Direct measurements of rate constants and activation volumes for the binding of H2, D2, N2, C2H 4, and CH3CN to W(CO)3(PCy3) 2: Theoretical and experimental studies with time-resolved step-scan FTIR and UV-Vis spectroscopy

Article abstract of DOI:10.1021/ja064627e

Pulsed 355 nm laser excitation of toluene or hexane solutions containing W-L (W = mer,trans-W(CO)₃(PCy₃)₂; PCy ₃ = tricyclohexylphosphine; L = H₂, D₂, N ₂, C₂H₄, or CH₃CN) resulted in the photoejection of ligand L and the formation of W. A combination of nanosecond UV-vis flash photolysis and time-resolved step-scan FTIR (s²-FTIR) spectroscopy was used to spectroscopically characterize the photoproduct, W, and directly measure the rate constants for binding of the ligands L to W to reform W-L under pseudo-first-order conditions. From these data, equilibrium constants for the binding of L to W were estimated. The UV-vis flash photolysis experiments were also performed as a function of pressure in order to determine the activation volumes, ΔV?, for the reaction of W with L. Small activation volumes ranging from -7 to -3 cm³ mol^-1 were obtained, suggesting that despite the crowded W center an interchange mechanism between L and the agostic W...H-C interaction of one of the PCy₃ ligands (or a weak interaction with a solvent molecule) at the W center takes place in the transition state. Density functional theory (DFT) calculations were performed at the B3LYP level of theory on W with/without the agostic C-H interaction of the PCy₃ ligand and also on the series of model complexes, mer,trans-W(CO)₃(PH₃)₂L (W′-L, where L = H₂, N₂, C₂H₄, CO, or n-hexane) in an effort to confirm the infrared spectroscopic assignment of the W-L complexes, to simulate and assign the electronic transitions in the UV-vis spectra, to determine the nature of the HOMO and LUMO of W-L, and to understand the agostic C-H interaction of the ligand vs solvent interaction. Our DFT calculations indicate an entropy effect that favors agostic W...H-C interaction over a solvent σ C-H interaction by 8-10 kcal mol ^-1.

Full text of DOI:10.1021/ja064627e

Published on Web 11/15/2006
Direct Measurements of Rate Constants and Activation
Volumes for the Binding of H2, D2, N2, C2H4, and CH3CN to
W(CO)3(PCy3)2: Theoretical and Experimental Studies with
Time-Resolved Step-Scan FTIR and UV-Vis Spectroscopy
David C. Grills,† Rudi van Eldik,‡ James T. Muckerman,† and Etsuko Fujita*,†
Contribution from the Chemistry Department, BrookhaVen National Laboratory,
Upton, New York 11973-5000, and Institute for Inorganic Chemistry, UniVersity of
Erlangen-Nu¨rnberg, Egerlandstrasse 1, 91058 Erlangen, Germany
Received June 29, 2006; E-mail: fujita@bnl.gov
Abstract: Pulsed 355 nm laser excitation of toluene or hexane solutions containing W-L (W ) mer,trans-
W(CO)3(PCy3)2; PCy3 ) tricyclohexylphosphine; L ) H2, D2, N2, C2H4, or CH3CN) resulted in the
photoejection of ligand L and the formation of W. A combination of nanosecond UV-vis flash photolysis
and time-resolved step-scan FTIR (s2-FTIR) spectroscopy was used to spectroscopically characterize the
photoproduct, W, and directly measure the rate constants for binding of the ligands L to W to reform W-L
under pseudo-first-order conditions. From these data, equilibrium constants for the binding of L to W were
estimated. The UV-vis flash photolysis experiments were also performed as a function of pressure in
order to determine the activation volumes, ∆Vq, for the reaction of W with L. Small activation volumes
ranging from -7 to +3 cm3 mol-1 were obtained, suggesting that despite the crowded W center an
interchange mechanism between L and the agostic W‚‚‚H-C interaction of one of the PCy3 ligands (or a
weak interaction with a solvent molecule) at the W center takes place in the transition state. Density functional
theory (DFT) calculations were performed at the B3LYP level of theory on W with/without the agostic C-H
interaction of the PCy3 ligand and also on the series of model complexes, mer,trans-W(CO)3(PH3)2L
(W′-L, where L ) H2, N2, C2H4, CO, or n-hexane) in an effort to confirm the infrared spectroscopic
assignment of the W-L complexes, to simulate and assign the electronic transitions in the UV-vis spectra,
to determine the nature of the HOMO and LUMO of W-L, and to understand the agostic C-H interaction
of the ligand vs solvent interaction. Our DFT calculations indicate an entropy effect that favors agostic
W‚‚‚H-C interaction over a solvent σ C-H interaction by 8-10 kcal mol-1
.
Introduction
since it demonstrated an entirely new form of chemical bonding
involving the donation of a σ-bonding electron pair to a vacant
metal orbital. The family of M(CO)3(PR3)2(H2) (M ) Cr, Mo,
W) complexes is now often referred to as the “Kubas com-
plexes”.1 In M-(η2-H2) the metal center stabilizes H2 binding
by back-donation from a filled d orbital to the empty σ* orbital
of the H-H bond.1,6 Prior to this breakthrough, dihydrogen
complexes were assumed to be unobservable intermediates in
the oxidative addition of dihydrogen to form metal dihydrides.
However, by systematic, mainly electronic, variations of the
ligand set, a large series of dihydrogen complexes has since
been isolated, each one exhibiting a different degree of H-H
bond elongation and each representing a snapshot along the
H-H bond activation pathway, ultimately leading to classical
metal dihydride complexes.
The efficient activation and conversion of small, stable
molecules, such as H2, N2, CH4, C2H4, and CO2, into industrially
useful, functionalized compounds via metal-mediated catalytic
processes remains a key challenge for organometallic chemists
in the 21st century. The successful development of such
processes would provide major economic and environmental
benefits. One of the key steps in these molecular activation
processes involves the binding of the substrate molecule to a
vacant coordination site at a transition-metal center. The kinetics
and thermodynamics of this step can have a major effect on
the overall mechanism and efficiency of the catalytic process.
Therefore, it is vital that these binding steps are fully character-
ized and understood to allow the development of better and more
efficient catalysts.
In 1983, Kubas and co-workers isolated the first example of
a nonclassical molecular dihydrogen complex, mer,trans-
W(CO)3(PiPr3)2(η2-H2).1-5 This was a remarkable discovery
(2) Kubas, G. J.; Ryan, R. R.; Swanson, B. I.; Vergamini, P. J.; Wasserman,
H. J. J. Am. Chem. Soc. 1984, 106, 451-452.
(3) Kubas, G. J.; Unkefer, C. J.; Swanson, B. I.; Fukushima, E. J. Am. Chem.
Soc. 1986, 108, 7000-7009.
(4) Kubas, G. J.; Ryan, R. R.; Wrobleski, D. A. J. Am. Chem. Soc. 1986, 108,
1339-1341.
† Brookhaven National Laboratory.
‡ University of Erlangen-Nu¨rnberg.
(5) Wasserman, H. J.; Kubas, G. J.; Ryan, R. R. J. Am. Chem. Soc. 1986, 108,
2294-2301.
(1) Kubas, G. J. Metal Dihydrogen and σ-Bond Complexes: Structure, Theory,
and ReactiVity; Kluwer Academic/Plenum Publishers: New York, 2001.
(6) Kubas, G. J. Acc. Chem. Res. 1988, 21, 120-128.
9
15728
J. AM. CHEM. SOC. 2006, 128, 15728-15741
10.1021/ja064627e CCC: $33.50 © 2006 American Chemical Society

