Catalytic nitrile hydration with [Ru(η6- p -cymene)Cl 2(PR2R′)] complexes: Secondary coordination sphere effects with phosphine oxide and phosphinite ligands

Article abstract of DOI:10.1021/om400380j

The rates of nitrile hydration reactions were investigated using [Ru(η⁶-p-cymene)Cl₂(PR₂R′)] complexes as homogeneous catalysts, where PR₂R′ = PMe ₂(CH₂P(O)Me₂), PMe₂(CH ₂CH₂P(O)Me₂), PPh₂(CH ₂P(O)Ph₂), PPh₂(CH₂CH ₂P(O)Ph₂), PMe₂OH, P(OEt)₂OH. These catalysts were studied because the rate of the nitrile-to-amide hydration reaction was hypothesized to be affected by the position of the hydrogen bond accepting group in the secondary coordination sphere of the catalyst. Experiments showed that the rate of nitrile hydration was fastest when using [Ru(η⁶-p-cymene)Cl₂PMe₂OH]: i.e., the catalyst with the hydrogen bond accepting group capable of forming the most stable ring in the transition state of the rate-limiting step. This catalyst is also active at pH 3.5 and at low temperatures - conditions where α-hydroxynitriles (cyanohydrins) produce less cyanide, a known poison for organometallic nitrile hydration catalysts. The [Ru(η⁶-p-cymene) Cl₂PMe₂OH] catalyst completely converts the cyanohydrins glycolonitrile and lactonitrile to their corresponding α-hydroxyamides faster than previously investigated catalysts. [Ru(η⁶-p-cymene) Cl₂PMe₂OH] is not, however, a good catalyst for acetone cyanohydrin hydration, because it is susceptible to cyanide poisoning. Protecting the -OH group of acetone cyanohydrin was shown to be an effective way to prevent cyanide poisoning, resulting in quantitative hydration of acetone cyanohydrin acetate.

Full text of DOI:10.1021/om400380j

Article
pubs.acs.org/Organometallics
Catalytic Nitrile Hydration with [Ru(η6‑p‑cymene)Cl2(PR2R′)]
Complexes: Secondary Coordination Sphere Effects with Phosphine
Oxide and Phosphinite Ligands
Spring Melody M. Knapp,†,§ Tobias J. Sherbow,† Robert B. Yelle,† J. Jerrick Juliette,‡
and David R. Tyler*,†
†Department of Chemistry, University of Oregon, Eugene, Oregon 97403, United States
‡Dow Advanced Materials-Performance Monomers, The Dow Chemical Company, Deer Park, Texas 77536, United States
S
* Supporting Information
ABSTRACT: The rates of nitrile hydration reactions were investigated
using [Ru(η6-p-cymene)Cl2(PR2R′)] complexes as homogeneous catalysts,
where PR2R′ = PMe2(CH2P(O)Me2), PMe2(CH2CH2P(O)Me2),
PPh2(CH2P(O)Ph2), PPh2(CH2CH2P(O)Ph2), PMe2OH, P(OEt)2OH.
These catalysts were studied because the rate of the nitrile-to-amide
hydration reaction was hypothesized to be affected by the position of the
hydrogen bond accepting group in the secondary coordination sphere of
the catalyst. Experiments showed that the rate of nitrile hydration was
fastest when using [Ru(η6-p-cymene)Cl2PMe2OH]: i.e., the catalyst with the hydrogen bond accepting group capable of forming
the most stable ring in the transition state of the rate-limiting step. This catalyst is also active at pH 3.5 and at low
temperaturesconditions where α-hydroxynitriles (cyanohydrins) produce less cyanide, a known poison for organometallic
nitrile hydration catalysts. The [Ru(η6-p-cymene)Cl2PMe2OH] catalyst completely converts the cyanohydrins glycolonitrile and
lactonitrile to their corresponding α-hydroxyamides faster than previously investigated catalysts. [Ru(η6-p-cymene)Cl2PMe2OH]
is not, however, a good catalyst for acetone cyanohydrin hydration, because it is susceptible to cyanide poisoning. Protecting the
−OH group of acetone cyanohydrin was shown to be an effective way to prevent cyanide poisoning, resulting in quantitative
hydration of acetone cyanohydrin acetate.
molecule on the coordinated nitrile.9,10 However, decreasing
the pH of the solution will decrease the amount of hydroxide
INTRODUCTION
■
The hydration of acetone cyanohydrin to α-hydroxyamide
present in solution, potentially leading to a corresponding
products is an important step in the production of acrylic
monomers.1 Acetone cyanohydrin hydration is carried out
decrease in the rate of the hydration reaction. To increase the
rate of hydration, we hypothesized that incorporation of a
industrially using sulfuric acid, but considerable amounts of
hydrogen bond accepting ligand in the secondary coordination
ammonium hydrogen sulfate byproduct are produced, which
requires significant energy and effort to recycle.2 A potentially
desirable alternative process would be catalytic hydration using
sphere (e.g., Figure 1) would activate water for nitrile
hydration, even under acidic conditions, meaning that these
catalysts might be sufficiently fast under acidic conditions to
hydrate cyanohydrins.
a homogeneous catalyst. Catalytic hydration of cyanohydrins is
challenging because cyanohydrins degrade readily under basic
conditions and at high temperatures to produce HCN.3−5 The
cyanide thus produced reacts irreversibly with organometallic
catalysts, poisoning the catalysts.6−8 Cyanohydrins can be
stabilized under acidic conditions and at low temperatures,
although a survey of the catalytic nitrile hydration literature
shows that few nitrile hydration catalysts are functional under
these conditions.9 For example, platinum phosphinito,6
molybdocene,6 and ruthenium7,8 nitrile hydration catalysts
have shown limited success for cyanohydrin hydration,
especially for the hydration of bulky cyanohydrins.
On paper, complete cyanohydrin hydration would be
possible with a homogeneous catalyst if the catalyst was fast
under conditions that stabilize the cyanohydrinnamely, low
pH and low temperatures. The rate-determining step of nitrile
hydration is typically attack by an external water or hydroxide
Figure 1. Proposed structure showing how hydrogen bonding of a
water molecule to the secondary coordination sphere of a catalyst will
activate the water molecule for nitrile hydration.
Received: May 2, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/om400380j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Co. glovebox or on a Schlenk line using N2. HPLC grade THF,
dichloromethane, hexanes, acetonitrile, and diethyl ether (Burdick and
Jackson) were dried and deoxygenated by passing them through
commercial columns of CuO, followed by alumina, under an argon
atmosphere. HPLC grade water was degassed by sparging with N2.
Chloroform was distilled under N2 from CaH2 and degassed by three
freeze−pump−thaw cycles. Petroleum ether (35−60 °C; Mackron
Chemicals) was degassed by three freeze−pump−thaw cycles.
[RuCl2(η6-p-cymene)]2, 1,2-bis(diphenylphosphino)methane monox-
ide (dppmO), 1,2-bis(diphenylphosphino)ethane monoxide (dppeO),
1 , 2 - b i s ( d i m e t h y l p h o s p h i n o ) m e t h a n e , a n d 1 , 2 - b i s -
(dimethylphosphino)ethane were obtained from Strem. Acetone
cyanohydrin, lactonitrile, and glycolonitrile (55 wt % in H2O) were
obtained from Sigma Aldrich. Cyanohydrins were distilled prior to use.
All other commercially available reagents were used as received.
[Ru(η6-p-cymene)Cl2(dppeO)],14 [Ru(η6-p-cymene)Cl2(dppmO)],14
[Ru(η6-p-cymene)Cl2P(NMe2)3],15 dimethylphosphine oxide,16 dieth-
ylphosphinate,17 and 2-acetoxy-2-methylpropoanenitrile18 were syn-
thesized following literature procedures. The ligands 1,2-bis-
(dimethylphosphino)methane monoxide (dmpmO) and 1,2-bis-
(dimethylphosphino)ethane monoxide (dmpeO) were synthesized
following a modified literature procedure19 as described in the
Supporting Information.
