Radical Cation Probes for Photoinduced Intramolecular Electron Transfer in Metal-Organic Complexes

Article abstract of DOI:10.1021/jp953537l

Two transition metal complexes of the type fac-(bpy)Re(I)(CO)3(DA)(1+) (where bpy = 2,2'-bipyridine and DA is a pyridine ligand that is substituted with a 1,2-diamine electron donor) have been prepared.The 1,2-diamine serves as a "reactive donor ligand" o

Full text of DOI:10.1021/jp953537l

5408
J. Phys. Chem. 1996, 100, 5408-5419
Radical Cation Probes for Photoinduced Intramolecular Electron Transfer in
Metal-Organic Complexes
Yingsheng Wang and Kirk S. Schanze*
Department of Chemistry, UniVersity of Florida, P.O. Box 117200, GainesVille, Florida 32611
ReceiVed: NoVember 17, 1995X
Two transition metal complexes of the type fac-(bpy)ReI(CO)3(DA)+ (where bpy ) 2,2′-bipyridine and DA
is a pyridine ligand that is substituted with a 1,2-diamine electron donor) have been prepared. The 1,2-
diamine serves as a “reactive donor ligand” owing to its propensity to undergo rapid C-C bond fragmentation
when activated by single electron transfer oxidation. Photoexcitation of the diamine complexes affords a
ligand-to-ligand charge transfer (LLCT) state via intramolecular electron transfer quenching of a metal-to-
ligand charge transfer (MLCT) state, [(bpy)ReI(CO)3(DA)]+ + hν f [(bpy•-)ReII(CO)3(DA)]+*(MLCT) f
[(bpy•-)ReI(CO)3(DA•+)]+*(LLCT). Photochemical product and quantum efficiency studies indicate that the
diamine reactive donor ligand undergoes photoinduced C-C bond fragmentation with high efficiency,
presumably via the radical cation (DA•+) which is present in the LLCT excited state. Laser flash photolysis
allows direct detection of the metal complex based radicals that are formed by C-C bond fragmentation.
Quantitative kinetic information gathered through luminescence, laser flash photolysis, and quantum yield
studies allows estimation of the rates for formation of the LLCT state by forward electron transfer (kFET),
decay of the LLCT state by back electron transfer (kBET), and the rate of diamine radical cation bond
fragmentation in the LLCT state (kBF). The relationship between these kinetic parameters and the driving
force for electron transfer and bond fragmentation as well as the structure of the reactive donor ligands is
discussed.
Introduction
Transition metal complexes have played a prominent role
during the development of photoinduced electron transfer
(ET).1-6 Recently, research on photoinduced ET in transition
metal systems has emphasized charge separation and recombi-
nation in metal complex “dyads” and “triads” that are festooned
with organic electron donor and/or acceptor groups.7-12 The
As anticipated on the basis of the reactivity of aminoalcohol
study of these dyads and triads has led to increased understand-
radical cations, near-UV irradiation of two [(bpy)ReI(CO)3-
ing of the factors that control the efficiency for formation and
(AA)]+ complexes led to C-C bond fragmentation of the
lifetime of charge-separated ligand-to-ligand charge transfer
(LLCT) states that are formed by ET between an organic donor-
acceptor pair that are bridged by a transition metal center.
Our group has been examining the properties of LLCT excited
states in transition metal complexes that feature diimine acceptor
aminoalcohol reactive donor ligand, albeit with low quantum
efficiency.13a The C-C bond fragmentation reaction of the
aminoalcohol moiety was presumed to occur via the LLCT state
that is formed by forward ET quenching of the dπ (Re) f π*
(bpy) metal-to-ligand charge transfer (MLCT) excited state,
ligands such as 2,2′-bipyridine and a variety of organic amine
donor ligands.12 Through this effort we have developed
[(bpy)ReI(CO)3(AA)]+ + hν f
complexes that undergo photochemistry from the LLCT state
through the use of “reactive donor ligands” that are designed
to undergo a rapid bond fragmentation reaction triggered by
*[(bpy•-)ReII(CO)3(AA)]+
f
single ET oxidation (Scheme 1).13 We have demonstrated that
since the bond fragmentation reaction occurs from the LLCT
state and must compete with charge recombination via back
ET, detailed study of the efficiency of the permanent photo-
chemical reaction affords insight concerning the dynamics of
the forward and back ET reactions.13
MLCT
*[(bpy•-)ReI(CO)3(AA•+)]+
LLCT
f
bond fragmentation products (2)
While the [(bpy)ReI(CO)3(AA)]+ complexes clearly demon-
strated the utility of using a reactive donor ligand to probe the
dynamics of the LLCT state, the aminoalcohol system was
limited because the C-C bond fragmentation reaction of the
aminoalcohol radical cation is comparatively slow (kBF ≈ 105-
106 s-1) and therefore does not compete effectively with decay
of the LLCT state by back ET (kBET ≈ 108 s-1).13a The
comparatively slow rate of C-C bond fragmentation of the
aminoalcohol cation radical was believed to be the reason that
Recently, we presented a detailed study of a series of
complexes of the type fac-[(bpy)ReI(CO)3(AA)]+, where bpy
) 2,2′-bipyridine and AA is a reactive donor ligand which
features the 1,2-aminoalcohol unit.13a These complexes were
designed on the basis of previous reports which indicated that
aminoalcohol radical cations undergo rapid C-C bond frag-
mentation, eq 1, X ) OH:14
X Abstract published in AdVance ACS Abstracts, March 1, 1996.
0022-3654/96/20100-5408$12.00/0 © 1996 American Chemical Society

Photoinduced ET in Metal-Organic Complexes
J. Phys. Chem., Vol. 100, No. 13, 1996 5409
SCHEME 1
CHART 1
the overall quantum efficiency for reaction of the [(bpy)ReI-
(CO)3(AA)]+ complexes was low.
More recently we have turned our attention to the photo-
chemistry and photophysics of complexes of the type [(bpy)-
ReI(CO)3(DA)]+, where DA is a 1,2-diamine reactive donor
ligand.13b,c The decision to change from the aminoalcohol to
the diamine reactive donor ligand was predicated by studies of
bimolecular photoinduced ET which implied that C-C bond
fragmentation in 1,2-diamine cation radicals (eq 1, X ) NR2)
is substantially more rapid than in analogously substituted
aminoalcohol radical cations.15
This report presents a detailed study of the photochemical
and photophysical properties of e-1a and e-1b (Chart 1) that
contain the fac-(bpy)ReI(CO)3-chromphore covalently linked to
1,2-diamine reactive donor ligands. Steady state photochemical
studies demonstrate that photoexcitation of e-1a and e-1b into
the dπ (Re) f π* (bpy) MLCT absorption induces efficient
C-C bond fragmentation of the diamine reactive donor ligand.
Product studies also reveal that irradiation of e-1a and e-1b in
degassed solution produces t-1a and t-1b, respectively, presum-
ably via a mechanism involving reversible C-C bond frag-
mentation. Luminescence and nanosecond transient absorption
studies indicate that (1) in both complexes formation of the
LLCT state occurs rapidly (k g 109 s-1) after MLCT excitation;
(2) C-C bond fragmentation occurs within the LLCT state with
kBF ≈ 3 × 108 s-1. Quantitative analysis of the steady state
quantum yield data allows estimation of the rate of back ET to
be kBET ≈ 1 × 108 s-1 in both compounds. This observation
indicates that despite the significant difference in donor ligand
to metal electronic coupling in e-1a and e-1b, there is little
difference in the rate of back ET in the two compounds. The
significance of this finding with respect to the mechanism of
LLCT excited state decay via back ET is discussed.
steam bath for 4 h. During this period the reaction mixture
gradually solidified. After the 4-h heating period, the mixture
was cooled, whereupon 30 mL of concentrated HCl was added.
The acidified reaction mixture was then subjected to steam
distillation for a 4-h period. The reaction mixture was then
made basic by the addition of NaOH (15 g, 0.38 mol) dissolved
in 50 mL of H2O. At this point the product precipitated as a
light brown mass. The brown product was collected by suction
filtration, dried, and then purified by repeated sublimation. The
purified product was obtained as a pale yellow solid, yield 2 g
(10%): 1H NMR (300 MHz, CDCl3) δ 3.12 (s, 6H), 6.81 (d,
2H), 7.72 (d, 2H), 7.83 (d, 2H), 8.53 (d, 2H).
erythro-1-{p-[4-(Pyridyl)methyl]anilino}-2-piperidino)-1,2-
diphenylethane (5). A mixture of erythro-2-piperidino-1,2-
diphenylethanol17 (562 mg, 2.0 mmol), 10 mL of freshly distilled
CH2Cl2, and triethylamine (0.8 mL, 6.0 mmol) was placed in a
100-mL three-necked round bottom flask, and the resulting
solution was purged with N2 for 30 min. During this period
the temperature was adjusted to -20 °C by using a dry ice-
2-propanol bath. A solution of methanesulfonyl chloride (0.155
mL, 2.0 mmol) dissolved in 2 mL of freshly distilled CH2Cl2
was added dropwise by using a syringe. The resulting mixture
was maintained at -20 °C and stirred gently for 1 h. After
this period of time a solution of p-[(4-pyridyl)methyl]aniline13a
(370 mg, 2.0 mmol) in 3 mL of freshly distilled CH2Cl2 was
added dropwise by using a syringe. After the temperature of
the solution increased to ambient the solution was heated to
reflux. During the reflux period the reaction was monitored
by TLC (silica, ethyl acetate/hexane, 4:6 v/v). Refluxing was
discontinued after 16 h, at which point p-[(4-pyridyl)methyl]-
aniline was no longer visible by TLC. The solvent was removed
under reduced pressure, and the crude product was purified by
chromatography (silica, ethyl acetate/hexane, 4:6 v/v). Pure 5
was obtained as a yellow solid in 35% yield after purification:
Experimental Section
Solvents and chemicals used for synthesis were of reagent
grade and used without purification unless noted. Silica gel
(Merck, 230-400 mesh) and neutral alumina (Fisher, Brockman
grade III) were used for chromatography. NMR spectra were
run on either GE QE-300 MHz or Varian VXR-300 MHz
spectrophotometers. Metal complexes 2,16 3a,13a and (bpy)-
ReI(CO)3(CF3SO3)13b were prepared as described previously.
1
TLC (silica, ethyl acetate/hexane, 4:6 v/v) Rf ) 0.4; H NMR
(300 MHz, CDCl3) δ 1.45 (m, 2H), 1.56 (m, 4H), 2.45 (br,
4H), 3.45 (d, J ) 5.7 Hz, 1H), 3.80 (s, 2H), 4.76 (d, J ) 5.7
Hz, 1H), 5.0 (br, 1H), 6.45 (d, J ) 8.1 Hz, 2H), 6.86 (d, J )
8.1 Hz), 6.9-7.25 (m, 12H), 8.45 (d, J ) 6.0 Hz); 13C NMR
(75 MHz, CDCl3) δ 24.6, 26.2, 40.4, 52.3, 58.9, 75.9, 114.4,
124.1, 126.5, 127.0, 127.3, 127.4, 127.5, 127.7, 129.1, 129.4,
137.6, 141.5, 146.7, 149.6, 150.8.
-
All Re(I) complexes were isolated as PF6 salts.
p-(4-Pyridyl)-N,N-dimethylaniline (4). Freshly distilled
pyridine (16 g, 0.2 mol), benzoyl chloride (14 g, 0.1 mol), and
Cu powder (0.5 g, 8 mmol) were combined in a 1-L three-
necked flask which was fitted with a mechanical stirrer and a
reflux condensor. The reaction mixture was purged with dry
N2 and then heated and stirred for 1 h over a steam bath. After
this period the reaction mixture was cooled, whereupon N,N-
dimethylaniline (12 g, 0.1 mol) was added. The resulting
solution was again purged with N2 gas and then heated over a
Complex e-1a. (All procedures described below were carried
out in a dimly lighted room to minimize exposure of e-1a to
light.) A mixture of (bpy)Re(CO)3(TFMS) (115 mg, 0.2 mmol),
5 (134 mg, 0.3 mmol), NH4PF6 (326 mg, 2 mmol), and 5-mL
of freshly distilled THF was stirred at 25 °C under a blanket of
dry N2 for 4 h. Excess NH4PF6 was removed by filtration, and

