Photochemistry of (η5-C5H5)(η5-C4H4N)Fe and (η5-C5H5)(η1-N-C4H4N)Fe(CO)2 in low-temperature matrixes and room-temperature solution. Evidence for a photoinduced haptotropic shift of the π-coordinated pyrrolyl ligand

Article abstract of DOI:10.1021/om0002632

Azaferrocene, (η⁵-C₅H₅)(η⁵-C₄H₄N)Fe, undergoes a η⁵→η¹ haptotropic shift of the pyrrolyl ligand upon long-wavelength photolysis (λ_exc>495 nm) both in alkane solvents at room temperature and in frozen matrixes at 12 K. Room-temperature photolysis (λ_exc>495 nm) in CO-saturated cyclohexane solution generated (η⁵-C₅H₅) (η¹-N-C₄H₄N)Fe(CO)₂. Irradiation with λ_exc = 532 nm also produced an allyl monocarbonyl species, exo-(η⁵-C₅H₅)(η³-C-C₄H₄N)Fe(CO), identified by IR spectroscopy. In CO-doped matrixes at 12 K both (η⁵-C₅H₅)(η¹-N-C₄H₄N)Fe(CO) and (η⁵-C₅H₅)(η¹-N-C₄H₄N)Fe(CO)₂ are formed following broad-band irradiation (λ_exc>495 nm) of (η⁵-C₅H₅)(η⁵-C₄H₄N)Fe, in a ratio dependent on the concentration of CO in the matrix. Initial irradiation with λ_exc = 538 nm followed by broad-band photolysis (λ_exc>495 nm) in CO-doped matrixes formed additional monocarbonyl species, exo-(η⁵-C₅H₅) (η³-C-C₄H₄N)Fe(CO), and a species absorbing at 1962 cm^-1, which is either the appropriate endo-isomer or aza-allyl species. Laser flash photolysis experiments of (η⁵-C₅H₅)(η¹-N-C₄H₄N)Fe(CO)₂ in either CO-saturated cyclohexane or toluene produced (η⁵-C₅H₅)(η¹-N-C₄H₄N)Fe(CO), which reacted with CO with rate constants measured at 298 K of (3.0±0.3)×10⁸ and (3.3±0.3)×10⁸ M^-1 s^-1, respectively, regenerating (η⁵-C₅H₅)(η¹-N-C₄H₄N)Fe(CO)₂.

Full text of DOI:10.1021/om0002632

Organometallics 2000, 19, 3867-3873
3867
P h otoch em istr y of (η5-C5H5)(η5-C4H4N)F e a n d
(η5-C5H5)(η1-N-C4H4N)F e(CO)2 in Low -Tem p er a tu r e
Ma tr ixes a n d Room -Tem p er a tu r e Solu tion . Evid en ce for
a P h otoin d u ced Ha p totr op ic Sh ift of th e π-Coor d in a ted
P yr r olyl Liga n d
Davnat P. Heenan,† Conor Long,† Virginia Montiel-Palma,‡
Robin N. Perutz,‡ and Mary T. Pryce*,†
Inorganic Photochemistry Centre, Dublin City University, Dublin 9, Ireland, and Department
of Chemistry, University of York, Heslingon, York YO10 5DD, U.K.
Received March 27, 2000
Azaferrocene, (η5-C5H5)(η5-C4H4N)Fe, undergoes a η5 f η1 haptotropic shift of the pyrrolyl
ligand upon long-wavelength photolysis (λexc > 495 nm) both in alkane solvents at room
temperature and in frozen matrixes at 12 K. Room-temperature photolysis (λexc > 495 nm)
in CO-saturated cyclohexane solution generated (η5-C5H5)(η1-N-C4H4N)Fe(CO)2. Irradiation
with λexc ) 532 nm also produced an allyl monocarbonyl species, exo-(η5-C5H5)(η3-C-C4H4N)-
Fe(CO), identified by IR spectroscopy. In CO-doped matrixes at 12 K both (η5-C5H5)(η1-N-
C4H4N)Fe(CO) and (η5-C5H5)(η1-N-C4H4N)Fe(CO)2 are formed following broad-band irradia-
tion (λexc > 495 nm) of (η5-C5H5)(η5-C4H4N)Fe, in a ratio dependent on the concentration of
CO in the matrix. Initial irradiation with λexc ) 538 nm followed by broad-band photolysis
(λexc > 495 nm) in CO-doped matrixes formed additional monocarbonyl species, exo-(η5-C5H5)-
(η3-C-C4H4N)Fe(CO), and a species absorbing at 1962 cm-1, which is either the appropriate
endo-isomer or aza-allyl species. Laser flash photolysis experiments of (η5-C5H5)(η1-N-C4H4N)-
Fe(CO)2 in either CO-saturated cyclohexane or toluene produced (η5-C5H5)(η1-N-C4H4N)Fe-
(CO), which reacted with CO with rate constants measured at 298 K of (3.0 ( 0.3) × 108
and (3.3 ( 0.3) ×108 M-1 s-1, respectively, regenerating (η5-C5H5)(η1-N-C4H4N)Fe(CO)2.
In tr od u ction
at 10 K produced (η1-C5H5N)Cr(CO)5, while irradiation
at 308 nm also produced the more familiar CO-loss
The photochemical reactivity of half-sandwich com-
pounds of the type (ηx-aromatic ligand)M(CO)y (x ) 4,
5, or 6; y ) 2, 3, or 4; M ) V, Cr, Mn, or Co) can differ
from the thermal chemistry.1,2 For example, (η6-C6H6)-
Cr(CO)3 undergoes thermally induced arene exchange,
while CO loss is the dominant photochemical process.3,4
Arene exchange proceeds via a thermally induced hap-
totropic shift of the π-ligand,3 but to date the only report
of a photoinduced haptotropic shift for these systems
comes from our investigations into the photochemistry
of π-coordinated pyridine compounds.5 This work dem-
onstrated that these processes occur following low-
energy photolysis. For instance, photolysis of (η6-C5H5N)
Cr(CO)3 (λexc ) 460 nm) in a CO-doped methane matrix
product (η6-C5H5N)Cr(CO)2.
