Photoinduced charge transfer processes along triarylamine redox cascades

Article abstract of DOI:10.1021/ja0511570

In this paper, we describe the synthesis and photophysical properties of a series of acridine-triarylamine redox cascades. These cascades were designed in order to promote photoinduced hole transfer from an acridine fluorophore into an adjacent triarylamine. The excited dipolar state then injects a hole into the triarylamine redox cascade. Subsequently, the hole migrates along the redox gradient which was tuned by the substituents attached to the triarylamine redox centers. The rate of hole migration was determined by fluorescence lifetime measurements and is in the ns regime and depends strongly on the solvent polarity. The photophysical processes were also investigated by femtosecond broadband pump-probe spectroscopy. Our studies reveal different dynamic processes in the cascades depending on the solvent polarity, e.g., direct charge separation after photoexcitation vs a two step hole transfer mechanism.

Full text of DOI:10.1021/ja0511570

Published on Web 07/07/2005
Photoinduced Charge Transfer Processes along Triarylamine
Redox Cascades
Christoph Lambert,*,† Ju¨rgen Schelter,† Torsten Fiebig,*,‡ Daniela Mank,‡ and
Anton Trifonov‡
Contribution from the Institut fu¨r Organische Chemie, Bayerische Julius-Maximilians-UniVersita¨t
Wu¨rzburg, Am Hubland, D-97074 Wu¨rzburg, Germany, and Boston College, Department of
Chemistry, Eugene F. Merkert Chemistry Center, 2609 Beacon Street, Chestnut Hill, MA 02467
Received February 23, 2005; E-mail: lambert@chemie.uni-wuerzburg.de; Fiebig@bc.edu
Abstract: In this paper, we describe the synthesis and photophysical properties of a series of acridine-
triarylamine redox cascades. These cascades were designed in order to promote photoinduced hole transfer
from an acridine fluorophore into an adjacent triarylamine. The excited dipolar state then injects a hole into
the triarylamine redox cascade. Subsequently, the hole migrates along the redox gradient which was tuned
by the substituents attached to the triarylamine redox centers. The rate of hole migration was determined
by fluorescence lifetime measurements and is in the ns regime and depends strongly on the solvent polarity.
The photophysical processes were also investigated by femtosecond broadband pump-probe spectroscopy.
Our studies reveal different dynamic processes in the cascades depending on the solvent polarity, e.g.,
direct charge separation after photoexcitation vs a two step hole transfer mechanism.
Introduction
groups have been investigated but also cascades with many
triarylamines in a row27-29 up to polymers11,30 and den-
Electron transfer (ET) or, more precisely, hole transfer (HT)
processes in triarylamine based systems have thoroughly been
investigated in the past.1-4 Owing to the relatively simple
synthetic accessibility and the stability of oxidized triarylamines
these units are widely used as hole transport components in
optoelectronic devices.5-10 But also on a molecular level,
triarylamines have attracted considerable interest: the triaryl-
amine group was used as the charge bearing unit in organic
mixed valence compounds for intramolecular ET studies1-4,11-15
as well as in organic high spin systems for organic ferro-
magnets.16-26 One-dimensional systems with two triarylamine
drimers.31-36 In the present study, we will focus on triarylamine
cascades in which a hole can be transferred along a redox
gradient. The redox centers of these cascades are built up from
triarylamines connected by acetylene spacers. The hole is
injected in the cascade by photoinduced electron transfer (PET)
of an excited acridine chromophore. Triarylamines were used
as redox centers for the above-mentioned reasons but also
(16) Blackstock, S. C.; Selby, T. D. In Magnetic Properties of Organic Materials;
Lahti, P. M., Ed.; Marcel Dekker: New York, 1999; pp 165-178.
(17) Stickley, K. R.; Blackstock, S. C. Tetrahedron Lett. 1995, 36, 1585-1588.
(18) Stickley, K. R.; Blackstock, S. C. J. Am. Chem. Soc. 1994, 116, 11576-
11577.
(19) Selby, T. D.; Blackstock, S. C. Org. Lett. 1999, 1, 2053-2055.
(20) Wienk, M. M.; Janssen, R. A. J. Am. Chem. Soc. 1997, 119, 4492-4501.
(21) Bushby, R.-J. In Magnetism: Molecules to Materials II; Miller, J. S.,
Drillon, M., Eds.; Wiley-VCH: Weinheim, Germany, 2001; pp 149-187.
(22) Selby, T. D.; Blackstock, S. C. J. Am. Chem. Soc. 1999, 121, 7152-7153.
(23) Selby, T. D.; Stickley, K. R.; Blackstock, S. C. Optics Lett. 2000, 2, 171-
174.
(24) Ito, A.; Urabe, M.; Tanaka, K. Angew. Chem. 2003, 115, 951-954.
(25) Hauck, S. I.; Lakshmi, K. V.; Hartwig, J. F. Org. Lett. 1999, 1, 2057-
2060.
† Bayerische Julius-Maximilians-Universita¨t Wu¨rzburg.
‡ Boston College.
(1) Lambert, C.; No¨ll, G.; Schelter, J. Nat. Mater. 2002, 1, 69-73.
(2) Lambert, C.; No¨ll, G. Chem. Eur. J. 2002, 8, 3467-3477.
(3) Lambert, C.; No¨ll, G. J. Am. Chem. Soc. 1999, 121, 8434-8442.
(4) Lambert, C.; Amthor, S.; Schelter, J. J. Phys. Chem. A 2004, 108, 6474-
6486.
(5) Zhao, H. T. C.; Thayumanavan, S. Tetrahedron Lett. 2001, 42, 4421-
4424.
(6) Fujikawa, H.; Tokito, S.; Taga, Y. Synth. Met. 1999, 91, 161-162.
(7) Thayumanavan, S.; Barlow, S.; Marder, S. R. Chem. Mater. 1997, 9, 3231-
3235.
(26) Michinobu, T.; Takahashi, M.; Tsuchida, E.; Nishide, H. Chem. Mater.
1999, 11, 1969-1971.
(27) Selby, T. D.; Kim, K.-Y.; Blackstock, S. C. Chem. Mater. 2002, 14, 1685-
(8) Koene, B. E.; Loy, D. E.; Thompson, M. E. Chem. Mater. 1998, 10, 2235-
1690.
2250.
(28) Kim, K.-Y.; Hassenzahl, J. D.; Selby, T. D.; Szulczewski, G. J.; Blackstock,
S. C. Chem. Mater. 2002, 14, 1691-1694.
(9) Weiss, D. S.; Cowdery, J. R.; Young, R. H. In Electron Transfer in
Chemistry; Balzani, V., Ed.; Wiley-VCH: Weinheim, 2001; Vol. 5; pp
379-471.
(29) Plater, M. J.; Jackson, T. Tetrahedron 2003, 59, 4687-4692.
(30) Ohsawa, Y.; Ishikawa, M.; Miyamoto, T.; Murofushi, Y.; Kawai, M. Synth.
Met. 1987, 18, 371-374.
(10) Thelakkat, M. Macromol. Mater. Eng. 2002, 287, 442-461.
(11) Lambert, C.; No¨ll, G. Synth. Met. 2003, 139, 57-62.
