Gas phase study of the kinetics of formation and dissociation of Fe(CO)4L and Fe(CO)3L2 (L = C2H4 and C2F4)

Article abstract of DOI:10.1021/jp970275+

The bond dissociation energy for loss of C₂H₄ from Fe(CO)₃(C₂H₄)₂, produced by the reaction of C₂H₄ + Fe(CO)₃(C₂H₄), has been determined as 21.3 ± 2.0 kcal/mol. An estimate is made for a lower limit for the bond dissociation energy of Fe(CO)₄(C₂H₄), which can be formed by reaction of CO + Fe(CO)₃(C₂H₄) or Fe(CO)₄ + C₂H₄ with rate constants of (4.3 ± 0.8) × 10^-12 and (1.7 ± 0.2) × 10^-13 cm³/(molecule s) at 24 °C, respectively. The values for these bond dissociation energies are compared with those determined in prior studies of these systems. A new compound with infrared absorptions at 2147, 2091, and 2068 cm^-1 is best assigned as Fe(CO)₃(C₂F₄)₂. A rate constant of (5.4 ± 1.7) × 10^-12 cm³/(molecule s) at 24°C is reported for the reaction of C₂F₄ with Fe(CO)₃(C₂F₄) to form Fe(CO)₃(C₂F₄)₂. Fe(CO)₄(C₂F₄) can be formed by reaction of C₂F₄ and Fe(CO)₄, with a rate constant of (1.8 ± 0.4) × 10^-14 cm³/(molecule s) at 24°C. Infrared absorptions observed at 2135, 2074, and 2043 cm^-1 are assigned to this species. The relative stabilities of the mono- and bisethylene and perfluoroethylene compounds of iron are compared. Where possible, they are also compared to the corresponding chromium compounds and are discussed in the context of current concepts regarding metal-olefin bonding.

Full text of DOI:10.1021/jp970275+

2988
J. Phys. Chem. A 1997, 101, 2988-2995
Gas Phase Study of the Kinetics of Formation and Dissociation of Fe(CO)4L and Fe(CO)3L2
(L ) C2H4 and C2F4)
Paul G. House and Eric Weitz*
Department of Chemistry, Northwestern UniVersity, EVanston, Illinois 60208-3113
ReceiVed: January 22, 1997X
The bond dissociation energy for loss of C2H4 from Fe(CO)3(C2H4)2, produced by the reaction of C2H4 +
Fe(CO)3(C2H4), has been determined as 21.3 ( 2.0 kcal/mol. An estimate is made for a lower limit for the
bond dissociation energy of Fe(CO)4(C2H4), which can be formed by reaction of CO + Fe(CO)3(C2H4) or
Fe(CO)4 + C2H4 with rate constants of (4.3 ( 0.8) × 10-12 and (1.7 ( 0.2) × 10-13 cm3/(molecule s) at 24
°C, respectively. The values for these bond dissociation energies are compared with those determined in
prior studies of these systems. A new compound with infrared absorptions at 2147, 2091, and 2068 cm-1 is
best assigned as Fe(CO)3(C2F4)2. A rate constant of (5.4 ( 1.7) × 10-12 cm3/(molecule s) at 24 °C is reported
for the reaction of C2F4 with Fe(CO)3(C2F4) to form Fe(CO)3(C2F4)2. Fe(CO)4(C2F4) can be formed by reaction
of C2F4 and Fe(CO)4, with a rate constant of (1.8 ( 0.4) × 10-14 cm3/(molecule s) at 24 °C. Infrared absorptions
observed at 2135, 2074, and 2043 cm-1 are assigned to this species. The relative stabilities of the mono- and
bisethylene and perfluoroethylene compounds of iron are compared. Where possible, they are also compared
to the corresponding chromium compounds and are discussed in the context of current concepts regarding
metal-olefin bonding.
I. Introduction
II. Experimental Section
Olefins have a long history as ligands in organometallic
compounds, dating to Zeise’s salt.1 Olefins can participate in
a variety of catalytic transformations induced by organometallic
species including isomerization, hydrogenation, and hydrosila-
tion processes.2-4 Efforts have been made to delineate the
microscopic kinetics involved in the complex kinetic mecha-
nisms that lead to such transformations.5-8 Recently, using
The apparatus used to monitor CO and C2H4 addition to Fe-
(CO)3(C2H4) has been described previously.12,13 Coordinatively
unsaturated metal carbonyls are photogenerated by the output
of either a frequency-tripled Nd:YAG laser (355 nm) or an
excimer laser, operating on either XeF (351 nm) or XeCl (308
nm). In each case, the laser delivered 7-10 mJ/cm2 at the cell
window. Photolysis of Fe(CO)5 at either 355 or 351 nm is
expected to produce an ∼60:40 mixture of Fe(CO)3 and Fe-
(CO)414 while photolysis at 308 nm produces almost exclusively
Fe(CO)3.15
time-resolved infrared spectroscopy, it has been possible to
5,7,8
identify and demonstrate that Fe(CO)3(C2H4)2
and H2Fe-
(CO)3(C2H4)7,8 are crucial intermediates in the gas phase
catalytic hydrogenation of ethylene induced by photolysis of
Fe(CO)5. It has also been pointed out that a knowledge of the
rates of dissociation of these species, which depend on the bond
dissociation energies, is crucial to formulating a complete model
of the kinetics of this system.8
The infrared beam from either a home-built, line-tunable, CO
laser or a diode laser was used to probe the kinetics of
association reactions. The beam double passed either a 16 or
a 42 cm long gas cell terminated with CaF2 windows. Detection
of the infrared laser beam was by a fast (τ1/2 ≈ 70 ns) InSb
detector. The output of the detector was amplified and sent to
a digital oscilloscope which averaged 20-40 wave forms. The
digitized, averaged signal was sent to a computer for fitting
and analysis.
For more than 40 years, the Dewar-Chatt-Duncanson model
has provided a framework for discussing metal-olefin bond-
ing.9,10 As implied above, a knowledge of the magnitudes of
olefin-metal bonds can be crucial to a full understanding of
the reaction mechanisms involving such species. Though there
has been progress in the ability to experimentally measure and
to quantitatively calculate bond energies, information in this
area is still limited.
This study employs time-resolved infrared spectroscopy to
monitor the rates of formation and dissociation of a number of
iron carbonyl-olefin species. Where possible, comparisons are
made to corresponding Cr compounds. An objective is to obtain
a better understanding of how the nature of the olefin, the
number of olefin ligands, and the nature of the metal center
influence bonding and bond dissociation energies. Systems
studied include Fe(CO)4(C2H4), Fe(CO)3(C2H4)2, Fe(CO)4(C2F4),
and an iron carbonyl compound containing multiple C2F4
ligands. The infrared absorptions of this latter compound are
not compatible with those previously assigned to iron tetracar-
bonyl-cyclooctafluorotetramethylene.11 As such, the absorp-
tions identified in this study are best assigned to a new
compound, Fe(CO)3(C2F4)2, an analog of the previously identi-
fied compound, iron bisethylene-tricarbonyl.