We have extended these studies to sandwich com-
pounds related to ferrocene. Ferrocene is photoinert in
dry alkane solvents, but it readily photooxidizes in
haloalkanes.6 Quantum yields for photodecomposition
of ferrocene of up to unity are obtained in CCl4/EtOH
mixtures.6 Substituted ferrocenes have been reported
to photodecompose in solvents such as acetonitrile,
methanol, DMSO, decalin, chloroform, or carbontetra-
chloride; the nature of this decomposition remains
uncertain, however.7 In all cases products are complex,
and more recent work has suggested that the presence
of traces of water is a prerequisite for photoactivity in
some solvents.8
In this paper, we describe the photochemistry of
azaferrocene, (η5-C5H5)(η5-C4H4N)Fe, the closest and
most accessible heterocyclic analogue of ferrocene. The
reason for this study is primarily to determine if the
photoinduced hapticity changes observed for π-coordi-
nated pyridine, are also observed for π-coordinated
† Dublin City University.
‡ University of York.
(1) Basolo, F. New J . Chem. 1994, 18, 19.
(2) O’Connor, J . M.; Casey, C. D. Chem. Rev. 1987, 87, 307.
(3) (a) Trayler, T. G.; Stewart, K. J .; Goldberg, M. J . J . Am. Chem.
Soc. 1984, 106, 4445. (b) Trayler, T. G.; Stewart, K. J .; Goldberg, M.
J . Organometallics 1986, 5, 2062. (c) Trayler, T. G.; Stewart, K. J . J .
Am. Chem. Soc. 1986, 108, 6977. (d) Trayler, T. G.; Goldberg, M. J .
Organometallics 1987, 6, 2413. (e) Trayler, T. G.; Goldberg, M. J .
Organometallics 1987, 6, 2531.
(4) (a) Strohmeier, S.; von Hobe, D. Z. Naturforsch. 1963, 18B, 770.
(b) Wrighton, M. S.; Haverty, J . L. Z. Naturforsch. 1975, 30B, 254. (c)
Nasielski, J .; Denishoff, O. J . Organomet. Chem. 1975, 102, 65. (d)
Gilbert, A.; Kelly, J . M.; Budzwait, M.; Koerner von Gustorf, E. Z.
Naturforsch. 1976, 31B, 1091. (e) Rest, A. J .; Sodeau, J . R.; Taylor, D.
J . J . Chem. Soc., Dalton Trans. 1978, 651.
(6) Traverso, O.; Scandola, F. Inorg. Chim. Acta 1970, 4, 493. (b)
Tarr, A. M.; Wiles, D. M. Can. J . Chem. 1968, 46, 2725. (c) von Gustorf,
E. K.; Koller, H.; J un, M.-J .; Schenck, G. O. Chem.-Ing-Tech. 1963,
35, 591.
(7) (a) Ali, L. H.; Cox, A.; Kemp, T. S. J . Chem. Soc., Dalton Trans.
1973, 1468. (b) Traverso, O.; Rossi, R.; Sostero, S.; Carassiti, V. Mol.
Photochem. 1973, 5, 457.
(5) Breheny, C. J .; Draper, S. M.; Grevels, F.-W.; Klotzbu¨cher, W.
E.; Long, C.; Pryce, M. T.; Russell, G. Organometallics 1996, 15, 3679.
(8) Davis, J .; Vaughan, D. H.; Cardosi, M. F. Electroanalysis 1997,
9, 650.
10.1021/om0002632 CCC: $19.00 © 2000 American Chemical Society
Publication on Web 08/17/2000

3868 Organometallics, Vol. 19, No. 19, 2000
Heenan et al.
Ta ble 1. In fr a r ed Ba n d s Obser ved in th e CO
Str etch in g Region for (η5-C5H5)F e(CO)2X
Com p ou n d s a n d Th eir P h otop r od u cts in Va r iou s
Low -Tem p er a tu r e Ma tr ixes a n d
Room -Tem p er a tu r e Solu tion
complex
νCO (cm-1
)
medium
(η5-C5H5)(η1-C4H4N)Fe(CO)2
(η5-C5H5)(η1-C4H4N)Fe(CO)2
(η5-C5H5)(η1-C4H4N)Fe(CO)2
(η5-C5H5)(η1-C4H4N)Fe(CO)
(η5-C5H5)(η1-C4H4N)Fe(CO)
(η5-C5H5)Fe(CO)2Cla
2053, 2007
2052, 2006
2048, 2002
1974
Ar matrix
N2 matrix
C6H12
Ar matrix
N2 matrix
CH4 matrix
N2 matrix
CH4 matrix
N2 matrix
C6H12
CCl4
thin film
C6H12
CO/Ar matrix
CO/Ar matrix
1974
2054, 2010
2056, 2013
1977
1979
2038, 2005
1950
1949
ca. 1949
1948
(η5-C5H5)Fe(CO)2Cla
(η5-C5H5)Fe(CO)Cla
F igu r e 2. Spectral changes observed in the carbonyl
region following photolysis (λexc > 500 nm) of (η5-C5H5)(η5-
C4H4N)Fe in CO-saturated cyclohexane at room tempera-
ture, showing formation of (η5-C5H5)(η1-N-C4H4N)Fe(CO)2.
