Formation and reactions of tetracarbonyl intermediates of the Fischer carbene complex (CO)5WqqC(OMe)Ph. A laser flash photolysis study using time-resolved infrared and UV/vis spectroscopy

Article abstract of DOI:10.1021/om990708q

The results of nanosecond laser flash photolysis (XeCl excimer irradiation at 308 nm) of the Fischer carbene complex [(CO)₅WqqC(OMe)Ph] (C_Z) in di-n-butyl ether and n-hexane are reported. Spectrokinetical detection of the intermediates was based on transient absorptions simultaneously recorded in the UV/vis and the IR spectral regions. The primary photoproducts are the unstable isomer C_E produced with a quantum yield of about 30% and a tetracarbonyl complex I-S, probably of structure [(CO)₄(S)WqqC(OMe)Ph], where S is a solvent molecule. The quantum yield of CO photoelimination was estimated as 1.7%. The C_E isomer was found to return to the stable C_Z isomer within a few microseconds, confirming previous results of McGarvey and co-workers. The tetracarbonyl species reacts with cosolutes present in the solution such as N₂, H₂O, acetonitrile, and methyl trans-crotonate to form secondary complexes of cis-tetracarbonyl complex structure [(CO)₄(L)WqqC(OMe)Ph] (I-L). The tetracarbonyl complexes react readily with CO. For I-N₂ in n-hexane the rate constant is 3.5 × 10⁸ M^-1 s^-1. The rate constants of the corresponding reactions with acetonitrile (MeCN) and the olefin are larger by a factor of 2. In the presence of CO the complexes I-MeCN and I-olefin revert to C_Z in a dissociative reaction mechanism. The rate constants for the elimination of MeCN or olefin are 5 and 10⁵ s^-1, respectively. The I-olefin complex irreversibly decomposes with a rate constant of 8 × 10³ s^-1.

Full text of DOI:10.1021/om990708q

2354
Organometallics 2000, 19, 2354-2364
F or m a tion a n d Rea ction s of Tetr a ca r bon yl
In ter m ed ia tes of th e F isch er Ca r ben e Com p lex
(CO)5WdC(OMe)P h . A La ser F la sh P h otolysis Stu d y
Usin g Tim e-Resolved In fr a r ed a n d UV/Vis Sp ectr oscop y†
Hans-Peter Gut, Nicolas Welte, Uwe Link, Helmut Fischer, and
Ulrich E. Steiner*
Fakulta¨t fu¨r Chemie, Universita¨t Konstanz, 78457 Konstanz, Germany
Received September 7, 1999
The results of nanosecond laser flash photolysis (XeCl excimer irradiation at 308 nm) of
the Fischer carbene complex [(CO)5WdC(OMe)Ph] (CZ) in di-n-butyl ether and n-hexane
are reported. Spectrokinetical detection of the intermediates was based on transient
absorptions simulaneously recorded in the UV/vis and the IR spectral regions. The primary
photoproducts are the unstable isomer CE produced with a quantum yield of about 30% and
a tetracarbonyl complex I-S, probably of structure [(CO)4(S)WdC(OMe)Ph], where S is a
solvent molecule. The quantum yield of CO photoelimination was estimated as 1.7%. The
CE isomer was found to return to the stable CZ isomer within a few microseconds, confirming
previous results of McGarvey and co-workers. The tetracarbonyl species reacts with cosolutes
present in the solution such as N2, H2O, acetonitrile, and methyl trans-crotonate to form
secondary complexes of cis-tetracarbonyl complex structure [(CO)4(L)WdC(OMe)Ph] (I-L).
The tetracarbonyl complexes react readily with CO. For I-N2 in n-hexane the rate constant
is 3.5 × 108 M-1 s-1. The rate constants of the corresponding reactions with acetonitrile
(MeCN) and the olefin are larger by a factor of 2. In the presence of CO the complexes
I-MeCN and I-olefin revert to CZ in a dissociative reaction mechanism. The rate constants
for the elimination of MeCN or olefin are 5 and 105 s-1, respectively. The I-olefin complex
irreversibly decomposes with a rate constant of 8 × 103 s-1
.
In tr od u ction
In contrast, the primary reaction step in the photo-
chemical reactions of tungsten carbene complexes
[(CO)5WdC(XR)R′] seems to be the photoinduced loss
of a CO ligand.5,6 Subsequent addition of nucleophiles
to the resulting coordinatively unsaturated species
affords substitution products cis-[(CO)4(L)WdC(XR)R′]
(L ) PR3,7 MeCN,8 THF,8 or alkynes8). The quantum
yield of CO loss from [(CO)5WdC(XR)R′] is fairly low
(approximately 1% at 366 and 313 nm) and drops to less
than 10-4 at 436 nm.8,9 This has been explained by the
inactivity in photodissociation of the long-wavelength
While thermal reactions of transition-metal carbene
complexes are widely employed in synthetic chemistry,1
the synthetic potential and, in particular, the mecha-
nisms of their photochemical reactions are much less
explored. Only recently2 McGuire and Hegedus intro-
duced photochemical reactions of chromium carbene
complexes into organic synthesis. Photolysis of [(CO)5Crd
C(XR)R′] was proposed3 to proceed by a thermorevers-
ible intramolecular coupling of a CO and the carbene
ligand to form chromium ketene complexes, which are
then trapped by, for example, imines, olefins, aldehydes
and ketones, or alcohols to afford â-lactams, cyclobu-
tanones, â-lactones, and amino acid esters, respectively.4
† Dedicated to Professor J u¨rgen Felsche on the occasion of his 60th
birthday.
* To whom correspondence should be addressed. Tel: +49-7531-
883570. Fax: +49-7531-883014. E-mail: ulrich.steiner@uni-konstan-
z.de.
(1) See for example: (a) Do¨tz, K. H. In Transition Metal Carbene
Complexes; Do¨tz, K. H., Fischer, H., Hofmann, P., Kreissl, F. R.,
Schubert, U., Weiss, K., Eds.; Verlag Chemie: Weinheim, 1983; p 191.
(b) Do¨tz, K. H. Angew. Chem. 1984, 96, 573; Angew. Chem., Int. Ed.
Engl. 1984, 23, 587. (c) Aumann, R. Angew. Chem. 1988, 100, 1512;
Angew. Chem., Int. Ed. Engl. 1988, 27, 1456. (d) Wulff, W. D. In
Comprehensive Organic Synthesis; Trost, B. M., Fleming, J ., Eds.;
Pergamon Press: Oxford, 1991; Vol. 5, p 1065. (e) Wulff, W. D. In
Comprehensive Organometallic Chemistry II; Abel, E. W., Stone, F.
G. A., Wikinson, G., Eds.; Pergamon Press: Oxford, 1995; Vol. 12, p
469.
(4) For recent reviews see: (a) Hegedus L. S. In Comprehensive
Organometallic Chemistry II; Abel, E. W., Stone, F. G. A., Wikinson,
G., Eds.; Pergamon Press: Oxford; 1995; Vol. 12, p 549. (b) Hegedus,
L. S. Tetrahedron 1997, 53, 4105.
(5) See for example: (a) Casey, C. P.; Vollendorf, N. W.; Haller, K.
J . J . Am. Chem. Soc. 1984, 106, 3754. (b) Casey, C. P.; Hornung, N.
L.; Kosar, W. P. J . Am. Chem. Soc. 1987, 109, 4908. (c) Fong, L. K.;
Cooper, J . N. J . Am. Chem. Soc. 1984, 106, 2595. (d) Toledano, C. A.;
Rudler, H.; Daran, J .-C.; J eannin, Y. J . Chem. Soc., Chem. Commun.
1984, 574. (e) Nandi, M.; Sathe, K. M.; Sarkar, A. J . Chem. Soc., Chem.
