FTIR studies of iron-carbonyl intermediates in allylic alcohol photoisomerization

Article abstract of DOI:10.1002/chem.200600053

The 532 or 355 nm laser-induced photoisomerization of allylic alcohols to aldehydes catalyzed by [Fe₃(CO)₁₂] or [Fe(CO) ₄PPh₃] in hexane was investigated. The Fourier transform infrared (FTIR) absorption spectra of iron-carbonyl intermediate species such as [Fe(CO)₅], [Fe(CO)₄(RC₃H₄OH)], and more importantly the π-allyl iron-carbonyl hydride species [FeH(CO) ₃(R-C₃H₃OH)] (R = H, Me, Ph) were recorded during the catalytic process using [Fe₃(CO)₁₂] as the catalytic precursor. When [Fe(CO)₄PPh₃] was photolyzed with 355 nm, [FeH(CO)₃(R-C₃H₃OH)] was also generated indicating the common occurrence of the species in these two systems. The π-allyl hydride species is long believed to be a key intermediates and its detection here lends support to the π-allyl mechanism of the photoisomerization of allyl alcohols.

Full text of DOI:10.1002/chem.200600053

DOI: 10.1002/chem.200600053
FTIR Studies of Iron–Carbonyl Intermediates in Allylic Alcohol
Photoisomerization
[
a]
Thiam Seong Chong, Sze Tat Tan, and Wai Yip Fan*
Abstract: The 532 or 355 nm laser-in-
duced photoisomerization of allylic al-
cohols to aldehydes catalyzed by
[FeH(CO)3
Ph) were recorded during the catalytic
process using [Fe (CO) ] as the cata-
A
H
R
U
(R-C H OH)] (R=H, Me,
[FeH(CO) (R-C H OH)] was also gen-
A
T
E
N
3
3
3 3 3
erated indicating the common occur-
rence of the species in these two sys-
tems. The p-allyl hydride species is
long believed to be a key intermediates
and its detection here lends support to
the p-allyl mechanism of the photo-
3
12
[
Fe (CO) ] or [Fe(CO) PPh ] in
lytic precursor. When [Fe(CO) PPh ]
4 3
3
12
4
3
hexane was investigated. The Fourier
transform infrared (FTIR) absorption
spectra of iron–carbonyl intermediate
was
photolyzed
with
355 nm,
Keywords: alcohols
pounds · IR spectroscopy · iron ·
isomerization
·
allyl com-
species such as [Fe(CO) ], [Fe(CO)
A
H
R
U
G
isomerization of allyl alcohols.
A
H
R
U
G
5
C H OH)], and more importantly the
3
4
p-allyl iron–carbonyl hydride species
Introduction
ate species could be detected directly during the room-tem-
perature process itself with low catalytic loading.
Photoisomerization of alkenes by iron–carbonyl species has
been the subject of intense study for many years. Several as-
pects such as the catalytic efficiency of different metal pre-
cursors and detection of key intermediates generated sepa-
rately in low-temperature matrices and in the gas phase
In this work, we have largely focused on triiron dodeca-
carbonyl [Fe (CO) ] as the catalytic precursor, because of
its much milder toxicity as well as the ease of handling the
3
12
solid. Since the toxic properties of volatile [Fe(CO) ] could
5
limit its use as a versatile reagent in most organic syntheses,
photoisomerization in the presence of [Fe (CO) ] at room
[1–5]
have been explored.
[Fe(CO) ]) has been the precursor of choice for photoisom-
By far, iron–pentacarbonyl
3
12
[4]
(
temperature may offer a better alternative. [Fe (CO) ] is
3 12
5
A
C
H
T
R
E
U
N
G
erization, and in the case of allylic alcohols, aldehydes or
also a much less studied catalytic precursor in terms of its
mechanism, as it is often assumed to behave in a similar
aldol products are produced irreversibly upon tautomeriza-
[6–11]
tion.
key intermediate [FeH(CO)3
mainly under low-temperature matrix or gas-phase condi-
tions and a computational study on the mechanism of
iron–carbonyl-mediated isomerization of allylic alcohols to
saturated carbonyls has been carried out, more convincing
evidence to support the catalytic photoisomerization of al-
lylic alcohols could be obtained if some of these intermedi-
While the proposed active catalyst [Fe(CO) ] and
manner to [Fe(CO) ]. Another catalytic percursor, [Fe-
3
5
A
H
R
U
G
A
H
R
U
G
[12]
studies.