5410 J. Phys. Chem., Vol. 100, No. 13, 1996
Wang and Schanze
the THF solvent was removed under reduced pressure. The
crude product was purified by chromatography on alumina
eluting with CH2Cl2/CH3CN (95:5 v/v). After removal of the
chromatography solvent, the residue was dissolved in a minimal
amount of CH2Cl2 and precipitated from n-pentane to produce
a yellow powder, yield 102 mg (50%) after purification: TLC
(300 MHZ, CD3CN) δ 3.76 (s, 2H), 6.53 (d, 2H), 6.82 (d, 2H),
7.08 (d, 2H), 7.77 (t, 2H), 8.06 (d, 2H), 8.27 (t, 2H), 8.38 (d,
2H), 9.19 (d, 2H).
e-1a Photoproduct Isolation and Characterization.
A
solution of e-1a in CH3CN (50 mg in 50 mL, c ) 0.7 mM)
was placed in a Pyrex test tube and degassed by bubbling with
a stream of argon for 2 h. The resulting degassed solution was
irradiated with a 450-W medium-pressure Hanovia mercury arc
lamp for 80 min. The 366-nm emission of the Hg arc was
isolated by using Corning 7-54 and Schott LG 350 glass filters.
The photolyzed solution was concentrated and the residue was
chromatographed on alumina using CH2Cl2/CH3CN (95:5 v/v)
as eluant. The fractions that contained the predominant yellow
band from the column were combined, and the solvent was
evaporated. The residue was dissolved in a minimal volume
of CH2Cl2 and then precipitated by addition to n-pentane to yield
5 mg of a bright yellow powder: 1H NMR (300 MHz, CD3-
CN) δ 1.30 (m, 2H), 1.50 (m, 2H), 1.60 (m, 2H), 2.3 (m, the
integral is obscured by a nearby H2O resonance), 3.66 (1H,
coupling obscured by nearby singlet), 3.71 (s, 2H), 4.62 (d, J
≈ 10.5 Hz, 1H), 6.07 (1H), 6.48 (d, 2H), 6.77 (d, J ) 8.1 Hz),
7.0-7.3 (m, 12H), 7.74 (t, 2H), 8.03 (d, J ) 5.1 Hz), 8.24 (t,
2H), 8.35 (d, J ) 8.1 Hz), 9.17 (d, 2H); 13C NMR (75 MHz,
CD3CN) δ 24.0, 26.0, 39.1, 49.3, 57.4, 75.1, 124.3, 126.2, 126.4,
127.1, 127.2, 127.6, 127.8, 128.4, 129.2, 129.6, 133.4, 140.7,
142.8, 147.4, 151.2, 153.5, 155.5, 156.0; FAB-HRMS calcd for
C44H41N5O3Re (M+) 874.277, found 874.284. The NMR and
mass analysis are consistent with structure t-1a.
1
(silica, CHCl3/MeOH, 96:4 v/v) Rf ) 0.2-0.3; H NMR (300
MHz, CD3CN) δ 1.28 (br, 2H), 1.36 (br, 4H), 2.15 (br, 2H),
2.43 (br, 2H), 3.62 (d, J ) 8.1 Hz, 1H), 3.67 (s, 2H), 4.8-4.9
(m, 2H, NH), 6.38 (d, J ) 8.1 Hz, 2H), 6.72 (d, J ) 8.1 Hz,
2H), 7.02 (d, J ) 6.3 Hz, 2H), 7.1-7.3 (m, 10H), 7.75 (t, J )
5.1 Hz, 2H), 8.02 (d, J ) 6.3 Hz, 2H), 8.24 (t, J ) 8.1 Hz,
2H), 8.36 (d, J ) 8.1 Hz, 2H), 9.16 (d, J ) 5.1 Hz, 2H); 13C
NMR (75 MHz, CD3CN) δ 24.1, 25.8, 39.1, 51.1, 57.4, 74.8,
124.4, 125.8, 126.2 (2C’s), 126.7, 127.1, 127.2, 127.4, 128.5,
128.9, 129.2, 136.8, 140.8, 142.6, 146.5, 151.2, 153.5, 155.5,
156.0; HRMS (FAB, positive ion) calcd for C44H41N5O3Re,
874.277 (M+); obsd, 874.273.
Complex e-1b. This complex was prepared according to the
procedure described above for e-1a, except erythro-1-{p-[4-
(pyridyl)]anilino}-2-piperidino-1,2-diphenylethane (6a)17 was
used in place of 5. The crude product was purified by
chromatography on alumina eluting with CH2Cl2/CH3CN (9:1
v/v). After removal of the chromatography solvent, the residue
was dissolved in a minimal amount of CH2Cl2 and precipitated
from n-pentane to produce a yellow powder, yield 90 mg (45%)
after purification: TLC (silica, CH2Cl2/CH3CN 4:1 v/v) Rf )
1
0.3-0.6; H NMR (300 MHz, CD3CN) δ 1.25 (br, 2H), 1.33
Procedures for Quantitative Photolyses. Quantum yield
studies were carried out using equipment and procedures that
have been described in detail in previous publications.13a HPLC
analysis were carried out using a Whatman ODS-3 reversed
phase column with a mobile phase consisting of THF/CH3CN/
H2O (5:45:50 v/v/v) with 40 mM sodium heptane sulfonate and
50 mM triethylamine. All analyses were conducted with a
detector wavelength of 250 nm.
As an example, the following is the procedure used for
quantum efficiency measurements on e-1a. A 2-mL aliquot of
an argon-degassed CH3CN solution containing e-1a (c ) 0.5
mM) was irradiated for 3 min (366 nm, 75-W high-pressure
Hg lamp). Following light exposure, 0.5 mL of an internal
standard solution (2 mM benzophenone in CH3CN/H2O, 9:1
v/v, pH ) 3) was added to 2 mL of the photoproduct solution,
and the mixture was analyzed by triplicate HPLC analysis.
For determination of quantum yields for e-1a f t-1a and
e-1b f t-1b isomerization, it was assumed that the molar
absorptivity of the threo and erythro isomers is the same at
250 nm.
(br, 4H), 2.15 (br, 2H), 2.46 (br, 2H), 3.70 (d, J ) 8.7 Hz),
5.00 (m, 1H), 5.43 (br, 1H), 6.54 (d, J ) 8.7 Hz, 2H), 7.1-7.4
(m, 14H), 7.77 (t, J ) 5.1 Hz, 2H), 8.01 (d, J ) 6.3 Hz, 2H),
8.24 (t, J ) 7.5 Hz, 2H), 8.34 (d, J ) 7.5 Hz, 2H), 9.20 (d, J
) 5.1 Hz, 2H); 13C NMR (75 MHz, CD3CN) δ 24.2, 25.9, 51.2,
57.3, 74.8, 113.5, 121.5, 122.2, 124.5, 126.6, 127.0, 127.4,
127.6, 127.9, 128.6, 129.0, 136.5, 141.0, 142.1, 150.2, 150.6,
151.3, 153.6, 155.6; HRMS (FAB, positive ion) calcd for
C43H39N5O3Re, 860.261 (M+); obsd, 860.264.
[(bpy)Re(CO)3(p-(4-pyridyl)-N,N-dimethylaniline)][PF6]
(3b). This complex was prepared according to the procedure
described above for e-1a, except p-(4-pyridyl)-N,N-dimethyla-
niline (4) was used in place of 5. The crude product was
purified by chromatography on alumina eluting with CH2Cl2/
CH3CN (9:1 v/v). After removal of the chromatography solvent,
the residue was dissolved in a minimal amount of CH2Cl2 and
precipitated from n-pentane to produce a yellow powder, yield
125 mg (81%) after purification: TLC (silica, CH2Cl2/CH3CN,
9:1 v/v) Rf ) 0.6-0.8; 1H NMR (300 MHz, CD3CN) δ 2.97 (s,
6H), 6.72 (d, J ) 8.7 Hz, 2H), 7.42 (d, J ) 6.3 Hz, 2H), 7.55
(d, J ) 8.7 Hz, 2H), 7.79 (t, J ) 5.1 Hz, 2H), 8.05 (d, J ) 6.3
Hz, 2H), 8.26 (t, J ) 8.1 Hz, 2H), 8.37 (d, J ) 8.1 Hz, 2H),
9.23 (d, J ) 5.1 Hz, 2H); 13C NMR (75 MHz, CDCL3) δ 39.2,
112.2, 120.8, 121.4, 124.6, 127.9, 128.7, 141.0, 150.7, 151.4,
152.4, 153.7, 155.7; HRMS (FAB, positive ion) calcd for
C26H22N4O3Re, 625.125 (M+); obsd, 625.128.
Methods for Emission, Transient Absorption, and Elec-
trochemical Experiments. Time-resolved emission, transient
absorption, steady state emission, and electrochemistry were
carried out by using methods and equipment that has been
described in previous publications.18 All transient absorption
experiments were carried out by using a recirculating cell with
a 100-mL volume to minimize the effects of sample decomposi-
tion during data acquisition.
[(bpy)Re(CO)3(p-{4-(pyridyl)methyl}aniline)][PF6] (7a). A
mixture of (bpy)Re(CO)3Cl (90 mg, 0.2 mmol), p-[(4-pyridyl)-
methyl]aniline13a (74 mg, 0.4 mmol), and AgPF6 (100 mg, 0.4
mmol) in 4 mL of DMF was heated to 80 °C for 3 h. After
cooling, the solution was filtered through a medium-porosity
fritted filter, and then the DMF was evaporated under reduced
pressure. The residue was chromatographed on alumina using
CHCl3/MeOH (95:5 v/v) as eluant. After removal of the
chromatography solvent, the residue was dissolved in a minimal
amount of CH2Cl2 and precipitated from n-pentane to produce
Results
Structures. The structures of the Re(I) complexes discussed
herein are illustrated in Chart 1. Complexes e-1a,b are the focus
of the present study. These compounds contain the fac-(bpy)-
Re(CO)3-chromophore covalently linked to a 1,2-diamine “reac-
tive donor”. In e-1a,b the 1,2-diamine ligands are comprised
of a 2° aromatic amine and a 3° aliphatic amine, and the diamine
unit has erythro stereochemistry. The distinction between these
1
a yellow powder, yield 70 mg (50%) after purification: H NMR

