Mechanism of the photochemical ligand substitution reactions of fac-[Re(bpy)(CO)3(PR3)]+ complexes and the properties of their triplet ligand-field excited states

Article abstract of DOI:10.1021/ja017032m

We report herein the mechanism of the photochemical ligand substitution reactions of a series of fac-[Re(X₂bpy)(CO)₃(PR₃)]⁺ complexes (1) and the properties of their triplet ligand-field (³LF) excited

Full text of DOI:10.1021/ja017032m

Published on Web 08/30/2002
Mechanism of the Photochemical Ligand Substitution
Reactions of fac-[Re(bpy)(CO)3(PR3)]+ Complexes and the
Properties of Their Triplet Ligand-Field Excited States
Kazuhide Koike,*,† Nobuaki Okoshi,† Hisao Hori,† Koji Takeuchi,† Osamu Ishitani,*,‡,|
Hideaki Tsubaki,‡ Ian P. Clark,§ Michael W. George,§ Frank P. A. Johnson,§ and
James J. Turner§
Contribution from the National Institute of AdVanced Industrial Science and Technology,
Onogawa 16-1, Tsukuba 305-8569, Japan, Department of Chemistry, Graduate School of
Science and Engineering, Tokyo Institute of Technology, O-okayama 2-12-1, Meguro-ku,
Tokyo 152-8551, Japan, CREST, JST (Japan Science and Technology Corporation), 1-8 Honcho,
Kawaguchi-shi, Saitama Prefecture 332-0012, Japan, and Department of Chemistry,
UniVersity of Nottingham, Nottingham NG7 2RD, U.K.
Received September 8, 2001
Abstract: We report herein the mechanism of the photochemical ligand substitution reactions of a series
of fac-[Re(X2bpy)(CO)3(PR3)]+ complexes (1) and the properties of their triplet ligand-field (3LF) excited
states. The reason for the photostability of the rhenium complexes [Re(X2bpy)(CO)3(py)]+ (3) and [Re(X2-
bpy)(CO)3Cl] (4) was also investigated. Irradiation of an acetonitrile solution of 1 selectively gave the
biscarbonyl complexes cis,trans-[Re(X2bpy)(CO)2(PR3)(CH3CN)]+ (2). Isotope experiments clearly showed
that the CO ligand trans to the PR3 ligand was selectively substituted. The photochemical reactions
proceeded via a dissociative mechanism from the 3LF excited state. The thermodynamical data for the 3LF
excited states of complexes 1 and the corrective nonradiative decay rate constants for the triplet metal-
to-ligand charge-transfer (3MLCT) states were obtained from temperature-dependence data for the emission
lifetimes and for the quantum yields of the photochemical reactions and the emission. Comparison of 1
3
with [Re(X2bpy)(CO)3(py)]+ (3) and [Re(X2bpy)(CO)3Cl] (4) indicated that the LF states of some 3- and
3
4-type complexes are probably accessible from the MLCT state even at ambient temperature, but these
complexes were stable to irradiation at 365 nm. The photostability of 3 and 4, in contrast to 1, can be
explained by differences in the trans effects of the PR3, py, and Cl- ligands.
Rhenium diimine complexes of the type fac-[ReI(LL)-
(CO)3L′]n+ (n ) 0, 1; LL ) diimine ligand) have been well
Introduction
The photochemical and photophysical properties of diimine
complexes with d6 metal ions are tremendously interesting from
the standpoint of basic science as well as for their practical
applications.1-7 Many of these complexes exhibit the following
energetically accessible excited states: charge-transfer (CT),
ligand-centered (LC), and ligand-field (LF). An intimate un-
derstanding of their energetics and dynamics is important for
the design of new useful photochemical devices.
studied because they are excellent emitters,1,8-10 photo-
catalysts,6,11-16 and building blocks for supramolecules.5,17-19
Although a tremendous number of fundamental studies of the
photophysics of these rhenium complexes have been carried
out over the last two decades, the targets of these studies have
been only triplet CT excited statesssuch as metal-to-ligand
(8) Wrighton, M. S.; Morse, D. L. J. Am. Chem. Soc. 1974, 96, 998.
(9) Casper, J. V.; Meyer, T. J. J. Phys. Chem. 1983, 87, 952-954.
(10) Kalyanasundaram, K. J. Chem. Soc., Faraday Trans. 2 1986, 82, 2401.
(11) Hawecker, J.; Lehn, J.-M.; Ziessel, R. HelV. Chim. Acta 1986, 69,
1990.
* To whom correspondence should be addressed. E-mail: k-koike@
aist.go.jp (K.K.); ishitani@chem.titech.ac.jp (O.I.).
† National Institute of Advanced Industrial Science and Technology.
‡ Tokyo Institute of Technology.
(12) (a) Hori, H.; Johnson, F. P. A.; Koike, K.; Ishitani, O.; Ibusuki, T. J.
Photochem. Photobiol., A 1996, 96, 171-174. (b) Hori, H.; Koike, K.;
Ishizuka, M.; Takeuchi, K.; Ibusuki, T.; Ishitani, O. J. Organomet. Chem.
1997, 530, 169-176.
§ University of Nottingham.
| CREST, JST.
(1) Kalyanasundaram, K. Photochemistry of polypyridine and porphyrine
complexes; Academic Press Ltd.: London 1992.
(13) Hori, H.; Johnson, F. P. A.; Koike, K.; Takeuchi, K.; Ibusuki, T.; Ishitani,
O. J. Chem. Soc., Dalton Trans. 1997, 1019-1023.
(2) Meyer, T. J.; Casper, J. V. Chem. ReV. 1985, 85, 187-218.
(3) Roundhill, D. M. Photochemistry and photophysics of metal complexes;
Plenum Press: New York; 1994.
(14) Koike, K.; Hori, H.; Westwell, J. R.; Ishizuka, M.; Takeuchi, K.; Ibusuki,
T.; Enjouji, K.; Konno, H.; Sakamoto, K.; Ishitani, O. Organometallics
1997, 16, 5724-5729.
(4) Stufkens, D. J. Coord. Chem. ReV. 1990, 104, 39-112.
(5) Balzani, V.; Juris, A.; Ventyri, M.; Campagna, S.; Serroni, S. Chem. ReV.
1996, 96, 759-833.