Commun. 1992, 793. (f) Choi, Y. H.; Rhee, K. S.; Kim, K. S.; Shin, G.
C.; Shin, S. C. Tetrahedron Lett. 1995, 36, 1871.
(6) The photolysis of [(CO)5WdC(OMe)R′] in the presence of cis-RNd
NR (R ) C6H4X-p) proceeds by a different pathway. The proposed
mechanism involves initial addition of the azo compound to the carbene
carbon to give an adduct that rapidly rearranges to form an isolable
zwitterion (CO)5W-[NR-NR+dC(OMe)R′]. Photolysis of the zwitterion
finally affords diaziridines and imino ethers. See: Maxey, C. T.;
Sleiman, H. F.; Massey, S. T.; McElwee-White, L. J . Am. Chem. Soc.
1992, 114, 5153, and literature cited therein.
(2) McGuire M. A.; Hegedus L. S. J . Am. Chem. Soc. 1982, 104, 5538.
(3) (a) Merlic, C. A.; Xu, D. J . Am. Chem. Soc. 1991, 113, 7418. (b)
Merlic, C. A.; Xu, D.; Gladstone, B. G. J . Org. Chem. 1993, 58, 538.
(7) Fischer, E. O.; Fischer, H. Chem. Ber. 1974, 107, 657.
(8) Foley, H. C.; Strubinger, L. M.; Targos, T. S.; Geoffroy, G. L. J .
Am. Chem. Soc. 1983, 105, 3064.
10.1021/om990708q CCC: $19.00 © 2000 American Chemical Society
Publication on Web 05/16/2000

Tetracarbonyl Intermediates
Organometallics, Vol. 19, No. 12, 2000 2355
metal-to-ligand charge transfer (MLCT) transition and
by the only small activity of the ligand-field (LF)
excitation due to its rapid radiationless decay to the
lower MLCT state. Such small quantum yields may be
a major drawback that can limit possible synthetic
applications.
Many important thermal reactions of pentacarbonyl
carbene complexes such as the Do¨tz reaction,1,10 the
transfer of the carbene ligand of heteroatom-stabilized
carbene complexes to acceptor-substituted olefins1,11 or
dienes,12 and the formation of carbo- and heterocycles
from carbene complexes and enynes13 are initiated by
loss of a CO ligand. Since in these reactions CO
dissociation is rate-limiting, the succeeding product-
developing steps are not accessible to conventional
kinetic investigations. Therefore, the photoinduced ac-
cess to such CO dissociation products should open up
the possibility of detailed studies of these product-
developing reaction steps using kinetic laser flash
photolysis methods.
Transients generated by photolysis of [(CO)5Wd
C(XR)R′] have already been investigated spectroscopi-
cally by the groups of McGarvey14-16 and Stufkens.17
At room temperature, photolysis of [(CO)5WdC(OCH3)-
Ph] in solution yielded a species with a lifetime of
several microseconds. The transient was initially as-
signed a tetracarbonyl carbene structure.14,15 It was
suggested that the vacant coordination site of the
intermediate is blocked by an agostic interaction with
a C-H bond of the methoxy substituent.
Continued irradiation of the matrix led to CO loss and
to a tetracarbonyl carbene complex with a syn confor-
mation of the carbene ligand. Further irradiation of the
tungsten complex gave a species “X”, which was tenta-
tively assigned by Stufkens and co-workers to the
agostic tetracarbonyl species previously invoked by
Bechara et al.14 Recently, several [(CO)5WdC(OR)R′]
complexes with (R, R′) ) (Me, Me), (Me, p-tolyl), and
(Et, Ph) were studied in a variety of solvents.15 The
formation of the syn conformer in its ground state was
observed during the laser flash of several nanoseconds.
The subsequent thermal regeneration of the anti con-
former was found to occur within a few microseconds
to several thousands of microseconds depending on the
substituents, solvent, and temperature.
We now report results obtained from photolysis of
[(CO)5WdC(OCH3)Ph] using a newly set up laser flash
spectrometer allowing for simultaneous time-resolved
transient detection in the UV/vis and in the 1700-2200
cm-1 infrared region, from which the most significant
information on the CO coordination can be obtained.
Resu lts
As a general demonstration of the spectrokinetical
features observed on laser flash photolysis of complex
CZ and of the possibility of parallel recording of tran-
sient UV/vis and IR signals, characteristic kinetic traces
at different wavelengths in di-n-butyl ether are depicted
in Figure 1. Measurements with 100 µs, 10 ms, and 1 s
full scale have been combined to yield signal traces
spanning 5 orders of magnitude in time. In the UV/vis
spectral range the main changes are observed in the
region corresponding to the MLCT band of complex CZ.
Therefore, positive as well as negative ∆A steps can be
observed. The rise time of these signals is on the order
of 20 ns and corresponds to the width of the laser pulse.
The subsequent decay of the signals indicates two
different transformations: with time constants of about
10 µs and about 0.1 s. After several seconds the signals
have decayed essentially to zero, establishing the overall
reversibility of the photoinduced processes.
On the basis of the results of matrix isolation studies
with [(CO)5{W,Cr}dC(OCH3)Ph] the primary photore-
action was assigned to an anti/syn isomerization by
Servaas et al.17
In the ν˜(CO) range of the IR spectrum (cf. Figure 2),
too, positive and negative initial signal steps are ob-
served (Figure 1b). The fastest initial signal rise is
determined by the time constant of about 1 µs of the
detection system (cf. signals at 1938 and 1957 cm-1),
but there are also frequencies where the rise of the
signal is clearly slower than the instrumental time
resolution. The rise of the signal at 1891 cm-1 agrees
well with the first time-resolved kinetic step in the UV/
vis signals showing a time constant of about 10 µs. As
with the UV/vis measurements, the slow final decay of
the transients is also borne out in the IR. Thus, the
signals in Figures 1a and 1b clearly demonstrate that
the transient kinetics observed in both spectral regions
monitor the same chemical intermediates.
An equilibrium between anti and syn conformers in
solution had already been reported before by Kreiter and
Fischer18 on the basis of NMR spectroscopic studies.
(9) Pourreau, D. B.; Geoffroy, G. L. Adv. Organomet. Chem. 1985,
24, 249.
(10) Do¨tz, K. H. Angew. Chem. 1975, 87, 672; Angew. Chem., Int.
Ed. Engl. 1975, 14, 644.
(11) (a) Fischer, E. O.; Do¨tz, K. H. Chem. Ber. 1970, 103, 1273. (b)
Do¨tz, K. H.; Fischer, E. O. Chem. Ber. 1972, 105, 1356. (c) Cooke, M.
D.; Fischer, E. O. J . Organomet. Chem. 1973, 56, 279. (d) Wienand,
A.; Reissig, H.-U. Organometallics 1990, 9, 3133, and literature cited
therein.
(12) (a) Wulff, W. D.; Yang, D. C.; Murray, C. K. J . Am. Chem. Soc.
1988, 110, 2653. (b) Wienand A.; Reissig, H.-U. Tetrahedron Lett. 1988,
29, 2315. (c) Harvey, D. F.; Lund, K. P. J . Am. Chem. Soc., 1991, 113,
8916. (d) Hoffmann, M.; Buchert, M.; Reissig, H.-U. Chem. Eur. J .
1999, 5, 876, and references therein.
(13) (a) Korkowski, P. F.; Hoye, T. R.; Rydberg, D. B. J . Am. Chem.
Soc. 1988, 110, 2676. (b) Harvey, D. F.; Lund, K. P. J . Am. Chem. Soc.
1991, 113, 5066. (c) Recent review: Harvey, D.; Sigano, D. M. Chem.