Instrumentation. Nuclear magnetic resonance spectra were
recorded on a Varian Unity/Inova 300 MHz (1H, 300 MHz; 31P,
121 MHz) spectrometer or a 500 MHz (1H, 500 MHz; 31P, 202 MHz;
13C, 126 MHz) spectrometer. The 1H chemical shifts were referenced
to the solvent peak or TMS (0.00 ppm). The 13C chemical shifts were
referenced to the solvent peak, and the 31P chemical shifts were
referenced externally to H3PO4 (0.00 ppm). Elemental analyses were
conducted by Complete Analysis Laboratories, Inc. pH measurements
were taken using an Accumet AB15 Basic pH meter. All hydration
reaction samples were prepared in a glovebox under an atmosphere of
N2 in Wilmad 9 in. precision NMR tubes or in 1 dram screwcap vials
fitted with septum caps. Reactions carried out in the Wilmad 9 in.
NMR tubes were flame-sealed. Reaction tubes and vials were heated in
an oil bath. Cyanohydrin hydration reactions were performed using a
variety of reaction conditions. Representative procedures that gave the
best results are given below.
Secondary coordination sphere effects in nitrile hydration
reactions have been previously investigated by our group using
[Ru(η6-p-cymene)Cl2PR3] complexes, where R = NMe2, OMe,
Et.7,8 These studies showed that the catalyst with the best
hydrogen bond accepting ligand (R = NMe2) has the fastest
rate of nitrile hydration. Furthermore, we showed that [Ru(η6-
p-cymene)Cl2P(NMe2)3] operates by a general base catalysis
mechanism, meaning that the amine functionality on the ligand
does not simply act as a base and raise the concentration of
hydroxide in solution.8 Cadierno11 and Frost12 also investigated
the effect of amines in the secondary coordination sphere with
similar [Ru(η6-arene)Cl2(PR2R′)]-type complexes. Cadierno
and co-workers found that nitrile hydration reactions using the
amine-containing complexes [Ru(η6-arene)Cl2PPh2R], where R
= 2-C6H5CH2NHR′, 3-C6H5CH2NHR′, 4-C6H5CH2NHR′ and
R′ = iPr, tBu, were significantly faster than nitrile hydration
reactions with the corresponding triphenylphosphine complex
(R = Ph).11 However, addition of 1 equiv of PhCH2NHiPr or
PhCH2NHtBu to the [Ru(η6-arene)Cl2PPh3]-catalyzed reac-
tions increased the hydration rates to the same rates observed
for the aminophosphine-containing complexes. They concluded
that the rate enhancement observed was likely due to the
inherently basic nature of the amine group, which would
generate more hydroxide in solution.
Frost and co-workers investigated nitrile hydration with a
series of β-aminophosphine-derived [Ru(η6-toluene)Cl2PR3]
complexes, where PR3 = PTA-CHPhNHPh, PTA-CH(p-
C6H4OCH3)NHPh, PTA-CPh2NHPh. Nitrile hydration with
these complexes was slightly faster than that with the parent
[Ru(η6-toluene)Cl2PTA] catalyst.12 These complexes did not
show the expected enhanced activity that had previously been
observed for nitrile hydration catalysts containing amine groups
in the secondary coordination sphere. The explanation was
shown to be that the [Ru(η6-toluene)Cl2PR3] catalysts are
hemilabile. The amine group can therefore bond to the Ru
center; this prevents nitrile coordination and thereby decreases
the catalyst activity.
These results suggest there may be a limit to the distance
from the active site where a hydrogen bond accepting group
can be placed in order to increase the rate of nitrile hydration.
We decided, therefore, to investigate the effect of the hydrogen
bond accepting group position on the rate of hydration. In
order to limit any possibility of the catalyst acting as a base (as
was seen with the amine-containing ligands used by
Cadierno11) or coordination of the hydrogen bond accepting
group (as was seen with the amine-containing ligands used by
Frost12), a different hydrogen bond accepting group was used:
namely, phosphine oxides. Phosphine oxides are some of the
strongest hydrogen bond accepting groups, with an average
R3PO···H−OH hydrogen bond length of 1.846(4) Å (in
comparison to a trialkylamine and a dialkyl ether with average
donor···H−OH distances of 1.96(1) and 1.978(9) Å,
respectively).13 To investigate the effect of hydrogen bond
accepting group position on the rate of hydration, the [Ru(η6-p-
cymene)Cl2(PR2R′)] complexes were synthesized, where
PR2R′ = Ph2P(CH2P(O)Ph2) (dppmO), Me2P(CH2P(O)Me2)
(dmpmO), Ph2P(CH2CH2P(O)Ph2) (dppeO), Me2P-
(CH2CH2P(O)Me2) (dmpeO), Me2P(OH), (EtO)2P(OH).
The results of our study are reported herein.
Preparation of [Ru(η6-p-cymene)Cl2(dmpmO)] (1-Me). [Ru-
(η6-p-cymene)Cl2]2 (0.083 g, 0.14 mmol) and dmpmO (0.11 g, 0.28
mmol) were dissolved in 10 mL of CH2Cl2, and then the solution was
stirred at room temperature for 4 h. The CH2Cl2 was removed under
vacuum, and then hexanes was added to the orange solid, which was
filtered over Celite. The solid was eluted off the Celite with CH2Cl2.
The product was precipitated from solution by addition of an excess of
light petroleum ether and then filtered to obtain an orange powder
1
(0.21 g, 73% yield). H NMR (500 MHz, CDCl3): δ 5.48 (q, J = 6.2
Hz, 4H), 2.84 (heptet, J = 7.0 Hz, 1H), 2.53 (dd, J = 11.3, 10.4 Hz,
2H), 2.07 (d, J = 3.8 Hz, 6H), 2.05 (s, 3H), 1.55 (dd, J = 12.8, 5.8 Hz,
6H), 1.23 (d, J = 6.9 Hz, 6H). 13C NMR (151 MHz, CDCl3): δ
107.43, 94.91, 89.55 (d, J = 4.8 Hz), 85.00 (d, J = 6.1 Hz), 30.74, 26.10
(dd, J = 57.3, 25.7 Hz), 22.11, 20.28 (d, J = 69.8 Hz), 18.36, 17.13 (d, J
= 31.7 Hz). 31P{1H} NMR (202 MHz, CDCl3): δ 37.93 (d, J = 21.9
Hz), 10.35 (d, J = 20.6 Hz) ppm. Anal. Found: C, 39.25; H, 6.06; P,
13.36. Calcd for C15H28Cl2OP2Ru: C, 39.51; H, 6.16; P, 13.52.
Preparation of [Ru(η6-p-cymene)Cl2(dmpeO)] (2-Me). [Ru(η6-
p-cymene)Cl2]2 (0.06 g, 0.098 mmol) and dmpeO (0.034g, 0.2 mmol)
were dissolved in 10 mL of CH2Cl2, and then the solution was stirred
at room temperature for 4 h. The CH2Cl2 was removed under vacuum,
and then hexanes was added to the orange solid, which was filtered
over Celite. The solid was eluted off the Celite with CH2Cl2. The solid
was precipitated from solution by addition of an excess of light
petroleum ether and then filtered to obtain an orange powder (0.04 g,
47% yield). 1H NMR (500 MHz, CDCl3): δ 5.50 (q, J = 5.9 Hz, 4H),
2.83 (heptet, J = 7.1 Hz, 1H), 2.29 (m, 2H), 2.10 (s, 3H), 2.00 (m,
2H), 1.60 (d, J = 10.5 Hz, 6H), 1.52 (d, J = 12.5 Hz, 6H), 1.24 (d, J =
6.9 Hz, 6H). 13C NMR (151 MHz, CDCl3): δ 107.46, 94.71, 89.45 (d,
J = 4.5 Hz), 85.12 (d, J = 5.7 Hz), 30.80, 25.95 (d, J = 66.3 Hz), 23.25
(dd, J = 27.7, 4.6 Hz), 22.20, 18.33, 16.36 (d, J = 68.5 Hz), 13.99 (d, J
EXPERIMENTAL SECTION
Materials and Methods. Unless stated otherwise, all manipu-
lations were carried out in either an N2-filled Vacuum Atmospheres
■
B
dx.doi.org/10.1021/om400380j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