(6) Kalyanasundaram, K. In Photosensitization and Photocatalysis using
Inorganic and Organometallic Compounds; Kalyanasundaram, K., Gra¨zel,
M., Eds.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1993.
(7) Juris, A.; Balzani, V.; Barigeketti, F.; Belser, P.; v. Zelewsky, A. Coord.
Chem. ReV. 1988, 84, 85-277.
(15) Pac, C.; Ishii, K.; Yanagida, S. Chem. Lett. 1989, 765.
(16) Tajik, M.; Detellier, C. J. Chem. Soc., Chem. Commun. 1987, 1824.
(17) Slone, R. V.; Benkstein, K. D.; Be´langer, S. J.; Hupp, T.; Guzei, I. A.;
Rheingold, A. L. Coord. Chem. ReV. 1998, 171, 221-243.
(18) Ziessel, R.; Juris, A.; Venturi, M. Inorg. Chem. 1998, 37, 5061-5069.
(19) Rajendran, T.; Manimaran, B.; Lee, F.-Y.; Lee, G.-H.; Peng, S.-M.; Wang,
C. M.; Lu, K.-L. Inorg. Chem. 2000, 39, 2016-2017.
9
11448
J. AM. CHEM. SOC. 2002, 124, 11448-11455
10.1021/ja017032m CCC: $22.00 © 2002 American Chemical Society

Ligand Substitution Reactions of [Re(bpy)(CO)3(PR )]+
A R T I C L E S
3
(3MLCT),8,20-25 ligand-to-ligand (3LLCT),18,26 and σ-to-ligand
(3σ-π*)2,27-29 statessand 3LC excited states.30-32 In contrast,
little is known about the LF states of these rhenium diimine
complexes. To our knowledge, only one report discusses their
triplet ligand-field (3LF) states: this report reveals that the
emission lifetimes of fac-[Re(LL)(CO)3(CNR)]+ complexes
strongly depend on temperature and that the temperature-
dependence profiles can be fitted by means of a model based
on three thermally accessible excited states: 3MLCT, 3LC, and
3LF.33 However, this report presents no unambiguous evidence
supporting the existence of another excited state, i.e., a 3LF state,
other than 3MLCT and 3LC, because fac-[Re(LL)(CO)3(CNR)]+
complexes are photostable and because their photochemical
ligand substitution reactions, a typical reaction proceeded via
bonds, these reactions are probably the first reported examples
of photochemical ligand substitution reactions of rhenium
diimine tricarbonyl complexes via their 3LF excited states. Rich
information on the “unknown” LF excited states of rhenium
complexes might be obtained from detailed investigation of the
photophysical and photochemical behaviors of 1.
3
In this study, we aim to (1) clarify the mechanism of the
photochemical ligand substitution reactions of 1, (2) obtain
3
thermodynamical data on the LF states of 1 by using photo-
substitution reactions as probes, and (3) investigate why rhenium
complexes such as fac-[Re(X2bpy)(CO)3Cl] and fac-[Re(X2bpy)-
(CO)3py]+ are photostable.
Experimental Section
3
the LF excited state of d6 metal complexes, have not been
Materials. Spectral-grade acetonitrile and CH2Cl2 purchased from
Kanto Chemical Co., Inc., were dried and distilled over CaH2 prior to
use. 13C-enriched (99.8 atom %) carbon monoxide was purchased from
ISOTECH Inc. The synthesis of and spectral data for the rhenium
complexes have been reported elsewhere.12,14
reported so far.
There had been no reports of photochemical ligand substitu-
tion reactions of other rhenium diimine tricarbonyl complexes,
until we recently reported the first examples as described
below,34 except for complexes with a Re-M (M ) metal
carbonyl), Re-C, or Re-H bond. The 3σ-π* excited states of
the complexes are accessible from the 3MLCT states at ambient
temperature, and the homolytic dissociation of these bonds pro-
ceeds via the 3σ-π* excited states.2,27-29 There are two possible
explanations for the photochemical stability of these rhenium
complexes: either (1) the energy gap between the lowest excited
Measurements. UV-vis absorption spectra were recorded on a
Hitachi 330 spectrophotometer or an Otsuka-Denshi Photal-1000
multichannel spectrophotometer with a D2 (25 W)/I2 (25 W) mixed
lamp. IR spectra were obtained in acetonitrile with a JEOL JIR-6500
FTIR spectrophotometer. Time-resolved infrared (TRIR) absorption
spectra were measured at the University of Nottingham by using a third
harmonic wave of the Quanta-Ray GCR-12S Nd3+-YAG laser (∼7 ns
fwhm, ∼50 mJ/pulse) as the photolysis source. The changes in IR
absorption at particular wavelengths were monitored with CW Mu¨tek
MDS4 diode laser elements and a HgCdTe detector (Infrared Associates
HCT-100) with “point-to-point” operation. The details of the TRIR
apparatus have been described elsewhere.22,35,36
3
3
state and the LF state is so large that LF is not accessible at
3
ambient temperature or (2) the LF state is accessible but not
reactive. We have no way to distinguish between these two
possibilities. It is surprising that participation of the 3LF excited
state has not been considered in most of the studies of rhenium
diimine complexes; if it is accessible, the photophysics and
photochemistry of these compounds might be affected.29
Emission Studies. Emission spectra were measured on a Hitachi
F-3000 or a JASCO FP-6600 spectrofluorometer. The spectra were
corrected for the detector sensitivity by using Rhodamine B, quinine
sulfate, and 4-(dimethylamino)-4′-nitrostilbene as standards (Hitachi
F-3000) or by using correction data supplied by JASCO (JASCO FP-
6600). Emission quantum yields were evaluated with quinine disulfate
as a standard. In the emission spectral measurements, the absorbance
of all sample solutions was less than 0.1 at the excitation wavelength.
The temperatures of the sample solutions in the 1 cm × 1 cm quartz
cell were controlled to within 0.1 °C by an EYELA CTP-101 cooling
thermo pump. Emission lifetimes were measured with a Horiba NAES-
1100 time-correlated single-photon-counting system (the excitation
source was a nanosecond H2 lamp, NFL-111, and the instrument
response time was less than 1 ns). Table 1 summarizes the emission
data measured at 298 K.