Rates of Small-Molecule Binding to W(CO)3(PCy3)2
A R T I C L E S
Thermodynamic and kinetic parameters for the addition
of small molecules such as H2, D2, and N2, to W (W )
mer,trans-W(CO)3(PCy3)2 and PCy3 ) tricyclohexylphosphine),
eq 1, have not been reported. However, Hoff’s group elegantly
employed stopped-flow kinetics7,8 to measure the indirect
substitution reaction of W-L (L ) H2 and N2) with pyridine
(py), evaluating kf(H ) and kf(N ) as (2.2 ( 0.3) × 106 and (5.0 (
assignments1,3,5,10 and our current IR observations, to simulate
and assign the electronic transitions in the UV-vis spectra, to
determine the nature of the HOMO and LUMO of W-L, and
to understand the agostic C-H interaction of the ligand vs
solvent σ C-H interaction.
Experimental Section
2
2
1.0) × 105 M-1 s-1, respectively, at 25 °C. Since reaction 1
with L ) py was too fast to follow under pseudo-first-order
conditions on a stopped-flow apparatus, they first accurately
determined the rate constant of reaction 1 with L ) P(OMe)3.
On the basis of the substitution reaction mechanism and its rate
Materials. W was prepared as previously described5 and character-
ized by NMR, UV-vis, and IR spectroscopy. No impurities were found
in the NMR and IR spectra except for a trace amount of W-CO, which
has previously been reported5 to be present in solutions of W. CH3CN
and pyridine were purified in the published manner,11 stored under
vacuum over activated molecular sieves, and then vacuum-distilled and
transferred in an Ar-filled glovebox just prior to sample preparation.
THF was purified in the published manner11 and stored under vacuum
over NaK. Toluene and hexane were distilled over sodium benzophe-
none ketyl and CaH2, respectively, under an Ar atmosphere and stored
in the glovebox. Ultrahigh purity H2, D2, N2, C2H4, CO, and Ar (Praxair)
were used without further purification. The solubilities of the various
gases in toluene and hexane were calculated from previously published
data.12 Accordingly, the following values at 25 °C and 1 atm partial
pressure of gas are used in this study: [H2] ) 2.97 and 4.80 mM in
toluene and hexane, respectively; [D2] ) 3.06 and 4.91 mM in toluene
and hexane,13 respectively; [N2] ) 5.37 and 10.7 mM in toluene and
hexane, respectively; [C2H4] ) 139 and 178 mM in toluene and hexane,
respectively.
law outlined in eqs 2 and 3, respectively, and k2(P(OMe) ) ) 5.45
3
× 104 M-1 s-1 in toluene at 25 °C, they calculated k2(py) ) 8.7
× 105 M-1 s-1 7,8
.
Substitution reaction 4 was used for the
evaluation of kf(H ) and kf(N )
.
2
2
kf(L)
W + L y z W-L
(1)
kr(L)
k2(P(OMe)
W-py yk z W + py
3)8 W-P(OMe)3 (2)
2(py)
k-2(py)
P(OMe)3
k2(py)k2(P(OMe) )[W - py][P(OMe)3]
d[W - py]
dt
3
)
(3)
(4)
k-2(py)[py] + k2(P(OMe) )[P(OMe)3]
3
Spectroscopic Measurements. UV-vis spectra were measured on
a Hewlett-Packard 8452A diode-array spectrophotometer. FTIR spectra
were recorded on a Bruker IFS 66/S spectrometer. NMR spectra were
measured on a Bruker UltraShield 400 MHz spectrometer. UV-vis
flash photolysis experiments were conducted using an apparatus
described previously,14,15 with excitation being provided by the third
harmonic (355 nm, ∼6 ns, ∼20 mJ/pulse) of a Continuum Surelite
I-10 Nd:YAG laser. These experiments were performed in toluene at
25 °C under 1 atm of Ar or under a known pressure of H2, D2, N2,
C2H4, or CO. The solutions containing 0.1-0.5 mM W-py or
0.5-1.0 mM W-NCCH3 were prepared using a known amount of py
or CH3CN, respectively, in a glovebox. The solutions containing
0.5-1.5 mM W-L (L ) H2, D2, N2, C2H4 or CO) were prepared by
adding a known pressure of L after having degassed the Ar-saturated
toluene solution of W. The kinetics of the signal-averaged data were
analyzed using Levenberg-Marquardt nonlinear least-squares routines
written in MATLAB. For the high-pressure work, the solution contain-
ing W-L (L ) H2, N2, or C2H4) was transferred via a syringe into a
vacuum-tight quartz pillbox optical cell,16-18 which had been flushed
with H2, N2, or C2H4 as in previous investigations.18,19 The solution
containing W-py or W-NCCH3 was transferred in a similar manner
under Ar. After removing the excess gas, the pillbox was immediately
placed inside the high-pressure cell, which was mounted on the laser
table.
W-L + py f W-py + L
Furthermore, they determined the enthalpies9 for reaction 1
in toluene by solution calorimetry (-9.9, -13.5, -15.1, -18.9,
-26.6, and -30.4 kcal mol-1 for L ) H2, N2, CH3CN, py,
P(OMe)3, and CO, respectively).
While the activation enthalpy (∆Hq) and entropy (∆Sq), and
the overall enthalpy (∆H°) and entropy (∆S°) changes for these
reactions using W and other complexes have been extensively
investigated using both calorimetric and stopped-flow tech-
niques, there is no information on the activation volume (∆Vq)
for the formation of dihydrogen, dinitrogen, or ethylene
complexes of W and other metal centers. Activation volumes
describe the volume changes that occur along the reaction
coordinate of a chemical process upon going to the transition
state, and can provide further information on the underlying
mechanism and the nature of the transition state, e.g. for small-
molecule addition to W with highly bulky PCy3 ligands. In the
present study we report direct measurements of rate constants
and activation volumes for the formation of W-L from W and
L. When a toluene or hexane solution containing W-L (L )
H2, D2, C2H4, N2, or CH3CN) and a known amount of L is
irradiated with a 355 nm laser pulse, W-L bond dissociation
takes place to form W. Subsequently, clean regeneration of
W-L is observed using time-resolved UV-vis and step-scan
FTIR (s2-FTIR) spectroscopy. In addition, we have performed
density functional theory (DFT) calculations at the B3LYP level
of theory on W with/without the agostic C-H interaction of
the PCy3 ligand and also on a series of model complexes,
mer,trans-W′-L (W′ ) W(CO)3(PH3)2; L ) H2, C2H4, N2, or
CO) in an effort to resolve discrepancies between earlier spectral
(10) Bender, B. R.; Kubas, G. J.; Jones, L. H.; Swanson, B. I.; Eckert, J.; Capps,
K. B.; Hoff, C. D. J. Am. Chem. Soc. 1997, 119, 9179-9190.
(11) Riddick, J. A.; Bunger, W. B.; Sakano, T. K. Organic SolVents: Physical
Properties and Methods of Purification, 4th ed.; Wiley: New York, 1986.
(12) Wilhelm, E.; Battino, R. Chem. ReV. 1973, 73, 1-9.
(13) Since we could not find any published data on the solubility of D2 in hexane,
this value has been estimated by scaling the solubility of H2 in hexane
according to the relative solubilities of D2 and H2 in heptane.
(14) Hamada, T.; Brunschwig, B. S.; Eifuku, E.; Fujita, E.; Korner, M.; Sasaki,
S.; van Eldik, R.; Wishart, J. F. J. Phys. Chem. A 1999, 103, 5645-5654.
(15) Thompson, D. W.; Wishart, J. F.; Brunschwig, B. S.; Sutin, N. J. Phys.
Chem. A 2001, 105, 8117-8122.
(16) le Noble, W. J.; Schlott, R. ReV. Sci. Instrum. 1976, 47, 770-771.
(17) van Eldik, R. High-Pressure Kinetics: Fundamental and Experimental
Aspects. In Inorganic High-Pressure Chemistry: Kinetics and Mechanisms;
van Eldik, R., Ed.; Elsevier: Amsterdam, 1986; Vol. 7, pp 8-11.
(18) Fujita, E.; van Eldik, R. Inorg. Chem. 1998, 37, 360-362.
(19) Fujita, E.; Brunschwig, B. S.; Creutz, C.; Muckerman, J. T.; Sutin, N.;
Szalda, D.; van Eldik, R. Inorg. Chem. 2006, 45, 1595-1603.
(7) Zhang, K.; Gonzalez, A. A.; Hoff, C. D. J. Am. Chem. Soc. 1989, 111,
3627-3632.
(8) Gonzalez, A. A.; Zhang, K.; Hoff, C. D. Inorg. Chem. 1989, 28, 4285-
4290.
(9) Gonzalez, A. A.; Zhang, K.; Nolan, S. P.; Delavega, R. L.; Mukerjee, S.
L.; Hoff, C. D.; Kubas, G. J. Organometallics 1988, 7, 2429-2435.
9
J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006 15729