(η5-C5H5)Fe(CO)Cla
(η5-C5H5)Fe(CO)2Ib
(η5-C5H5)(η3-C3H5)Fe(CO)c
(η5-C5H5)(η3-C4H7)Fe(CO)c
(η5-C5H5)(η3-C-C4H4N)Fe(CO)
(η5-C5H5)(η3-C-C4H4N)Fe(CO)
(η5-C5H5)(η3-N-C4H4N)Fe(CO)
Resu lts
1962
The IR spectroscopic data for (η5-C5H5)(η5-C4H4N)Fe
and (η5-C5H5)(η1-N-C4H4N)Fe(CO)2 and their photoprod-
ucts are given in Table 1, while the UV/vis spectra of
the starting materials in cyclohexane are represented
in Figure 1. Both compounds exhibit weak absorptions
in the visible region. The spectrum of (η5-C5H5)(η5-
C4H4N)Fe in cyclohexane contains features centered at
326 nm (ꢀ ) 46 M-1 cm-1) and 423 nm (ꢀ ) 38 M-1
cm-1), while (η5-C5H5)(η1-N-C4H4N)Fe(CO)2 has an ab-
sorption at 340 nm (ꢀ ) 142 M-1 cm-1) and a weaker
feature at 484 nm (ꢀ ) 35 M-1 cm-1).
a
b
Reference 17. Reference 10. c Reference 11.
Stea d y-Sta te P h otolysis Exp er im en ts
Br oa d -Ba n d P h ot olysis of (η5-C5H5)(η5-C4H4N)-
Fe in CO-Satu r ated Cycloh exan e Solu tion at Room
Tem p er a tu r e. Visible photolysis of a CO-saturated
solution of (η5-C5H5)(η5-C4H4N)Fe in cyclohexane with
broad-band radiation from a xenon-arc lamp (275 W)
fitted with a filter that only transmits wavelengths
greater than 500 nm produced (η5-C5H5)(η1-N-C4H4N)-
Fe(CO)2 (νCO ) 2048 and 2002 cm-1, Table 1; reaction
1; Figure 2).9 Experiments using a λexc > 400 nm filter
F igu r e 1. UV/vis spectra for (η5-C5H5)Fe(η5-C5H5) (1.79
× 10-2 M; spectrum A) and (η5-C5H5)(η1-N-C4H4N)Fe(CO)2
(5.91 × 10-3 M; spectrum B) in cyclohexane.
pyrroles. The presence of the nitrogen atom in the
π-coordinated ligand has a number of effects. The
proximate nitrogen can act as an intramolecular “trap”
for ring-slip intermediates. More fundamentally, its
presence reduces the symmetry of the molecule. As a
consequence, the degeneracy of valence orbitals, a
feature of the homocyclic analogues, is lifted. This
reduces the energy of low-energy excited states, permit-
ting access to novel photochemical processes, such as
haptotropic shift reactions.
The experimental results reported here were obtained
using a combination of room-temperature techniques in
dry cyclohexane solution and low-temperature matrix
isolation methods. Evidence is presented to demonstrate
that, in common with η6-coordinated pyridine, the η5-
pyrrolyl ligand also undergoes facile photoinduced hap-
totropic shifts, opening multiple coordination sites on
the metal center. Moreover, experiments using mono-
chromatic irradiation indicate that the haptotropic
shifts may occur by a stepwise η5 f η3 f η1 transforma-
tion involving the absorption of two photons. To assist
with the assignment of intermediate species observed
in these experiments, the photochemistry of (η5-C5H5)-
(η1-N-C4H4N)Fe(CO)2 was investigated by flash pho-
tolysis techniques in solution and also by matrix isola-
tion.
also produced the dicarbonyl species as the major
product, but small amounts of [(η5-C5H5)Fe(CO)2]2 were
formed in addition (νCO ) 1960 and 1793 cm-1; the high-
energy band at 2002 cm-1 is obscured by the low-energy
absorption of (η5-C5H5)(η1-N-C4H4N)Fe(CO)2). In all
cases the yield of the dinuclear species appeared low
(based on IR band intensities), although it was greatest
in experiments where the solution was simply flushed
with CO rather than rigorously degassed by freeze-
pump-thaw procedures followed by admission of CO to
the solution cell. It should be noted that formation of
(9) (a) We have confirmed that this reaction is photochemical by
control experiments in which the solution of (η5-C5H5)(η5-C4H4N)Fe in
CO-saturated cyclohexane was kept in the dark; no (η5-C5H5)(η1-N-
C4H4N)Fe(CO)2 was formed under these conditions. (b) Martin, K. F.;
Hanks, T. W. Organometallics 1997, 16, 4857.

Photochemistry of (η5-C5H5)(η5-C4H4N)Fe
Organometallics, Vol. 19, No. 19, 2000 3869
the dinuclear species is also observed following photoly-
sis of (η5-C5H5)Fe(CO)2I in alkane solvents.10
Br oa d -Ba n d P h ot olysis of (η5-C5H5)(η5-C4H4N)-
F e in Dega ssed Cycloh exa n e Solu tion a t Room
Tem p er a tu r e. Broad-band irradiation (λexc > 500 nm)
of a concentrated solution of (η5-C5H5)(η5-C4H4N)Fe in
degassed cyclohexane caused the solution to become
turbid and eventually produced a precipitate. After the
admission of CO, the precipitate slowly disappeared and
the solution clarified. Monitoring of the solution by
infrared spectroscopy indicated the formation (over
several hours at room temperature) of bands at 2048
and 2002 cm-1, assigned to (η5-C5H5)(η1-N-C4H4N)Fe-
(CO)2.
The observation of solution turbidity following ir-
radiation in cyclohexane renders time-resolved studies
on this system both difficult and prone to artifacts
resulting from photoacoustic effects. As a consequence,
the photochemistry of azaferrocene was not investigated
by time-resolved techniques.
F igu r e 3. Infrared spectrum recorded after monochro-
matic irradiation (λexc ) 532 nm) of (η5-C5H5)(η5-C4H4N)-
Fe in CO-saturated cyclohexane at room temperature.