Rate constants for association reactions were measured under
the indicated conditions in a static cell: Fe(CO)3(C2H4) + C2H4
- 0.200-0.210 Torr of Fe(CO)5, 0.49-11.42 Torr of C2H4,
and enough He to bring the total pressure to at least 90 Torr;
Fe(CO)3(C2H4) + CO - 0.200-0.211 Torr of Fe(CO)5, 1.22-
2.02 Torr of C2H4, 0.00-8.36 Torr of CO, and He to raise the
total pressure to at least 85 Torr; Fe(CO)4 + C2H4 - 0.030-
0.037 Torr of Fe(CO)5, 11.2-153.5 Torr of C2H4, and enough
He to bring the total pressure to at least 90 Torr; Fe(CO)3 +
C2F4 - ∼0.025 Torr Fe(CO)5, 0.000-1.010 Torr C2F4 and
enough He to bring the total pressure to at least 80 Torr; Fe-
(CO)4 + C2F4 - ∼0.1 Torr of Fe(CO)5, 3.66-51.11 Torr of
C2F4, and either ∼2.4 or ∼3.9 Torr of CO and He to bring the
total pressure to at least 75 Torr; Fe(CO)3(C2F4) + C2F4 - ∼
0.050 Torr of Fe(CO)5, 1.055-9.00 Torr of C2F4, and enough
He to bring the total pressure to at least 85 Torr. All of the
experiments that involved ethylene as a ligand were performed
using 351 nm photoysis pulses while those with perfluoroeth-
ylene as a ligand were performed using 308 nm photoysis pulses.
X Abstract published in AdVance ACS Abstracts, March 15, 1997.
S1089-5639(97)00275-2 CCC: $14.00 © 1997 American Chemical Society

Kinetics of Fe(CO)4L and Fe(CO)3L2
J. Phys. Chem. A, Vol. 101, No. 16, 1997 2989
Dissociation of Fe(CO)3(C2H4)2 was monitored with a Matt-
son RS1 FTIR operating at 4 cm-1 resolution and averaging 25
scans. A spectrum was obtained every 3 s for the first 35 s
and every 5.7 s thereafter. Samples were followed out to 0.5-
10 min, depending on the ratio of C2H4:CO and the temperature.
Samples of 0.098-0.134 Torr of Fe(CO)5, 19.6-200 Torr of
CO, and 404-641 Torr of C2H4 were placed in a 42 cm long,
2 1/2 cm diameter static gas cell. The cell contents were allowed
to thermally equilibrate and mix for 30 min before being
photolyzed for ∼20 s at 10 Hz with the output of the frequency-
tripled Nd:YAG laser.
Experiments directed toward a measurement of the bond
dissociation energy for Fe(CO)4(C2H4) involved photolysis of
mixtures of 0.020-0.203 Torr of Fe(CO)5, 20-330 Torr of CO,
and 100-565 Torr of C2H4 with ∼200 laser pulses from a
frequency-tripled Nd:YAG laser. Analogous experiments di-
rected toward measurements of the bond dissociation energy
for Fe(CO)4(C2F4) and Fe(CO)3(C2F4)2 involved photolysis of
mixtures of 0.100-0.200 Torr of Fe(CO)5, 8.3-111.0 Torr of
CO, and 39.5-65.6 Torr of C2F4 with between ∼200 and 1000
pulses from the frequency-tripled Nd:YAG laser. The cells used
for these experiments were covered with opaque tape to inhibit
photolytic processes due to ambient light.
For experiments above room temperature, the cells were
wrapped with heating tape and insulated with cotton batting.
The windows of the 16 cm cell were warmed with two 60 W
incandescent light bulbs to inhibit the condensation of iron
carbonyls. Chromel-alumel thermocouples, attached to the
outside of the cells, monitored the temperature. The temperature
uncertainty for the 16 cm cell was (1 K and (2 K for the 42
cm cell.
Fe(CO)5 was obtained from Aldrich Chemical and subjected
to a series of freeze-pump-thaw cycles before use. At the
beginning of each day, the iron pentacarbonyl was briefly
pumped on to remove CO and any volatile polynuclear iron
species that were present due to decomposition. C2F4 (97%)
was purchased inhibited from PCR Corp. The gas was passed
through a copper tube packed with activated charcoal to remove
the D-limonene inhibitor. Based on a calibration with neat
D-limonene, FTIR measurements of the C2F4, subsequent to this
procedure, showed the D-limonene concentration was less than
0.07%. The following gases were obtained from Matheson, at
the indicated purity, and were used as received: C2H4, 99.5%;
CO, 99.9%; He, 99.999%.
Figure 1. Plot of the negative of the change in relative intensity of
the probe beam at 1966 cm-1 versus time following 355 nm photolysis
of 0.205 Torr of Fe(CO)5, 2.02 Torr of C2H4, 3.10 Torr of CO, and
88.8 Torr of He. The rise is due to formation of Fe(CO)3(C2H4) which
is depleted on reaction with CO or C2H4.
(C2H4). Fe(CO)3(C2H4)2 is initially produced by sequential
addition of C2H4 to Fe(CO)3. Fe(CO)4(C2H4), the product in
reaction 3, dissociates on a much longer time scale than Fe-
(CO)3(C2H4)2. As a result, addition of CO to Fe(CO)3(C2H4)
is effectively an irreversible step, taking the coordinatively
unsaturated species out of the formation-dissociation cycle for
Fe(CO)3(C2H4)2. Under experimental conditions, as applied to
the dissociation of Fe(CO)3(C2H4)2, the process in eq 4 is not
expected to be significant and can be neglected. Applying the
steady state approximation to Fe(CO)3(C2H4) results in the
following expression for the dissociation rate constant, kd, for
Fe(CO)3(C2H4)2:
k2[C2H4]
kd ) kobs
+ 1
(5)
(
)
k3[CO]
kobs is a first-order phenomenological rate constant for the
disappearance of the substituted species and the appearance of
the compound resulting from CO addition, in this case Fe(CO)4-
(C2H4). The temperature dependence of kd can be related to
12
the bond dissociation energy;
therefore, it is necessary to
measure the temperature dependence of k2, k3, and kobs to
determine a bond dissociation energy.
A. Addition Rate Constants. The rate constants for the
various addition reactions were determined from plots of the
rate of the addition process versus pressure of the appropriate
ligand (under pseudo-first-order conditions), monitored at an
absorption of either or both the iron carbonyl reactant or the
reaction product. Experiments were performed to assure
measurements on all reactions were in the high-pressure limit
below the total pressure listed in section II.13
1. Fe(CO)3(C2H4) + CO. A plot of a typical signal,
monitored at the 1966 cm-1 absorption of Fe(CO)3(C2H4) used
to probe the reaction of Fe(CO)3(C2H4) + CO, is shown in
Figure 1. Figure 2 shows a plot of the rate for this process
versus CO pressure at 35 °C, which yields a rate constant of
(4.3 ( 0.8) × 10-12 cm3/(molecule s). Within experimental
error the rate constant is independent of temperature from 24
to 47 °C. The errors on measured rate constants are (2σ and
are based solely on precision.
III. Results
The kinetic scheme used to determine bond dissociation
energies (BDEs), based on dissociative loss of a ligand bound
to a metal carbonyl, has been previously described.16 Conditions
are set up such that the rate of loss of a weakly bound ligand
is rate-limiting in the regeneration of parent or a product
produced by addition of another ligand, typically CO. For the
case of Fe(CO)3(C2H4)2, the relevant reactions are
kd
Fe(CO)3(C2H4)2
98 Fe(CO)3(C2H4) + C2H4
(1)
k2
Fe(CO)3(C2H4) + C2H4
Fe(CO)3(C2H4) + CO
9
8 Fe(CO)3(C2H4)2
(2)
(3)
k3
98 Fe(CO)4(C2H4)
2. Fe(CO)3(C2H4) + (C2H4) f Fe(CO)3(C2H4)2. The rate
constant for ethylene addition to Fe(CO)3(C2H4), k2, has been
previously measured to be (10.8 ( 0.6) × 10-12 cm3/(molecule
s) at 23 °C.7 In this work k2 was measured to be (11.1 ( 0.8)
× 10-12 cm3/(molecule s) at 24 °C, in good agreement with
the prior work, by plotting the rate of loss of Fe(CO)3(C2H4) at
k4
Fe(CO)3(C2H4) + Fe(CO)5
98 Fe2(CO)8(C2H4) (4)
where it is assumed that the polynuclear compound formed as
a result of reaction of Fe(CO)3(C2H4) with Fe(CO)5 is Fe2(CO)8-

2990 J. Phys. Chem. A, Vol. 101, No. 16, 1997
House and Weitz
Figure 2. Rate of disappearance of Fe(CO)3(C2H4) probed at 1966
cm-1 plotted versus CO pressure.