(12) Lambert, C.; No¨ll, G.; Hampel, F. J. Phys. Chem. A 2001, 105, 7751-
7758.
(31) Louie, J.; Hartwig, J. F.; Fry, A. J. J. Am. Chem. Soc. 1997, 119, 11695-
11696.
(32) Katsuma, K.; Shirota, Y. AdV. Mater. 1998, 10, 223-226.
(33) Wu, I.-Y.; Lin, J. T.; Tao, Y.-T.; Balasubramaniam, E. AdV. Mater. 2000,
12, 668-669.
(13) Lambert, C.; No¨ll, G.; Kriegisch, V.; Zabel, M.; Hampel, F.; Schma¨lzlin,
E.; Bra¨uchle, C.; Meerholz, K. Chem. Eur. J. 2003, 9, 4232-4239.
(14) Bonvoisin, J.; Launay, J.-P.; Verbouwe, W.; Van der Auweraer, M.; De
Schryver, F. C. J. Phys. Chem. 1996, 100, 17079-17082.
(15) Bonvoisin, J.; Launay, J.-P.; Van der Auweraer, M.; De Schryver, F. C. J.
Phys. Chem. 1994, 98, 5052-5057, see also correction 1996, 100, 18006.
(34) Chen, C. H.; Shi, J.; Tang, C. W. Macromol. Symp. 1997, 125, 1-48.
(35) Selby, T. D.; Blackstock, S. C. J. Am. Chem. Soc. 1998, 120, 12155-
12156.
(36) Krys, B.; S, T. J. Org. Chem. 2003, 68, 5559-67.
9
10600
J. AM. CHEM. SOC. 2005, 127, 10600-10610
10.1021/ja0511570 CCC: $30.25 © 2005 American Chemical Society

Photoinduced Charge Transfer Processes
A R T I C L E S
Chart 1
because triarylamines possess relatively low internal reorganiza-
tion energy.37,38 Therefore, HT processes are expected to be
rather quick which will favor photoinduced charge separation
processes over other competing processes. On the other hand,
if the back electron transfer is in the Marcus inverted region, a
small reorganization energy will slow the back electron transfer
and will lead to long-lived charge separated states. Our
investigations illustrate a first step how to construct an artificial
light-driven system based on triarylamines that might be able
to induce the long-range separation of opposite charges upon
irradiation. So far, most systems that aim to mimic basic aspects
of natural photosynthetic reaction centers use porphyrine
chromophores in close analogy to their natural examples.39 Only
a few systems are known that incorporate triarylamine units.40-48
(39) Gust, D.; Moore, T. A.; Moore, A. L. In Electron Transfer in Chemistry;
Balzani, V., Ed.; Wiley-VCH: Weinheim (Germany), 2001; Vol. 3; p 272.
(40) Ramos, A. M.; Meskers, S. C. J.; van Hal, P. A.; Knol, J.; Hummelen, J.
C.; Janssen, R. A. J. J. Phys. Chem. A 2003, 107, 9269-9283.
(41) Ramos, A. M.; Beckers, E. H. A.; Offermans, T.; Meskers, S. C. J.; Janssen,
R. A. J. J. Phys. Chem. A 2004, 108, 8201-8211.
(37) Eberson, L. Electron-Transfer Reactions in Organic Chemistry; Springer:
(42) Ohno, T.; Moriwaki, K.; Miyata, T. J. Org. Chem. 2001, 66, 3397-3401.
(43) Lor, M. V.; Lucien; Pilot, R.; Fron, E.; Jordens, S.; Schweitzer, G.; Weil,
T.; Mu¨llen, K.; Verhoeven, J. W.; Van der Auweraer, M. V.; De Schryver,
F. C. J. Phys. Chem. B 2004, 108, 10721-10731.
Berlin (Germany), 1987.
(38) Lin, B. C.; Cheng, C. P.; Lao, Z. P. M. J. Phys. Chem. A 2003, 107, 5241-
5251.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10601

A R T I C L E S
Lambert et al.
In this study, we refrain from employing subunits that resemble
the well-known natural systems and turn to completely artificial
chromophores and redox centers which are more likely to be
incorporated in future optoelectronic devices that shall build
up an electrical potential upon irradiation.
The cascade structures 1-5 which we have designed are
drawn in Chart 1 together with some reference systems. The
principal features of the cascade systems are the fluorescent
donor substituted acridine chromophore and the triarylamine
cascade. The triarylamine units are connected by triple bonds
in order to ensure an essentially rigid structure of the whole
cascade with fixed HT distances. To illustrate our conceptual
approach we divide the chromophores into the following
subunits: acridine acceptor (A), the first triarylamine attached
to the acridine (Tara1), the second, and third triarylamine
(Tara2 and Tara3) which are farther apart; two triarylamine
groups connected by a triple bond yield a tolandiamine (e.g.,
Tara1-≡-Tara2). While the HT distances are fixed in 1-3 and
5 by the almost rigid structure it is not in 4 in which the
N(acridine)...N(Tara3) distance can vary between ca. 23 and 29
Å, depending on the conformer.
The donor substituted acridine fluorophore (A-Tara1) is a
strong oxidizing agent in the first excited charge transfer (CT)
singlet state. This locally excited singlet state is expected to
oxidize the adjacent triarylamine moiety (Tara2) which is
attached to the donor substituted acridine by an acetylene spacer.
We use donor substituted acridine dyes because their photo-
physics are reasonably well understood thanks to the work of
Herbich and Kapturkiewicz49 (hereafter called HK). These
authors investigated a series of dialkylamino substituted phenyl-
acridines by fluorescence spectroscopy. The local redox potential
of the triarylamine units is tuned by substituents (Cl, Me, MeO)
in para position of the phenyl rings so as to form two short (1,
2) and one long (4) cascade with downhill hole transport
gradient, that is, the triarylamine moiety which is farthest apart
from the acridine has the lowest redox potential, i.e., it is more
easily oxidized than the triarylamines adjacent to the acridine.
In a recent study, we showed that the redox potentials of
triarylamine in tolandiamines may span a range of 400 mV
depending on the substituents.50 For comparison, we also
synthesized a branched dendrimeric system (5) as well as a short
cascade with uphill redox gradient (3). The photophysical
properties of these cascades will be compared with A-Tara1
subunits (6-8) as well as with tolandiamine (9) and phenyl-
acridine (10).
Figure 1. Absorption spectra of 6-8 in CH2Cl2.
Table 1. Absorption Maxima of 6-8
1
1
1
ν
ꢀ
˜
abs/cm-
ν
ꢀ
˜
abs/cm-
ν˜
abs/cm-
ꢀ
1
1
1
(
/M-1cm-
)
(
/M-1cm-
)
(
/M-1cm-
)
6
7
8
C6H12
Bu2O
26200 (10400)
26100 (10300)
26200 (10600)
25900 (9980)
26100 (10700)
26000 (10200)
26000 (9950)
25800 (9780)
25900 (8980)
25700 (8560)
25100 (9050)
25100 (8490)
25300 (8640)
24900 (8030)
25300 (8210)
25200 (8100)
24900 (7810)
24600 (7160)
24700 (7710)
24400 (6550)
24700 (6700)
24600 (6830)
24600 (7270)
24400 (7410)
24700 (7580)
24600 (7630)
24300 (7070)
24100 (6940)
24200 (6880)
MTBE
1,4-dioxane
Et2O
EtOAc
THF
CH2Cl2
DMF
DMSO
The sharp peaks at ca. 26 000, 28 000, and 29 000 cm-1 stem
from localized excitations of phenylacridine (10) and the broad
and very intense peak at ca. 33 000 cm-1 is due to a localized
triarylamine excitation. The CT bands are somewhat stronger
red shifted with MeO substituents than with Me or Cl, that is,
the donor strength of the triarylamine unit can be tuned by the
para substituents attached to the phenyl rings.