A R T I C L E S
Grills et al.
Table 1. UV-Vis and FTIR Spectroscopic Properties Observed
Time-resolved step-scan FTIR experiments were performed by
operating the Bruker IFS 66/S FTIR spectrometer in step-scan mode
and using the third harmonic of the Nd:YAG laser to excite the sample
at a repetition rate of 5 Hz (Laser master/FTIR slave). A 1 mm
photovoltaic HgCdTe infrared detector equipped with a 20 MHz
preamplifier (Kolmar, KMPV11-1-LJ2/239) was used in these experi-
ments. The detector has both AC- and DC-coupled outputs, which were
amplified and/or attenuated by external preamplifiers and attenuators
(Phillips Scientific, model nos. 6931 and 5010 and Stanford Research
Systems, SR560) and digitized by an external 100 MHz 8-bit digitizer
(Spectrum, PAD82a). Before commencing the laser excitation measure-
ment, the DC output was collected in step-scan mode with Mertz phase
correction, to generate a set of single beam background spectra and
also to store a set of phase spectra in memory, which were later used
to automatically phase correct the subsequent time-resolved AC spectra.
During the laser measurement, the AC output was monitored and
amplified by ×10-20, to use the full dynamic range of the digitizer.
Optical band-pass filters (OCLI, W05342-8 ca. 2000-1765 cm-1 band-
pass and W04944-8 ca. 2260-1830 cm-1 band-pass) were used in order
to minimize the number of interferogram mirror positions. Single-sided
interferograms, with 4 or 8 laser shots at each mirror position, were
collected at a spectral resolution of 4 cm-1, and a phase resolution of
32 cm-1. Delta absorbance TRIR spectra, ∆(OD)t, at each of the 100
time delays, t following the laser flash, were generated using the
equation, ∆(OD)t ) -log10(1 + γ(AC)t/DC), where γ is the ratio of
the amplifications applied to the DC and AC signals, (AC)t is the single-
beam spectrum generated from the detector’s AC output at time t after
the laser flash and DC is the single-beam spectrum that was gen-
erated from the DC output before the laser measurement. Due to in-
tense absorptions of toluene in the metal carbonyl infrared region, the
s2-FTIR experiments were performed in 25-50 mL of hexane, at a
concentration of ca. 0.5 mM. The dry, degassed, violet-colored hexane
solutions of W were saturated with a known pressure of the appropriate
gas, L, in order to generate W-L, which typically had a pale-yellow
color. The solutions were handled in an air-sensitive manner and flowed
cyclically through an airtight, home-built flow system consisting of
1/8 in. o.d. stainless steel tubing and Swagelok fittings, a glass sample
reservoir vessel fitted with glass-to-metal 1/4 in. Kovar joints, a 1 mm
path length CaF2 IR flow cell (Harrick Scientific Products, DLC-S25),
and a recirculating gear pump (Micropump, GA-V23JFSG with
HG-0024 drive).
for W-L Complexes in Toluene and Hexane, Respectively, at 25
°C
L
properties
UV-visa
none
noneb
H2
312 (9700), 350 sh (830), 452 sh (900), 566 (1100)
452 sh, 550
366 (1300), 390 sh (1100)
N2
350 sh (2400), 370 (2500), 408 (3300)
350 sh (3200), 400 sh (1300)
340 sh (3200), 430 sh (1300), 485 sh (540)
346 (2500), 382 (3000)
C2H4
CH3CN
CO
py
560 (6600)
IRc
none
H2
D2
1819m, 1837s, 1949w
1847s, 1851s, 1861s, 1912w, 1963w, 1988w
1847s, 1851s, 1860s, 1908w, 1965w, 1988w
N2
C2H4
CO
1853s, 1857sh, 1953w, 2113w (νNN),2117w (νNN
)
1841s, 1865m, 1962w
1866s, 1871vw, sh
a λmax, nm (ꢀ, M-1 cm-1) in toluene. b In THF. c ν(CO), cm-1 in hexane,
s ) strong, m ) medium, w ) weak, sh ) shoulder, v ) very.
of the various species. Calculated geometries are shown in Table 3
and Figures 4 and 5, and binding energies are summarized in Table 4.
Results
UV-Vis Spectra of W and W-L Species. When an
appropriate gas L (L ) H2, N2, C2H4, CO) was added at a partial
pressure of 1 atm to a degassed violet solution containing W at
room temperature, the 566 nm absorption of W completely
disappeared due to the high stability of the W-L species being
formed (eq 1). The final solution is yellow as reported
previously.5 UV-vis spectra of W and W-L are summarized
in Table 1 and shown in Figure 1. Upon removal of the gases,
such as H2, N2 and C2H4, by freeze-pump-thaw cycles from
the solutions containing the corresponding W-L species in
toluene or THF, the original W spectra were cleanly reproduced.
However, CO binding in W-CO was found not to be reversible.
In the case of W-py (py ) pyridine), upon addition of py, the
566 nm absorption is replaced by a much more intense
absorption band at 560 nm, and the solution turned bright red.
Flash Photolysis. Flash photolysis experiments on W-L in
toluene were carried out using 355 nm pulsed laser excitation.
Figure 2 shows the transient absorption spectrum of a short-
lived species observed 2 µs after excitation of a toluene solution
containing 1.40 mM W-C2H4 at 25 °C at an ethylene partial
pressure of 740 Torr. By comparison with the UV-vis spectrum
of W in Figure 1, this transient species can easily be assigned
to W, which is then observed to decay as W-L is reformed
under pseudo-first-order conditions. Similar transient spectra
were also observed for L ) H2, D2, N2, or CH3CN. The observed
pseudo-first-order rate constant for the reformation process is
given by eq 5.
DFT Calculations. Gas-phase DFT calculations on the series of
model complexes, mer,trans-W′-L (L ) none, H2, C2H4, N2, CO, or
n-hexane), and on W and W(CO)3(PH3)(PH2Cy) with and without an
agostic interaction, were carried out using the hybrid B3LYP method20-22
and the Hay-Wadt VDZ (n+1) ECP (LANL2DZ ECP) basis set for
W23-25 and the 6-31G** (5d) basis set for H, C, N, O, and P.26-28 Note
that the computed “net agostic interaction energy” discussed in this
paper is the sum of the intrinsic agostic interaction energy and any
steric and/or electronic energy differences associated with the two
different conformers. Although this is not the intrinsic “agostic”
interaction, it is nevertheless a relevant number. Geometry optimizations
and frequency analyses were performed using the Gaussian 03 package
of programs.29 Time-dependent B3LYP calculations were carried out
with the Gaussian 03 program in order to predict the UV-vis spectra
(20) Becke, A. D. Phys. ReV. A 1988, 38, 3098-3100.
(21) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B. 1988, 37, 785-789.
(22) Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157,
200-206.
kobs ) kf[L] + kr
(5)
The forward and reverse rate constants (kf and kr in eqs 1
and 5) in toluene were determined at 25 °C. Figure 3 shows a
plot of the observed rate constants for formation of W-L (L )
H2 and D2) as a function of H2 and D2 concentration in toluene.
The slopes and intercepts of these plots yield kf ) (2.0 ( 0.1)
× 106 M-1 s-1 and kr ) 950 ( 100 s-1 for L ) H2, and kf )
(1.6 ( 0.1) × 106 M-1 s-1 and kr ) 680 ( 70 s-1 for L ) D2.
(23) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270-283.
(24) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284-298.
(25) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299-310.
(26) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971, 54, 724-
728.
(27) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56, 2257-
2261.
(28) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213-222.
(29) Pople, J. A.; et al. Gaussian 03, revision B.04; Gaussian, Inc.: Wallingford
CT, 2004.
9
15730 J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006

Rates of Small-Molecule Binding to W(CO)3(PCy3)2
A R T I C L E S
Figure 3. Plot of observed rate constants for formation of W-H2 (red)
and W-D2 (blue) as a function of gas concentration in toluene at 25 °C.
[H2] ) 2.97 mM/atm, [D2] ) 3.06 mM/atm at 25 °C.
kobs is small, it can be assumed that kf[L] ≈ kobs under the
experimental conditions used. The effect of pressure on the
relaxation process was found to be reproducible for a complete
pressure cycle, i.e. increasing pressure from ambient to 2000
atm and decreasing the pressure back to ambient conditions. A
relatively small increase or decrease in the rate was observed
in all cases (Table S1, Supporting Information). The activa-
tion volume, ∆Vq(kf), was estimated from the initial slope ()
-∆Vq/RT) of the plot of ln(kobs) vs pressure, and the results are
included in Table 2. The activation volumes range from -7
cm3 mol-1 for C2H4 to +3 cm3 mol-1 for CH3CN.
DFT Calculations: Geometric and Electronic Structures
and Vibrational Analyses. The optimized Cartesian coordinates
for mer,trans-W′-L (L ) none, H2, C2H4, N2, CO, or n-hexane),
W with and without an agostic interaction, and W(CO)3-
(PH3)(PH2Cy) with and without an agostic interaction are
presented in Tables S2-S13 (Supporting Information). The
calculations on W′-L were performed mainly in an effort to
confirm our assignment of three ν(CO) bands in the IR spectra
of the W-L complexes under study. This is due to a discrepancy
that exists over the assignment of these ν(CO) bands in the
literature1,3,5,10 vs the spectra measured in this work. For
example, while our IR experiments for W-C2H4 clearly show
three ν(CO) bands, Wasserman et al. only reported two of these.5
Bender et al.10 also missed one of the three ν(CO) bands30,31 in
their normal-coordinate analysis of W-H2, probably based on
previously reported experimental values (1963 and 1843 cm-1).3,5
In accordance with previous theoretical and experimental
studies,32 our DFT calculations show that the ligand, L in mer,-
trans-W′-L, and hence presumably in W-L, is oriented parallel
to the P-W-P axis for L ) H2 and C2H4 (see Figure 4, b and
f). For L ) N2, the dinitrogen molecule binds end-on via a single
N atom, which is a common coordination mode for N2 (see
Figure 4e).33-35 (However, N2 can act as a bridging ligand to
form dimeric species such as (PiPr3)(CO)3W-N2-W(CO)3-
((PiPr3), the structure of which has been solved.47) In further
Figure 1. UV-vis spectra of various W-L complexes in toluene: (top)
W, black; W-H2, red; W-C2H4, blue; and W-NCCH3, green; (bottom)
W, black; W-N2, red; W-CO, blue; and W-py, green; * ) minor
contamination by W-CO.
Figure 2. Transient absorption spectrum obtained 2 µs after 355 nm pulsed
laser excitation of a solution containing 1.4 mM W-C2H4 and 740 Torr
C2H4 in toluene at 25 °C.
The equilibrium constants, K calculated from kf/kr for W-H2
and W-D2 are thus 2100 and 2300 M-1, respectively. Figures
S1-S3 (Supporting Information) show plots of the observed
rate constants for the formation of W-L (L ) N2, C2H4,
CH3CN) as a function of the concentration of L in toluene. The
forward and reverse rate constants and the equilibrium constants
for W-L, calculated in a similar manner, are summarized in
Table 2.
(30) Tomas, J.; Lledos, A.; Jean, Y. Organometallics 1998, 17, 190-195.
(31) Tomas, J.; Lledos, A.; Jean, Y. Organometallics 1998, 17, 4932-4939.
(32) Kubas, G. J. J. Organomet. Chem. 2001, 635, 37-68.
Activation Volumes. The forward rate constants for the
formation of W-L (eq 1) in toluene were determined as a
function of pressure at 25 °C. Since the contribution of kr to
(33) Hidai, M.; Mizobe, Y. Chem. ReV. 1995, 95, 1115-1133.
(34) Leigh, G. J. Acc. Chem. Res. 1992, 25, 177-181.
(35) Yandulov, D. V.; Schrock, R. R. Science 2003, 301, 76-78.
9
J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006 15731

