A tungsten-mediated closed cycle of reactivity for the reduction of CO 2 to CO

Article abstract of DOI:10.1039/c0dt00489h

The recycling of CO₂ by reduction to CO is an important objective in the context of renewable carbon feedstock chemicals. A tungsten-mediated reduction of CO₂ to CO reported by Mayer and coworkers has been re-examined, and it is shown that a series of four well-defined stoichiometric steps can be executed which form a closed cycle and sum as CO₂ + 2H⁺ + 2e^- → CO + H ₂O. Energetic parameters of this system are probed by cyclic voltammetry, by calculations of gas-phase reaction enthalpies for each of the four steps, and by calculation of the WO bond dissociation energy for the tungsten species that results from oxidation addition of CO₂.

Full text of DOI:10.1039/c0dt00489h

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
A tungsten-mediated closed cycle of reactivity for the reduction of
CO2 to CO†
Upul Jayarathne, Perumalreddy Chandrasekaran,‡ Heiko Jacobsen, Joel T. Mague and James P. Donahue*
Received 17th May 2010, Accepted 23rd June 2010
DOI: 10.1039/c0dt00489h
The recycling of CO2 by reduction to CO is an important objective in the context of renewable carbon
feedstock chemicals. A tungsten-mediated reduction of CO2 to CO reported by Mayer and coworkers
has been re-examined, and it is shown that a series of four well-defined stoichiometric steps can be
executed which form a closed cycle and sum as CO2 + 2H+ + 2e- → CO + H2O. Energetic parameters
of this system are probed by cyclic voltammetry, by calculations of gas-phase reaction enthalpies for
each of the four steps, and by calculation of the W O bond dissociation energy for the tungsten
species that results from oxidation addition of CO2.
form the corresponding cis oxo-carbonyl tungsten(IV) complexes
Introduction
(eqn (2)).22 More recently, the formation of CO and an oxo-bridged
Concerns about the nonrenewable nature of fossil fuel reserves
diuranium(IV) species arising from the reaction of CO2 with a U(III)
have stimulated broad ranging efforts aimed at the recycling of
triazacyclononane compound was reported by Meyer and Castro-
Rodriguez (eqn (3)).23 However, further studies of these Group
4, Group 6 and actinide systems were not undertaken, possibly
being dissuaded by a belief that the resulting metal oxides are
thermodynamic dead-ends.
CO2. Established uses for CO2 include its incorporation into
organic polymers, particularly polycarbonates,1–2 its niche use
in certain small molecule syntheses, such as salicylic acid and
urea production,3–4 and its application as a supercritical fluid
in extractions5–7 and coatings8–9 and as a medium for chemical
reactions.10–12 Reduction of CO2 to the valuable lower oxide CO is,
however, a much greater challenge owing to the thermodynamic
stability of CO2 and the appreciable kinetic barrier to its activation.
Thus, a successful practical system for reduction of CO2 must
implement some form of catalyst.
Metal-mediated catalytic schemes for the reduction of CO2 to
lower carbon oxides that are chemically driven,13–15 electrolytically-
driven,16–19 or photolytically driven20 have been described, each
with distinctive advantages and shortcomings. Of these systems,
the electrolytic and photolytic are more desirable because of their
ability to more directly use energy from a renewable source. Among
the many issues demanding consideration with any such catalyst
are catalyst lifetime and cost, catalyst rate, efficiency of energy
usage, and product selectivity.
Compared to complexes of the late and middle-late transition
metals, notably Ni, Pd, Co, Fe, Ru, and Re, complexes of the
early and early-middle transition metals and of the f-block metals
have been subject to considerably less scrutiny for their ability to
reduce CO2. The reduction of CO2 to CO and oxo-bridged di- and
trimetallic products was reported by Floriani with titanocene and
zirconocene compounds (eqn (1)),21 while Mayer and coworkers
described the reaction of chlorophosphine complexes of W(II) to
3[Cp2ZrII(CO)2] + 3CO2 → [Cp2Zr]3(m-O)3 + 9CO 21
[WIICl2(PMePh2)4] + CO2 → [WIVO(CO)Cl2(PMePh2)2]
(1)
(2)
(3)
22
+ 2PMePh2
2LU + CO2 ꢀ LU-(CO2)-UL → LU-O-UL + CO 23
L
=
1,4,7-tris(3,5-di-tert-butyl-2-hydroxybenzylate)-1,4,7-
triazacyclonone.
Motivated by (i) the high specificity of the tungsten chlorophos-
phine complexes for CO formation from CO2 (ii) the greater
exposure, and hence greater opportunity for further reactiv-
ity, of a terminal oxo versus a bridging oxo ligand and
(iii) the relatively affordability of tungsten metal, we initi-
ated an investigation of the possibility for turnover of the
[WIV( O)(CO)Cl2(PMePh2)2] product that results in the Mayer
system. The rationale underpinning this effort was that success-
ful reduction of [WIV( O)(CO)Cl2(PMePh2)2] back to non-oxo
tungsten(II) species would at least demonstrate stoichiometric
catalysis, which is a requisite first step if this system could hope,
with appropriate modifications, to operate electrocatalytically in
the reduction of CO2 to CO. Described here are experimental
results that extend Mayer’s observations into a closed cycle
of four stoichiometric steps that provide for the reduction of
CO2 to CO and H2O. These reactivity studies are augmented
with electrochemical data and computational work that bring
additional clarification to the energetic parameters within which
this closed cycle is constrained. While this work was in the course of
preparation for publication, a conceptually similar cycle mediated
with a niobium tris(amido) platform was reported by the Cummins
Group.24
Department of Chemistry, Tulane University, 6400 Freret Street, New
Orleans LA 70118-5698
† Electronic supplementary information (ESI) available: X-ray diffraction
data collection procedures, complete crystallographic data in CIF format,
thermal ellipsoid plots for all compounds with complete atom labeling,
and computational details (optimized geometries, convergence criteria,
final energies, thermodynamic data, Table S6). CCDC reference numbers
769389–769400. For ESI and crystallographic data in CIF or other
electronic format see DOI: 10.1039/c0dt00489h
‡ Current address: Department of Chemistry and Chemical Biology,
Cornell University, Ithaca, NY 14853-1301.
9662 | Dalton Trans., 2010, 39, 9662–9671
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
UV/vis (THF), lmax (eM): 342 nm (5500), 548 nm (230). Anal.
Calcd for C39H39Cl2OP3W: C, 53.75; H, 4.51; P, 10.66. Found: C,
53.48; H, 4.53; P, 10.58.
Experimental
Syntheses
A literature procedure was employed for the generation of
gaseous HCl.25 This procedure was modified by passing the
HCl(g) through a trap of concentrated H2SO4 immediately
before use in order to scavenge traces of moisture. All other
reagents were purchased from commercial sources and used as
received (PMe3 (1.0 M in toluene), PMe2Ph, PMePh2, PPh3, dppe
(dppe = 1,2-bis(diphenylphosphino)ethane), P(NMe2)3, Cp*2Co,
WCl4, WCl6). Solvents either were dried with a system of drying
columns from the Glass Contour Company (CH2Cl2, n-pentane,
hexanes, Et2O, THF, C6H6, toluene) or distilled according to
standard procedures (1,2-dichloroethane).26 A numbering system
for compound identification is defined in Chart 1. Compounds 1
and 2 were prepared by literature methods (by Na/Hg reduction
[WOCl2(dppe)(PMePh2)], 3e. A 50 mL Schlenk flask with
magnetic bar was charged with [W(O)(CO)Cl2(PMePh2)2] (0.208 g,
0.297 mmol) and dppe (0.118 g, 0.296 mmol) under an Ar
atmosphere. Freshly distilled toluene (15 mL) was added, and the
resulting reaction mixture was heated to 70 ◦C for 4 h with constant
stirring. After being cooled to room temperature, the solution was
reduced to dryness under reduced pressure. The residual solid
was washed with pentane and dried under vacuum. Redissolution
of this crude product in CH2Cl2 (10 mL), followed by addition of
hexanes (30 mL) led to formation of a microcrystalline precipitate,
which was isolated by cannula filtration and dried under vacuum
1
for 24 h. Yield: 0.210 g, 0.242 mmol, 81%. H NMR (400 MHz,
CDCl3): d 7.97–7.92 (m, 2H, Ph), 7.86 (t, JHH = 8.4 Hz, 2H, Ph),
7.66–7.61 (m, 2H, Ph), 7.55 (t, JHH = 8.8 Hz, 2H, Ph), 7.41–7.28
(m, 12H, Ph), 7.20 (t, JHH = 7.6 Hz, 4H, Ph), 7.11–7.03 (m, 4H, Ph),
6.84 (t, JHH = 8.8 Hz, 2H, Ph), 3.20–2.91 (m, 2H, PCH2CH2P),
2.77–2.62 (m, 1H, PCH2CH2P), 2.11–1.99 (m, 1H, PCH2CH2P),
1.84 (d, JPH = 8.4 Hz, 3H, PMePh2). 31P NMR (d, ppm in CDCl3):
32.01 (d of d, 2JPP = 4.3 Hz, 17.8 Hz; 1JPW = 391 Hz, PCH2CH2P
27
of 4c in the presence of additional PMePh2 and by oxidative
addition of CO2 to 1,22 respectively.). Compounds 4b,c,d,e were
synthesized by published procedures from WCl6 and the corre-
sponding phosphine.27b,28–29 All reactions and manipulations were
conducted under an atmosphere of Ar or N2 unless indicated
otherwise.