[1]
We have used pulsed laser FTIR absorption spectroscopy
to monitor directly the spectral evolution of the precursor,
intermediates, and final product of the photoisomerization
of allyl alcohols to aldehydes during isomerization. Com-
pared to a broadband source, the intense laser light, either
in the visible (532 nm) or UV region (355 nm) could en-
hance the probability of detecting low concentrations of in-
termediate species. We have chosen the study of allylic alco-
hol isomerization rather than that of a simple alkene, be-
[6]
[
a] T. S. Chong, S. T. Tan, Dr. W. Y. Fan
Department of Chemistry
National University of Singapore
cause of the ease of monitoring the strong n stretch at
CO
ꢀ1
1
700–1730 cm of aldehyde by using IR spectroscopy. At
3
Science Drive 3, 117543 (Singapore)
the same time, transition-metal–carbonyl intermediates pro-
duced during the catalytic cycles could be identified by their
Fax : (+65)677-91-691
E-mail: chmfanwy@nus.edu.sg
ꢀ
1
distinctive CO stretches around 1900–2100 cm .
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
5128
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 5128 – 5133

FULL PAPER
Experimental Section
3
0
All the manipulations were carried out in the dark under a nitrogen envi-
ronment to avoid decomposition of the catalysts. Hexane and toluene
a
2
0
were distilled before use and [Fe(CO)
4
PPh ] was prepared according to
3
[
12]
literature methods.
[Fe
3
(CO)12] and [Ru (CO)12] were obtained from
3
Strem Chemicals, while allyl alcohol, methylallyl alcohol, and phenylallyl
alcohol were from Sigma–Aldrich and used without further purification.
A typical preparation for each kinetic run would be to mix the allyl alco-
10
T%
–10
–20
ꢀ
5
a
hol (~0.01mol) and the metal–carbonyl complex ( ~10 mol, loading=
.1–1.0%) in hexane or toluene (30 mL).
0
d
d
We originally intended to use a broadband xenon lamp source (200 W)
to initiate the catalysis, but it was found to be too weak to produce any
measurable amount of iron–carbonyl intermediates, although conversion
of allyl alcohol to aldehyde was observed. Hence a Nd-YAG pulsed laser
system (10 Hz Continuum Surelite III-10, 8 ns pulse width) was used in-
stead to generate either 532 nm (ꢁ10–20 mJ per pulse) or 355 nm (5–
c
b
c
b & c
1
5 mJ per pulse) wavelength.
A brief description of the cell used for performing FTIR monitoring on
the reaction mixture is given here and can also be found in refer-
ence [13]. It wa a home-made stainless steel cell of volume 40 cm fitted
2100
2050
2000
1950
–1
Wavenumbers / cm
3
Figure 2. The absolute IR bands of a) [Fe
c) [Fe(CO) (C OH)], and d) [FeH(CO) (C
(CO)12] (2 mg) and allyl alcohol (1mL) in
3
5
(CO)12], b) [Fe(CO) ],
2
with CaF windows for passage of the IR probe beam. Absorbance path-
lengths of 0.2 to 5 mm was accommodated by manual adjustment of the
window holders. The cell also allowed for placement of quartz or glass
windows to permit photodissociation by a laser beam propagated at right
angles to the IR probe beam. A magnetic bar was used to provide contin-
uous stirring, while inlet and outlet ports on each side of the cell allowed
4
A
H
R
U
G
3
H
5
3
A
H
R
U
3 4
H OH)] during 355 nm ir-
radiation of a solution of [Fe
hexane (35 mL) after 1hour.
3
Beer–Lambertꢁs law (Abs=ecl) was used for concentration determina-
tion. The vibrational band extinction coefficients (e) for the allyl alcohols
and aldehydes were determined experimentally in our laboratory by
using pure samples of each species as the calibrant. Since most of the
iron–carbonyl intermediates here have not been previously detected, we
have used the data for the analogous alkene species in which their extinc-
tion coefficients of vibrational bands have been determined (Table 1).
2
for N bubbling through the solution if required.
Infrared spectra during irradiation were obtained by using a Nicolet
Nexus 870 FTIR spectrometer operating in rapid scan mode (1000–
ꢀ
1
ꢀ1
4
000 cm , 2 cm resolution, 16 scans co-added for spectral averaging).