A R T I C L E S
Grills et al.
Table 2. Reaction Rate Constants, Equilibrium Constants (calculated from kf/kr) and Activation Volumes for the Formation of W-L with
Various Ligands, L, Using UV-Vis Flash Photolysis at 25 °C or Step-Scan FTIR Spectroscopy at 23 ( 2 °C
1
1
1
1
ligand
solvent
method
kf, M-1 s-
kr, s-
calcd K, M-
∆
Vq, cm3 mol-
H2
H2
D2
D2
C2H4
C2H4
N2
N2
CH3CN
toluene
hexane
toluene
hexane
toluene
hexane
toluene
hexane
toluene
UV-vis
s2-FTIR
UV-vis
s2-FTIR
UV-vis
s2-FTIR
UV-vis
s2-FTIR
UV-vis
(2.0 ( 0.1) × 106
950 ( 100
680 ( 70
570 ( 70
99 ( 10
<200
2100 ( 300
2300 ( 300
60 ( 10
-3 ( 1
a3 × 106
(1.6 ( 0.1) × 106
a2 × 106
(3.4 ( 0.3) × 104
(7 ( 1) × 104
(3.0 ( 0.2) × 105
a3 × 105
-7 ( 1
-3 ( 1
+3 ( 1
3000 ( 350
>2 × 104
(3.6 ( 0.3) × 106
a These rate constants were estimated from a single experiment at one concentration of ligand, L using the equation, kf ) (kobs - kr)/[L], where kobs is the
observed rate of decay of the photoproduced W and kr is the reverse rate constant that we obtained by UV-vis flash photolysis in toluene.
(H)2, and W′-cis-(H)2) are in agreement with results published
previously.30,31 The free energies (∆G°298) for the formation of
W′-H2, W′-distal-(H)2, and W′-cis-(H)2 are -6.5, -1.2, and
-0.1 kcal mol-1, respectively (See Table 4). The corresponding
three sets of ν(CO) bands were calculated and are shown in
Figure 8b.
Calculations on W, W′, W′-n-hexane, W(CO)3(PH3)-
(PH2Cy), and W(CO)3(PCy3)(PH3) were carried out to inves-
tigate the strength of an agostic interaction of a C-H of a PCy3
ligand vs the hydrocarbon solvent interaction. Figures 4a and
5a-e show optimized geometries of these complexes. In the
five-coordinate square pyramidal, W′, the basic structural
features of W′-L remain the same except that a large unoc-
2
cupied dz orbital exists on the opposite side of the unique CO
with a shorter W-CO distance (1.934 Å) than those calculated
for other W′-L complexes (av 2.020 Å). We calculated the
geometric and electronic structure of W′-n-hexane to compare
with W′: the shortest W-H and W-C distances are 2.208 and
3.025 Å, respectively and the W′-n-hexane electronic (∆E)
and free (∆G°298) energies of formation are -5.4 and +5.3 kcal
mol-1, respectively. Furthermore, geometry optimizations were
performed for both the agostic and nonagostic W(CO)3(PH3)-
(PH2Cy) complexes (where nonagostic refers to a local minimum-
energy structure in a conformation that does not allow an agostic
W‚‚‚H-C interaction from the Cy group): the shortest W-H
and W-C distances are 2.217 and 3.242 Å, respectively, in the
agostic complex and 3.472 and 4.065 Å, respectively, in the
nonagostic complex. The calculated gas-phase net agostic
interaction (∆E) and free energy (∆G°298) are -0.5 and +1.3
kcal mol-1. The difference is small due to a small entropic
contribution associated with the intramolecular interaction.
Geometry optimizations were also performed for both the agostic
and nonagostic W complexes: the shortest W-H and W-C
distances are 2.488 and 3.206 Å, respectively, in the agostic
complex and 3.196 and 3.812 Å, respectively, in the nonagostic
complex. The calculated gas-phase net agostic interaction (∆E)
and free energy (∆G°298) are -4.7 and -5.2 kcal mol-1. The
difference is again small due to a small entropic contribution
associated with the intramolecular interaction. We should also
point out that while an agostic R-C-H interaction was observed
in both the X-ray structure and the DFT calculations for W
and W(CO)3(PCy3)(PH3), respectively, an agostic â-C-H
interaction was observed in the DFT calculations for W(CO)3-
(PH3)(PH2Cy) (Figure 5).
Figure 4. DFT optimized structures of (a)W′, (b) W′-H2 (dihydrogen),
(c) W′-(H)2 (distal dihydride complex), (d) W′-(H)2 (cis-dihydride
complex), (e) W′-N2 and (f) W′-C2H4.
agreement with previous theoretical studies,30,31 our DFT
calculations establish that the H ligands in the lowest-energy
configuration of the dihydride form (W′-(H)2) are arranged
in a distal fashion, either side of one of the PH3 ligands (see
Figure 4c), as opposed to remaining in a cis orientation (Fig-
ure 4d), which might otherwise have been anticipated fol-
lowing a simple elongation of the H-H bond during the
oxidative addition process. While we used a slightly different
basis set for the geometry optimizations, our results on the
electronic energies without zero-point energy corrections (∆E)
and geometries of the three species (W′-H2, W′-distal-
The DFT calculations also allow the nature of the HOMO
and LUMO of W-L to be determined. Our DFT calculations
9
15732 J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006

Rates of Small-Molecule Binding to W(CO)3(PCy3)2
A R T I C L E S
Figure 5. DFT optimized structures of (a) W′-n-hexane, (b) W(CO)3(PH3)(PH2Cy) without a C-H agostic interaction of Cy, (c) W(CO)3(PH3)(PH2Cy)
with a C-H agostic interaction of Cy, (d) W without a C-H agostic interaction of PCy3, and (e) W with a C-H agostic interaction of PCy3.
on W′-N2 have been published previously.36 DFT calculations
on W′-L (L ) H2, C2H4 and CO) show that the HOMO is
mainly metal-based and that the LUMO is largely located on
the three CO ligands and L, with some contribution of P* as
shown in Table 3, Figure 6 and Figure S5. This general result
suggests that the lowest-energy transition observed in the UV-
vis spectra of W-L (see Figure 1) is metal-to-ligand charge
transfer (MLCT) in nature. In the case of the five-coordinate
W′, while the HOMO is dπ mixed with the unique CO*, the
2
LUMO is dσ (dz ) mixed with the trans CO* (Figure 6).
(36) Grills, D. C.; Huang, K.-W.; Muckerman, J. T.; Fujita, E. Coord. Chem.
ReV. 2006, 250, 1681-1695.
9
J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006 15733

A R T I C L E S
Grills et al.
Table 3. DFT Calculated Geometrical Parameters for mer,trans-W(CO)3(PH3)2L (W′-L) Model Complexes
properties
W
′
W
′−H2
W
′−distal-(H)2
W
′−(C2H4)
W
′−N2
W′−CO
HOMOa
LUMOa
dπ + (CO*) dπ + (CO*)
dπ + (CO*)
dπ + (CO*)
dπ + (CO*)
dπ + (CO*)
dz + (CO*) CO* + (P* + H2*) CO* + H-* + P* C2H4* + CO* + (P*) N2* + CO* + P* CO* + (P*)
2
W-L (H2, H-, C2H4, N2) distance, Å
H-H distance, Å
C-C (C2H4) distance, Å
N-N distance, Å
-
-
-
-
1.908
0.837
-
1.752
3.355
-
2.396
-
1.403
-
2.110
-
-
-
-
-
-
-
-
1.121
W-P distance, Å
2.457
1.934
2.039
1.172
1.161
175.14
92.43
90.02
2.466
1.995
2.040
1.163
1.160
178.01
89.00
90.03
2.469
2.042
2.035
1.155
1.160
139.55
134.69
88.18
2.465
2.011
2.037
1.162
1.161
172.20
86.10
90.15
2.468
2.005
2.046
1.164
1.159
179.19
90.40
90.01
2.473
2.047
-
W-CO (1) distance, Å
W-CO (2) distance, Å
C-O (1) distance, Å
C-O (2) distance, Å
P-W-P angle, deg
P-W-C (1) angle, deg
P-W-C (2) angle, deg
1.159
-
180.00
90.00
-
bν(CO), cm-1
1898(38)
1919(91)
1999(1)
1927(25)
1941(9)
2009(1)
2846
1933(7)
1973(2)
2029(1)
1917(14)
1937(6)
2002(1)
1934(9)
1938(5)
2006(1)
1940(4359)c
1974(6)
2035(1)
bν(HH), cm-1
bν(NN), cm-1
2184
a The orbitals in parentheses are minor contributions. b Frequencies scaled by 0.9613 (relative intensities for ν(CO)). c This band comprises two overlapping
bands with almost identical frequencies and intensities.
Figure 6. HOMO and LUMO for W′ (a and e), W′-dihydrogen (b and f), W′-distal dihydride (c and g), and W′-CO (d and h), respectively.
Time-dependent DFT (TDDFT) calculations were carried out
for the W′-H2 and W′-(H)2 (distal- and cis-dihydride)
complexes because both the η2-dihydrogen and the dihydride
forms of W-H2 exist in equilibrium in solution.7,10 The
UV-vis spectrum of W-H2 in toluene shows an absorption
band at 366 nm with a shoulder at 390 nm. TDDFT calculated
spectra for W′-H2 (blue curve), W′-distal-(H)2 (red curve)
and W′-cis-(H)2 (green curve), shown in Figure 7a, indicate
that λmax of the lowest-absorption band of the W′-(H)2
complexes (∼290 nm) is considerably blue-shifted compared
to that of W′-H2 (∼335 nm). The calculated absorptions for
W′-H2 are at higher energies than the experimental data,
probably due to the electronic and steric effects caused by
replacing PCy3 with PH3. However, the shape of the spectrum
matches well with the observed spectrum. The lowest-energy
transition is the HOMO-1 to LUMO excitation mixed with
substantial HOMO-1 to LUMO+2 and HOMO-2 to
LUMO+3 contributions at 340 nm, with an oscillator strength
of 0.0059. The next lowest transition is the HOMO-1 to
LUMO+1 excitation at 335 nm, with an oscillator strength of
0.0109. The third lowest transition is the HOMO to LUMO+2
excitation mixed with substantial HOMO-2 to LUMO+1 and
HOMO to LUMO contributions at 301 nm, with an oscillator
strength of 0.0141. Our theoretical investigation on W′-N2 has
been published previously.36 In W′-C2H4, the lowest-energy
transition is the HOMO to LUMO excitation mixed with a
substantial HOMO-1 to LUMO+1 contribution at 373 nm
with an oscillator strength of 0.0006 (See Figure S5 for the
Frontier orbitals). The next lowest transition is the HOMO-2
to LUMO excitation mixed with substantial HOMO-2 to
LUMO+2 and HOMO-1 to LUMO+3 contributions at 333
nm with an oscillator strength of 0.0105. The third lowest
transition is the HOMO to LUMO+2 excitation mixed with
substantial HOMO-1 to LUMO+1 and HOMO to LUMO
contributions at 327 nm with an oscillator strength of 0.0039.
The calculated absorptions are again at slightly higher energies,
however, the shape of the spectrum matches well with the
observed spectrum.
9
15734 J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006