= 32.3 Hz). 31P{1H} NMR (202 MHz, CDCl3): δ 42.37 (d, J = 37.1
Hz), 11.98 (d, J = 39.7 Hz) ppm. Anal. Found: C, 40.60; H, 6.51; P,
12.97. Calcd for C16H30Cl2OP2Ru: C, 40.69; H, 6.40; P, 13.12.
Preparation of [Ru(η6-p-cymene)Cl2(PMe2OH)] (3-Me). [Ru-
(η6-p-cymene)Cl2]2 (0.40 g, 0.65 mmol) and dimethylphosphine oxide
(0.26 g, 3.4 mmol) were dissolved in 10 mL of CH2Cl2, and then the
solution was stirred at room temperature for 12 h. The CH2Cl2 was
removed under vacuum, and then hexanes was added to the orange
solid, which was filtered over Celite. The solid was eluted off the Celite
with CH2Cl2. The solid was precipitated from solution by addition of
an excess of light petroleum ether and then filtered to obtain an orange
powder (0.42 g, 85% yield). 1H NMR (500 MHz, CDCl3): δ 5.50 (d, J
= 5.7 Hz, 2H), 5.45 (d, J = 5.4 Hz, 2H), 2.91−2.72 (m, 1H), 2.10 (s,
3H), 1.93 (d, J = 9.5 Hz, 6H), 1.63 (dd, J = 13.8, 3.5 Hz, 1H), 1.23 (d,
J = 6.9 Hz, 6H). 13C NMR (151 MHz, CDCl3): δ 107.12, 95.87, 89.41
(d, J = 5.0 Hz), 86.44 (d, J = 5.7 Hz), 30.79, 22.17 (d, J = 36.5 Hz),
22.14, 18.61. 31P{1H} NMR (202 MHz, CDCl3): δ 115.33 ppm. Anal.
Found: C, 37.45; H, 5.59; P, 7.93. Calcd for C12H21Cl2OPRu: C,
37.51; H, 5.51; P, 8.06.
spectroscopy by observing the disappearance of the methyl peak of
acetone cyanohydrin at 1.57 ppm (s, 6H, HO(CH3)2CCN) and the
appearance of the α-hydroxyisobutyramide resonance at 1.34 ppm (s,
HO(CH3)2CC(O)NH2). Specific details for the hydration of other
cyanohydrins are described in the Supporting Information.
Titration with KCN. [Ru(η6-p-cymene)Cl2(PMe2OH)] (0.087 g,
0.23 mmol) was dissolved in 5.5 mL of degassed D2O to form a 0.041
M stock solution, and KCN (0.11 g, 1.68 mmol) was dissolved in 1.0
mL of D2O to form a 1.68 M stock solution. In a 1 dram screwcap vial
fitted with a septum cap, aliquots of KCN dissolved in D2O (0−110
μL; 0−0.19 mmol, 0−50 mol %) were added to acetonitrile (19.5 μL,
0.37 mmol) and 0.45 mL of the [Ru(η6-p-cymene)Cl2(PMe2OH)]
stock solution (0.019 mmol, 5 mol %). Each solution was heated to
100 °C with stirring. Aliquots (0.1 mL) were removed periodically
using a gastight syringe and were combined in an NMR tube with 0.6
mL of D2O and 0.1 mL of a 3.8 mM NMe4PF6 in D2O internal
1
standard solution. The progress of the reaction was monitored by H
NMR spectroscopy by observing the disappearance of the acetonitrile
resonance at 2.01 ppm (s, CH3CN) and the appearance of acetamide
at 1.93 ppm (s, CH3C(O)ND2).
Preparation of [Ru(η6-p-cymene)Cl2(P(OH)(OEt)2)] (3-OEt).
[Ru(η6-p-cymene)Cl2]2 (0.23 g, 0.38 mmol) and diethylphosphine
oxide (0.13 g, 0.97 mmol) were dissolved in 10 mL of CH2Cl2, and
then the solution was stirred at room temperature for 12 h. The
CH2Cl2 was removed under vacuum and then hexanes were added to
the orangeish solid, which was filtered over Celite. The product was
eluted off the Celite with CH2Cl2, precipitated from solution by
addition of an excess of light petroleum ether, and then filtered to
Control Experiments for Ester Hydrolysis. [Ru(η6-p-cymene)-
Cl2(PMe2OH)] (0.014 g, 0.04 mmol) was dissolved in 1.5 mL of
degassed D2O to form a 0.025 M stock solution, and 3-(N-
morpholino)propanesulfonic acid (MOPS, 0.20 g, 0.97 mmol) was
dissolved in 3 mL of degassed D2O to form a 0.32 M stock solution.
The following solutions were made up in a 1 dram screwcap vial fitted
with a septum cap: (1) 610 μL of Ru(η6-p-cymene)Cl2(PMe2OH)]
stock solution (0.015 mmol, 5 mol %) was combined with ethyl
acetate (30 μL, 0.31 mmol) and 1.39 mL of D2O to form a pH 3.5
solution; (2) 610 μL of Ru(η6-p-cymene)Cl2(PMe2OH)] stock
solution (0.015 mmol, 5 mol %) was combined with ethyl acetate
(30 μL, 0.31 mmol), 200 μL of MOPS stock solution (0.065 mmol),
and 1.20 mL of D2O to form a pH 7.0 solution; (3) 1 M HCl was
added dropwise to 3 mL of D2O until pH 3.5 was measured and 2005
μL of this solution was combined with ethyl acetate (30 μL, 0.31
mmol) to form a pH 3.5 solution; (4) 200 μL of MOPS stock solution
(0.065 mmol) was combined with ethyl acetate (30 μL, 0.31 mmol),
and 1.80 mL of D2O to form a pH 7.0 solution. Each solution was
heated to 100 °C with stirring. Aliquots (0.1 mL) were removed
periodically using a gastight syringe and were combined in an NMR
tube with 0.6 mL of D2O. The progress of the reaction was monitored
by 1H NMR spectroscopy by observing the disappearance of the ethyl
acetate resonance at 1.17 ppm and the appearance of the ethanol
1
obtain an orange powder (0.26 g, 78% yield). H NMR (500 MHz,
CDCl3): δ = 5.57 (d, J = 6.0 Hz, 2H), 5.44 (d, J = 6.1 Hz, 2H), 4.22(p,
J = 7.2 Hz, 4H), 2.97−2.72 (m, 1H), 2.17 (s, 3H), 1.34 (t, J = 7.1 Hz,
4H), 1.25 (d, J = 6.9 Hz, 4H). 13C NMR (151 MHz, CDCl3): δ
107.77, 101.28, 88.89, (d, J = 6.3 Hz), 88.21 (d, J = 5.6 Hz), 63.26 (d, J
= 8.2 Hz), 30.84, 22.19, 18.75, 16.51 (d, J = 6.0 Hz). 31P{1H} NMR
(202 MHz, CDCl3): δ 111.79 ppm. Anal. Found: C, 37.93; H, 5.76; P,
6.79. Calcd for C14H25Cl2O3PRu: C, 37.85; H, 5.67; P, 6.97.
General Procedure for the Hydration of Acetonitrile. [Ru(η6-
p-cymene)Cl2(PR2R′)] (0.02 mmol) was added to 2.8 mL of degassed
D2O in a 1 dram screwcap vial fitted with a septum cap. Dissolution of
3-Me or 3-OEt in water produces an acidic solution with pH 3.5, while
solutions of 1 and 2 had a measured pH of 7.0. To this was added 0.45
mmol of acetonitrile to form a 150 mM nitrile solution. This solution
was heated to 100 °C with stirring. Aliquots (0.1 mL) were removed
periodically using a gastight syringe and then combined in an NMR
tube with 0.6 mL of D2O and 0.1 mL of a 3.8 mM NMe4PF6 in D2O
internal standard solution. The progress of the reaction was monitored
by 1H NMR spectroscopy by observing the disappearance of the
acetonitrile resonance at 2.01 ppm (s, CH3CN) and the appearance of
acetamide at 1.93 ppm (s, CH3C(O)ND2).
1
resonance at 1.06 ppm. H NMR (400 MHz, D2O): ethyl acetate δ
4.03 ppm (q, 2H), 1.99 ppm (s, 3H), 1.17 ppm (t, 3H); ethanol δ 3.44
ppm (q, 2H), 1.06 ppm (t, 3H); acetic acid δ 1.91 ppm (s, 3H).
Computational Methods. Density functional theory (DFT)
calculations were performed on the [Ru(η6-p-cymene)Cl(MeCN)-
(PMe2OH)···H−OH]+ complex. Geometry optimizations were first
performed on the complex without the water molecule. A water
molecule was then added to promote hydrogen bonding to the
PMe2OH ligand, and the entire complex was optimized. Hydrogen
bond distances were obtained from the optimized structures.