We have found that rhenium polypyridine tricarbonyl com-
plexes with a phosphorus ligand, such as fac-[Re(X2bpy)(CO)3-
(PR3)]+ (1; X2bpy ) 4,4′-X2-2,2′-bipyridine), are photoactive
even at ambient temperature and that excitation of these
complexes in acetonitrile solution selectively gives biscarbonyl
complexes cis,trans-[Re(X2bpy)(CO)2(PR3)(CH3CN)]+ (2) (eq
1).34 Because these complexes have no Re-M, Re-C, or Re-H
Emission Spectral Fitting. A single-mode Franck-Condon line-
shape analysis was used to fit the emission spectra.37-40 The spectra
were calculated with Wavemetrics Igor software on an Apple Macintosh
computer, and the parameters were optimized by comparison of
calculated and experimental spectra by means of the nonlinear least-
(20) Smothers, W. K.; Wrighton, M. S. J. Am. Chem. Soc. 1983, 105, 1067.
(21) Stripin, D. R.; Crosby, G. A. Chem. Phys. Lett. 1994, 221, 426-430.
(22) Ishitani, O.; George, M. W.; Ibusuki, T.; Johnson, F. P. A.; Koike, K.;
Nozaki, K.; Pac, C.; Turner, J. J.; Westwell, J. R. Inorg. Chem. 1994, 33,
4712-4717.
(32) Wallace, L.; Rillema, D. P. Inorg. Chem. 1993, 32, 3836-3843.
(33) Sacksteder, L. A.; Lee, M.; Demas, J. N.; Degraff, B. A. J. Am. Chem.
Soc. 1993, 115, 8230-8238.
(34) Koike, K.; Tanabe, J.; Toyama, S.; Tsubaki, H.; Sakamoto, K.; Westwell,
J. R.; Johnson, F. P. A.; Hori, H.; Saitoh, H.; Ishitani, O. Inorg. Chem.
2000, 39, 2777.
(23) Worl, L. A.; Duesing, R.; Chen, P.; Ciana, L. D.; Meyer, T. J. J. Chem.
Soc., Dalton Trans. 1991, 849-858.
(24) Vanhelmont, F. W. M.; Hupp, J. T. Inorg. Chem. 2000, 39, 1817-
1819.
(25) (a) Casper, J. V.; Meyer, T. J. J. Phys. Chem. 1983, 13, 359. (b) Carlson,
D. L.; Murphy, W. R., Jr. Inorg. Chim. Acta 1991, 181, 61-64.
(26) Schanze, K. S.; MacQueen, D. B.; Perkins, T. A.; Cabana, L. A. Coord.
Chem. ReV. 1993, 122, 63-89.
(35) Dixon, A. J.; Healy, M. A.; Hodges, P. M.; Moore, B. D.; Poliakoff, M.;
Simpson, M. B.; Turner, J. J.; West, M. A. J. Chem. Soc., Faraday Trans.
2 1986, 82, 2083.
(36) George, M. W.; Poliakoff, M.; Turner, J. J. Analyst (London) 1994, 119,
551.
(27) Rossenaar, B. D.; Kleverlaan, C. J.; M. C. E. Vandeven, M. C. E.; Stufkens,
D. J.; Vlcˇek, A., Jr. Chem.sEur. J. 1996, 2, 228-237.
(37) Worl, L. A.; Duesing, R.; Chen, P.; Della Ciana, L.; Meyer, T. J. J. Chem.
Soc., Dalton Trans. 1991, 849.
(28) Stufkens, D. J. Comments Inorg. Chem. 1992, 13, 359-385.
(29) Stufkens, D. J.; Vlcˇek, A., Jr. Coord. Chem. ReV. 1998, 177, 127-179.
(30) Giordano, P. J.; Wrighton, M. S. J. Am. Chem. Soc. 1979, 101, 2888.
(31) Sacksteder, L.-A.; Zipp, A. P.; Brown, E. A.; Streich, J.; Demas, J. D.;
DeGraff, B. A. Inorg. Chem. 1990, 29, 4335-4340.
(38) Caspar, J. V.; Kober, E. M.; Sullivan, B. P.; Meyer, T. J. J. Am. Chem.
Soc. 1982, 104, 630.
(39) Wang, Y.; Schanze, K. S. Inorg. Chem. 1994, 33, 1354.
(40) Walters, K. A.; Schanze, K. S. Spectrum 1998, 11, 1.
9
J. AM. CHEM. SOC. VOL. 124, NO. 38, 2002 11449

A R T I C L E S
Koike et al.
Table 1. Emission Data for the Rhenium Bipyridine Tricarbonyl
Complexes and Quantum Yields of the Photochemical Ligand
Substitution Reactions (Φr) of 1 in CH3CN at 298 K
Table 2. UV-Vis Absorption Spectral Data for
[Re(X2bpy)(CO)2(PR3)(CH3CN)]+ (2) in CH3CN at 298 K
2
λmax
ꢀ2
[Re(X bpy)(CO)3(PR )]+
2
3
X
PR
3
(nm)
(M-1 cm-1
)
λe
τe
X
PR
(nm)
(ns)
Φe
Φr
3
a
b
ca
d
e
H
P(OEt)3
P(OEt)3
P(OEt)3
P(n-Bu)3
PEt3
375
375
3900
4000
1a
1b
1c
1d
1e
1f
1g
1h
1i
H
P(OEt)3
P(OEt)3
P(OEt)3
P(n-Bu)3
PEt3
542
533
623
561
561
540
542
543
543
1034
936
162
621
654
416
644
952
1076
0.155
0.175
0.031
0.091
0.153
0.097
0.127
0.247
0.216
0.089
0.16
0.0003
0.095
0.15
0.55
0.25
0.099
0.118
CH3
CF3
H
H
H
H
H
H
CH3
CF3
H
H
H
H
H
H
427
422
405
392
382
374
3600
3600
3600
3100
3100
2500
f
PPh3
PPh3
g
h
i
P(OMe)Ph2
P(O-i-Pr)3
P(OMe)3
P(OMe)Ph2
P(O-i-Pr)3
P(OMe)3
a See the text.