Rates of Small-Molecule Binding to W(CO)3(PCy3)2
A R T I C L E S
Figure 7. Calculated UV-vis spectra with oscillator strengths: (a) W′-dihydrogen (blue), W′-distal-dihydride complex (red) and W′-cis-dihydride
complex (green); (b) W′ (red) and W′-n-hexane (blue). The line spectra were converted with a Gaussian broadening function with a width parameter of
0.18 eV to obtain the continuous spectra shown.
We also conducted TDDFT calculations on the W′ and
W′-n-hexane complexes in the gas phase to understand the
nature of the transition and the spectral change due to a
weak solvent (such as a hexane) interaction. The results are
shown in Figure 7b. The lowest transition in W′ (red curve in
Figure 7b) is the HOMO to LUMO excitation at 709 nm
with an oscillator strength of 0.0198. The next lowest transi-
tion is the HOMO-2 to LUMO excitation mixed with a
substantial HOMO-1 to LUMO+1 contribution at 551 nm
with an oscillator strength of 0.0195. The third lowest transition
is the HOMO-1 to LUMO+1 excitation mixed with a
substantial HOMO-2 to LUMO+2 contribution at 309 nm with
an oscillator strength of 0.0476. The calculated absorptions are
in this case at slightly lower energies; however, the shape of
the spectrum matches well with the observed spectrum in
toluene. This could be due to this being a gas-phase calculation
without any solvent interaction at the sixth coordination site.
To prove this, we calculated the spectrum of W′-n-hexane
(Figure 7b, blue curve). The lowest transition is the HOMO to
LUMO transition at 497 nm with an oscillator strength of
0.0198. The second lowest transition is the HOMO-2 to LUMO
excitation mixed with the HOMO-1 to LUMO+1 excita-
tion at 424 nm with an oscillator strength of 0.0112. The
calculated spectrum is shifted to a significantly higher energy
than that of the experimental spectrum in either toluene or
hexane. The experimental spectra of W in toluene and n-hexane
are identical with λmax at 566 nm. However, in THF λmax is at
550 nm. The high-energy shift in the calculated spectra is
9
J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006 15735

A R T I C L E S
Grills et al.
Figure 8. (a) Experimental IR spectra for five W-L complexes in hexane solution and (b) gas-phase DFT calculated IR spectra for five W′-L complexes.
The calculated line spectra were converted with a Gaussian broadening function with a width parameter of 3.6 cm-1. * ) minor contamination by W-CO.
observed for all W′-L (L ) H2, N2,36 C2H4, n-hexane)
complexes and could be due to the different electronic effect
of PH3 relative to that of PCy3, combined with the fact that
these are gas-phase calculations.
features immediately following photoexcitation of the W-L
complex. In some cases the third, much weaker band of W was
also observed at 1949 cm-1 if a negative bleach band of W-L
was not obscuring it. The ν(CO) bands of W were then observed
to decay exponentially via pseudo-first-order kinetics at the same
rate as the recovery of the negative bleach bands of W-L during
the reaction of W with excess L present in the solution. Some
extremely minor, static features were observed in the baselines
of the s2-FTIR spectra, between ca. 1940 and 1900 cm-1. These
features are assigned to long-lived products associated with the
photodecomposition of trans-W(CO)4(PCy3)2, which is a minor,
unavoidable contaminant in the samples.
Figure 8a shows the experimentally observed FTIR spectra
of W and W-L (L ) N2, H2, C2H4, or CO) in hexane solution.
The DFT calculated gas-phase IR spectra of the W′-L model
complexes are shown in Figure 8b and summarized in Table 3.
Since DFT with the B3LYP functional typically overestimates
vibrational frequencies, the frequencies of the calculated spectra
have been scaled by a factor of 0.9613, in accordance with an
earlier recommendation37 for the B3LYP functional when used
with the 6-31G* basis set. As expected, the absolute frequency
scale is still predicted too high since the calculations were
performed on PH3 complexes in the gas phase, as opposed to
on PCy3 complexes in hexane solution. However, the relative
frequencies and intensities (see Table 3 and Figure 8) of the
bands are in excellent agreement with our experimental results,
thus confirming the presence of three ν(CO) bands in the IR
spectra of W-L. For the H2 complex, three additional ν(CO)
bands are observed and calculated, due to the presence of the
cis- and distal-dihydride complexes, which are in equilibrium
with the η2-dihydrogen σ-bonded species. It is clear from Figure
8 that trans-W(CO)4(PCy3)2 exhibits a single, intense ν(CO)
band that almost coincides with the middle of the three ν(CO)
bands of the W-L complexes.
Using the reverse rate constants, kr that were measured in
the UV-vis flash photolysis experiments in toluene (eq 1, Table
2), the forward rate constants, kf for the s2-FTIR experiments
in hexane under a single partial pressure of L were estimated
from kf ) (kobs - kr)/[L], where kobs is the observed rate of
decay of the photogenerated W or observed rate of recovery of
W-L, and [L] is the concentration of L. Despite the fact that
it was necessary to perform the s2-FTIR experiments in hexane
rather than in toluene (due to intense IR absorption bands of
toluene in the ν(CO) region), the rate constants estimated by
s2-FTIR are in reasonable agreement with those obtained by
UV-vis flash photolysis in toluene (see Table 2). Band positions
and rate constants for all of the complexes studied by s2-FTIR
are reported in Tables 1 and 2.
Step-Scan FTIR Spectroscopy. In all of the s2-FTIR
experiments, at least two of the ν(CO) bands of W (at 1837
and 1819 cm-1) were observed as newly formed positive
The FTIR spectrum of W-C2H4 in hexane under 1.8 atm
partial pressure of C2H4 exhibits three ν(CO) bands at 1962,
1865, and 1841 cm-1 (Figure 9a). The s2-FTIR spectra obtained
at 2 µs intervals following 355 nm excitation of this solution
(37) Wong, M. W. Chem. Phys. Lett. 1996, 256, 391-399.
9
15736 J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006

Rates of Small-Molecule Binding to W(CO)3(PCy3)2
A R T I C L E S
Figure 10. (a) FTIR of W-H2 in hexane under 1.8 atm partial pressure
of H2, showing six overlapped ν(CO) bands due to the equilibrium mixture
of W-(H2), W-distal-(H)2 and W-cis-(H)2; * ) minor contamination by
W-CO. (b) Step-scan FTIR spectra recorded between 2 and 192 µs
following 355 nm excitation of this solution. Arrows show directions of
band decay and recovery following excitation.
Figure 9. (a) FTIR of W-C2H4 in hexane under 1.8 atm partial pressure
of C2H4, (using hexane saturated with 1.8 atm of C2H4 as a background
spectrum). (b) Step-scan FTIR spectra recorded between 2 and 192 µs
following 355 nm excitation of this solution. Arrows show directions of
band decay and recovery following excitation.
are shown in Figure 9b. Three ν(CO) bands of W-C2H4 are
observed to deplete, together with the simultaneous formation
of the three characteristic bands of W at 1949, 1837, and 1819
cm-1. A minor, static baseline artifact, which does not change
in intensity throughout the duration of the experiment, is also
observed at 1902 cm-1. The transient ν(CO) bands of W are
observed to decay exponentially at a rate (kobs ) (2.3 ( 0.2) ×
104 s-1) identical to the rate of recovery (kobs ) (2.2 ( 0.2) ×
104 s-1) of W-C2H4. The value of kf(C H ) estimated (as
observed in Nujol mulls for complexes containing perdeuterio-
phosphine ligands, due to interference by the ν(CH) modes of
the PCy3 ligands and common solvents.3,10 The five remaining
W-H2 vibrations are typically weak or medium in intensity
and occur at frequencies significantly lower than those of the
ν(CO) bands. As mentioned above, the dihydride species exists
in equilibrium. However, W-hydride vibrations are very weak
and lower in frequency than ν(CO). For these reasons, only the
ν(CO) vibrations were probed in the present s2-FTIR experi-
ments. By comparison with our DFT calculations and previously
published data,3 the bands observed at 1963, 1861, and 1847
cm-1 are assigned to the η2-dihydrogen complex and those at
1988, 1912, and 1851 cm-1 to the distal-dihydride species that
exists in equilibrium with the dihydrogen complex in solution.
According to our DFT calculations, the bands for the cis-
dihydride species, which is the least stable of the hydrides, will
overlap with the bands at 1988, 1912, and 1861 cm-1. The
slightly unsymmetrical shape of the band at 1912 cm-1 and the
calculated bond energy suggest the existence of the cis-dihydride
species, however, we expect that this is a minor species.
Upon 355 nm excitation of a hexane solution of W-H2
saturated under 1.8 atm partial pressure of H2, the ν(CO) bands
at 1963 and 1861-1847 cm-1 are observed to bleach together
with the concomitant formation of the three characteristic bands
of W (see Figure 10). However, the dihydride bands at 1988
and 1912 cm-1 are not observed to bleach throughout the
experiment, which suggests that 355 nm excitation selectively
dissociates dihydrogen from the η2-dihydrogen complex, leaving
the dihydride complexes intact. Since only the η2-dihydrogen
complex is bleached upon 355 nm excitation, the dihydrogen/
2
4
described above) from this experiment was ca. 2× higher than
that obtained by UV-vis flash photolysis of W-C2H4 in
toluene. Therefore, the s2-FTIR experiment was repeated at five
different concentrations of C2H4 in order to more accurately
determine kf(C H ) for the reaction of W with C2H4 to reform
2
4
W-C2H4 in hexane. The observed rate of decay, kobs, of the
photogenerated W complex varies linearly with respect to
[C2H4] (see Figure S4) and from the slope of this graph, kf(C H )
2
4
was found to be (7 ( 1) × 104 M-1 s-1
.
For the dihydrogen adduct of W, the experiment is compli-
cated by the presence of both the η2-dihydrogen and the two
dihydride forms of W-H2, all of which exist in equilibrium
with one another in solution. It has previously been estimated7,10
that ca. 10-30% of W-H2 in solution is present in the dihydride
forms. This equilibrium gives rise to the six ν(CO) bands that
are observed in the IR spectrum of W-H2 in hexane and which
were described earlier. In addition to the ν(CO) modes, the η2-
dihydrogen complex should exhibit six additional IR active
modes involving the H2 ligand, one of which is a ν(HH)
stretching vibration.10 The ν(HH) mode, which occurs at 2690
cm-1, is extremely weak and broad and has only been clearly
9
J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006 15737