Spectral acquisition with the total collection time of 180 min at an inter-
val of 3 min was typically carried out. Difference spectra (solvent, allyl
alcohol, and metal precursor as background) were recorded so that the
changes in IR absorption with respect to the initial conditions could be
observed. Spectral depletion of the reactants was represented by the IR
bands pointing in one direction, whereas the production of newly-formed
species was shown by the IR bands pointing in the other (as shown in
Figure 1). Absolute spectra (only solvent as background) were scanned
so that the concentrations of all the detected species could be estimated.
In this case, all vibrational bands pointed in the same direction in the
spectrum (see Figure 2).
For example, we used the coefficients for [Fe(CO)
the concentration of [Fe(CO) (C OH)]. However, the absolute values
of e are not known for the p-allyl hydride species. As a first approxima-
tion, the band intensity was also based on the [Fe(CO) (C )] for
FeH(CO) (C OH)] species. Unfortunately the value for the phos-
phine-substituted species, [Fe(CO) (PPh (C OH)] was unavailable,
4
A
H
R
U
G
3 6
(C H )] to estimate
4
A
H
R
U
G
3 5
H
4
A
H
R
U
G
3 6
H
[
3
A
H
R
U
G
3 4
H
3
A
H
R
U
G
3
)
A
H
R
U
G
3 5
H
hence its concentration could not be estimated. Usually the area under
the curve for only one band of each species (the one of least spectral in-
terference from other species) was used for the concentration determina-
tion as well as monitoring the progress of the species.
b & c
0
.08
.04
Results and Discussion
0
b
c
A
H
R
U
G
c
Abs
wavelength laser pulses: The 532 nm photolysis of
Fe (CO) ] and allyl alcohol in hexane (35 mL) was first
[
3
12
-
0.04
a
carried out to verify that catalysis was indeed ocurring and
to characterize the main transition-metal–carbonyl products
formed. Figure 3 (top) shows the concentration profiles of
propionaldehyde and allyl alcohol during 532 nm photoisom-
-
-
-
0.08
0.12
0.16
ꢀ
5
ꢀ1
AHCTREUNG
a
3
5
ꢀ
5
ꢀ1
ꢁ
1.3”10 molmL
of propionaldehyde was produced
2
080
2040
Wavenumbers / cm
2000
1960
–1
under a catalytic loading of 0.5%. This was partly due to
the high vapor pressure of aldehyde; a certain amount of it
has escaped to fill the volume above the solution within the
cell. Other species might have formed, but since the detec-
tion of aldehyde in our case was primarily used to verify the
Figure 1. The difference IR bands of a) [Fe
3
(CO)12], b) [Fe(CO)
OH)] during 532 nm irradiation of
(CO)12] (2 mg) and allyl alcohol (1mL) in hexane (35 mL) after
hour.
5
], and c)
a solution of
[
[
1
Fe(CO)
Fe
4
A
C
H
T
R
E
U
N
G
(C
3
H
5
3
Chem. Eur. J. 2006, 12, 5128 – 5133
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5129

W. Y. Fan et al.
Table 1. Molecular species observed during the photoisomerization of allyl alcohol under various conditions.
[Fe (CO) ] to [Fe(CO) ] and
presumably other species in-
cluding [Fe(CO)4] and
3
12
5
ꢀ
1
Species
Observed IR Bands [cm ] and selected
Selected values
ꢀ1
ꢀ1 [a]
e in parenthesis [mLmg mm
]
from reference [1]
A
C
H
T
R
E
U
N
G
[Fe
Fe(CO)
Fe(CO)
3
(CO)12
]
2047 (4.95), 2028
2022 (4.90), 1999
–
[Fe(CO) ]. Since [Fe(CO)5]
has no absorptions in the green
region, the formation of
3
[
5
]
[Fe(CO)
[Fe(CO)
5
4
]
2023, 1996
2081, 2000, 1979
2064, 1994
–
–
–
–
2022, 1961, 1931
2052, 1979, 1946
–
–
[
4
A
C
H
T
R
E
U
N
G
(C
(C
(Me-C
(Me-C
(Ph-C
[FeH(CO) (Ph-C
(C
PPh
3
H
5
OH)]
OH)]
OH)]
2085, 2001, 1985
A
H
R
U
G
A
H
R
U
G
3
H
6
)]
A
C
H
T
R
E
U
N
G
[FeH(CO)
3
A
C
H
T
R
E
U
N
G
3
H
4
2062, 1989
2072, 2013, 1983
A
H
R
U
G
3
A
H
R
U
3
H
5
)]
[
Fe(CO) (C H OH)] was at-
A
T
E
N
[
Fe(CO)
4
3
H
4
4 3 5
A
C
H
T
R
E
U
N
G
[FeH(CO)
3
3
H
3
OH)] 2060
OH)] 2070, 2011, 1987
OH)] 2059, 1987
tributed to the coordination of
[2]
[
Fe(CO)
4
3
H
4
allyl alcohol with [Fe(CO) ].