Photoinduced ET in Metal-Organic Complexes
J. Phys. Chem., Vol. 100, No. 13, 1996 5411
two complexes lies in the linkage between the pyridyl ligand
and the aromatic amine portion of the diamine: in e-1a there
is a methylene spacer, while in e-1b the methylene is absent,
and therefore the aromatic amine and the pyridine are conju-
gated. This difference is significant, since conjugation in e-1b
provides a pathway for strong electronic coupling between the
aniline donor and the Re(I) center. The aniline to Re coupling
is anticipated to be comparatively much weaker in e-1a due to
the methylene spacer.
CHART 2
Complex 2 was studied to provide information concerning
the dπ (Re) f π* (bpy) MLCT excited state of the (bpy)-
Re(CO)3-chromophore, while complexes 3a,b were examined
to provide information concerning the effect of the aniline donor
on the MLCT state without the complications associated with
the diamine reactive donor. It is also important to note that a
recent detailed report was published that examined the photo-
physical and photochemical properties of ligand 6a (Chart 2).17
decay rate in model complex 2. The nonradiative decay
pathway is attributed to forward ET from the donor ligand to
the photoexcited Re center (kFET path, Scheme 3).
Photochemistry of Diamine-Substituted Complexes. De-
tailed studies were carried out to identify the products formed
by irradiation of e-1a in air-saturated and argon-degassed CH3-
CN solution. In both cases irradiation was effected using 366-
nm light, which corresponds to the dπ (Re) f π* (bpy) MLCT
absorption. Scheme 2 provides a summary of the outcome of
the photochemical reactions, and Table 2 provides a listing of
the pertinent quantum efficiencies. Typical HPLC chromato-
grams for samples of e-1a which were irradiated to ap-
proximately 10% conversion are available as supporting infor-
mation.
Photophysics. The absorption spectra of e-1a,b, 2, 3a,b, and
6a were determined in CH3CN. Table 1 contains a listing of
absorption maxima and molar extinction coefficients, and Figure
1 illustrates the spectra of several representative complexes. The
spectrum of 2 (Figure 1a), which illustrates the spectral features
of the (bpy)Re(CO)3-chromophore, is characterized by a mod-
erately intense band with λmax ≈ 350 nm that is due to the dπ
(Re) f π* (bpy) MLCT absorption.19 This complex also
displays bands at 320 and 308 nm, which are assigned to π,π*
(bpy) intraligand transitions, while the intense features below
300 nm are a composite of π,π* intraligand and MLCT
transitions. The absorption of diamine complex e-1a is very
similar to that of 2 at wavelengths longer than 300 nm (Figure
1a), except for a slightly increased absorptivity, which is likely
due to weak n,π* absorption of the aniline moiety in the diamine
Photolysis of e-1a in air-saturated solution leads quantitatively
to formation of pyridyl-aniline complex 7a and benzaldehyde.
These products were identified by comparison of HPLC
retention times for peaks in the photoproduct solution with
HPLC retention times of authentic samples of the two com-
pounds. Benzeldehyde and 7a are believed to form by photo-
induced oxidative C-C bond fragmentation of the 1,2-diamine
ligand in e-1a, followed by hydrolysis of the resulting imine
complex 14a and iminium ion 11 (Scheme 3). Quantum
efficiency studies reveal that 7a and benzaldehyde are formed
in 1:1 and 2:1 ratios, respectively, relative to disappearence of
e-1a, as expected on the basis of the reaction stoichiometry
(Table 2). The quantum yield studies indicate that disappearence
of e-1a under air-saturated conditions is quite efficient, with Φ
) 0.54.
The reaction mixture obtained by photolysis of e-1a in argon-
degassed solution is more complicated. HPLC analysis of the
argon-degassed reaction mixture indicates that, in addition to
benzaldehyde and 7a, another product is formed which has a
slightly longer retention time than that of e-1a (see supporting
information). Semipreparative photolysis of e-1a followed by
silica gel chromatography afforded a small amount of the
unknown material. Analysis by 1H NMR, 13C NMR, and high-
resolution mass spectroscopy indicates that the additional
product is the threo diastereomer, t-1a (see the Experimental
Section). The threo diastereomer is likely formed by stereo-
random recombination of the metal complex radical 10a and
ion 11, which are formed by photoinduced oxidative C-C bond
fragmentation of the 1,2-diamine ligand (Scheme 3).15 Quantum
efficiency studies indicate that the disappearance of e-1a is
suppressed significantly in argon-degassed solution (Table 2).
Fragmentation products 7a and benzaldehyde are still formed
in the expected 1:1 and 2:1 stoichiometry relative to disap-
pearence of e-1a; however, under argon-degassed conditions
the predominant reaction pathway becomes e-1a f t-1a
photoisomerization. The suppressed quantum efficiency for
disappearence of e-1 under argon-degassed conditions is likely
due to the fact that there is a significant fraction of bond
ligand.13a Complex e-1a also exhibits a strong band at λmax
254 nm, which is attributed to the π,π* transition of the aniline
≈
moiety.
The absorption spectrum of e-1b (Figure 1b) is distinct from
that of model 2 or diamine e-1a. The most prominent feature
in the absorption of e-1b is a band at λmax ≈ 370 nm that is
considerably more intense than the near-UV MLCT absorption
observed in e-1a and 2. Comparison of the absorption of e-1b
with that of diamine ligand 6a (Figure 6b) indicates that the
370-nm absorption in the metal complex is likely due primarily
to the intraligand π,π* charge transfer (IL-CT) transition of the
conjugated pyridyl-aniline chromophore. It is likely that there
also is some contribution to the near-UV band from an
underlying dπ (Re) f π* (bpy) MLCT transition. Interestingly,
the IL-CT transition of the pyridyl-aniline chromophore is
significantly red-shifted and more intense in e-1b than in free
ligand 6a, which indicates that there is substantial electronic
interaction between the aniline donor and the (bpy)Re(CO)3-
chromophore in the metal complex.
The luminescence properties of e-1a,b, 3a,b, and 6a were
examined in degassed CH3CN solution, and the results are also
summarized in Table 1. Complex 2 exhibits a moderately
intense and long-lived yellow luminescence which is typical of
the (bpy)ReI(CO)3-chromophore and is attributed to the dπ (Re)
3
f π* (bpy) MLCT excited state.19 By contrast, each of the
electron donor-substituted complexes e-1a,b and 3a,b are
virtually nonluminescent. Studies carried out with steady state
and time-resolved photon counting indicate that in each case
3
the upper limit for the MLCT emission quantum yield and
lifetime is 10-4 and 0.5 ns, respectively. This observation
indicates that a nonradiative decay process is operative in the
electron donor-substituted complexes, which accelerates the
3
decay of the MLCT state by greater than 102 relative to the