2
1
cis to PMePh2), 25.91 (d of d, JPP = 17.8 Hz, 211.8 Hz; JPW
=
=
2
327.84 Hz, PCH2CH2P trans to PMePh2), 0.50 (d of d, JPP
4.5 Hz, 211.8 Hz, 1JPW = 346.6 Hz, PMePh2). IR (KBr, cm-1): 3050
(w), 1481 (m), 1435 (s), 1092 (m), 940 (m), 742 (m), 742 (s), 690
(vs), 525 (m).
trans-[WCl4(PMe3)2], 4a. The following is a simplified alter-
native to a literature preparation for 4a.30 A 100 mL Schlenk flask
with stir bar was charged with [WCl4]• (1.00 g, 3.07 mmol) and
30 mL of CH2Cl2 under an Ar atmosphere. To this mixture was
slowly added a solution of PMe3 in toluene (6.2 mL, 6.2 mmol) via
syringe at ambient temperature. The resulting dark red reaction
mixture was stirred for 4 h at ambient temperature. The solvent was
removed under reduced pressure to afford a brownish red solid.
The formation of a mixture of [WCl4(PMe3)2] and [WCl4(PMe3)3]
1
was indicated by H NMR spectroscopy. The flask containing
the brownish red solid was heated with an oil bath to 120 ◦C
under vacuum for 24 h. During this period, an orange solid
progressively deposited on the sides of the flask. After being cooled
to room temperature, the flask was taken inside a N2 box. The dark
orange solid was collected from the walls of the flask, dissolved in
50 mL of CH2Cl2 and filtered through a Celite pad on a medium
porosity frit. The dark orange filtrate was reduced to a volume
of 20 mL and cooled to -30 ◦C for 24 h to afford square-shaped
red crystals, which were collected by filtration, washed with 2 ¥
10 mL of hexanes and dried under vacuum for 24 h. Yield: 0.645 g,
1.35 mmol, 44%. 1H NMR (400 MHz, CDCl3): d -26.67 (s). Anal.
Calcd for C6H18P2Cl4W: C, 15.08; H, 3.80; P, 12.96. Found: C,
15.16; H, 3.64; P, 13.01.
Chart 1 Numbering system for compound identification.
[WOCl2(PMePh2)3],
3c. Preparations
of
3c
from
WCl6/PMePh2 in EtOH28 and from [WCl4(PMePh2)] in wet
acetone or THF30 have been described earlier. A solution of
[WIVO(CO)Cl2(PMePh2)2] (0.201 g, 0.287 mmol) and PMePh2
(0.258 mL, 0.278 g, 1.39 mmol) in toluene (40 mL) under Ar was
heated to 55 ◦C and bubbled continuously with a slow stream of
Ar over an 8 h period. During this time, the initial reddish-purple
color of the solution gradually changed to a mauve color. The
toluene solution was cooled to room temperature and evaporated
to dryness under reduced pressure, and the resulting solid residue
was washed with Et2O (5 ¥ 10 mL). Recrystallization from
benzene–hexane produced block-shaped crystals with a blue-red
dichroic aspect. Yield: 0.192 g, 0.220 mmol, 77%. 31P NMR (d,
ppm in C6D6): 1.05 (s, JPW = 336 Hz), -15.75 (s, JPW = 415 Hz).
Conversion of [WOCl2(PMePh2)3] (3c) to [WCl4(PMePh2)2] (4c).
A 200 mL three-neck flask with a stir bar was charged with
[WOCl2(PMePh2)3] (0.330 g, 0.38 mmol) and connected to a
Schlenk line. The middle neck was fitted to a special glass tube for
the delivery of gaseous HCl, and the remaining neck was closed
by a septum. The system was placed under Ar, and 50 mL of
dry toluene were added to the flask via syringe. A mauve color
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9662–9671 | 9663
©

View Article Online
solution resulted, and the mixture was stirred for several minutes
until the solid completely dissolved. Gaseous HCl was bubbled
through the solution and vented through the Schlenk line for 1 h,
during which time the solution assumed a bright yellow color. The
solvent was removed under reduced pressure. The resulting wet
solid was washed with dry pentane (3 ¥ 10 mL) and dried in vacuo
overnight. Diffusion of hexanes vapor into a THF solution of the
product produced single crystals with unit cell parameters that
matched those obtained in an earlier structural characterization
with Et2O (4 ¥ 10 mL), and dried in vacuo. Yield: 0.152 g,
1
0.21 mmol, 15%. H NMR (400 MHz, C6D6): d 8.76 (d, JHH
=
7.6 Hz, 8H, Ph), 8.51 (t, JHH = 7.4 Hz, 8H, Ph), 7.54 (t,
JHH = 7.6 Hz, 4H, Ph), -18.00 (s, 6H PMePh2). NMe2 peaks
were not visible. 31P NMR (d, ppm in C6D6): 0.52 (s, JPW
=
337 Hz, PMePh2), -16.34 (s, JPW = 416 Hz, signal tentatively
attributed to a ditungsten impurity). UV/vis (C6H6), lmax (eM):
~308 nm (sh, 7000), ~362 nm (sh, 2600). MS (ESI, + mode): 699.3
(M+ – Cl). Electrochemistry: [WCl3(NMe2)(PMePh2)2]1+ + e- →
1
of [WCl4(PMePh2)2]. Yield: 0.240 g, 0.33 mmol, 87%. H NMR
[WCl3(NMe2)(PMePh2)2], (reversible, E 1 = +0.52 V vs. Ag/AgCl,
2
(400 MHz, CDCl3): d 11.26 (d, 8H, Ph), 8.35 (t, 8H, Ph), 8.02
(t, 4H, Ph), -27.42 (s, 6H, PMePh2); UV/vis (THF), lmax (eM):
306 nm (5383), 410 nm (1449). Anal. Calcd for C26H26P2Cl4W:
C, 43.01; H, 3.61; P, 8.53. Found: C, 42.81; H, 3.52; P,
8.78.
-0.02 V vs. [Cp2Fe]+/Cp2Fe, DEp = 0.101 V). Anal. Calcd for
C28H32Cl3NP2W: C, 45.77; H, 4.39; N, 1.91. Found: C, 42.47: H,
4.28; N, 1.27.
Physical methods
[Cp*2Co][WCl4(PMePh2)2], [Cp*2Co][4c] and [Cp*2Co]-
[WCl5(PMePh2)], [Cp*2Co][5]. A 200 mL Schlenk flask with a
stir bar was charged with [WCl4(PMePh2)2] (0.200 g, 0.28 mmol)
UV-vis spectra (molar absorptivities reported in M-1 cm-1)
were obtained at ambient temperature with a Hewlett-Packard
8452A diode array spectrometer, while IR spectra were taken
as pressed KBr pellets with a Thermo Nicolet Nexus 670 FTIR
instrument in absorption mode. All NMR spectra were recorded
at 25 ◦C with a Varian Unity Inova spectrometer operating
at 400 and 161.8 MHz for 1H and 31P, respectively. Spectra
and 40 mL of dry toluene. A solution of Cp* Co (0.090 g,
2
0.27 mmol) in 30 mL dry toluene was transferred via cannula
to the [WCl4(PMePh2)2] solution. Formation of a light brown
precipitate was observed. This mixture was stirred for 2 h, and
the solvent was reduced to ~10 mL under reduced pressure.