1966 cm-1 versus pressure of C2H4. The rate constant is
temperature independent from 24 to 45 °C.
3. Fe(CO)5 + Fe(CO)3(C2H4) f Fe2(CO)8(C2H4). The
reaction of Fe(CO)5 and Fe(CO)3(C2H4) is included in the kinetic
scheme discussed above, as eq 4, and can potentially act as a
termination step in the catalytic hydrogenation of C2H4.8 The
kinetic equation for the loss of [Fe(CO)3(C2H4)] is
Figure 3. Spectra generated on 355 nm photolysis of Fe(CO)5, C2F4,
and CO. A prephotolysis spectrum was used as the background in both
experiments. (A) The solid line spectrum was recorded after 200 laser
pulses irradiated a mixture of 0.106 Torr of Fe(CO)5, 41.6 Torr of C2F4,
and 48.7 Torr of CO. The dotted line spectrum was recorded after 1000
laser pulses irradiated a mixture of 0.106 Torr of Fe(CO)5, 39.6 Torr
of C2F4, 6.4 Torr of CO, and 54.1 Torr of He. (B) The solid line in
panel A, assigned as Fe(CO)4(C2F4), was subtracted from the dotted
line spectra in (A). The resulting spectrum (dotted line) is assigned to
Fe(CO)3(C2F4)2. In both panels a reference spectrum of Fe(CO)5 has
been added in to compensate for photolytic loss of Fe(CO)5.
-d[Fe(CO)3(C2H4)]
) (k3[CO] + k2[C2H4] +
dt
k4[Fe(CO)5])[Fe(CO)3(C2H4)] (6)
C2F4 pressure to give a rate constant of (5.4 ( 1.7) × 10-12
Experiments to determine the rate constant for addition of C2H4
to Fe(CO)3(C2H4), at 1966 cm-1, were performed at a constant
pressure of Fe(CO)5, in the absence of added CO, under the
conditions indicated in section II. Only a small portion of the
initial concentration of Fe(CO)5 is photolyzed during each
experiment. Thus, [Fe(CO)5] . [CO], where the CO is
produced photolytically, and the first term on the right-hand
side of eq 6 can be ignored. The intercept of a plot of reaction
rate, in eq 6, versus C2H4 pressure gives the rate of formation
of polynuclear species which on division by the Fe(CO)5
concentration yields k4 ) (4 ( 2) × 10-11 cm3/(molecule s) at
24 °C.
4. Fe(CO)4 + C2H4 f Fe(CO)4(C2H4). The rate constant
for addition of ethylene to Fe(CO)4 was determined to be (1.7
( 0.2) × 10-13 cm3/(molecule s) by plotting the rate of the rise
of Fe(CO)4(C2H4), at 2024 cm-1, versus the pressure of C2H4
at 24 °C. The rate constant is temperature independent up to
46 °C.
cm3/(molecule s) at 24 °C.
B. Fe(CO)3(C2F4)2 and Fe(CO)4(C2F4) Absorptions. The
spectrum indicated by the solid line in Figure 3A was generated
by photolysis of 0.106 Torr of Fe(CO)5, 41.6 Torr of C2F4, and
48.7 Torr of CO with 200 pulses from the tripled Nd:YAG laser.
The spectrum indicated by the dotted line was generated by
photolysis of 0.106 Torr of Fe(CO)5, 39.6 Torr of C2F4, 6.4
Torr of CO, and 54.1 Torr of He photolyzed with 1000 pulses.
In each case a full background spectrum was taken before
photolysis, and an Fe(CO)5 reference spectrum has been used
to compensate for photolytic loss of Fe(CO)5.
Iron carbonyl compounds, containing C2F4 as ligands, have
been previously reported.11,17-20 Fe(CO)4(C2F4) has been
observed and studied by electron diffraction19 and NMR.20 Fe-
(CO)4(C2F4)2, which is thought to have two C2F4 ligands
arranged in a cyclic structure17 so that its formula can be written
as Fe(CO)4(CF2)4, has absorptions at 2150, 2092, 2072, and
2056 cm-1 in hexane.11 The absorptions, in the carbonyl stretch
region, of Fe(CO)4(C2H4), have been recorded in hydrocarbon
solution and in the gas phase. The average shift is ∼10 cm-1
to lower frequency for absorptions observed in solution relative
to the gas phase.21,22 The experimentally observed peaks at
2147, 2135, 2091, 2074, and 2043 cm-1 do not provide a good
correspondence relative to the previously reported absorptions
of Fe(CO)4(CF2)4 when the expected liquid to gas phase shifts
are considered. In addition, there is no absorption reported for
Fe(CO)4(CF2)4 that corresponds to the 2043 cm-1 absorption
observed in these experiments.
5. Fe(CO)3 + C2F4 f Fe(CO)3(C2F4). The rate of loss of
Fe(CO)3 at 1955 cm-1 was plotted against C2F4 pressure to give
a rate constant of (3.3 ( 1.2) × 10-11 cm3/(molecule s) at 24
°C.
6. Fe(CO)4 + C2F4 f Fe(CO)4(C2F4). The rate constant
for Fe(CO)4 + C2F4 was determined by monitoring the rate of
loss of Fe(CO)4 and the rise of Fe(CO)4(C2F4). CO present in
the cell can react with nascent Fe(CO)3 to generate Fe(CO)4
subsequent to 308 nm photolysis of Fe(CO)5. A plot of the
rate of loss of Fe(CO)4 at 1990 cm-1 versus C2F4 pressure gives
a rate constant of (1.3 ( 0.5) × 10-14 cm3/(molecule s) at 24
°C. A plot of the rate of rise of product at 2076 cm-1 versus
C2F4 pressure gave a rate constant of (2.2 ( 0.6) × 10-14 cm3/
(molecule s) at 24 °C. These measurements yield an error
weighted average rate constant of (1.8 ( 0.4) × 10-14 cm3/
(molecule s).
Another indication that Fe(CO)4(CF2)4 is not observed in the
present experiments is the changes in relative peak intensities
as a function of conditions. As can be seen in Figure 3A,
changes in photolysis conditions and C2F4:CO ratio result in
changes in the relative intensities of the peaks at 2091 and 2074
cm-1. This clearly indicates that both of these absorptions
cannot belong to the same species and thus the observed
absorptions are due to more than one compound. Additionally,
the absorptions that best correlate with some of those that would
7. Fe(CO)3(C2F4) + C2F4 f Fe(CO)3(C2F4)2. The rate of
rise of an absorption at 2092 cm-1, belonging to a species
assigned as Fe(CO)3(C2F4)2 (see section B), was plotted against

Kinetics of Fe(CO)4L and Fe(CO)3L2
J. Phys. Chem. A, Vol. 101, No. 16, 1997 2991
Fe(CO)4(C2HF3) were studied in light petroleum solutions.23 As
such, these absorptions are plotted with the aforementioned
anticipated 10 cm-1 shift toward higher frequency to allow the
positions of these absorptions to be more directly comparable
to gas phase numbers. The general trend for the effect of
additional fluorines on the positions of the absorption of these
compounds is clear and consistent with fluorine acting as an
electron-withdrawing moiety. Most likely, the B2 modes of Fe-
(CO)4(C2H4) and Fe(CO)4(C2F4) are not observed because they
are masked by the strong Fe(CO)5 absorption. However, even
if the lowest observed Fe(CO)4(C2F4) absorption is the B2 mode,
the overall conclusions regarding trends in the frequency of a
given mode as a function of the number of fluorine atoms
remains unchanged, as does the assignment of the absorptions
at 2135, 2074, and 2043 cm-1 to Fe(CO)4(C2F4).
