Stereoelectronic factors in iron catalysis: Synthesis and characterization of aryl-substituted iron(II) carbonyl P-N-N-P complexes and their use in the asymmetric transfer hydrogenation of ketones

Article abstract of DOI:10.1021/om2005172

A series of five (S,S)-trans-[Fe(CO)(Br)(PR₂-CH ₂CH=NCH(Ph)CH(Ph)N=CHCH₂-PR₂)][X] compounds (1a-c, X = BPh₄; 1d,e, X = BF₄) were synthesized and tested for the asymmetric transfer hydrogena

Full text of DOI:10.1021/om2005172

ARTICLE
pubs.acs.org/Organometallics
Stereoelectronic Factors in Iron Catalysis: Synthesis and
Characterization of Aryl-Substituted Iron(II) Carbonyl PÀNÀNÀP
Complexes and Their Use in the Asymmetric Transfer Hydrogenation
of Ketones
Peter E. Sues, Alan J. Lough, and Robert H. Morris*
Department of Chemistry, University of Toronto, Toronto, Ontario M5S 3H6, Canada
S Supporting Information
b
ABSTRACT:
A series of five (S,S)-trans-[Fe(CO)(Br)(PR2-CH2CHdNCH(Ph)CH(Ph)NdCHCH2-PR2)][X] compounds (1aÀc, X = BPh4;
1d,e, X = BF4) were synthesized and tested for the asymmetric transfer hydrogenation (ATH) of acetophenone. Three of the
complexes had methyl-substituted aryl groups (a, R = para-CH3C6H4; b, R = ortho-CH3C6H4; c, R = 3,5-(CH3)2C6H3), and two
had trifluoromethyl-substituted aryl groups (d, R = para-CF3C6H4; e, R = 3,5-(CF3)2C6H3). Using both known and new
phosphonium dimers, [cyclo-(PR2CH2CH(OH)À)2][Br]2 (2aÀc), in a one-pot template reaction, the corresponding (S,S)-
trans-[Fe(CH3CN)2(PR2-CH2CHdNCH(Ph)CH(Ph)NdCHCH2-PR2)][BPh4]2 complexes (3aÀc) were generated and then
converted to precatalysts 1aÀc via CO addition reactions. While investigating compounds 1aÀc, an alternative route for
synthesizing phosphonium dimers was developed that allowed the facile introduction of tetrafluoroborate counterions. Compounds
1d and 1e could not be synthesized using previously developed methods; phosphinoacetaldehyde diethyl acetal precursors (5d, 5e)
were isolated because trifluoromethyl-substituted phosphonium dimers did not form. Precursors 5d and 5e were incompatible with
a base-catalyzed template approach, so a new acid-catalyzed template procedure was developed to generate the tetrafluoroborate
salts (S,S)-trans-[Fe(CH3CN)2(PR2-CH2CHdNCH(Ph)CH(Ph)NdCHCH2ÀPR2)][BF4]2 (3d, 3e). Both 3d and 3e were
converted to precatalysts 1d and 1e via CO addition reactions. Complexes 1b, 1d, and 1e were inactive for the ATH
of acetophenone, while complexes 1a and 1c were active. Compound 1a showed very high activity, with a turnover frequency of
30 000 hÀ1 at 28 °C, and is currently the most active iron ATH catalyst. Compound 1c produced more enantiopure (R)-1-
phenylethanol, with an ee of 90%, and is the most selective iron catalyst reported to date for the ATH of acetophenone. The activity of
complexes 1aÀe for ATH was compared to those of known complexes 1f (R = Ph), 1g (R = Et), 1h (R = i-Pr), and 1i (R = Cy), and the
most active catalysts were defined by a narrow range of electronic (ν(CO)) as well as steric (Tolman cone angles) parameters.
’ INTRODUCTION
transformations, and many well-defined iron catalysts have been
reported that have applications in commercially relevant reductions
of polar bonds.12,13 Several promising hydrosilylation,14,15 direct
hydrogenation,16,17 and transfer hydrogenation (TH)18À23 sys-
tems based on iron have been developed for the reduction of
ketones, as well as imines.24,25
Economic and political pressure for more sustainable, envir-
onmentally friendly practices has provided a driving force for the
development of greener catalysts that incorporate cheap, abun-
dant metals. Iron is an attractive candidate for replacing platinum
group metals for these reasons; it is inexpensive, plentiful, and
environmentally benign.1 In fact, iron is an essential element that
is found in the enzymes of every living system. Inspired by nature,
many oxidation catalysts based on iron2À7 were developed in
the past, several of which employed macrocyclic ligands.8À11 More
recently, however, research focus has shifted toward reductive
The asymmetric transfer hydrogenation (ATH) of prochiral
ketones is a valuable transformation because it allows the
Received: June 16, 2011
Published: July 20, 2011
r
2011 American Chemical Society
4418
dx.doi.org/10.1021/om2005172 Organometallics 2011, 30, 4418–4431
|

Organometallics
ARTICLE
the complexes through their phosphorus donors while maintain-
ing an S,S-dpen backbone, we endeavored to develop more active
and selective catalysts. In addition, by comparing the catalytic
activities of the electron-donating and -withdrawing groups it was
anticipated we would gain some insight into the ATH mechanism.
With respect to nomenclature, throughout this paper compounds
will be named according to their aryl substitution pattern: a for
para-tolyl; b for ortho-tolyl; c for meta-xylyl, d for para-trifluoro-
methylphenyl, and e for 3,5-ditrifluoromethylphenyl.
Figure 1. Iron(II) carbonyl PÀNÀNÀP complexes developed by
’ RESULTS AND DISCUSSION
Morris and co-workers for ATH.
Conventional Synthesis (Route A) of Phosphonium Di-
mers 2aÀc. The new dimeric phosphonium salts 2b and 2c were
synthesized according to a known literature procedure
(Scheme 1, route A), which has already been successfully
employed to generate 2a.42 Much like 2a, the synthesis of 2b
and 2c yielded air-stable, white solids in high yields (80À90%).
The 31P{1H} NMR spectrum of 2c showed two characteristic
singlets: a major peak at 12.9 ppm and a minor peak at 14.8 ppm.
These peaks indicate that both the rac- and meso-diastereomers
of this compound are present.
Dimer 2b, on the other hand, showed a dramatically different
behavior in solution. The 1H NMR spectrum of 2b, in DMSO-d6,
showed three species in solution: an aldehyde (doublet of triplets
at 9.6 ppm) and an enol (two doublets of doublets in a 1:1
integration ratio at 6.9 and 5.3 ppm) in a 1:2 ratio, as well as a very
broad singlet around 12.3 ppm. The 31P{1H} NMR spectrum
showed two sharp singlets at 25.1 (enol) and 28.1 ppm -
(aldehyde), indicating that both the enol and aldehyde phosphori
were protonated and that interconversion of the two tautomers
occurred relatively slowly. No coupling was observed, however,
between phosphorus and hydrogen in the 31P NMR spectrum.
This can be explained if the protonated phosphori are in
equilibrium with protonated DMSO and the rate of exchange
between the two is similar or greater than the frequency of the
coupling between the phosphori and hydrogen (approximately
500 Hz). This exchange process would also account for the broad
singlet seen in the 1H NMR spectrum, which corresponds to acid
in solution. These findings are summarized in Scheme 2. Con-
trary to these findings, an IR spectrum of dimer 2b (KBr disk) did
not exhibit any characteristic absorptions for aldehydes or
protonated phosphonium salts.
In another attempt to characterize 2b in solution, additional
NMR spectra of the compound were run in MeOH-d4, which has
a lower dielectric constant than DMSO. The spectra obtained
were quite different than those run previously in DMSO-d6. The
31P{1H} NMR spectrum revealed a very broad peak at À21.0 ppm,
which suggests the presence of a three-coordinate phosphine in
solution undergoing a dynamic process on a time scale near that
of the NMR experiment. We believe this is a fast exchange of
protons between the phosphorus and the solvent. In addition,
both the 13C{1H} and 1H NMR spectra of 2b in CD3OD were
indicative of a hemiacetal, and the high-resolution ESI+ MS
(MeOH) spectrum showed a peak at 289.1 m/z for the proto-
nated hemiacetal [C19H20O2P]+.
Further contradictions arose when examining the elemental
analysis (EA) of 2b, which was consistent with that of a dimer,
and the high-resolution DART-MS (direct analysis in real time
mass spectrometry) spectrum, which also gave evidence for a
dimer with a peak at 257.1 m/z for the doubly charged species
[C32H36O2P2]2+. On the basis of the EA and DART-MS results,
synthesis of enantiopure alcohols for use in the pharmaceutical,
fragrance, and food industries.26,27 This field is dominated by
systems that employ expensive as well as toxic platinum group
metals,28 such as the ruthenium and iridium ATH catalysts
reported by Noyori29 and Gao.30,31 A common theme of these
catalysts is that they employ ligands with two phosphorus donors
and two nitrogen donors.29 A particularly attractive ligand used
by Gao and Noyori is a chiral tetradentate ligand, which
coordinates via two phosphorus donors and two nitrogen donors
(a PÀNÀNÀP ligand).32
To date, the only iron catalysts that challenge platinum metal
ATH catalysts on both activity and selectivity were developed
by the Morris group.33,34 These complexes utilize chiral
PÀNÀNÀP ligands different from those of Noyori’s ruthenium
catalysts in that they contain two imines and have only one
methylene linker between the phosphorus and imine function-
alities (see Figure 1).34,35 Complex 1f in particular is exceptional;
when activated with base, it is the most active and selective iron
ketone ATH catalyst reported thus far, with turnover frequencies
(TOF) of 28 000 hÀ1 at 28 °C to give (R)-1-phenylethanol in
82% ee.34 On the other hand, more hindered ketones, such as
tert-butyl phenyl ketone, are reduced to alcohols in up to 99%
ee.34 The high activity and selectivity of 1f has been attributed to
the (S,S)-1,2-diphenylethylenediamine (S,S-dpen) backbone in
combination with the phenyl substituents on the phosphorus
donors.34 With ethyl substituents on the phosphorus donors
the catalyst 1g is much less active and selective.35
Despite the high TOF and high enantioselectivity of complex
1f, further improvements are needed to make iron catalysts
competitive with existing ruthenium catalysts. Noyori and co-
workers discovered that they could increase the activities of their
ruthenium bis-chelate direct hydrogenation catalysts by replacing
the phenyl groups on BINAP with more electron-donating para-
tolyl groups.29 In addition, for the same catalytic systems, as well
as for related ATH systems reported by Morris and co-workers, it
was found that replacing the phenyl groups with meta-xylyl
groups afforded more enantiopure products.29,36 Conversely,
an increase in activity was seen for rhodium hydroformylation
catalysts when electron-withdrawing groups were installed on the
phosphorus donors, especially para- or 3,5-ditrifluoromethyl-
substituted phenyl groups.37,38 Furthermore, a similar trend was
seen for rhodium arylation catalysts39 and ruthenium oxidation
catalysts,40 as well as for a nickel asymmetric hydrocyanation
catalyst, which also saw increased enantioselectivity with 3,5-
ditrifluoromethyl-substituted phenyl groups.41
In this paper we explore the synthesis of a new series of
iron(II) carbonyl PÀNÀNÀP complexes containing methyl-
and trifluoromethyl-substituted aryl groups and examine their
activity for ATH. By tuning the steric and electronic properties of
4419
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
Scheme 1. Conventional (Route A) and Alternative (Route B) Syntheses of Aryl-Substituted Phosphonium Dimersa
a DIBAL refers to diisobutylaluminum hydride, and THF refers to tetrahydrofuran. b N/A: not applicable. Dimers 2b and 6b decomposed upon
dissolving in common NMR solvents, so the ratio of diastereomers could not be determined.