The precipitate was collected on a frit under Ar and dried
overnight (0.230 g). Diffusion of diethyl ether vapor into a
1,2-dichloroethane solution produced a mixture of orange
prism and pale yellow plate crystals, which were identified
1
were referenced to the solvent residual for H and to external
85% H3PO4 for 31P. Electrochemical measurements were made
with a CHI620C electroanalyzer workstation using a Ag/AgCl
reference electrode, a platinum disk working electrode, Pt wire as
auxiliary electrode, and [Bu4N][PF6] as the supporting electrolyte.
The [Cp2Fe]+/Cp2Fe couple occurred at +0.54 mV (by CV)
in 0.10 M [Bu4N][PF6] in CH2Cl2. Elemental analyses were
performed by Midwest Microlab, LLC of Indianapolis, IN. Details
regarding methods of crystallization, X-ray diffraction data col-
lection, and structure solution and refinement are deferred to the
ESI.†
by X-ray crystallography as [Cp* Co][WCl4(PMePh2)2] and
2
[Cp* Co][WCl5(PMePh2)], respectively. A manual separation of
2
crystals produced 3.0 mg of [Cp*2Co][4c], 0.9% yield, and 3.3 mg
[Cp*2Co][5], 1.1% yield (based on tungsten). Approximately
0.025 g of mixed crystals remained, which were not separable by
1
hand. [Cp*2Co][4c]: H NMR (400 MHz, CDCl3): d 11.26 (d,
JHH = 7.2 Hz, 8H, Ph), 8.36 (t, JHH = 7.4 Hz, 8H, Ph), 8.03 (t,
JHH = 7.4 Hz, 4H, Ph), 1.75 (s, 30H, [(C5(CH3)5)2Co]+), -27.40
(s, 6H, PMePh2). UV/vis (1,2-dichloroethane), lmax (eM): 416 nm
(1450), ~470 nm (sh, 460). MS (MALDI-TOF): 726.0 (M-, -
mode)), 329.2 [(C5(CH3)5)2Co]+ (+ mode). HRMS (MALDI-
TOF) monoisotopic m/z: 723.9785 (Calcd for C26H26Cl4P2W
(M-) 723.9773). [Cp*2Co][5]: 1H NMR (400 MHz, CDCl3): d
12.09 (d, JHH = 7.2 Hz, 4H, Ph), 8.59 (t, JHH = 7.2 Hz, 4H, Ph),
8.02 (d, JHH = 7.2 Hz, 2H, Ph), 1.70 (s, 30H, [(C5(CH3)5)2Co]+),
-19.21 (s, 3H, PMePh2). UV/vis (1,2-dichloroethane), lmax (eM):
~344 nm (sh, 3800), ~388 nm (sh, 1400). MS (MALDI-TOF):
555.2 (M-). HRMS (MALDI-TOF) monoisotopic m/z: 558.8701
(Calcd for C13H13Cl5PW (M-) 558.8707). Electrochemistry:
Computational methods
Geometries have been optimized and final energies have been eval-
uated in DFT calculations, using the PBE exchange–correlation
functional as prescribed by Perdew, Burke and Ernzerhof.31 For all
calculations, tungsten was modeled by a small-core, quasirelativis-
tic effective core potential with an associated (8s7p6d)/[6s5p3d]
valence basis set contracted according to a (311111/22111/411)
scheme.32 All main group elements except for those that are part
of a phenyl group have been described by contracted Gaussian
basis sets of triple zeta valence quality augmented by polarization
functions (TZVP).33 For phenyl groups, the C atom bonded to
phosphorus has been described by a TZVP basis, whereas the
remaining atoms were described by contracted Gaussian split-
valence basis sets (SV).34 The nature of stationary points located
on the potential energy surface has been checked by harmonic
frequency calculations, and all molecules considered in this work
possess a harmonic spectrum without any imaginary frequencies.
Enthalpies H have been evaluated at zero K from total energies
E and zero-point energy correction terms (ZPE). The calculated
frequency data provide thermodynamic corrections to calculate
free energies G at 298 K. The Gaussian 03 suite of programs
constitutes the underlying computational engine of the present
work.35
[WCl5(PMePh2)] + e- → [WCl5(PMePh2)]1- (quasireversible, E 1
=
2
+0.69 V vs. Ag/AgCl, +0.15 V vs. [Cp2Fe]+/Cp2Fe).
[WCl3(NMe2)(PMePh2)2], 6. A solution of 4c (0.976 g,
1.3 mmol) in toluene (70 mL) was treated with neat P(NMe2)3
via syringe (0.51 mL, 0.45 g, 2.8 mmol). The resulting mixture
was allowed to stir 12 h at ambient temperature, during which
time a white precipitate and orange-yellow supernatant formed.
The white precipitate was removed by filtration, and the filtrate
was concentrated to a volume of 15 mL under reduced pres-
sure. Diethyl ether (50 mL) was layered on top of the toluene
concentrate. Small yellow prism-shaped crystals deposited over
a period of 24–48 h, which were isolated by filtration, washed
9664 | Dalton Trans., 2010, 39, 9662–9671
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Results and discussion
The [WCl2(PMePh2)4] system, which was studied in some detail
by Mayer in oxidative addition reactions with various heterocu-
mulenes, is to the best of our knowledge, the only instance in
which a terminal oxo ligand results from the reduction of CO2
to CO by a metal complex. Other outcomes for this reaction
type are the formation of sterically protected oxo-bridged di- or
trimetallic species21,23 or products in which oxygen atom is inserted
into another metal–ligand bond.36 Despite the formidable strength
of the W O bond in [WIVO(CO)Cl2(PMePh2)2], we believed that
the terminal oxo-ligand resulting from oxidative addition of CO2
in this system could be amenable to some form of electrophilic
attack, ligand exchange, or other pathway leading back to reduced
[WIICl2(PMePh2)4].
Beginning with air- and moisture-sensitive [WIICl2(PMePh2)4],
we re-examined the CO2 reactions described by Mayer and, as
reported, found the oxidative addition reaction with CO2 to
occur in a timescale of approximately ten minutes under the
mild conditions of 1 atm pressure CO2 and ambient temperature
(1 → 2, Scheme 1). The preparation, isolation, and handling of
[WIICl2(PMePh2)4] was most effectively accomplished under an
Ar atmosphere up to the moment of CO2 introduction. The oxo
carbonyl species (2 in Scheme 1) that is the immediate product
of oxidative addition of CO2 is relatively robust and isolable in
spectroscopically pure form. Displacement of CO by PMePh2
liberated in the initial oxidative addition reaction creates
O + H2S → M S + H2O 51
O + RNH2 → M NR + H2O 52
O + 2HCl → MCl2 + H2O 53
(4)
M
(5)
(6)
M
M
an equilibrium that was reported by Mayer as being a 9 : 1 mixture
favoring the oxo-carbonyl species, 2.22c Complete formation of
[WIVOCl2(PMePh2)3] (3c in Scheme 1) is readily accomplished by
mild heating of [WIVO(CO)Cl2(PMePh2)2] in the presence of excess
PMePh2 under a slow purge of Ar gas that sweeps away substituted
CO(g). With the chelating effect enjoyed by dppe, the displacement
of CO from isolated 2 is accomplished without the additional drive
afforded by excess phosphine and Ar purge to quantitatively afford
[WIVOCl2(dppe)(PMePh2)], 3e. While structural determinations
for 122b and 222a have been previously reported, atomic coordinates
useful for DFT calculations are unavailable. Thus, new, higher
resolution structural determinations were undertaken (Scheme 1,
Table 1); unit cells obtained for 1 and 2 are different than
those reported by Mayer owing to differences with co-crystallized
solvent molecules.