5412 J. Phys. Chem., Vol. 100, No. 13, 1996
Wang and Schanze
TABLE 1: Photophysical Dataa
absorption
emission
Φem
complex
λ
max/nm (ꢀmax/103 M-1 cm-1
)
assignment
λmax/nm
τem/ns
e-1a
254 (38.5)
308 (17.8)
320 (17.2)
350 (5.3)
242 (29.9)
312 (15.9)
320 (18.8)
370 (40.1)
250 (20.0)
266 (19.4)
306 (11.9)
320 (12.4)
350 (3.7)
262 (29.8)
308 (14.4)
320 (12.8)
350 (4.6)
244 (28.1)
310 (14.4)
320 (16.5)
382 (40.0)
318 (25.0)
π,π* IL (aniline)
π,π* bpy
b
<10-4 b
<0.5b
<0.5b
210
π,π* bpy
Re f bpy MLCT
π,π* IL
π,π* bpy
π,π* bpy
e-1b
b
<10-4 b
IL CT & Re f bpy MLCT
π,π* IL
2
595
4.5 × 10-2
π,π* IL
π,π* bpy
π,π* bpy
Re f bpy MLCT
π,π* IL (aniline)
π,π* bpy
3a
3b
6a
b
b
<10-4 b
<10-4 b
<0.5b
<0.5b
1.1
π,π* bpy
Re f bpy MLCT
π,π* IL
π,π* bpy
π,π* bpy
IL CT & Re f bpy MLCT
π,π* CT
392
a All data in degassed CH3CN solution at 25 °C. b Emission too weak for accurate determination.
As for complex e-1a, irradiation of e-1b in argon-degassed
CH3CN leads to a more complex reaction mixture. Benzalde-
hyde and 7b are still formed under argon-degassed conditions;
however, a new compound with a slightly longer HPLC
retention time compared to that of e-1b is detected. Although
this compound was not isolated and identified by semiprepara-
tive irradiation, the close similarity of the HPLC chromatograms
for the e-1a and e-1b air-saturated photoproduct solutions
strongly implies that the unidentified peak at longer retention
time is the threo diastereomer, t-1b. Quantum yield studies
indicate that the efficiency for disappearance of e-1b is
suppressed under argon-degassed conditions and that the
dominant reaction pathway appears to be e-1b f t-1b photo-
isomerization (Table 2).
Transient Absorption Spectroscopy. Nanosecond-micro-
second laser flash photolysis studies were carried out on e-1a,b
and 3a,b to examine the spectroscopic properties and dynamics
of the excited states and reactive intermediates involved in the
photochemistry and photophysics of these complexes. Pico-
second pump-probe studies were also carried out on e-1a and
3a, and the findings of this work are described elsewhere.13c
All nanosecond-microsecond laser flash photolysis studies were
carried out by using the third harmonic of a Nd:YAG laser for
excitation (355 nm, 10 ns fwhm, 10 mJ/pulse). Listed delay
times correspond to delay after the leading edge of the laser
excitation pulse.
Figure 1. UV-visible absorption spectra of CH3CN solutions: (a)
solid line, 2; broken line, e-1a; (b) solid line, e-1b; broken line, 6a.
fragmentation events which are unproductive due to recombina-
tion to form the starting complex e-1a.17
Steady state photochemical studies of e-1b were also effected
by using 366-nm photolysis, which excites largely the IL-CT
transition of the pyridyl-aniline chromphore of the diamine
ligand. However, despite this feature, the product distribution
and overall quantum efficiency of the photochemistry of e-1b
is remarkably similar to that of e-1a (see Scheme 2 and Table
2). Irradiation of e-1b in air-saturated CH3CN solution leads
to exclusive formation of benzaldehyde and pyridyl-aniline
complex 7b, which were identified by comparison of HPLC
retention times for peaks in the photoproduct solution with
HPLC retention times of authentic samples of the two com-
pounds. Benzaldehyde and 7b presumably form via a pathway
analogous to that described for fragmentation of the diamine
ligand in e-1a (Scheme 3). Quantum yield studies indicate that
benzaldehyde is formed in approximately 2:1 stoichiometry
relative to disappearence of e-1b, as anticipated on the basis of
the reaction stoichiometry. The overall quantum efficiency for
disappearence of e-1b under air-saturated conditions is slightly
lower compared to that of e-1a, but the reaction remains
comparatively efficient with Φ- ) 0.43.
Figure 2a illustrates the transient absorption spectrum of 3a
obtained at 10-ns delay following the leading edge of the laser
pulse. This spectrum is characterized by a comparatively weak,
broad absorption centered at λmax ) 490 nm (∆Amax ) 0.078).
The transient absorption decays too rapidly for accurate deter-
mination of the decay rate with the nanosecond apparatus (see
inset to Figure 2a); however, picosecond pump-probe spec-
troscopy indicates that the decay rate of this transient is 1 ×
108 s-1 13c
The transient absorption spectrum of 3a is substan-
.
tially different from that of the dπ (Re) f π* (bpy) MLCT
excited state in model complex 2, which displays a strong,
narrow absorption band with λmax ≈ 370 nm.20 On this basis,
the transient absorption spectrum of 3a is attributed to LLCT
excited state 16a (eq 3), which is formed by a sequence that
involves initial excitation to the MLCT state 15a, followed by

Photoinduced ET in Metal-Organic Complexes
J. Phys. Chem., Vol. 100, No. 13, 1996 5413
SCHEME 2
TABLE 2: Photochemical Quantum Efficienciesa
air saturated
argon degassed
complex
Φ-
Φthreo
ΦPhCHO
Φamine
Φ-
Φthreo
ΦPhCHO
Φamine
e-1a
e-1b
0.54 ( 0.03
0.43 ( 0.02
<0.01
<0.01
1.10 ( 0.05
0.75 ( 0.04
0.48 ( 0.01
b
0.29 ( 0.04
0.26 ( 0.03
0.16 ( 0.01
0.09 ( 0.02
0.19 ( 0.02
0.14 ( 0.02
0.08 ( 0.01
b
a All data for CH3CN solutions at 25 °C. Φ-, quantum yield for disappearence of starting complex; Φthreo, quantum yield for erythro f threo
photoisomerization; ΦPhCHO, quantum yield for benzaldehyde formation; Φamine, quantum yield for formation of 7a or 7b. b Not determined.
SCHEME 3
rapid forward ET (kFET . 109 s-1). An important point for the
present study is that the picosecond measurement of the decay
rate of LLCT state 16a indicates that the rate of back ET in
kinetic findings are summarized in Figure 2b. The transient
absorption of this complex is dominated by a moderately strong
bleach at λmax ≈ 390 nm and weak aborption in the mid-visible.
The bleach at 390 nm corresponds to the ground state absorption
band that is assigned to the π,π* IL-CT absorption of the
pyridyl-aniline donor ligand. Two observations suggest that
the transient absorption spectrum observed for 3b is due to the
LLCT state 16b, eq 3. First, the spectrum differs significantly
from that of the dπ (Re) f π* (bpy) MLCT state.20 Second,
the strong bleach of the IL-CT absorption band indicates that
there is depletion of electron density at the aniline donor in the
transient, which is consistent with the electronic structure of
this aniline-substituted complex is kBET ) 1 × 108 s-1 13c
.
Nanosecond transient absorption spectroscopy was carried out
on pyridyl-aniline complex 3b, and the spectroscopic and