Frequency calculations were performed on the optimized structures
to confirm they were at a true minimum, and all yielded zero imaginary
frequencies.
Hydration of Propionitrile. In a 1 dram screwcap vial fitted with
a septum cap, propionitrile (22 μL, 0.31 mmol) was combined with
[Ru(η6-p-cymene)Cl2(PMe2OH)] (1.9 mg, 0.005 mmol) and 2.03 mL
of D2O, forming a solution with pH 3.5. The solution was heated to
100 °C, and 0.1 mL aliquots were removed periodically and combined
with 0.5 mL of d6-DMSO and 10 μL of 255 mM NMe4PF6 in d6-
DMSO internal standard solution. The progress of the reaction was
1
monitored by H NMR spectroscopy by observing the disappearance
of the propionitrile resonances at 2.54 (q, J = 7.6 Hz, 2H,
CH3CH2CN) and 1.26 ppm (t, J = 7.6 Hz, 3H, CH3CH2CN) and
the appearance of the propionamide resonances at 2.19 (q, J = 7.6 Hz,
2H, CH3CH2C(O)ND2) and 1.10 ppm (t, J = 7.6 Hz, 3H,
CH3CH2C(O)ND2). Specific details for the hydration of other nitriles
are described in the Supporting Information.
Hydration of Acetone Cyanohydrin. In a 1 dram screwcap vial
fitted with a septum cap, [Ru(η6-p-cymene)Cl2(PMe2OH)] (10.7
μmol) was combined with 0.67 mL of H2O and 0.77 mL of acetone,
forming a solution with pH 3.5. Freshly distilled acetone cyanohydrin
(20 μL, 0.22 mmol) was added to the solution, which was allowed to
react at 25 °C, and 0.1 mL aliquots were removed periodically and
combined with 0.5 mL of D2O and 0.1 mL of 3.55 mM NMe4PF6 in
D2O. The progress of the reaction was monitored by 1H NMR
Modeling of the [Ru(η6-p-cymene)Cl(MeCN)(PMe2OH)···H−
OH]+ complex was done using the program Ecce v6.1.20 Calculations
were performed using NWChem version 6.0.21 For all atoms except
ruthenium, the 6-311G** basis set22 was used for the geometry
optimizations and frequency calculations. For ruthenium, the basis set
and effective core potential developed by Andrae et al.,23 augmented
with one diffuse f function (ζ = 1.666) determined by Martin and
Sundermann24 and resulting in a (8s7p6d1f)/[6s5p3d1f] contraction,
was used for both the optimizations and frequency calculations. All
calculations employed the B3LYP functional.25−27
X-ray Crystallography: General Methods. Structure determi-
nations were performed on an Oxford Diffraction Gemini-R
diffractometer, using Mo Kα radiation (0.71073 Å). Single crystals
were mounted on Hampton Research Cryoloops using Paratone-N oil.
C
dx.doi.org/10.1021/om400380j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Scheme 1. Synthesis of the [Ru(η6-p-cymene)Cl2(PR2R′)] Complexes
Unit cell determination, data collection and reduction, and analytical
absorption correction were performed using the CrysAlisPro software
package.28 Direct methods structure solution was accomplished using
SIR9229 (1-Me) or Superflip30 (3-Me), and full-matrix least-squares
refinement was carried out using CRYSTALS.31
All non-hydrogen atoms were refined anisotropically. Hydrogen
atoms were placed in calculated positions. Hydrogen positions were
initially refined using distance and angle restraints and were fixed in
place for the final refinement cycles.32 Crystallographic data for 1-Me
and 3-Me and the details of data collection and refinement of the
crystal structure are given in the Supporting Information.
Diffraction-quality crystals were obtained by slow diffusion of
pentane into solutions of 1-Me and 3-Me in CH2Cl2.
RESULTS AND DISCUSSION
Synthesis of the [Ru(η6-p-cymene)Cl2(PR2R′)] Com-
■
Figure 2. X-ray crystal structure of 1-Me, showing 50% probability
ellipsoids. Hydrogen atoms have been omitted for clarity. Selected
metric data (bond lengths in Å and angles in deg): Ru(1)−Cl(2),
2.4219(8); Ru(1)−Cl(3), 2.3991(9); Ru(1)−P(4), 2.3387(8); Cl(2)−
Ru(1)−Cl(3), 89.98(3); Cl(2)−Ru(1)−P(4), 82.56(3); Cl(3)−
Ru(1)−P(4), 86.45(3); Ru(1)−Cavg, 2.211(4); Ru(1)−centroid,
1.705; P(4)−C(5), 1.828(3); P(6)−C(5), 1.817(3); P(4)−C(5)−
P(6), 119.26(16); P(6)−O(7), 1.487(2).
plexes. To explore the effect of the hydrogen bond accepting
group position on the rate of nitrile hydration, the series of
[Ru(η6-arene)Cl2(PR2R′)] catalysts 1-Me, 1-Ph, 2-Me, 2-Ph,
3-Me, and 3-OEt was investigated (Scheme 1). The complexes
[Ru(η6-p-cymene)Cl2(dppmO)] (1-Ph) and [Ru(η6-p-
cymene)Cl2(dppeO)] (2-Ph) are known, having been
previously investigated for their use as potential anticancer
agents.33 [Ru(η6-arene)Cl2PR3]-type complexes with bulky
phosphine ligands are slower nitrile hydration catalysts than
catalysts with smaller phosphine ligands,8 and for that reason,
the new, less bulky complexes 1-Me, 2-Me, 3-Me, and 3-OEt
were also prepared. The respective reactions of dmpmO,
dmpeO, Me2P(O)H, and (EtO)2P(O)H with [Ru(η6-p-
cymene)Cl2]2 in CH2Cl2 at room temperature cleanly
produced the monomeric complexes [Ru(η6-p-cymene)-
Cl2(dmpmO)] (1-Me), [Ru(η6-p-cymene)Cl2(dmpeO)] (2-
Me), [Ru(η6-p-cymene)Cl2(PMe2OH)] (3-Me), and Ru(η6-p-
cymene)Cl2(P(OEt)2OH)] (3-OEt) (Scheme 1). The com-
plexes 1-Me, 1-Ph, 2-Me, and 2-Ph showed two doublet
resonances in the 31P NMR spectrum, where the signal
corresponding to the phosphine is shifted upfield (Δ = 50−60
ppm) relative to the free phosphine. The signal corresponding
to the phosphine oxide is shifted downfield by 2−4 ppm,
consistent with η1 coordination of the bisphosphine monoxide
ligand.34 Complexes 3-OEt and 3-Me both showed a singlet 31P
NMR signal around 110 ppm, corresponding to P coordination
of the phosphinite ligands.
The structures of compounds 1-Me and 3-Me were
determined by X-ray diffraction methods. Figure 2 shows the
structure of 1-Me, and Figure 3 shows that of 3-Me. Each
complex adopts the expected three-legged “piano-stool”
geometry. The Ru−P (2.3387(8) Å) and P−O (1.487(2) Å)
bond lengths of 1-Me are similar to those of 1-Ph (2.355(7) Å
and 1.49(1) Å)35 and the related pendant oxide complex
[Ru(η6-p-cymene)Cl2(η1-PPh2CH(CH3)P(O)Ph2)] (2.374(1)
and 1.484(3) Å).14 3-Me has a slightly shorter Ru−P bond
Figure 3. X-ray crystal structure of 3-Me, showing 50% probability
ellipsoids (selected molecule from the asymmetric cell). Hydrogen
atoms have been omitted for clarity. Selected metric data (bond
lengths in Å and angles in deg, averaged over asymmetric cell):
Ru(1)−Cl(2), 2.4231(9); Ru(1)−Cl(3), 2.4160(5); Ru(1)−P(4),
2.3078(1); Cl(2)−Ru(1)−Cl(3), 87.00(8); Cl(2)−Ru(1)−P(4),
86.16(3); Cl(3)−Ru(1)−P(4), 86.51(3); Ru(1)−Cavg, 2.217(1);
Ru(1)−centroid, 1.707; P(4)−O(5), 1.613(3).