Figure 4. Frequencies of the carbonyl stretching absorptions in
Fe(CO)4(C2H4-xFx) are plotted versus X. The spectra of Fe(CO)4(C2H4)
were obtained in the gas phase, and the positions indicated for the low-
frequency A1 and B1 absorptions of Fe(CO)4(C2H4) are the average of
the positions reported in refs 6 and 21. The absorptions reported in
light petroleum have been shifted by +10 cm-1 to facilitate comparison
with gas-phase absorptions. Symbols: b, high-frequency A1 absorption;
9, low frequency A1 absorption; 2, B1 absorption.
In principle, the second compound, which we have indicated
is likely Fe(CO)3(C2F4)2, could be an Fe(CO)x(C2F4)5-x (x e
2) species. This compound would have to form by sequential
photolysis of Fe(CO)4(C2F4) and/or Fe(CO)3(C2F4)2. It would
not be expected to form as a result of an initial photolytic process
since 355 nm photolysis of Fe(CO)5 is expected to produce
significant quantities of only Fe(CO)4 and Fe(CO)3.14 However,
there is nothing clearly apparent in the data in Figure 3. and
related experiments that would indicate the presence of more
than two compounds. Thus, if one of the two compounds is
Fe(CO)x(C2F4)5-x (x e 2), then an intermediate must be
photolyzed to produce this species before the concentration of
the intermediate can build up to significant concentrations. The
most likely intermediate for this process would be Fe(CO)3-
(C2F4)2, which would have to have a large photolysis cross
section at 355 nm relative to both Fe(CO)4(C2F4) and the
putative Fe(CO)x(C2F4)5-x (x e 2) species. However, Fe-
(CO)x(olefin)5-x (x e 2) species are anticipated to be unstable.
In fact, even Fe(CO)2(olefin)3 compounds are very limited.24-26
Though Fe(CO)3(C2F4)2 is clearly more stable than Fe(CO)3-
(C2H4)2, it would not be surprising based on the aforementioned
data that even if Fe(CO)2(C2F4)3 were to form in the experiments
reported in Figure 3, it would be short-lived and not observed.25
Thus, though we cannot rule out the possible formation and
observation of an Fe(CO)x(C2F4)5-x (x e 2) species, we feel it
is very likely that the simplest interpretation of this data is
accurate: we are observing Fe(CO)4(C2F4) and Fe(CO)3(C2F4)2.
However, if one of the species we are observing is Fe-
(CO)x(C2F4)5-x (x e 2) instead of Fe(CO)3(C2F4)2, then, as
commented on again in section IV.D, its stability is even more
remarkable.
be expected for Fe(CO)4(CF2)4 are favored by lower C2F4
concentrations and shorter irradiation time: opposite what would
be anticipated if these absorptions belonged to Fe(CO)4(CF2)4.18
Thus, we have no evidence to indicate that Fe(CO)4(CF2)4 is
produced in the present experiments. Experiments in which
this compound was reported involved relatively long time scale
solution phase photolysis11 compared to the pulsed gas phase
photolysis experiments described herein.
The simplest interpretation that is consistent with the data in
Figure 3, and other related data generated by photolysis of Fe-
(CO)5 in the presence of different concentrations of CO and
C2F4, is that two different C2F4-containing species are produced
by photolysis. The solid line spectrum in Figure 3A shows
peaks at 2135 (0.04), 2074 (1.00), and 2043 cm-1 (0.78), which
are assigned to Fe(CO)4(C2F4) due to their greater relative
absorption with shorter photolysis times and lower C2F4:CO
ratios. Additionally as indicated in section III.A.7, the rate of
loss of Fe(CO)4 in the presence of C2F4 and the rate of rise of
the 2074 cm-1 absorption (monitored at 2076 cm-1) assigned
to Fe(CO)4(C2F4) agree within experimental error. The numbers
in parentheses are the approximate relative peak heights. The
peaks at 2147 (0.08), 2091 (1.00), and 2068 cm-1 (0.60) (Figure
3B), which are favored by longer photolysis times and higher
C2F4:CO ratios, are assigned to another C2F4-containing iron
carbonyl compound. This compound has not been previously
reported, but its behavior is consistent with it being Fe(CO)3-
(C2F4)2, an analog of Fe(CO)3(C2H4)2. It would be expected
to form, under conditions used to generate the spectra in Figure
3, by addition of a second C2F4 to a Fe(CO)3(C2F4) moiety that
is generated by photolytic loss of CO from Fe(CO)4(C2F4) or
directly by addition of two molecules of C2F4 to photolytically
generated Fe(CO)3. As indicated in section III.A.7, and
consistent with this assignment, the reaction that occurs involv-
ing Fe(CO)3, which is almost the exclusive photoproduct of 308
nm photolysis of Fe(CO)5, and C2F4 was monitored at 2092
cm-1. This frequency lies within the strongest absorption of
the species observed in Figure 3B.
It would, in principle, be possible for C2F4 to oxidatively add
to Fe(CO)3 to produce Fe(CO)3(CF2)2. This possibility is made
more plausible because C2F4 has a carbon-carbon bond energy
of 76.3 kcal/mol.27 However, if oxidative addition were to take
place, the rate of formation of Fe(CO)3(CF2)2 would equal the
rate of reaction of Fe(CO)3 with C2F4. A comparison of rate
constants reported in section III.A indicates this is not the case,
and thus formation of Fe(CO)3(CF2)2 by this route can be ruled
out.
C. kobs for Fe(CO)3(C2H4)2 f Fe(CO)4(C2H4). The FTIR
was used to monitor Fe(CO)3(C2H4)2 absorptions at 2070 cm-1
and 1997 cm-1 and an Fe(CO)4(C2H4) absorption at 2095 cm-1
.
Both species have absorptions between 2020 and 2000 cm-1 6
that could not be easily resolved because they overlap with
absorptions of Fe(CO)5. kobs was obtained from single-
exponential fits of absorbance versus time for each of the three
observed peaks. Figure 5 is a plot of the time evolution of the
spectrum, in the 2050-2150 cm-1 region, which depicts the
decay of the Fe(CO)3(C2H4)2 absorption at 2070 cm-1 and the
The assignment of absorptions at 2135, 2074, and 2043 cm-1
to Fe(CO)4(C2F4), which appears to be the first literature report
of infrared absorptions for this compound, is supported by their
positions relative to the IR absorptions of related compounds.
