.
ACKNOWLEDGMENTS
■
This work was supported by SNF Grant 2-77063-16 and by
ETH Grant 0-20356-17.
(24) The common deboronation reaction with amines proceeds
sluggishly with trialkyl phosphines. See: Lloyd-Jones, G. C.; Taylor, N.
P. Chem. - Eur. J. 2015, 21, 5423−5428.
REFERENCES
■
(1) Casey, C. P.; Guan, H. J. Am. Chem. Soc. 2007, 129, 5816−5817.
(2) (a) Junge, K.; Schroder, K.; Beller, M. Chem. Commun. 2011, 47,
(25) Gorgas, N.; Stoger, B.; Veiros, L. F.; Pittenauer, E.; Allmaier, G.;
̈
̈
Kirchner, K. Organometallics 2014, 33, 6905−6914.
(26) Dub, P. A.; Ikariya, T. J. Am. Chem. Soc. 2013, 135, 2604−2619.
(27) Li, L.; Pan, Y.; Lei, M. Catal. Sci. Technol. 2016, 6, 4450−4457.
(28) Sonnenberg, J. F.; Wan, K. Y.; Sues, P. E.; Morris, R. H. ACS
Catal. 2017, 7, 316−326.
4849−4859. (b) Bauer, I.; Knolker, H.-J. Chem. Rev. 2015, 115, 3170−
̈
3387. (c) Darcel, C.; Sortais, J. B. Top. Organomet. Chem. 2015, 50,
173−216. (d) Ollevier, T.; Keipour, H. Top. Organomet. Chem. 2015,
50, 259−309. (e) Furstner, A. ACS Cent. Sci. 2016, 2, 778−789.
̈
(3) Magano, J.; Dunetz, J. R. Org. Process Res. Dev. 2012, 16, 1156−
(29) Schrock, R. R.; Osborn, J. A. J. Chem. Soc. D 1970, 9, 567−568.
(30) Clapham, S. E.; Hadzovic, A.; Morris, R. H. Coord. Chem. Rev.
2004, 248, 2201−2237.
1184.
(4) Selected papers: (a) Sui-Seng, C.; Freutel, F.; Lough, A. J.;
Morris, R. H. Angew. Chem., Int. Ed. 2008, 47, 940−943. (b) Mikhai-
line, A.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2009, 131, 1394−
1395. (c) Sonnenberg, J. F.; Coombs, N.; Dube, P. A.; Morris, R. H. J.
Am. Chem. Soc. 2012, 134, 5893−5899. (d) Mikhailine, A. A.; Maishan,
M. I.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2012, 134, 12266−
12280. (e) Zuo, W. W.; Lough, A. J.; Li, Y. F.; Morris, R. H. Science
2013, 342, 1080−1083. (f) Zuo, W. W.; Prokopchuk, D. E.; Lough, A.
J.; Morris, R. H. ACS Catal. 2016, 6, 301−314.
(5) Li, Y. Y.; Yu, S. L.; Shen, W. Y.; Gao, J. X. Acc. Chem. Res. 2015,
48, 2587−2598.
(6) (a) Bigler, R.; Huber, R.; Mezzetti, A. Angew. Chem., Int. Ed.
2015, 54, 5171. (b) Bigler, R.; Mezzetti, A. Org. Process Res. Dev. 2016,
20, 253. (c) Bigler, R.; Huber, R.; Stockli, M.; Mezzetti, A. ACS Catal.
2016, 6, 6455. (d) De Luca, L.; Mezzetti, A. Angew. Chem., Int. Ed.
2017, 56, 11949−11953.
(7) Fe(II) complexes containing tetradentate ligands only give slow
H2 activation: Morris, R. H. Chem. Soc. Rev. 2009, 38, 2282−2291.
(8) (a) 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.
(b) Zirakzadeh, A.; Kirchner, K.; Roller, A.; Stoger, B.; Widhalm, M.;
Morris, R. H. Organometallics 2016, 35, 3781−3787. (c) Smith, S. A.
M.; Lagaditis, P. O.; Lupke, A.; Lough, A. J.; Morris, R. H. Chem. - Eur.
(31) Samec, J. S. M.; Backvall, J.-E.; Andersson, P. G.; Brandt, P.
̈
Chem. Soc. Rev. 2006, 35, 237−248.
(32) (a) Hashiguchi, S.; Fujii, A.; Takehara, J.; Ikariya, T.; Noyori, R.
J. Am. Chem. Soc. 1995, 117, 7562−3. (b) Haack, K.-J.; Hashiguchi, S.;
Fujii, A.; Ikariya, T.; Noyori, R. Angew. Chem., Int. Ed. Engl. 1997, 36,
285−288. (c) Noyori, R.; Yamakawa, M.; Hashiguchi, S. J. Org. Chem.
2001, 66, 7931−7944. (d) Ikariya, T.; Blacker, A. J. Acc. Chem. Res.
2007, 40, 1300−1308. (e) Dub, P. A.; Gordon, J. C. ACS Catal. 2017,
7, 6635−6655.
(33) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J.
Phys. Chem. 1994, 98, 11623−11627.
(34) (a) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys.
2010, 132, 154104−154119. (b) Grimme, S.; Ehrlich, S.; Goerigk, L. J.
Comput. Chem. 2011, 32, 1456−1465.
̈
(35) (a) Dolg, M.; Wedig, U.; Stoll, H.; Preuss, H. J. Chem. Phys.
1987, 86, 866−872. (b) Martin, J. M. L.; Sundermann, A. J. Chem.