D
dx.doi.org/10.1021/om400380j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Table 1. Hydration of Acetonitrile to Acetamide Catalyzed by [Ru(η6-p-cymene)Cl2(PR2R′)] and [Ru(η6-p-cymene)Cl2(PR3)]
a
Complexes in Water
b
c
cg
,
entry
catalyst PR3
TOF
hydration (%)
reacn time (h)
Ni(CO)3PR3 ν(CO) (cm−1
)
PR3 cone angle (deg)
1
2
3
4
dmpmO (1-Me)
dmpeO (2-Me)
dppmO (1-Ph)
dppeO (2-Ph)
PMe2OH (3-Me)
P(OEt)2OH (3-OEt)
P(OEt)3
0.054
0.051
0.033
0.030
31.95
0.96
92
88
56
50
98
5
331
331
331
331
0.9
1
∼2063
∼2063
∼2067
∼2067
∼123
∼123
∼140
∼140
∼118
∼109
109
d
e
5
2072 (2006 for deprotonated form)
d
ef
,
6
2087 (2028 for deprotonated form)
2076.3
2062
7
0.01
18
97
26
d
8
P(NMe2)3
25.3
1.1
157
a
b
Reactions were performed under an N2 atmosphere at 100 °C and pH 7 using acetonitrile (0.15 M in water), with 5% catalyst loading. Turnover
frequencies ((mol of amide)/((mol of catalyst) time)) were determined by taking the initial rate ((M amide)/time) divided by catalyst
concentration (M). Electronic and steric parameters for the various PR2R′ ligands are taken from ref 38. Reactions were performed at pH 3.5. The
pH of entry 8 was adjusted with HCl. Dissolution of 3-Me or 3-OEt in water produces an acidic solution with pH 3.5. Reference 39. Electronic
parameters are for the similar ligand P(OCH2CH2O)OH. The cone angle and Tolman electronic parameter for the PR2R′ ligands in 1-Ph and 2-Ph
are assumed to be similar to those for PEtPh2, and the cone angle and Tolman electronic parameter for the PR2R′ ligands in 1-Me and 2-Me are
assumed to be similar to those for PMe2Et. The cone angles for the PR2R′ ligands in 3-Me and 3-OEt are similar to those for PMe3 and P(OEt)3.
c
d
e
f
g
Figure 4. Transition state structures for the hydration of acetonitrile with (a) [Ru(η6-p-cymene)Cl2(PR2(CH2P(O)R2))] (1-Me and 1-Ph), (b)
[Ru(η6-p-cymene)Cl2(R2P(CH2CH2P(O)R2))] (2-Me and 2-Ph), (c) [Ru(η6-p-cymene)Cl2(PR2(OH))] (3-Me and 3-OEt), and (d) [Ru(η6-p-
cymene)Cl2(P(OMe)3)]. The transition state ring sizes for each are 7, 9, 10, and 7 for a−d, respectively.
length (2.3078(1) Å) and a longer P−O bond length (1.613(3)
Å), consistent with a P−O single bond.
If the steric properties (and to a lesser extent the electronic
properties) of the ligand solely determine the efficiency of the
catalysts, then 3-OEt should be faster than 3-Me (because
P(OEt)2OH is less bulky than PMe2OH and because PMe2OH
is a better donor than P(OEt)2OH; see Table 1). However, 3-
Me is a far better catalyst, and the implication is that other
factors are also important in determining the rate. It is
suggested that the superior catalytic ability of 3-Me in
comparison to 3-OEt is due to the H bonding that occurs in
the secondary coordination sphere between the PR2R′ ligand
and the entering water nucleophile (Figure 4). Although
P(OEt)2OH could potentially H bond to water, the electron-
withdrawing inductive effect of the OEt groups will decrease
the electron density on the O(H) atom in P(OEt)2OH, making
this O atom a poorer H bond acceptor than the corresponding
O atom in the PMe2OH ligand.
The very low TOFs with catalyts 1-Me, 1-Ph, 2-Me, and 2-
Ph likely indicate that H bonding to the nucleophilic water is
not occurring with these catalysts. This result is not surprising.
The activation of water by H bonding in the secondary
coordination sphere results in the formation of a cyclic
transition state in the rate-limiting step (Figures 1 and 4) . It
is well-established that ring-forming reactions became more
entropically favorable as the ring size increases from 3- to 6-
membered rings, then less favorable as the ring size increases
from 7- to 10-membered rings, with 5-, 6-, and 7-membered
rings being the most favorable.40,41 It is also known that the
strength of H bonding interactions is sensitive to entropy.42
Consequently, in H bonding situations involving the formation
of a cyclic transition state, the rate of the reaction will decrease
as the ring-forming reaction becomes less entropically
Secondary Coordination Sphere H-Bonding Effects.
The effect on catalyst activity of the position of the H bond
accepting group in the secondary coordination sphere was
investigated by hydrating acetonitrile using 1-Me, 1-Ph, 2-Me,
2-Ph, 3-Me, and 3-OEt as catalysts. Of the complexes
investigated, 3-Me was by far the best catalyst, with a turnover
frequency (TOF) of 31.95 h−1 (Table 1).36 3-OEt had the
second highest activity with a TOF of 0.96 h−1. Complexes 2-
Ph, 1-Ph, 2-Me, and 1-Me had much lower catalytic activity,
with TOFs of 0.030, 0.033, 0.051, and 0.053 h−1, respectively.
Previous work with the [Ru(η6-p-cymene)Cl2PEt3] and [Ru(η6-
p-cymene)Cl2PiPr3] catalysts showed that nitrile hydration is
faster with less bulky phosphines,8 and the general trend
observed in Table 1 is consistent with this prior finding. Thus,
catalysts 1-Ph and 2-Ph are the slowest because they have the
bulkiest ligands, catalysts 1-Me and 2-Me have smaller PR2R′
ligands than 1-Ph and 2-Ph and are therefore slightly faster, and
3-Me and 3-OEt have the smallest phosphines and are the
fastest. (The exception to this ordering is that 3-Me is faster
than 3-OEt, despite the fact that 3-Me is the sterically bulkier
ligand. Reasons for this deviation are discussed below.)
Previous work also showed that the rate decreased with
increasing electron-donating ability of the ligands on the
catalyst. The general trend in the rates for the catalysts in Table
1 also follow this general rule, but the changes in electron-
donating ability (as indicated by the Tolman ν(CO)
parameter) are small, and it is suggested that the major
influence on the rate is the steric bulkiness of the PR2R′
ligand.37
E
dx.doi.org/10.1021/om400380j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
favorable. 3-Me and the previously investigated complex
[Ru(η6-p-cymene)Cl2P(NMe2)3] (which has a comparable
rate; Table 1) would form 7-membered transition states,
which are more favorable than the 9- and 10-membered rings
that form in the transition state with catalysts 1-Me, 1-Ph, 2-
Me, and 2-Ph.
Table 2. Calculated Binding Energies and H-Bond Lengths
a
of the [P(R)3···H−OH] Free Ligand and the
Corresponding [Ru(η6-p-cymene)Cl(MeCN)(PR3)···H−
OH]+ Complexes using 6-311G**(B3LYP)
H bond
length (Å)
H2O···MeCN
b
complex
dist (Å)
It might be argued that the decreased activity of 1-Me, 1-Ph,
2-Me, and 2-Ph may simply be due to the ability of the
phosphine oxide group to coordinate to the ruthenium, making
the R2P(CH2)nP(O)R2 (R = Ph, Me and n = 1, 2) ligands
bidentate (Figure 5).14 This type of η2 coordination through
c
[P(NMe2)3···H−OH]
[P(OMe)3···H−OH]
[PMe2OH···H−OH]
2.016
2.012
1.957
1.974
c
[Ru(η6-p-cymene)Cl(MeCN)
2.923
3.005
2.942
c
P(NMe2)3···H−OH]+
[Ru(η6-p-cymene)Cl(MeCN)
2.038
2.014
c
P(OMe)3···H−OH]+
[Ru(η6-p-cymene)Cl(MeCN)
(PMe2OH)···H−OH]+
a
The water is hydrogen bonded to the heteroatom on the free
b
phosphine ligand. This describes the calculated distance between the
oxygen of the water that is hydrogen bonded to the coordinated
phosphine ligand and the nitrile carbon of the coordinated acetonitrile.
c
See the structures in Figure 4. See ref 8.