CpFe(CO)2 ) (η5-C5H5)(η1-N-C4H4N)Fe(CO)2; CpFe(CO)(η3-
pyrrolyl) ) (η5-C5H5)(η3-C-C4H4N)Fe(CO).
Mon och r om a tic P u lsed P h otolysis of (η5-C5H5)-
(η5-C4H4N)F e in CO-Sa tu r a ted Cycloh exa n e Solu -
tion a t Room Tem p er a tu r e. Pulsed photolysis of (η5-
C5H5)(η5-C4H4N)Fe in CO-saturated cyclohexane with
λexc ) 355 nm produced (η5-C5H5)(η1-N-C4H4N)Fe(CO)2
as the sole carbonyl-containing product. However, pulsed
photolysis using λexc ) 532 nm generated an additional
species, with a single νCO band at 1949 cm-1 (Figure 3).
We have assigned this band to (η5-C5H5)(η3-C-C4H4N)-
Fe(CO) (reaction 2), as it is close to that observed for
(η5-C5H5)(η3-C3H5)Fe(CO) or (η5-C5H5)(η3-C4H7)Fe(CO)
(νCO ) 1950 or 1948 cm-1 respectively) in CCl4 or as a
thin film.11
F igu r e 4. Infrared spectrum recorded following photolysis
(λexc > 500 nm) of (η5-C5H5)(η5-C4H4N)Fe in 0.5% CO/Ar
matrix at 12 K. CpFe(CO)2 ) (η5-C5H5)(η1-N-C4H4N)Fe-
(CO)2; CpFe(CO)(eta-1-pyrrolyl) ) (η5-C5H5)(η1-N-C4H4N)-
Fe(CO).
and the other with a single band at 1974 cm-1 (Figure
4).12 The ratio of the intensity of these product bands
varied with the concentration of CO in the matrix. For
instance, the band at 1974 cm-1 band was 4 times as
intense as that at 2053 and 2008 cm-1 for a matrix
containing 0.5% CO, while it was only twice as large in
a 2% CO matrix. Within individual experiments, this
ratio remained constant.
The species absorbing at 2053 and 2008 cm-1 was
assigned to (η5-C5H5)(η1-N-C4H4N)Fe(CO)2 by compari-
son with the infrared spectrum of an authentic sample
of (η5-C5H5)(η1-N-C4H4N)Fe(CO)2 in the same matrix (cf.
Table 1).13 The band at 1974 cm-1 was assigned to the
16-electron (η5-C5H5)(η1-N-C4H4N)Fe(CO) species (a
strong absorption at 415 nm, a shoulder at 317 nm, and
a weak broad feature at ca. 730-750 nm correlate with
this IR band), and support for this assignment came
from matrix experiments on (η5-C5H5)(η1-N-C4H4N)Fe-
(CO)2, which produced (η5-C5H5)(η1-N-C4H4N)Fe(CO)
and free CO on photolysis (see below). The band at 1975
cm-1 is asymmetric, with two further unresolved com-
ponents at 1962 and 1948 cm-1 (see below for assign-
ments).
Ma tr ix Isola tion Stu d ies
P h otolysis (λexc > 495 n m ) of (η5-C5H5)(η5-C4H4N)-
F e in Ar gon Ma tr ixes a t 12 K Con ta in in g Eith er
0.5% or 2% CO. Azaferrocene exhibits UV/vis bands at
325 and 424 nm in a CO-doped argon matrix. Photolysis
(12) The 1975 cm-1 band also exhibited shoulders to the low-energy
side at ca. 1962 and 1948 cm-1, possibly indicating the formation of
further carbonyl-containing species.
of (η5-C5H5)(η5-C4H4N)Fe in these matrixes with λexc
>
495 nm produced two species, one with two νCO bands
(13) It should be noted that the values for the νCO bands for (η5-
C5H5)(η1-N-C4H4N)Fe(CO)2 quoted in the literature are incorrect. We
have confirmed our values by correlating our infrared spectral data
obtained from a single sample of the compound with the results of high-
resolution mass spectrometry and single-crystal X-ray diffraction
techniques.
of approximately equal intensity (2053 and 2008 cm-1
)
(10) Pan, X.; Philbin, C. E.; Castellani, M. P.; Tyler, D. R. Inorg.
Chem. 1988, 27, 671.
(11) Green, M. L. H.; Nagy, P. L. I. J . Chem. Soc. 1963, 189.

3870 Organometallics, Vol. 19, No. 19, 2000
Heenan et al.
F igu r e 5. Infrared spectra following photolysis of (η5-
C5H5)(η5-C4H4N)Fe in a 2% CO/Ar matrix at 12 K (a) at λ
exc ) 538 nm for 200 min (s) and (b) subsequent photolysis
at λexc > 495 nm (- - -).
Sch em e 1. Br oa d -Ba n d Ma tr ix P h otolysis of
(η5-C5H5)(η5-N-C4H4N)F e a n d
(η5-C5H5)(η1-N-C4H4N)F e(CO)2
F igu r e 6. UV/vis spectra following photolysis of (η5-C5H5)-
(η5-C4H4N)Fe in a 2% CO/Ar matrix at 12 K, (a) following
deposition (- - -), (b) at λ exc ) 538 nm (200 min; ‚‚‚), and (c)
subsequent photolysis at λexc > 495 nm (30 min; s).