Figure 4 shows the literature values for the infrared absorptions,
in the CO stretching region, for Fe(CO)4(C2H4-xFx) compounds
where x ) 0-4. Fe(CO)4(C2H3F), Fe(CO)4(C2H2F2), and

2992 J. Phys. Chem. A, Vol. 101, No. 16, 1997
House and Weitz
A possible explanation for the decay of the Fe(CO)4(C2H4)
signal, to a nonzero absorption value, is an equilibrium between
Fe(CO)4(C2H4) + CO and Fe(CO)5 + C2H4. However, there
was no evidence for production of Fe(CO)4(C2H4) when Fe-
(CO)5 and C2H4 were allowed to stand for 24 h in the dark,
making this explanation unlikely. The initial fast decay of the
Fe(CO)4(C2H4) signal, to a nonzero value, could be a result of
heterogeneous decay and/or adsorption of Fe(CO)4(C2H4) on
the cell walls and/or windows. The cell surfaces then become
passivated, and since there is no significant homogeneous gas
phase dissociation of Fe(CO)4(C2H4) taking place on the
experimental time scale, the Fe(CO)4(C2H4) absorbance levels
off. However, the rate and amplitude of the decay did not
change significantly when the cell was partially filled with glass
beads. Thus, it is possible that the decay process occurs
preferentially on the windows of the cell. Nevertheless,
whatever the initial loss mechanism, it is consistent with an
initial acceleration of the rate of loss of Fe(CO)4(C2H4).
Therefore, these observations can be used to estimate an upper
limit for the gas phase dissociation rate constant for Fe(CO)4-
(C2H4).
Figure 5. Time-resolved spectra generated by 355 nm photolysis of
0.101 Torr of Fe(CO)5, 494 Torr of C2H4, and 160 Torr of CO. The
four spectra are separated by 30 s intervals starting at t ) 0. Arrows
show the direction of evolution of the peaks. The inset is a plot of the
absorbance versus time for the peaks at 2095 (rise) and 2070 cm-1
(fall).
Using the measured rate constant for addition of C2H4 to Fe-
(CO)4 and the previously reported rate constant for CO addition
28
to Fe(CO)4 in conjunction with eq 5, an upper limit for the
rate constant for Fe(CO)4-C2H4 dissociation of 6 × 10-6 s-1
was estimated at 24 °C. However, the decay may be almost
exclusively due to a heterogeneous reaction, and thus the actual
rate constant for homogeneous gas phase dissociation could be
much smaller than this limiting value.
F. Fe(CO)3(C2F4)2 and Fe(CO)4(C2F4) Dissociation. At-
tempts were also made to directly determine the rate constant
for dissociation of Fe(CO)3(C2F4)2 and Fe(CO)4(C2F4). The time
dependence of the absorptions for both of these species behaved
very similarly to what was observed with Fe(CO)4(C2H4). They
exhibited a relatively rapid partial decay, at a rate essentially
independent of temperature, and C2F4:CO ratio, which leveled
off to a nonzero value and exhibited very little subsequent
change over a period of days.
Figure 6. Arrhenius plot for the loss of C2H4 from Fe(CO)3(C2H4)2.
The decomposition rate constant, kd, is plotted versus 103/T.
TABLE 1: Rate Constant, kd, for Loss of C2H4 from
Fe(CO)3(C2H4)2
The observed decay rate of Fe(CO)3(C2F4)2 is consistent with
the perfluoroethylene being more strongly bound than ethylene
in Fe(CO)3(C2H4)2. Under similar reaction conditions, Fe(CO)3-
(C2H4)2 completely decays away, as a result of the homogeneous
gas phase processes in eqs 1-4, in a few minutes, while the
Fe(CO)3(C2F4)2 peak does not fully disappear even after a
number of days. An upper limit for kobs for Fe(CO)3(C2F4)2
requires knowledge of the rate constant for addition of C2F4 to
Fe(CO)3(C2F4), which has been measured, and the rate constant
for addition of CO to Fe(CO)3(C2F4), which was not measured
since an absorption for Fe(CO)3(C2F4) was not identified.
temp (°C)
kd (s-1
)
temp (°C)
kd (s-1
)
24
30
33
0.15 ( 0.08
0.34 ( 0.12
0.5 ( 0.3
37
42
0.6 ( 0.3
1.6 ( 0.9
growth of the Fe(CO)4(C2H4) absorption at 2095 cm-1, at 30 s
intervals. The inset in this figure shows the absorbance of the
two peaks vs time over a longer time period.
D. Fe(CO)3(C2H4)-C2H4 Bond Dissociation Energy.
Table 1 shows the average kd for a range of temperatures for
the dissociation of Fe(CO)3(C2H4). For unactivated association
reactions, the energy of activation for the corresponding
dissociation reaction can be directly related to the change in
internal energy, ∆U, and the bond dissociation energy.12 The
bond dissociation energy for loss of C2H4 from Fe(CO)3(C2H4)2,
based on the Arrhenius plot in Figure 6, is 21.3 ( 2.0 kcal/
mol, and ln A ) 33 ( 3, where A is the preexponential for kd.
E. Fe(CO)4(C2H4) Dissociation. Attempts to directly
measure the rate for loss of C2H4 from Fe(CO)4(C2H4), as a
result of a dissociative homogeneous gas phase process, were
unsuccessful. The Fe(CO)4(C2H4) peak at 2095 cm-1 was
monitored for an extended period of time. Over the first few
hours, the peak amplitude dropped significantly and reached a
nonzero absorption with little additional change over the next
few days. There was no apparent trend in the rate of the initial
fall as the C2H4:CO ratio and temperature were varied.
Based on the relative positions of absorptions of ethylene
versus perfluoroethylene-containing iron compounds, the ab-
sorption of Fe(CO)3(C2F4), which corresponds to the absorption
of Fe(CO)3(C2H4) that was used to monitor the addition rate of
C2H4 to Fe(CO)3, could be masked by the intense Fe(CO)5 peak.
As such, a gas kinetic rate constant, which would result in the
smallest value of kd, was assumed for addition of CO to Fe-
(CO)3(C2F4). Based on these conditions and the lack of
significant long-term decay of Fe(CO)3(C2F4)2, eq 5 can be used
to provide an estimate of an upper bound for kd of ∼6 × 10-6
s-1, the same value as the estimate for Fe(CO)4(C2H4). Though
this same limiting value was also determined for Fe(CO)4(C2H4),
it should be emphasized that these limits may be almost
exclusively for heterogeneous processes, and the homogeneous
gas phase processes are likely to be slower and their rates may
be different for the different compounds.

Kinetics of Fe(CO)4L and Fe(CO)3L2
J. Phys. Chem. A, Vol. 101, No. 16, 1997 2993
TABLE 2: Some IR Absorptions of Fe(CO)5, Fe(CO)4, and
Fe(CO)4(C2H4)
IV. Discussion
A. Fe(CO)3(C2H4)-C2H4 Bond Dissociation Energy. The
dissociation of Fe(CO)3(C2H4)2 has been previously investi-
gated.5,6 In ref 6, kd and the ratio k3/k2 were reported as (2.9 (
0.3) × 10-3 s-1 and 35 ( 5, respectively, at ambient temper-
ature. In the current work, kd has been measured as 0.15 (
0.08 s-1 at 24 °C, and the values for k3 and k2, measured
independently, give 0.40 ( 0.08 for the ratio k3/k2, at 24 °C.
Using the values of kobs reported in ref 5 and the values of k2
and k3 determined in this work, a value for kd of 0.15 ( 0.04
s-1 can be calculated, in good agreement with the value
determined in this work. The value of 27 ( 6 kcal/mol, reported
in ref 5, as an approximate activation energy for loss of C2H4
from Fe(CO)3(C2H4)2, agrees, within experimental error, with
the value of 20.7 ( 2.0 kcal/mol determined in this study for
the activation energy.
mode
point
group
symmetryj
compound
CO stretch (cm-1
)
Fe(CO)5a (vapor)
D3h
2118b
A1′
A1′
A1
A1
2045b
2034
A2′′ B1
2013
2087
E′
A1
A1 + B2
Fe(CO)4c (SF6 matrix)
C2V
1999
A1
1994
B1
1974
2000
1984
2095 (s)
2020 (vs)
2003 (vs)
1988 (vs)
2095 (0.11)
2024 (0.74)
2020 (1.00)
B2
Fe(CO)4d (vapor)
C2V
B1(o)e
B2(o)
A1
Fe(CO)4(C2H4)f (vapor) C2V
Fe(CO)4(C2H4)g (vapor) C2V
Fe(CO)4(C2H4)h (vapor) C2V
A1
B1
B2
A1
A1
B1
B. Fe(CO)4-C2H4 Bond Dissociation Energy. There are
two previous reports of the ethylene-Fe bond energy in Fe-
(CO)4(C2H4).29,30 Laser pyrolysis experiments by Lewis and
co-workers give a value of 37.4 kcal/mol.29 Earlier work by
Brown et al., using data from thermal decomposition experi-
ments, puts the BDE at 23.1 kcal/mol30 assuming all five Fe-C
bonds had the same bond dissociation energy of 28.1 kcal/mol.