Figure 5. Structure of the bidentate complex [Ru(η6-p-cymene)Cl(η2-
R2P(CH2)nP(O)R2)]Cl, where R = Ph, Me and n = 1, 2.
the phosphine oxide was previously shown for the related
complex [Ru(η6-p-cymene)Cl2(η1-PPh2CH(R′)P(O)Ph2)] (R′
≠ H) in polar solutions, where equilibrium mixtures of the
bidentate and monodentate complexes were observed. It is
noted that complexes with bulkier R′ groups in the methylene
position showed greater concentrations of the bidentate
complex at equilibrium.14 Crucially, the bidentate complex
was not observed for 1-Ph,14 and therefore the bidentate
complex is also not expected to form for 1-Me and, by
extrapolation, for 2-Me because both are less sterically bulky
than 1-Ph and 2-Ph.
Figure 6. Nitriles hydrated with 3-Me: (a) propionitrile; (b)
isobutyronitrile; (c) trimethylacetonitrile; (d) acrylonitrile; (e)
methacrylonitrile; (f) benzonitrile; (g) methoxyacetonitrile; (h)
methyl cyanoacetate; (i) acetone cyanohydrin acetate.
To summarize this section, the general trend in catalyst rates
is primarily determined by the steric properties of the PR2R′
ligand in the Ru(η6-p-cymene)Cl2(PR2R′)-type complexes.
However, ligands with an OH group bonded to the P atom
(as in 3-Me and 3-OEt) are faster catalysts, which is probably
attributable to their ability to activate the water nucleophile by
forming a relatively stable, cyclic, H bonded transition state.
The location of the H bond acceptor atom is crucial. When the
formation of an H bond results in a large ring, the rate is not
enhanced, likely due to the entropic unfavorability of forming
the large ring.
DFT Analysis of Nitrile Hydration. The hydrogen bond
accepting ability of [Ru(η6-p-cymene)Cl(MeCN)(PMe2OH)]+
was investigated by DFT analysis at the 6-311G** level of
theory. The H···O(ligand) and O···C(nitrile) bond lengths in
the cyclic transition state (Figure 4c) were calculated and
compared to prior results obtained for the complexes [Ru(η6-p-
cymene)Cl(MeCN)(P(NMe2)3)]+ and [Ru(η6-p-cymene)Cl-
(MeCN)(P(OMe)3)]+ (Table 2).8,13 The calculations showed
that the complex [Ru(η6-p-cymene)Cl(MeCN)(P(OMe)3]+,
which is the slowest nitrile hydration catalyst of the three, has
the longest bond distances. The bond lengths in the catalysts
[Ru(η6-p-cymene)Cl(MeCN)(P(NMe2)3)···H−OH]+ and
[Ru(η6-p-cymene)Cl(MeCN)(PMe2OH)···H−OH]+ are short-
er, implying greater hydrogen bonding and consistent with their
faster activity.
probe the effect of nitrile bulk on catalyst activity. These nitriles
were hydrated with TOFs of 14.0, 3.0, 6.1, and 20.7 h−1,
respectively (Table 3). The general trend is for decreased TOF
with increasing steric bulk in the nitrile, but the cause of the
reduced activity with 3-Me toward isobutyronitrile is not
known. In this case, the catalyst solution turned gray as
hydration progressed, indicating that the catalyst mixture may
not remain homogeneous.
The hydration reactions of acrylonitrile, methacrylonitrile,
methylcyanoacetate, and acetone cyanohydrin acetate with 3-
Me were also investigated because ester and alkene functional
groups are known to be susceptible to hydration or hydrolysis
in the presence of some nitrile hydration catalysts.9 The CN
groups in acrylonitrile and methacrylonitrile were completely
hydrated, and no alkene hydration or polymerization was
observed. Some (∼5%) ester hydrolysis was observed during
the hydration of methyl cyanoacetate (Table 1, entry 9), as
indicated by the formation of methanol in the NMR spectrum.
However, the hydration of the resulting mixture of methyl
cyanoacetate and its hydrolyzed cyanoacetic acid went to
completion. The P−O−H group has a pKa of 4.11
0.08,
meaning that at pH 3.5 approximately one-third of the catalyst
in solution is deprotonated. Because the hydration reactions
with 3-Me were carried out at pH 3.5, the ester hydrolysis may
be catalyzed by the acid in solution, rather than by 3-Me. To
investigate this hypothesis, control hydrolysis reactions of ethyl
acetate were carried out at pH 3.5 (100 °C) both with catalyst
(5 mol % of 3-Me) and without catalyst. In addition, control
hydrolysis reactions of ethyl acetate were carried out at pH 7.0
Nitrile Hydration with 3-Me. Because 3-Me was the
fastest catalyst for CH3CN hydration, its activity toward other
nitriles was investigated (Figure 6). Propionitrile, isobutyroni-
trile, trimethylacetonitrile, and benzonitrile were investigated to
F
dx.doi.org/10.1021/om400380j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
a
Table 3. Selected Nitrile Hydration Results Using [Ru(η6-p-cymene)Cl2(P(OH)Me2)]
b
entry
substrate
acetonitrile
[substrate] (mM)
[catalyst] (mM)
pH
hydration (%)
reacn time (h)
1
150.8
150.3
150.2
149.0
150.0
149.7
150.2
152.3
150
7.6
7.5
7.5
7.5
7.5
7.5
7.5
7.6
7.5
7.5
7.5
7.5
0
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
7.0
3.5
7.0
3.5
7.0
98
1.0
2.2
7.5
7.5
2.0
3.0
2.0
1.0
2.0
1.7
1.8
1.8
1.8
1.8
2
propionitrile
93
3
isobutyronitrile
trimethylacetonitrile
acrylonitrile
74
4
90
5
>99
98
6
methacrylonitrile
benzonitrile
7
>99
>99
8
methoxyacetonitrile
methyl cyanoacetate
methyl cyanoacetate
ethyl acetate
9
>99 (5% ester hydrolysis)
10
11
12
13
14
150.2
150.7
150.7
150.7
150.7
>99 (25% ester hydrolysis)
46
26
46
ethyl acetate
ethyl acetate
ethyl acetate
0
0
a
b
1
Reactions performed under an N2 atmosphere in D2O with 5% catalyst loading at 100 °C. Yields determined by H NMR. No formation of
carboxylic acid byproduct was observed in any hydration trial, and no deuteration of C−H protons was observed in any hydration trial.
a
b
Table 4. Selected Cyanohydrin Hydration Results Using [Ru(η6-p-cymene)Cl2(P(NMe2)3)] or [Ru(η6-p-
cymene)Cl2(PMe2OH)]
[cosolvent]
(M)
[catalyst]
(mM)
calcd [HCN] at equilibrium
hydration
c
d
entry
1
cyanohydrin
glycolonitrile
catalyst PR3
(mM)
(%)
reacn time (h)
PMe2OH
P(NMe2)3
PMe2OH
P(NMe2)3
PMe2OH
PMe2OH
PMe2OH
P(NMe2)3
PMe2OH
PMe2OH
7.5
7.6
7.5
7.6
7.5
7.5
7.4
7.5
7.5
7.4
0.01
0.07
2.01
2.01
0.93
1.48
2.10
1.01
>99
>99
94
6.2
e
2
glycolonitrile
49
3
4
5
6
7
lactonitrile
17.5
lactonitrile
>99
3.3
112
ACH
11.6
7.2
0.8
ACH
6.5
1.0
ACH
5.0
3.9
1.1
f
8
ACH
10.6
14
149
9
acetone cyanohydrin acetate
acetone cyanohydrin acetate
60.6
>99
2.0 (100 °C)
3.7 (100 °C)
g
10
a
Reactions performed under an N2 atmosphere in D2O at 25 °C with 0.15 M cyanohydrin, 5% catalyst loading, and pH 3.5. ACH hydrations
conducted with acetone cosolvent. See ref 7. Concentration of cyanide at equilibrium was calculated using published equilibrium constants.3
b
c
d
Yields determined by 1H NMR. No formation of carboxylic acid byproduct was observed in any hydration trial, and no deuteration of C−H protons
e
f
g
was observed in any hydration trial. Reaction done at pH 8.5. Reaction done at pH 3.0. Reaction done at pH 7.0.