In itia l P h otolysis (λexc > 495 n m ) of (η5-C5H5)(η5-
C4H4N)F e in Ar gon Ma tr ixes a t 12 K, Con ta in in g
Eith er 0.5% CO or 2% CO, a n d Su bsequ en t Br oa d -
Ba n d P h otolysis (λexc > 325 n m ). Initial photolysis
of an argon matrix containing (η5-C5H5)(η5-C4H4N)Fe
again produced the two carbonyl-containing species as
outlined above. Changing to shorter wavelengths re-
sulted in a change in ratio of the product bands, the
intensity of the band at 1974 cm-1 increasing more
rapidly than those at 2053 and 2008 cm-1. This suggests
that the dicarbonyl product is undergoing photoinduced
CO loss at these excitation wavelengths (Scheme 1).
This was confirmed by matrix and time-resolved experi-
ments on (η5-C5H5)(η1-N-C4H4N)Fe(CO)2 (see below).
Mon och r om a tic P h otolysis (λexc ) 538 n m ) of (η5-
C5H5)(η5-C4H4N)F e in Ar gon Ma tr ixes a t 12 K
Con ta in in g 2% CO a n d Su bsequ en t Br oa d -Ba n d
P h otolysis w ith λ exc > 495 n m . In contrast with the
matrix experiments described above, where samples
were irradiated with broad-band light sources, mono-
chromatic irradiation of (η5-C5H5)(η5-C4H4N)Fe in a CO-
doped argon matrix with λexc ) 538 nm produced four
carbonyl-containing species (Figure 5a). In addition to
(η5-C5H5)(η1-N-C4H4N)Fe(CO)2 and (η5-C5H5)(η1-N-C4-
H4N)Fe(CO) previously observed, further species with
νCO ) 1962 and 1948 cm-1 were formed, the identity of
which will be discussed later. Subsequent irradiation
of this matrix with λexc > 495 nm resulted in a sub-
stantial increase in the intensity of the bands at 2053,
2008, 1962, and 1948 cm-1, while the intensity of the
band assigned to (η5-C5H5)(η1-N-C4H4N)Fe(CO) de-
creased slightly (Figure 5b). This experiment and
related ones showed a feature at 660 nm in the UV/vis
spectrum (Figure 6) which is associated with the species
absorbing at 1962 and 1948 cm-1 in the infrared
spectrum.
F igu r e 7. IR difference spectrum obtained following
photolysis (λexc > 495 nm) of (η5-C5H5)(η5-C4H4N)Fe in an
argon matrix at 12 K.
P h otolysis of (η5-C5H5)(η5-C4H4N)F e in a n Ar gon
or a Nitr ogen Ma tr ix a t 12 K. The spectroscopic
changes observed following photolysis of (η5-C5H5)(η5-
C4H4N)Fe (λexc > 495 nm) in an argon matrix are
presented in Figure 7. We have assigned these changes
to the depletion of (η5-C5H5)(η5-C4H4N)Fe and the
formation of a ring-slip product. The nature of the
coordination mode of the pyrrolyl ligand is uncertain.
However ab initio calculations on (η5-C5H5)(η1-N-C4H4N)-
Fe were undertaken to estimate the infrared properties
of this 14-electron species. The starting geometry was
based on the molecular structure published for (η5-
C5H5)(η1-N-C4H4N)Fe(CO)2 but with both carbonyl
14
ligands removed from the atom list. A full geometry
optimization was undertaken at B3LYP/LANL2DZ model
chemistry using the Gaussian-98 program suite.15 The
Hessian matrix, again calculated using the same model
chemistry, revealed only one trivial negative frequency
(-55 cm-1) associated with the rotation of the (η5-C5H5)
ligand. The results of these calculations together with
the observed band positions are presented in Table 2.16
On the basis of these results we can assign the spectral
features observed in these matrix experiments to (η5-
C5H5)(η1-N-C4H4N)Fe. Photolyses in a dinitrogen matrix
(14) Powell, M. A.; Bailey, R. D.; Eagle, C. T.; Schimel, G. L.; Hanks,
T. W.; Pennington, W. T. Acta Crystallogr. 1997, C53, 1611.

Photochemistry of (η5-C5H5)(η5-C4H4N)Fe
Organometallics, Vol. 19, No. 19, 2000 3871
Ta ble 2. Obser ved (for p h otolysis p r od u ct of
a za fer r ocen e in a r gon m a tr ixes) a n d Ca lcu la ted
Ba n d P osition s for (η5-C5H5)(η1-N-C4H4N)F e Ba sed
on a n Op tim ized Geom etr y a t B3LYP /LANL2DZ
Mod el Ch em istr ya
observed band
positions (cm-1
calculated band
)
positions (cm-1
)
719
772
715
749
781
763
788
777
831
834
1035
1009
1012
1023
1049
1055
1142
1154
1440
1079
1167
1464
F igu r e 8. UV/vis difference transient absorption spectrum
obtained after 2 µs following photolysis of (η5-C5H5)(η1-N-
C4H4N)Fe(CO)2 at λexc ) 355 nm in cyclohexane under 1
atm of CO. The inset shows a typical transient signal
observed at 420 nm.
a
The calculated values have been corrected using an empirical
Sch em e 2. F la sh P h otolysis of
(η5-C5H5)(η1-N-C4H4N)F e(CO)2 Resu ltin g in
Deca r bon yla tion F ollow ed by Ra p id
Recom bin a tion w ith CO
factor of 0.98 (see ref 16).
gave rise to spectral changes, essentially identical to
those observed in an argon matrix.