The reduction of [WIVOCl2(PMePh2)3] back down to
[WIICl2(PMePh2)4] is a challenge that may be addressed in
several distinctly different ways. One approach is via successive
coupled proton–electron transfer reactions that, by analogy
to the mechanism by which molybdenum and tungsten
oxotransferase and hydroxylase enzymes37–38 are known to
function, ultimately remove the oxo ligand as H2O. Another
approach is by direct oxygen atom excision, a two electron
reduction process, using an oxygen atom acceptor more oxophilic
even than W(II). Although phosphines are common oxygen
atom acceptors from transition metal complexes,39 a report
Scheme 1 A W-mediated cycle for CO2-to-CO reduction by direct O
atom excision. The oxo-for-chloride ligand exchange and the two electron
reduction that regenerates active W(II) are accomplished as separate
steps. All compounds are isolable and have been characterized by X-ray
crystallography. Thermal ellipsoid plots are drawn at the 50% probability
level. H atoms are omitted for clarity.
of oxygen atom abstraction from Ph2P(O)CH2CH2PPh2
in
[WIICl2(Ph2P(O)CH2CH2PPh2)(PMePh2)2]
to
afford
[WIVOCl2(Ph2PCH2CH2PPh2)(PMePh2)] indicates that the
thermodynamics are unfavorable for this process for phosphines
with basicity comparable to PMePh2.40 Available thermodynamic
data41 suggest that silanes such as Me3SiSiMe3 may have the
thermodynamic strength to compete with low valent tungsten
for an oxygen atom. Furthermore, direct silylation of a terminal
metal-oxo ligand by an electrophilic silane reagent, with a
concomitant lowering of the metal-oxo bond order, is amply
precedented.42–49 However, efforts to apply this reaction to
[WIVOCl2(PMePh2)3] using Me3SiI or Me3SiOTf resulted in facile
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9662–9671 | 9665
©

View Article Online
Table 1 Crystal and refinement data for compounds 1, 2, 3c,e, 4a,b,c,d,e, [4c]1-, [5]1-, and 6
Compound
Formula
FW
1·C6H5CH3
C59H60Cl2P4W
1147.70
2·0.5C6H6
C30H29Cl2O2P2W
738.22
3c
3e·CH2Cl2
C40H39Cl4OP3W
954.27
C39H39Cl2OP3W
871.36
Crystal system
Space group
Color, habit
Triclinic
P1
Monoclinic
P21/n
Triclinic
P1
Monoclinic
C2/c
¯
¯
Orange block
11.4069(8)
14.3468(9)
17.092(1)
76.523(1)
71.996(1)
84.321(1)
2585.9(3)
100
Purple column
9.5417(6)
14.6428(9)
20.463(1)
90
90.775(1)
90
2858.8(3)
100
4
Purple column
9.8636(8)
10.3088(9)
18.707(2)
78.530(1)
83.485(1)
73.102(1)
1780.5(3)
100
Blue plate
23.532(2)
14.265(1)
25.620(2)
90
105.896(1)
90
8271(1)
100
8
˚
a/A
˚
b/A
˚
c/A
a (◦)
b (◦)
g (◦)
3
˚
V/A
T/K
Z
2
2
R1,a wR2 b
GoF
0.0194, 0.0466
1.048
0.0233, 0.0456
1.066
0.0217, 0.0493
1.046
0.0265, 0.0618
1.050
Compound
Formula
FW
4a
4b
4c
4e
C6H18Cl4P2W
477.79
C16H22Cl4P2W
601.93
C26H26Cl4P2W
726.06
C26H24Cl4P2W
724.04
Crystal system
Space group
Color, habit
Monoclinic
P21/c
Monoclinic
P21/c
Triclinic
P1
Monoclinic
P21/c
¯
Orange block
6.5351(6)
8.4615(8)
13.326(1)
90
99.717(1)
90
726.3(1)
100
2
Orange block
9.644(1)
13.781(2)
8.203(1)
90
106.744(2)
90
1044.0(2)
100
2
Orange block
8.860(3)
9.491(3)
9.765(3)
66.983(4)
89.510(4)
63.041(4)
659.3(3)
100
Orange plate
11.300(4)
21.711(9)
13.943(5)
90
109.633(4)
90
3222(2)
100
4
˚
a/A
˚
b/A
˚
c/A
a (◦)
b (◦)
g (◦)
3
˚
V/A
T/K
Z
1
R1,a wR2 b
GoF
0.0208, 0.0605
1.144
0.0146, 0.0336
1.032
0.0152, 0.0388
1.048
0.0548, 0.1278
1.065
Compound
Formula
FW
[Cp*2Co][4c]
C46H56Cl4CoP2W
1055.43
Monoclinic
C2/c
Orange block
27.23(1)
11.087(4)
17.943(6)
90
125.829(4)
90
4392(3)
100
4
[Cp*2Co][4c]
C46H56Cl4CoP2W
1055.43
Monoclinic
P21/n
Orange column
9.600(2)
11.047(2)
21.228(3)
90
93.333(2)
90
2247.3(6)
100
2
[Cp*2Co][5]
C33H43Cl5CoPW
890.67
6
C28H32Cl3NP2W
734.69
Crystal system
Space group
Color, habit
Triclinic
Monoclinic
C2/c
¯
P1
Yellow plate
9.654(1)
10.162(1)
17.949(2)
80.677(2)
80.642(2)
89.713(2)
1714.2(3)
100
Yellow column
13.8088(8)
13.2473(8)
16.0074(9)
90
93.997(1)
90
2921.1(3)
100
4
˚
a/A
˚
b/A
˚
c/A
a (◦)
b (◦)
g (◦)
3
˚
V/A
T/K
Z
2
R1,a wR2 b
GoF
0.0414, 0.1039
1.340
0.0196, 0.0442
1.033
0.0385, 0.0967
1.030
0.0161, 0.0339
1.065
ꢀ
ꢀ
ꢀ
ꢀ
a R1 = ꢀFo| - |Fcꢀ/ |Fo|. b wR2 = { [w(Fo - Fc2)2]/ [w(Fo2)2]}1/2; w = 1/[s (Fo2) + (aP)2 + bP], where P = [2Fc + Max(Fo2,0)]/3.
2
2
2
chloride abstraction to afford Me3SiCl instead of addition of the
Me3Si+ electrophile to the WIV O ligand. Although an undesired
outcome here, chloride abstraction by electrophilic silylium
cations has been described by Heinekey and coworkers as useful
for the preparation of otherwise inaccessible iridium and rhodium
hydrido complexes.50
Another reaction type to which metal oxo ligands are subject is
ligand exchange without reduction of the metal. Because these
reactions are not attended by a change in oxidation state at
the metal, they are often, if not generally, easier to accomplish
than direct atom abstractions. The reaction of H2S or a primary
amine (RNH2) upon a terminal metal oxo ligand, M O, is
widely applicable for the synthesis of corresponding terminal
sulfido (M S)51 and imido (M NR) compounds52 (eqn (4)–(5)).