However, there are a number of more recent determinations
which report bond energies for the Fe(CO)4-CO bond relative
to either the singlet or triplet (ground state) of Fe(CO)4 that are
significantly greater than this average value.29,31-33 Recent
calculations conclude that the energy difference between the
singlet and triplet states of Fe(CO)4 is small.34 As such, the
bond energy for Fe(CO)4-C2H4, recalculated based on a Fe-
(CO)4-CO bond energy of 40 kcal/mol,29,31 is likely to be
encompassed by the range 36.0 ( 2.5 kcal/mol, irrespective of
whether the singlet or triplet state of Fe(CO)4 is taken as a
reference. Larger bond energies for Fe(CO)5 will result in a
larger recalculated energy for the Fe(CO)4-C2H4 bond. Taking
the value of kd of 6 × 10-6 s-1 and any reasonable value for
the preexponential for dissociation of Fe(CO)4(C2H4) leads to
a lower limit for the bond dissociation energy for Fe(CO)4(C2H4)
of less than 36 kcal/mol. Thus, our results are compatible with
both the Fe(CO)4-C2H4 bond dissociation energy reported in
ref 29 and the recomputed value based on work in ref 30.
C. Carbonyl Stretching Frequencies. Table 2 contains
information on the infrared absorptions of Fe(CO)4 and Fe(CO)4-
(C2H4). To compare the two A1 absorptions of Fe(CO)4, which
have been only observed in an SF6 matrix, to the corresponding
absorptions of Fe(CO)4(C2H4), the absorptions reported in the
matrix for Fe(CO)4 were incremented by +8 cm-1, the average
shift between the positions of the two Fe(CO)4 absorptions
which have been observed in both a matrix and the gas phase.
The two lower frequency absorptions of Fe(CO)4 have been
reported in the gas phase and thus they can be directly compared
to gas phase absorptions for Fe(CO)4(C2H4). Comparing the
Fe(CO)4 absorptions with the Fe(CO)4(C2H4) absorptions of
Andrews et al.,21 there is no shift of the highest frequency A1
mode, a +13 cm-1 shift for the other A1 mode, a +3 cm-1
shift for the B1 mode, and a +4 cm-1 shift for the B2 mode in
going from Fe(CO)4 to Fe(CO)4(C2H4). Grant et al. have also
reported gas phase absorptions for Fe(CO)4(C2H4).6 These
absorptions indicate no shift for the highest frequency A1 mode
of Fe(CO)4 relative to Fe(CO)4(C2H4), but shifts of +17 to +20
cm-1 for the three lower frequency modes Fe(CO)4(C2H4)
relative to Fe(CO)4. These shifts represent an interesting
contrast to the Cr(CO)5 system, where on complexation of C2H4
2007 (0.50), 2002 (0.75) B2
2095 (0.1)
2020 (1.0)
A1(o)
A1(o)
∼2000i
a Jones, L. H.; McDowell, R. S. Spectrochim. Acta 1964, 20, 248.
b Calculated frequencies for gas phase A1′ absorptions. c Poliakoff, M.;
Turner, J. J. J. Chem. Soc., Dalton Trans. 1974, 2276. d Ryther, R. J.;
Weitz, E. J. Phys. Chem. 1991, 95, 9841. e (o) indicates the symmetry
label is our assignment. f Andrews, D. C.; Davidson, G. J. Organomet.
Chem. 1972, 35, 161. g Weiller, B. H.; Miller, M. E.; Grant E. R. J.
Am. Chem. Soc. 1987, 109, 352. h This work. i Fe(CO)4(C2H4) and
Fe(CO)3(C2H4)2 absorptions near 2000 cm-1 were not resolved in the
present work. As a result, a relative peak height was not determined.
j All modes symmetries are indicated for the C2V point group except
for Fe(CO)5, where they are indicated for D3h (left), and their correlated
symmetries in the C2V point group are also indicated (right).
withdrawing ligand while when complexed to Cr(CO)5 it is a
net donator of electron density.
A study of Fe(CO)4(C2H4) using photoelectron spectroscopy
found the iron center in Fe(CO)4(C2H4) to be more positiVe than
the iron center in Fe(CO)5.35 The photoelectron study concluded
that C2H4 has σ-donating and π-accepting character that are
either comparable to or better than that of CO. Similar
σ-donating and π-accepting character for CO and C2H4 are
consistent with the similar bond dissociation energies for Fe-
(CO)4-(C2H4) and Fe(CO)4-CO. The shifts in the frequencies
of the modes in Fe(CO)4(C2H4) relative to Fe(CO)4 are
compatible with C2H4 being a net electron-withdrawing ligand
in Fe(CO)4(C2H4), which is further consistent with significant
π-accepting character for C2H4 in Fe(CO)4(C2H4). However,
CO stretching frequencies in Fe(CO)4(C2H4) are shifted to lower
frequency relative to Fe(CO)5.. This is compatible with C2H4
being a better σ donor and/or a poorer π acceptor than CO and
Fe(CO)4(C2H4) having a more negative iron center than Fe-
(CO)5.
The three IR absorptions of Fe(CO)4(C2F4) are shifted 40-
70 cm-1 to higher frequency from the corresponding Fe(CO)4
absorptions, taking into account the effect of the matrix. The
three highest frequency absorptions of Fe(CO)4(C2F4) are shifted
21-40 cm-1 to higher frequency than the corresponding
absorptions of Fe(CO)5 (see Table 2), indicating less electron
density on the metal center when C2F4 replaces CO. Similar
trends are observed for Fe(CO)3(C2F4)2. Thus, from these data,
C2F4 is a poorer σ donor and/or a better π acceptor than CO.
This is consistent with the idea that the electron-withdrawing
nature of the F atoms on C2F4 make it a relatively poor electron
donor and a relatively good π acceptor.
the E mode absorptions shift to lower frequency by ∼10 cm-1
.
D. Comparison of Bond Dissociation Energies. Compari-
sons of the bond dissociation energy for Fe-C2H4 bonds with
Thus, when complexed to Fe(CO)4 ethylene is a net electron-

2994 J. Phys. Chem. A, Vol. 101, No. 16, 1997
House and Weitz
Cr-C2H4 bonds support accepted ideas regarding metal-olefin
bonding. As discussed, Fe(CO)4(C2H4) has a bond dissociation
energy that is expected to be in the range 36.0 ( 2.5 kcal/mol.
The corresponding chromium compound, Cr(CO)5(C2H4), has
been reported to have a BDE of 24.7 ( 2.4 kcal/mol.12 The
BDE for Fe(CO)3(C2H4)2 is reported in this work as 21.3 (
2.0 kcal/mol, and the BDE for cis-Cr(CO)4(C2H4)2 has been
reported as ∼15 kcal/mol.36 Two trends are evident: (1) the
monosubstituted compounds are more stable than the disubsti-
tuted compounds, and (2) the iron compounds are more stable
than the corresponding chromium compounds.
additional electrons, forms stronger bonds with ethylene and
perfluoroethylene than chromium.