proceeded quantitatively. On the basis of these observations, we
determined the molecular absorption coefficients of products 2 (ꢀ2) from
the final spectra of the irradiated solutions, as shown in Table 2. We
have isolated 2a and 2f,34 and their molecular absorption coefficients
are in good agreement with the ꢀ2 values in Table 2. Using these data
and eq 3, we calculated the concentrations of product 2 and starting
A(λ) - A0(λ)
[2] ) [1]0 - [1] )
(3)
(ꢀ2(λ) - ꢀ1(λ))d
material 1 after irradiation. d is the path length (1 cm), A(λ) is the
absorbance at the monitoring wavelength, A0(λ) is the initial absorbance
12b
at λ, and ꢀ1 and ꢀ2 are the molecular absorption coefficients at λ of
1 and 2, respectively. We also observed a set of isosbestic points in
the first stage of the photochemical reaction of 1c. However, the
photoreactivity of 1c was low, and some byproducts were produced
upon prolonged irradiation. Therefore, we determined the concentration
of 1c using HPLC.41
Figure 1. Corrected emission spectrum of 1a. The solid line was obtained
from the experimental data, and the dotted line is a spectrum calculated by
means of a single-mode Franck-Condon line-shape analysis.
13C NMR Study of the Photochemical Reaction of 1a. A CDCl3
solution (1 mL) containing 1a (16.5 mM) was degassed by means of
the freeze-pump-thaw method and transferred into a 5 mm L NMR
tube (total volume ∼4 mL). After 550 mmHg of 13CO gas (13C content
99.8 atom %) was introduced through a vacuum line, the NMR tube
was shaken for several minutes, and the tube was then sealed with a
torch. The solution was irradiated with 365 nm light in the apparatus
described above. The 13C NMR spectra of the solution at various times
during irradiation were measured on a JEOL Lambda 500 system (125
MHz) by using CDCl3 as an internal standard. Integration of the peaks
obtained by the NOE complete 1H-decoupling method (EXMOD: nne)
was used for determining the relative concentrations of the 13CO ligands;
the methylene carbons of the P(OEt)3 ligand were used as an internal
standard.
squares method. Equation 2 was used for the calculated spectra. I(ν˜) is
I(ν˜) )
νm
ν˜00 - νmh$m 3Sm
ν˜00 - νmh$m
∆ν˜00,1/2
5
2
exp -4 ln 2
(2)
∑
{
}
}
]
(
)
(
)
[
ν˜00
νm
νm)0
the relative emission intensity at frequency ν˜, ν˜00 is the 0-0 emission
energy, h$m is the average of medium-frequency acceptor modes that
are coupled to the electronic transition, Sm is the electron-vibration
ν
m
coupling constant (Huang-Rhys factor), and ∆ν˜00,1/2 is the half-width
of the 0-0 vibronic band; the sum is taken over the quantum number
of the average medium-frequency vibrational mode (νm). A value for
h$m of 1450 cm-1 was used in all of the fits.37 As a typical example,
the fit for 1a is shown in Figure 1. Similarly good fits were obtained
for other complexes. The 0-0 band energy gaps between 3MLCT and
the ground state (E00(3MLCT)) were obtained from the calculated
spectra.
Photolysis. An Ushio 350 W high-pressure Hg lamp was used as a
light source, and the 365 nm monochromatic light was introduced into
a Pyrex vessel with a 10-cm path length solution filter of CuSO4 (0.46
M) and a Shimazu M250 monochromator. The irradiation light intensity,
4.24 × 1015 photons s-1, was calculated with a K3Fe(C2O4)3 chemical
actinometer. A gentle stream of argon was bubbled into an acetonitrile
solution of 1 (0.12 mM) for 15 min, and the solution was then
photolyzed. The temperature of the solution was controlled by means
of the cooling thermo pump described above. The photochemical
reactions were monitored with the Otsuka-Denshi Photal-1000 multi-
channel spectrophotometer. In all reactions except for that of 1c, one
set of isosbestic points was observed until the starting complexes
completely disappeared, and therefore, the photochemical reactions
Results and Discussion
Mechanism of the Photochemical Ligand Substitution
Reactions. As a typical example of the spectra observed for
the photochemical ligand substitution reactions, the in situ UV-
vis absorption spectra obtained during irradiation of an aceto-
nitrile solution containing 1a are shown in Figure 2. A set of
isosbestic points was observed, and the final spectrum was
identical to that of isolated 2a, which indicates that the reaction
proceeded quantitatively and that product 2a was photochemi-
cally stable. Irradiation of acetonitrile solutions containing other
fac-[Re(X2bpy)(CO)3(PR2R′)]+-type complexes (1; X ) H, CH3,
CF3; R ) R′ ) Ph, Et, n-Bu, O-i-Pr, OPh, OMe; R ) Ph, and
R′ ) OEt) caused similar ligand substitution reactions; i.e., cis,-
trans-[Re(X2bpy)(CO)2(PR2R′)(CH3CN)]+ complexes were
(41) Hori, H.; Koike, K.; Ishizuka, M.; Westwell, J. R.; Takeuchi, K.; Ibusuki,
T.; Ishitani, O. Chromatographia 1996, 43, 251.
9
11450 J. AM. CHEM. SOC. VOL. 124, NO. 38, 2002

Ligand Substitution Reactions of [Re(bpy)(CO)3(PR )]+
A R T I C L E S
3
Figure 2. In situ UV-vis absorption spectral changes in a CH3CN solution
containing 1a (0.12 mM) under an Ar atmosphere during irradiation at 365
nm.
Figure 4. Intensity change in the 13C NMR signals for the axial (O) and
the equatorial (b) CO ligands of 1a in a CDCl3 solution during irradiation
(365 nm) under a 13CO atmosphere. The methylene carbon signal of the
P(OEt)3 ligand was used as an internal standard.
the phosphorus atom of the ligand. Irradiation of the solution
caused a tremendous increase in the axial 13CO peaks but almost
no change in the equatorial 13CO peaks, as shown in Figure 3b.
Figure 4 shows the changes in the 13C peak intensities of the
CO ligands at various times during irradiation. Even after a
photochemical stationary state was almost reached ()500 min),
we observed only a small increase in the peak areas of the
equatorial 13CO ligands. If the photochemical ligand substitution
reaction with 13CO proceeded by means of a shift of the axial
CO to the equatorial position, the 13C content at the equatorial
CO ligand should increase at least after the photochemical
stationary state is reached (eq 6). These results clearly indicate
that the operative mechanism is that shown in eq 4, i.e., the
site-selective substitution of the axial CO ligand.