A R T I C L E S
Grills et al.
dihydride complex equilibrium will be significantly perturbed
immediately following the laser pulse. However, conversion
from the dihydride complexes to the η2-dihydrogen complex
has previously been estimated to occur on a significantly longer
By contrast, there is no clear evidence for such coordination in
toluene and hexane. We calculated UV-vis spectra of the W′
complex without any agostic C-H interaction and W′-n-hexane
in the gas phase. The lowest-energy absorption of the W′
complex without any agostic C-H interaction and of W′-n-
hexane is a dπ f dσ transition heavily mixed with CO*. The
second transition in both species is a metal-to-ligand charge
transfer (MLCT) transition (dπ to CO*) mixed with some
contribution of the dπ f dσ transition. The lowest-energy
absorption of all W-L complexes is MLCT in nature, i.e. a dπ
(with small contribution of L* and CO*) f L* + CO*
transition. Upon pulsed laser excitation into this band, the weak
W-L bond cleaves and the W species forms. The transient
UV-vis and FTIR spectra obtained 2 µs after excitation in
toluene and hexane are identical to those of W as shown in
Figures 1, 2, 8, 9, and 10. The formation of W-L from W and
L was directly observed, and all reactions (L ) H2, D2, N2,
C2H4, and CH3CN) are cleanly first-order in [L]. The second-
order rate constants (kf) for the reaction between W and L (L
) H2, D2, N2) in toluene obtained by UV-vis detection are in
good agreement with the values estimated by s2-FTIR detection
in hexane.
FTIR of W-L. Three IR-active ν(CO) bands (2A1 + B2)
are expected to be observed for the mer,trans-W(CO)3(PCy3)2L
complexes in accordance with their C2V symmetry (see Table
1). The A1(1) mode is the highest in frequency and extremely
weak, but is observable due to coupling with the A1(2) mode
and geometric distortions from the ideal mer,trans environ-
ment.39 The results of the frequency analysis in our DFT
calculations confirm that the A1(2) mode is the next lowest in
frequency and of medium intensity. Finally, the B2 mode is the
most intense of the three and occurs at the lowest frequency.
When this family of W-L complexes was initially reported,40
three ν(CO) bands were identified in their IR spectra. However
as described earlier, W-CO is generally present to a varying
degree in samples of W-L due to a disproportionation reaction
and/or minor oxidation. The single, intense ν(CO) band of
W-CO almost coincides with the A1(2) band of the W-L
complexes. For example, in hexane it is less than 0.5 cm-1
higher in frequency than the A1(2) mode of W-C2H4. In
subsequent publications the A1(2) band of the W-L complexes
has always been dismissed as merely a contamination by
W-CO, and only two of the ν(CO) bands (A1(1) and B2) have
since been reported and assigned for these complexes.1,3,5,10 Our
DFT calculations have now confirmed the expected presence
and relative intensities of all three IR-active ν(CO) modes for
W-L (Figure 8 and Table 1). The agostic species, W also has
C2V symmetry and exhibits three ν(CO) bands, similar to the
W-L complexes (see Figure 8 and Table 1). However, the
intensity ratio of these bands is different from that of W-L.
Despite the fact that it is not possible to model the agostic
interaction using PH3 ligands, the relative frequencies and
intensities of the IR bands predicted by the DFT calculations
agree well with the experimental IR spectrum of W.
i
time scale (τ ) 30 ms for W-H2 and τ ) 16 ms for the Pr
analogue)7,38 than that being probed in our experiments.
Therefore, any interconversion between the dihydride and η2-
dihydrogen complexes would be far too slow to contribute to
the measured kinetics. The conversion from the cis-dihydride
species may be faster, but the contribution from this minor
species has not been observed in our experiments. The transient
ν(CO) bands of W decay exponentially at a rate (kobs ) (2.7 (
0.3) × 104 s-1) identical to the recovery rate (kobs ) (2.4 (
0.2) × 104 s-1) of the W-H2 bands. However, on the time
scale of this experiment, the bleach at 1861 cm-1 only partially
recovers (by ca. 50%) and then remains constant. We assign
this to the overlap of the 1861 cm-1 band of the dihydrogen
complex, W-H2, with the ν(CO) band of the trans-W(CO)4-
(PCy3)2 (W-CO) contaminant. The cis-dihydride species also
has a band which overlaps the 1861 cm-1 band of W-H2, but
the contribution should be almost negligible. The long-lived
component of the bleach is likely to be associated with the
photochemistry of trans-W(CO)4(PCy3)2, which we are currently
in the process of investigating and which will be the subject of
a future report.
Analogous results were obtained for the s2-FTIR of W-D2
in the presence of 1.8 atm partial pressure of D2. W-D2 exhibits
six ν(CO) bands in hexane solution at frequencies almost
identical to those observed for the η2-dihydrogen/dihydride
complexes. These correspond to the η2-dideuterium and the
dideuterido forms of W-D2, which exist in equilibrium with
one another in solution. The ν(DD) stretching vibration of
W-D2 has previously been observed10 by Raman spectroscopy
as an extremely weak and broad feature at ca. 1900 cm-1
.
However, in the IR spectrum it is obscured by the ν(CO) band
at 1908 cm-1 and could not be probed by s2-FTIR. Similar to
the results obtained with W-H2, only the ν(CO) bands
corresponding to the η2-dideuterium complex were observed
to bleach upon excitation. The average lifetime for the decay
of W and the regeneration of W-D2 (τ ≈ 47 µs) in the s2-
FTIR of W-D2 is slightly longer than that observed for the
decay of W and the regeneration of W-H2 (τ ≈ 44 µs) in the
s2-FTIR of W-H2.
Our experimental investigation with step-scan FTIR and UV-
vis techniques on W-N2, together with a theoretical investiga-
tion of W′-N2, has been published previously.36
Discussion
UV-vis Spectra of W-L. Whereas the FTIR and NMR
spectra of W-L species have been extensively investigated,1
their UV-vis spectra have rarely been discussed. Only the UV-
vis spectrum of W in toluene has been previously reported.5
The visible spectra of W in toluene and hexane are practically
identical, showing λmax at 566 nm with a shoulder at 452 nm.
A small peak/shoulder at 384 nm is due to minor contamination
by W-CO. However, λmax in THF shifts to 550 nm, and the
shoulder at 452 nm is less pronounced, indicating that THF
weakly coordinates to the W center as suggested previously.5
In the s2-FTIR experiment performed on W-H2, only the
ν(CO) bands that correspond to the η2-dihydrogen complex were
observed to bleach upon 355 nm excitation, while those of the
dihydride species were unaffected. The photoejection of dihy-
(38) Khalsa, G. R. K.; Kubas, G. J.; Unkefer, C. J.; Vandersluys, L. S.; Kubat-
(39) Braterman, P. S. Metal Carbonyl Spectra; Academic Press: London, 1975.
(40) Kubas, G. J. J. Chem. Soc., Chem. Commun. 1980, 61-62.
Martin, K. A. J. Am. Chem. Soc. 1990, 112, 3855-3860.
9
15738 J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006