Analogous to the foregoing reaction types, but less extensively
developed, are halide-for-oxo exchange reactions (eqn (6)).53
Treatment of [WIVOCl2(PMePh2)3] with anhydrous HCl(g)
does indeed effect chloride-for-oxo exchange (Scheme 1) and
remove the oxo ligand as H2O to produce the known compound
trans-[WCl4(PMePh2)2], a species that is easily identified by its
distinctive NMR spectrum.27b,28 We have also identified trans-
[WCl4(PMePh2)2] by X-ray crystallography since its structure has
9666 | Dalton Trans., 2010, 39, 9662–9671
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
II
˚
Table 2 Selected bond lengths (A), angles (deg.) for [W Cl2(PMePh2)4], compound 1
W(1)–Cl(1)
W(1)–Cl(2)
W(1)–P(1)
W(1)–P(2)
W(1)–P(3)
W(1)–P(4)
2.4270(4)
2.4289(4)
2.5282(5)
2.5199(5)
2.5191(5)
2.5214(5)
Cl(1)–W(1)–Cl(2)
Cl(1)–W(1)–P(1)
Cl(1)–W(1)–P(2)
Cl(1)–W(1)–P(3)
Cl(1)–W(1)–P(4)
Cl(2)–W(1)–P(1)
Cl(2)–W(1)–P(2)
Cl(2)–W(1)–P(3)
178.78(2)
83.14(2)
96.21(2)
84.12(2)
95.58(2)
98.05(2)
84.11(2)
94.70(2)
Cl(2)–W(1)–P(4)
P(1)–W(1)–P(2)
P(1)–W(1)–P(3)
P(1)–W(1)–P(4)
P(2)–W(1)–P(3)
P(2)–W(1)–P(4)
P(3)–W(1)–P(4)
84.12(2)
90.09(2)
167.26(2)
90.85(2)
91.04(2)
168.20(2)
90.63(2)
IV
˚
Table 3 Selected bond lengths (A), angles (deg.) for [W O(CO)Cl2(PMePh2)2], compound 2
W(1)–O(1)
W(1)–C(27)
W(1)–Cl(1)
W(1)–Cl(2)
W(1)–P(1)
W(1)–P(2)
C(27)–O(2)
1.713(2)
2.043(2)
2.4326(7)
2.4964(7)
2.5284(7)
2.5429(7)
1.135(3)
O(1)–W(1)–C(27)
O(1)–W(1)–Cl(1)
O(1)–W(1)–Cl(2)
O(1)–W(1)–P(1)
O(1)–W(1)–P(2)
W(1)–C(27)–O(2)
C(27)–W(1)–Cl(1)
C(27)–W(1)–Cl(2)
88.3(1)
C(27)–W(1)–P(1)
C(27)–W(1)–P(2)
Cl(1)–W(1)–Cl(2)
Cl(1)–W(1)–P(1)
Cl(1)–W(1)–P(2)
Cl(2)–W(1)–P(1)
Cl(2)–W(1)–P(2)
P(1)–W(1)–P(2)
92.28(8)
91.62(8)
92.75(2)
87.37(2)
85.61(2)
80.04(2)
81.01(2)
159.42(2)
100.83(7)
166.33(7)
98.88(7)
101.43(7)
178.6(3)
170.83(8)
78.18(8)
IV
˚
Table 4 Selected bond lengths (A), angles (deg.) for [W OCl2(PMePh2)3], compound 3c
W(1)–O(1)
W(1)–Cl(1)
W(1)–Cl(2)
W(1)–P(1)
W(1)–P(2)
W(1)–P(3)
1.716(2)
O(1)–W(1)–Cl(1)
O(1)–W(1)–Cl(2)
O(1)–W(1)–P(1)
O(1)–W(1)–P(2)
O(1)–W(1)–P(3)
Cl(1)–W(1)–Cl(2)
Cl(1)–W(1)–P(1)
Cl(1)–W(1)–P(2)
97.99(6)
172.19(6)
99.25(6)
88.03(6)
95.26(6)
89.46(2)
80.54(2)
173.54(2)
Cl(1)–W(1)–P(3)
Cl(2)–W(1)–P(1)
Cl(2)–W(1)–P(2)
Cl(2)–W(1)–P(3)
P(1)–W(1)–P(2)
P(1)–W(1)–P(3)
P(2)–W(1)–P(3)
86.63(2)
84.24(2)
84.63(2)
82.77(2)
96.24(2)
161.80(2)
95.18(2)
2.4668(6)
2.4915(6)
2.5530(6)
2.5172(6)
2.5463(6)
IV
IV
1- a
˚
Table 5 Selected bond lengths (A), angles (deg.) for [W Cl4(PMePh2)2], 4c, and [Cp*2Co][W Cl4(PMePh2)2], [4c] .
4c
[4c]1-
4c
[4c]1-
W(1)–Cl(1)
W(1)–Cl(2)
W(1)–P(1)
P(1)–Caryl
P(1)–Cmethyl
2.3379(7)
2.3466(6)
2.5672(7)
1.821[1]
2.4355(6)
2.4092(6)
2.5303(6)
1.829[1]
Cl(1)–W(1)–Cl(1A)
Cl(1)–W(1)–Cl(2)
Cl(1)–W(1)–Cl(2A)
Cl(1)–W(1)–P(1)
Cl(1)–W(1)–P(1A)
Cl(2)–W(1)–Cl(2A)
Cl(2)–W(1)–P(1)
Cl(2)–W(1)–P(1A)
P(1)–W(1)–P(1A)
180
180
89.84(3)
90.16(3)
92.10(2)
87.90(2)
180
90.34(3)
89.66(3)
180
89.41(2)
90.59(2)
88.71(2)
91.29(2)
180
93.93(2)
86.07(2)
180
b
1.817(2)
1.820(2)
a Data are for the P21/n polymorph. b Average of two P–Caryl bond lengths.
not been previously reported. This transformation of 3c → 4c
(Scheme 1) has been performed on a scale of 0.20 g with an isolated
yield of 87%, product identity and purity being confirmed spectro-
scopically and analytically. A particularly fortunate aspect of this
reaction is that [WCl4(PMePh2)2], which is an easily handled air-
stable molecule, is the immediate precursor to [WCl2(PMePh2)4]
and is transformed to it by Na/Hg reduction in the presence of
PMePh2.27 Thus, a tungsten-mediated reactivity cycle for CO2-
to-CO reduction, with an oxygen atom removed as H2O, is
formed in a series of stoichiometric steps that are well-defined
and unambiguous. The tungsten species involved at each point
along the way is definitively established by X-ray crystallography.
Crystal and refinement data for 1, 2, 3c and 4c, are presented
in Table 1, while selected metrical parameters are collected in
Tables 2–5, respectively.
H2O + 2Cl-. The essential reducing equivalents for CO2-to-CO
reduction are provided, at present, by sodium metal. However,
the reduction of 4c to 1 in Scheme 1 should be feasible with
reducing agents less potent than Na0 metal or, in principle, via
bulk electrolytic means at an electrode surface that is poised
at the minimal potential necessary. Redox potentials for simple
chlorophosphine complexes of the type [WCl4(PMe3-xPhx)2] (x =
0–3) appear not to have been previously reported. With the
aim of disclosing the minimal energetic parameters within which
the closed cycle for CO2-to-CO reduction that is defined by
Scheme 1 might be made to operate, a cyclic voltammetry study
of the series [WCl4(PMe3-xPhx)2] (x = 0–3) was undertaken. This
electrochemical study establishes specific W(IV) → W(III) and
W(III) → W(II) reduction potentials and clarifies the extent to
which the window of redox potentials may be manipulated in these
simple chlorophosphine complexes by variation of the phosphine
substituents.
The overall balanced equation obtained from adding the
individual steps of Scheme 1 is CO2 + 2e- + 2HCl → CO +
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9662–9671 | 9667
©

View Article Online
IV
˚
Table 6 Selected bond lengths (A) and angles (deg.) for [W Cl3(NMe2)(PMePh2)2], 6
W(1)–Cl(1)
W(1)–Cl(2)
W(1)–P(1)
W(1)–N(1)
N(1)–C(14)
2.3953(4)
2.4500(6)
2.5638(5)
1.928(2)
1.470(2)
Cl(1)–W(1)–Cl(1A)
Cl(1)–W(1)–Cl(2)
Cl(1)–W(1)–P(1)
Cl(1)–W(1)–P(1A)
Cl(1)–W(1)–N(1)
Cl(2)–W(1)–P(1)
176.58(2)
88.29(1)
92.67(2)
86.99(2)
91.71(1)
84.38(1)
Cl(2)–W(1)–N(1)
P(1)–W(1)–P(1A)
P(1)–W(1)–N(1)
W(1)–N(1)–C(14)
C(14)–N(1)–C(14A)
180
168.76(2)
95.62(1)
125.6(1)
108.9(2)
138.9
a
j
a
j = Angle between W(1)–Cl(1)–Cl(1A)–Cl(2)–N(1) and N(1)–C(14)–C(14A) mean planes.
The chlorophosphine complexes trans-[WCl4(PMe3-xPhx)2] (x =
0–3, compounds 4a–4d, respectively) and [WCl4(dppe)] (4e), were
prepared by literature methods either from WCl4 or from WCl6
with excess phosphine or Zn metal as reductant.27b,28–29 Synthesis
of trans-[WCl4(PMe3)2] is invariably attended by formation of
the related seven-coordinate species [WCl4(PMe3)3].30,54 As its
preparation in pure form has only been via a sealed tube reaction,30
a more convenient, simplified, synthesis procedure is described
here in some detail. The identity of these compounds has been
established by X-ray crystallography (Fig. S1†). Crystallographic
data are deferred to ESI.† Only trans-[WCl4(PPh3)2] has been
previously identified by X-ray diffraction.55
In the continuum of available phosphine ligands, those of
the type P(Me3-xPhx)3 (x = 0–3) exert appreciably less p-acidity
than phosphines of the type P(OR)3 (R = alkyl, aryl), P(NR2)3,
and PX3 (X = halide).56 The new tungsten chlorophosphine
complex [WCl4(P(NMe2)3)2] was thus anticipated as a complex
that would reveal less negative W(IV) → W(III) and W(III) →
W(II) couples because of the electron-withdrawing character of
the supporting phosphine. Efforts to prepare [WCl4(P(NMe2)3)2]
from [WCl4(PMePh2)2] via simple ligand displacement instead
resulted in facile chloride-for-amido ligand exchange to produce
[WCl3(NMe2)(PMePh2)2], 6, and, possibly, PCl(NMe2)2. Multiple
chloride-for-amido ligand exchanges were observed in reactions
of WCl6 with P(NMe2)3. A similar chloride-for-dimethylamido
ligand exchange at tungsten has been described using Me2N–
SiMe3.57
Compound 6 has been identified structurally (Tables 1 and 6,
Fig. 1), and its structure reveals the NMe2 plane to be oriented at a
~139◦ angle with respect to the WCl3N mean plane, presumably a
balance between the conflicting demands for optimal p-donation
into the tungsten t2g orbitals and minimization of steric crowding
with the phosphine ligands. The appreciably greater p-donor
ability of the NMe2 ligand is underscored by the absence of any
reversible reduction for 6, as was seen for 4c, but instead the
occurrence of a reversible oxidation corresponding to formation
of [WVCl3(NMe2)(PMePh2)2]1+ (0.51 V, Fig. S2†).