The CT bands of 6-8 are moderately positive solvatochromic
(e.g., 6: 26 200 cm-1 in C6H12 and 25 700 cm-1 in DMSO)
which indicates an increase in dipole moment upon excitation.
Fluorescence spectra of all compounds were measured at the
excitation energy of the CT band maximum. Much in contrast
to 9-phenylacridine 10 which is practically nonfluorescent, the
donor substituted compounds 6-8 show strong fluorescence
with high quantum yields (see Table 2). The fluorescence spectra
of 6-8 are strongly positive solvatochromic and display a large
Stokes shift (see e.g., Figure 2 for the fluorescence spectra of
6, the full data set for all compounds is given in Table 2) which
is stronger for 8 and 7 than for 6. This Stokes shift indicates a
major reorganization in the excited state. HK49 explain this
behavior by a solvent dependent mixing of a locally excited
state and a CT state with a different degree of planarization in
the excited state while the ground state is generally nonplanar.
Here, we simply state that AM1 computations yield two ground-
state minima for 8 that are very close in energy but differ in
their relative orientation of the dianisylamino group to the
acridine moiety, coplanar vs perpendicular, the latter being in
accordance with the X-ray crystal structure investigation (see
Supporting Information). If cyclohexane is excluded, then a plot
of the fluorescence energy vs the solvent parameters according
to Lippert and Mataga (eq 1)51 yields a reasonable linear
correlation (Figure 3 and Table 3). Cyclohexane is excluded as
Results and Discussion
A. Stationary Optical Properties. The absorption spectra
of the A-Tara1 species 6-8 show a moderately strong CT band
around 25 000 cm-1 which can be ascribed to a charge transfer
from the triarylamine donor to the acridine acceptor (see Figure
1 and Table 1).49
(44) Sandanayaka, A. S. D.; Sasabe, H.; Araki, Y.; Furusho, Y.; Ito, O.; Takata,
T. J. Phys. Chem. A 2001, 108, 5145-5155.
(45) Luo, H.; Fujitsuka, M.; Araki, Y.; Ito, O.; Padmawar, P.; Chiang, L. Y. J.
Phys. Chem. B 2003, 107, 9312-9318.
(46) Komamine, S.; Fujitsuka, M.; Ito, O. M.; Miyata, K.; Toshiyuki; Ohno, T.
J. Phys. Chem. A 2000, 104, 11497-11504.
(47) Lor, M.; Thielemans, J.; Viaene, L.; Cotlet, M.; Hofkens, J.; Weil, T.;
Hampel, C.; Mu¨llen, K.; Verhoeven, J. W.; Van der Auweraer, M.; De
Schryver, F. C. J. Am. Chem. Soc. 2002, 124, 9918-9925.
(48) Sandanayaka, A. S. D.; Matsukawa, K.; Ishii, T.; Mataka, S.; Araki, Y.;
Ito, O. J. Phys. Chem. B 2004, 108, 19995-20004.
(49) Herbich, J.; Kapturkiewicz, A. J. Am. Chem. Soc. 1998, 120, 1014-1029.
(50) Lambert, C.; No¨ll, G. J. Chem. Soc., Perkin Trans. 2 2002, 2039-2043.
9
10602 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Photoinduced Charge Transfer Processes
A R T I C L E S
Table 2. Steady State and Time Resolved Optical Properties of 1-8
1
1
1
1
1
1
ν
˜f/cm-
Φf
τf/ns
kf/108 s-
k
nr/108 s-
ν
˜f/cm-
Φf
τf/ns
kf/108 s-
k
nr/108 s-
1
3
5
7
2
4
6
8
C6H12
Bu2O
MTBE
1,4-dioxane
Et2O
EtOAc
THF
CH2Cl2
PrCN
22400
20500
19400
19500
18100
17800
17300
15700
0.23
0.38
0.37
0.49
0.10
0.08
0.06
0.01
1.5
1.5
5.1
21900
20000
19100
18900
17800
17500
17300
15500
0.42
0.53
0.49
0.67
0.22
0.27
0.22
0.02
1.7
2.5
3.4
5.9
5.4
0.63
0.91
1.1
0.94
4.8
6.4
1.0
1.0
1.0
0.52
4.7
5.0
0.8
0.16
0.12
0.13
2.0
1.9
12
5.9
0.7
0.37
0.28
1.3
14
DMF
DMSO
C6H12
Bu2O
MTBE
1,4-dioxane
Et2O
EtOAc
THF
CH2Cl2
PrCN
22000
20000
19000
18900
17800
17300
17200
15700
0.42
0.51
0.67
0.58
0.50
0.36
0.46
0.18
1.8
2.4
3.2
22600
21000
20000
19800
18900
18400
18300
16000
0.18
0.30
0.32
0.42
0.20
0.22
0.13
0.02
1.0
1.8
8.2
4.9
6.3
6.9
1.4
0.92
0.73
0.68
0.67
0.72
4.3
4.7
0.75
0.90
1.6
1.2
8.0
9.2
4.8
0.27
0.14
0.032
0.98
0.95
2.1
7.2
7.4
0.63
0.25
0.76
1.1
DMF
DMSO
14800
0.01
C6H12
Bu2O
MTBE
1,4-dioxane
Et2O
EtOAc
THF
CH2Cl2
PrCN
21800
20300
19100
19000
17900
17700
17400
15300
0.40
0.74
0.31
0.47
0.23
0.18
0.15
0.01
1.9
3.4
4.5
6.0
2.1
2.2
0.69
0.78
3.2
0.77
1.5
23000
21800
20900
20600
20100
19100
19200
18200
17100
16500
15900
0.14
0.26
0.38
0.52
0.47
0.86
0.74
0.90
0.74
0.72
0.65
0.55
2.6
16
1.9
3.2
2.0
1.6
3.3
1.5
0.88
6.3
6.9
1.0
0.29
0.22
0.10
1.3
1.2
9.9
6.0
5.7
9.3
1.2
1.5
0.97
0.43
0.24
0.10
DMF
DMSO
11.4
11.6
0.63
0.56
0.25
0.30
C6H12
Bu2O
MTBE
1,4-dioxane
Et2O
EtOAc
THF
CH2Cl2
PrCN
22100
20400
19400
19000
18300
17700
17700
16100
15600
15000
14800
0.28
0.41
0.56
0.72
0.63
0.67
0.76
0.58
0.28
0.10
0.07
1.3
4.9
2.2
1.1
5.5
21300
18900
17800
17600
16600
16100
16000
14800
0.34
0.56
0.66
0.71
0.34
0.20
0.29
0.05
2.0
6.7
1.7
3.3
0.90
0.99
0.51
8.9
12.4
0.86
0.46
0.27
0.34
6.2
2.4
0.47
0.21
1.1
4.0
DMF
DMSO
2.8
0.23
3.3
solvent. In eq 1 ν˜f is the fluorescence energy maximum, ν˜vac is
the fluorescence energy maximum in vacuum, µg and µef are
the ground state and excited-state dipole moments, respectively,
which are assumed to be parallel, a0 is the effective radius of
the solute, n is the refractive index and ꢀ is the permittivity of
the solvent. The slopes of the Lippert-Mataga correlations for
6-8 are quite similar (see Figure 3 and Table 3). An effective
solute radius of 7 Å (estimated from an AM1 computation)
yields a dipole moment of 31 D for the excited state. The ground
state has a dipole moment of 4-5 D depending on the
conformation (see above) estimated by AM1 computations.