Rev. 1996, 96, 271.
(14) Bechara, J . N.; Bell, S. E. J .; McGarvey, J . J .; Rooney, J . J . J .
Chem. Soc., Chem. Commun. 1986, 1785.
The spectral characteristics of the transients in di-
n-butyl ether and n-hexane are similar. The results of
the experiments differ mainly in the kinetics of the
intermolecular reactions with CO and other solutes
added as reactants. In n-hexane interaction of the
(15) Bell, S. E. J .; Gordon, K. C.; McGarvey, J . J . J . Am. Chem. Soc.
1988, 110, 3107.
(16) Rooney, A. D.; McGarvey, J . J .; Gordon, K. C. Organometallics
1995, 14, 107.
(17) Servaas, P. C.; Stufkens, D. J .; Oskam, A. J . Organomet. Chem.
1990, 390, 61.
(18) (a) Kreiter, C. G.; Fischer, E. O. Angew. Chem. 1969, 81, 780;
Angew. Chem., Int. Ed. Engl. 1969, 8, 761. (b) Kreiter, C. G.; Fischer,
E. O. Pure Appl. Chem. 1971, 6, 151.

2356 Organometallics, Vol. 19, No. 12, 2000
Gut et al.
F igu r e 2. IR spectra of CZ in n-hexane (c ) 2 × 10-3 M,
d ) 0.05 cm). Solid line: measured on a FTIR spectrometer
with 2 cm-1 spectral resolution. Dashed line: spectrum
calculated from the FTIR spectrum by simulating a trian-
gular slit function of 14 cm-1 halfwidth as determined by
the slit width setting of the IR monochromator used in the
time-resolved measurements.
F igu r e 1. Transient absorption signals obtained in a 0.5
mm cuvette upon laser photolysis of CZ (2 × 10-3 M) in
N2-saturated di-n-butyl ether at selected wavelengths in
the UV/vis (a) and IR (b). Each signal trace spanning 5
orders of magnitude in time was amalgamated from three
individual signals measured separately in three different
time domains. To avoid distortions of the initial stage of
the signal due to the log scale, the zero time of the signal
was shifted to t ) 10 µs.
F igu r e 3. UV/vis transient absorption spectra obtained
on laser photolysis of CZ (2 × 10-3 M) in n-hexane. b, +
(left ∆A axis) at a delay time of 1 µs after the laser pulse:
(b) solution saturated with N2, (+) solution saturated with
CO. 0 (right ∆A axis) at a delay time of 40 µs in
N2-saturated solution. (In a CO-saturated solution all
transient absorptions have completely decayed by 40 µs.)
solvent with coordinatively unsaturated intermediates
is weaker, and thus reactions of these intermediates
with added solutes are faster than in di-n-butyl ether.
Therefore, we will focus on the experiments in n-hexane.
According to the general spectrokinetical picture in
di-n-butyl ether (Figure 1), the primary intermediates
revert to the starting compound and/or transform to
secondary intermediates within less than 20 µs. Thus,
recording transient spectra 1 and 40 µs after the laser
pulse should provide the spectral characteristics of the
primary and secondary intermediates, respectively.
Corresponding transient spectra are shown in Figures
3 and 4.
At 1 µs delay, the striking feature of both UV/vis and
IR spectra is a derivative-type absorption band centered
with its inflection point at the wavelength of the main
absorption maximum of the starting complex CZ, i.e.,
at 402 nm in the UV and at 1948 cm-1 in the IR (Figure
2). Such spectral features indicate a transformation of
the starting complex into a transient with an absorption
spectrum only slightly shifted with respect to the former
one. In both the UV/vis and the IR region the phase
(sign sequence) of the ∆A curve signifies a shift of the
transient bands to higher frequencies.
Apart from the dominant derivative-type character-
istics at 1948 cm-1 the 1 µs transient spectrum exhibits
several other noteworthy features in the IR:
(i) There are positive ∆A changes, i.e., new absorp-
tions around 2045 cm-1 and in the region between 1850
and 1910 cm-1, all of which are clearly outside the
region of IR lines of the precursor complex CZ (cf. Figure
2).
(ii) Around 2070 cm-1, i.e., at the position of the weak
A1(1) absorption of CZ, the transient spectrum seems to
indicate a superposition of a derivative-type and a
bleaching feature. However, due to the weakness of this
signal, details cannot be assessed with certainty.
From these spectral features we conclude that at least
two early transients are formed in the photolysis of CZ.
One whose UV/vis and IR spectra are only slightly
shifted with respect to CZ gives rise to the derivative-
type band. This transient is assigned to an isomer of

Tetracarbonyl Intermediates
Organometallics, Vol. 19, No. 12, 2000 2357
respectively. For the frequencies of the IR lines assigned
to I-N2 and I-H2O see Table 1.
All observations are consistent with an assignment
of the initial transient with the derivative-type ∆A
spectrum as the E isomer (CE) of the pentacarbonyl-
carbene complex CZ. According to matrix isolation
experiments by Servaas et al.17 in the IR the E isomer
shows only very small (1-3 cm-1) high-frequency shifts
relative to the Z isomer. The ∆A spectrum deduced from
such a shift is perfectly in line with the derivative-type
band shape in the initially observed transient IR
absorption spectrum. Our pertinent experiments afford
the first time-resolved infrared detection of the photo-
induced CZ f CE process at room temperature. Some
flash spectroscopic investigations with UV/vis transient
absorption measurements of complex CZ and related
compounds have already been reported by Rooney et
al.16,19 Our UV/vis transient spectrum at 1 µs delay (cf.
Figure 3) is consistent with their spectra, although their
observations were restricted to a spectral range between
370 and 470 nm. According to the latter authors, the
decay time of CE at 20 °C is 2.9 µs in n-hexane and 25
µs in acetonitrile. Our experiments at somewhat higher
temperature of approximately 25 °C yielded slightly
shorter values of 1.9 µs for n-hexane and 14.8 µs for
acetonitrile. These deviations seem reasonable in view
of the activation energy assessed for the E f Z process.
The decay times exhibit the same solvent dependence
as observed by McGarvey’s group.
F igu r e 4. IR transient spectra obtained on laser photolysis
of CZ (2 × 10-3 M) in n-hexane. b, + at 1 µs delay time,
(b) with N2-saturated solution, (+) with CO-saturated
solution. 0 at 40 µs delay time with N2-saturated solution.
complex CZ. The other one (denoted I-S for reasons to
be given below) has IR bands clearly distinct from those
of CZ. The IR band positions of I-S are listed in Table
1.
After 40 µs the spectral intensity in the UV/vis has
strongly decreased. Instead of the derivative-type band
with the inflection point at the absorption maximum of
CZ, we now see a bleaching at this position and positive
absorption changes to the right and left of it. This
implies that the new transient has an absorption band
with a maximum at the same position as CZ, although
with a somewhat lower molar extinction coefficient. The
absorption is broader and extends to much longer
wavelengths. The decrease in intensity most probably
implies that a great fraction of the primary intermedi-
ates has reverted to the initial complex CZ.
In the IR the ∆A spectrum after 40 µs clearly
indicates bleaching at the positions of the main bands
of CZ (2070, 1950 cm-1) and positive absorption changes
due to new absorptions at the low-frequency sides of
these bands.
It seems likely that the long-lived transient is formed
in an intermolecular process between the primary
transient and the solvent or some other solute. Since
in all experiments described so far N2 was the only other
solute, its possible involvement was tested by substitut-
ing argon for nitrogen as protecting gas. This did not
cause any change of the transient spectra at delay times
as short as 1 µs, but it suppressed the formation of the
long-lived transient.
In Ar-saturated solutions (cf. Figure 5) after 40 µs the
spectrum outside the derivative-type band was the same
as that of Ar- or N2-saturated solutions at 1 µs (cf.