P h otolysis of (η5-C5H5)(η1-N-C4H4N)F e(CO)2 in
Ar gon a n d Din itr ogen Ma tr ixes a t 12 K. Photolysis
(λexc > 400 nm) of (η5-C5H5)(η1-N-C4H4N)Fe(CO)2 in an
argon matrix at 12 K resulted in the depletion of the
parent bands (2053 and 2007 cm-1) and the formation
of new bands at 2134 and 1974 cm-1 (the 1974 cm-1
band exhibited shoulders on the low-energy side).12
These bands are assigned to free CO and (η5-C5H5)(η1-
N-C4H4N)Fe(CO), respectively. The spectral changes
observed in a dinitrogen matrix were identical to those
in an argon matrix. This suggests that (η5-C5H5)(η1-N-
C4H4N)Fe(CO) does not react with dinitrogen. The
matrix photochemistry of (η5-C5H5)Fe(CO)2Cl reported
by Rest and co-workers17 also identified CO loss as the
dominant photochemical process, and the resulting
monocarbonyl (η5-C5H5)Fe(CO)Cl species was inert to
dinitrogen (exhibiting similar band positions in both
methane and nitrogen matrixes, νCO ) 1977 and 1979
cm-1, respectively).
and a difference spectrum are given in Figure 8. Under
these conditions, this species decays to the preirradia-
tion baseline in less than 2 µs, indicating that the
process is reversible. The steady-state UV/vis spectrum
of the photolysis solution was monitored throughout the
experiment and showed no significant change, confirm-
ing that the overall process is reversible.
A plot of the observed rate constant (kobs) for the decay
of the transient species against [CO] was linear. The
slope of this plot gave the second-order rate constant
(k2) of (3.0 ( 0.3) × 108 M-1 s-1 at 298 K for the reaction
of (η5-C5H5)(η1-N-C4H4N)Fe(CO) with CO. Flash pho-
tolysis in CO-saturated toluene again produced a tran-
sient species with a maximum centered at 430 nm which
reacted with CO with a similar rate constant of (3.3 (
0.3) × 108 M-1 s-1 (see Scheme 2). These results indicate
that neither the alkane nor the aromatic solvents
interact to any significant extent with (η5-C5H5)(η1-N-
C4H4N)Fe(CO).
UV/Vis F la sh P h otolysis Exp er im en ts of (η5-
C5H5)(η1-N-C4H4N)Fe(CO)2 in Cycloh exan e or Tolu -
en e Solu tion a t Room Tem p er a tu r e. Flash photol-
ysis of (η5-C5H5)(η1-N-C4H4N)Fe(CO)2 produced a tran-
sient species with a λ max at approximately 430 nm (λexc
) 355 nm) assigned to (η5-C5H5)(η1-N-C4H4N)Fe(CO).
A typical transient signal in CO-saturated cyclohexane
(15) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, M. R.; Zakrzewski, V. G.; Montgomery, J .
A., J r.; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Abuck, A. D.; Raghavachari, K.;
Foresman, J . B.; Cioslowski, J .; Ortiz, J . V.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.;
Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; J ohnson, B.; Chen, W.;
Wong, M. W.; Andres, J . L.; Gonzalez, C.; Head-Gordon, M.; Replogle,
E. S.; Pople, J . A. Gaussian 98, Revision A.3; Gaussian, Inc.: Pitts-
burgh, PA, 1998.
(16) The calculated band positions were corrected using a factor of
0.98, derived from calculations using the same model chemistry on
half-sandwich compounds of the type (η6-arene)Cr(CO)3, which exhibit
unambiguous spectral features particularly in the νCO region.
(17) Hooker, R. H.; Mahmoud, K. A.; Rest, A. J . J . Chem. Soc.,
Dalton Trans. 1990, 1231.
Discu ssion
In contrast with ferrocene, (η5-C5H5)(η5-C4H4N)Fe is
photoactive in dry hydrocarbon solvents. Broad-band
photolysis of (η5-C5H5)(η5-C4H4N)Fe at λ
> 500 nm
exc
in degassed cyclohexane at room temperature produced
a precipitate together with turbidity and a deepening
of the solution color. Subsequent addition of CO to the
photolyzed solution yielded (η5-C5H5)(η1-N-C4H4N)Fe-
(CO)2 over several hours, presumably by the slow
reaction of CO with the precipitate. Attempts to isolate
and characterize this precipitate failed.
Broad-band photolysis (λexc > 500 nm) of a CO-
saturated cyclohexane solution ([CO] ) 9.0 × 10-3 M)

3872 Organometallics, Vol. 19, No. 19, 2000
Heenan et al.
of (η5-C5H5)(η5-C4H4N)Fe formed (η5-C5H5)(η1-N-C4H4N)-
Fe(CO)2, confirming that the η5-coordinated pyrrolyl
ligand undergoes a photoinduced haptotropic shift reac-
tion. Another carbonyl compound, observed only follow-
ing monochromatic irradiation at 532 nm, was identified
as the exo-isomer of (η5-C5H5)(η3-C-C4H4N)Fe(CO) (spe-
cies 1). This assignment was based on a comparison of
and 2% CO in argon matrixes at 12 K. The former
assignment was based on the known νCO band positions
for (η5-C5H5)(η1-N-C4H4N)Fe(CO)2 in the same matrix
(Table 1). The latter depended on its similarity to the
spectroscopic properties of (η5-C5H5)Fe(CO)Cl measured
by Rest and co-workers (Table 1)17 and the results of
our investigations of the photochemistry of (η5-C5H5)-
(η1-C4H4N)Fe(CO)2 reported here. Experimental data in
the literature confirm that a η1-N-coordinated pyrrolyl
ligand has electronic properties similar to those of the
halogens, resulting in similar νCO frequencies (Table
1).17,18
In low-temperature matrixes the monocarbonyl and
dicarbonyl species were formed concomitantly and in the
same ratio throughout the photolysis with λ > 495 nm.
This observation suggests that they were both formed
by the reaction of CO with a single photoproduct, i.e.,
the 14-electron species (η5-C5H5)(η1-N-C4H4N)Fe. The
observation that the ratio of mono- to dicarbonyl prod-
ucts was strongly dependent on the concentration of CO
further supported this conclusion, the relative yield of
the dicarbonyl species being greatest in matrixes con-
taining a greater concentration of CO.