Scheme 2. Dynamic Behavior of Dimer 2b in DMSO and MeOH Solutions
as well as the IR spectrum, we propose that 2b is a phospho-
Scheme 3. One-Pot Template Synthesis of Complexes 3aÀc
nium dimer. It is a kinetic product that quickly forms and
precipitates out of solution, but decomposes upon dissolution
to give products that depend on the nature of the solvent
(summarized in Scheme 2).
Synthesis of Bis-acetonitrile Complexes 3aÀc. Despite the
peculiar behavior of dimer 2b, all three of the aforementioned
phosphonium dimers, 2aÀc, were successfully employed in the
previously reported template synthesis of iron(II) PÀNÀNÀP
bis-acetonitrile complexes.43,33 In a three-step, one-pot reaction,
phosphonium dimer 2a, 2b, or 2c, NaOMe, [Fe(H2O)6][BF4]2,
and (S,S)-dpen were stirred in an CH3CN/MeOH solution for
several days at room temperature to yield the desired metal
complexes 3aÀc (see Scheme 3). In the case of 3a and 3b this
process took 1À2 days, while for 3c it took 10À12 days. In order
to accelerate conversion, the reaction mixture of 3c was refluxed
for several hours.
isolated as BPh4À salts. The yields of isolated products were poor,
ranging from 10% to 37% due to the high solubility of the
complexes. However the conversions were excellent without
isolation, and these reaction mixtures were used directly in
carbonylation reactions to produce 1aÀc.
A salt metathesis reaction with NaBPh4 was required to
remove the mixture of counterions (BrÀ, FeBr4À, and BF4
)
Compounds 3aÀc all gave characteristic singlets in the 31P-
{1H} NMR spectra (61.7, 66.8, and 72.4 ppm for 3a, 3b, and 3c,
À
present in the reaction mixtures of 3aÀc, and the complexes were
4420
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
Table 1. Selected Bond Lengths (Å) and Angles (deg) for 3a
3a 0.5MeOH
3
Bond Lengths (Å)
FeÀP(1)
FeÀP(2)
FeÀN(1)
FeÀN(2)
FeÀN(3)
FeÀN(4)
2.259(2)
2.248(2)
1.965(5)
1.974(5)
1.910(5)
1.915(5)
Bond Angles (deg)
109.63(6)
82.9(2)
P(1)ÀFeÀP(2)
N(1)ÀFeÀN(2)
N(3)ÀFeÀN(4)
P(1)ÀFeÀN(1)
P(1)ÀFeÀN(2)
P(1)ÀFeÀN(3)
P(1)ÀFeÀN(4)
P(2)ÀFeÀN(1)
P(2)ÀFeÀN(2)
P(2)ÀFeÀN(3)
P(2)ÀFeÀN(4)
177.6(2)
84.0(1)
165.9(2)
85.0(1)
97.3(1)
166.3(1)
83.7(1)
Figure 2. ORTEP3 representation (thermal ellipsoids at 50%
probability) and atom numbering for 3a. The solvent, counterion, and
most hydrogens are omitted for clarity.
93.4(2)
85.0(1)
1
respectively) as well as imine peaks in the H NMR spectra
and 4c, respectively. Furthermore, the 1H NMR spectra for 4aÀc
reveal a characteristic doublet of doublets around 2.7 ppm
corresponding to the methylene protons alpha to phosphorus.
The structure of 4b was confirmed by single-crystal X-ray
diffraction (see Supporting Information, Figure S2).
(around 8.2 ppm for 3aÀc). Furthermore, two sets of inequi-
valent methyl groups (around 2.2 ppm) were seen for all three
complexes.
Crystals of 3a as a tetrabromoferrate salt and a methanol
solvate, suitable for X-ray diffraction studies, were grown from
the slow diffusion of ether into an CH3CN/MeOH solution of
the crude reaction mixture. The geometry around iron in 3a,
shown in Figure 2, is a distorted octahedron with acetonitrile
ligands in the apical positions trans to each other. The PÀFeÀP
angle is obtuse, 109.63(6)°, which suggests that the three five-
membered rings formed by the tetradentate PÀNÀNÀP ligand
constrain the geometry of the metal center considerably. Other
notable bond lengths and angles are given in Table 1. Another
important feature of this structure is that the phenyl rings of the
diamine backbone are equatorial, while the less sterically de-
manding hydrogens occupy the axial positions. Compound 3b
was also characterized by single-crystal X-ray diffraction and
displayed a similar geometry (see Supporting Information,
Figure S1).
Alternative Synthesis (Route B) of Phosphonium Dimers
2aÀc and 6aÀc. A drawback of synthetic route A (see
Scheme 1) is that highly air-sensitive diarylphosphines are
needed as starting materials. The literature synthesis of these
compounds involves the use of Grignard reagents and diethyl
phosphite to generate diarylphosphine oxides, which are subse-
quently reduced to phosphines using DIBAL.44 The reduction
step in this synthesis is tedious, labor intensive, and, in some
cases, slow and low-yielding. This was the impetus for the
development of an alternative, more convenient, and efficient
method of synthesizing phosphonium dimers (Scheme 1, route
B). In the first step, a diarylphosphine oxide is deprotonated and
acts as a nucleophile to displace the BrÀ of bromoacetaldehyde
diethyl acetal in an SN2 reaction. This affords diarylphosphine
oxide acetaldehyde diethyl acetals 4aÀc as air-stable solids in
moderate yields (60À75%). The 31P{1H} NMR spectra are
diagnostic with singlets at 28.6, 30.1, and 28.7 ppm, for 4a, 4b,
In the second step of route B, phosphine oxides 4aÀc are
reduced to their corresponding phosphines, 5aÀc, using lithium
aluminum hydride. Colorless oils of 5aÀc can be isolated in
moderate yields (45À60%). Despite the limitations of the yield,
this synthetic route has practical advantages: it is easy to set up
and requires simple product workup. In addition, the crude
product can be used directly to synthesize phosphonium dimers
because all impurities present in the crude reagent are removed
during dimer purification. Moreover, if a pure form of the
phosphinoacetaldehyde diethyl acetal is desired, the crude
product can be purified using a silica gel column in air. Unlike
secondary diarylphosphines, compounds 5aÀc are only mildly
air-sensitive, even in solution, and only begin to oxidize upon
1
extended exposure to oxygen. The H NMR spectra of 5aÀc
were similar to those of 4aÀc, but the 31P{1H} NMR chemical
shifts showed diagnostic peaks around À24.2 ppm for com-
pounds 5a and 5b and À45.5 ppm for compound 5c.
In the final step, crude products of 5aÀc were dissolved in
THF and an excess of aqueous acid was added. After stirring
overnight, phosphonium dimers 2aÀc or 6aÀc could be isolated
as mixtures of meso- and rac-diastereomers in excellent yields
(87À94%, see Scheme 1). The spectra of 2aÀc were identical to
those of 6aÀc, and the structures of both 6a and 6c were
determined by single-crystal X-ray diffraction (see Supporting
Information, Figures S3 and S4). Much like 2a, the isomers that
were crystallized for 6a and 6c were meso with six-membered
rings in the chair conformation.
The synthesis of BF4À salts 6aÀc demonstrates the utility of
synthetic route B versus route A. In route B, any acid can be used
to generate a phosphonium dimer from phosphines 5aÀc
because the air-stable intermediates 4aÀc can be isolated free
4421
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
Scheme 4. Template Synthesis of 3aÀc Using 6aÀc
templating reaction: these compounds have the opposite proper-
ties of phosphonium dimers. Whereas phosphonium dimers are
stable with respect to acid and generate phosphinoacetaldehydes
upon exposure to base, 5d and 5e are stable with respect to base
and generate phosphinoacetaldehydes upon exposure to acid.
The conventional template approach illustrated in Scheme 3 is
base-catalyzed and thus is unsuitable for these starting materials.
Instead, an acid-catalyzed templating approach was developed
(see Scheme 6).
A diarylphosphinoacetaldehyde diethyl acetal precursor, 5d or
5e, was dissolved in CH3CN with [Fe(H2O)6][BF4]2, followed
by addition of S,S-dpen. A small amount of aqueous HBF4 was
subsequently introduced into the system in order to deprotect 5d
or 5e and generate a diarylphosphinoacetaldehyde in situ. The
solution was then neutralized with NaOMe, and upon workup,
dark red solids could be isolated in moderate yields (45À50%).
Both 3d and 3e showed characteristic singlets in their 31P{1H}
NMR spectra at 69.0 and 72.0 ppm, respectively, as well as
distinctive imine peaks in the 1H NMR spectra around 8.2 ppm.
A notable difference between 3d and 3e and their electron-
donating counterparts 3aÀc was that there were no inequivalent
methyl peaks in the 1H NMR spectra. Instead, two inequivalent
trifluoromethyl peaks were evident in the 19F NMR spectra (at
À63.4 and À63.6 ppm). It is interesting to note that the singlets
in the 31P{1H} NMR spectra and the imine peaks in the 1H NMR
spectra of 3d and 3e were not shifted upfield by a significant
amount relative to those of 3aÀc, indicating that these chemical
shifts are insensitive to the electronic changes made at
phosphorus.
The mechanism of the acid-catalyzed template synthesis is still
poorly understood. Throughout the reaction, before arriving at
the final product, the major peak in the 31P{1H} NMR spectra is a
broad singlet around À10.1 ppm. This seems to indicate that the
phosphorus is at most loosely bound to the iron, rapidly
associating and dissociating such that this peak appears at the
average of the chemical shifts of bound and unbound phos-
phorus. In addition, the reagents were added in several different
sequences, but the given permutation was the only procedure
that yielded the desired product. If diamine addition is withheld
until after neutralization with base, a mixture of products is
generated. If, on the other hand, the deprotection of 5d or 5e is
attempted in the presence of both iron(II) and diamine, neu-
tralization of the solution yields the desired PÀNÀNÀP com-
plex 3d or 3e.