Although highly crystalline, samples of 6 appeared, on the
combined evidence of elemental analysis, NMR spectroscopy, and
mass spectrometry to be adulterated both with small amounts of
an unidentified organophosphine species and with a ditungsten
species of color similar to 6, which is tentatively formulated
as the edge-sharing bioctahedron [(Ph2MeP)WCl3(m-NMe2)(m-
Cl)WCl3(PMePh2)]. Recrystallization was ineffective in removing
these minor species, nor was an exhaustive screening of crystals
by X-ray diffraction successful in identifying anything different
from 6.
Fig.
1
Thermal ellipsoid plots at the 50% probability level
for (a) [WCl3(NMe2)(PMePh2)2], 6, and (b) P21/n polymorph of
[Cp*2Co][WCl4(PMePh2)2]. H atoms are omitted for clarity.
[WCl4(dppe)] (4e) in THF. Little effect is seen upon the potentials
for the W(IV) → W(III) reduction for the trans-[WCl4(PMe3-xPhx)2]
(x = 0–3) series, consistent with the orbital into which the electron
is added being a tungsten dxz,dyz orbital. The W(IV) → W(III)
reduction generally occurs at quite mild potentials and reveals
reversible behavior (Fig. 2); the mildest reduction potential is
observed for [WCl4(dppe)]. The observed DEp of 0.10 V and ia/ic
of 1.1 for 4c + e- → [4c]1- (Fig. 2), while greater than the theoretical
values of 0.059 V and 1.0, respectively, for a fully reversible process,
are typically observed for [Cp2Fe]+ + e- → Cp2Fe under these same
conditions with our system. For this reason, and because [4c]1- can
be chemically prepared and isolated (vide infra), we attribute the
departure of DEp from its ideal value to uncompensated solution
resistance effects rather than to chemical instability. Further, a
Table 7 summarizes the reduction potentials measured for trans-
[WCl4(PMe3-xPhx)2] (x = 0–3, compounds 4a–4d, respectively) and
9668 | Dalton Trans., 2010, 39, 9662–9671
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Table 7 Reduction potentials, E (V vs. Ag/AgCl), for [WIVCl4(PRR¢2)2]
1
The mild potentials and reversible nature found for the
W(IV) → W(III) couples suggested the feasibility of isolating
a [WIIICl4(PMePh2)2]1- anion, a species which has not been
previously isolated. Introduction of a toluene solution of Cp2*Co
to a solution of [WCl4(PMePh2)2] induces immediate formation
of a light brown precipitate. Isolation and crystallization of this
solid from 1,2-dichloroethane/ether produced a mixture of orange
columnar and pale yellow plate crystals, both of which were
interrogated by X-ray diffraction. The former proved to be the
intended [Cp*2Co][WIIICl4(PMePh2)2], which was found to occur
in both primitive (P21/n) and C-centered (C2/c) polymorphs.
An ellipsoid plot of [Cp*2Co][4c] in the P21/n polymorph is
presented in Fig. 1. The average W–Cl bond distance in this
anion is 2.4195[7] A, in contrast to the value of 2.3423[5] A
for the corresponding charge neutral compound 4c. This change
is consistent with a diminution of chloride p-bonding into the
t2g-type orbitals due to the presence of an additional tungsten
d electron. Further, the W–P and P–C bond distances slightly
decrease and increase, respectively, changes consistent with p-acid
character of the phosphine ligand.58 The pale yellow species was
determined to be [Cp*2Co][WIVCl5(PMePh2)], [Cp*2Co][5], also a
new chlorophosphine complex of tungsten (See ESI†).
2
complexes.a
W(IV)→W(III)b
W(IV)→W(III)c
W(III)→W(II)
[WCl4(PMe3)2]
[WCl4(PMe2Ph)2]
[WCl4(PMePh2)2]
[WCl4(PPh3)2]
[WCl4(dppe)]
-0.40
-0.35
-0.36
-0.44
-0.21
-0.39
-0.37
-0.37
-0.45
-0.22
-1.69 cd
-1.70 cd
-1.60 de
-1.43 cd
-1.43 cd
a Solvent and supporting electrolyte were THF and [Bu4N][PF6]. All data
were obtained under a blanketing atmosphere of Ar. Under the conditions
employed, the [Cp2Fe]1+ + e- → Cp2Fe coupled occurred at +0.54 V. b Data
1
obtained by cyclic voltammetry (CV) by the relationship E = (Ea + Ec)/2.
2
c
1
2
Data obtained by differential pulse voltammetry by the relationship E
=
E
max + DE/2, where DE is the pulse amplitude. d All W(III) → W(II) couples
˚
˚
were irreversible. e Estimated from CV.
An important aspect of the closed reaction cycle illustrated
in Scheme 1 is a quantification of the thermodynamics for
each step. The reaction enthalpies for each step were evaluated
computationally in the gas phase for tungsten complexes bearing
the PH3 and PMe3 supporting ligands as well as the PMePh2 ligand
that is used for the reactions of Scheme 1. The first reaction, labeled
as (Rxn 1) on the x axis in Fig. 3, is the oxidative addition reaction
[WIICl2(phosphine)4] + CO2 → [WIVO(CO)Cl2(phosphine)2] +
2(phosphine). For convenience, since it is reaction enthalpies that
are of immediate interest, the enthalpy of [WIICl2(phosphine)4] +
CO2 is assigned a relative value of zero. Reaction 1 is exothermic
for all three systems (Fig. 3), but the calculated exothermicity
increases by ~50 kJ mol-1 in going from [WIICl2(PH3)4] to
[WIICl2(PMePh2)4]. A possible reason for the greater exothermicity
for the latter is suggested in the thermal ellipsoid plot of 1
in Scheme 1 by its appreciable distortions from an idealized
octahedral coordination geometry. The high degree of steric
crowding in [WIICl2(PMePh2)4] appears to diminish each PMePh2
ligand’s ability to form a good dative bond to tungsten and thus
Fig. 2 Cyclic voltammogram of [WCl4(PMePh2)2] in CH2Cl2 with
[Bu4N][PF6] supporting electrolyte and scan rate of 100 mV s-1. Under
these conditions, the [Cp2Fe]+ + e- → Cp2Fe coupled occurred at +0.54 V.
plot of a ic/v1/2 (ic = current in the forward (reducing) direction;
n = scan speed) is linear in the range 20–500 mV s-1 (Fig.
S3†). For [WCl4(PPh3)2] and [WCl4(PMePh2)2], voltammetric data
had to be obtained quickly owing to facile sample degradation.
The remaining compounds were stable toward repeated scanning
because the more basic (or chelating) nature of the phosphine
ligand diminishes the rate of phosphine dissociation.
Fig. 3 Calculated enthalpy profile (kJ mol-1) for reactions that are part of the cycle of Scheme 1. Reactions 1, 2 and 3 along the x axis in the plot are
as defined in the line equations at the right, where reactions with differing phosphine supporting ligands tungsten are distinguished by use of different
colors (red for PH3, blue for PMe3, green for PMePh2). All calculations are for species in the gas phase.
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9662–9671 | 9669
©

View Article Online
Table 8 Calculated bond dissociation energies, kJ mol-1
first described by Mayer and elaborated upon here has some
attraction in this context because of the mild conditions under
which it functions and because of its high specificity for CO
formation. A series of four well-defined stoichiometric steps
have been demonstrated that collectively form a closed cycle.
Significantly, the tungsten–oxygen multiple bond that results from
oxidative addition of CO2 has been shown both experimentally and
computationally to be amenable to ligand exchange. These data
indicate that, if removed in the form of H2O and substituted with
two chloride ligands, the multiply bonded terminal oxo ligand
is not at all a thermodynamic dead-end that obviates turnover.