4
A
C
H
T
R
E
U
N
G
3
3
H
4
The existence of [Fe(CO) ] in
4
[
[
Fe(CO)
Fe(CO)
3
A
C
H
T
R
E
U
N
G
3
H
5
OH)PPh
3
]
2022, 1958, 1930
2051, 1977, 1944
1741 (0.40)
[Fe(CO)
[Fe(CO)
3
4
A
H
R
U
G
2
H
4
)PPh
3
]
our catalytic system was fur-
ther supported with the results
obtained from visible irradia-
tion of a solution containing
4
3
]
PPh
3
]
propionaldehyde
allyl alcohol
1647 (0.026)
[
a] The values of e were assumed to be the same as the alkene iron carbonyl counterparts.
[
Fe (CO) ], allyl alcohol and a
3 12
trace amount of PPh3 in
hexane, in which [Fe(CO) PPh ] was detected.
Allyl Alcohol
Propionaldehyde
4
3
20
15
15
From Figure 3 (bottom) the gradual decay of [Fe (CO) ]
3
12
can be seen over the irradiation time. The IR bands of
Fe(CO) ] were found to evolve faster and achieved three to
[
5
four times higher concentration than those of [Fe(CO)4-
A
H
R
N
cies, we found that [Fe (CO) ] mostly dissociated to yield
5
0
3
12
[
Fe(CO) ] and [Fe(CO) (C H OH)] under 532 nm photoly-
A
H
R
U
G
5
4 3 5
sis. We also carried out photoisomerization reactions of
other allyl alcohols, such as methylallyl alcohol and phenyl-
allyl alcohol, and similar results were obtained.
0
500
1000 1500
2000 2500 3000 3500
4000
Time / s
Since the aldehyde has been shown to be generated cata-
lytically, the active species should be present in the reaction
[
Fe (CO)12]
3
[
Fe(CO)5]
mixture. As [Fe(CO) (C H OH)] does not absorb in the
532 nm region, it is unlikely that this species will further dis-
4 3 5
A
T
E
N
9
6
3
.0x10-8
[Fe(CO) (C H OH)]
4
3
5
sociate to yield [Fe(CO) ], presumably the active catalytic
3
[2,14]
species.
However the direct formation of [Fe(CO) ] from
3
.0x10-8
[
Fe (CO) ] might account for the production of the 16-elec-
3
12
tron [Fe(CO) (C H OH)] species, which in turn undergoes
A
H
R
U
G
3 3 5
.0x10-8
rapid intramolecular rearrangement to form the 18-electron
[6]
p-allyl iron hydride species [FeH(CO) (C H OH)] species.
A
T
E
N
3 3 4
Unfortunately we have failed to observe any IR signals at-
tributable to [FeH(CO)3(C H OH)]. This might be due to
the small absorption cross-section of [Fe (CO) ] producing
0
.0
ACHTREUNG
3
4
3
12
1x103
2x103
Time / s
3x103
4x103
0
only a small amount of intermediate species at each pulse.
Figure 3. Top: Concentration profiles of propionaldehyde and allyl alco-
hol during 532 nm photoisomerization of allyl alcohol in hexane with
Fe (CO) -catalyzed photoisomerization with 355 nm wave-
3
12
length laser pulses: When the rise of the aldehyde signal
was compared for the 532 and 355 nm cases, it appeared
that under similar experimental conditions the isomerization
proceeds about three times faster for the latter case (see
Figure 4, top). In addition, a new pair of signals at 2062 and
[
Fe
3
(CO)12]. Bottom: Concentration profiles of [Fe
(C OH)] under the same conditions.