It is surprising that deprotection of 5d or 5e must occur in the
presence of both iron(II) and diamine because protonation of
the basic diamine will occur rapidly upon addition of acid to this
mixture. The protonated diamine, however, can act as a weak acid
and deprotect the diarylphosphinoacetaldehyde diethyl acetal.
After the diethyl acetal is deprotected, iron can act as a Lewis acid
to increase the electrophilicity of the carbonyl carbon and
facilitate Schiff base reactions. The amount of acid added is only
enough to protonate each diamine once, so attack on the
carbonyl carbon by the unprotonated amino group of the
diamine would form a protonated tridentate ligand. This could
explain the 31P{1H} NMR spectrum because a protonated
tridentate ligand with an electron-deficient phosphorus donor
would be prone to dissociation. After deprotonating the pro-
posed tridentate ligand, it could coordinate to iron, forming
complex 7, which was characterized by X-ray crystallography (see
Supporting Information, Figure S5). Upon further condensation
with another equivalent of diarylphosphinoacetaldehyde, 7 could
of all halide salt byproduct. Conversely, in route A, only the use of
aqueous HCl or HBr is practical, depending on the haloacetal-
dehyde diethyl acetal used in the synthesis because, otherwise,
the product would have a mixture of counterions.42,45 Thus,
route B permits the facile introduction of a variety of counterions
into the phosphonium salts and greatly increases their value as
ligand precursors. Halide counterions can coordinate to transi-
tion metals and can compete with desired ligands to give
unwanted side products. For example as illustrated in Scheme 3,
an extra half-equivalent of iron(II) starting material is needed to
accommodate the formation of FeBr42À when using a bromide
salt of the phosphonium dimer.43,33
With a BF4À salt, however, the excess iron starting material is
unnecessary, and the template synthesis is more atom-economic-
al (see Scheme 4). With route B, phosphonium dimers can be
tailored to different reaction conditions and metal pecursors by
simply changing the acid used to generate them. This luxury is
often lacking in synthetic inorganic chemistry, where it is
generally the metal precursor that is changed when a reaction
does not give the desired product, or a sacrificial reagent such as
AgBF4 is used to remove unwanted halides.
Synthesis of Precursors 5d and 5e, as Well as Bis-acetoni-
trile Complexes 3d and 3e. In addition to substituting the aryl
groups on phosphorus with electron-donating methyl groups,
efforts were made to substitute them with electron-withdrawing
trifluoromethyl groups in the para- and 3,5-positions. Unfortu-
nately, with these substituents, neither route A nor B gave the
desired products. The strongly electron-withdrawing aryl groups
reduced the nucleophilicity of the bonded phosphorus center to
such an extent that it was unable to displace a bromide as a
phosphine oxide anion in route B and unable to attack an
aldehyde after acid addition as a trisubstituted phosphine in
route A (summarized in Scheme 5).
Fortunately, the diarylphosphinoacetaldehyde diethyl acetals
5d and 5e could be synthesized cleanly if the temperature of the
reaction mixture was controlled (see Experimental Section for
details). Moreover, these compounds can be purified using silica
gel chromatography in air because the oxidation potential of the
phosphorus is raised by the electron-withdrawing substituents.
To synthesize 5d and 5e, the corresponding diarylphosphine was
deprotonated by KH in order to generate a potassium phosphide
salt. After addition of bromoacetaldehyde diethyl acetal and
workup, a dark oil was obtained in moderate yields (72À78%).
Both 5d and 5e gave singlets in the 19F NMR spectra at À63.3 ppm,
as well as in the 31P{1H} NMR spectra at À22.0 and À19.0 ppm,
respectively. The 1H NMR spectra of5d and 5e were very similar to
those of compounds 5aÀc.
A problem arose when contemplating the use of diarylpho-
sphinoacetaldehyde diethyl acetal precursors 5d and 5e in a
4422
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
Scheme 5. Synthesis and Lack of Reactivity of Trifluoromethyl-Substituted Precursors 5d and 5e
The bis-acetonitrile intermediates used to generate 1aÀc were
not isolated in order to reduce the number of purification steps
and to increase atom economy.35
Scheme 6. Acid-Catalyzed Synthesis of 3d and 3e
For 1d and 1e, on the other hand, the CO addition reaction
was performed in CH2Cl2 with only one equivalent of KBr. The
reaction progressed very slowly, taking over a week to proceed to
completion, and required several cycles of solvent removal under
reduced pressure, followed by addition of fresh solvent to remove
any traces of CH3CN from solution. The products, yellow solids,
À
were isolated as BF4 salts in moderate yields (40À50%).
Dichloromethane was chosen over acetone to prevent decom-
position of 3d and 3e, and only one equivalent of KBr was used in
order to prevent the formation of a mixture of products. The
formation of 1d and 1e proceeded slowly undoubtedly because
of the electron-deficient nature of the iron. Without much
electron density to π back-donate into the CO ligand, the
ironÀcarbon bond is much weaker. This would make other,
more basic ligands, such as CH3CN or BrÀ, more competitive for
coordination sites than CO and explains why the stoichiometry
of KBr needs to be controlled as well as why the CH3CN needs to
be continually removed from the system.
Indeed, the electron-deficient nature of the iron in complexes
1d and 1e compared to 1aÀc is reflected in the CO vibrations
listed in Table 2, which also lists the CO vibrations for literature
complexes 1fÀi (see Figure 1 for structures), in addition to the
steric parameters of the substituted phosphorus donors of 1aÀi.
There is a 15À20 wavenumber difference between the CO
vibrations of the approximately isosteric complexes with elec-
tron-donating and electron-withdrawing substituents, while
there is a 45 wavenumber range overall for 1aÀi. The difference
in wavenumber, however, is not reflected in the 31P{1H} NMR
spectra. All of the CO complexes show diagnostic pairs of
doublets between 60.0 and 70.0 ppm. An interesting feature to
note for 1c is that the 31P{1H} NMR spectrum shows an extreme
second-order pattern at 61.3 ppm. Additionally, 1b exhibits no
phosphorus peaks at room temperature. This is due to limited
rotation of the ortho-methyl-substituted aryl groups about the
phosphorus donors as well as fast interconversion of roto-
isomers on the time scale of the NMR experiment. If the
spectrum is acquired at low temperatures, four sets of doublets
appear: one major set, one minor set, and two sets with similar,
intermediate intensities. This indicates there are four roto-
isomers in solution with slightly different energies. The 31P{1H}
generate 3e (summarized in Scheme 7). Presumably, the triden-
tate ligand is too sterically demanding to allow the formation of a
bis-tridentate ligand complex, as was reported previously by our
group.43,46
The acid-catalyzed deprotection of a precursor in the presence
of a template is not an unfamiliar concept. This type of synthesis
is used extensively by materials chemists in the preparation of
mesoporous silica.47 Acid is used to hydrolyze alkoxy-silane
precursors in the presence of a liquid crystal template in order
to produce silica with periodic channels.47 This synthetic
approach, however, is completely unknown in inorganic chem-
istry, where, in general, an acid-catalyzed template synthe-
sis refers to a Lewis-acid-catalyzed process.48 In addition,
there are several Brønsted-base-catalyzed template syntheses
known,43,46,49,50 but the literature on Brønsted-acid-catalyzed
templating is nonexistent. With the success of this process, a
whole new range of compounds can be targeted for ligand
synthesis, especially those that form unstable species upon
treatment with acid. Under templating conditions, these spe-
cies can be generated in situ and used to produce previously
unknown series of ligands.
Synthesis of Iron(II) Carbonyl PÀNÀNÀP Complexes
1aÀe. The new iron(II) carbonyl complexes 1aÀc were synthe-
sized according to a known literature procedure,35,34 while small
changes were made to generate complexes 1d and 1e (see
Scheme 8). For 1aÀc the CO addition reaction was performed
in acetone in the presence of an excess of KBr, and after several
À
days, the products were isolated as BPh4 salts in low yields
(10À35% overall with respect to the phosphonium dimer
precursors).
4423
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
Scheme 7. Proposed Mechanism for the Formation of 3e Using the Acid-Catalyzed Template Approach
Scheme 8. General Method for Synthesizing 1aÀe
Figure 3. Reaction profiles of the ATH of acetophenone to (R)-1-
phenylethanol using 1aÀe, in iPrOH, and in the presence of KOtBu. C/
B/S = 1:8:1000, T = 28.2 °C, [Fe] = 92.4 μM. (9) 1a, ee = 84%; (2) 1c,
ee = 90%; (b) 1b, 1d, and 1e. Each data point represents the average of
three runs, and error bars are smaller than the symbols themselves.
NMR spectra of all the CO complexes show pairs of doublets due
to the C1 symmetry produced by the chiral S,S-dpen backbone
and the different trans ligands on iron. This lack of symmetry also
manifests in the 1H and 19F NMR spectra of 1aÀe. For 1aÀc, the
methyl substituents are inequivalent and four methyl peaks
around 2.1 ppm can be seen (for 1b, there are 16 methyl peaks:
four for each roto-isomer). The trifluoromethyl-substituted
complexes 1d and 1e show a similar set of four peaks but in
the 19F NMR instead (around À63.5 ppm).
different. The para-methyl-substituted system, 1a, showed high
activity and selectivity for the asymmetric hydrogenation of
acetophenone, while the ortho-methyl-substituted system, 1b,
was completely inactive (see Figure 3).
One possible mechanism of action for these catalysts was
thought to be the dissociation of a phosphorus donor to create an
open site for substrate coordination so that reduction could
occur in a MeerweinÀPonndorfÀVerley fashion.51 This seems
unlikely, though, in light of the inactivity of 1b. The increased
steric bulk around the phosphorus donors should favor dissocia-
tion, and indeed, the catalytic mixture displays numerous peaks
in the negative region of the 31P{1H} NMR spectrum. Thus we
conclude that under catalytic conditions the decoordination of
phosphorus donors for 1b occurs, but instead of leading to an
active catalytic species, this process leads to numerous decom-
position products.