Figure 4 for N2). When argon or nitrogen was saturated
with water vapor, the resultant spectrum was un-
changed at 1 µs and was very similar to the spectrum
obtained in solution saturated with dry N2. However,
at 40 µs the spectral intensity of the H2O-containing
solutions (cf. Figure 5) was significantly enhanced in
the region between 1970 and 1850 cm-1. To distinguish
between the long-lived species in solutions with dry
nitrogen and with H2O-saturated gases we will refer to
them by the short-hand notations I-N2 and I-H2O,
The intermediates I-S, I-N2, and I-H2O exhibit
very similar IR bands, four each, shifted to the low-
(1)
frequency side of the A1 and E bands of CZ. The
absorptions of I-N2 and I-H2O are at somewhat lower
frequencies than those of I-S. Thus the lines of the
former two species are further away from those of the
initial complex CZ and therefore less prone to compen-
sating superposition by negative ∆A contributions due
to the bleaching of CZ. Thus the signals of I-N2 and
I-H2O appear stronger than those of I-S. When looking
for possible assignments of transients originating from
a photoinduced reaction from a pentacarbonyl carbene
complex, the possibility of CO loss leading to a tetra-
carbonyl complex is obvious. The IR transient spectra
of I-L (L ) solvent S, N2, H2O) can be readily assigned
to cis-tetracarbonyl carbene complexes. For comparison,
the IR band positions and symmetry assignments of two
isolated and structurally well-characterized cis-tetra-
carbonyl carbene complexes cis-[(CO)4{X}WdC(OMe)-
7
Ph] (X ) P(n-C4H9)3 , CH3CN8,20) have been included
in Table 1. The acetonitrile complex cis-[CO)4(CH3-
CN)WdC(OMe)Ph] is formed when the pentacarbonyl
carbene complex CZ is irradiated in acetonitrile. The
formation of the acetonitrile adduct is accompanied by
a UV/vis spectral change with an isosbestic point at 420
nm. The absorption maximum of the acetonitrile com-
plex is bathochromically shifted with respect to the
MLCT maximum of CZ.8,15 Our ∆A spectra of intermedi-
ates I-L are in accord with such a spectral change.
Thus both IR and UV/vis transient information support
the assignment of I-L as cis-tetracarbonyl carbene
(19) Rooney, A. D.; McGarvey, J . J .; Gordon, K. C.; McNicholl, R.-
A.; Schubert, U.; Hepp, W. Organometallics 1993, 12, 1277.

2358 Organometallics, Vol. 19, No. 12, 2000
Ta ble 1. In fr a r ed Sp ectr a l Da ta for cis-Tetr a ca r bon yl Com p lexes
Gut et al.
(1)
(2)
A1
A1
B1
B2
ref
solvent
n-hexane
n-hexane
n-hexane
n-hexane
n-hexane
n-hexane
n-hexane
cis-[(CO)4{P(n-C4H9)3}WdC(OMe)Ph]
cis-[(CO)4(CH3CN)WdC(OMe)Ph]
2018 m
2024 m
2040
2040
2025
2020
2020
2040
1925 m, sh
1940 s
∼1945b
∼1940b
1938b
1919 s
1912 m
∼1910b
1910b
1890 vs
1880 m
1880
1875
1875
1880
1863
1876
7
8a
I-S
this workc
this workc
this workc
this workc
this workc
17
I-N2
I-H2O
1910b
I-MeCN
1935b
1910b
I-olefind
1935b
1908b
syn-[(CO)4WdC(OMe)Ph]e
1948
CH4-matrix at 10 K
a
b
Assignments of the original paper are corrected. Determined from a superposition of the transient ∆A spectrum with the spectrum
of CZ. c From transient absorption measurements. Methyl trans-crotonate. e From matrix isolation experiments.
d
F igu r e 5. IR transient spectra obtained on laser photolysis
of CZ (2 × 10-3 M) in n-hexane at a delay time of 40 µs.
Solutions saturated with N2, Ar, and “wet” Ar, respectively.
F igu r e 6. IR transient signals observed on laser photolysis
of CZ (2 × 10-3 M) in n-hexane at 1956 cm-1 in N2- and
CO-saturated solution, respectively.
complex species. In I-S the solvent (S ) n-hexane)
occupies the sixth coordination site.
about 10 µs and reaches a plateau of negative ∆A
corresponding to a partial transformation of CZ into
transient I-N2. The initial amplitude of the signal and
the decay time are independent of the presence or
absence of CO. However, in the presence of CO, the
decay only reverts to the zero level of ∆A and no long-
term bleaching occurs. Obviously, the formation of the
tetracarbonyl complex I-N2 is inhibited due to a rapid
recombination of tetracarbonyl complex I-S with CO.
In N2-saturated solutions the intermediate I-N2
formed from I-S finally also decays. The kinetics is
second-order, indicating recombination of I-N2 with
photoeliminated CO in the solution. The rate constant
In flu en ce of CO. To confirm the structural assign-
ments of the transient intermediates, additional flash
photolysis experiments in CO-saturated solutions were
performed. Spectral traces recorded at 1 µs delay time
in n-hexane saturated with 1 bar of CO are shown in
Figures 3 and 4. All spectral features assigned to
transient CE in the UV/vis as well as in the IR are
unaffected by the presence of CO, whereas those parts
of the ∆A spectra assigned to I-S are significantly
weakened. Moreover, in the presence of CO intermedi-
ate I-N2 is not formed. Examples of two kinetic traces
in the absence and presence of CO are given in Figure
6. At 1956 cm-1 ∆A of the derivative-type band due to
CE is a maximum, whereas I-N2 does not absorb. In
solutions without added CO, the signal decays within
kI-N ,CO of this process was estimated as 3 × 108 M-1
2
s-1 21
This is in reasonable agreement with a value of
.
4.3 × 108 M1 s-1 that we obtained from a pseudo-first-
order analysis of kinetic experiments with various
concentrations of excess CO.
(20) We note here that the assignment of a band at 1790 cm-1 to
the B2-vibration of the complex cis-[(CO)4(CH3CN)WdC(OMe)Ph] in
acetonitrile as given by Foley et al.8 is incorrect. This band was reported
to grow in when monitoring a series of IR spectra during the irradiation
of CZ in acetonitrile. When we repeated this experiment in our
laboratory, we found that the time dependence of this band was clearly
different from that of the other bands assigned to cis-[(CO)4(CH3-
CN)WdC(OMe)Ph]. While the latter increased linearly at the begin-
ning of the irradiation and passed through a maximum after some time,
the rise of the 1790 cm-1 band was preceded by an induction period
and then increased monotonically. This is a clear indication that this
band belongs to a secondary product. Actually, after long irradiation
R ea ct ion of cis-Tet r a ca r bon yl In t er m ed ia t es
w ith Aceton itr ile. Since acetonitrile was known to
yield a stable cis-tetracarbonyl carbene adduct,8,15 the
reaction of this nucleophile with the tetracarbonyl
intermediate I-S or its successor I-N2 was investi-
(21) The rate constant kI-N
and the rate constant obtained from
2,CO
the second-order analysis of the transient absorption signal (k∆A) are
related by kI-N
) k∆A∆ꢀd where d ) 0.05 cm is the optical path
2,CO
this band persists together with another sharp band at 1911 cm-1
.
length of the cuvette and ∆ꢀ ≡ ꢀI-N ,CO - ꢀC the difference of absorption
coefficients of I-N2 and CZ. To fin2d the przoper conversion factor from
the observed ∆A spectrum to the ∆ꢀ spectrum, it was assumed that
The corresponding product, which no longer exhibits the NMR
resonances of the carbene ligand, is the fac-W(CO)3(CH3CN)3 complex,
as can be verified by comparing it with the product obtained by
irradiation of W(CO)6 in acetonitrile. The band positions of the cis-
[(CO)4(CH3CN)WdC(OMe)Ph] complex in acetonitrile are 2018 cm-1
(m), 1899 cm-1 (s), 1838 cm-1 (sh). These are practically the same as
in ref 6: 2017 cm-1 (m), 1900 cm-1 (s), 1840 cm-1 (sh). (Note that in
the latter reference the last value is misprinted as 1804 cm-1.)