Thus, the dipole moment difference of ca. 26 D roughly
corresponds to the transfer of a unit charge from the triarylamine
nitrogen to the acridine center. Therefore, we can best describe
the CT state by (A)--(Tara1)+. In general, the stationary
Figure 2. Normalized fluorescence spectra (solid lines) of 6 from left to
right in DMSO, PrCN, THF, MTBE, C6H12.
it gives values which are significantly too low; this effect might
have to do with changes in solute geometry in this very apolar
(51) Suppan, P.; Ghoneim, N. SolVatochromism; The Royal Society of Chem-
istry: Cambridge, 1997.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10603

A R T I C L E S
Lambert et al.
Figure 3. Lippert-Mataga plot for 6 (circles), 7 (squares), and 8 (triangles).
The solvents are from left to right: C6H12, Bu2O, Et2O, MTBE, EtOAc,
THF, CH2Cl2, DMSO, PrCN, DMF. Cyclohexane is excluded from the
correlation.
Figure 5. Absorption spectra of 4 and 5 in CH2Cl2.
Table 3. Fitted Parameters from Eq 1
(2bµe
4
(bµe − bµg))/
3
νvf ac
˜
/cm-
πꢀ0hca0/cm-
1
1
1
2
3
4
5
6
7
8
27120
26120
25840
27620
26730
28200
26380
24820
-33330
-30470
-29570
-32710
-32520
-31020
-30060
-30050
3
Figure 6. Reduced (I(ν˜f)/ν˜f vs ν˜f) and normalized fluorescence spectra
(solid lines) of 3 from left to right in THF, Et2O, Bu2O, C6H12, and
simulation by eq 2 (symbols).
lution. From Figure 4, it is obvious that the substitution pattern
has only a marginal influence on the absorption spectra with
exception of the low-energy tail which is more intense in 3 due
to the MeO substituent at Tara1 which makes it a stronger donor
compared to Tara1 in 1 and 2. The absorption spectra of 4 and
5 are quite similar because both species have the same number
of Tara groups and tolandiamine units. This observation
supports the conceptual approach of dividing the cascades into
the above-mentioned moderately interacting units.
The fluorescence spectra of 1-5 (see Figure 6) are strongly
solvatochromic and are very similar in energy and band shape
to those of 6-8, irrespective of the excitation energy. However,
while there are significant differences in the Stokes shift of 6-8
due to strong substituent effects, the fluorescence energy maxima
for 1-5 are very similar in a given solvent. The slopes of the
Lippert-Mataga plots of 1-5 are also in good agreement with
those of 6-8. Excitation spectra of all compounds are identical
to the absorption spectra and prove complete energy transfer to
the S1 state, irrespective of excitation energy. In addition, no
fluorescence from other states, e.g., local tolandiamine states
was observed which should occur between 25 000 and 20 000
cm-1. For comparison, the fluorescence of tolandiamine 9 occurs
between 24 700 cm-1 (cyclohexane) and 20 000 cm-1 (DMSO)
with quantum yields of 0.60 and 0.70. Altogether, these features
prove that the fluorescent state in the cascades has the same
electronic nature as that in the A-Tara1 species 6-8, i.e., all
cascades emit from a more or less dipolar CT state localized in
the A-Tara1 moiety.
Figure 4. Absorption spectra of 1-3 in CH2Cl2.
absorption and fluorescence behavior of 6-8 is analogous to
the dialkylamino substituted species described by HK.49
2µe(µe - µg)
1
2
ν˜f ) ν˜vf ac
-
f(ꢀ) - f(n2) with
(
)
3
4πꢀ0hca0
ꢀ - 1
2ꢀ + 1
n2 - 1
2n2 + 1
f(ꢀ) )
and f(n2) )
(1)
The absorption spectra of the cascades 1-5 (see Figures 4
and 5) are much less structured than those of 6-8 due to severe
band overlap. The main feature that is added is a very intense
band at ca. 27 000 cm-1 which is due to the tolandiamine
moiety. This band overlaps strongly with the CT band at ca.
25 000 cm-1 which is only seen as a low-energy tail of the
tolandiamine band which prevents a reasonable band deconvo-
9
10604 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Photoinduced Charge Transfer Processes
A R T I C L E S
Table 4. Reorganization Energies and Free Energy Difference of
3 as Derived from Eq 2
Table 5. Redox Potentials of 1-8 in CH2Cl2/0.2 M TBAH vs
Fc/Fc+, v ) 250 mV s-1a
Er1ed2/mV
Eo1x2,1/mV
Eo1x2,2/mV
Eo1x2,3/mV
Eo1x2,4/mV
1
1
1
1
λ
v/cm-
λ
s/cm-
∆G
00/cm-
ν˜v/cm-
/
/
/
/
/
C6H12
Bu2O
1100
1240
1210
1100
1140
1170
1090
1440
1240
2080
2620
2600
3670
3740
3650
3400
-23250
-22510
-21980
-21740
-21790
-21340
-21080
-19520
1450
1400
1400
1400
1400
1400
1400
1400
1
2
3
4
5
6
7
8
-2160b
-2130b
-2150b
-2160b
-2160b
-1730b
-2080b
-2140b
260 (Tara2)
250 (Tara2)
400 (Tara1)
240 (Tara3)
270c (Tara2)
640 (Tara1)
1120b
1100b
530 (Tara1)
480 (Tara2)
MTBE
1,4-dioxane
Et2O
EtOAc
THF
620 (Tara2)
680 (Tara1)
630 (Tara1)
690 (Tara1)
1030b
1050b
1320b
460 (Tara1)
a
CH2Cl2
290 (Tara1)
1020b
a Bad signal-to-noise ratio.
a The assignments in parentheses refer to the triarylamine group that is
oxidized at the given potential. b Irreversible process, peak potential. c Two
electron transfer, not resolved.
To extract the reorganization energy λs associated with any
low-frequency vibrations (such as solvent and low energy solute
motions) and the reorganization energy λv associated with any
high energy vibrations (bond length and angle distortions) as
well as the free energy difference ∆G00 between the relaxed
excited CT and the ground-state we applied the vibrational
coupling theory52-54 based on a Golden rule expression where
µfl is the fluorescence transition moment and S is the Huang-
Rhys factor λv/ν˜V.The fluorescence spectra were fitted by eq 2,
where ν˜v is an averaged high-frequency mode associated with
λv.