We also synthesized a meta-xylyl version of our catalyst, 1c, in
the hopes that the increased steric bulk of these groups, which is
slightly removed from the metal center, would afford more
enantiopure products. Whereas 1a produced (R)-1-phenyletha-
nol with an ee of 84%, very similar to the 82% ee previously
reported for 1f, 1c gave an improved ee of 90% for the production
of (R)-1-phenylethanol. Only a few other catalysts produce such
ATH Using Precatalysts 1aÀe. We have previously shown
that by replacing the phenyl groups on phosphorus with alkyl
groups, the catalytic activity of the iron(II) carbonyl complexes
for the ATH of ketones is either completely deactivated (1h and
1i) or drastically decreased (1g).35 It was believed that the steric
bulk of the cyclohexyl and isopropyl groups was the cause of this
deactivation, but it was difficult to definitively conclude whether
or not this was true because these groups impose drastic
electronic as well as steric effects.35 Furthermore, the catalysts
with less sterically demanding ethyl substituents showed only
modest turnover frequencies at elevated temperatures, illustrat-
ing that electronic factors are playing a role in the lower activity of
the alkyl complexes.35
In order to clearly separate the electronic from the steric
effects, we prepared complexes with ortho- and para-substituted
phenyl groups on the phosphorus donors. In this way, the
electronics of the two systems would be very similar, but the
steric crowding about the iron center would be drastically
4424
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
Table 2. IR CO Stretches and Tolman Cone Angles of Iron(II) Carbonyl Complexes 1aÀi
compound
substituent on phosphorus
CO stretch (cmÀ1
)
cone anglea (deg)
ref
1i
cyclohexyl
1945
1951
1956
1970
1972
1974
1975
1989
1990
157
132
151
148b
141
173
141
141
148b
35
1g
1h
1c
1a
1b
1f
ethyl
35
isopropyl
35
3,5-(CH3)2C6H3
para-CH3C6H4
ortho-CH3C6H4
phenyl
this work
this work
this work
34
1d
1e
para-CF3C6H4
3,5-(CF3)2C6H3
this work
this work
a Tolman cone angles for the phosphorus donors were calculated using parameters from Tolman (ref 52), Giering et al. (ref 53), and Po€e and co-workers
(ref 54) (see Supporting Information for sample calculations). The linkage between the phosphorus and imine moieties is modeled as an ethyl group.
b Cone angles approximated from crystal structure 7. The FeÀP bond length is 2.22 Å for 7, which is close to the 2.28 Å NiÀP bond length in the model
Ni0 complex used to determine Tolman cone angles (ref 52). Note that this is a tentative estimate for both 1c and 1e (we approximate that CF3 and CH3
are similar in size).
high enantioselectivities for the ATH of unsubstituted acetophe-
none, and none are as active as 1c (see Table 3).28 Generally,
such high ee are observed with other, bulkier ketones, with
isopropyl or tert-butyl groups, as was seen by our group while
screening substrates for 1f.33 For all of our iron(II) carbonyl
catalysts, however, the ee begins to drop after approximately 10
minutes. More studies are underway to determine the nature of
this loss of enantioselectivity.
We propose that all of the aforementioned active catalysts
operate under the same mechanism because of some key
similarities: they all have distinctive sigmoidal reaction profiles,
and they all show the same color changes upon activation. In
Figure 3, the reaction profiles of 1a and 1c show a brief induction
period followed by a linear increase in product formation, which
then tapers off as the concentration of substrate decreases and the
system reaches equilibrium. This is consistent with results
previously reported for 1f.34 Furthermore, 1a and 1c need to
be activated by a strong base to be catalytically active, and upon
reacting with base, the compounds turn green in color. This
colored transition is also seen with the alkyl-substituted catalysts
where a doubly deprotonated, five coordinate, bis-ene-amido
iron(II) carbonyl complex was recently identified.55 Yet another
result in support of this proposal is that inactive 1b does not turn
green upon reacting with base. Instead, the solution turns nearly
colorless as the complex decomposes.
After determining that substitutions in the ortho-position led
to inactive catalysts and substitutions in the 3,5-positions led to
increased selectivity, we rationalized that some mechanistic
insights might be obtained if substitutions were made with
electron-withdrawing trifluoromethyl groups in the para- and
3,5-positions. Precatalysts 1d and 1e were tested for activity
under similar conditions to 1aÀc, but much to our surprise, these
catalysts were completely inactive. If the mechanism involves
transferring a hydride equivalent to an iron intermediate, 1d and 1e
should favor and stabilize these species due to the electron-deficient
nature of the metal center, but an increase in activity is not seen. In
fact, upon addition of base to 1d or 1e, they do not turn green in
color; instead, they turn orange. IR spectra of these reaction
mixtures indicated that the CO ligand was still present after the
base was added. We are currently exploring the reasons why 1d and
1e are completely inactive for the ATH of acetophenone.
catalytic conditions, to 1f. The initial rates of 1a, 1c, and 1f were
determined using high substrate loadings in order to saturate the
catalysts and ensure a linear increase in product formation for an
extended period of time (see Figure 4).
The TOF of 1f was an impressive 28 000 ( 200 hÀ1, with an ee
of 82% as previously reported.46 The TOF of isosteric 1a was
even higher, at 30 000 ( 1500 hÀ1 with an ee of 84%, while the
TOF of 1c was slightly lower at 26 000 ( 1000 hÀ1 with an
elevated ee of 90%.
A striking and peculiar feature of the behavior of 1a and 1c
stands out when Figures 3 and 4 are examined closely. In Figure 3,
if 1a and 1c are compared, it is evident that 1c is slightly faster. In
Figure 4, on the other hand, 1a is faster. This result seems
somewhat counterintuitive at first; by increasing the concentra-
tion of substrate, the rate of 1c does not increase as much as the
rate of 1a. If both systems operate under the same mechanism,
this observation could be explained by a larger equilibrium
constant for the formation of an inactive enolate complex for
1c than for 1a. This explanation is not favored, however, because
transfer hydrogenation was conducted using benzophenone as
the substrate, which cannot form an enolate, and the results
showed that the rate of 1c still had a lower dependence on the
concentration of substrate than 1a when increasing the substrate
loading 6-fold (see Supporting Information, Figures S6ÀS11).
The mechanism of ATH is still under investigation, and further
studies may elucidate the cause of this interesting behavior.
Complexes 1a, 1c, and 1f were all exceptionally active, but
there does not seem to be a dramatic difference between them.
We previously reported that electron-rich, alkyl-substituted
phosphorus donors inhibited catalysis, yet 1a is more active than
1f, and 1c is only slightly slower, which is surprising, as both 1a
and 1c should be more electron-donating than 1f. The IR spectra,
however, show that catalysts 1a, 1c, and 1f have similar carbonyl
stretches (see Table 2), indicating that the iron centers have
comparable electron densities. It is only when the carbonyl
stretches vary drastically from these three complexes, as with
1d and 1e or the alkyl-substituted complexes 1gÀi (see Table 3),
that the activity decreases drastically or disappears entirely.
In general, it seems that our iron(II) carbonyl catalysts require
a delicate balance of both sterics and electronics at phosphorus to
keep them active, as well as selective. If they are too electron-rich,
their activity begins to drop, but if the compounds are too
electron-deficient, they are completely inactive. At the same time,
After determining that 1a and 1c were active for the ATH of
acetophenone, they were compared directly, under identical
4425
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
Table 3. Comparison of Activity, Selectivity, and Reaction Conditions between 1c and Several Highly Selective Platinum-Metal-
Based ATH Catalysts Reported in the Literaturea
catalyst
S/C
temp (°C)
yield (%)/time
% ee
ref
1c/base
1000
200
200
200
200
100
200
100
200
200
100
100
100
28
22
45
28
22
22
28
22
30
30
22
20
22
94/7 min
95/15 h
93/7 h
90
97
97
92
94
91
96
90
97
96
94
92
93
this work
56
[RuCl2(mes)]2/Tsdpen/base
RuCl2(8)/base
32
[RuCl2(hmb)]2/9/base
94/1 h
57
[RuCl2(p-cym)]2/10/base
[RuCl2(p-cym)]2/11/base
RuCl(12)/base
92/1.5 h
70/1.5 h
>99/18 h
80/48 h
85/12 h
36/12 h
96/48 h
97/3 h
58
59
60
Cp*RhCl(Tsdpen)/base
Cp*RhCl(Tscydn)/base
Cp*IrCl(Tscydn)/base
23
61
61
[RuCl2(p-cym)]2/13/base
RuHCl(xylbinop)(dpen)/base
RuH(BH4)(xylbinop)(dpen)
a Where the ligands are abbreviated as.
62
63
92/8 h
36
if they are too sterically hindered, they become completely
inactive, but if they are not sterically hindered enough, the
enantioselectivity of the catalysts drops significantly. This delicate
balance of stereoelectronics can be represented graphically by
plottingthe catalytic activity ofthe ironcomplexes, inTOF, against
CO stretching frequencies and Tolman coneangles (see Figure5).
This distinctive plot, a “volcano plot”, reveals a narrow region of
steric and electronic parameters that produces highly active ATH
catalysts. Deviation from this region leads to completely inactive
catalysts or severely diminished activity in the case of 1g.35 The
phenyl groups, as well as para-methyl- and 3,5-dimethyl-substi-
tuted phenyl groups, must provide the right balance of sufficient
Lewis basicity and steric bulk on the phosphorus donors to ensure
fast turnover and high selectivity, but not so much electron density
that the active species becomes too electron-rich to effectively
abstract a hydrogen equivalent from i-PrOH.
however, the system is homogeneous, and the metal center is
kept constant, while the phosphorus donors (L) are varied.
Similar studies were conducted by Giering et al.53 and Po€e
et al.,54 where the steric and electronic effects of the phosphorus
donors on the reduction potentials of η5-Cp(CO)(L)Fe-
(COMe) and the fragmentation of Os3(CO)9(μ-C4Ph)(L) were
examined, respectively. In both studies, the influence of the
stereoelectronic factorsofthephosphorusdonorson thetransition
states of these processes was explored in a qualitative manner. The
approach in this paper is somewhat unique in that the activity
of a catalytic system is examined with respect to both steric
and electronic parameters on phosphorus donors. Rajanbabu
et al.41,67À69 and Gebbink et al.70 have performed analogous
studies with respect to the influence of phosphorus donors on
catalytic activity, except that only electronic parameters were
examined in the first case, and a qualitative approach regarding
activity was used in the latter case. In the current work the TOFs
of the catalysts, which are dictated by the transition states of the
“Volcano plots” have been extensively used to understand and
design systems in heterogeneous catalysis.64À66 In this case,
4426
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
Figure 4. Linear, initial portions of the ATH of acetophenone to (R)-1-
phenylethanol using 1f, 1a, and 1c. C/B/S = 1:8:6000, T = 28.2 °C, [Fe] =
92.4 μM. (9) 1a, ee = 84%; (2) 1c, ee = 90%; (b) 1f, ee = 82%. Each data
point represents the average of three runs.
rate-limiting steps, are compared to the stereoelectronic factors
of the phosphorus donors. This type of analysis, in conjunction
with density functional theory (DFT) and kinetic data, can be
used to gain insights into the nature of the transition state of the
rate-determining step for catalytic cycles. Moreover, with this
type of analysis, “hot-spots” for activity can be identified and
targeted for future ligand design and catalyst optimization.