3 5
(CO)12], [Fe(CO) ],
and [Fe(CO)
4
A
C
H
T
R
E
U
N
G
3 5
H
occurrence of catalysis, we did not attempt to detect other
products of this system.
ꢀ
1
1989 cm (shoulder peak) were observed, which we have
assigned to the p-allyl iron carbonyl hydride [FeH(CO) (R-
Figure 3 (bottom) shows the depletion of [Fe (CO) ] and
ACHTREUNG
3
12
3
the concurrent production of [Fe(CO) ] and [Fe(CO) -
C H OH)] (Figure 2). The two signals possessed the same
3 3
5
4
A
C
H
T
R
E
U
N
G
(C H OH)] as the irradiation progressed. The IR bands of
spectral evolution as the isomerization progressed, thus sup-
porting their assignment to a single species. IR assignments
3
5
[
Fe(CO) (C H OH)] species were identified based on the
A
C
H
T
R
E
U
N
G
4 3 5
similarity in the band positions with those belonging to
Fe(CO) (alkene)] species, such as [Fe(CO) (C H )]
of the [FeH(CO) (R-C H OH)] species were also based on
A
T
E
N
3 3 3
[
A
C
H
T
R
E
U
N
G
A
T
E
N
the similarity in the IR band positions with those found for
[FeH(CO) (C H )] species detected in a matrix as shown in
4
4
3
6
(
(
Table 1). The 532 nm irradiation even at higher energies
>20 mJ per pulse) resulted in a slow conversion of
ACHTREUNG
3
3
5
[1]
Table 1. As mentioned earlier, the concentration of this
5130
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 5128 – 5133

Photochemistry
FULL PAPER
sorption cross-section of [Fe (CO) ], thus yielding more in-
3
12
termediate species to effect the process. [Fe(CO) ] and
5
[
Fe(CO) (C H OH)] have also been detected in this system
A
H
R
U
G
4 3 5
and unlike visible light irradiation, photodissociation of
Fe(CO) ] and [Fe(CO) (C H OH)] could now take place
[
ACHTREUNG
5
4
3
5
upon 355 nm irradiation to produce [Fe(CO) ] containing
3
species in addition to its direct production from
[14]
[
Fe (CO) ]. This would provide one of the main factors
3 12
for the successful detection of [FeH(CO)3
this UV wavelength. Subsequent loss of CO from [Fe(CO)4-
(C H OH)] species upon further 355 nm photolysis might
A
H
R
U
G
AHCTREUNG
3
5
also enhance the production of [Fe(CO) (C H OH)] species,
A
T
E
N
3 3 5
which then rearrange to form more [FeH(CO)
molecules.
A
H
R
U
G
Figure 4 (bottom) shows the decay of [Fe (CO) ] after
3
12
1
hour of 355 nm photolysis, together with the concurrent
production of [Fe(CO) ], [Fe(CO)4(C H OH)], and
AHCTREUNG
5
3
5
[
FeH(CO) (R-C H OH)]. In contrast to the 532 nm case,
A
H
R
U
G
3 3 3
the concentration of [Fe(CO) ] was roughly equal to that of
5
[
Fe(CO) (C H OH)]; this fact indicates that the former spe-
A
T
E
N
4 3 5
cies undergoes further photolysis to yield the latter species
as one of its main products. However the amount of
[
Fe(CO) (C H OH)] did not increase indefinitely, since it
A
T
E
N
4 3 5
was also subjected to subsequent photolysis. As expected,
the reactive [FeH(CO) (C H OH)] species appeared in
AHCTREUNG
Figure 4. Top: Concentration profile of propionaldehyde and allyl alcohol
3
3
4
during 355 nm photoisomerization of allyl alcohol with [Fe
Bottom: Concentration profiles of [Fe (CO)12], [Fe(CO) ], [Fe(CO)
(C OH)], and [FeH(CO) (C OH)] under the same conditions.