Comparing the bonding of ethylene and perfluoroethylene
to chromium and iron highlights the differences between the
two metals. Previous work in this laboratory determined that
the bond dissociation energy for Cr(CO)5-C2H4 is ∼5 kcal/
mol larger than Cr(CO)5-C2F4.12 Although we were not able
to make quantitative measurements on the C2F4 system, it
appears that C2F4 in Fe(CO)3(C2F4)2 is more strongly bound
than C2H4 in Fe(CO)3(C2H4)2.
The rate constant for dissociative loss of C2F4 from Fe(CO)3-
(C2F4)2 is no more than 1/25 000 that of kd for loss of C2H4
from Fe(CO)3(C2H4)2. This difference in rate constants could,
in principle, be due to either a difference in preexponential
factors, activation energies, or a combination of both factors. It
seems implausible that the entire difference in rate constants
could be due to the preexponential since the reverse reactions
are unactivated with rate constants that differ by only a factor
of 2. Thus, for the entire difference in kd’s to be due to the
pre-exponentials, the ratio of preexponentials for the dissociation
would have to differ from the ratio of preexponentials for the
reverse process by over 12 000. This seems unlikely for
reactions that, a priori, would be expected to have similar shaped
transition states. If this difference in kd’s is solely due to the
activation energies, the bond energies for the two systems differ
by at least ∼6 kcal/mol.
To make these comparisons, it must be justified that it is more
appropriate to compare the bond dissociation energy for loss
of C2H4 from cis-Cr(CO)4(C2H4)2 than from trans-Cr(CO)4-
(C2H4)2 to that for Fe(CO)3(C2H4)2. There are large differences
in the stability of cis and trans isomers of the bis-olefin-
substituted group VI metal carbonyls that have been explained
as being due to a competition between CO and olefin π* orbitals
for electron density from the same metal orbital.37,38 The
explanation focuses on the fact that olefins are single-faced π
acceptors, i.e., they have a single π* orbital to accept electron
density, while CO has two orthogonal π* orbitals. In the trans
bis-olefin group VI metal carbonyls, the olefins orient them-
selves 90° to each other, and because each has only one
π-accepting orbital, the orbitals are orthogonal and do not back-
bond with the same metal orbital.37,38 A diagram illustrating
this is in ref 38. With an olefin trans to CO, independent of
how the olefin is oriented, the CO will have one of its
π-accepting orbitals in the same plane as the olefin’s π-accepting
orbital, leading to competition for electron density from the same
metal orbital. In Fe(CO)3(C2H4)2 the ethylene ligands occupy
two of the equatorial sites and compete with each other and a
CO group, in the other equatorial site, for electron density from
the same metal orbitals. cis-Cr(CO)4(C2H4)2 has carbonyl and
olefin groups competing for electron density from the same
metal orbitals, while in the trans isomer the ethylenes do not
compete with carbonyl groups, or each other, for electron density
from a common metal orbital. For this reason, the bond
dissociation energy for loss of C2H4 from cis-Cr(CO)4(C2H4)2
has been used for comparison purposes with Fe(CO)3(C2H4)2.
The stronger perfluoroethylene-iron bond in Fe(CO)3(C2F4)2
relative to the ethylene-iron bond in Fe(CO)3(C2H4)2 is
consistent with a similar trend found in a series of Fe(CO)4-
(haloolefin) compounds.23 PRDDO calculations predicted C2H4
to be slightly (∼1 kcal/mol) more strongly bound to Fe(CO)4
than C2F4.42 However, as previously indicated, our results do
not allow us to determine whether the Fe(CO)4-C2H4 or Fe-
(CO)4-C2F4 BDE is larger. Nevertheless, only a 1 kcal/mol
difference in bond energies is much closer in energy than for
the corresponding Cr compounds.
As we have indicated in section III.B, we feel it is highly
likely that the second compound we observe is Fe(CO)2(C2F4)2
and have discussed this species in these terms. However, since
we cannot rule out the possibility that the second compound is
actually a more highly substituted perfluoroethylene species, it
seems warranted to indicate once again, that if the second
compound we observe is actually an Fe(CO)x(C2F4)5-x (x e 2)
species, then its stability is all the more remarkable.
It was previously concluded, based on bond dissociation
energies and CO stretching frequencies, that for Cr(CO)5L
compounds, C2F4 is a poorer σ donor and also may be a better
π acceptor than C2H4.12 With Cr(CO)5, the better σ donor ability
of C2H4 appears to be the dominant factor in determining which
of the olefins is more strongly bound. In the iron carbonyls,
because of the increased electron density on iron compared to
chromium, the π-accepting ability of C2F4 appears to be more
important than the relatively greater σ-donating ability of C2H4.
This competition for electron density has also been invoked
to explain the observed decrease in lifetime of iron carbonyl
complexes as the number of olefins increases. There is a limited
back-bonding capability of the metal, with more olefins drawing
electron density from the same metal orbitals.24,39 The olefins
occupy in-plane equatorial sites when substituted in iron
carbonyl, and as a result they compete with each other for
electron density from the same metal orbitals.
Experimental results are consistent with this explanation. A
number of stable Fe(CO)4(olefin) compounds are known,40 but
only two stable Fe(CO)3(olefin)2 compounds39 and one stable
Fe(CO)2(olefin)3 compound, Fe(CO)2((E)-cyclooctene)3, has
been reported.24 As discussed, previous work has also shown
that Cr(CO)5(C2H4) is more stable than cis-Cr(CO)4(C2H4)2.36
In fact, Cr(CO)5(C2H4) has been isolated as a stable solid after
UV photolysis of Cr(CO)6 in supercritical ethylene.41 Also,
M(CO)5(C2H4) has been observed to be more stable than cis-
M(CO)4(C2H4)2 (M ) Cr, Mo, W) in alkane solutions.37
The greater stability of the ethylene-substituted iron carbonyls
compared to the ethylene-substituted chromium carbonyls is
consistent with established ideas of metal-olefin bonding.
Metals possessing more bonding electrons have the capability
to form a stronger π back-bond and thus more stable complexes
with π acceptors.1 Thus, it is not surprising that iron, with two
V. Conclusions
The bimolecular rate constants for the processes CO + Fe-
(CO)3(C2H4), C2H4 + Fe(CO)3(C2H4), Fe(CO)3(C2H4) + Fe-
(CO)5, and Fe(CO)4 + C2H4 have been measured as (4.3 (
0.8) × 10-12, (11.1 ( 0.8) × 10-12, (4 ( 2) × 10-11, and (1.7
( 0.2) × 10-13 cm3 /( molecule s), respectively. All are
temperature independent from 24 °C to at least 45 °C. The
rate constant for the reaction of Fe(CO)3(C2H4) + C2H4 agrees
well with a previous determination.7
Infrared absorptions observed at 2147, 2091, and 2068 cm-1
belong to a compound best assigned as Fe(CO)3(C2F4)2, which

Kinetics of Fe(CO)4L and Fe(CO)3L2
J. Phys. Chem. A, Vol. 101, No. 16, 1997 2995
(6) Weiller, B. H.; Miller, M. E.; Grant, E. R. J. Am. Chem. Soc. 1987,
109, 352.
is expected to be an analog of Fe(CO)3(C2H4)2. Infrared
absorptions observed at 2135, 2074, and 2043 cm-1 are assigned
to Fe(CO)4(C2F4). The rate constant for addition of C2F4 to
Fe(CO)3 is (3.3 ( 1.2) × 10-11 cm3/(molecule s) at (24 ( 1)
°C. The rate constants for addition of C2F4 to Fe(CO)3(C2F4)
and Fe(CO)4 were determined to be (5.4 ( 1.7) × 10-12 and
(1.8 ( 0.4) × 10-14 cm 3/(molecule s) respectively at (24 ( 1)
°C.