5414 J. Phys. Chem., Vol. 100, No. 13, 1996
Wang and Schanze
Figure 2. Transient absorption difference spectra following laser
excitation (355 nm, 10 ns fwhm, 10 mJ/pulse). All spectra are for
degassed CH3CN solutions. Delays refer to delay time from leading
edge of laser excitation pulse. (a) 3a, 10 ns delay. Inset: transient
absorption kinetics of 3a at 470 nm. (b) 3b, 10 ns delay. Inset:
transient absorption kinetics of 3b at 390 nm.
Figure 3. Transient absorption difference spectra of e-1a in degassed
CH3CN following laser excitation (355 nm, 10 ns fwhm, 10 mJ/pulse).
Delays refer to delay time from leading edge of laser excitation pulse.
(a) Delay times: 5 ns (-9-), 100 ns (- -), 200 ns (‚‚‚), and 1000 ns
(- - -). INSET: transient absorption kinetics at 460 nm. (b) Delay
times (in order of decreasing transient absorption intensity): 5, 50, 100,
200, 250, and 325 µs. INSET: transient absorption kinetics at 365
nm.
LLCT state 16b. The transient absorption of 3b decays too
rapidly for accurate determination by using the nanosecond
system (see inset to Figure 2b); however, the fact that a
moderately strong transient absorption is observed with nano-
second laser excitation implies that the decay rate of the transient
is e109 s-1
.
Figure 3 provides a summary of the transient absorption data
for e-1a in degassed CH3CN solution. Parts a and b of Figure
3 illustrate spectra at delay times ranging from 5 to 1000 ns
and 5 to 325 µs, respectively, while the insets to parts a and b
of Figure 3 illustrate the temporal profile of the transient
absorption at 460 and 365 nm, respectively. Taken together,
these data clearly indicate that near-UV excitation of e-1a leads
to rapid formation of a strongly absorbing transient with λmax
≈ 360 nm. For reasons that will be fully outlined below, this
transient absorption spectrum is assigned to metal complex
radical 10a (Scheme 3) that is produced by C-C bond
fragmentation of the diamine ligand in e-1a. The 360-nm
transient is formed almost completely during the laser pulse;
however, the early time transient absorption data (refer to Figure
3a) imply that there is a spectral evolution that occurs which is
slightly too fast to be temporally resolved by the nanosecond
system. Recently published picosecond transient absorption
studies of e-1a allow resolution of these fast dynamics and
indicate that the rise time of the 360-nm absorption and the
fast decay of the absorption at 460 nm (see inset, Figure 3a)
occur with the same first-order rate of k ) 4 × 108 s-1 13c
.
The microsecond time domain data shown in Figure 3b
indicate that the 360-nm transient attributed to metal complex
radical 10a decays comparatively slowly. Kinetic analysis
reveals that the decay can be modeled by assuming second-
order, equal concentration kinetics. A fit of the data provides
k2/∆ꢀ ) 3.0 × 104 cm s-1, where k2 is the second-order decay
rate constant (units ) M-1 s-1) and ∆ꢀ is the difference molar
absorptivity (units ) M-1 cm-1) of the transient absorption at
360 nm. Assuming that ∆ꢀ ≈ 2 × 104 M-1 cm-1 leads to an
Figure 4. Transient absorption difference spectra of e-1b in degassed
CH3CN solution following laser excitation (355 nm, 10 ns fwhm, 10
mJ/pulse). Delays refer to delay time from leading edge of laser
excitation pulse. (a) Delay times (in order of increasing transient
absorption intensity at 425 nm): 10, 30, 50, and 100 ns. Inset: transient
absorption kinetics at 370 and 420 nm. (b) Delay times (in order of
decreasing transient absorption intensity at 425 nm): 5, 50, 100, 200,
250, and 325 µs. Inset: transient absorption kinetics at 370 and 420
nm.
estimate of k2 ≈ 6 × 108 M-1 s-1
.
for e-1b in degassed CH3CN solution. Parts a and b of Figure
4 illustrate spectra at delay times ranging from 5 to 100 ns and
Figure 4 provides a summary of the transient absorption data

Photoinduced ET in Metal-Organic Complexes
J. Phys. Chem., Vol. 100, No. 13, 1996 5415
TABLE 3: Electrochemical Potentialsa
compound
Epox/Vb E1/2red/Vc ∆GFET/eV ∆GBET/eV
5 to 325 µs, respectively, while the insets to parts a and b of
Figure 4 illustrate the temporal profile of the transient absorption
at 420 and 370 nm, respectively. The transient absorption
spectra clearly indicate that photoexcitation of e-1b leads very
rapidly to (1) significant bleaching of the ground state absorption
band at 370 nm, which is due to the IL-CT absorption of the
pyridyl-aniline chromophore, and (2) production of a strongly
absorbing transient with an apparent absorption maximum at λ
≈ 425 nm. For reasons that will be fully explained in the
Discussion section, the transient absorption features observed
for e-1b are ascribed to photoinduced fragmentation of the
diamine ligand and production of metal complex radical 10b
(Scheme 3). Bleaching of the ground state absorption at 370
nm is due to loss of the IL-CT absorption of the pyridyl-aniline
chromophore, while the transient absorption feature at 425 nm
is assigned to 10b. While the photoinduced fragmentation is
very fast on the nanosecond time scale, the early time data
(Figure 4a) indicate that there is a dynamic process which occurs
during and immediately following the 10-ns laser pulse.
Significantly, the spectrum at the earliest delay time (10-ns
delay) exhibits a fully developed bleach at 370 nm. By contrast,
the 425-nm transient absorption rises comparatively more
slowly, taking ≈30 ns to fully develop. These features suggest
that bond fragmentation (e.g., 9b f 10b) occurs on a time scale
that is very close to the temporal resolution of the nanosecond
flash photolysis system (e.g., kBF < 109 s-1).
The microsecond time domain transient absorption data for
e-1b (Figure 4b) reveal that the transient absorption and bleach
signals that are attributed, respectively, to 10b and depletion of
e-1b decay within approximately 500 µs of the pulse. Interest-
ingly, a substantial fraction of the ground state bleaching in the
near-UV region recovers, which is consistent with recombination
of the radicals formed by bond fragmentation to regenerate the
starting complex. The transient absorption and bleach recovery
kinetics can be modeled by assuming second-order, equal
concentration kinetics. A fit of the transient absorption kinetic
data at 420 nm provides k2/∆ꢀ ) 2.0 × 105 cm s-1. Assuming
that ∆ꢀ420 ≈ 2 × 104 M-1 cm-1 leads to k2 ≈ 4 × 109 M-1 s-1
for recovery of the bleach signal.
e-1a
e-1b
2
+0.84
+0.85
-1.17
-1.17
-1.16
-1.17
-0.37
-0.36
-2.01
-2.02
3a
+0.82f
+0.97
+0.72g
-0.39
-1.99
N-benzylpiperidine
N-methylaniline
N,N-dimethylaniline +0.77g
a All potentials for CH3CN solutions with 0.1 M TBAH supporting
electrolyte. Potentials relative to saturated sodium chloride calomel
reference electrode. b Peak potential for irreversible anodic wave at 100
mV/s sweep rate. c Half-wave potential for reversible cathodic wave.
d Free energy change for photoinduced forward electron transfer. e Free
f
energy change for back electron transfer. Half-wave potential for quasi-
reversible anodic wave. g Potentials from ref 30.
of peak potentials for irreversible oxidation of N-benzylpiperi-
dine, N-methylaniline, and N,N-dimethylaniline.21 This data
reveals that the irreversible oxidation of N-benzylpiperidine (a
3° aliphatic amine) is shifted several hundred millivolts anodic
compared to the oxidation of N-methylaniline and N,N-di-
methylaniline (typical aromatic amines).
All of the available information suggests that the oxidation
of the diamine ligands in e-1a,b is localized primarily at the
aniline unit (and not at the piperidine site). First, the peak
anodic potentials of e-1a,b occur at approximately the same
potential as the quasi-reversible oxidation of the N,N-dimethyl-
aniline unit in 3a. Note that the peak anodic potentials for
e-1a,b and 3a are slightly more positive than the oxidation
potentials of N-methylaniline or N,N-dimethylaniline; however,
the positive shift of the aniline-based oxidation in the metal
complexes is likely due to a net +1 charge on the complexes.
Despite the slight positive shift, the anodic waves for e-1a and
e-1b still occur at a less positive potential compared to oxidation
of the piperidine model compound N-benzylpiperidine, which
further substantiates the view that the oxidation in the diamine
complexes occurs at the aniline donor unit.
The free energy change that accompanies photoinduced
forward and back ET (∆GFET and ∆GBET, respectively) in e-1a,b
and 3a,b can be estimated by the following expressions:22
Nanosecond transient absorption studies carried out on e-1a
and e-1b in air-saturated CH3CN solution indicate that under
these conditions metal complex radicals 10a and 10b are formed
by laser excitation. However, in the presence of O2 these
transients are quenched exceedingly rapidly, with k ≈ 1 × 107
s-1. Assuming that [O2] ≈ 1.9 mM in air-saturated CH3CN
solution, the second-order rate constant for reaction of these
radicals with O2 is ≈5 × 109 M-1 s-1, which is close to the
diffusion-controlled limit in CH3CN.
∆GFET ≈ Ep(D/D•+) - E1/2(bpy/bpy•-) - E00 (4a)
∆GBET ≈ E1/2(bpy/bpy•-) - Ep(D/D•+)
(4b)
where Ep(D/D•+) is the peak oxidation potential of the aniline
or diamine donor ligand, E1/2(bpy/bpy•-) is the half-wave
reduction potential of the bpy acceptor ligand, and E00 is the
0-0 energy of the relaxed MLCT excited state of the (bpy)-
ReI(CO)3-chromophore. By using the electrochemical potentials
listed in Table 3 along with E00 estimated from the emission
spectrum of 2 in CH3CN (E00 ≈ 17 550 cm-1),16 ∆GFET and
∆GBET for each complex have been estimated, and the values
are listed in Table 3. These calculations indicate that for each
complex forward ET is moderately exothermic and back ET is
strongly exothermic.
Electrochemistry and Thermodynamics for Intramolecu-
lar ET. Cyclic voltammetry was carried out on e-1a,b, 2, 3a,
and various model amines to determine (1) the potentials for
oxidation of the diamine (or amine) donor ligand and for
reduction of the bpy acceptor ligand and (2) to attempt to
identify the site at which oxidation occurs within the diamine
unit. Table 3 lists the electrochemical potentials for these
compounds relative to the saturated sodium chloride calomel
electrode (SSCE). First, cathodic scans for each of the metal
complexes reveal a reversible wave at E1/2 ≈ -1.17 V. In each
case this wave is ascribed to reduction of the coordinated bpy
ligand.19 Anodic scans for the two diamine complexes e-1a
and e-1b and the dimethylaniline-substituted complex 3a display
an irreversible wave each, with Ep ≈ +0.85 V. This wave is
not observed in model complex 2 (which does not contain an
amine donor ligand), and on this basis it is attributed to oxidation
of the aniline or diamine donor. Table 3 also contains a listing
Discussion
Overall Mechanism for Photochemistry. Scheme 3 pre-
sents an overall mechanism that is consistent with the observed
photochemical and photophysical properties of e-1a,b. Initial
photoexcitation produces the relaxed dπ (Re) f π* (bpy) MLCT
excited state 8.23 The MLCT state relaxes via radiative and
nonradiative decay to the ground state (kd path) or by forward
ET from the diamine donor to the photoexcited Re(I) center