Figure 5. “Volcano plot” of catalyst activity (TOF, hÀ1) versus CO
stretch (cmÀ1), as an electronic parameter, and Tolman cone angle
(deg), as a steric parameter, for precatalysts 1aÀi.
1aÀe for ATH was compared to the activities of previously
reported complexes 1fÀi, and the most active catalysts displayed
a narrow range of electronic and steric parameters for their
phosphorus donors, as defined by CO stretches and Tolman
cone angles, respectively. This analysis may give insights into the
mechanism of catalysis and lead to more rational catalyst design.
’ CONCLUDING REMARKS
We have synthesized a series of five aryl-substituted iron(II)
carbonyl PÀNÀNÀP compounds and tested them for the ATH
of prochiral ketones. Three of these complexes, 1a, 1b, and 1c,
had electron-donating methyl groups in the para-, ortho-, and 3,5-
positions, respectively, while the other two complexes, 1d and 1e,
had electron-withdrawing trifluoromethyl groups in the para-
and 3,5-positions, respectively. Compounds 1aÀc were synthe-
sized according to known literature procedures: starting with
phosphonium dimers, 2aÀc, the bis-acetonitrile complexes,
3aÀc, were generated in a one-pot template synthesis and
converted to precatalysts in a CO addition reaction. While
investigating the methyl-substituted compounds, an alternative
and more convenient route for synthesizing phosphonium
dimers was developed, which allows facile introduction of any
desired counterion, thus making the phosphonium dimers more
synthetically useful and valuable. Compounds 1d and 1e, on the
other hand, could not be synthesized using previously developed
methods. Phosphonium dimers of the electron-withdrawing
groups could not be isolated, so phosphinoacetaldehyde diethyl
acetal precursors, 5d and 5e, were isolated instead. These
precursors were incompatible with the previously reported
template synthesis, so an unprecedented one-pot, acid-catalyzed
template procedure was developed to generate bis-acetonitrile
complexes 3d and 3e. Although the mechanism of the acid-
catalyzed template is not completely known, it may go through a
tridentate, trisacetonitrile intermediate. Compounds 3d and 3e
were then converted to precatalysts in a standard CO addition
reaction. Of the five CO compounds synthesized, only 1a and 1c
were active catalysts; 1b had too much steric crowding about the
iron center, while 1d and 1e were too electron-poor. Complexes
1a and 1b were very successful: 1a showed increased activity
compared to 1f and is now the most active iron ATH catalyst
reported to date, while 1c produced more enantiopure (R)-1-
phenylethanol with an ee of 90% and is now the most selective
iron ATH catalyst developed thus far. The activity of complexes
’ EXPERIMENTAL SECTION
General Considerations. All procedures and manipulations were
performed under an argon or nitrogen atmosphere using standard
Schlenk-line and glovebox techniques unless stated otherwise. All
solvents were degassed and dried using standard procedures prior to
all manipulations and reactions unless stated otherwise. The synthesis of
the para-methyl-substituted phosphonium dimer 2a has previously been
reported.71 Deuterated solvents were purchased from Cambridge Iso-
tope Laboratories or Sigma Aldrich, degassed, and dried over activated
molecular sieves prior to use. All other reagents were purchased from
commercial sources and utilized without further purification. The IR
spectra of the KBr pellets containing precatalysts 1aÀi were measured
on a Paragon 500 spectrometer (spectral range 4600 to 400 cmÀ1) at
25 °C using Perkin-Elmer’s SPECTRUM for data collection and
processing. The ESI-MS data were collected on an AB/Sciex QStar
mass spectrometer with an ESI source, the EI-MS data were collected on
a Waters GC ToF mass spectrometer with an EI/CI source, and the
DART-MS data were collected on a JEOL AccuTOF-DART mass
spectrometer with a DART-ion source (no solvent is required). NMR
spectra were recorded at ambient temperature and pressure using a
Varian Gemini 400 MHz spectrometer (400 MHz for 1H, 100 MHz for
13C, 376 MHz for 19F, and 161 MHz for 31P) unless stated otherwise.
1
The H and 13C NMR were measured relative to partially deuterated
solvent peaks but are reported relative to tetramethylsilane. All 19F
chemical shifts were measured relative to trichlorofluoromethane as an
external reference. All 31P chemical shifts were measured relative to 85%
phosphoric acid as an external reference. The elemental analyses were
performed at the Department of Chemistry, University of Toronto, on a
Perkin-Elmer 2400 CHN elemental analyzer. Some complexes gave
inconsistent carbon analyses but acceptable hydrogen and nitrogen
4427
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
contents; we attribute this to a combustion problem due to boron-
containing counterions.72
134.20 (ArÀCH), 134.11 (Ar-C), 131.74 (Ar-CH), 130.18 (Ar-CH),
129.90 (Ar-CH), 129.79 (Ar-C), 129.63 (Ar-CH), 125.87 (B-C-CH-
CH), 122.04 (B-C-CHÀCH-CH), 78.76 (CH-NH), 46.08 (CH2-P),
and 20.76 (Ar-CH3) ppm. Anal. Calcd for [C50H52N4P2Fe]-
[B(C6H5)4]2: C, 80.33; H, 6.33; N, 3.82. Found: C, 76.31; H, 6.71;
N, 3.77. MS (ESI, methanol/water; m/z+): 372.1 [C50H52N4P2Fe]2+.
Synthesis of [C50H52N4P2Fe][B(C6H5)4]2 (3b). Similar to the
synthesis of 3a; see Supporting Information S3ÀS4.
Synthesis of [C32H36O2P2][Br]2 (2b). A Schlenk flask was
charged with KH (0.379 g, 9.5 mmol) and dry THF (9.5 mL). Di-
(ortho-tolyl)phosphine (1.688 g, 7.9 mmol) was added, and the solution
turned red in color. The solution was stirred for 90 min and then cooled
to À78 °C. Bromoacetadehyde diethyl acetal (1.22 mL, 10.9 mmol) was
added over 20 min, and the solution turned yellow. The solution was
warmed to room temperature, and 48% HBr (1.8 g, 10.7 mmol) was
added. A white precipitate formed and the solution turned colorless. The
mixture was left stirring for 2 h and then placed in a freezer (À40 °C)
overnight. The precipitate was filtered off and washed with H2O (2 Â
15 mL), as well as ethyl acetate (15 mL). The precipitate was dried under
high vacuum. Yield: 83.1% (1.69 g). 1H NMR (400 MHz, DMSO-d6) δ:
12.60À12.05 (vb s, 3H), 9.59 (dt, 2H, Ar-H, J = 3.3, 1.1 Hz), 7.65 (ddd,
4H, Ar-H, J = 13.7, 7.7, 1.3 Hz), 7.54 (ddd, 2H, Ar-H, J = 13.7, 7.6,
1.4 Hz), 7.47 (tt, 4H, Ar-H, J = 7.5, 1.5 Hz), 7.41 (tt, 2H, Ar-H, J = 7.5,
1.4 Hz), 7.36À7.24 (m, 10H, Ar-H), 7.24À7.13 (m, 2H, Ar-H), 6.91
(dd, 1H, OHCHdCH, J = 13.1, 10.1 Hz), 5.32 (dd, 1H, CHdCHP, J =
17.6, 13.1 Hz), 3.90 (dd, 4H, CH2P, J = 14.1, 3.3 Hz), 2.21 (s, 12H,
CH3), and 2.19 (s, 6H, CH3) ppm. 31P{1H} NMR (161 MHz, DMSO-
d6) δ: 28.15 (s) and 25.13 (s) ppm. 13C NMR (100 MHz, DMSO-d6) δ:
197.13 (d, CdO, J = 4.1 Hz), 159.70 (d, HOCHdCH, J = 14.7 Hz),
141.26 (d, Ar-CP, J = 9.0 Hz), 140.95 (d, Ar-CP, J = 8.6 Hz), 134.41 (s,
Ar-C), 133.39 (s, Ar-C), 132.65 (d, Ar-CH, J = 2.7 Hz), 132.28À130.76
(m, Ar-CH), 126.34 (d, Ar-CH, J = 12.2 Hz), 126.10 (d, Ar-CH, J =
12.0 Hz), 93.03 (d, CHdCHP, J = 116.02 Hz), 46.35 (d, CH2P, J =
59.4 Hz), 21.05 (d, Ar-CH3, J = 4.2 Hz), and 20.93 (d, Ar-CH3, J = 4.5 Hz)
Synthesis of [C54H60N4P2Fe][B(C6H5)4]2 (3c). Similar to the
synthesis of 3a; see Supporting Information S4ÀS5.
Synthesis of [C20H27O3P] (4a). A Schlenk flask was charged with
KH (0.166 g, 4.13 mmol) and dry THF (8 mL). A solution of bis(para-
tolyl)phosphine oxide (1.00 g, 4.13 mmol) in THF (8 mL) was added.
Gas evolved, and the solution turned yellow. After 30 min of stirring, gas
evolution ceased, and the solution was cooled to 0 °C. Bromoacetalde-
hyde diethyl acetal (0.725 g, 4.13 mmol) was added to the mixture over
5 min. The solution turned lighter yellow and cloudier. After 20 h, the
solution was warmed to room temperature and ether (20 mL) was
added. The solution was filtered, and the solvent was removed under
reduced pressure to give a cloudy oil. The oil was dissolved in 3:1
hexanes/ether and cooled in a freezer (À40 °C) overnight. The solution
was filtered, and the solvent removed under reduced pressure to give a
clear, colorless oil. Yield: 54.6% (0.788 g). 1H NMR (400 MHz, CDCl3)
δ: 7.66À7.59 (m, 4H, ArÀ-H), 7.26À7.20 (m, 4H, Ar-H), 4.99 (q, 1H,
O-CH-O, J = 5.5 Hz), 3.62À3.54 (m, 2H, O-CH2), 3.45À3.37 (m, 2H,
O-CH2), 2.69 (dd, 2H, P-CH2-CH, J = 11.4, 5.5 Hz), 2.38 (s, 6H, Ar-
CH3), and 0.99 (t, 6H, CH2-CH3, J = 7.0 Hz) ppm. 31P{1H} NMR
(CDCl3, 161 MHz) δ: 28.56 (s) ppm. 13C NMR (100 MHz, CDCl3) δ:
141.85 (d, Ar-CP, J = 2.9 Hz), 130.92 (s, Ar-CH), 130.83 (s, Ar-CH),
129.91 (s, Ar-C), 129.15 (s, Ar-CH), 129.03 (s, Ar-CH), 98.80 (s,
CHOO), 62.44 (s, CH2O), 36.16 (d, OOCHCH2P, J = 71.3 Hz), 21.53
(d, Ar-CH3, J = 1.3 Hz), and 14.94 (s, CH3) ppm. Anal. Calcd for
[C20H27O3P]: C, 69.35; H, 7.86. Found: C, 69.08; H, 7.83. MS (DART,
dichloromethane; m/z+): 347.2 [C20H28O3P]+.