3 n(n2 + 2)2
-S
j
16 × 106π
3ꢀ0
e S
1
∞
Ifl/ν˜3 )
µfl
2
∑
9
j!
4πhcλ kT
x
j)0
s
hc(jν˜v + λs + ν˜ + ∆G00)2
4λskT
exp -
(2)
[
]
Figure 7. Rate constant of the radiative process kf vs fluorescence transition
energy ν˜f for 1-8 in different solvents.
The fitted reorganizational parameters are collected in Table 4
for 3 and the fits are displayed in Figure 6. An averaged
vibration ν˜v ) ca. 1400 cm-1 was used for the fits which roughly
correlates to a C-N vibration. Although the partitioning of the
total reorganization energy into λs and λv is somewhat arbitrary,
the free energy ∆G00 is rather accurate and decreases with
increasing solvent polarity. However, the sum of both re-
organization energies compares well with those of tetraaryl-
The redox potentials for 1, 2, 4, and 5 show that the energy
cascade for an excited state with predominantly localized
triarylamine radical cation character is downhill for the migration
of a hole from Tara1 to Tara2 and Tara3. Much in contrast it
is uphill for 3 where the triarylamine substituents (Cl, MeO)
are exchanged in comparison with 1.
C. Time-Resolved Fluorescence Properties. Time-resolved
fluorescence decays τf for 1-8 were measured in the nanosecond
regime in selected solvents that cover the range from totally
apolar (cyclohexane) to highly polar (DMSO). In all cases a
single-exponential decay was observed. In addition, we mea-
sured the fluorescence quantum yield Φf in all solvents. The
quantum yields vary depending on the solvent from being low
in apolar solvents and in rather polar solvents with a maximum
of quantum yield in between up to 0.90 (see Table 2). From
both quantities, τf and Φf we calculated the rate constant for
the radiative (kf) and nonradiative decay (knr) according to eqs
3 and 4.
tolandiamine radical cations in CH2Cl2 which is ca. 6000 cm-1 50
.
Similar observations are true for all other derivatives 1, 2, and
4-8.
B. Redox Properties. To gain insight into the energy of
localized states with radical cation or radical anion character
we measured the redox potentials of 1-8 by cyclic voltammetry
(CV) in CH2Cl2/0.2 M TBAH solution vs Fc/Fc+. The corre-
sponding redox potentials are collected in Table 5 where they
are grouped according to their redox potential range. An
interpretation as to which triarylamine moiety the given
potentials refer is also given in parentheses. The relatively
narrow range in which the redox potential for a given type of
triarylamine groups occurs shows that the coupling between the
triarylamine groups is relatively small compared to the influence
of the substituents (Cl, Me, MeO). The redox processes at ca.
-2200 mV refer to chemically irreversible reductions of the
acridine acceptor, those above 1000 mV refer to the second
oxidation of the triarylamine that is most easily oxidized.
kf
Φf )
τf )
(3)
(4)
kf + knr
1
kf + knr
Both a plot of knr and of kf for compounds 1-8 in different
solvents vs the fluorescence energy shows a relatively narrow
scattering range of data points (see Figures 7 and 8), Therefore,
it is obvious that the photophysics of 1-8 in the nanosecond
(52) Jortner, J.; Bixon, M. J. Chem. Phys. 1988, 88, 167-170.
(53) Gould, I. R.; Noukakis, D.; Gomez-Jahn, L.; Young, R. H.; Goodman, J.
L.; Farid, S. Chem. Phys. 1993, 176, 439-456.
(54) Marcus, R. A. J. Phys. Chem. 1989, 93, 3078-3086.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10605

A R T I C L E S
Lambert et al.
Table 6. Hole Transfer Rate from Tara1 to Tara2 in 1-5 from
Fluorescence Lifetime Measurementsa
1
kHT
)
knr(x)
−
knr(y)/108 s-
x/y
1/6
2/7
3/8
4/6
5/7
C6H12
(-)
(-)
(-)
1.6
1.7
12
(-)
(-)
(-)
(-)
(-)
0.55
0.71
2.0
(-)
MTBE
1,4-dioxane
EtOAc
THF
0.1
0.17
0.6
1.0
14
(-)
(-)
0.93
9.6
CH2Cl2
a (-) indicates no hole transfer (kHT is negative).
quenching effect we employ the Rehm-Weller equation.
Equation 5 can be used to estimate the free energy for the PET-
hole migration process ∆G° by subtracting the redox potential
difference ∆E of donor (Tara2) and acceptor (A) from the free
energy difference of the ground and excited CT state potential
energy surface minima ∆G00.56 The redox potential difference
can be obtained from the values in Table 5 which refer to
CH2Cl2/0.2 TBAH solution: ∆E ) 0.260 V (Tara2) - (-2.160
V (A)) ) 2.420 V for 1 and ∆E ) 0.620 V (Tara2) - (-2.150
V (A)) ) 2.770 V for 3. The free energy ∆G00 in CH2Cl2 is
taken from the fits of eq 2 (Table 4) which can be estimated to
be ca. 20 000 cm-1 for 1 (the ∆G00 value for CH2Cl2 could not
be determined for 1 because of the bad signal-to-noise ratio of
the fluorescence spectrum in CH2Cl2) and 19 520 cm-1 for 3.
Figure 8. Rate constant of the nonradiative process knr vs fluorescence
transition energy ν˜f for 1-8 in different solvents.
time domain are energy controlled, that is, the energy of the
fluorescent A-Tara1 located CT state dominates its deactivation
kinetics. The fluorescence energy can either be tuned by
substituents (Cl, Me, MeO, Tara2, etc.) or by the solvent (apolar
> polar).
As one can see in Figure 8 knr is practically constant between
16 000 and 20 000 cm-1 while above 20 000 cm-1 knr rises
strongly. This increase with increasing fluorescence energy was
also found by HK in carbazole derivatives55 and was interpreted
by a dominating intersystem crossing (ISC) from the fluorescent
CT state to a lower lying triplet state. However, as we will see
below we did not find any evidence in the transient absorption
(TA) spectra for the population of a long-lived triplet state. The
reason of the knr increase for ν˜f > 20 000 cm-1 thus remains
cumbersome. For the energies below 16 000 cm-1 the set of
derivatives 1-8 splits into two groups: those derivatives whose
nonradiative rate constant remains essentially constant or
increases slightly (6-8 and 3) and those of which knr dramati-
cally increases by roughly one order of magnitude (1,2,5).
Obviously, the above-mentioned energy control of deactivation
kinetics fails for the latter group of compounds. For compound
4 there is no data point low enough in energy in order to decide
to which group it belongs, however, we will present evidence
below that 4 also is likely to belong to the latter group.
∆G° ) 8065.5[E° (Tara2) - E° (A)] - ∆G00
(5)
Eq 5 yields ∆G° ) -500 cm-1 for 1 but +2820 cm-1 for 3.