3
(CO)12].
lower concentrations, at least three times less, throughout
the spectral-monitoring time although its concentration
gradually increased during the irradiation period. According
to many proposed mechanisms, the p-allyl hydride species
comes directly from the intramolecular rearrangement of
3
5
4
-
A
C
H
T
R
E
U
N
G
3
H
5
3
A
H
R
N
3 4
H
species was based on the extinction coefficient values de-
rived from [Fe(CO) (C H )]. Although the error associated
A
H
R
U
G
[Fe(CO) (C H OH)] (isomer A, bonded to C=C not directly
A
H
R
U
G
4
3
6
3 3 5
with such an estimate could be large, we believe that this
species was indeed present in very low concentrations be-
cause of the greater difficulties in its detection with respect
to [Fe(CO) (C H OH)]. When methylallyl alcohol or phe-
A
C
H
T
R
E
U
N
G
4 3 5
nylallyl alcohol were used as the substrate, the correspond-
ing p-allyl hydrides were also detected, but at lower signal-
to-noise ratios. Due to spectral overlap, the lower frequency
band of the p-methylallyl hydride species was not observa-
ble. Thus subsequent investigations of this species were con-
ducted using the parent allyl alcohol (C H OH) system
3
5
itself.
We have also investigated the possibility that the new sig-
nals might have come from a bis-allyl alcohol species such
attached to OH), while it decays by means of the reverse
process in which the hydride attached to the iron atom
transfers back to the allyl alcohol system (isomer B, C=C di-
rectly bonded to OH); see also Equation (1).
as [Fe (CO) (C H OH) ], since its alkene counterpart also
A
C
H
T
R
E
U
N
G
2
6 3 5 2
possesses two vibrational bands in the same region
ꢀ
1
[1]
(
Fe (CO) (C H ) , 2050, 1975 cm ). However the intensity
A
C
H
T
R
E
U
N
G
2
6 3 6 2
of the two detected bands were roughly equal in the spec-
trum and hence they clearly did not match those of the bis-
alkene species, for which the lower frequency vibration is at
least ten times more intense than the higher frequency vi-
bration. A detailed search of the literature did not reveal
any other matches for the bands observed here. Interesting-
ly, we found that these two bands also occurred in the [Fe-
k1
k2
½
FeðCOÞ ðC H OHފ ðAރ!HFeðCOÞ ðC H OHރ!
3
3
5
3
3
4
ð1Þ
½
FeðCOÞ ðC H OHފ ðBÞ
3
3
5
Although it is assumed that the p-allyl hydride species is
extremely short-lived, under certain conditions in which k <
k1, a buildup of this species could occur that would account
for its slight concentration increase towards the later part of
photolysis. According to computational models reported in
reference [6], the first and second steps of the reaction pos-
2
A
C
H
T
R
E
U
N
G
(CO) PPh ] catalytic system as will be discussed later.
4 3
For the 355 nm irradiation system, a much faster isomeri-
zation process was observed primarily due to the larger ab-
Chem. Eur. J. 2006, 12, 5128 – 5133
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5131

W. Y. Fan et al.
hn
ꢀ
1
sess Gibbs activation energies of ꢁ8 kcalmol
and
½FeðCOÞ
4
PPh
3
Šƒ!½FeðCOÞ
3
PPh
3
Š þ CO
ð2Þ
ð3Þ
ð4Þ
ꢀ
1
ꢁ
11 kcalmol respectively. (isomer A here is equivalent to
hn
½
½
½
FeðCOÞ PPh Šƒ!½FeðCOÞ Š þ PPh
4
3
4
3
structure 2b and isomer B is structure 4a in reference [6]).
This indicates that the conversion of p-allyl hydride to
isomer B is slightly hindered and thus would have increased
its lifetime.