1
ppm. H NMR (400 MHz, CD3OD) δ: 7.56À7.49 (m, 2H, Ar-H),
7.46À7.41(m, 2H,Ar-H),7.32À7.26(m, 4H,Ar-H),4.49(dt, 1H, O-CH-
O, J = 8.4, 5.2 Hz), 3.05À2.98 (m, 2H, P-CH2-CH), and 2.33 (s, 6H, Ar-
CH3) ppm. 31P{1H} NMR (161 MHz, CD3OD) δ: À21.04 (s) ppm. 13
C
NMR (100 MHz, CD3OD) δ: 142.52 (d, Ar-CP, J = 16.1 Hz), 135.58 (s,
Ar-C), 132.63 (d, Ar-CH, J = 6.5 Hz), 132.19 (s, Ar-CH),131.03 (d, Ar-
CH, J = 8.3 Hz), 126.56 (d, Ar-CH, J = 8.2 Hz), 101.00 (d, CHOO, J = 4.5
Hz), 26.99 (d, OOCHCH2P, J = 25.8 Hz), and 19.58 (d, Ar-CH3, J = 13.6
Synthesis of [C20H27O3P] (4b). Similar to the synthesis of 4a; see
Supporting Information S5ÀS6.
Hz) ppm. Anal. Calcd for [C32H36O2P2][Br]2 0.25[H2O]: C, 56.24; H,
3
Synthesis of [C22H31O3P] (4c). Similar to the synthesis of 4a; see
Supporting Information S6.
5.46, Found: C, 56.27; H, 5.35. MS (DART; m/z+):257.1[C32H36O2P2]2+.
MS (ESI, methanol/water; m/z+): 257.1 [C16H18OP]+, 289.1
[C19H20O2P]+.
Synthesis of [C20H27O2P] (5a). A Schlenk flask was charged with
LiAlH4 (0.204 g, 5.37 mmol) and dry ether (10 mL) and then cooled to
0 °C. A solution of 4a (0.603 g, 1.74 mmol) in ether (5 mL) was added
slowly. Gas evolved. After stirring for 45 min, the solution turned yellow
and gas evolution ceased. The solution was stirred overnight and cooled
to 0 °C, and degassed H2O (0.20 mL) was added slowly. Gas evolved,
and the solution turned clear with a gray precipitate. The solution was
filtered, and the solvent was removed under reduced pressure, yielding a
clear oil. Crude yield: 59.6% (0.343 g). Analytically pure samples were
obtained from silica gel chromatography, eluting with 10:1 hexanes/
ethyl acetate. 1H NMR (300 MHz, CD2Cl2) δ: 7.33 (app t, 4H, Ar-H, J =
7.7 Hz), 7.15 (d, 4H, Ar-H, J = 7.7 Hz), 4.54 (q, 1H, O-CH-O, J =
5.9 Hz), 3.65À3.54 (m, 2H, O-CH2), 3.50À3.38 (m, 2H, O-CH2), 2.40
(d, 2H, P-CH2-CH, J = 5.9 Hz), 2.34 (s, 6H, Ar-CH3), and 1.12 (t, 6H,
CH3, J = 7.0 Hz) ppm. 31P{1H} NMR (161 MHz, CD2Cl2) δ: À24.40
(s) ppm. 13C NMR (100 MHz, CD2Cl2) δ: 138.71 (s, Ar-C), 135.81 (d,
Ar-CP, J = 12.1 Hz), 132.88 (d, Ar-CH, J = 19.7 Hz), 129.58 (d, Ar-CH,
J = m 7.1 Hz), 101.44 (d, CHOO, J = 21.3 Hz), 61.40 (s, CH2O), 34.18
(d, OOCHCH2P, J = 14.1 Hz), 21.13 (s, Ar-CH3), and 15.20 (s, CH3)
ppm. Anal. Calcd for [C20H27O2P]: C, 72.70; H, 8.24. Found: C, 72.36;
H, 8.23. MS (DART; m/z+): 331.2 [C20H28O2P]+, 213.1[C14H14P]+.
Synthesis of [C20H27O2P] (5b). Similar to the synthesis of 5a; see
Supporting Information S6ÀS7.
Synthesis of [C36H44O2P2][Br]2 (2c). Similar to the synthesis of
2b; see Supporting Information S2ÀS3.
Synthesis of [C50H52N4P2Fe][B(C6H5)4]2 (3a). A vial was
charged with 2a (0.235 g, 0.324 mmol) and CH3CN (4 mL). A yellow
solution of [Fe(H2O)6][BF4]2 (0.164 g, 0.485 mmol) in CH3CN
(2 mL) was added to the white slurry, followed by NaOMe (0.035 g,
0.647 mmol) in MeOH (1 mL). The color of the solution changed from
yellow to colorless. After 20 min of stirring, (1S,2S)-(À)-1,2-dipheny-
lethylenediamine (0.069 g, 0.323 mmol) in CH3CN (0.5 mL) was added
over 5 min, and the solution turned deep purple. After 48 h, the mixture
was filtered to remove a white precipitate. The solvent was removed
under reduced pressure to give a red-pink residue. The residue was
dissolved in a minimum of MeOH (∼1 mL) and added to a solution of
NaBPh4 (0.250 g, 0.658 mmol) in MeOH (1 mL) to cause the
precipitation of a pale pink solid. The product was filtered and washed
with MeOH (2 Â 5 mL) and dried under vacuum. Yield: 26.3%
(0.120 g). Crystals suitable for X-ray diffaction studies were obtained
1
by slow diffusion of Et2O into CH3CN/MeOH (1:5 by volume). H
NMR (400 MHz, CD3CN) δ: 8.20À8.09 (m, 2H, HCdN), 7.75À6.80
(m, 66H, Ar-H), 5.42 (s, 2H, HC-N), 4.30À4.15 (m, 2H, HC-P),
4.03À3.90 (m, 2H, HC-P), 2.40 (s, 6H, Ar-CH3), 2.35 (s, 6H, Ar-CH3),
and 1.53 (s, 6H, CH3CN) ppm. 31P{1H} NMR (161 MHz; CD3CN) δ:
72.42 (s) ppm. 13C NMR (CD3CN, 100 MHz) δ: 177.83 (CdNH),
164.08 (B-C), 143.05 (Ar-CP), 141.82 (Ar-CP), 136.01 (BÀC-CH),
Synthesis of [C22H31O2P] (5c). Similar to the synthesis of 5a; see
Supporting Information S7.
4428
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
Alternative Synthesis of [C32H36O2P2][Br]2 (2a). A Schlenk
flask was charged with 5a (0.754 g, 2.28 mmol) and dry THF (5 mL). An
excess of 48% HBr (0.60 mL, 7.20 mmol) was added. Over 5 min a white
precipitate formed. The solution was stirred overnight and then cooled
in a freezer (À40 °C) for 2 h. The white solid was filtered, washed with
H2O (10 mL) and ether (2 Â 10 mL), and dried under high vacuum to
yield a white powder. Yield: 86.7% (0.683 g). All spectroscopic data were
identical to 2a.
Synthesis of [C32H36O2P2][BF4]2 (6a). A Schlenk flask was
charged with 5a (0.823 g, 2.09 mmol) and dry THF (5 mL). An excess
of 48% HBr (0.70 mL, 8.40 mmol) was added. Over 30 min a white
precipitate formed. The solution was stirred overnight and then cooled
in a freezer (À40 °C) for 5 h. The white solid was filtered, washed with
H2O (10 mL) and ether (2 Â 10 mL), and dried under high vacuum to
yield a white powder. Crystals suitable for X-ray diffraction studies
were obtained from slow diffusion of ether into MeOH. Yield: 93.5%
(0.804 g). All spectroscopic data were identical to 2a. Anal. Calcd for
14.87 (s, CH3) ppm. Anal. Calcd for [C20H21F6O2P]: C, 54.80; H, 4.83.
Found: C, 54.69; H, 4.76. MS (EI, dichloromethane; m/z+): 438.1
[C20H21F6O2P]+, 392.1 [C18H15F6OP]+, 322.0 [C14H9F6P]+.
Synthesis of [C22H19F12O2P] (5e). Similar to the synthesis of 5d;
see Supporting Information S9ÀS10.
Synthesis of [C50H40F16FeN4P2][BF4]2 (3d). A Schlenk flask
was charged with 5d (0.284 g, 0.646 mmol), [Fe(H2O)6][BF4]2
(0.109 g, 0.324 mmol), and CH3CN (4 mL). The yellow solution was
left for 1 h, and it turned darker in color. (1S,2S)-(À)-1,2-Dipheny-
lethylenediamine (0.069 mg, 0.323 mmol) in CH3CN (2 mL) was
added, and the solution was left for 6 h. The solution slowly turned pink.
A small amount of 48% HBF4 (0.1 mL, 1.2 mmol) was added, and the
solution was stirred for 6 h. The solution turned cloudy and colorless.
The solvent was removed under reduced pressure to give a pale yellow
residue. The residue was dissolved in CH3CN (4 mL), and a slurry of
NaOMe (0.065 g, 1.2 mmol) in CH3CN (2 mL) was added, along with 4
drops of MeOH. The solution turned dark red over 20 min. The solution
was stirred overnight and then filtered. The solvent was removed under
reduced pressure, and the red residue was dissolved in a minimum of
CH2Cl2 (∼5 mL) and then filtered. The solvent was removed under
reduced pressure, and the red solid was washed with 5:1 ether/THF
(2 Â 5 mL). Yield: 44.2% (0.174 g). 1H NMR (400 MHz, CD3CN) δ:
8.30À8.15 (m, 2H, HCdN), 7.85 (app d, 4H, Ar-H), 7.67 (app d, 4H,
Ar-H), 7.51 (app q, 8H, Ar-H), 7.43 (s, 10H, dpen Ar-H), 5.48 (s, 2H,
HC-N), 4.54À4.40 (m, 2H, HC-P), 4.34À4.20 (m, 2H, HC-P), and
1.71 (s, 6H, CH3CN) ppm. 31P{1H} NMR (161 MHz, CD3CN) δ:
69.01 (s) ppm. 19F NMR (376 MHz, CD3CN) δ: À63.74 (s) and
À63.87 (s) ppm. 13C NMR (100 MHz, CD3CN) δ: 178.53 (CdNH),
137.92 (CH3CN), 137.3 (CF3), 134.78 (Ar-CH), 133.88 (Ar-C), 132.72
(Ar-CH), 130.27 (Ar-C), 132.24 (CF3), 129.96 (Ar-C), 129.61 (Ar-
CH), 126.31 (Ar-CH), 126.05 (Ar-CH), 125.32 (Ar-CP), 78.72 (CH-
NH), 45.31 (CH2P), and 4.53 (CH3CN) ppm. Anal. Calcd for
[C50H40F16FeN4P2][BF4]2: C, 49.38; H, 3.31; N, 4.61. Found: C,
44.70; H, 3.53; N, 4.06. MS (ESI, dichloromethane; m/z+): 480.2
[C46H34F12FeN2P2]2+.