The νCO band (1974 cm-1) of (η5-C5H5)(η1-N-C4H4N)-
Fe(CO) was asymmetric (Figure 4). This was true
whether the monocarbonyl species was formed by pho-
tolysis of (η5-C5H5)(η5-C4H4N)Fe in a CO-doped matrix
or from (η5-C5H5)(η1-N-C4H4N)Fe(CO)2 in an argon
matrix. The shoulders on the low-energy side indicated
the presence of further CO-containing species rather
than a matrix effect, as only this band exhibited the
asymmetry. The species absorbing at approximately
1948 cm-1 was assigned to (η5-C5H5)(η3-C-C4H4N)Fe-
(CO) (species 1) on the basis of the observation of a
similar band in the room-temperature photolysis of (η5-
C5H5)(η5-C4H4N)Fe in CO-saturated cyclohexane (see
above). The species absorbing at 1962 cm-1 is more
difficult to assign however. This species could be either
the aza-allyl isomer (η5-C5H5)(η3-N-C4H4N)Fe(CO) or
alternatively the endo-isomer of (η5-C5H5)(η3-C-C4H4N)-
Fe(CO) (species 3 and 4, respectively). It is generally
accepted that the endo-isomers of allyl compounds are
less stable than the exo-isomers and that the carbonyl
stretching absorptions of the endo-isomers occur at
lower energy than those of exo-isomers. Consequently,
an assignment of the 1962 cm-1 band to species 4 is not
consistent with previously published data for related
systems,19,20 and we favor the assignment of the 1962
cm-1 band to species 3.
The yield of the species absorbing at 1962 cm-1 was
significantly increased by initially using monochromatic
irradiation (λexc ) 538 nm) followed by broad-band
photolysis (λexc > 495 nm) (Figure 5). This suggests that
monochromatic irradiation of (η5-C5H5)(η5-C4H4N)Fe
produced a precursor species, containing an η3-coordi-
nated pyrrolyl ligand, which then reacted with CO,
producing the observed species. Therefore, the ultimate
observation of a particular carbonyl-containing species
is dependent on the initial photolysis wavelength be-
cause all precursor species are themselves photoactive.
its νCO band (1948 cm-1) with that of (η5-C5H5)(η3-C3H5)-
Fe(CO) (1950 cm-1).11 The observation of what appears
to be an intermediate η3 species suggests that two
photons are required for the overall η5-to-η1 transforma-
tion. Indeed, subsequent broad-band visible photolysis
of the η3-intermediate indicates that it is photosensitive,
presumably yielding the η1 species. However it is not
possible to be certain of this, as the additional dicarbonyl
species produced could be the result of further photolysis
of the parent species still present in solution.
Matrix experiments on (η5-C5H5)(η5-C4H4N)Fe were
performed to identify the nature of ring-slip intermedi-
ates. These experiments can be divided into two groups,
those in which the isolating matrix is inert and those
in which the matrix material reacts with the photo-
fragments. For this system dinitrogen, frequently used
as a reactive matrix, is inert. Consequently, active
matrixes used consisted of various concentrations (ex-
pressed as %) of CO in argon. Coordination of CO to
the photofragments provided information on their na-
ture by examination of the νCO absorption bands.
Long-wavelength (λexc > 495 nm) photolysis of (η5-
C5H5)(η5-C4H4N)Fe in an argon matrix resulted in
depletion of the infrared bands of the parent compound,
and the formation of new bands in the fingerprint region
(Figure 7). Because the photoproduct(s) lack diagnostic
spectroscopic features, we calculated the IR spectrum
of a likely product by DFT methods. The band positions
calculated for (η5-C5H5)(η1-C4H4N)Fe provide the best
match to those observed in the matrix experiment
(Table 2). As mentioned above, experiments in a dini-
trogen matrix produced almost identical spectral changes
to those observed in argon, with no evidence for species
with coordinated dinitrogen.
Experiments using CO-doped matrixes provided con-
siderably more information on the nature of the photo-
products. In these experiments, both (η5-C5H5)(η1-
N-C4H4N)Fe(CO)2 and (η5-C5H5)(η1-N-C4H4N)Fe(CO)
(species 2) were identified following low-energy broad-
band photolysis of (η5-C5H5)(η5-C4H4N)Fe in both 0.5%
(18) Alway, D. G.; Barnett, K. W. Inorg. Chem. 1978, 17, 2826.
(19) Belmont, J . A.; Wrighton, M. S. Organometallics 1986, 5, 1421.
(20) Fish, R. W.; Giering, W. P.; Marten, D.; Rosenblum, M. J .
Organomet. Chem. 1976, 105, 101.

Photochemistry of (η5-C5H5)(η5-C4H4N)Fe
Organometallics, Vol. 19, No. 19, 2000 3873
and used without further purification. The parent species, (η5-
C5H5)(η5-C4H4N)Fe, was synthesized by the method reported
by Pauson,23 and (η5-C5H5)Fe(CO)2(η1-N-C4H4N) was synthe-
sized according to the method of Zakrzewski.24
Laser flash photolysis of (η5-C5H5)(η1-N-C4H4N)Fe-
(CO)2 produced a transient species with a λmax centered
at 420 nm. The primary photoproduct in solution was
confirmed as the 16-electron CO-loss species. The
observed rate constant for its reaction with CO varied
linearly with [CO], providing estimates of the second-
order rate constant of (3.0 ( 0.3) × 108 M1 s-1 in
cyclohexane and (3.3 ( 0.3) × 108 M-1 s-1 in toluene at
298 K. The magnitude of the second-order rate constants
indicates that this intermediate does not interact with
the solvent to any significant extent. This is in contrast
with 16-electron species of the group 6 metals. For
instance, the rate of displacement of solvent by CO in
(η6-C6H6)Cr(CO)2(toluene)21 is an order of magnitude
slower than that for (η6-C6H6)Cr(CO)2(cyclohexane).22
These experiments confirmed that (η5-C5H5)(η1-N-
C4H4N)Fe(CO)2 undergoes photoinduced CO loss fol-
lowing photolysis at 355 nm, which in turn explains why
the ratios of (η5-C5H5)(η1-N-C4H4N)Fe(CO) to (η5-C5H5)-
(η1-N-C4H4N)Fe(CO)2 formed following photolysis of (η5-
C5H5)(η5-C4H4N)Fe in CO-doped argon matrixes varied
with photolysis time when λexc > 325 nm was used. With
photons of these wavelengths the dicarbonyl product
undergoes a photoinduced decarbonylation.