(100 °C) both with catalyst (5 mol % of 3-Me) and without
catalyst. At pH 3.5, the catalyst had no effect on the extent of
hydrolysis; 46% of ethyl acetate was hydrolyzed to acetic acid
and ethanol in each trial (Table 3, entries 11 and 13; see the
Supporting Information for details).43 In comparison, the
catalyst does have an effect at pH 7: no hydrolysis was observed
in the absence of catalyst (Table 3, entry 14), but 26%
hydrolysis occurred with the catalyst present (Table 3, entry
12). The ability of 3-Me to function as a hydrolysis catalyst at
pH 7 likely explains the more extensive ester hydrolysis
observed at pH 7 than at pH 3.5 for the hydration of methyl
cyanoacetate (Table 3, entries 9 and 10).
Hydration of Cyanohydrins. The hydration of the
cyanohydrins glycolonitrile, lactonitrile, and acetone cyanohy-
drin was investigated with 3-Me. Previous investigations with
the [Ru(η6-p-cymene)Cl2(P(NMe2)3)] catalyst showed that the
optimal conditions for cyanohydrin hydration are low temper-
atures (25 °C) and low pH (3−4). Acetone cyanohydrin
hydration also requires the use of acetone cosolvent.4,7,8 The
results of the cyanohydrin hydrations with 3-Me at 25 °C, pH
3.5, 5% catalyst loading, and 150 mM nitrile are shown in Table
4.
with >99% and 94% yields, respectively (Table 4). These
results correlate with the superior catalytic activity of 3-Me in
comparison to [Ru(η6-p-cymene)Cl2(P(NMe2)3)] at pH 3.5.
The hydration of acetone cyanohydrin, by comparison, was
worse than expected. Under all conditions in which the
hydration was attempted, no more than 4% conversion of
acetone cyanohydrin to 2-hydroxyisobutyramide was observed
(Table 4, entries 5−7). It is suggested that this low conversion
is likely due to a higher sensitivity of 3-Me to cyanide, as shown
by cyanide poisoning studies (see below).
To limit the degradation of acetone cyanohydrin, the alcohol
group on acetone cyanohydrin was protected by acetate.
Hydration of acetone cyanohydrin acetate with 3-Me at pH 3.5
and 100 °C gave a 60% conversion to a mixture of the desired
α-hydroxyisobutyramide and α-acetoxyisobutyramide products
(Table 4, entry 9). However, hydrolysis of the acetone
cyanohydrin acetate also occurred to form acetone cyanohydrin
(Scheme 2a), which degraded and subsequently caused the
catalyst activity to cease due to cyanide poisoning. The previous
control studies with ethyl acetate showed that the rate of ester
hydrolysis was significantly decreased at higher pH; therefore,
the hydration of acetone cyanohydrin acetate was also carried
out at pH 7.0. Under these conditions, complete hydration of
acetone cyanohydrin acetate occurred, with the amide products
Under these conditions, glycolonitrile and lactonitrile
hydration both went to completion within 6.2 and 17.5 h
G
dx.doi.org/10.1021/om400380j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Scheme 2. (a) Hydrolysis of Acetone Cyanohydrin Acetate,
Reforming Acetone Cyanohydrin, Which Quickly Degrades
To Form Cyanide, and (b) Hydration of Acetone
Cyanohydrin Acetate to α-Acetoxyisobutyramide, Which
May Be Hydrolyzed To Form α-Hydroxyisobutyramide
SUMMARY
■
With the catalysts 1-Me, 1-Ph, 2-Me, 2-Ph, 3-Me, and 3-OEt,
the general trend in nitrile hydration rates is primarily
determined by the steric properties of the PR2R′ ligands.
However, the proper placement of an H bond accepting group
in the PR2R′ ligand can effect the rate. The TOF is faster if the
cyclic, H bonded transition state (Figures 1 and 4) is
entropically favored. For example, the transition states formed
with 3-Me and 3-OEt have 7 total atoms, and these catalysts
are faster than reactions with 1-Me, 1-Ph, 2-Me, and 2-Ph,
where 9- and 10-membered rings are formed. Although the
catalysts 3-Me and 3-OEt both form transition states with 7
atoms, the faster rate for 3-Me is attributed to the greater H
bond accepting ability of the PMe2OH ligand in comparison to
the P(OEt)2OH ligand.
being a mixture of α-hydroxyisobutyramide and α-acetoxyiso-
butyramide (Table 4, entry 10; Scheme 2b).
Catalyst 3-Me was able to hydrate a variety of nitriles,
including several cyanohydrins. In fact, 3-Me was a faster
catalyst for the hydration of glycolonitrile and lactonitrile (both
cyanohydrins) than [Ru(η6-p-cymene)Cl2P(NMe2)3], the best
previously investigated homogeneous cyanohydrin hydration
catalyst.7,8 Unfortunately, as was the case with the other
previously studied homogeneous nitrile hydation catalysts, the
catalysts investigated here are susceptible to cyanide poisoning.
Thus, 3-Me was not a good acetone cyanohydrin hydration
catalyst for this reason. What is abundantly clear is that
homogeneously catalyzed acetone cyanohydrin hydration will
never be viable unless a way can be found to eliminate cyanide
poisoning. One strategy to prevent catalyst poisoning is to
protect the −OH group of the cyanohydrin, and in fact this
strategy does work: protecting acetone cyanohydrin with an
acetate group quantitatively gave the hydration product at pH 7
and with >60% yield at pH 3.5. Of course, protecting the −OH
group may not be industrially viable. Another strategy to
prevent catalyst poisoning is to use a catalyst with labile metal−
cyanide bonds; our next paper will report the results of using
this approach.
Catalyst Inhibition Tests. Previous control experiments
with the PtCl(PMe2OH)(PMe2O)2H catalyst showed that as
the [KCN]/[PtCl(PMe2OH)(PMe2O)2H] ratio increased, the
rate of acetonitrile hydration decreased because of cyanide
poisoning.6 To determine if cyanide similarly inhibits the nitrile
hydration activity of 3-Me, aliquots of potassium cyanide were
added to acetonitrile hydration solutions. Interestingly, the rate
of hydration did not decrease linearly, as was observed
previously with [PtCl(PMe2OH)(PMe2O)2H].6 Instead, as
the number of equivalents of cyanide increased from 0 to 1,
the rate of acetonitrile hydration increased. Only after 1 equiv
of cyanide was added did the rate of hydration decrease, and all
catalyst activity practically ceased with 5 equiv of cyanide
(Figure 7).
ASSOCIATED CONTENT
■
S
* Supporting Information
Text, figures, and CIF files giving experimental details for the
hydration of nitriles and cyanohydrins with 3-Me, H, 31P, and
1
13C NMR spectra of dmpeO, dmpmO, and catalysts 1-Me, 2-
Me, 3-Me, and 3-OEt, crystallographic data for 1-Me and 3-Me
and details of the data collection and refinement of the crystal
structures, and time course data for the hydrolysis of ethyl
acetate and the hydration and hydrolysis of acetone
cyanohydrin acetate at pH 3.5 and 7.0. This material is
available free of charge via the Internet at http://pubs.acs.org.
Figure 7. Plot of acetonitrile hydration rate versus [KCN]/[Ru]
showing the effect of added cyanide on the catalytic activity of [Ru(η6-
p-cymene)Cl2(PMe2OH)].
To explain this behavior, it is proposed that the active
catalyst during acetonitrile hydration in the absence of cyanide
is [Ru(η6-p-cymene)Cl(NCMe)(PMe2OH)].44 Addition of 1
equiv of cyanide yields the complex [Ru(η6-p-cymene)(CN)-
(NCMe)(PMe2OH)]+ (eq 1), which is likely a faster catalyst
AUTHOR INFORMATION
Corresponding Author
*E-mail for D.R.T.: dtyler@uoregon.edu.
■
Present Address
§Department of Chemistry, Colgate University, Hamilton, NY
13346.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
than its chloro or aquo precursor because the cyanide ligand is
more electron withdrawing than chloride or water and can
therefore better activate nitriles for hydration.
We thank Anthony R. Chianese at Colgate University for his
help in obtaining the X-ray crystal structures. We would also
like to thank Rohm and Haas Chemical Co. and the NSF
H
dx.doi.org/10.1021/om400380j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
(25) Becke, A. D. Phys. Rev. A 1988, 38, 3098−3100.
(26) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652.
(27) Becke, A. D. J. Chem. Phys. 1993, 98, 1372−1377.