hn
FeðCOÞ PPh Šƒ!½FeðCOÞ Š þ CO þ PPh
4
3
3
3
FeðCOÞ PPh Š þ C H OH ! ½FeðCOÞ PPh ðC H OHފ
3
3
3
5
3
3
3
5
ð5Þ
ð6Þ
[
Fe(CO) PPh ]-catalyzed photoisomerization with 355 nm
4 3
wavelength laser pulses: We have also extended our photo-
isomerization studies to the [Fe(CO) PPh ]-catalyzed
system. Figure 5 (top) shows the concentration profiles of
the allyl alcohol and aldehyde during 355 nm photoisomeri-
½
FeðCOÞ Š þ C H OH ! ½FeðCOÞ ðC H OHފ
4
3
5
4
3
5
ACHTREUNG
4
3
hn
½FeðCOÞ
PPh
ðC
H
5
OHފƒ!½FeHðCOÞ
ðC
H
OHފ þ PPh
3
3
3
3
3
4
3
ð7Þ
hn
½
FeðCOÞ PPh ðC H OHފƒ!½FeHðCOÞ PPh ðC H OHފ þ CO
3
3
3
5
2
3
3
4
ð8Þ
hn
½
½
½
FeðCOÞ ðC H OHފƒ!½FeHðCOÞ ðC H OHފ þ CO
ð9Þ
4
3
5
3
3
4
FeðCOÞ Š þ C H OH ! ½FeðCOÞ ðC H OHފ
ð10Þ
ð11Þ
3
3
5
3
3
5
FeðCOÞ ðC H OHފ ! ½FeHðCOÞ ðC H OHފ
3
3
5
3
3
4
Upon 355 nm photolysis of a hexane solution of [Fe-
A
H
R
U
G
4
3
[
Fe(CO)
A
H
R
U
G
A
T
E
N
3
3
5
3
observed, as shown in Figure 6. Figure 5 (bottom) shows the
0
0
.08
.04
a
b
b
a
a
Abs
–
–
–
–
–
0.04
0.08
0.12
0.16
0.20
c
c
c
Figure 5. Top: Concentration profiles of propionaldehyde and allyl alco-
hol during 355 nm photoisomerization of allyl alcohol with [Fe-
A
C
H
T
R
E
U
N
G
(CO)
Fe(CO)
conditions. The concentration axis only refers to that of [Fe(CO)
and [FeH(CO) (C OH)], since the extinction coefficient of
Fe(CO) PPh (C OH)] is not known.
4
PPh
3
]. Bottom: Concentration profiles of [Fe(CO)
(C OH)], and [FeH(CO) (C OH)] under the same
3
PPh ]
4
PPh
3
]
,
2080
2040
2000
Wavenumbers / cm
1960
1920
[
3
PPh
3
A
C
H
T
R
E
U
N
G
3
H
5
3
A
T
U
G
3
H
4
–1
4
A
C
H
T
R
E
U
N
G
3
H
4
Figure 6. Production of a) [Fe(CO)
3
PPh
3
A
H
R
U
G
3
H
5
OH)] and b) [FeH(CO)
PPh
3
3
-
]
3
[
3
3
A
C
H
T
R
E
U
N
G
3
H
5
A
T
E
N
3
H
4
4
(
10 mg) and allyl alcohol (1 mL) in hexane (35 mL). Spectrum shown was
recorded after 10 min of irradiation.
zation with [Fe(CO) PPh ] as the catalytic precursor. As ex-
concentration profiles of [Fe(CO) PPh ], [Fe(CO) PPh -
4
3
4
3
3
3
pected from previous literature, the catalytic activity was
about three to four times less effective compared to those
A
T
E
N
A
H
R
U
G
3
5
merization process. Both [Fe(CO) PPh (C H OH)] and
[FeH(CO) (C H OH)] species were found to be in relatively
A
H
E
N
3
3 3 5
[2]
catalyzed by [Fe (CO) ] under the same catalytic loading.
A
H
R
U
G
3
12
3 3 4
The photochemistry of [Fe(CO) PPh ] resembles that of
low concentrations throughout the isomerization process.
Attempts were made for the detection of both [Fe(CO)4-
4
3
[
Fe(CO) ] rather than [Fe (CO) ] for which UV light exci-
5 3 12
tation is required for CO-loss activation. A series of possible
pathways [Eqs. (2)–(11)] involving detachment of either
A
T
E
N
A
H
E
N
3
5
2
their vibrational bands were not seen despite changing the
allyl alcohol concentrations, increasing the laser energy, or
adding free phosphines into the reaction mixture. The detec-
PPh or CO ligand could be proposed for this system.
3
5132
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 5128 – 5133

Photochemistry
FULL PAPER
tion of [Fe(CO) PPh
A
H
R
U
G
copy under 355 nm photolysis of [Fe (CO) ] and [Fe-
3
3
3
5
3
12
suggested that the reaction shown in Equation (5) is prefer-
red over reaction that in Equation (6), and this in turn sug-
gested that the detachment of CO [Eq. (2)] is more likely to
A
T
E
N
ous literature reports have indirectly pointed towards the
existence of this species by examining the nature of the final
products, its direct detection here lends good support to the
p-allyl mechanism operating for the isomerization of allyl al-
cohols as opposed to the metal–hydride addition–elimina-
occur than the detachment of PPh [Eq. (3)]. However, the
3
sensitivity of FTIR monitoring is not high enough to prove
the complete absence of these two reactions [Eqs. (3) and
[8,10]
(
6)].