Ap p a r a tu s. Spectra were recorded on the following instru-
ments: IR, Perkin-Elmer 2000 FT-IR (2 cm-1 resolution), UV/
vis Hewlett-Packard 8452A; NMR, Bruker AC 400. The laser
flash photolysis apparatus has been described previously.5 For
this work, both the 355 and the 532 nm lines of a pulsed Nd:
YAG laser were used (energy approximately 35 and 80 mJ per
pulse respectively; system response 20 ns). Solutions for
analysis were placed in a fluorescence cuvette (d ) 1 cm)
attached to a degassing bulb and were degassed by three cycles
of freeze-pump-thaw to 10-2 Torr, followed by liquid pumping
to remove traces of water and carbon dioxide (this typically
removed half the original volume of solvent). The absorbance
of the solution at the excitation wavelength was adjusted to
lie in the range 0.5-1.0. The UV/vis spectrum of the sample
solution was monitored throughout the experiments to monitor
changes in absorbance. The concentration of CO was deter-
mined by the pressure of CO admitted to the cell. The
solubilities of CO in cyclohexane and toluene are 9.0 × 10-3
and 7.5 × 10-3 M, respectively, under 1 atm of CO at 298 K.25,26
The matrix isolation apparatus has been described in detail
elsewhere.27 Samples for infrared spectroscopy or for UV/vis
spectra were deposited onto a CsI or BaF2 window cooled by
an Air Products CS202 closed-cycle refrigerator to 12-20 K.
The outer windows of the vacuum shroud were chosen to
match. Both (η5-C5H5)(η5-C4H4N)Fe and (η5-C5H5)(η1-N-C4H4N)-
Fe(CO)2 were sublimed from right-angled tubes at 289-294
and 313-316 K, respectively, as the gas stream entered the
vacuum shroud. The samples were deposited onto the windows
at 20 K, which were then cooled to 12 K before recording the
IR spectra on a Mattson Unicam Research Series FTIR
spectrophotometer fitted with a TGS detector and a CsI beam
splitter, which was constantly purged with dry CO2-free air.
Spectra were recorded at 1 cm-1 resolution with 128 scans.
UV/vis spectra were recorded on a Perkin-Elmer Lambda 7G
spectrophotometer. Matrixes were photolyzed through a quartz
window with a 300 W Xe arc lamp and a water filter.
Photolysis wavelengths were selected with cutoff or interfer-
ence filters.
Con clu d in g Rem a r k s
This study demonstrates that (η5-C5H5)(η5-N-C4H4N)-
Fe has an extensive photochemistry which is dominated
by haptotropic shifts of the coordinated pyrrolyl ligand.
A photoinduced stepwise shift from η5 through η3
ultimately yields an η1-coordinated species. The relative
importance of the various intermediates is strongly
dependent on the photolysis conditions. This reflects the
photosensitive nature of many of the intermediates
formed. The coordinatively unsaturated intermediates
also appear to be discriminating with what they will
react, a feature also observed with other iron systems.
The ability of the pyrrolyl ligand to adopt a variety of
coordination modes highlights its importance in eluci-
dating the mechanistic details of photoinduced hapto-
tropic shift reactions. Other compounds containing
π-coordinated pyrrolyl ligands are currently under
investigation.
Ack n ow led gm en t. The authors gratefully acknowl-
edge the support of Enterprise Ireland (D.P.H.) and
CONACYT (Mexico) (V.M.P.). We also appreciate the
help of Mr. V. Lafond with matrix experiments.
Su p p or tin g In for m a tion Ava ila ble: A listing of the final
refined geometric parameters and the results of a frequency
calculation for the 14-electron (η5-C5H5)(η1-N-C4H4N)Fe spe-
cies. This information is available free of charge via the
Internet at http://pubs.acs.org.
Exp er im en ta l Section
Ma ter ia ls. All operations were performed under inert gas
atmospheres, and purity of all isolated products was verified
by microanalysis. Spectroscopic grade cyclohexane and toluene
were used as obtained (Aldrich spectroscopic grade). Gases
used for the matrix experiments and for the flash photolysis
experiments (Ar, N2, and CO) were BOC research grade
(99.999% purity). (η5-C5H5)Fe(CO)2I was supplied by Aldrich
OM0002632
(23) Pauson, P. L.; Qazi, A. R.; Rockett, B. J . Organomet. Chem.
1967, 7, 325.
(24) Zakrzewski, J . J . Organomet. Chem. 1987, 327, C41-C42.
(25) Boese, W. T.; Ford, P. C. Organometallics 1994, 13, 3525.
(26) Makranczy, J .; Megyery-Balog, K.; Rosz, L.; Rosz, L.; Patyt, D.
Hung. J . Ind. Chem. 1976, 4, 269.
(21) Farrell, G. Ph.D. Thesis, Dublin City University, 1992.
(22) Breheny, C. J .; Kelly, J . M.; Long, C.; O’Keeffe, S.; Pryce, M.
T.; Russell, G.; Walsh, M. M. Organometallics 1998, 17, 3690.
(27) Haddleton, D. M.; McCamley, A.; Perutz, R. N. J . Am. Chem.
Soc. 1988, 110, 1810.