5416 J. Phys. Chem., Vol. 100, No. 13, 1996
Wang and Schanze
TABLE 4: Electron Transfer and Bond Fragmentation
Ratesa
(kFET path). Forward ET produces LLCT state 9. In a one-
electron approximation, 9 comprises a bpy anion radical (e.g.,
bpy•-) and a diamine radical cation held in proximity by the
transition metal center. LLCT state 9 relaxes either by back
ET (kBET path) via transfer of an electron from bpy•- to the
diamine radical cation or by C-C bond fragmentation of the
diamine radical cation (kBF path). Bond fragmentation within
9 occurs in a formally heterolytic fashion (Vide infra) to produce
metal complex radical 10 and iminium ion 11, which is derived
from the piperidine portion of the diamine unit. Under air-
saturated conditions radical 10 is quenched by rapid ET to O2,13b
followed by loss of a proton to form imine complex 14. Imine
complex 14 and iminium ion 11 hydrolyze under HPLC analysis
conditions to form the corresponding free amines (7 and
piperidine) and 2 equiv of benzaldehyde.
Under argon-degassed conditions, metal complex radical 10
survives for a sufficient period to allow it to react in a
bimolecular reaction with iminium ion 11 in a two-step sequence
that results in net regeneration of starting complex e-1a (or e-1b)
and the corresponding diastereomer t-1a (or t-1b). The first
step of this recombination sequence presumably involves ET
from bpy•- in 10 to iminium ion 11, and the second step
involves stereorandom recombination of 12 and 13 to produce
the diastereomeric complexes e-1 and t-1 in approximately
equivalent proportions.17 The mechanism of this recombination
reaction will be discussed in more detail in a succeeding section.
Kinetics of ET and Bond Fragmentation in e-1a and e-1b.
The kinetics for forward ET in the donor-substituted complexes
are given by the expression22
compound
kFET/s-1
kBET/s-1
kBF/s-1
e-1a
e-1b
3a
>2 × 109
>2 × 109
>2 × 109
1.0 × 108
1.3 × 108
1.0 × 108
3.0 × 108
3.0 × 108
a Data for CH3CN solution. Rates estimated as described in text.
3, the overall efficiency for bond fragmentation (φBF) is given
by the product of the efficiencies for formation of LLCT state
9b by forward ET (φLLCT) and for formation of the radical 10b
by bond fragmentation (âBF),25
φBF ) φLLCTâBF
(6a)
where
kFET
kFET + kd
kBF
φLLCT
)
and âBF )
(6b)
kBF + kBET
Transient absorption indicates that O2 is an effective scavenger
for radical 10b, which is the primary product of bond fragmen-
tation. Therefore, it is reasonable to assume that bond frgmen-
tation is irreversible under air-saturated conditions and hence
air
that φBF ) Φ-air, where Φ- is the quantum yield for
disappearence for e-1b in air-saturated solution. Taking this
into account, rearranging and combining eqs 6a and 6b leads
to eq 7,
air
kBET ) kBF(1 - Φ-
)
(7)
kFET ) 1/τ - 1/τ(2)
(5)
air
where it has also been assumed that φBF ) Φ- and φLLCT
)
where τ and τ(2) are, respectively, the MLCT emission lifetimes
of the donor-substituted complexes and model complex 2.
Emission lifetime studies indicate that the MLCT emission
lifetimes of all of the donor-substituted complexes (e.g., e-1a,b
and 3a,b) are less than 0.5 ns (Table 1). Using eq 5 and the
MLCT emission lifetime of model complex 2 (210 ns) leads to
the conclusion that kFET g 2 × 109 s-1 in each of the donor-
substituted complexes.
1. The latter assumption is justified by the fact that kFET . kd.
Now, substituting kBF ) 1.0 × 108 s-1 and Φ- ) 0.43 into
air
eq 7 leads to kBET ≈ 1.3 × 108 s-1 for e-1b.
A summary of the rate constants for forward ET, back ET,
and bond fragmentation in e-1a, e-1b, and 3a is provided in
Table 4. Several points merit discussion concerning these data.
First, given that forward ET is moderately exothermic (≈-0.4
eV) and that the separation distance between the aniline donor
and the Re(I) center is relatively small (ca. 5-7 Å), the fact
that kFET g 2 × 109 s-1 is not unreasonable. Further, the rapid
forward ET rates observed for e-1a,b and 3a are consistent with
The rate of back ET (kBET) was directly determined in model
complex 3a to be 1 × 108 s-1 by using picosecond transient
absorption.13c While kBET cannot be directly determined in
diamine complex e-1a, it is reasonable to assume that 3a serves
as an excellent model for e-1a and, therefore, that kBET ≈ 1.0
× 108 s-1 for e-1a as well. Further support for this conclusion
comes from the fact that the rate of back ET is very similar in
the structurally similar aminoalcohol complex 17 (Chart 2),
the donor-acceptor electronic coupling matrix element (HAB
)
and reorganization energy (λ) determined by a study of the
temperature dependence for photoinduced forward ET in the
structurally similar donor-substituted complex 18 (Chart 2),
where HAB ) 7 cm-1 and λ ) 1.0 eV.12d
where kBET ) 8.3 × 107 s-1 13a
.
The rates for back ET in e-1a,b and 3 are comparable to
values previously observed for strongly exothermic back ET in
the related Re(I) complexes 18 (kBET ) 1.3 × 107 s-1, ∆GBET
) -2.13 eV)13d and 19 (kBET ) 4.0 × 107 s-1, ∆GBET ) -1.96
eV).7b However, despite the similarity among the rates for
charge recombination in these structurally related Re(I) com-
plexes, there are some surprising features. First, electronic
coupling between the Re center and the aniline donor unit is
likely to be significantly larger in e-1b than in e-1a because of
the structural difference in the donor ligands for the two
complexes.26,27 Evidence of strong donor ligand-metal elec-
tronic coupling in e-1b is provided by the significant shift in
the energy of the π,π* IL-CT absorption of the pyridyl-aniline
ligand in complex e-1b compared to the energy of the same
transition in free ligand 6a. However, despite the significant
difference in donor ligand-metal coupling between e-1a and
e-1b, kBET is virtually the same within experimental error in
the two complexes. Another surprise is that kBET for the group
With a value for kBET for e-1a in hand, it is possible to
determine kBF for e-1a from the picosecond transient absorption
data. The kinetic model presented in Scheme 3 indicates that
disappearance of 9a and appearance of 10a should occur with
the same first-order rate constant, kobs, where kobs ) kBET
+
kBF. Consistent with this premise, the picosecond transient
absorption studies reveal that the decay of LLCT state 9a and
the grow-in of radical 10a both occur with kobs ) 4 × 109 s-1 13c
.
Hence, the assumption that kBET ) 1.0 × 108 s-1 affords a value
of kBF ) 3.0 × 108 s-1 for e-1a.
Picosecond transient absorption studies have not been carried
out on e-1b or the relevant model complex 3b to evaluate kBET
or kBF in these complexes. However, by assuming that the rate
of bond fragmentation is the same in the LLCT excited state of
both e-1a and e-1b (kBF ) 3.0 × 108 s-1),24 it is possible to
apply the steady state photochemical results on e-1b to provide
an estimate for kBET in this complex. By reference to Scheme