(7) Hayes, D. M.; Weitz, E. J. Phys. Chem. 1991, 95, 2723.
(8) Wells, J. R.; Weitz, E. J. Phys. Chem. 1993, 97, 3084.
(9) Dewar, M. J. S. Bull. Soc. Chim. Fr. 1951, 18, C79.
(10) Chatt, J.; Duncanson, L. A. J. Chem. Soc. 1953, 2939.
(11) Fields, R.; Germain, M. M.; Haszeldine, R. N.; Wiggans, P. W. J.
Chem. Soc. A 1970, 1964.
(12) Wells, J. R.; House, P. G.; Weitz, E. J. Phys. Chem. 1994, 98,
8343.
(13) Weitz, E. J. Phys. Chem. 1987, 91, 3945.
(14) Ryther, R. J.; Weitz, E. J. Phys. Chem. 1992, 96, 2561.
(15) Long, G. T. Northwestern University, unpublished results.
(16) Bogdan, P. L.; Wells, J. R.; Weitz, E. J. Am. Chem. Soc. 1991,
113, 1294.
Two prior determinations29,30 of the bond dissociation energy
for the loss of C2H4 from Fe(CO)4(C2H4) are compared with
current experimental results.
The activation energy for loss of C2H4 from Fe(CO)3(C2H4)2
is in agreement, within experimental error, with a previous
determination5 and leads to a bond dissociation energy of 21.3
( 2.0 kcal/mol. This Fe(CO)3(C2H4)-C2H4 bond energy is
compared to previously determined bond dissociation energies
for similar metal-olefin bonds. With the same metal center,
the decreased stability of bisethylene-substituted species as
compared to monoethylene-substituted compounds has been
ascribed to an increase in olefin competition for metal electron
density from common d orbitals.24,37-39 With the same number
of ethylene ligands, the iron compounds are more stable than
the chromium compounds due to iron being more electron rich
and therefore having more electron density available for back-
bonding. There appears to be a reversal in the order of stability
of the disubstituted compounds of C2H4 and C2F4 on going from
chromium to iron, again explainable based on the relative
electron density of the two metals. Though our data do not
allow us to definitively reach a similar conclusion regarding
the monosubstituted (C2H4 and C2F4) species, the data are not
inconsistent with such a conclusion. Interestingly, the shifts in
the three lowest frequency CO modes in Fe(CO)4(C2H4) relative
to those in Fe(CO)4 indicate that C2H4 is a net electron-
withdrawing ligand while when C2H4 is bound to Cr(CO)5 the
shifts in the E mode CO frequencies12 indicate that C2H4 is a
net electron donor.
(17) Hoehn, H. H.; Pratt, L.; Watterson, K. F.; Wilkinson, G. J. Chem.
Soc. 1961, 2738.
(18) Fields, R.; Germain, M. M.; Haszeldine, R. N.; Wiggans, P. W. J.
Chem. Soc. A 1970, 1969.
(19) Beagley, B.; Schmidling, D. G.; Cruickshank, D. W. J. Acta
Crystallogr. 1973, B29, 1499.
(20) Karel, K. J.; Tulip, T. H.; Ittel, S. D. Organometallics 1990, 9,
1276.
(21) Andrews, D. C.; Davidson, G. J. Organomet. Chem. 1972, 35, 161.
(22) Murdoch, H. D.; Weiss, E. HelV. Chim. Acta 1963, 46, 1588.
(23) Fields, R.; Godwin, G. L.; Haszeldine, R. N. J. Chem. Soc., Dalton
Trans. 1975, 1867.
(24) Angermund, H.; Bandyopadhyay, A. K.; Grevels, F. W.; Mark, F.
J. Am. Chem. Soc. 1989, 111, 4656.
(25) Weiller, B. H.; Grant, E. R. J. Am. Chem. Soc. 1987, 109, 1051
(26) Wuu, Y. M.; Bentsen, J. G.; Brinkley, C. G.; Wrighton, M. S. Inorg.
Chem. 1987, 26, 530.
(27) Handbook of Chemistry and Physics, 77th ed.; Lide, D. R., Ed.;
CRC Press: Boca Raton, FL. Zmbov, K. F.; Uy, O. M.; Margrave, J. L. J.
Am. Chem. Soc. 1968, 90, 5090.
(28) Seder, T. A.; Ouderkirk, A. J.; Weitz, E. J. Chem. Phys. 1986, 85,
1977.
(29) Lewis, K. E.; Golden, D. M.; Smith, G. P. J. Am. Chem. Soc. 1984,
106, 3905.
(30) Brown, D. L. S.; Conner, J. A.; Leung, M. L.; Paz-Andrade, M. I.;
Skinner, H. A J. Organomet. Chem. 1976, 110, 79.
(31) Shen, J.; Gao, Y.; Shi. Q.; Basolo, F. Inorg. Chem. 1989, 28, 4304.
(32) Engelking, P. C.; Lineberger, W. C. J. Am. Chem. Soc. 1979, 101,
5569.
(33) Majima, T.; Matsumoto, Y.; Takami, M. J. Photochem. Photobiol.,
A 1993, 71, 213.
(34) Lyne, P. D.; Mingos, D. M. P.; Ziegler, T.; Downs, A. J. Inorg.
Chem. 1993, 32, 4785.
(35) Beach, D. B.; Jolly, W. L. Inorg. Chem. 1983, 22, 2137.
(36) Gregory, M. F.; Jackson, S. A.; Poliakoff, M.; Turner, J. J. J. Chem.
Soc., Chem. Commun. 1986, 1175.
(37) Grevels, F. W.; Jacke, J.; O¨ zkar,S. J. Am. Chem. Soc. 1987, 109,
7536.
Acknowledgment. We acknowledge support of this work
by the National Science Foundation under Grant CHE90-24509.
We thank Prof. C. Mirkin for useful discussions.
References and Notes
(38) Grevels, F. W.; Jurgen, J.; Kotzbu¨cher, W. E.; O¨ zkar, S.; Skibbe,
V. Pure Appl. Chem. 1988, 60, 1017.
(1) Pruchnik, F. P. Organometallic Chemistry of the Transition
Elements (English translation by S. A. Duraj); Plenum Press: New York,
1990; Chapter 6.
(2) Wrighton, M. S.; Hammond, G. S.; Gray, H. B. J. Organomet.
Chem. 1974, 70, 283.
(3) Geoffroy, G. L.; Wrighton, M. S. Organometallic Photochemistry;
Academic: New York, 1979.
(39) Angermund, H.; Grevels, F. W.; Moser, R.; Benn, R.; Kru¨eger,
C.; Roma˜o, M. J. Organometallics 1988, 7, 1994.
(40) King, R. B. In The Organic Chemistry of Iron; Koerner Von
Gustorf, E. A.; Grevels, F. W.; Fischler, I., Eds.; Academic: New York,
1978: Vol. 1, pp 397-462.
(41) Banister, J. A.; Howdle, S. M.; Poliakoff, M. J. Chem. Soc., Chem.
Commun. 1993, 1814.
(42) Axe, F. U.; Marynick, D. S. J. Am. Chem. Soc. 1984, 106, 6230.
(4) Chalk, A. J.; Harrod, J. F. J. Am. Chem. Soc. 1967, 89, 1640.
(5) Miller, M. E.; Grant, E. R. J. Am. Chem. Soc. 1987, 109, 7951.