Conformational dynamics of charge-transfer states in donor-bridge-acceptor systems

Article abstract of DOI:10.1002/1099-0690(200108)2001:16<3105::AID-EJOC3105>3.0.CO;2-I

Donor-acceptor compounds containing a phenylenediamine electron donor and a naphthalene, a cyanobenzene, or a cyanonaphthalene acceptor were studied. The two chromophores are connected by three different bridging units, consisting of CH₂ groups

Full text of DOI:10.1002/1099-0690(200108)2001:16<3105::AID-EJOC3105>3.0.CO;2-I

FULL PAPER
Conformational Dynamics of Charge-Transfer States in
Donor؊Bridge؊Acceptor Systems
Xavier Y. Lauteslager,[a] Ivo H. M. van Stokkum,[b] H. John van Ramesdonk,[a]
Dick Bebelaar,[c] Jan Fraanje,[d] Kees Goubitz,[d] Henk Schenk,[d] Albert M. Brouwer,*[a] and
Jan W. Verhoeven*[a]
Keywords: Charge-transfer / Conformation analysis / DonorϪacceptor systems / Electron transfer / Fluorescence
Donor−acceptor compounds containing a phenylenediamine
ible in a nonpolar polymer matrix and also in methylcyclo-
electron donor and a naphthalene, a cyanobenzene, or a cy- hexane at low temperature. In the cases of the piperidine-
anonaphthalene acceptor were studied. The two chromoph- and piperazine-bridged donor−acceptor compounds, evid-
ores are connected by three different bridging units, con-
sisting of CH2 groups linked to a semiflexible piperidine or
piperazine ring or to a rigid 2,5-diazabicyclo[2.2.1]heptane
group. All donor−acceptor compounds show photoinduced
charge separation, resulting in the formation of a compact
ence for the involvement of an extended charge-transfer
(ECT) species as a precursor of the CCT species was ob-
tained from time-resolved fluorescence measurements. In
contrast to harpooning systems studied previously, the en-
ergy differences between the species in the present case are
charge-transfer (CCT) state in nonpolar solvents. The con- so small that their interconversions are reversible within the
formational change needed to arrive at this species is imposs-
excited state lifetimes.
Introduction
pounds 1 and 2, Scheme 2) is replaced by a piperazine
bridge (e.g., compound 3).[8]
Electron transfer between an electron donor D and an
electron acceptor A can create charge-transfer states
(DϩAϪ) with charge separation distances as large as ca.
1 nm.[1] Because of the electrostatic attraction between Dϩ
and AϪ, structural changes that bring the charged groups
closer together can occur subsequently. This ‘‘harpooning’’
process was first described for the interaction between alkali
metal and halide atoms,[2] but has also been found to occur
in molecular systems in solution.[3Ϫ7] The prototype com-
pound 1 illustrates the interconversion between a spatially
extended charge-transfer excited state (ECT) and a more
compact CT state labeled CCT (Scheme 1). We have previ-
ously shown that the photoinduced harpooning process is
effectively suppressed when the piperidine bridge (e.g., com-
Scheme 1. The harpooning process, illustrated for compound 1
[a]
Institute of Molecular Chemistry, University of Amsterdam,
Although the conformational flexibility of the piperazine
ring is similar to that of the piperidine ring, the charge dis-
tribution in the initially formed charge-transfer state is very
different for the two systems: In donorϪacceptor com-
pounds 1 and 2 the positive charge resides on the (methoxy)-
aniline group, yielding an ECT species with a dipole mo-
ment of ca. 30 D.[9,10] Piperazine-bridged systems such as
3, on the other hand, can be regarded as (D2ϪD1ϪA)
triads, the trialkyl-substituted nitrogen atom acting as an
electron donor group D1.[11] Trapping of the positive charge
on D1 to yield a D2ϪD1ϩϪAϪ species prevents the forma-
tion of an extended CT (ECT) state (D2ϩϪD1ϪAϪ) and
thereby also precludes the electrostatically driven folding
from ECT to CCT. In radical cations of piperazines
Laboratory of Organic Chemistry,
Nieuwe Achtergracht 129, 1018 WS Amsterdam, The
Netherlands
Fax: (internat.) ϩ 31-20/525-5670
E-mail: fred@org.chem.uva.nl
[b]
Faculty of Sciences, Vrije Universiteit Amsterdam,
De Boelelaan 1081, 1081 HV Amsterdam, The Netherlands
[c]
Institute of Molecular Chemistry, University of Amsterdam,
Laboratory of Physical Chemistry,
Nieuwe Achtergracht 129, 1018 WS Amsterdam, The
Netherlands
[d]
Institute of Molecular Chemistry, University of Amsterdam,
Laboratory of Crystallography,
Nieuwe Achtergracht 166, 1018 WV Amsterdam, The
Netherlands
Supporting information for this article is available on the
WWW under http://www.wiley-vch.de/home/eurjoc or from
the author.
Eur. J. Org. Chem. 2001, 3105Ϫ3118
WILEY-VCH Verlag GmbH, 69451 Weinheim, 2001
1434Ϫ193X/01/0816Ϫ3105 $ 17.50ϩ.50/0
3105