Photoinduced ET in Metal-Organic Complexes
J. Phys. Chem., Vol. 100, No. 13, 1996 5417
of donor-substituted complexes e-1a,b, 3a, 18, and 19 (and
others) is comparatively slow compared to the rates of charge
recombination in other donor-acceptor systems where the
driving force and electronic coupling are comparable. For
example, back ET in covalently linked porphyrin-quinone
systems with donor-acceptor separation distances ranging from
5 to 10 Å and with ∆GBET ≈ 1.5 eV typically occurs with kBET
SCHEME 4
≈ 109-1010 s-1 28
.
Taken together, the features outlined above imply that the
rate of back ET in donor-substituted Re(I) complexes such as
1a,b, 3a, 18, and 19 may be controlled, at least in part, by the
dynamics of triplet f singlet intersystem crossing in the LLCT
state. Intersystem crossing may play a role in determining the
rate of back ET in the LLCT state because it is formed by
forward ET from the MLCT state, which is believed to possess
absorption maximum of the RC6H4NHC˙ HC6H5 chromophore
is red-shifted substantially in 10b compared to 10a because of
the effect of conjugation between the pyridine and aniline rings.
Note that R-amino radical 13, which would be produced by
homolytic fragmentation, exhibits a very strong absorption at
λmax ≈ 340 nm.13b This band does not appear in the spectra of
either e-1a or e-1b, as would be expected if homolytic
fragmentation occurs.
A thermodynamic cycle that can be used to estimate the free
energy for fragmentation of diamine radical cation 21 is
illustrated in Scheme 4. The free energy can be calculated by
eq 9:32
3
3
considerable triplet spin character (e.g., MLCT f LLCT).29
Since the product of back ET (the ground state) has singlet spin
multiplicity, intersystem crossing must precede decay of 3LLCT
via back ET.
That spin state dynamics play an important role in determining
the lifetimes of charge-separated states formed by ET quenching
in d6 transition metal complex systems has been recently
demonstrated by studies of heavy atom and magnetic field
3
effects on ET reactions between Ru(bpy)32+* and electron
donors or acceptors.30 These studies indicate that the geminate
radical pairs formed by oxidative or reductive ET quenching
3
of Ru(bpy)32+* have a high degree of triplet spin character
∆GCC•+ ) ∆GCC - F[E1/2(11/13) - E1/2(21/23)] (9)
and, furthermore, that back ET in the geminate pairs is controlled
largely by the dynamics of intersystem crossing.30 Since
spectroscopic studies indicate that the orbital and spin charac-
•+
where ∆GCC and ∆GCC are, respectively, the free energies for
bond fragmentation of radical cation 21 and the corresponding
neutral 23, the E1/2’s represent electrochemical half-wave
potentials for the indicated one-electron reductions, and F is
the Faraday constant. E1/2(21/23) is given approximately by
Ep for e-1a and e-1b (+0.85 V), while E1/2(11/13) is available
from the literature (-0.94 V).33 An estimate for ∆GCC is arrived
at through the following reasoning. First, the literature value
for ∆HCC for the 1,2-CC bond in N,N,N′-trimethyl-1,2-diami-
noethane is +60 kcal/mol.34 Analysis of bond homolysis data
for a variety of aryl-substituted ethanes indicates that vicinal
diphenyl substitution typically decreases ∆HCC by ≈22 kcal/
mol compared to the corresponding ethane.34 These values lead
to an estimate of ∆HCC ≈ 38 kcal/mol for e-1a and e-1b.
Finally, an estimate of ∆SCC ≈ 36 eu for e-1a and e-1b is based
on ∆SCC values for ethane (38 eu), 1,2-diphenylethane (36 eu),
and 1,2-diaminoethane (36 eu).35 By using the estimated values
teristics of the luminescent 3MLCT excited states of Ru(bpy)3
2+
and (bpy)Re(CO)3L+ complexes are virtually identical,29 it is
reasonable to expect that intersystem crossing dynamics will
play an important role in photoinduced ET reactions in the
(bpy)Re(CO)3L+ system as well.
Mechanism and Thermodynamics for Bond Fragmenta-
tion in the LLCT State. In Scheme 3, C-C bond fragmenta-
tion of the diamine radical cation within LLCT state 10 is
implicitly considered to occur in a heterolytic mode. However,
as illustrated by eq 8, fragmentation may occur in either a
heterolytic or homolytic mode. These two bond fragmentation
modes differ in outcome: heterolytic fragmentation (eq 8a)
would produce metal complex 10 and piperidine iminium ion
11, while homolytic fragmentation (eq 8b) would produce metal
complex iminium ion 20 and piperidine R-amino radical 13.
for ∆HCC and ∆SCC for e-1a and e-1b, we calculate that ∆GCC
-
(298 K) ≈ +27 kcal/mol.
Substitution of the estimated values for E1/2(21/23), E1/2(11/
•+
13), and ∆GCC into eq 9 leads to a calculated value of ∆GCC
) -14 ( 5 kcal/mol for 21 (and by analogy for the radical
cations of e-1a and e-1b), where the estimated error of 5 kcal/
mol takes into account errors arising from estimation of ∆GCC
-
(298 K) and from using an irreversible peak potential for e-1a
and e-1b rather than a reversible potential. The estimate of
•+
∆GCC ≈ -14 kcal/mol indicates that bond fragmentation of
the diamine cation radical is moderately exothermic. In view
of this fact, it is remarkable that the reaction is thermally
activated: the rate of bond fragmentation at 298 K (3 × 108
CC•+
) of +3 kcal/
The transient absorption spectra of the radical products
produced by C-C bond fragmentation in e-1a and e-1b (Figure
3a,b) indicate that heterolytic fragmentation predominates or
occurs exclusively. The spectra produced by photolysis of e-1a
and e-1b exhibit absorption bands with maxima at 365 and 425
nm, respectively. These absorption bands are attributed to the
R-amino radical chromophore, RC6H4NHC˙ HC6H5,13b,17,31 which
is present in metal complex radicals 10a and 10b. The
s-1) indicates an activation free energy (∆G*
mol. That fragmentation of the diamine cation radical is weakly
activated despite being moderately exothermic implies that bond
cleavage has an associated “reorganization energy”, or intrinsic
barrier. This energetic barrier may be due to free energy costs
associated with the stereoelectronic demands of the transition
state36,37 and solvent repolarization that occurs along the reaction
coordinate. Indeed, the fragmentation kinetics of diastereomeric

5418 J. Phys. Chem., Vol. 100, No. 13, 1996
Wang and Schanze
aminoalcohol cation radicals (eq 1, X ) OH) imply that in the
transition state for fragmentation there is a stereoelectronic
requirement for maximization of orbital overlap between the p
orbital on the heteroatom of the electrofugal group (X in eq 1)
and the singly occupied p orbital on N.14,31 Overlap is
maximized if the two heteroatoms are oriented either syn or
anti with respect to rotation about the 1,2-CC bond. It is
plausible that the intrinsic barrier for fragmentation of the 1,2-
diamine cation radical is due to an unfavorable enthalpy or
probability factor (activation entropy) associated with achieve-
ment of the proper orbital alignment in the transition state.36,37
Mechanism for e-1 to t-1 Isomerization. One of the
surprising outcomes of the photochemical studies of e-1a and
e-1b is the observation of erythro f threo photoisomerization.
This reaction provides unequivocal evidence that bond frag-
mentation is reversible, albeit in an operational but not neces-
sarily strictly mechanistic sense, Vide infra. An important
observation is that e-1 f t-1 isomerization is not observed under
air-saturated conditions (see Φthreo, Table 2), which indicates
that geminate recombination of R-amino radical 10 and iminium
11 does not occur. That geminate recombination does not occur
implies that direct recombination of 10 and 11 to produce 9 is
not feasible.38 Furthermore, it suggests that recombination
requires a preceding ET step from bpy•- in 10 to iminium ion
11 to produce the radicals 12 and 13, which then recombine
(Scheme 3). The thermodynamic driving force for this outer
sphere ET reaction is estimated as ∆G ≈ -0.23 eV on the basis
of the reduction potentials of the coordinated bpy ligand (E1/2
) -1.17 V, Table 3) and iminium ion 11 (E1/2 ≈ -0.94 V).33
A subtle, but interesting point to consider is that the lack of
geminate recombination indicates that bpy•- to iminium ET does
not occur within the geminate pair [10,11] that is initially
produced by C-C bond fragmentation. The lack of ET at this
stage may be due to poor electronic coupling between the donor
(e.g., bpy•-) and the iminium ion, which are likely to be
separated by 10-15 Å immediately following bond fragmenta-
tion.
the conclusion that recombination does not occur between the
primary radical-ion pair formed by fragmentation of the
diamine radical cation, but rather involves a two-step sequence
involving electron transfer followed by radical-radical recom-
bination.
A primary objective of this study was to develop a reactive
donor ligand system which is useful for probing the electronic
structure and dynamics of charge transfer excited states in metal
complex systems. The 1,2-diamine ligands which provide the
basis for the unique reactivity of e-1a and e-1b clearly satisfy
this objective. Future studies seek to apply the diamine and
other reactive donor systems to explore the electronic structure,
dynamics, and reactivity of MLCT, LLCT, and ion-pair CT
states in inorganic and organometallic complexes.
Acknowledgment. We gratefully acknowledge support for
this project from the National Science Foundation (Grant Nos.
CHE 91-23000 and CHE-9401620).
Supporting Information Available: HPLC chromatograms
for photolyzed solutions of e-1a and e-1b under argon-degassed
and air-saturated conditions (2 pages). See any current masthead
page for ordering information.
References and Notes
(1) (a) Bock, C. R.; Meyer, T. J.; Whitten, D. G. J. A. Chem. Soc.
1974, 96, 4710. (b) Young, R. C.; Meyer, T. J.; Whitten, D. G. J. Am.
Chem. Soc. 1975, 97, 4782. (c) Bock, C. R.; Connor, J. A.; Guitierrez, A.
R.; Meyer, T. J.; Whitten, D. G.; Sullivan, B. P.; Nagle, J. K. J. Am. Chem.
Soc. 1979, 101, 4815.
(2) (a) Sutin, N.; Creutz, C. J. Chem. Ed. 1983, 10, 809. (b) Marcus,
R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265.
(3) (a) Scandola, F.; Balzani, V. J. Chem. Ed. 1983, 10, 814. (b) Juris,
A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; Von Zelewsky,
A. Coord. Chem. ReV. 1988, 84, 85.
(4) (a) Hoffman, M. Z. J. Phys. Chem. 1988, 92, 3458. (b) Ohno, T.;
Yoshimura, A.; Prasad, D. R.; Hoffman, M. Z. J. Phys. Chem. 1991, 95,
4723.
(5) Photoinduced Electron Transfer; Fox, M. A., Chanon, M. D., Eds.;
Elsevier: Amsterdam, 1988; Parts A-D.
Finally, it is important to note that the behavior of metal
complexes e-1a and e-1b contrasts with that of free ligand 6a
and the related diamine 6b.17 Compounds 6a and 6b undergo
1,2-CC bond fragmentation and erythro f threo stereoisomer-
ization when irradiated in air-saturated solution, presumably
via a mechanism involving geminate recombination. However,
the mechanism for bond fragmentation in 6a and 6b differs in
such a way that a radical pair, rather than a radical-ion pair, is
produced by bond fragmentation. In this case, recombination
directly produces the neutral diamine in a reaction that is
exoergonic by ca. 27 kcal/mol (the C-C bond free energy).
(6) Photoinduced Electron Transfer II; Mattay, J., Ed.; Top. Curr.
Chem. 1990, 158, 1-199.
(7) (a) Westmoreland, T. D.; Le Bozec, H.; Murray, R. W.; Meyer, T.
J. J. Am. Chem. Soc. 1983, 105, 3932. (b) Chen, P.; Westmoreland, T. D.;
Danielson, E.; Schanze, K. S.; Anthon, D.; Neveaux, P. E., Jr.; Meyer, T.
J. Inorg. Chem. 1987, 26, 1116. (c) Chen, P.; Duesing, R.; Graff, D. K.;
Meyer, T. J. J. Phys. Chem. 1991, 95, 5850. (d) Jones, W. E., Jr.; Bignozzi,
C. A.; Chen, P.; Meyer, T. J. Inorg. Chem. 1993, 32, 1167. (e)
Mecklenburg, S. L.; Peek, B. M.; Schoonover, J. R.; McCafferty, D. G.;
Wall, C. G.; Erickson, B. W.; Meyer, T. J. J. Am. Chem. Soc. 1993, 115,
5479.
(8) (a) Elliott, C. M.; Freitag, R. A.; Blaney, D. D. J. Am. Chem. Soc.
1985, 107, 4647. (b) Cooley, L. F.; Headford, C. E. L.; Elliott, C. M.;
Kelley, D. F. J. Am. Chem. Soc. 1988, 110, 6673. (c) Ryu, C. K.; Wang,
R.; Schmehl, R. H.; Ferrere, S.; Ludwikow, M.; Merkert, J. W.; Headford,
C. E. L.; Elliott, C. M. J. Am. Chem. Soc. 1992, 114, 430. (d) Larson, S.
L.; Elliott, C. M.; Kelley, D. F. J. Phys. Chem. 1995, 99, 6530.
(9) (a) Yonemoto, E. H.; Riley, R. L.; Kim, Y. I.; Atherton, S. J.;
Schmehl, R. H.; Mallouk, T. E. J. Am. Chem. Soc. 1992, 114, 8081. (b)
Yonemoto, E. H.; Saupe, G. B.; Schmehl, R. H.; Hubig, S. M.; Riley, R.
L.; Iverson, B. L.; Mallouk, T. E. J. Am. Chem. Soc. 1994, 116, 4786.
(10) (a) Fox, L. S.; Kozik, M.; Winkler, J. R.; Gray, H. B. Science 1990,
247, 1069. (b) Farid, R. S.; Chang, I.-J.; Winkler, J. R.; Gray, H. B. J.
Phys. Chem. 1994, 98, 5176.
(11) Goulle, V.; Harriman, A.; Lehn, J.-M. J. Chem. Soc., Chem.
Commun. 1993, 1034.
(12) (a) Schanze, K. S.; Sauer, K. J. Am. Chem. Soc. 1988, 110, 1180.
(b) Schanze, K. S.; Cabana, L. A. J. Phys. Chem. 1990, 94, 2740. (c)
Perkins, T. A.; Hauser, B. T.; Eyler, J. R.; Schanze, K. S. J. Phys. Chem.
1990, 94, 8745. (d) MacQueen, D. B.; Schanze, K. S. J. Am. Chem. Soc.
1991, 113, 7470.
(13) (a) Wang, Y.; Hauser, B. T.; Rooney, M. M.; Burton, R. D.;
Schanze, K. S. J. Am. Chem. Soc. 1993, 115, 5675. (b) Wang, Y.; Lucia,
L. A.; Schanze, K. S. J. Phys. Chem. 1994, 99, 1961. (c) Lucia, L. A.;
Wang, Y.; Nafisi, K.; Netzel, T. L.; Schanze, K. S. J. Phys. Chem. 1995,
99, 11801.
Summary and Conclusions
The photophysics and photochemistry of diamine-substituted
complexes e-1a and e-1b have been carefully examined.
Photoexcitation of these complexes affords the LLCT state with
high efficiency via a sequence involving initial population of
the MLCT state followed by intramolecular diamine donor to
metal ET. The diamine radical cation which is present in the
LLCT state undergoes irreversible carbon-carbon bond frag-
mentation with kBF ) 3 × 108 s-1, while the LLCT state decays
by back ET with kBET ) 1 × 108 s-1. Bond fragmentation
competes very effectively with back ET in both e-1a and e-1b,
and as a result, the overall quantum efficiency for irreversible
photochemistry of the complexes is high. The surprising
observation of erythro f threo photoisomerization for e-1a and
e-1b indicates that bond fragmentation is reversible. However,
the fact that geminate recombination does not occur leads to