By effecting the removal of the oxygen atom in a nonredox
ligand exchange reaction, as opposed to a redox atom abstraction
reaction, we have an insight via the electrochemistry into the
approximate overpotential diminution offered by this tungsten
system over the noncatalyzed reduction of CO2. Compared
C
O, CO2
543
610
617
624
W
W
W
O, [WOCl2(PH3)3]
O, [WOCl2(PMePh2)3]
O, [WOCl2(PMe3)3]
primes the complex for ligand dissociation and oxidative addition
of CO2. Reaction 2, [WIVO(CO)Cl2(phosphine)2] + phosphine →
[WIVOCl2(phosphine)3] + CO, is essentially isoenergetic for all
three systems. The surprisingly large endotherm calculated for
reaction 2 reflects a special stability arising from a synergistic
“push–pull” effect of the cis-oriented p-donating oxo and p-acidic
carbonyl ligands interacting through the same d orbital(s) of
tungsten.22,52
Movement from the third to fourth position along the x axis
in Fig. 3 corresponds to the enthalpy of the gas phase reaction
[WOCl2(phosphine)3] + 2HCl → [WCl4(phosphine)2] + H2O +
to a potential of ~-2.6 V (vs [Cp2Fe]+/Cp2Fe) for the direct,
1- 19
uncatalyzed reduction of CO2 to ·CO2
,
the W(II) oxidation
phosphine (Rxn 3). Despite the formidable strength of the W
O
state that is competent for oxidative addition of CO2 is accessible
from W(IV) at ~-2.1 V (vs [Cp2Fe]+/Cp2Fe).62 Thus, the lowering
of the overpotential barrier to CO2 reduction that is offered by
this tungsten system is ~0.5 V.
bond, substitution of the multiply-bonded oxo ligand for two
chloride ligands is calculated to be slightly exothermic or, at
worst, near thermoneutral (Fig. 3). While solvation effects are
certainly important to these reactions, the stabilizing solvation
effects enjoyed by reactants and products are likely to be largely
offsetting. Thus, the general trend that reaction 3 is slightly
exothermic is a plausible one. Although the two additional W–
Cl single bonds in [WCl4(phosphine)2] may not, at first thought,
appear to be enough to compensate for a W O bond and a
W←(phosphine) dative bond, it is probably the case that the
degree of p-bonding with tungsten enjoyed by each of the four
chloride ligands in [WCl4(phosphine)2] is significant enough that
these W–Cl bonds are worth appreciably more than bona fide
single bonds. The substantial W–Cl bond lengthening in reducing
4c to [4c]1- is consistent with a disruption of this p-bonding
(Table 5). The significance of this computational result is first
that it corroborates the immediate experimental observation that
the oxo-for-chloride ligand exchange does occur and is relatively
facile. More generally, the result suggests that metal oxo multiple
bonds may be transformable via nonredox ligand exchange into
intermediates that offer greater tractability for designed turnover.
Considering the perception that metal–oxo multiple bonds in
early and middle transition metal complexes are an obstacle
against turnover in a cycle for small molecule reduction, it
Although each of the reactions in Scheme 1 has some degree
of precedent, taken together they make the point that early and
middle transition metal systems such as this deserve some renewed
consideration as potential catalysts for reduction of CO2. The
perceived disadvantage of metal–ligand multiple bond formation
is, we submit, more than compensated for by the metal’s ability to
activate a thermodynamically and kinetically stable small molecule
and to effect a multi-electron reduction under mild conditions.
Among our continuing interests with the Mayer system are efforts
to perform clean, bulk electrolytic reduction of [WCl4(PMePh2)2],
or a suitable variant thereof, to W(II), an exploration of other
spectator ligand types that can support the oxidative addition of
CO2 to W(II), and a temperature-dependent spectrophotometric
study of the oxidative addition of CO2 to [WCl2(PMePh2)4] such
that an experimental Ea may be determined.
Acknowledgements
Support for this research by the Eppley Foundation of New York
and by the National Science Foundation (Grant CHE-0845829
to J.P.D.) is gratefully acknowledged. The Louisiana Board of
Regents is thanked for the support with which Tulane University’s
CCD diffractometer was obtained (Grant LEQSF-(2002-03)-
ENH-TR-67), and Tulane University is thanked for its ongoing
support of the X-ray diffraction laboratory. The authors thank the
Louisiana Optical Network Initiative (LONI) for granting access
to its computational facilities.
would be useful to have an experimentally determined W
O
bond dissociation energy. The W O bond strength has been
estimated by Mayer to be at least 577 kJ mol-1,22c,59 moderately
greater than the 531 kJ mol-1 determined for the C O bond
of CO2.60–61 We have computationally assessed the W O bond
dissociation energies in [WOCl2(PH3)3], [WOCl2(PMePh2)3] and
[WOCl2(PMe3)3] in the gas phase (Table 8) and found them to
agree reasonably well with Mayer’s estimate. These W O bond
strengths appear to correlate in a plausible way with phosphine
ligand basicity.
References
1 D. J. Darensbourg, Chem. Rev., 2007, 107, 2388–2410.
2 H. Sugimoto and S. Inoue, Pure Appl. Chem., 2006, 78, 1823–1834.
3 T. Sakakura, J.-C. Choi and H. Yasuda, Chem. Rev., 2007, 107, 2365–
2387.
Summary and conclusions
4 R. P. A. Sneedon, in Comprehensive Organometallic Chemistry, ed.
G. Wilkinson, Pergamon, New York, 1982, Vol. 8, pp. 225–283.
5 E. L. C. Cheah, L. W. Chan and P. W. S. Heng, Asian J. Pharm. Sci.,
2006, 1, 59–71.
In the face of accumulating atmospheric CO2 levels and dimin-
ishing carbon feedstocks, systems that catalytically activate CO2
are highly desirable. The tungsten-mediated CO2-to-CO reduction
9670 | Dalton Trans., 2010, 39, 9662–9671
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
6 C. M. Wai, Reprocessing Spent Nuclear Fuel with Supercritical Carbon
Dioxide, in Separations for the Nuclear Fuel Cycle in the 21st Century,
ed. G. J. Lumetta, K. L. Nash, S. B. Clark and J. I. Friese, ACS
Symposium Series 933, American Chemical Society, Washington, DC,
2006, pp 57–67.
7 G. Perretti, Supercritical Carbon Dioxide Extraction of Agricultural
and Food Processing Wastes and Byproducts, in Modern Extraction
Techniques, ed. C. Turner, ACS Symposium Series 926, American
Chemical Society, Washington, DC, 2006, Chapter 2, pp 23–35.
8 J. N. Hay and A. Khan, J. Mater. Sci., 2002, 37, 4743–4752.
9 Y. Chernyak, F. Henon, E. Hoggan, B. Novick, J. M. DeSimone and R.
Carbonell, Coatings from Liquid and Supercritical Carbon Dioxide,
in Green Chemistry Using Liquid and Supercritical Carbon Dioxide, ed.
J. M. DiSimone and W. Tumas, Oxford University Press, New York,
2003, Chapter 12, pp 193–214.
10 R. A. Sheldon, Reactions in Non-Conventional Media for Sustain-
able Organic Synthesis, in New Methodologies and Techniques for a
Sustainable Organic Chemistry, ed. A. Mordini and F. Faigl, NATO
Science Series 246. II. Math, Physics and Chemistry, Springer Verlag,
The Netherlands, 2008, pp 1–28.
11 T. Ikariya and R. Noyori, Carbon Dioxide as a Reactant and Solvent
in Catalysis, in Green Chemistry Using Liquid and Supercritical Carbon
Dioxide, ed. J. M. DiSimone and W. Tumas, Oxford University Press,
New York, 2003, Chapter 3, pp 48–63.
Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi,
C. Pomelli, J. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P.
Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D.
Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S.
Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz,
I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y.
Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. G. Johnson,
W. Chen, M. W. Wong, C. Gonzalez and J. A. Pople, GAUSSIAN 03
(Revision D.02), Gaussian, Inc., Wallingford, CT, 2004.
36 L. J. Procopio, P. J. Carroll and D. H. Berry, Organometallics, 1993, 12,
3087–3093.
37 R. Hille, Chem. Rev., 1996, 96, 2757–2816.
38 M. K. Johnson, D. C. Rees and M. W. W. Adams, Chem. Rev., 1996,
96, 2817–2839.