Figure 3. 13C NMR spectra of 1a in a CDCl3 solution under a 13CO
atmosphere: (a) before irradiation (10000× accumulation) and (b) after 60
min of irradiation (18000× accumulation).
formed quantitatively. The quantum yields of the photochemical
ligand substitution reactions at 25 °C (Φr) are shown in Table
1. In all cases, neither cis,cis-[Re(X2bpy)(CO)2(PR2R′)(CH3-
CN)]+ nor [Re(X2bpy)(CO)3(CH3CN)]+ was observed, which
indicates that acetonitrile was selectively introduced into a
position trans to the phosphorus ligand.
There are two possible mechanisms for this selectivity: in
one mechanism, only the axial CO ligand, which is trans to the
phosphorus ligand, is photolabile (eq 4), and in the other
mechanism, the dissociation of the equatorial CO ligands is
accompanied by a shift of the axial CO to the open equatorial
position and subsequent occupation of the axial position by the
acetonitrile molecule (eq 5).
Another mechanistic question is whether the photochemical
ligand substitution reactions proceed via a dissociative or an
associative mechanism. Irradiation of a CH2Cl2 solution of 1a
in the presence of reagents L (240 mM) having different
nucleophilicitiesse.g., CH3CN, pyridine, and P(OEt)3sgave
cis,trans-[Re(bpy)(CO)2{P(OEt)3}(L)]+ quantitatively, and nei-
ther the type nor the concentration of the added reagent affected
the quantum yields of the formation of these products. Conse-
quently, the photochemical ligand substitution reaction of 1a,
and presumably also that of 1b-i, probably proceeds via the
dissociative mechanism shown in eqs 7 and 8.
We employed 13C NMR spectroscopy to determine which
mechanism was operating. Figure 3a shows the 13C peaks in
the carbonyl region of a 13CO-saturated CDCl3 solution contain-
ing 1a before irradiation. We attributed the two sets of peaks,
at 193.2 and 193.1 ppm and at 187.8 and 187.1 ppm, to one
axial and two equivalent equatorial CO ligands, respectively,
on the basis of their peak areas and the coupling constants with
hν
[Re(bpy)(CO)3(PR3)]+ 98 [Re(bpy)(CO)2(PR3)]+ + CO (7)
[Re(bpy)(CO)2(PR3)]+ + L f [Re(bpy)(CO)2(PR3)(L)]+ (8)
Reactive Excited State. We used TRIR spectroscopy to
obtain information about the excited state from which the
9
J. AM. CHEM. SOC. VOL. 124, NO. 38, 2002 11451

A R T I C L E S
Koike et al.
Figure 6. Temperature dependence of the emission yield (Φe) and lifetime
(τe), and the quantum yield of the photochemical ligand substitution reaction
(Φr) of 1a in a degassed CH3CN solution.
Scheme 1. Jablonski Diagram Describing the Mechanism of the
Photoinduced Ligand Substitution Reaction of 1
Figure 6. We also found that the lifetime (τe) and the quantum
yield (Φe) of the emission that occurred from the 3MLCT excited
state5,6 depended strongly on the temperature of the solution
(Figure 6). We should point out that, unlike Φr, both Φe and τe
decreased at higher temperature. These results strongly suggest
that the photochemical ligand substitution reaction occurred
via an excited state (RS) thermally accessible from the 3MLCT
state. Photochemical ligand substitution reactions of many
transition-metal diimine complexes are known to proceed via
the 3LF excited state that is thermally accessible from the
3MLCT state, especially when the photosubstitution occurs by
a dissociative mechanism.45 We can interpret all of the
experimental results for the photosubstitutions of 1a-i with the
energy diagram shown in Scheme 1. However, more-detailed
thermodynamic studies are required before we can definitively
assign the RS state to 3LF because, as discussed below, we lack
information about the 3LF states of the rhenium diimine
complexes.
Figure 5. Transient IR spectra of 1a in a degassed CH3CN solution taken
at 50 ns (a) and 5 µs (b) after 355 nm laser excitation. The three insets
represent the typical decay traces for different IR bands.
photochemical ligand substitution reaction occurs. Figure 5
shows the ν(CO) region in the TRIR spectra after laser excitation
of 1a in an Ar-saturated acetonitrile solution. Immediately after
the laser flash, we observed three negative peaks (at 1928, 1962,
and 2047 cm-1), which indicate a decrease in the concentration
of the ground state of 1a, and three positive peaks at higher
wavenumbers attributable to the 3MLCT excited state (∼1970,
2017, and ∼2070 cm-1) (Figure 5a).42 Figure 5b shows the
3
spectrum 5 µs after the laser excitation: the MLCT state has
disappeared, a set of new peaks at 1876 and 1951 cm-1, which
coincide with the peaks for product 2a, has increased, and there
is partial recovery of the peaks of the ground state. Figure 5
also illustrates the temporal behavior of the three peaks
3
attributable to the MLCT state (2012 cm-1), the ground state
On the basis of the energy diagram in Scheme 1, the
temperature dependence of the quantum yields of the emission
and the reaction (Φe and Φr) and the emission lifetime (τe) can
be described by eqs 9-13.43
(2050 cm-1) of 1a, and the ground state of 2a (1871 cm-1).
The good agreement among the first-order rate constants
obtained from these data (1.2 × 10-6, 1.1 × 10-6, and 1.6 ×
10-6 s-1, respectively) clearly indicates that the photochemical
ligand substitution reaction of 1a to form 2a proceeded via either
Φe ) ke/(ke + kd1 + k(T))
Φr ) kr′/(ke + kd1 + k(T))
(9)
(10)
(11)
(12)
(13)
3
the MLCT state of 1a or another excited state (RS) that was
3
thermally accessible from the MLCT state.
Studies of the temperature dependence of the reactions and
the emissions gave additional information. The quantum yields
of the photochemical reactions of 1a-i (Φr) increased at higher
temperatures. A typical example, the case of 1a, is shown in
τe ) 1/(ke + kd1 + k(T))
kr′ ) Φr/τe ) {kr/(kr + kd2)}k(T)
k(T) ){kth(kr + kd2)}/(k-th + kr + kd2)
(42) Similar frequency shifts of the ν(CO) peaks have been reported for the
3MLCT states of other rhenium bipyridine carbonyl complexes, such as
Re(bpy)(CO)3Cl44 and [Re(bpy)(CO)2{P(OEt)3}2]+.22 These shifts can be
explained as follows: the lower electron density on the rhenium metal in
the 3MLCT state causes a decrease in π-back-donation to the π* orbitals
of the carbonyl ligands, which weakens the CO bonds.