Although the Rehm-Weller equation only allows a very rough
estimate of the free energy of a PET the exergonic value for 1
suggests that a hole migration from the A-Tara1 CT state to
Tara2 is energetically favorable while the endergonic value for
3 explains why we cannot observe hole migration, and,
consequently no additional fluorescence quenching pathway. The
same holds true for 2 for which ∆G0 ) ca. -200 cm-1. Because
the cascade 4 and the dendrimer 5 possess very similar
fluorescence energy properties and redox potential values we
expect hole migration also to be possible in polar solvents. In
fact, for 5 the rather high nonradiative rate constant suggests a
similar process. For the cascade 4, we were unable to measure
the rate constant in a solvent more polar than CH2Cl2 for which
the hole transfer is distinctly favorable. In any case, the free
energy values derived by the Rehm-Weller equation for 1 and
2 are rather small so that small changes in the medium polarity
can switch the system in a way that hole migration to Tara1 or
Tara2 is favorable or unfavorable while for 3 it is an
unfavorable uphill process independent of the solvent. From
the fluorescence lifetime data (see Table 2) we can estimate
the hole transfer rate kHT from Tara1 to Tara2 in 1-5 by the
difference between knr of these compounds and the reference
compounds 6-8 as given in Table 6. The reference compounds
were chosen as to match the A-Tara1 moiety in 1-5 as close
as possible. A negative sign of kHT indicates that no hole
migration is possible either because it is against the redox
gradient as in 3 or it is because less polar solvents disfavor
charge separation.
The increase in knr for 1, 2, 4, and 5 is due to an additional
deactivation pathway. Though superficially very similar, the
contrasting knr behavior of 3 vs 1 and 2 shows that the lack of
an additional deactivation pathway in 3 is due to the fact that
3 displays an uphill gradient for hole migration while 1 and 2
possess a downhill gradient. Therefore, we assume that the
additional nonradiative deactivation pathway is a hole transfer
(HT) from Tara1 to Tara2 in 1 and 2. This hole migration is
only downhill in rather polar solvents, because, in contrast to
the oxidation processes observed in CV, hole migration from
an excited A-Tara1 CT state involves a charge separation which
is unfavorable in apolar solvents. Much in contrast, the hole
migration is uphill for 3 in all solvents and, thus, no additional
nonradiative deactivation is observed. To assess whether a
photoexcitation process followed by a hole migration can be
responsible for the above-mentioned additional fluorescence
(55) Kapturkiewicz, A.; Herbich, J.; Karpiuk, J.; Nowacki, J. J. Phys. Chem. A
1997, 101, 2332-2344.
(56) Kavarnos, G. J. Fundamentals of Photoinduced Electron Transfer; VCH:
Weilheim (Germany), 1993.
9
10606 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Photoinduced Charge Transfer Processes
A R T I C L E S
100 fs, however, the evolution becomes strongly solvent
dependent. In acetonitrile (Figure 9a) the initial two peak
distribution becomes more pronounced while the band at 690
nm undergoes a shift to 725 nm. Reference measurements on
dianisylphenylamine have shown that the lowest excited singlet
state of these compounds absorb at ∼700 nm. At the same time
dianisylarylamine radical cations absorb at around 730 nm.3
Hence, upon charge transfer in 8 one would expect to observe
a dynamical shift in this spectral region. Acridine radical anions
are known to have an absorption band at ∼610 nm.57 Thus, the
temporal evolution of 8 in acetonitrile reflects an ultrafast
photoinduced (as opposed to optical) charge transfer from
triarylamine to the acridine unit in less than 2 ps.
In benzonitrile (Figure 9b) the initial spectral bifurcation is
merging into a pronounced band at 700 nm which reaches its
maximum intensity after 500 fs. The onset of charge transfer is
marked by a spectral shift of the amine radical cation band
(accompanied by a drop in intensity) as well as by the rise of
the band of the acridine radical anion (610 nm). In contrast to
acetonitrile, where the charge-transfer process leads to a
continuous spectral shift the results in benzonitrile reveal
characteristic “two-level” kinetics with a well pronounced
isosbestic point at 660 nm.
Finally, in MTBE we observe the rise of the triarylamine
excited-state absorption band at 710 nm. However, the rise is
not followed by a spectral shift toward 730 nm (triarylamine
radical cation) and the acridine radical anion band around 610
remains absent as well. Consequently, there is no evidence for
the population of a strongly dipolar CT state in MTBE.
From the results shown in Figure 9 the following mechanistic
picture emerges. Excitation of 8 leads to a largely delocalized
excited state with similar electronic (and geometrical) structure
in acetonitrile, benzontrile and MTBE. Immediately after
excitation, coupling to various nuclear modes (both internal and
solvent modes) leads to energy localization on the time scale
of 100 fs to 1 ps. The spectroscopic signatures found in the
transient absorption spectra suggest thatsin all three solventss
a locally excited (LE) triarylamine state is formed shortly after
excitation. Reference measurements on dianisylphenylamine
have shown that the lowest excited singlet state of these
compounds absorb at ∼700 nm. In the polar acetonitrile, ultrafast
energy localization competes with charge transfer which occurs
on the same time scale. In benzonitrile, the thermodynamic
driving force for CT is much smaller which leads to a reduced
CT rate. Hence, energy localization and CT are clearly separable
processes. Finally, in MTBE the thermodynamics for CT are
highly unfavorable and thus the triarylamine excited-state cannot
undergo efficient CT.
Figure 9. Temporal evolution of the pump-probe spectra of 8 in different
solvents; (a) acetontrile, (b) benzonitrile, (c) MTBE, within the first 20 ps
after excitation (step size: 60 fs). Early spectra are shown in blue/green
and late spectra are shown in orange/red colors.
D. Femtosecond Pump-Probe Spectra. Excited-State
Dynamics of 8. To elucidate the photophysical mechanisms in
the model redox cascades it is crucial to investigate first the
dynamics after photoexcitation in the acridine-triarylamine D-A
systems without the tolandiamine “antenna”. Figure 9 shows
the temporal evolution of the pump-probe-spectra of 8 mea-
sured in three different solvents after excitation at 360 nm
(27 800 cm-1). Immediately after excitation there is a broad
transient absorption band showing a weak bifurcation with
weakly pronounced maxima centered around 610 and 690 nm.
Note that this initial spectral distributionswhich rises within
the pulse duration of our experimental setup (∼80 fs)sis almost
solvent independent (Figure 9a-c, blue spectra). After several
CT and Effect of Driving Force in Tolandiamine-Bridged
Systems. Photoexcitation of 1-5 at the pump wavelength (360
nm ) 27 800 cm-1) initially populates a tolandiamine state (see
Figure 10) because this unit has the highest absorptivity at this
energy (see Figures 4 and 5). Hence, the tolandiamine units
serve as antenna system. These tolandiamine states are expected
to be the precursor states for ultrafast intramolecular energy
transfer to the corresponding charge transfer states. Immediately
after photoexcitation of 1 and 3 we observe a broad, structureless
absorption between 500 and 700 nm (see Figure 10a). This band
(57) Shida, T. Electronic Absorption Spectra of Radical Ions; Elsevier: Am-
sterdam, 1988.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10607

A R T I C L E S
Lambert et al.