At first it seems surprising that we observed the same two
vibrational bands attributed to [FeH(CO)3(C H OH)] in this
tion mechanism.
AHCTREUNG
3
4
phosphine-containing catalytic system. It appears that a
common iron allyl hydride species acts as the key intermedi-
ate in these two systems. The simplest explanation to ac-
Acknowledgements
T.S.C. thanks the Institute of Chemical Engineering and Sciences (ICES),
ASTAR Singapore for a research scholarship. This work is supported
under an NUS grant (143-000-210-112).
count for the formation of [FeH(CO) (C H OH)] is due to
A
H
R
U
G
3 3 4
the production of [Fe(CO) ] formed from a direct dissocia-
3
tion of two ligands (CO and PPh ) from [Fe(CO) PPh ],
3
4
3
since the photolysis of [Fe(CO) ] itself is also known to pro-
5
[1] J. C. Mitchener, M. S. Wrighton, J. Am. Chem. Soc. 1983, 105, 1065.
[2] J C. Mitchener, M. S. Wrighton, J. Am. Chem. Soc. 1981, 103, 975.
duce [Fe(CO) ]. If it were the major pathway, it would ex-
3
[
3] A. J. Ouderkirk, P. Wermer, N. L. Schultz, E. Weitz, J. Am. Chem.
Soc. 1983, 105, 3354.
plain the presence of the hydride species in this phosphine-
containing system as well.
However there is an alternate way to consider how the
[
[
4] N. Iranpoor, E. Mottaghinejad, J. Organomet. Chem. 1992, 423, 399.
5] D. B. Chase, F. J. Weigert, J. Am. Chem. Soc. 1981, 103, 977.
hydride species could be formed. Since [Fe(CO)4
was not detected in this system or is present only as a minor
component, the possibility of forming [FeH(CO) (C H OH)]
A
H
R
U
G
[6] V. Branchadell, C. Crevisy, R. Gree, Chem. Eur. J. 2003, 9, 2062.
[7] C. Crevisy, M. Wietrich, V. Le Boulaire, R. Uma, R. Gree, Tetrahe-
dron Lett. 2001, 42, 395.
3
5
AHCTREUNG
3
3
4
[
8] R. C. van der Drift, E. Bouwman, E. Drent, J. Organomet. Chem.
002, 650, 1 .
from [Fe(CO) (C H OH)] would have to be excluded
A
C
H
T
R
E
U
N
G
4 3 5
2
[
Eq. (9)]. The reaction shown in Equation (7), however, re-
mains a possible pathway for generating the hydride species,
since [Fe(CO) PPh (C H OH)] has been spectrally ob-
[
9] R. Uma, N. Gouault, C. Crevisy, R. Gree, Tetrahedron Lett. 2003,
44, 6187.
[
[
[
10] R. Uma, C. Crevisy, R. Gree, Chem. Rev. 2003, 103, 27.
11] H. Cherkaoui, M. Soufiaoui, R. Gree, Tetrahedron 2001, 57, 2379.
12] J. L. Graff, R. D. Sanner, M. S. Wrighton, Organometallics 1982, 1,
AHCTREUNG
3
3
3
5
served. The release of a bulky group like PPh instead of a
3
CO ligand upon photolysis is quite a common occurrence in
metal–phosphine systems as it places less constraint on
steric congestion upon intramolecular hydrogen migration.
We have shown that the p-allyl iron carbonyl hydride spe-
8
37.
[13] T. S Chong, P. Li, W. K. Leong, W. Y. Fan, J. Organomet. Chem.
2005, 690, 4132.
14] P. Portius, J. Yang, X. Sun, D. C. Grills, P. Matousek, A. W. Parker,
[
M. Towrie, M. W. George, J. Am. Chem. Soc. 2004, 126, 10713.
cies [FeH(CO) (C H OH)] is present and most probably
A
C
H
T
R
E
U
N
G
3 3 4
takes part in the catalytic cycle for all the three systems in-
vestigated here. It was detected directly using FTIR spectros-
Received: January 14, 2006
Published online: April 27, 2006
Chem. Eur. J. 2006, 12, 5128 – 5133
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5133