(43) Thorburn, I. S.; Rettig, S. J.; James, B. R. Inorg. Chem. 1986, 25, 227-
234.
(44) George, M. W.; Johnson, F. P. A.; Westwell, J. R.; Hodges, P. M.; Turner,
J. J. J. Chem. Soc., Dalton Trans. 1993, 2977.
9
11452 J. AM. CHEM. SOC. VOL. 124, NO. 38, 2002

Ligand Substitution Reactions of [Re(bpy)(CO)3(PR )]+
A R T I C L E S
3
temperature-dependence results for kr′ and eq 15, are sum-
marized in Table 3, along with the 0-0 band energy gaps
3
between the ground state and the MLCT states, E00(3MLCT).
Complexes 1a-i all show similar and positive ∆Sq values, which
suggests that their transition states are similar and that the ligand
substitution reactions proceed via the same dissociative mech-
anism. The sum of E00(3MLCT) and the activation free energy
change, i.e., (E00 + ∆Gq)298, is the energy gap between the
transition state of the ligand substitution reactions and the ground
state at 298 K, and because of the structural similarities among
1a-i, it is reasonable that this value closely approximates the
energy of the RS state or the thermal activation barrier to the
RS state. The introduction of strongly electron-withdrawing CF3
groups into the 4 and 4′ positions of the bipyridine ligand (1c)
greatly decreased the 3MLCT energy level of this complex (by
nearly 2200 cm-1) compared to that of bpy complex 1a. In
contrast, the 3MLCT energy level of 1b, which bears electron-
donating CH3 groups in the bpy ligand, was 170 cm-1 higher
than that of 1a. However, the differences in the activated state
Figure 7. Arrhenius plots of the radiative decay rate (ke) and the reaction
rate of the photochemical ligand substitution reaction (kr′) of 1a in a degassed
CH3CN solution.
Constants kd1 and kd2 are the nonradiative decay rate constants
for the 3MLCT and 3LF states, respectively. The rate constants
kth and k-th are the forward and backward internal conversion
3
rates between the MLCT state and the RS state, and depend
on the thermal activation between these two states. The
temperature-dependent component can be summarized in a
single term, k(T) (eq 13). We can analyze the temperature
dependence of the observed reaction rate kr′ by considering the
two limiting cases (1) k-th , kr + kd2 and (2) k-th . kr + kd2.
In case 1, the temperature-dependent term is the forward
thermal activation rate kth; i.e., k(T) ) kth. Consequently, kr′
can be described by eq 14, where ∆Er is the activation energy
energies for these three complexes were less than 1200 cm-1
.
Therefore, the reactive state energies, unlike the 3MLCT
energies, are not sensitive to the substituents on the bipyridine
ligand. However, the electronic properties of the phosphorus
ligands affected both the reactive state energies and the 3MLCT
energies similarly. These results unambiguously support the
3
hypothesis that the RS state is LF.
There are three possible relaxation pathways through the
3LF state: (1) a photodissociation that gives reaction products,
(2) a photodissociation and successive recombination, and (3)
a direct nonradiative decay to the ground state. An apparent
nonradiative decay rate from 3LF (kd2′ in eq 16) that occurs by
pathway 2 or pathway 3 or both can be described by eq 17, on
the basis of a treatment similar to that of kr′ in eq 15.
Consequently, eqs 9-11 can be rewritten as eqs 9′-11′. In eqs
0
and kth is the preexponential factor for the forward internal
conversion process.
kr′ ) {kr/(kr + kd2)}kth0 exp(-∆Er/RT)
(14)
In case 2, where thermal equilibrium is achieved, k(T) ) (kth/
k-th)(kr + kd2) and, consequently, kr′ ) (kth/k-th)kr. In this
equilibrium, kth/k-th should have an Arrhenius-type temperature
dependence and kr′ can be described by eq 14′, which is similar
to the equation used for case 1 except that the preexponential
factor kr0 becomes (kth0/k-th0)kr.
0
0
0
0
16 and 17, kd2 ) {kd2/(kr+kd2)}kth (case 1) or kd2 ) (kth
/
k-th0)kd2 (case 2).
kd2′ ) {kd2/(kr + kd2)}k(T)
kd2′ ) kd20 exp(-∆Er/RT)
(16)
(17)
0
kr′ ) (kth /k-th0)kr exp(-∆Er/RT)
(14′)
Consequently, the temperature dependence of the observed
reaction rate kr should be represented by eq 15 in both cases.
kr0 ) {kr/(kr + kd2)}kth0 (case 1), or kr0 ) (kth0/k-th0)kr (case 2).
Φe ) ke/{ke + kd1 + (kr0 + kd20) exp(-∆Er/RT)} (9′)
Φr ) kr0 exp(-∆Er/RT)/
kr′ ) kr0 exp(-∆Er/RT)
(15)
{ke + kd1 + (kr0 + kd20) exp(-∆Er/RT)} (10′)
As a typical example, the results for kr′ of 1a are shown in
Figure 7. The emission rate constant ke, which can be calculated
as Φe/τe, is also plotted. The Arrhenius plots for both kr′ and ke
show good linear relationships, as do the plots for the other
complexes. However, the activation energies for the ligand
substitution reactions were far different from those for the
emissions. In the case of 1a, the values of ∆Er and kr0 were
4620 ( 314 cm-1 and (4.11 ( 14.0) × 1014 s-1, respectively,
which represent the values for the thermal activation of 3MLCT
to RS. However, ke was almost independent of the temperature
in the range from 273 to 333 K (1.5 × 105 s-1). This result is
τe ) 1/{ke + kd1 + (kr0 + kd20) exp(-∆Er/RT)} (11′)
The temperature dependence of Φe, Φr, and τe was well
simulated by using eqs 9′-11′ and the obtained values of ke,
kr0, and ∆Er but could not be simulated if the temperature-
dependent nonradiative decay pathways 2 and 3 were excluded.