ꢀI-N (1948) , ꢀC ) 8600 M-1 cm-1, which seems reasonable since most
2
of the IR bandszof I-N2 appear well separated from the lines of CZ
and 1948 cm-1 is the position of the absorption maximum of CZ.
Normalizing the ∆A spectrum by the appropriate factor, it was found
that at 1891 cm-1, where the second-order kinetics was evaluated, the
absorption coefficient ∆ꢀ adopts a value of 2315 M-1 cm-1
.

Tetracarbonyl Intermediates
Organometallics, Vol. 19, No. 12, 2000 2359
F igu r e 7. (a) Effect of acetonitrile (2 5 × 10-5 M, 9 9.5 ×
10-5 M, b 1.9 × 10-4 M) on IR transient spectra observed
on laser photolysis of CZ (2 × 10-3 M) in N2-saturated
n-hexane at delay time of 40 µs. (b) (- - -) IR spectrum of
I-MeCN obtained by correcting the final transient ∆A
spectrum (symbols b in a) for the ∆A due to the depletion
of CZ. (s) FTIR spectrum of the photoproduct obtained on
stationary illumination of CZ at 313 nm in i-octane 0.1 M
in MeCN.
F igu r e 8. (a) Series of IR transient signals measured at
1903 cm-1 on photolysis of CZ (2 × 10-3 M) in n-hexane
saturated with 1 bar of CO and containing various con-
centrations of acetonitrile. (b) Kinetic simulation. For
details see text.
gated. Pertinent transient experiments were carried out
with increasing concentrations of acetonitrile in n-
hexane.
× 104 s-1, from which a second-order rate constant of 4
× 108 M-1 s-1 is derived.
The transient IR spectra at a delay time of 40 µs for
various concentrations of acetonitrile are displayed in
Figure 7. There are two isosbestic regions at about 2030
and 1930 cm-1. The appearing spectrum is assigned to
a complex I-MeCN. In the region 1925-2000 cm-1 its
spectrum overlaps with the bleaching of CZ. Here the
exact positions of the I-MeCN maxima were evaluated
by adding to the observed ∆A spectrum a spectrum of
CZ multiplied by a suitable factor. The factor was chosen
in such a way that no negative absorption remained in
the resultant spectrum. The positions of the four-band
maxima of I-MeCN thus obtained are listed in Table
1. They agree well with those determined from the
stationary IR spectrum of the cis-acetonitrile carbene
tetracarbonyl complex.
Direct kinetic information on the association reaction
of the cis-tetracarbonyl intermediate I-N2 with aceto-
nitrile has been derived from kinetic traces at 2046 cm-1
(a band maximum of I-N2) and the corresponding
maximum of I-MeCN at 2019 cm-1. After the buildup
of the I-N2 absorption within approximately 10 µs (due
to the I-S + N2 f I-N2 process) a decay at 2046 cm-1
and a further rise at 2019 cm-1 follow. Both have the
same time constant. For [MeCN] ) 1.9 × 10-4 M the
pseudo-first-order rate constant was determined as 7.6
The acetonitrile concentrations employed for observ-
ing the conversion from I-N2 to I-MeCN were fairly
low. At higher concentrations of acetonitrile, formation
of I-N2 is prevented and I-MeCN is formed directly
from I-S. However, the rate of this process is within
the time resolution of our IR detection system. Never-
theless, kinetic information about the formation of
I-MeCN from I-S and the reverse of this reaction could
be obtained by means of competition experiments with
CO. The IR band structure of the transient obtained
with acetonitrile in CO-saturated solutions of CZ was
indistinguishable from that in CO-free solution, indicat-
ing the formation of the same product I-MeCN either
from I-N2 or from its precursor I-S.
A representative set of curves at 1903 cm-1 for various
concentrations of acetonitrile in CO-saturated solutions
is shown in Figure 8a. The significant features are as
follows:
(i) The maximum of the initial signal increases with
acetonitrile concentration and approaches a saturation
value at high acetonitrile concentrations.
(ii) The decay time of the I-MeCN absorption, which
varies between fractions of seconds to tens of seconds,
increases also with increasing acetonitrile concentration.
(iii) The rise time (not resolved in Figure 8) of all

2360 Organometallics, Vol. 19, No. 12, 2000
Gut et al.
maximum of CZ, the transient signals exhibit negative
∆A, and a recovery of CZ would be conveniently observed
by a returning of the signal to zero (cf. for example
Figure 1b).
In accord with these time-dependent observations, a
permanent photoinduced loss of CZ is detected when
measuring stationary IR spectra after laser photolysis.
In n-hexane (2 × 10-3 M in CZ and 0.2 M of methyl
trans-crotonate) a bleaching of 5 × 10-5 M per laser
pulse of 100 mJ cm-2 results. From these data a
quantum yield of 1.7% can be estimated for the photo-
bleaching of CZ by the olefin. From the IR spectra it
can be shown that upon this photolysis the pentacar-
bonyl complex (CO)5W(olefin) is formed.22 The yield of
this decomposition product amounts to about 20%.
Other products containing the carbene moiety, in par-
ticular cyclopropanes, have not been detected.
Discu ssion
Our observations can be accounted for by the pro-
cesses represented in Scheme 1. Excitation of the Z
isomer of the carbene complex (CZ) by laser light of 308
nm affords an upper ligand field excited state CZ* with
three possible relaxation pathways: (i) radiationless
decay back to the ground state most probably through
the intermediacy of the lowest excited MLCT state; (ii)
isomerization to form CE in appreciable amounts; (iii)
loss of CO in an excited ligand field state and formation
of a pentacoordinated tetracarbonyl carbene complex
(CO)4W(cb) (I).
The coordinatively unsaturated complex I either
irreversibly recombines with CO to regenerate the
initial complex CZ or reversibly combines with L to form
a complex (CO)5W(cb)L (L denoting the solvent S or
other coordinating species such as N2, H2O, MeCN, or
an olefin). It is unlikely that we can detect the penta-
coordinated complex I in the microsecond regime since
W(CO)5 or Cr(CO)5 complexes generated by photolysis
of the corresponding hexacarbonyls25-29 have been
reported to coordinate a solvent molecule within pico-
seconds. Therefore the first tetracarbonyl-type transient
that we see is assigned to the solvent adduct I-S. The
transformation of I-S observed during the first 10 µs
in nitrogen-purged solutions can be assigned to the
replacement of a solvent ligand by N2 to yield the
complex I-N2 (or by H2O yielding I-H2O in the case of
solutions purged with wet argon). This assignment is
substantiated by the spectral features characteristic for
F igu r e 9. (a) Series of IR transient signals measured at
1903 cm-1 on photolysis of CZ (2 × 10-3 M) in n-hexane
saturated with 1 bar of CO and containing various con-
centrations of methyl trans-crotonate. The olefin concentra-
tions are (from below) 6 × 10-4, 2 × 10-3, 6 × 10-3, 1.9 ×
10-2, 1.9 × 10-1 M. (b) Kinetic simulation. For details see
text.
signals is within the time resolution of the IR detection
system. The UV/vis channel could not be used to obtain
further kinetic information. In this spectral region the
short-time kinetics are dominated by the strong signals
due to the CE f CZ reversion process.