[C32H36O2P2][BF4]2 1.5[H2O]: C, 53.74; H, 5.50. Found: C, 53.89;
3
H, 5.17.
Alternative Synthesis of [C32H36O2P2][Br]2 (2b). Similar to
the alternative synthesis of 2a; see Supporting Information S8.
Synthesis of [C32H36O2P2][BF4]2 (6b). Similar to the synthesis
of 6a; see Supporting Information S8.
Alternative Synthesis of [C36H44O2P2][Br]2 (2c). Similar to
the alternative synthesis of 2a; see Supporting Information S8.
Synthesis of [C36H44O2P2][BF4]2 (6c). Similar to the synthesis of
6a; see Supporting Information S8.
Alternative Synthesis of [C50H52N4P2Fe][B(C6H5)4]2 (3a). A
vial was charged with 6a (0.237 g, 0.324 mmol) and CH3CN (4 mL). A
yellow solution of [Fe(H2O)6][BF4]2 (0.109 g, 0.324 mmol) in CH3CN
(2 mL) was added to the white slurry, followed by NaOMe (0.035 g,
0.647 mmol) in CH3CN (1 mL). The color of the solution changed from
yellow to orange. After 20 min of stirring, (1S,2S)-(À)-1,2-dipheny-
lethylenediamine (0.069 g, 0.323 mmol) in CH3CN (0.5 mL) was added
over 5 min, and the solution turned purple. After 48 h, the mixture was
filtered to remove a white precipitate. The solvent was removed under
reduced pressure to give an orange-red residue. The residue was
dissolved in MeOH (∼2 mL) and added to a solution of NaBPh4
(0.250 g, 0.658 mmol) in MeOH (1 mL) to cause precipitation of a pale
pink solid. The product was filtered and washed with MeOH (2 Â 5 mL)
and dried under vacuum. Yield: 39.9% (0.182 g). All spectroscopic data
were identical to 3a.
Synthesis of [C54H36F24FeN4P2][BF4]2 (3e). Similar to the
synthesis of 3d; see Supporting Information S10ÀS11.
Synthesis of [C53H26BrFeN2OP2][B(C6H5)4] (1a). The crude
reaction mixture of 2a (either starting with the bromide or tetrafluor-
oborate phosphonium dimer) was filtered. The solvent was removed
under reduced pressure, and acetone (9 mL) was added, along with an
excess of KBr. The solution was stirred under a CO headspace for 3 days.
The resulting yellow-brown solution was evaporated to dryness, dis-
solved in a minimum of MeOH (∼2 mL), filtered, and added to a
solution of NaBPh4 (0.125 g, 0.329 mmol) in MeOH (1 mL) to cause
the formation of a yellow precipitate. The precipitate was washed with
ether (2 Â 5 mL) and MeOH (5 mL) and dried under high vacuum.
Yield: 28.5% (0.108 g). 1H NMR (400 MHz, (CD3)2CO) δ: 7.81À7.72
(m, 1H, HCdN), 7.67À7.58 (m, 1H, HCdN), 7.32À6.71 (m, 46H, Ar-
H), 5.68À5.61 (m, 1H, HC-N), 5.18À5.11 (m, 1H, HC-N), 4.09À3.99
(m, 1H, HC-P), 3.82À3.70 (m, 2H, HC-P), 3.62À3.52 (m, 1H, HC-P),
2.34 (s, 3H, Ar-CH3), 2.33 (s, 3H, Ar-CH3), 2.30 (s, 3H, Ar-CH3), and
2.26 (s, 3H, Ar-CH3) ppm. 31P{1H} NMR (161 MHz, (CD3)2CO) δ:
65.96 (d, J = 39.5 Hz) and 67.93 (d, J = 39.5 Hz) ppm. 13C NMR
(100 MHz, (CD3)2CO) δ: 175.00 (CdNH), 174.74 (CdNH), 164.04
(B-C), 142.39 (Ar-CP), 141.95 (Ar-CP), 141.25 (Ar-CP), 140.94 (Ar-
CP), 136.12 (B-C-CH), 134.96 (Ar-C), 134.86 (Ar-C), 134.27 (Ar-C),
134.10 (Ar-C), 133.73 (Ar-CH), 133.62 (Ar-CH), 133.57 (Ar-CH),
133.48 (Ar-CH), 131.34 (Ar-C), 131.25 (Ar-C), 129.82 (Ar-
CH),129.75 (Ar-CH), 129.48 (Ar-CH), 129.43 (Ar-CH), 129.40
(Ar-CH), 129.38 (Ar-CH), 129.33 (Ar-CH), 129.31 (Ar-CH), 129.29
(Ar-CH), 128.99 (Ar-CH), 128.89 (Ar-CH), 128.32 (Ar-CH),
128.21(Ar-CH), 125.06 (B-C-CH-CH), 121.30 (B-C-CH-CH-CH),
80.24 (CH-NH), 77.60 (CH-NH), 47.56 (CH2-P), 47.27 (CH2-P),
Alternative Synthesis of [C50H52N4P2Fe][B(C6H5)4]2 (3b).
Similar to the alternative synthesis of 3a; see Supporting Information
S8ÀS9.
Alternative Synthesis of [C54H60N4P2Fe][B(C6H5)4]2 (3c).
Similar to the alternative synthesis of 3a; see Supporting Information S9.
Synthesis of [C20H21F6O2P] (5d). A Schlenk flask was charged
with KH (0.079 g, 2.3 mmol) and dry THF (6 mL) and then cooled
to À40 °C. Di(para-trifluoromethylphenyl)phosphine (0.473 g, 1.5 mmol)
was added, and the mixture was stirred for 2 h, until the solution was dark
red in color. Bromoacetaldehyde diethyl acetal (0.23 mL, 1.5 mmol) was
added over 10 min. The solution was stirred for 30 min and then warmed to
room temperature. The solution turned yellow-brown, and the solvent was
removed under reduced pressure. The dark brown residue was dissolved in
ether, filtered, and dried under high vacuum to yield a brownish-red oil.
Yield: 72.0% (0.463 g). 1H NMR (400 MHz, CD2Cl2) δ: 7.60À7.57 (m,
4H, Ar-H), 4.64 (t, 1H, O-CH-O, J= 5.5 Hz), 3.65À3.29 (m, 4H, O-CH2),
2.95 (dd, 2H, P-CH2-CH, J= 5.8, 0.75 Hz), and 1.10 (t, 3H, CH3, J = 7.0
Hz) ppm. 31P{1H} NMR (161 MHz, CD2Cl2) δ: À22.03 (s) ppm. 19
F
NMR (376 MHz, CD2Cl2) δ: À63.20 (s) ppm. 13C NMR (100 MHz,
CD2Cl2) δ: 143.6 (app d, Ar-C), 133.12 (d Ar-CH, J = 19.5 Hz,), 130.43
(d, Ar-CP, J = 32.4 Hz), 124.99 (app sept, Ar-CH), 100.37 (d, CHOO, J =
21.1 Hz), 61.70 (s, CH2O), 33.75 (d, OOCHCH2P, J = 14.9 Hz), and
4429
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
and 20.50 (Ar-CH3) ppm. IR (KBr): 1972 cmÀ1 (νCtO). Anal. Calcd
for [C53H26BrFeN2OP2][B(C6H5)4]: C, 72.71; H, 5.68; N, 2.39.
Found: C, 70.72; H, 5.99; N, 2.11. MS (ESI, dichloromethane; m/z+):
851.2 [C47H46BrFeN2OP2]+.
Synthesis of [C53H26BrFeN2OP2][B(C6H5)4] (1b). Similar to
the synthesis of 1a; see Supporting Information S11ÀS12.
Synthesis of [C51H54BrFeN2OP2][B(C6H5)4] (1c). Similar to
the synthesis of 1a; see Supporting Information S12.
Synthesis of [C47H34BrF12FeN2OP2][BF4] (1d). Similar to the
synthesis of 1a; see Supporting Information S13.
Synthesis of [C51H30F24BrFeN2OP2][BF4] (1e). Similar to the
synthesis of 1a; see Supporting Information S13ÀS14.
(16) Casey, C. P.; Guan, H. J. Am. Chem. Soc. 2007, 129, 5816–5817.
(17) Langer, R.; Leitus, G.; Ben-David, Y.; Milstein, D. Angew.
Chem., Int. Ed. 2011, 50, 2120–2124.
(18) Kandepi, V. V. K. M.; Cardoso, J. M. S.; Peris, E.; Royo, B.
Organometallics 2010, 29, 2777–2782.
(19) Naik, A.; Maji, T.; Reiser, O. Chem. Commun. 2010, 46,
4475–4477.
(20) Enthaler, S.; Erre, G.; Tse, M. K.; Junge, K.; Beller, M.
Tetrahedron Lett. 2006, 47, 8095–8099.
(21) Buchard, A.; Heuclin, H.; Auffrant, A.; Le Goff, X. F.; Le Floch,
P. Dalton Trans. 2009, 1659–1667.
(22) Enthaler, S.; Hagemann, B.; Erre, G.; Junge, K.; Beller, M.
Chem. Asian J. 2006, 1, 598–604.
(23) Mashima, K.; Abe, T.; Tani, K. Chem. Lett. 1998, 27, 1199–1200.
(24) Sui-Seng, C.; Freutel, F.; Lough, A. J.; Morris, R. H. Angew.
Chem., Int. Ed. 2008, 47, 940–943.
(25) Zhou, S.; Fleischer, S.; Junge, K.; Das, S.; Addis, D.; Beller, M.
Angew. Chem., Int. Ed. 2010, 49, 8121–8125.
(26) Blaser, H.-U.; Studer, M. Acc. Chem. Res. 2007, 40, 1348–1356.
(27) Klingler, F. D. Acc. Chem. Res. 2007, 40, 1367–1376.
(28) Ikariya, T.; Blacker, A. J. Acc. Chem. Res. 2007, 40, 1300–
1308.