Rates of Small-Molecule Binding to W(CO)3(PCy3)2
A R T I C L E S
drogen from transition-metal dihydride complexes is a well-
known phenomenon and is a commonly used technique for the
generation of reactive unsaturated species.41-43 However, all
known examples of such photoelimination reactions occur with
cis-dihydrides. As discussed earlier, the results of our and
previous30,31 DFT calculations suggest that two dihydride
complexes exist: one with the H ligands in a cis geometry and
the other in a distal geometry with the H ligands separated by
one of the PCy3 ligands. The latter species is 0.7 kcal mol-1
more stable in our DFT calculations. The distal geometry may
prevent the photoelimination of dihydrogen from the dihydride
complex. However, it is possible that neither dihydride species
may absorb the 355 nm laser light. To consider the possibility
that neither dihydride species is excited by the 355 nm laser
pulse, we carried out TDDFT calculations to investigate the
UV-vis spectra of the dihydrogen and dihydride complexes of
W′, which are shown in Figure 7a. The calculated spectrum of
the W′-dihydrogen complex is blue-shifted compared to the
experimental spectrum of the W-dihydrogen complex. How-
ever, this is a general trend of TDDFT calculations and is also
due to the substitution of PCy3 by PH3 in the calculations. Since
the calculated absorption bands of both W′-dihydride com-
plexes are much more blue-shifted, the W-dihydride complexes
likely do not absorb the 355 nm laser light. These calculated
results are consistent with the observed transient FTIR spectra
in our s2-FTIR experiment performed on W-H2.
W-D2, and W-H2. Our B3LYP/6-31G** calculations for the
model complexes W′-L indicate that the W′-L binding free
energies are in the order W′-CO > W′-N2 > W′-C2H4 >
W′-H2 (dihydrogen) > W′-distal-(H)2 > W′-cis-(H)2. The
fact that W-C2H4 was found experimentally to be the least
stable species, whereas W′-C2H4 was calculated to have a
relatively large binding energy, may be due to a steric factor
associated with the bulky Cy groups in W-C2H4 and the η2-
coordination mode of C2H4. In fact, the crystal structure47 of
W-C2H4 indicates a long W-P distance to accommodate the
CdC bond of C2H4 parallel to P-W-P.
Isotope Effects. The kinetic isotope effects (KIEs) for the
forward and reverse reactions, kf(H )/kf(D ) and kr(H )/kr(D ), for the
2
2
2
2
formation of W-L (L ) H2 and D2), obtained by our
photoinduced method, are 1.3 ( 0.2 and 1.4 ( 0.3, respectively,
in toluene at 25 °C. The kinetic isotope effect of H2/D2
replacement by py obtained by Hoff’s group7 in toluene, kr(H )
/
2
kr(D ) ) 1.7 ( 0.1, is slightly larger than our H2/D2 removal
2
KIE (1.4 ( 0.3). However, these numbers are probably within
the respective experimental errors, because our number is
estimated from the intercepts (∼10% error) of Figure 3 for the
direct binding reactions, while Hoff’s value is obtained from
estimated rate constants (within (15% error) using complicated
substitution reactions, and also because of the existence of an
equilibrium between W-H2 and W-(H)2.
Calculated equilibrium constants for W-H2, W-D2, and
W-C2H4 are 2100 ( 300, 2300 ( 300, and 60 ( 10 M-1
,
Comparison of Binding Rate and Equilibrium Constants.
The second-order rate constant for the reaction of acetonitrile
respectively. The equilibrium isotope effect (EIE) for the W
complex has not been directly measured previously, but the
equilibrium shown in eq 6 has been used to obtain EIE )
KH/KD ) 0.70 ( 0.15 in THF at 22 °C using calibrated H2/N2
and D2/N2 gas mixtures.10 Our value in toluene at 25 °C is 0.91
( 0.15, which is larger than the value reported.10 Our
preliminary experiments in THF give an EIE of 0.80 ( 0.15 at
15 °C.
with W, kf(CH CN) ) (3.6 ( 0.3) × 106 M-1 s-1, is the highest
3
among the reactions we and others8 have measured for this series
of complexes. However, it is still only a factor of 2 larger than
that for H2 and D2. Hoff’s group evaluated kf(H ) and kf(N ) as
2
2
(2.2 ( 0.3) × 106 and (5.0 ( 1.0) × 105 M-1 s-1, respectively.7
Our directly measured values, kf(H ) ) (2.0 ( 0.1) × 106 M-1
2
s-1 and kf(N ) ) (3.0 ( 0.2) × 105 M-1 s-1 are similar to theirs,
2
but slightly lower. The rate constant for the formation of W-H2
in toluene is ca. 60× higher than that for the formation of
W-C2H4, mainly due to the small size of H2 compared to that
of C2H4, and therefore there is a decreased steric interaction
between the incoming ligand and the bulky PCy3 ligands.
Furthermore, this smaller steric repulsion is a major factor
contributing to the greater stability of W-H2 compared to that
of W-C2H4. Using the s2-FTIR technique, kf(C H ) in hexane
W-N2(soln) + H2(g) h W-H2(soln) + N2(g) (6)
KH/KD can be calculated from a mass and moment of inertia
factor (MMI), a vibrational excitation factor (EXC), and a zero-
point energy factor (ZPE) as shown in eq 7.10,44-46
EIE ) MMI × EXC × ZPE
(7)
2
4
was determined to be (7 ( 1) × 104 M-1 s-1, which is ca. 2×
The calculated value (EIE ) 0.78)10 for W with H2/D2 was
in agreement with their experimental value (0.70) and slightly
smaller than our experimental values.
higher than the value, kf(C H ) ) (3.4 ( 0.3) × 104 M-1 s-1
,
2
4
measured by UV-vis flash photolysis in toluene. While such a
36
difference was not observed in the formation of W-N2 and
Mechanistic Considerations of W-L Formation. When L
is photodissociated from W-L, a rapid equilibration with the
agostic W‚‚‚H-C interaction of a cyclohexyl group of one of
the PCy3 ligands with the metal center, and possibly a partial
coordination of the solvent via a σ C-H interaction of the
solvent with the metal center, will take place. The visible spectra
in hexane and toluene are identical, indicating that the agostic
W‚‚‚H-C interaction of the cyclohexyl group may be stronger
than the σ C-H interaction of solvents. Our calculated free
W-H2 (see Table 2), this suggests that the interaction of a
solvent molecule with the W center may be important in the
formation of W-C2H4. (See Mechanistic Considerations of
W-L Formation section).
While W-py and W-P(OMe)3 complexes are more stable
than W-NCCH3,9 the stabilities of the various adducts deter-
mined in this study are in the order W-NCCH3 > W-N2 >
W-D2 > W-H2 > W-C2H4, which match well with the
results obtained by Hoff’s group9 for W-NCCH3, W-N2,
(44) Bigeleisen, J.; Goeppert Mayer, M. J. Chem. Phys. 1947, 15, 261-267.
(45) Bullock, R. M. Isotope Effects in Reactions of Transition Metal Hydrides.
In Transition Metal Hydrides; Dedieu, A., Ed.; VCH Publishers: New York,
1992; pp 263-307.
(46) Bullock, R. M.; Bender, B. R. Isotope Methods - Homogeneous. In
Encyclopedia of Catalysis; Horva´th, I., Ed.; Wiley and Sons: New York,
2002; pp 281-348.
(41) Perutz, R. N. Pure Appl. Chem. 1998, 70, 2211-2220.
(42) Mawby, R. J.; Perutz, R. N.; Whittlesey, M. K. Organometallics 1995, 14,
3268-3274.
(43) Colombo, M.; George, M. W.; Moore, J. N.; Pattison, D. I.; Perutz, R. N.;
Virrels, I. G.; Ye, T. Q. J. Chem. Soc., Dalton Trans. 1997, 2857-2859.
9
J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006 15739

A R T I C L E S
Grills et al.
Table 4. Calculated Binding Energies, Enthalpies, Entropies, and Free Energies for Various W′-L, W, and Related Complexes
∆
E
∆
H
°
∆
S
°
∆G°
298
1
a
1
1
1
system
L
(kcal mol-
)
(kcal mol-
)
(cal K-1 mol-
)
(kcal mol-
)
binding to W(CO)3(PH3)2
binding to W(CO)3(PH3)2
binding to W(CO)3(PH3)2
binding to W(CO)3(PH3)2
binding to W(CO)3(PH3)2
binding to W(CO)3(PH3)2
binding to W(CO)3(PH3)2
binding to W(CO)5
CO
C2H4
N2
dihydrogen
distal-dihydride
cis-dihydride
n-hexane
n-hexane
agostic R-C-H of PCy3
agostic R-C-H of PCy3
agostic â-C-H of PH2Cy
-47.6
-26.9
-26.0
-16.8
-11.2
-10.5
-5.4
-45.8
-25.3
-24.3
-14.6
-9.9
-9.3
-4.4
-6.3
-5.1
-5.4
-0.9
-32.6
-41.9
-34.0
-27.0
-28.9
-30.9
-32.6
-32.1
0.6
-36.1
-12.8
-14.2
-6.5
-1.2
-0.9
5.3
-7.2
3.2
net agostic interactionb in W(CO)3(PCy3)2
net agostic interactionb in W(CO)3(PCy3)(PH3)
net agostic interactionb in W(CO)3(PH3)(PH2Cy)
-4.7
-5.2
-4.8
1.3
-5.2
-2.1
-7.4
-0.5
a Difference in total electronic energy without correction for zero-point energy. b The sum of the intrinsic agostic interaction energy and any steric and/or
electronic energy differences associated with the two different conformers.
energy for the formation of W′-n-hexane (∆G°298 ) 5.3 kcal
mol-1) indicates that such an interaction may be unfavorable
(Table 4). Hoff et al.9 observed a KIE of 1.20 ( 0.05 for the
reaction of W-py (and its deuterium-substituted species
W(CO)3[P(C6D11)3]2(py)) with P(OMe)3 in toluene and verified
the importance of the “agostic” W‚‚‚H-C interaction of a
cyclohexyl group in the ligand substitution reaction. This KIE
is similar to that of dissociation of py from W-py at low py
concentrations and indicates that dissociation of py is assisted
by partial formation of the C-H agostic bond at the transition
state.8 An inverse KIE of 0.87 for W with P(OMe)3 was
observed.8 Again, the agostic interaction must be involved in
the transition state.
been reported. However, those of metal dihydride complexes
were determined previously (-20 ( 1 and -16 ( 1 cm3 mol-1
for the formation of Ir(H)2Cl(CO)(PPh3) in toluene49 and
Rh(H)2(bpy)2 in methanol,19 respectively). The small activation
volumes found in the present study (viz., -3, -7, -3, and +3
cm3 mol-1 for the formation of W-H2, W-C2H4, W-N2, and
W-NCCH3, respectively) in toluene suggest that W-L bond
reformation involves the displacement of the fluxional agostic
W‚‚‚H-C interaction of a cyclohexyl group of the PCy3 ligand
and/or toluene in the transition state. These small absolute values
therefore result from a combination of bond formation and bond
cleavage processes as found in Ia and Id substitution mechanisms,
respectively.
Kubas’ group47 reported that the static complex, with an
agostic W‚‚‚H-C interaction of a cyclohexyl group of the PCy3
ligand, does not exist in toluene solution at room temperature
due to the fact that the 31P NMR shows only one signal at room
temperature and the high-field chemical shift for the agostic
hydrogen was not observed. Therefore, the agostic W‚‚‚H-C
interaction of a cyclohexyl group of the PCy3 ligand was
considered to be fluxional. While it is very important to consider
the agostic interaction of a cyclohexyl group of the PCy3 ligand
in ligand exchange or addition reactions, the concentration of
solvent is much higher than that of the metal species. Therefore,
a fluxional interaction of both a solvent molecule and an agostic
C-H interaction of a cyclohexyl group with the W center may
exist. Such a weak solvent interaction may be important in the
reaction of W with C2H4, in which the rate constant for
formation of W-C2H4 in hexane is 2× faster than that in
toluene. We did not observe solvent-dependent rate constants
for formation of W-L (L ) H2 and N2). Because the interaction
of C2H4 with W is much weaker than that of the other ligands
such as H2 and N2 (as evidenced by the much smaller
equilibrium constant for the reaction of W with C2H4), the rate
constant in the C2H4 experiment is likely to be more noticeably
affected by a weaker solvent coordination in hexane, whereas
for the other more strongly coordinating ligands, L, the rate
constants in hexane are not so perturbed from those measured
in toluene. Also, toluene offers the possibility of arene bonding
interactions that hexane does not,48 although such interaction
may be relatively weak due to the bulky phosphine groups
around the metal center.
Pulsed laser flash photolysis of M(CO)6 in solution with a
known amount of L (M ) Cr, Mo, W; L ) piperidine, 1-hexene,
pyridine, P(C6H5)3, P(OC2H5)3, etc.) has been investigated in
different solvents.50-53 A rapid formation of M(CO)5(solvent),
followed by the displacement of solvent by different entering
nucleophiles, L, was observed. For the W complex, small
positive activation volumes (0.4-2.7 cm3 mol-1) are found for
L ) 1-hexene in fluorobenzene, chlorobenzene, or heptane as
a solvent. This was interpreted in terms of a dissociative
interchange (Id) mechanism in which the coordinated solvent
is displaced by L. However, when THF was used as a solvent,
the reactions of W(CO)5(THF) with piperidine, P(C6H5)3, and
P(OC2H5)3 showed significantly negative activation volumes
(-4 to -15 cm3 mol-1). Thus, in the latter case there is real
evidence for an associative type (Ia) mechanism. It is worth
pointing out that negative activation volumes were observed in
reactions of all metal complexes performed in THF.50-53 Since
THF can bind to M(CO)5 more strongly than any other solvents
used in the experiments, an incoming ligand may need to form
a partial bond at the transition state, consistent with the Ia
mechanism. It seems that the reaction mechanism depends on
a balance between the interaction of W(CO)5 with solvents and
the nucleophilicity of the incoming ligand.
The photodissociation of L from W-L in the present systems
can cause a rapid equilibration with the agostic W‚‚‚H-C
(48) Cheng, T.-Y.; Szalda, D. J.; Bullock, R. M. J. Chem. Soc., Chem. Commun.
1999, 1629-1630.
(49) Schmidt, R.; Geis, M.; Kelm, H. Z. Phys. Chem. 1974, 92, 223-234.
(50) Zang, V.; Zhang, S.; Dobson, C. B.; Dobson, G. R.; van Eldik, R.
Organometallics 1992, 11, 1154-1158.
To our knowledge, activation volumes for the formation of
metal dihydrogen, -dinitrogen, or -ethylene complexes have not
(51) Zhang, S.; Dobson, G. R.; Zang, V.; Bajaj, H. C.; van Eldik, R. Inorg.
Chem. 1990, 29, 3477-3482.
(52) Zhang, S.; Zang, V.; Bajaj, H. C.; Dobson, G. R.; van Eldik, R. J.
Organomet. Chem. 1990, 397, 279-289.
(47) Butts, M. D.; Bryan, J. C.; Luo, X.-L.; Kubas, G. J. Inorg. Chem. 1997,
36, 3341-3353.
(53) Wieland, S.; van Eldik, R. Organometallics 1991, 10, 3110-3114.
9
15740 J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006