R ea ct ion of cis-Tet r a ca r bon yl In t er m ed ia t es
w ith a n Olefin ic Rea cta n t. The analogous experi-
ments have been performed with methyl trans-crotonate
as an olefinic reactant. The IR spectral change is very
similar to that shown for acetonitrile in Figure 7. The
frequencies of the IR-band maxima are included in
Table 1.
The competition experiment involving CO-saturated
solutions and increasing concentrations of olefin yielded
the signal traces shown in Figure 9a. In principle, the
findings were analogous to those with acetonitrile.
However remarkable quantitative differences were ob-
served:
(22) The following bands have been assigned to the [(CO)5W(methyl
trans-crotonate)] in n-hexane: A1g 2089 (2093) cm-1, Eg 2007 (2009)
cm-1, T1u 1977/1955 (1986/1962) cm-1 13C[E(C4v)] 1970 (1976). The
,
band positions are very similar to those obtained by Grevels et al.23
for [(CO)5W(methyl acroleate)] upon photolysis of [W(CO)6] in solutions
containing methyl acroleate. The latter data (values in parentheses)
have been verified and supplemented in our laboratory.
(23) Grevels, F.-W.; Goddard, R.; Kru¨ger, C. Z. Naturforsch. B 1980,
35, 1298.
(24) Gut, H.-P. Doctoral Thesis, University of Konstanz, 1996.
(25) (a) Simon, J . D.; Xie, X. J . Phys. Chem. 1986, 90, 6751. (b)
Simon, J . D.; Xie, X. J . Phys. Chem. 1987, 91, 5538.
(26) (a) J oly, A. G.; Nelson, K. A. J . Phys. Chem. 1989, 93, 2876. (b)
J oly, A. G.; Nelson, K. A. Chem. Phys. 1991, 152, 69.
(27) Lee, M.; Harris, C. B. J . Am. Chem. Soc. 1989, 111, 8963.
(28) Sprague, J . R.; Arrivo, S. M.; Spears, K. G. J . Phys. Chem. 1991,
95, 10528.
(i) The decay time of the I-olefin adduct is 4-5 orders
of magnitude shorter than with acetonitrile.
(ii) As the olefin concentration increases, the lifetime
of the adduct reaches a limiting value of 125 µs.
(iii) No recovery of CZ occurs with high olefin concen-
trations, even in the solutions saturated with CO at 1
bar. This may be concluded from observations in the IR
region around 1950 cm-1. Here, at the absorption
(29) (a) Dougherty, T. P.; Heilweil, E. J . Chem. Phys. Lett. 1994,
227, 19. (b) Dougherty, T. P.; Heilweil, E. J . J . Chem. Phys. 1994, 100,
4006.

Tetracarbonyl Intermediates
Organometallics, Vol. 19, No. 12, 2000 2361
Sch em e 1
tetracarbonyl complexes (cf. Table 1) and by the effect
of dry Ar in place of N2 used for purging the solution.
Whether the conversion of I-S into I-N2 is quantitative
or not cannot be definitely decided. The two species I-S
and I-N2 should interconvert through a dissociative
mechanism involving the coordinatively unsaturated
species I as the intermediate. According to the experi-
mental results, the equilibrium between the three
species must be established within 10 µs.
Qu a n tu m Yield s. The photochemical isomerization
CZ f CE has been characterized in some detail by the
groups of Stufkens17 and McGarvey.16 However, the
quantum yield of this process has not been reported.
According to Servaas et al.,17 the IR bands in the CO-
stretching region of CZ are shifted by 1-3 cm-1 with
respect to those of CZ. Qualitatively this is in line with
our observation of the derivative-type absorption band
observed by time-resolved IR at short delay times. By
using the ∆ν˜ values given in ref 17 the concentration of
CE produced by one laser flash can be estimated.
Since the ∆ν˜ shifts are small in comparison with the
spectral widths of the lines (determined by the spectral
resolution of our instrument), the following relation
applies:
the absorption band at 403 nm associated with the
transformation CZ f CE: By applying the relation
analogous to eq 1 and using ∆A430 ) -0.09 (cf. Figure
3), (dꢀ/dλ)430 ) -142 M-1 cm-1 nm-1, and d ) 0.05 cm,
we obtain ∆λ ) -(14 ( 7) nm.
To estimate the quantum yield of photoinduced loss
of CO, we need to quantify the resulting transient
concentration of tetracarbonyl complexes. As has been
described in the Results section, the tetracarbonyl
complex intermediate is quantitatively scavenged by the
olefin methyl trans-crotonate. There is no regeneration
of the initial complex from the intermediate I-olefin.
Thus the quantum yield of photolysis under such
conditions (1.7%) can be also assigned to the quantum
yield of primary CO loss. It should be noted that this
quantum yield is in fairly good agreement with the
value of 1.1% reported by Foley et al.8 for the photoin-
duced bleaching of CZ under continuous illumination at
313 nm in neat acetonitrile.
From their experiments carried out under low-tem-
perature matrix isolation conditions Servaas et al.17
concluded that CO loss occurred after formation of the
CE (syn) isomer and that the tetracarbonyl complex
generated from CE was still in the syn-conformation.
By contrast, our observations do imply that CO loss
occurs in a one-photon process with CZ. This follows
from the similarity of the quantum yields for pulsed and
continuous illumination. In liquid solution at room
temperature the CE isomer returns to CZ within a few
microseconds. Therefore, under stationary illumination
the quantum yield for irreversible decomposition initi-
ated by loss of CO should be negligible if only the CE
isomer could undergo this reaction. Furthermore, a
strict linearity between the laser pulse energy and the
observed ∆A for the tetracarbonyl intermediates is
observed. This would not be compatible with a two-
photon mechanism of CO loss which requires CZ f CE
photoisomerization as a first step. Nevertheless, in view
of the different orders of magnitude of the quantum
yields of photoisomerization and loss of CO, it seems
that our observations and those reported by the Stufkens
group are compatible. From our results one would
expect, for conditions where CE is long-lived, i.e., in a
low-temperature matrix, that the CZ f CE conversion
dꢀ
dν˜ max
∆Amax ) -∆cd
∆ν˜
(1)
(
)
where ∆c is the concentration change due to the conver-
sion of CZ to CE, d ) 0.05 cm is the optical path length,
and (dꢀ/dν˜)max is the steepest slope of the IR line. With
ν˜max ) 1956 cm-1, ∆ν˜ ≈ 1 ( 0.5 cm-1 17
,
(dꢀ/dν˜)1953 )
-(415 ( 30) M-1 (cf. Figure 2), and ∆Amax ≈ 0.01530 we
obtain cC ) -∆cC ) (9 ( 4.5) × 10-4 M for the
E
Z
experimental conditions pertaining to the signals in
Figure 4. From this value and the equivalent concentra-
tion of absorbed photons of 3 × 10-3 M the quantum
yield φC
(308 nm) ) 0.3 ( 0.15 for photochemical CZ
ZfCE
f CE isomerization is obtained. With the value of ∆cC
Z
we can also estimate the as yet unknown shift ∆λ of
(30) From Figure 4 we obtain ∆A1956 ) 0.006. However, it must be
taken into account that due to the short decay time of the signal of
only about twice the time constant of the IR detection system, the
apparent signal maximum reaches only about 40% of the true value.
Thus a corrected ∆A1956 value of 0.015 should be used.