(CHE-0719171) for the support of research carried out in the
authors’ laboratory. X-ray structure determination was
supported by the NSF (CHE-1057792 and CHE-0819686).
S.M.M.K. also acknowledges the U.S. Department of Education
(P200A070436) and the NSF Graduate STEM Fellows in K-12
Education (GK-12) program (DGE-0742540) for additional
support. “Extensible Computational Chemistry Environment
(ECCE), A Problem Solving Environment for Computational
Chemistry, Software Version 6.1” (2011), as developed and
distributed by Pacific Northwest National Laboratory, P.O. Box
999, Richland, WA 99352, USA, and funded by the U.S.
Department of Energy, was used to obtain some of these
results.
(28) CrysAlisPro Software system, Version 1.171.32 and Xcalibur CCD
system; Oxford Diffraction Ltd., 2007.
(29) Altomare, A.; Cascarano, M.; Giacovazzo, C.; Guagliardi, A.;
Burla, M. C.; Polidori, G.; Camalli, M. J. Appl. Crystallogr. 1994, 27,
435−436.
(30) Palatinus, L.; Chapuis, G. J. Appl. Crystallogr. 2007, 40, 786−
790.
(31) Betteridge, P. W.; Carruthers, J. R.; Cooper, R. I.; Prout, K.;
Watkin, D. J. J. Appl. Crystallogr. 2003, 36, 1487.
(32) Cooper, R. I.; Thompson, A. L.; Watkin, D. J. J. Appl. Crystallogr.
2010, 43, 1100−1107.
(33) Das, S.; Sinha, S.; Britto, R.; Somasundaram, K.; Samuelson, A.
G. J. Inorg. Biochem. 2010, 104, 93−104.
(34) Previous studies have shown that both 31P NMR signals of a
bisphosphine monoxide ligand shift upfield by ∼20−30 ppm upon
coordination of the phosphine oxide.14
REFERENCES
■
(1) Nagai, K. Appl. Catal., A 2001, 221, 367−377.
(2) Green, M. M.; Wittcoff, H. A. In Organic Chemistry Principles and
Industrial Practice; Wiley-VCH: Weinheim, Germany, 2003; pp 137−
156.
(35) Chaplin, A. B.; Scopelliti, R.; Dyson, P. J. Eur. J. Inorg. Chem.
2005, 2005, 4762−4774.
(36) Turnover frequencies (TOFs) are calculated by dividing the
initial rate of the hydration reaction by the catalyst concentration.
Therefore, TOF = mol of amide (mol cat)−1 h−1.
(37) Note that deprotonation of PMe2OH to the corresponding
phosphinito ligand PMe2O− produces a complex with a more electron
donating ligand (the Tolman electronic parameter of PMe2O− is 2006
cm−1).38 To probe which of the two species is present during catalysis,
titration of 3-Me with sodium hydroxide showed that the P−O−H
group has a pKa of 4.11 0.08. For the nitrile hydration reactions
done at pH 3.5, the ratio of the acidic to the basic form is thus about
2:1.
(38) Tolman, C. A. Chem. Rev. 1977, 77, 313−348.
(39) Martin, D.; Moraleda, D.; Achard, T.; Giordano, L.; Buono, G.
Chem. Eur. J. 2011, 17, 12729−12740.
(40) These thermodynamic properties have been observed for the
investigation of SN2 ring-closing reactions of linear α,ω-bromoalkyl-
amines to form cyclic amines and the SN2 ring-closing reactions of
α,ω-bromoalkylcarboxylic acids to form lactones. See: Charton, M. J.
Am. Chem. Soc. 1975, 97, 1552−1556.
(3) Schlesinger, G.; Miller, S. L. J. Am. Chem. Soc. 1973, 95, 3729−
3735.
(4) Stewart, T. D.; Fontana, B. J. J. Am. Chem. Soc. 1940, 62, 3281−
3285.
(5) Yates, W. F.; Heider, R. L. J. Am. Chem. Soc. 1952, 74, 4153−
4155.
(6) Ahmed, T. J.; Fox, B. R.; Knapp, S. M. M.; Yelle, R. B.; Juliette, J.
J.; Tyler, D. R. Inorg. Chem. 2009, 48, 7828−7837.
(7) Knapp, S. M. M.; Sherbow, T. J.; Juliette, J. J.; Tyler, D. R.
Organometallics 2012, 31, 2941−2944.
(8) Knapp, S. M. M.; Sherbow, T. J.; Yelle, R. B.; Zakharov, L. N.;
Juliette, J. J.; Tyler, D. R. Organometallics 2013, 32, 824−834.
(9) Ahmed, T. J.; Knapp, S. M. M.; Tyler, D. R. Coord. Chem. Rev.
2011, 255, 949−974.
(10) Kukushkin, V. Y.; Pombeiro, A. J. L. Inorg. Chim. Acta 2005,
358, 1−21.
́
(11) García-Alvarez, R.; Díez, J.; Crochet, P.; Cadierno, V.
Organometallics 2010, 29, 3955−3965.
(12) Lee, W.-C.; Sears, J. M.; Enow, R. A.; Eads, K.; Krogstad, D. A.;
Frost, B. J. Inorg. Chem. 2013, 52, 1737−1746.
(41) DeTar, D. F.; Luthra, N. P. J. Am. Chem. Soc. 1980, 102, 4505−
4512.
(42) Knowles, R. R.; Jacobsen, E. N. Proc. Natl. Acad. Sci. U.S.A.
2010, 107, 20678−20685.
(13) Steiner, T. Angew. Chem., Int. Ed. 2002, 41, 48−76.
(14) Faller, J. W.; Patel, B. P.; Albrizzio, M. A.; Curtis, M.
Organometallics 1999, 18, 3096−3104.
(43) Charton, M. J. Am. Chem. Soc. 1975, 97, 1552−1556.
(44) Previous work has shown that similar complexes of the formula
[Ru(η6-p-cymene)Cl(NCMe)(P(NMe2)3)]PF6 are chemically and
kinetically competent for the hydration of nitriles. See ref 8 and:
(15) Boshaala, A. M. A.; Simpson, S. J.; Autschbach, J.; Zheng, S.
Inorg. Chem. 2008, 47, 9279−9292.
(16) Hays, H. R. J. Org. Chem. 1968, 33, 3690−3694.
(17) Clement, J.-L.; Finet, J.-P.; Frejaville, C.; Tordo, P. Org. Biomol.
Chem. 2003, 1, 1591−1597.
́
́
Garcia-Alvarez, R.; Dıez, J.; Crochet, P.; Cadierno, V. Organometallics
2011, 30, 5442−5451.
(18) Hiyama, T.; Oishi, H.; Suetsugu, Y.; Nishide, K.; Saimoto, H.
Bull. Chem. Soc. Jpn. 1987, 60, 2139−2150.
(19) Mading, P.; Scheller, D. Z. Anorg. Allg. Chem. 1988, 567, 179−
̈
191.
(20) Black, G.; Chase, J.; Chatterton, J.; Daily, J.; Elsethagen, T.;
Feller, D.; Gracio, D.; Jones, D.; Keller, T.; Lansing, C.; Matsumoto,
S.; Palmer, B.; Peterson, M.; Schuchardt, K.; Stephan, E.; Sun, L.;
Swanson, K.; Taylor, H.; Thomas, G.; Vorpagel, E.; Windus, T.;
Winters, C. ECCE, A Problem Solving Environment for Computational
Chemistry, Software Version 6.3; Pacific Northwest National Labo-
ratory: Richland, WA 99352-0999, USA, 2011.
(21) Valiev, M.; Bylaska, E. J.; Govind, N.; Kowalski, K.; Straatsma,
T. P.; Van Dam, H. J. J.; Wang, D.; Nieplocha, J.; Apra, E.; Windus, T.
L.; de Jong, W. A. Comput. Phys. Commun. 2010, 181, 1477−1489.
(22) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys.
1980, 72, 650−654.
(23) Andrae, D.; Haußermann, U.; Dolg, M.; Stoll, H.; Preuß, H.
̈
Theor. Chem. Acc. 1990, 77, 123−141.
(24) Martin, J. M. L.; Sundermann, A. J. Chem. Phys. 2001, 114,
3408−3420.
I
dx.doi.org/10.1021/om400380j | Organometallics XXXX, XXX, XXX−XXX