Rates of Small-Molecule Binding to W(CO)3(PCy3)2
A R T I C L E S
interaction with the metal center, but it could also involve partial
coordination of solvent, as found for M(CO)5(solvent) studied
previously.50-53 It is estimated that the energy of the agostic
interaction8,9 is 10 ( 6 kcal mol-1 based on a bond strength of
10 kcal mol-1 for Cr(CO)5(heptane). Note that the error limit
seems to be assigned arbitrarily, but the bond energy is expected
mol-1) in ∆G°298, however, could be within the errors associated
with the calculations.
Conclusions
In the present study, we investigated, for the first time, direct
laser flash photolysis measurements of rate constants and
activation volumes for the formation of W-L from W and L
in toluene and hexane solutions containing W-L (L ) H2, D2,
C2H4, N2, or CH3CN) and a known amount of L. W-L bond
dissociation took place to form W and subsequently, clean
regeneration of W-L was observed by the use of time-resolved
UV-vis and s2-FTIR spectroscopy. The results of the s2-FTIR
experiments demonstrate how TRIR spectroscopy is an ex-
tremely powerful technique for probing photoinitiated bond
cleavage and formation reactions of transition-metal carbonyl
complexes. The data are complementary to those obtained from
UV-vis flash photolysis experiments, providing a more accurate
spectroscopic picture of the mechanism of the photochemical
reaction, thus allowing the simultaneous identification of
multiple species and the temporal evolution of each one to be
followed. In addition, the high-pressure kinetic data enable a
more detailed picture of the ligand exchange process in the
transition state to be made. The small negative activation
volumes suggest that W-L (L ) H2, N2, or C2H4) bond
reformation involves the displacement of a fluxional agostic W‚
‚‚H-C interaction of a cyclohexyl group of a PCy3 ligand and/
or the methyl group of toluene in the transition states. DFT
calculations have elucidated the nature of the electronic transi-
tions, the assignments of IR frequencies, and bond energies for
W′-L complexes and W species with and without agostic W‚
‚‚H-C interactions. Our DFT calculations indicate an entropy
effect that favors agostic W‚‚‚H-C interaction over a solvent
to be in the wide range of 4-16 kcal mol-1 9
. Our DFT
calculated net agostic interaction energies are ∆E ) -4.7 and
∆H° ) -5.1 kcal mol-1, which are at the lower end of the
estimated range, and are comparable to the binding energy of
n-hexane (∆H° ) -4.4 kcal mol-1) to W′. We should point
out that the net agostic C-H interaction of Cy in W(CO)3-
(PCy3)(PH3) is also rather small (∆H° ) -5.4 kcal mol-1). The
displacement reaction with L can follow an associative or
dissociative interchange (Ia or Id) pathway mode, depending on
the nucleophilicity of the solvent, agostic interaction, and
incoming L. In the case of H2, N2, and C2H4, the negative
activation volumes suggest that bond formation is more
important than bond breakage (presumably involving mainly
the agostic W‚‚‚H-C interaction of a cyclohexyl group of a
PCy3 ligand, and possibly the σ bonding interaction of the
methyl group of toluene since it is present in huge excess) and
favors an Ia substitution mechanism similar to W(CO)5(THF)
mentioned above. These results were unexpected for a crowded
metal center such as W(CO)3(PCy3)2, but there is the precedent
that the KIE observed for the reaction of W-py with P(OMe)3
verified the involvement of the “agostic” W‚‚‚H-C interaction
of a cyclohexyl group at the transition state in the ligand
substitution reaction.9 In the case of CH3CN, the positive
activation volume suggests that bond cleavage is more important,
similar to what was found for W(CO)5(solvent) with 1-hexene
in fluorobenzene, chlorobenzene, and heptane solvent as men-
tioned above. The W-NCCH3 complex (K > 2 × 104 M-1) is
much more stable than W-H2, W-N2, and W-C2H4. The
binding rate constant (3.6 × 106 M-1 s-1) of CH3CN to W is
the largest we and others have studied using various incoming
ligands in toluene. Since CH3CN is much stronger as an
incoming ligand than H2, N2, and C2H4, the W‚‚‚H-C agostic
interaction and toluene coordination are relatively weak as found
for the solvent interaction with W(CO)5 in fluorobenzene,
chlorobenzene, and heptane. Therefore, in the CH3CN case the
photogenerated intermediate may follow an Id substitution
mechanism. Thus, our photoproduced intermediate, which we
postulate to have a fluxional behavior with W‚‚‚H-C agostic
and possibly W-toluene interactions, may behave in a way
similar to that found for M(CO)5(solvent) intermediates.
Now let us discuss the likelihood of a solvent σ C-H bonding
interaction with W vs an agostic interaction in W or its transition
states for binding a ligand, L, after the photoejection of L from
W-L. Our ∆G°298 values reflect a large entropy effect favoring
an agostic interaction (-T∆S° ≈ 0-2 kcal mol-1) in W and
disfavoring an interaction of the solvent with W (-T∆S° ≈ 10
kcal mol-1). This effect, combined with the small binding
energies calculated for the net agostic and σ C-H binding, leads
to ∆G°298 for formation of a solvent σ C-H bond with W being
positive by several kcal mol-1. Small differences (a few kcal
σ C-H interaction by 8-10 kcal mol-1
.
Acknowledgment. We thank Dr. Gregory Kubas (Los Ala-
mos National Laboratory) and Professors Michael George (U.
Nottingham, UK) and Carl D. Hoff (U. Miami) for insightful
suggestions and valuable discussions. We also thank Dr. Bruce
S. Brunschwig for his initial assistance in installing the high
pressure vessel in the flash photolysis system. This work was
performed at Brookhaven National Laboratory, funded under
contract DE-AC02-98CH10886 with the U.S. Department of
Energy and supported by its Division of Chemical Sciences,
Geosciences, & Biosciences, Office of Basic Energy Sciences.
RvE gratefully acknowledges financial support from the Deut-
sche Forschungsgemeinschaft (SFB 583) that enabled his par-
ticipation in this work.
Supporting Information Available: Pressure-dependent rates
for formation of W-L, DFT-optimized Cartesian coordinates
for 12 W species, plots of rate constants for formation of W-L
(N2, C2H4, and CH3CN) in toluene, a rate plot for formation of
W-C2H4 in hexane obtained by s2-FTIR, and Frontier orbitals
for W′-C2H4; complete ref 29. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA064627E
9
J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006 15741