2362 Organometallics, Vol. 19, No. 12, 2000
Ta ble 2. Ra te Con sta n ts for Exch a n ge of Solven t or N2 Liga n d s by CO
Gut et al.
complex X
kX,CO[CO]sat,1bar, s-1
kX,CO, M-1 s-1
ref
[(CO)5W(c-hexane)]
[(CO)5W(n-heptane)]
trans-[(CO)4(PPh3)W(n-heptane)]
cis-[(CO)4(PPh3)W(n-heptane)]
cis-[(CO)4(N2)W(cb)] ) I-N2
“naked” W(CO)5 in C6F11CF3
2 × 103
3.5 × 104
1.4 × 105
2 × 105 a
31
32
32
32
1 × 104
1 × 106 a
3.5 × 106 a
1.4 × 107 a
3.5 × 108 b
2 × 109
this work
33
Assuming a saturation concentration of CO of 10-2 M. Average value of second-order analysis in solution without excess CO and
a
b
pseudo-first-order analysis of experiments with excess CO.
Ta ble 3. Kin etic P a r a m eter s Used in th e Globa l
Sim u la tion 34 of th e Com p etitive Kin etics betw een
CO a n d Liga n d L in F igu r es 8 a n d 9
should occur first followed, on a longer time-scale, by
formation of the CO loss product from CE. In fact our
results do not exclude photoinduced CO loss from the
CE isomer; however, under conditions of low degree of
conversion from CZ to CE this process is just not
significant.
ligand
rate constant
acetonitrile
5 s-1
methyl trans-crotonate
k-L
kirr
105 s-1
8 × 103 s-1
Recom bin a tion of Tetr a ca r bon yl Com p lex Sp e-
cies w ith CO. In Table 2 the CO recombination rate
constants are compared with several literature values
of related reactions. The reactivity of the coordinatively
unsaturated tetracarbonyl carbene complex is only a
factor of 10 less than that of a “naked” W(CO)5 complex
and thus ranges among the highest ones. As has been
pointed out by Dobson et al.,32 the relative order of such
rate constants of substitution of CO for a weakly bound
ligand can be understood on the basis of a dissociative
mechanism. The release of the ligand to be replaced is
enhanced by bulky substituents cis to the leaving ligand.
However, the carbene ligand in cis-[(CO)4(N2)W(cb)] (I-
N2) or cis-[(CO)4(n-hexane)W(cb)] (I-S) is certainly less
bulky than the PPh3 ligand in cis-[(CO)4(PPh3)W(n-
heptane)]. Thus other reasons must be responsible for
the fast reaction with CO. These reasons presumably
are related to the steric and electronic properties of the
carbene ligand and its interaction with the metal.
Com petition of CO with Oth er P oten tial Ligan ds.
In CO-saturated solution the lifetime of I-S is esti-
mated to be shorter than 0.5 µs. Nevertheless, other
potential ligands such as acetonitrile or methyl trans-
crotonate can compete with CO. The results of the
competition experiments, namely, the rise of the signal
amplitude of I-L and the increase of the decay time of
I-L with the concentration of L, provide evidence for
the reversibility of I-L formation and for the dissocia-
tive mechanism of the replacement of L by CO. Based
on the mechanism represented in Scheme 1 the depen-
dence of the I-L-related IR transient signals on [L]
shown in Figures 8a and 9a can be simulated and the
parameters κL ) kI,L/kI,CO, k-L, and kirr can be deter-
mined. For the details of the kinetic treatment see the
Experimental and Methods Section.
κL,CO
kI-N
2
2
7 × 108 M-1 s-1 a
7 × 108 M-1 s-1 a
2,L
4 × 108 M-1 s-1 b
a
Obtained by multiplying the respective value of kI-N
with
2,CO
b
κL,CO
.
From direct measurement in CO-free solution.
invariant set of parameters, the fit of some of the
individual kinetic traces is only approximate.
The ratio of rate constants κL for the reaction of the
tetracarbonyl intermediate I with the ligand L and CO
is 2 for both acetonitrile and the olefin.This means that
these ligands react 2 times faster with I (and conse-
quently with I-S and I-N2) than CO. The values for
kI-N ,L ) κLkI-N ,CO are given in Table 3. The value of
2
2
kI-N ,L for L ) MeCN compares well with the value of 4
2
× 108 M-1 s-1 estimated from a direct observation of
the rise time of the I-MeCN IR signal in CO-free
solution (cf. Results section). The kI-N ,L values for
2
MeCN and methyl trans-crotonate are not far from those
of diffusion-controlled reactions and are about 1 order
of magnitude larger than the rate constant of 5.5 × 107
M-1 s-1 observed for the reaction of [(CO)5W-
(cyclohexane)] with acetone31 or of 6.5 × 107 M-1 s-1
calculated from the published kinetic data of the reac-
tion of [(CO)5W(methylcyclohexane)] with 4-acetylpyri-
dine.35
The rate constants k-L characterize the kinetic stabil-
ity of the complexes I-L toward dissociation. The values
(cf. Table 3) indicate a much higher kinetic (and, since
the rate constants of formation are equal, also thermo-
dynamic) stability of the acetonitrile than of the olefin
complex.
Another remarkable difference between the two I-L
complexes is the existence of an effective irreversible
decay channel to a product other than CZ in case of the
olefin complex. The details of this decay mechanism are
not yet clear, except for the fact that the carbene ligand
is lost in the course of this reaction and that a [(CO)5W-
(olefin)] complex has been identified as a product in
about 20% yield in n-hexane. In di-n-butyl ether the
complex [(CO)5W(solvent)] is formed with about the
same yield, as evidenced by the in-situ analysis of the
IR spectra. It seems that an understanding of the
irreversible decomposition reaction of I-olefin is very
important for controlling photoinduced coupling of the
As can be seen in Figures 8 and 9, the overall behavior
of the kinetic experiments is very well reproduced by
the theoretical curves, although, with a globally34
(31) Hermann, H.; Grevels, F.-W.; Henne, A.; Schaffner, K. J . Phys.
Chem. 1982, 86, 5151.
(32) Dobson, G. R.; Hodges, P. M.; Healy, M. A.; Poliakoff, M.;
Turner, J . J .; Firth, S.; Asali, K. J . J . Am. Chem. Soc. 1987, 109, 4218.
(33) Kelly, J . M.; Long, C.; Bonneau, R. J . Phys. Chem. 1983, 87,
3344.
(34) This term indicates that the same set of values of rate constants
has to account for the behavior of a family of kinetic curves. Thus,
while suitable sets of individually adjusted rate constants would
produce more excellent fits to each curve, the global fit usually yields
less good fits to all curves of the family; however the ambiguity in
determining the values is reduced.
(35) Lees, A. J .; Adamson, A. W. Inorg. Chem. 1981, 20, 4381.

Tetracarbonyl Intermediates
Organometallics, Vol. 19, No. 12, 2000 2363
F igu r e 10. Schematics of the laser flash spectrometer for simultaneous UV/vis and IR transient absorption detection.
Abbreviations: Am p , ac/dc preamplifier; C, optical flow cell; Det, MCT detector; DO, digital oscilloscope; P C, personal
computer; P M photomultiplier. For details see text.
The flow system was fed from a 300 mL steel tank, wherein
the sample solutions were flushed with the appropriate gases
or gas mixtures made up in a second steel tank preceding the
one with the sample solution. The flow of the solution was
effected by the gas pressure and controlled by a magnetic valve
positioned downstream of the optical cell.
Transient absorption in the UV/vis was monitored by a
detection line comprising a pulsed xenon high-pressure lamp
(Osram XBO 150 W), quartz optics, monochromator (AMKO,
f/3.4), and a photomultiplier (Hamamatsu R 955). For IR
monitoring, continuous probe light from an IR source (Mu¨ller
DUO 150) with a cesiwid rod was focused by ZnSe lenses,
spectrally analyzed by a grating monochromator (AMKO), and
detected by a MCT solid-state device (Infrared Associates Ltd.)
operated at 77 K. The IR photosignal was preamplified by a
factor of approximately 1000 using a custom-built dc and ac
amplifier before feeding this signal and that from the photo-
multiplier of the UV/vis detection line into a LeCroy 9410 dual
channel digital oscilloscope.