The evaluated rate constants kd1 and kd20 were 7.3 × 105 and
8 × 1014 s-1 for 1a. Comparison of kr0 with kd2 shows that
0
nonradiative decay from the 3LF state competed with the ligand
substitution process. Both the nonradiative decay and the
reaction pathways from the 3LF state were extremely fast (with
3
reasonable because the emissive MLCT is the lowest excited
3
state in this temperature range.
rate constants >1014 s-1); this suggests that the LF excited
The thermodynamical data on the photochemical ligand
substitution reactions of 1a-i, which were calculated from the
states of the complexes have repulsive potential curves, as in
case 1, and that the temperature-dependent nonradiative decay
9
J. AM. CHEM. SOC. VOL. 124, NO. 38, 2002 11453

A R T I C L E S
Koike et al.
Table 3. Thermodynamical Data for the Photochemical Ligand Substitution Reactions of [Re(X2bpy)(CO)3(PR3)]+ (1) in CH3CN
1
q
b
q
E (3MLCT)a
kd1
∆G 298
(cm-1
(E00 + ∆G )298
00
(cm-1
q
q
X
PR
3
)
(105 s-1
)
∆H (kJ mol-1
)
∆S (J mol-1 K-1
)
)
(cm-1
)
a
b
c
d
e
f
g
h
i
H
P(OEt)3
P(OEt)3
P(OEt)3
P(n-Bu)3
PEt3
19470 ( 55
19637 ( 50
17272 ( 49
18672 ( 50
18649 ( 51
19000 ( 31
19553 ( 46
19239 ( 52
19678 ( 43
6.9
4.1
60
11
11
8.5
9.7
7.9
6.2
52.8 ( 3.8
47.2 ( 0.5
65.6 ( 0.9
48.7 ( 5.2
44.3 ( 3.9
46.1 ( 10.3
42.4 ( 2.8
49.8 ( 4.8
50.4 ( 4.8
26.5 ( 12.3
13.2 ( 1.6
29.3 ( 2.8
26.0 ( 17.0
14.0 ( 12.5
27.5 ( 34.6
13.3 ( 9.1
24.6 ( 15.6
20.9 ( 15.7
3810 ( 590
3650 ( 80
23280 ( 650
23290 ( 130
22090 ( 190
22150 ( 880
22030 ( 660
22230 ( 1680
22790 ( 490
22840 ( 810
23420 ( 800
CH3
CF3
H
H
H
H
H
H
4820 ( 140
3480 ( 830
3390 ( 610
3230 ( 1650
3240 ( 440
3610 ( 760
3740 ( 760
PPh3
P(OMe)Ph2
P(O-i-Pr)3
P(OMe)3
3
a 0-0 band energy gap between the MLCT and the ground states. b Free activation energy change at 298 K.
Scheme 2. Energy vs Re-CO Distance for Rhenium Complexes,
Illustrating the Three Lowest-Lying Electronic States: (a)
[Re{(CF3)2bpy}(CO)3{P(OEt)3}]+ (1c), (b) [Re(bpy)(CO)3(py)]+ (3a)
and [Re{(CH3O)2bpy}(CO)3Cl] (4c) for the Case of a Nonreactive
3LF State, and (c) 3a and 4c for the Case of a Large Activation
Energy from the 3MLCT to 3LF States
Figure 8. ln(τe-1) vs T-1 for the series of fac-[Re(X2bpy)(CO)3Y]n+
complexes (Y ) py, Cl-) in CH3CN.
Table 4. Photophysical and Thermodynamical Data for
[Re(X2bpy)(CO)3Y]n+ Complexes in CH3CN
q
b
kd1
E (3MLCT)a
∆G 298
(cm-1
)
00
(cm-1
complexes, the explanation is not so simple, for the following
reasons. The energy level of the 3MLCT state of 3a, i.e.,
E00(3MLCT) of [Re(bpy)(CO)3(py)]+, is 220 cm-1 higher than
that of 1c (Tables 3 and 4). Moreover, the d-d energy splitting
of 3a (∆dd2 or ∆dd3 in Scheme 2) should be smaller than that of
1c (∆dd1) because of the weaker ligand-field strengths of py
and bpy compared to P(OEt)3 and (CF3)2bpy, respectively.
Similar reasoning can be applied to the Cl- complex [Re-
{(CH3O)2bpy}(CO)3Cl] (4c). Thus, 3a and 4c are photostable
complex
X
Y
n
(105 s-1
)
)
1c
3a
3b
4a
4b
4c
CF3
H
CF3
H
P(OEt)3
py
1
1
1
0
0
0
60
40
400
320
4100
520
17270 ( 50 4820 ( 140
17490 ( 15
15090 ( 15
15660 ( 30
13680 ( 20
16320 ( 1
-36 ( 15
84 ( 8
py
Cl-
Cl-
250 ( 21
252 ( 94
252 ( 40
CF3
OMe Cl-
a 0-0 band energy gaps between the 3MLCT and the ground states. b Free
activation energy change at 298 K.
3
3
even though their MLCT energies are higher and their LF
energies are lower than those of 1c, respectively. This can be
explained in two ways.
processes proceed via CO ligand dissociation and successive
recombination of the produced species, i.e., [Re(X2bpy)(CO)2-
(PR3)]+ and CO, in a solvent cage.
Photostabilities of Chloride and Pyridine Complexes. In
this paper, we have shown that the 3LF excited states of rhenium
diimine tricarbonyl complexes with a phosphorus ligand are
Scheme 2a illustrates the potential energies of the low-lying
3
electronic states (the ground state, the MLCT state, and the
3LF state) of 1c vs the bond distance between the rhenium metal
3
center and the axial-CO carbon atom. The LF state exhibits a
3
thermally accessible from the MLCT states at ambient tem-
repulsive potential energy curve (PEC) and is accessible from
perature. To investigate whether the same is true for other
rhenium complexes, we checked the photostabilities of [Re(X2-
bpy)(CO)3(py)]+ (3; py ) pyridine; X ) H, CF3) and [Re(X2-
bpy)(CO)3Cl] (4; X ) H, CF3, OMe), which were confirmed
to be stable to irradiation at 365 nm, in accordance with
“common knowledge”. Figure 8 illustrates Arrhenius plots for
the emission lifetimes of 3 and 4. Their temperature depend-
encies were much smaller than those for 1a-i between 10 and
60 °C. The activation free energy changes (∆Gq298) obtained
from these data together with kd1 and E00(3MLCT) are sum-
marized in Table 4.