39 R. H. Holm, Chem. Rev., 1987, 87, 1401–1449.
40 S. L. Brock and J. M. Mayer, Inorg. Chem., 1991, 30, 2138–2143.
41 R. H. Holm and J. P. Donahue, Polyhedron, 1993, 12, 571–589.
42 Z. Xiao, R. W. Gable, A. G. Wedd and C. G. Young, J. Am. Chem. Soc.,
1996, 118, 2912–2921.
43 C. Lorber, J. P. Donahue, C. A. Goddard, E. Nordlander and R. H.
Holm, J. Am. Chem. Soc., 1998, 120, 8102–8112.
44 J. P. Donahue, C. R. Goldsmith, U. Nadiminti and R. H. Holm, J. Am.
Chem. Soc., 1998, 120, 12869–12881.
12 W. Leitner, Acc. Chem. Res., 2002, 35, 746–756.
13 K. M. Nicholas, J. Organomet. Chem., 1980, 188, C10–C12.
14 B. Bogdanovic´, W. Leitner, C. Six, U. Wilczok and K. Wittmann,
Angew. Chem., Int. Ed. Engl., 1997, 36, 502–504.
15 D. S. Laitar, P. Mu¨ller and J. P. Sadighi, J. Am. Chem. Soc., 2005, 127,
17196–17197.
45 A. L. Odom, D. J. Mindiola and C. C. Cummins, Inorg. Chem., 1999,
38, 3290–3295.
46 G. Parkin and J. Bercaw, J. Am. Chem. Soc., 1989, 111, 391–393.
47 G. Parkin and J. Bercaw, Polyhedron, 1988, 7, 2053–2082.
48 W. A. Howard, M. Waters and G. Parkin, J. Am. Chem. Soc., 1993, 115,
4917–4918.
16 E. E. Benson, C. P. Kubiak, A. J. Sathrum and J. M. Smieja, Chem.
Soc. Rev., 2009, 38, 89–99.
49 W. A. Howard, T. M. Trnka, M. Waters and G. Parkin, J. Organomet.
17 M. Rakowski DuBois and D. L. DuBois, Acc. Chem. Res., 2009, 42,
1974–1982.
Chem., 1997, 528, 95–121.
50 (a) V. Pons and D. M. Heinekey, J. Am. Chem. Soc., 2003, 125, 8428–
8429; (b) R. Gelabert, M. Moreno, J. M. Lluch, A. Lledo´s, V. Pons and
D. M. Heinekey, J. Am. Chem. Soc., 2004, 126, 8813–8822.
51 J. P. Donahue, Chem. Rev., 2006, 106, 4747–4783.
52 W. A. Nugent and J. M. Mayer, Metal–Ligand Multiple Bonds, John
Wiley & Sons, New York, 1988.
18 J.-M. Save´ant, Chem. Rev., 2008, 108, 2348–2378.
19 D. L. DuBois Electrochemical Reactions of Carbon Dioxide, in Ency-
clopedia of Electrochemistry, Inorganic Electrochemistry, ed. A. J. Bard,
M. Stratmann, F. Scholz and C. J. Pickett, Wiley-VCH, Weinheim,
Germany, 2006, Vol 7A, pp 202–225.
20 (a) A. J. Morris, G. J. Meyer and E. Fujita, Acc. Chem. Res., 2009, 42,
1983–1994; (b) E. Fujita, Coord. Chem. Rev., 1999, 185–186, 373–384.
21 G. Fachinetti, C. Floriani, A. Chiesi-Villa and C. Guastini, J. Am.
Chem. Soc., 1979, 101, 1767–1775.
22 (a) J. C. Bryan, S. J. Geib, A. L. Rheingold and J. M. Mayer, J. Am.
Chem. Soc., 1987, 109, 2826–2828; (b) F.-M. Su, J. C. Bryan, S. Jang
and J. M. Mayer, Polyhedron, 1989, 8, 1261–1277; (c) K. A. Hall and
J. M. Mayer, J. Am. Chem. Soc., 1992, 114, 10402–10411.
23 I. Castro-Rodriguez and K. Meyer, J. Am. Chem. Soc., 2005, 127,
11242–11243.
53 R. Poli, Coord. Chem. Rev., 2008, 252, 1592–1612.
54 R. D. Rogers, E. Carmona, A. Galindo, J. L. Atwood and L. G. Canada,
J. Organomet. Chem., 1984, 277, 403–415.
55 M. G. B. Drew, E. M. Page and D. A. Rice, J. Chem. Soc., Dalton
Trans., 1983, 61–63.
56 R. H. Crabtree, The Organometallic Chemistry of the Transition Metals,
Wiley, New York, 1988, p. 72.
57 S. Dietz, V. Allured and M. Rakowski DuBois, Inorg. Chem., 1993, 32,
5418–5420.
58 A. G. Orpen and N. G. Connelly, J. Chem. Soc., Chem. Commun., 1985,
24 J. S. Silvia and C. C. Cummins, J. Am. Chem. Soc., 2010, 132, 2169–
1310–1311.
2171.
59 T. J. Crevier and J. M. Mayer, J. Am. Chem. Soc., 1997, 119, 8485–
8491.
60 S. W. Benson, J. Chem. Educ., 1965, 42, 502–518.
25 F. J. Arna´iz, J. Chem. Educ., 1995, 72, 1139.
26 W. L. F. Armarego and D. D. Perrin, Purification of Laboratory
Chemicals, 4th ed.; Butterworth-Heinemann, Oxford, U.K., 2000.
27 (a) P. R. Sharp, Organometallics, 1984, 3, 1217–1223; (b) P. R. Sharp,
J. C. Bryan and J. M. Mayer, Inorg. Synth., 1990, 28, 326–332.
28 A. V. Butcher, J. Chatt, G. J. Leigh and P. L. Richards, J. Chem. Soc.,
Dalton Trans., 1972, 1064–1069.
61 (a) H. Y. Afeefy, J. F. Liebman and S. E. Stein, Neutral Ther-
mochemical Data, in NIST Chemistry WebBook, NIST Standard
Reference Database Number 69, ed. P. J. Linstrom and W. G. Mallard,
National Institute of Standards and Technology: Gaithersburg, MD
20899, http://webbook.nist.gov (retrieved October 15, 2009).; (b) D. D.
Wagman, W. H. Evans, V. B. Parker, R. H. Schumm, I. Halow, S. M.
Bailey, K. L. Churney and R. L. Nuttall, The NBS Tables of Chemical
Thermodynamic Properties: Selected Values for Inorganic and C1 and
C2 Organic Substances in SI Units, ed. D. R. Lide, Jr., D. R., American
Chemical Society and the American Institute of Physics, Washington,
DC, J. Phys. Chem. Ref. Data, 1982, 11, (Suppl. 2).
29 J. R. Dilworth and R. L. Richards, Inorg. Synth., 1990, 28, 33–43.
30 E. Carmona, L. Sa´nchez, M. L. Poveda, R. A. Jones and J. G. Hefner,
Polyhedron, 1983, 2, 797–801.
31 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77,
3865–3868.
32 D. Andrae, U. Ha¨ußermann, M. Dolg, H. Stoll and H. Preuß, Theor.
Chim. Acta, 1990, 77, 123–141.
33 A. Scha¨fer, C. Huber and R. Ahlrichs, J. Chem. Phys., 1994, 100, 5829–
5835.
62 The value of -2.6 V (vs [Cp2Fe]+/Cp2Fe) for the direct, uncatalyzed
1-
reduction of CO2 to ·CO2 is in N,N-dimethylformamide (DMF) as
solvent. In DMF, the [Cp2Fe]+ + e-→ Cp2Fe couple occurs at +0.50 V
vs. the Ag/AgCl reference electrode used in this study, corresponding
to a slight cathodic shift of 0.04 V for the ferrocene reference in moving
from CH2Cl2 (+0.54 V) to DMF (+0.50 V). The chlorophosphine
complexes trans-[WCl4(PMe3-xPhx)2] (x = 0–3) and [WCl4(dppe)] were
not generally stable in DMF such that reduction potentials could be
recorded. In the absence of actual measurements in DMF, we assume
the W(III) + e- → W(II) couple is still -2.1 V vs. [Cp2Fe]+/[Cp2Fe] and
34 A. Scha¨fer, H. Horn and R. Ahlrichs, J. Chem. Phys., 1992, 97, 2571–
2577.
35 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, J. A. Montgomery, Jr., T. Vreven, K. N. Kudin, J. C.
Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci,
M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M.
Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.
Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E.
1-
thus +0.5 V less cathodic than the CO2 to ·CO2 couple.
This journal is
The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 9662–9671 | 9671
©