Figure 10. Temporal evolution of the pump-probe spectra of 1 (a, b) and 3 (c,d) in acetonitrile, the upper panels (a) and (c) show early spectra from -1.0
to 3 ps after excitation (60 fs step size). The lower panels (b) and (d) represent the spectra 3 to 130 ps after excitation (various step sizes). Early spectra in
each panel are shown in blue/green and late spectra are shown in orange/red colors. The insets show the time-dependent absorption at 740 and 700 nm,
respectively (see text for details).
represents the absorption of the unrelaxed Franck Condon state
populations. After several 100 fs, there are three main transient
absorption bands, the 610 nm (acridine radical anion), the 700-
740 nm band (triarylamine radical cation), and a very intense
band at 490 nm which must be assigned to excited-state
absorption of the tolandiamine system. A similar transient
absorption band at 480 nm is also visible in the spectra of
tolandiamine 9 in MeCN. The fact that the tolandiamine band
at 490 nm is present throughout the lifetime of the excited states
of 1-5 must be viewed as a clear indication that the transient
absorption spectra cannot be reconstructed by (diabatic) mixing
of spectral reference components obtained from separate
measurements. Hence, the electronic coupling between the
chromophores is significant.
A comparison of the 740 and 700 nm transients of 1 and 3
(insets of Figure 10) reveals a significantly longer lifetime of
the CT state of 3. However, in addition to these differences the
intensity ratio reveals also interesting information about the
excited-state relaxation in 1 and 3. Most obvious is the difference
in the relevant contribution of the 610 nm band which is
significantly stronger in 3 than in 1, suggesting a more localized,
twisted acridine radical anion in 3 than in 1. In fact, this
interpretation is in agreement with the rest of our data: In 3
the charge is only being transferred between the acridine and
the adjacent Tara1 amine unit. To establish a large dipole
moment the acridine has to twist to electronically decouple from
the rest of the aromatic system. This will result in a fairly pure
acridine radical anion band. The CT state dipole moment of 1
is somewhat larger than that of 3 (see the larger slope of 1 vs
3 in the Lippert-Mataga plot, Table 3), however one has to
bear in mind that the positive charge is delocalized much farther
over the second amine unit. Hence, the negative charge on the
acridine will also be delocalized. As a result, the radical anion
character is “diluted” which is manifested in the reduced 610
nm band.
Figure 10 also displays the time-dependent transient absorp-
tion signal of the 740 nm band. In contrast to 8 (acetonitrile)
where this band rises within the time resolution of our
experimental setup the rise of this band in 1 is strongly
biexponential, i.e., it shows a rise time of 4 ps (40%). This rise
time must be interpreted as the hole transfer time from the
triarylamine (Tara1) which is adjacent to the acridine to the
terminal amine (Tara2).
Effect of the Redox Cascade Extension. Figure 11 shows
the temporal evolution of the pump-probe spectra of 4 and 5
in acetonitrile. The insets display the time-dependent transient
absorption at 740 nm (triarylamine radical cation band). From
a comparison of the symmetrical “doubly branched” system 5
with its single-branch analogue 1 it becomes clear that the
spectral properties are very similar (both in time evolution and
Because of the substitution scheme in 3 hole transfer between
the central and the terminal amine unit is thermodynamically
unfavorable. Hence, a long-lived charge transfer state of the
type A--(Tara1)+-Tara2 is formed. Entirely consistent with
this proposal is the fact, that the rise of the 700 nm band in 3
is not showing a ps component (see Figure 10 inset).
9
10608 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Photoinduced Charge Transfer Processes
A R T I C L E S
Figure 12. State diagram for 1, 2, and 5 (left-hand side) and for 4 (extension
right-hand side).
Table 7. Rise Time τHT of the TA Peak at Ca. 740 nm and kHT as
Well as Life Time τCR of the Final Charge Separated State and
kCR of the Charge Recombination.
acetonitrile (TA)
benzonitrile(TA)
CH2Cl2
τ
HT/ps
τ
CR/ps
τ
HT /ps
τ
CR/ps
(fluoresc.)
1
1
1
1
1
(kHT/109 s-
)
(kCR/109 s-
)
(kHT /109 s-
)
(kCR/109 s-
)
(kHT/109 s-
)
1
4
5
3.2 (310)
20 (50)
2.3 (440)
22 (45)
410 (2.5)
23 (44)
30 (33)
250 (4)
28 (36)
180 (5.5)
>2000 (<0.5)
480 (2.1)
(1.2)
(0.20)
(0.96)
issdependent on the conformation of the moleculesan initial
extent of charge separation which will immediately undergo
relaxation toward the final extended charge transfer state within
200-300 fs.
The rate constants for HT and charge recombination resulting
from spectra deconvolution of the pump-probe measurements
for 1, 2, 4, and 5 in acetonitrile and benzonitrile are collected
in Table 7. From these data it is apparent that the extension of
the cascade (4 vs 1) slows down both kHT and kCR by one order
of magnitude. In the same way, replacement of the polar
acetonitrile by the less polar benzonitrile slows down both rate
constants by one order of magnitude because back electron
transfer occurs in the Marcus inverted region. For comparison,
the rate constants kHT from fluorescence measurements in
CH2Cl2 are also given in Table 7. The values are still smaller
than those obtained in benzonitrile because CH2Cl2 is somewhat
less polar than benzonitrile.
Figure 11. Temporal evolution of the pump-probe spectra of (a) 5 and
(b) 4 in acetonitrile, 3 ps to 150 ps after excitation (various step sizes).
Early spectra are shown in blue/green and late spectra are shown in orange/
red colors. The insets show the time-dependent absorption at 740 nm (see
text for details).
in spectral distribution) suggesting that the two tolandiamine
branches are acting as independent substituents.
Figure 11b shows the pump-probe spectra of the extended
redox system 4. Most noticeable is the kinetic behavior (inset
11b) which differs qualitatively and quantitatively from all other
systems. The instant rise of the 740 nm signal is followed by
an ultrafast decay (200-300 fs) and then another slow rise with
a time constant of 20 ps. These characteristics can be explained
by assuming that the signal is a composite of two overlapping
spectral contributions. Given the structure of 4 it seems
reasonable to propose the following model: Photoexcitation at
360 nm of the tolandiamine units leads to two parallel
processes: a direct charge separation from the tolandiamine
excited-state yielding A--Tara1-Tara2-(Tara3)+ (which occurs
in 200-300 fs) and a slower indirect hole transfer from the CT
state A--(Tara1)+-Tara2-Tara3 to A--Tara1-Tara2-(Tara3)+
in 20 ps (see Figure 12). The latter process is analogous to the
ones observed in the shorter cascades. In 4 there are actually
two similar but distinctly different tolandiamine groups which
should result in (partially) overlapping absorption transitions.
Nevertheless, these two tolandiamine units are likely to exhibit
different reactivity with respect to hole transfer. The direct
charge separation process can be rationalized based on the fact
that the tetraaryl tolandiamine adjacent to the acridine is actually
an electron D/A moiety which can form biradical intermediates.