Photoinduced ET in Metal-Organic Complexes
J. Phys. Chem., Vol. 100, No. 13, 1996 5419
(14) Ci, X.; Kellett, M. A.; Whitten, D. G. J. Am. Chem. Soc. 1991,
113, 3893.
(25) Equation 5a assumes that the lowest MLCT excited state is
populated with unit quantum efficiency following excitation.
(26) The difference in electronic coupling between two aromatic rings
which are directly attached (e.g., conjugated) or separated by a single
methylene spacer is examplified by studies of the intensity of intervalence
charge transfer absorption bands in mixed-valence complexes of the type
(NH3)5RuII-L-RuIII(NH3)53+, where L ) 4,4′-bipyridine or bis(4-pyridyl)-
methane. The electronic coupling matrix element is ca. 10 fold times larger
in the 4,4′-bipyridine bridged complex compared to that of the bis(4-pyridyl)-
methane complex.27
(27) (a) Fischer, H.; Tom, G. M.; Taube, H. J. Am. Chem. Soc. 1976,
98, 5512. (b) Rieder, K.; Taube, H. J. Am. Chem. Soc. 1977, 99, 7891.
(28) Wasielewski, M. R. Chem. ReV. 1992, 92, 435.
(29) Striplin, D. R.; Crosby, G. A. Chem. Phys. Lett. 1994, 221, 426.
(30) (a) Steiner, U. E.; Wolff, H.-J.; Ulrich, T.; Ohno, T. J. Phys. Chem.
1989, 93, 5147. (b) Wolff, H.-J.; Bu¨rssner, D.; Steiner, U. E. Pure Appl.
Chem. 1995, 67, 167.
(15) (a) Kellett, M. A.; Whitten, D. G. J. Am. Chem. Soc. 1989, 111,
2314. (b) Leon, J. W.; Whitten, D. G. J. Am. Chem. Soc. 1993, 115, 8038.
(16) Wang, Y.; Schanze, K. S. Inorg. Chem. 1994, 33, 1354.
(17) Wang, Y.; Schanze, K. S. J. Phys. Chem. 1995, 99, 6876.
(18) (a) Perkins, T. A.; Pourreau, D. B.; Netzel, T. L.; Schanze, K. S.
J. Phys. Chem. 1989, 93, 4511. (b) Wang, Y.; Schanze, K. S. Chem. Phys.
1993, 176, 305.
(19) Schanze, K. S.; MacQueen, D. B.; Perkins, T. A.; Cabana, L. A.
Coord. Chem. ReV. 1993, 122, 63, and references therein.
(20) Lucia, L. A.; Schanze, K. S. Inorg. Chim. Acta 1994, 225, 41.
(21) Encyclopedia of the Electrochemistry of the Elements; Bard, A. J.,
Ed.; Marcel Dekker: New York, 1973.
(22) Oevering, H.; Paddon-Row, M. N.; Heppener, M.; Oliver, A. M.;
Cotsaris, E.; Verhoeven, J. W.; Hush, N. S. J. Am. Chem. Soc. 1987, 109,
3258.
(23) An interesting point is that while 366-nm excitation of e-1a directly
populates the dπ (Re) f π* (bpy) MLCT manifold, 366-nm excitation of
e-1b must lead initially to a state having substantial π,π* IL-CT character.
Despite this difference, two features indicate that the reactivity of e-1b
emanates from a state having diamine donor f π* (bpy) LLCT character,
and not one having π,π* IL-CT character. First, the reactivity of e-1a and
e-1b are nearly identical, in both qualitative and quantitative sense. Second,
C-C bond fragmentation is nearly 8-fold more efficient in e-1b than in
the free ligand 6b.17
(24) The assumption that kBF is the same for e-1a and e-1b is supported
by a recent unpublished study we carried out which examined the
relationship between the oxidation potentials (E1/2(ox)) for a series of
aminoalcohols and the rate of C-C bond fragmentation (kBF) for the
corresponding aminoalcohol radical cations. This study indicated that there
is an excellent correlation between log kBF and E1/2(ox) for the series. Since
the oxidation potentials of the diamine ligands in e-1a and e-1b are the
same within experimental error, the aforementioned correlation suggests
that kBF will be the same for the two diamine radical cations.
(31) Lucia, L. A.; Burton, R. D.; Schanze, K. S. J. Phys. Chem. 1993,
97, 9078.
(32) Wayner, D. D. M.; Parker, V. D. Acc. Chem. Res. 1993, 26, 287.
(33) (a) Andriux, C. P.; Saveant, J. M. Bull. Soc. Chim. Fr. 1968, 4671.
(b) Andriux, C. P.; Saveant, J. M. J. Electroanal. Chem. 1970, 26, 223.
(34) McMillin, D. F.; Golden, E. M. Annu. ReV. Phys. Chem. 1982, 33,
493.
(35) Stein, S. E.; Rukkers, J. M.; Brown, R. L. NIST Standard Reference
Data No. 25, Computer Program for MS-DOS, Version 1.2, 1991.
(36) Bellville, D. J.; Pabon, R. A.; Bauld, N. L. J. Am. Chem. Soc. 1985,
107, 4978.
(37) Camaioni, D. M. J. Am. Chem. Soc. 1990, 112, 9475.
(38) Note that the thermochemical cycle (Scheme 4) indicates that direct
recombination of metal complex radical 10 and iminium 11 is endoergonic
by ca. 14 kcal/mol.
JP953537L