Phys. 2001, 114, 3408−3420.
(36) Dunning, T. H., Jr. J. Chem. Phys. 1989, 90, 1007−1023.
(37) (a) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
2999−3093. (b) Cossi, M.; Scalmani, G.; Rega, N.; Barone, V. J. Chem.
Phys. 2002, 117, 43−54.
̈
̈
(38) Saebø, S.; Cordell, F. R.; Boggs, J. E. J. Mol. Struct.:
THEOCHEM 1983, 104, 221−232.
J. 2017, 23, 7212−7216.
(9) Grabulosa, A. P-Stereogenic Ligands in Enantioselective Catalysis;
RSC Catalysis Series No. 7; Royal Society of Chemistry: Cambridge,
2011.
(39) Cremer, D. Isr. J. Chem. 1980, 20, 12−19.
(40) This conformation was chosen according to Hopmann’s TS.17
(41) (a) Desiraju, G.; Steiner, T. The Weak Hydrogen Bond; Oxford
University Press: Oxford, 1999. (b) Steiner, T. Crystallogr. Rev. 2003,
9, 177−228.
(10) For a single example of a P-stereogenic pincer ligand used for
the hydrogenation of ketones with Ru(II), see: Arenas, I.; Boutureira,
O.; Matheu, M. I.; Díaz, Y.; Castillon
3666−3669.
́
, S. Eur. J. Org. Chem. 2015, 2015,
(42) (a) Ben-Ari, E.; Leitus, G.; Shimon, L. J. W.; Milstein, D. J. Am.
Chem. Soc. 2006, 128, 15390−15391. (b) Zhang, J.; Leitus, G.; Ben-
David, Y.; Milstein, D. J. Am. Chem. Soc. 2005, 127, 10840−10841.
(c) Milstein, D. Top. Catal. 2010, 53, 915−923.
(11) Langer, R.; Leitus, G.; Ben-David, Y.; Milstein, D. Angew. Chem.,
Int. Ed. 2011, 50, 2120−2124.
(12) Zell, T.; Ben-David, Y.; Milstein, D. Angew. Chem., Int. Ed. 2014,
53, 4685−4689.
(43) (a) Kilpatrick, J. E.; Pitzer, K. S.; Spitzer, R. J. Am. Chem. Soc.
1947, 69, 2483−2485. (b) Bauman, L. E.; Laane, J. J. Phys. Chem.
1988, 92, 1040−1051. (c) Cui, W.; Li, F.; Allinger, N. L. J. Am. Chem.
Soc. 1993, 115, 2943−2951. (d) Han, S. J.; Kang, Y. K. J. Mol. Struct.:
THEOCHEM 1996, 362, 243−255. (e) Lightner, D. A.; Gurst, J. E.
Organic Conformation Analysis and Stereochemistry from Circular
Dichroism Spectroscopy; Wiley-VCH: New York, 2000. (f) Anslyn, E.
V.; Dougherty, D. A. Modern Physical Organic Chemistry; University
Science Books: Herdon, VA, 2005. (g) Ocola, E. J.; Bauman, L. E.;
Laane, J. J. Phys. Chem. A 2011, 115, 6531−6542.
(13) Zell, T.; Ben-David, Y.; Milstein, D. Catal. Sci. Technol. 2015, 5,
822−826.
(14) Garg, J. A.; Chakraborty, S.; Ben-David, Y.; Milstein, D. Chem.
Commun. 2016, 52, 5285−5288.
(15) Langer, R.; Iron, M. A.; Konstantinovski, L.; Diskin-Posner, Y.;
Leitus, G.; Ben-David, Y.; Milstein, D. Chem. - Eur. J. 2012, 18, 7196−
7209.
(16) Yang, X. Z. Inorg. Chem. 2011, 50, 12836−12843.
(17) Morello, G. R.; Hopmann, K. H. ACS Catal. 2017, 7, 5847−
5855.
(44) Seeman, J. I. Chem. Rev. 1983, 83, 83−134.
(45) Curtin−Hammett conditions are assumed also for mechanisms
D and I with regard to the conformational equilibrium and for I with
respect to the protonation/deprotonation equilibrium at the benzylic
positions.
(18) Kozuch, S.; Shaik, S. Acc. Chem. Res. 2011, 44, 101−110.
(19) Miura, T.; Yamada, H.; Kikuchi, S.-I.; Imamoto, T. J. Org. Chem.
2000, 65, 1877−1880.
(20) Muci, A. R.; Campos, K. R.; Evans, D. A. J. Am. Chem. Soc. 1995,
117, 9075−9076.
(46) (a) Leffler, J. E. Science 1953, 117, 340−341. (b) Hammond, G.
J. Am. Chem. Soc. 1955, 77, 334−338.
(21) Nagata, K.; Matsukawa, S.; Imamoto, T. J. Org. Chem. 2000, 65,
4185−4188.
(47) Fernan
4953−4967.
́
dez, I.; Bickelhaupt, F. M. Chem. Soc. Rev. 2014, 43,
(22) The optical purity of (R)-tert-butyl(hydroxymethyl)methyl-
phosphine borane has been improved by crystallization of the benzoyl
ester. See: Imamoto, T.; Sugita, K.; Yoshida, K. J. Am. Chem. Soc. 2005,
(48) For a recent example, see: Santoro, S.; Liao, R.-Z.; Marcelli, T.;
Hammar, P.; Himo, F. J. Org. Chem. 2015, 80, 2649−2660.
J
DOI: 10.1021/acs.organomet.7b00816
Organometallics XXXX, XXX, XXX−XXX