Hence, upon excitation of the tolandiamine system in 4 there
Conclusions
From the fluorescence data in Table 6 it is obvious that the
hole transfer from Tara1 to Tara2 is rapid and is the same for
1, 2 and 5 in CH2Cl2 within experimental error. The hole transfer
can compete very efficiently with the deactivation pathways in
the reference chromophores as can be seen from the fluorescence
quantum yields which are very low for 1, 2, 4, and 5 (0.01-
0.02 in CH2Cl2) but high for the reference compounds 6 and 7
(0.90 and 0.58, respectively, in CH2Cl2). Interestingly, there is
no significant difference for the hole transfer rate in the branched
molecule 5 compared to the linear array 1. However, the hole
transfer from Tara1 to Tara2 in 4 is significantly slower
because of the lower free energy difference between Tara2 and
Tara1 in 4 compared to that in e.g. 1 due to the methyl
substitution in Tara2 of 4. This conclusion is fully supported
by the pump-probe measurements in acetonitrile and benzo-
nitrile. The hole transfer in 1, 2, 4, and 5 should lead to a charge
separation, i.e., a negative charge on acridine (A), and a positive
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10609

A R T I C L E S
Lambert et al.
The quantum yields are corrected for the index of refraction of the
solvent. Fluorescence decays were measured with a PTI TimeMaster
Model TM-2/2003 stroboscopic boxcar spectrometer and a flash lamp
charged with a H2/N2 1:1 mixture. Excitation wavelength was 381 nm
for all compounds. The instrument response function was obtained by
using a dilute coffee creamer as a scatterer. The decay fits were single
exponential in all cases as checked by the ø2 value, the Durbin-Watson
parameter and the final residuals.
charge on Tara2. In the case of 4 one can expect that the smaller
redox potential of Tara3 compared to Tara2 leads to a second
hole transfer step in which the hole migrates to the terminal
Tara3. If less polar solvents than CH2Cl2 are used, hole transfer
in 1, 2, 4, and 5 is either too slow or even unfavorable because
the free energy difference of charge separation is very sensitive
to the polarity of the solvent. In the case of 3, the hole transfer
is unfavorable irrespective of the solvent due to the wrong
direction of the redox gradient.
The temporal evolution of the fs-broadband pump-probe
spectra provides detailed insight into the photophysical proper-
ties of 1-8, e.g., the data show that the extent of charge
separation in the CT state of 8 strongly depends on the solvent.
While in acetonitrile rapid energy localization competes with
photoinduced charge transfer in benzonitrile these processes are
distinctly separated on the time scale, the latter process being
absent in MTBE.
The analysis of the TA spectra of 1 and 3 prove the HT
process of 1 in polar solvents along the redox gradient while
this process is uphill in 3. The TA spectra also reveal the close
analogy of the branched 5 to the single cascade 1. For the
extended cascade 4 the TA spectra suggest that two different
processes lead to the final charge separated state, a direct charge
separation which follows the excitation of the tolandiamine
antenna moieties and a much slower process which involves
the generation of a CT state followed by a HT step analogous
to those in 1 and 5. In the compounds 1, 2 and 5 but especially
in 4 the tolandiamine moieties not only serve as the cascade
backbone but also possess a distinct antenna effect which
collects all primary excitation energy.
In conclusion, we were able to show that acridine-triarylamine
redox cascades can be used in efficient photoinduced electron-
transfer processes. This PET can be governed by fine-tuning
of the local redox potentials of the triarylamine redox centers,
i.e., the redox gradient along the cascade molecules. Close
analogy to triarylamine compounds used in optoelectronic
devices make the presented arrays hopeful candidates for
applications in, e.g., solar cells or photoconductors.
Femtosecond Broadband Pump-Probe Setup. A detailed descrip-
tion of the setup has been been given elsewhere.58 Briefly, the
compounds were exited by pump pulses at 360 nm at 5 × 10-4 - 5 ×
10-6 M sample concentration. The changes in optical density were
probed by a femtosecond white-light continuum (WLC) generated by
tight focusing of a small fraction of the output of a commercial
Ti:Sapphire based pump laser (CPA-2001, Clark-MXR) into a 3 mm
CaF2 plate. The obtained WLC provided a usable probe source between
370 and 720 nm. The WLC was split into two beams (probe and
reference) and focused into the sample using reflective optics. After
passing through the sample both probe and reference were spectrally
dispersed and simultaneously detected on a CCD sensor. The pump
pulse (340 nm, 100-200 nJ) was generated by frequency doubling of
the compressed output of a commercial NOPA system (Clark-MXR,
680 nm, 8 µJ, 30 fs). To compensate for group velocity dispersion in
the UV-pulse we used an additional prism compressor. Independent
measurements of the chirp of the WLC were carried out to correct the
pump-probe spectra for time-zero differences. The overall time
resolution of the setup was obtained from the rise time of the signal
(above 580 nm). Assuming a Gaussian shape cross-correlation we
obtained a width of 100-120 fs (fwhm). A spectral resolution of 7-10
nm was obtained. Measurements were performed with magic angle
geometry (54.7°) for the polarization of pump and probe pulses to avoid
contributions from orientational relaxation. Pump energy and pump spot
size (∼200-400 µm) were adjusted to minimize contributions from
the solvent to the signal. Steady-state absorption and fluorescence
spectra of the samples measured before and after the time-resolved
experiments were compared with each other and no indications for
degradation were found. A sample cell with 1.25 mm fused silica
windows and a light path of 1 mm was used for all measurements. No
indications for aggregation could be observed in the measurements.
Acknowledgment. This work was supported by the Fonds
der Chemischen Industrie, the Deutsche Forschungsgemein-
schaft, the Volkswagenstiftung, and the Bayerisches Staatsmin-
isterium fu¨r Wissenschaft, Forschung und Kunst (FORMAT
project). We thank JASCO GmbH Deutschland as well as
Heraeus GmbH for kind support.
Experimental Section
Cyclic Voltammetry. The electrochemical experiments have been
performed using a conventional three-electrode setup with a platinum
disk working electrode in dry, argon-saturated CH2Cl2 with 0.2 M
tetrabutylammonium hexafluorophosphate (TBAH) as supporting elec-
trolyte and 0.001 M substrate. The potentials are referenced against
ferrocene (Fc/Fc+). The reversibility or irreversibility, respectively, of
all redox processes were checked by measurements at different scan
rates.
Note Added after ASAP Publication. After this article was
published ASAP on July 7, 2005, a production error in the redox
potential difference equation for 1 in the text just above eq 5
was discovered. The corrected version was published ASAP
on July 11, 2005.
UV-vis/NIR and Fluorescence Spectroscopy. The UV-vis/NIR
spectra were recorded with a JASCO V570 spectrometer in transmis-
sion. Steady-state fluorescence spectra were recorded with a PTI
QuantaMaster Model QM-2000-4 spectrometer using degassed, argon
saturated solvents in the µM concentration range. The spectra are
corrected for the wavelength sensitivity of the detector. Spectra are
converted to wavenumber scale by multiplication of the intensity with
λ2. Fluorescence quantum yields were determined relative to rhodamine
101 in EtOH (conc ) ca. 1 µM) which has a quantum yield of 1.00.
Supporting Information Available: Synthetic details and
characterization for all compounds. X-ray crystallographic
analysis data for 8. This material is available free of charge via
the Internet at http://pubs.acs.org.
JA0511570
(58) Raytchev, M.; Pandurski, E.; Buchvarov, I.; Modrakowski, C.; Fiebig, T.
J. Phys. Chem. A 2003, 107, 4592-4600.
9
10610 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005