3
the MLCT state by thermal activation (∆Gq ).
1
The PECs of 3a and 4c can be illustrated in two ways
(Scheme 2b,c):
(1) The PEC of the 3LF state crosses the PEC of the 3MLCT
state at a lower activation energy than that of 1c (∆Gq2 < ∆Gq ),
1
but there is an extra potential barrier along the Re-CO
dissociation coordinate, a barrier that has a rather high activation
energy (∆Gq ′) because of the trans effect (Scheme 2b). The
2
Re-C bond of the CO ligand trans to the phosphorus ligand is
much weaker than the Re-C bonds of the chloride complex
and the pyridine complex because of the stronger π-acidity of
PR3 compared to the π-base Cl- and the weaker π-acid py.46
When the complexes are excited to the 3LF states and the bonds
between the central rhenium and the ligands are weakened, the
weakest bonds, i.e., the Re-CO (axial) bonds of 1a-i, are
The photostabilities of the rhenium complexes have been
explained, without clear evidence, by large band gaps between
3
3
the LF and MLCT states.29,45 However, for at least some
(45) Vlcˇek, A., Jr. Coord. Chem. ReV. 1998, 177, 219.
9
11454 J. AM. CHEM. SOC. VOL. 124, NO. 38, 2002

Ligand Substitution Reactions of [Re(bpy)(CO)3(PR )]+
A R T I C L E S
3
broken, whereas the Re-CO bonds in 3 and 4 might not be
weak enough to be broken. (2) The PEC of the 3LF state crosses
the PEC of the 3MLCT state at an energy higher than that of 1c
3
because the potential well of the LF state is much shallower
than that of 1c (Scheme 2c). In this case, the activation energies
(∆Gq ) are potentially higher than the activation energy of 1c
3
(∆Gq ), even if they have smaller d-d energy splittings (∆dd
3
1
< ∆dd1).
Because compounds 1c, 3a, and 4c have similar molecular
3
structures, it is not unlikely that the curvatures of their LF
excited-state potential surfaces dramatically change despite
having the same CO leaving group in the reaction. The first
explanation seems more favorable and is also useful for
explaining the positional selectivity of the photochemical loss
of CO in 1a-i as described above: the PR3 ligand has a much
stronger trans effect than the X2bpy ligand.
Figure 9. ln kd1 (O) and ln τe (b) of 1a-i vs E00(3MLCT) at 25 °C in
-1
a degassed CH3CN solution. The inset depicts the same plots for 1, 3, and
4.
circles). The inset in Figure 9 shows all the data obtained in
our study, i.e., the data not only for 1 but also for 3 and 4. The
line has a slope of -8.4 ( 0.6 eV-1 and an intercept of 33.6 (
1.3, whereas the values reported by Casper and Meyer for the
rhenium bpy tricarbonyl complexes are -11.76 eV-1 and 40.21,
respectively.25a
We cannot clearly discriminate between the mechanisms for
the photostabilities of other Cl- and py complexes (Table 4) at
this point, because the energy levels of their 3LF states and their
3MLCT states are lower than those of 1a-i and because we
3
cannot obtain quantitative data for the LF states of the Cl-
and py complexes. However, we can assume that many of the
Cl- and py complexes might be stable even when excited to
Conclusion
3
the LF state for the reasons delineated in this section.
The photochemical ligand substitution reactions of fac-[Re-
(X2bpy)(CO)3(PR3)]+ (1) proceed via a dissociation mechanism.
The selective loss of the CO ligand from the position trans to
Energy Gap Law. Nonradiative decay processes for various
types of transition-metal complexes have been investigated by
using the energy gap law.25 Casper and Meyer have reported
that there is a good linear relationship between the energy gap
and the ln “kd” values, which are “nonradiative decay rate
3
the PR3 ligand occurs from the LF excited state. The thermo-
dynamical data for the photochemical reactions of 1 have been
obtained and are summarized in Table 3. In contrast to
complexes 1, complexes of the type fac-[Re(X2bpy)(CO)2(py)]+
and fac-Re(X2bpy)(CO)Cl were stable to 365 nm irradiation,
and the temperature dependence of their emission lifetimes was
considerably smaller than that for 1. However, the 3LF excited
states of fac-[Re(bpy)(CO)3(py)]+ and fac-Re{(MeO)2bpy}-
3
constants from MLCT of [Re(X2bpy)(CO)3Y]n+ ” calculated
from only emission quantum yields, lifetimes, and emission
energies.25a However, Casper and Meyer did not consider the
3
participation of decay from the LF excited state even though
they used a [Re(bpy)(CO)3(PR3)]+ complex. Now we know the
“appropriate” rate constants for the nonradiative decay from
3MLCTsi.e., the kd1 values (Table 3), which were calculated
with consideration of the participation of 3LF as described
abovesand the 0-0 band energy gaps E00(3MLCT), and these
data are plotted in Figure 9 according to the energy gap law.
The kd1 values obtained in our study for 1a-i (open circles)
show a much better linear relationship than those calculated
without consideration of the participation of the 3LF state (solid
3
(CO)3Cl at least are presumably accessible from the MLCT
states at ambient temperature. The photostabilities of these two
complexes relative to 1 can be explained by the difference in
the trans effects of the ligands PR3, py, and Cl-.
Acknowledgment. We thank Prof. Derk J. Stufkens (Uni-
versiteit van Amsterdam) and Prof. Yoko Kaizu (Tokyo Institute
of Technology) for useful discussions and comments. This work
was partially supported by Research and Innovative Technology
for the Earth (RITE).
(46) The Re-C (of the axial CO) bond lengths measured by X-ray crystal-
lographic analysis were 1.93 Å for 1a46a and 1.88 Å for 3a:46b (a) Ishitani,
O.; Tsubaki, H.; Toyama, S. Unpublished result. (b) Lucia, L. A.; Abboud,
K.; Schanze, K. S. Inorg. Chem. 1997, 36, 6224.
JA017032M
9
J. AM. CHEM. SOC. VOL. 124, NO. 38, 2002 11455