Synchronization or suitable time delay of action of the
various components of the transient spectrometer was achieved
by a home-built trigger generator controlling the following
functions: opening and closing of the magnetic valve in the
flow-system for a time allowing for the replacement of the
solution in the optical cell,36 opening and closing of the shutters
in the UV/vis and IR beams (restricting these opening times
to a minimum period required for recording a transient signal
prevented undesired photobleaching and heating of the sample
solution by the probe light), triggering of the UV/vis probe light
pulse, triggering of the laser and of the digital oscilloscope.
For further data processing and display the recorded signals
were transferred to a personal computer via an IEEE-488
interface. For improving the S/N ratio transient signals were
averaged over typically 8-16 single measurements performed
at a repetition rate of about 1 Hz.
W-coordinated olefin and the carbene ligand. This
reaction mechanism is presently under investigation.
Con clu sion
By using the combination of time-resolved UV/vis and
IR spectroscopy it was shown that short-lived carbonyl
complex transients can be spectroscopically character-
ized and their reactions studied even with formation
quantum yields as low as 1% and the occurrence of a
much more efficient competitive photoprocess. In the UV
photolysis of the Fischer carbene complex CZ at room
temperature, photoisomerization and CO loss occur as
independent single-photon processes. The quantum
yield of isomerization is more than 10 times higher than
that of CO photoelimination. Our observations confirm
the notion9,10 that CO elimination has a low quantum
yield even if the complex is photoexcited to an LF state.
This is due to the competition of CO elimination with
electronic relaxation to the lowest excited state, which
is predominantly of MLCT character. The LF to MLCT
electronic conversion increases the charge on the car-
bene ligand and seems to favor the Z/E isomerization
process.
Exp er im en ta l a n d Meth od s
La ser F la sh Sp ectr om eter (for details cf. ref 24). A special
setup was developed to allow for laser pulse photolysis with
simultaneous observation of UV/vis and IR transient absor-
bance detection. The schematics of this setup is shown in
Figure 10. A XeCl excimer laser (Lambda Physik, LPX 105)
was used as the flash photolysis light source. Typically the
laser pulse energy was around 100 mJ . The laser beam was
focused from an initially 8 × 23 mm2 cross section onto the
entrance window of the IR cell of 10 × 10 mm2 by means of
two cylindrical lenses. The optical cell was custom-built with
two polished, round CaF2 windows of 30 mm diameter and 5
mm thickness. The body of the cell was made of V41 steel with
connectors to two steel capillaries, allowing for the use of the
cuvette as a flow cell. The mechanical properties of the
construction allowed for operation of the cell at internal
pressures up to 10 bar.
Ma ter ia ls a n d Meth od s. Pentacarbonyl[methoxy(phenyl)-
carbene]tungsten ([(CO)5WdC(OMe)Ph], CZ) was synthesized
according to ref 37 and purified by fractionated sublimation
(36) At 1 Hz repetition frequency the volume exchanged between
two laser pulses amounts to 39 µL. Thereby about 80% of the
illuminated volume of the cuvette is replaced. In view of the low
quantum yield of less than 2% permanent decomposition per laser flash
this may be considered a sufficient renewal of solution.
(37) Fischer, E. O.; Schubert, U.; Kleine, W.; Fischer, H. Inorg.
Synth. 1979, 19, 164.

2364 Organometallics, Vol. 19, No. 12, 2000
Gut et al.
[I-L]0
in vacuo until admixtures of [W(CO)6] were no longer detect-
able. A very useful and sensitive indicator of [W(CO)6] is the
kI,CO[CO]
kC ) k-L
) k-L 1 -
(5)
(
)
z
kI,CO[CO] + kI,L[L]
[I]0
photochemical bleaching of the IR absorption at 1980 cm-1
.
Methyl trans-crotonate was obtained from Fluka (purum; 98%)
and used without further purification. Solvents used were
n-hexane (SPECTRANAL, Riedel-de Hae¨n, and UVASOL,
Merck), acetonitrile (UVASOL, Merck), and di-n-butyl ether
(Wacker-Chemie, >99.3%). Only the latter was further purified
by reflux boiling over metallic sodium for 4 h, distillation, and
rectification over metallic sodium under a nitrogen atmo-
sphere. Gases applied to flush and saturate the solutions at 1
bar were N2 (5.0 Sauerstoffwerk Friedrichshafen GmbH), Ar
(4.8, Messer Griesheim GmbH), and CO (4.7, Messer Griesheim
GmbH). Sample solutions were usually 2 × 10-3 M in the
carbene complex. To saturate the solvent with the applied gas
or gas mixture a moderate, solvent-vapor saturated gas stream
was passed through 200 mL of sample solution in the supply
tank of the flow system for 45 min.
From eqs 4 and 5 it follows that the rate constant kobs of the
decay of I-L should decrease linearly with the initial ampli-
tude [I-L]0.
A complete (“global”)34 analysis of the kinetic signals
obtained for a series of ligand concentrations requires the
determination of k-L, kirr, and κL. The value of κL can be
obtained from a plot of the inverse of the initial signal
amplitude ∆A0(ν˜I-L) at a wavenumber ν˜I-L, where I-L absorbs
versus the inverse of the ligand concentration:
[CO]
{∆A0(ν˜I-L)}max κL{∆A0(ν˜I-L)}max
1
1
1
[L]
)
+
(6)
∆A0(ν˜I-L
)
The intercept of this diagram yields the value {∆A0(ν˜I-L)}max
,
which corresponds to the complete primary capture of all
tetracarbonylcarbene intermediates by L. A plot of kobs versus
Kin etic An a lysis of Sch em e 1. The initial concentration
of the ligand adduct I-L is given by
the initial signal amplitude ∆A0(ν˜I-L) yields k-L and kirr
∆A0(ν˜I-L
{∆A0(ν˜I-L)}max
:
kI,L[L]
κL[L]
)
[I-L]0 ) [I]0
) [I]0
(2)
kobs ) kirr + k-L - k-L
(7)
kI,L[L] + kI,CO[CO]
κL[L] + [CO]
The curves shown in Figures 8b and 9b were calculated by
eqs 2-5 using parameters (cf. Table 3) obtained in the way
just pointed out. For L ) olefin (Figure 9) the shortest time
resolution available had to be employed for the kinetic
where [I]0 is the total concentration of CO-loss products
generated by the laser flash and κL ) kI,L/kI,CO. The decay of
I-L is exponential with rate constant kobs
:
measurements. The effect of the finite response time (r-1
≈
[I-L]t ) [I-L]0 exp(-kobst)
(3)
0.5 µs) of the IR detection system was taken into account when
calculating the kinetic curves. The pertinent equation is
Two channels contribute to the decay: (a) recombination with
CO to give CZ with an effective rate constant kCz and, possibly,
(b) an irreversible decay to other products with a rate constant
r
[I-L]t ) [I-L]0
{exp(-kobst) - exp(-rt)} (8)
r - kobs
k
irr, i.e.,
Ack n ow led gm en t . Financial support by the
Schwerpunktprogramm of the federal state of Baden
Wu¨rttemberg and by the Fonds der Chemischen Indus-
trie is gratefully acknowledged. U.E.S. thanks Professor
D. J . Stufkens for a stimulating discussion.
kobs ) kC + kirr
(4)
z
The rate constant kC is a product of two factors: the rate
z
constant k-L of dissociation of the ligand adduct and the
probability that the ensuing coordinatively unsaturated in-
termediate I does react with CO in the next step:
OM990708Q