(29) Noyori, R.; Ohkuma, T. Angew. Chem., Int. Ed. 2001, 40, 40–73.
(30) Dong, Z.-R.; Li, Y.-Y.; Chen, J.-S.; Li, B.-Z.; Xing, Y.; Gao, J.-X.
Org. Lett. 2005, 7, 1043–1045.
’ ASSOCIATED CONTENT
S
Supporting Information. Extended experimental sec-
b
tion. Crystal structures and crystallographic data for 3b, 4b, 6a,
6c, and 7 (including CIF files). Sample calculations for Tolman
cone angles. Catalytic and GC conditions for ATH of acetophe-
none and TH of benzophenone. Catalytic profiles for TH of
benzophenone for 1a and 1c. Comparison of the linear portions
of the catalytic profiles for 1a and 1c. This material is available
free of charge via the Internet at http://pubs.acs.org.
(31) Li, Y.-Y.; Zhang, X.-Q.; Dong, Z.-R.; Shen, W.-Y.; Chen, G.;
’ AUTHOR INFORMATION
Gao, J.-X. Org. Lett. 2006, 8, 5565–5567.
(32) Gao, J.-X.; Ikariya, T.; Noyori, R. Organometallics 1996,
15, 1087–1089.
Corresponding Author
*E-mail: rmorris@chem.utoronto.ca.
(33) Mikhailine, A.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc.
2009, 131, 1394–1395.
(34) Mikhailine, A. A.; Morris, R. H. Inorg. Chem. 2010, 49, 11039–
11044.
’ ACKNOWLEDGMENT
(35) Lagaditis, P. O.; Lough, A. J.; Morris, R. H. Inorg. Chem. 2010,
49, 10057–10066.
We thank NSERC Canada for a Discovery grant to R.H.M
(36) Guo, R.; Chen, X.; Elpelt, C.; Song, D.; Morris, R. H. Org. Lett.
2005, 7, 1757–1759.
’ REFERENCES
(37) Moser, W. R.; Papile, C. J.; Brannon, D. A.; Duwell, R. A.;
Weininger, S. J. J. Mol. Catal. 1987, 41, 271–292.
(38) Palo, D. R.; Erkey, C. Organometallics 1999, 19, 81–86.
(39) Wang, X.; Lane, B. S.; Sames, D. J. Am. Chem. Soc. 2005,
127, 4996–4997.
(40) Muller, J. G.; Acquaye, J. H.; Takeuchi, K. J. Inorg. Chem. 1992,
31, 4552–4557.
(41) Casalnuovo, A. L.; RajanBabu, T. V.; Ayers, T. A.; Warren, T. H.
J. Am. Chem. Soc. 1994, 116, 9869–9882.
(1) Gaillard, S.; Renaud, J.-L. ChemSusChem 2008, 1, 505–509.
(2) Oldenburg, P. D.; Que, J. L. Catal. Today 2006, 117, 15–21.
(3) Yeung, H.-L.; Sham, K.-C.; Tsang, C.-S.; Lau, T.-C.; Kwong,
H.-L. Chem. Commun. 2008, 3801–3803.
(4) Bruijnincx, P. C. A.; Buurmans, I. L. C.; Gosiewska, S.; Moelands,
M. A. H.; Lutz, M.; Spek, A. L.; van Koten, G.; Klein Gebbink, R. J. M.
Chem.—Eur. J. 2008, 14, 1228–1237.
(5) Suzuki, K.; Oldenburg, P. D.; Que, L. Angew. Chem., Int. Ed.
2008, 47, 1887–1889.
(42) Mikhailine, A. A.; Lagaditis, P. O.; Sues, P. E.; Lough, A. J.;
Morris, R. H. J. Organomet. Chem. 2010, 695, 1824–1830.
(43) Lagaditis, P. O.; Mikhailine, A. A.; Lough, A. J.; Morris, R. H.
Inorg. Chem. 2009, 49, 1094–1102.
(44) Busacca, C. A.; Lorenz, J. C.; Grinberg, N.; Haddad, N.;
Hrapchak, M.; Latli, B.; Lee, H.; Sabila, P.; Saha, A.; Sarvestani, M.;
Shen, S.; Varsolona, R.; Wei, X.; Senanayake, C. H. Org. Lett. 2005,
7, 4277–4280.
(45) Matt, D.; Ziessel, R.; De Cian, A.; Fischer, J. New J. Chem. 1996,
20, 1257–1263.
(46) Mikhailine, A. A.; Kim, E.; Dingels, C.; Lough, A. J.; Morris,
R. H. Inorg. Chem. 2008, 47, 6587–6589.
(47) Mizoshita, N.; Tani, T.; Inagaki, S. Chem. Soc. Rev. 2011,
(6) Gꢀomez, L.; Garcia-Bosch, I.; Company, A.; Benet-Buchholz, J.;
Polo, A.; Sala, X.; Ribas, X.; Costas, M. Angew. Chem., Int. Ed. 2009,
48, 5720–5723.
(7) Muthupandi, P.; Alamsetti, S. K.; Sekar, G. Chem. Commun.
2009, 3288–3290.
(8) Rose, E.; Ren, Q.-Z.; Andrioletti, B. Chem.—Eur. J. 2004, 10,
224–230.
(9) Ellis, W. C.; Tran, C. T.; Denardo, M. A.; Fischer, A.; Ryabov,
A. D.; Collins, T. J. J. Am. Chem. Soc. 2009, 131, 18052–18053.
(10) Ellis, W. C.; Tran, C. T.; Roy, R.; Rusten, M.; Fischer, A.;
Ryabov, A. D.; Blumberg, B.; Collins, T. J. J. Am. Chem. Soc. 2010,
132, 9774–9781.
(11) Popescu, D.-L.; Chanda, A.; Stadler, M. J.; Mondal, S.; Tehranchi,
J.; Ryabov, A. D.; Collins, T. J. J. Am. Chem. Soc. 2008, 130,
12260–12261.
(12) Morris, R. H. Chem. Soc. Rev. 2009, 38, 2282–2291.
(13) Junge, K.; Schroder, K.; Beller, M. Chem. Commun. 2011,
47, 4849–4859.
(14) Tondreau, A. M.; Lobkovsky, E.; Chirik, P. J. Org. Lett. 2008,
10, 2789–2792.
40, 789–800.
(48) Busch, D. H.; Stephenson, N. A. Coord. Chem. Rev. 1990,
100, 119–154.
(49) Battle, A. R.; Edwards, P. G.; Haigh, R.; Hibbs, D. E.; Li, D.;
Liddiard, S. M.; Newman, P. D. Organometallics 2006, 26, 377–386.
(50) Albers, T.; Edwards, P. G. Chem. Commun. 2007, 858–860.
(51) Doering, W. V. E.; Young, R. W. J. Am. Chem. Soc. 1950, 72,
631–631.
(15) Tondreau, A. M.; Darmon, J. M.; Wile, B. M.; Floyd, S. K.;
Lobkovsky, E.; Chirik, P. J. Organometallics 2009, 28, 3928–3940.
(52) Tolman, C. A. Chem. Rev. 1977, 77, 313–348.
4430
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431

Organometallics
ARTICLE
(53) Fernandez, A. L.; Reyes, C.; Prock, A.; Giering, W. P. J. Chem.
Soc., Perkin Trans. 2 2000, 1033–1041.
(54) Bunten, K. A.; Moreno, C.; Poe, A. J. Dalton Trans.
2005, 1416–1421.
(55) Lagaditis, P. O.; Lough, A. J.; Morris, R. H. J. Am. Chem. Soc.
2011, 133, 9662–9665.
(56) Hashiguchi, S.; Fujii, A.; Takehara, J.; Ikariya, T.; Noyori, R.
J. Am. Chem. Soc. 1995, 117, 7562–7563.
(57) Takehara, J.; Hashiguchi, S.; Fujii, A.; Inoue, S.-I.; Ikariya, T.;
Noyori, R. Chem. Commun. 1996, 233–234.
(58) Alonso, D. A.; Nordin, S. J. M.; Roth, P.; Tarnai, T.; Andersson,
P. G.; Thommen, M.; Pittelkow, U. J. Org. Chem. 2000, 65, 3116–3122.
(59) Palmer, M. J.; Kenny, J. A.; Walsgrove, T.; Kawamoto, A. M.;
Wills, M. J. Chem. Soc. Perkin Trans. 1 2002, 416–427.
(60) Hannedouche, J.; Clarkson, G. J.; Wills, M. J. Am. Chem. Soc.
2004, 126, 986–987.
(61) Murata, K.; Ikariya, T.; Noyori, R. J. Org. Chem. 1999, 64,
2186–2187.
(62) Bubert, C.; Blacker, J.; Brown, S. M.; Crosby, J.; Fitzjohn, S.;
Muxworthy, J. P.; Thorpe, T.; Williams, J. M. J. Tetrahedron Lett. 2001,
42, 4037–4039.
(63) Guo, R.; Elpelt, C.; Chen, X.; Song, D.; Morris, R. H. Chem.
Commun. 2005, 3050–3052.
(64) Bligaard, T.; Nørskov, J. K.; Dahl, S.; Matthiesen, J.; Christensen,
C. H.; Sehested, J. J. Catal. 2004, 224, 206–217.
(65) Cheng, J.; Hu, P. J. Am. Chem. Soc. 2008, 130, 10868–10869.
(66) Jacobsen, C. J. H.; Dahl, S.; Clausen, B. S.; Bahn, S.; Logadottir,
A.; Nørskov, J. K. J. Am. Chem. Soc. 2001, 123, 8404–8405.
(67) RajanBabu, T. V.; Ayers, T. A.; Halliday, G. A.; You, K. K.;
Calabrese, J. C. J. Org. Chem. 1997, 62, 6012–6028.
(68) RajanBabu, T. V.; Radetich, B.; You, K. K.; Ayers, T. A.;
Casalnuovo, A. L.; Calabrese, J. C. J. Org. Chem. 1999, 64, 3429–3447.
(69) Yan, Y.-Y.; RajanBabu, T. V. J. Org. Chem. 2001, 66, 3277–3283.
(70) Snelders, D. J. M.; van Koten, G.; KleinGebbink, R. J. M. Chem.—
Eur. J. 2011, 17, 42–57.
(71) Mikhailine, A. A.; Lagaditis, P. O.; Sues, P. E.; Lough, A. J.;
Morris, R. H. J. Organomet. Chem. 2010, 695, 1824À1830.
(72) Marcꢀo, A.; Compa~nꢀo, R.; Rubio, R.; Casals, I. Microchim. Acta
2003, 142, 13–19.
4431
dx.doi.org/10.1021/om2005172 |Organometallics 2011, 30, 4418–4431