A. M. Brouwer, J. W. Verhoeven et al.
FULL PAPER
the photophysical properties of the donor unit in the ab-
sence of the acceptor required special attention. Electron
acceptor groups employed were the 1-naphthyl group
(compounds a; Ered ϭ Ϫ2.58 V vs. SCE[18]), the 4-cyano-
phenyl group (compounds b; Ered ϭ Ϫ2.2 V[18]), and the 4-
cyano-1-naphthyl group (compounds c; Ered ϭ Ϫ1.91 V[5]).
Results and Discussion
Synthesis
N-(4-Dimethylaminophenyl)piperazine was synthesized
as shown in Scheme 4. N-Benzylpiperazine was coupled
with 1-fluoro-4-nitrobenzene,[19] and the nitro group of the
obtained compound was reduced with sodium borohyd-
ride.[20] Methylation of the obtained N-(4-aminophenyl)-NЈ-
benzylpiperazine with (para)formaldehyde and sodium
borohydride was performed using the method of Giumanini
et al.[21]
Scheme 2. Compounds studied in previous work
(D2ϪD1), the positive charge tends to be delocalized over
both nitrogen atoms,[12] but stabilizing methoxyphenyl sub-
stituents in the free radical cations suffice to localize the
charge on the neighboring amino group, giving rise to a
D2ϩϪD1 charge distribution (as in compound 4 in
Scheme 2).[13Ϫ15]
The localization of the positive charge on D1
(D2ϪD1ϩϪAϪ) in piperazine 3 is a result of the electro-
static interaction with the nearby AϪ. If the oxidation po-
tential of D2 were lower, it should be possible to reach the
fully extended CT state (D2ϩϪD1ϪAϪ) and so restore the
harpooning mechanism. In this study we decided to incorp-
orate a phenylenediamine unit as the electron donor. This
is easier to oxidize than a methoxyaniline, by roughly
0.4 eV, and can form stable radical cations.[16,17] We com-
pare the piperazine systems 5aϪ5d with corresponding
compounds 6a, 6b, and 6d, all with a piperidine bridge, and
compounds 7aϪ7d, all with a conformationally rigid
The compound 5d thus obtained was able to serve nicely
as a model donor compound, incorporating the phenylene-
diamine donor, but lacking an electron-accepting group.
Debenzylation of this compound yielded the desired
piperazine, which could be coupled to different bromome-
thyl-substituted aromatic compounds. As (1S,4S)-N-benzyl-
2,5-diazabicyclo[2.2.1]heptane dihydrobromide was com-
mercially available, we were able to follow the same route
as just described for N-benzylpiperazine. Compound 7d in
this route could be used as a reference donor compound.
The synthesis of piperidine-bridged donorϪacceptor sys-
tems (Scheme 5)[22] starts with the preparation of the appro-
priate piperidone by means of a cyclization reaction be-
tween dimethylamino-p-phenylenediamine and 1,5-
dichloro-3-pentanone. The piperidone can be coupled by
WadsworthϪEmmons condensation to various phosphon-
ates to introduce the aromatic chromophores. Reduction of
the exocyclic double bond finally produces the desired sys-
tems. A model donor compound (6d) was available in our
laboratory.[14]
(1S,4S)-2,5-diaza[2.2.1]bicycloheptane
bridge
(see
Scheme 3). Although ring inversion is precluded in com-
pounds 7aϪ7d, nitrogen inversions and rotation around the
CϪC bonds of the methylene group are still possible.
Model Donor Compounds
UV-Absorption Spectra of the Model Donor Compounds
Compounds 5d, 6d, 7d, and N,N,NЈ,NЈ-tetramethyl-p-
phenylenediamine (TMPD) served as models for the p-
phenylenediamine unit. The maxima and molar absorption
1
1
coefficients of the La and Lb bands observed for all these
compounds are given in Table 1.
The positions of the 1Lb bands of 5d and 6d are identical,
but those of TMPD and 7d are bathochromically shifted.
The largest difference (1400 cmϪ1) is observed between 5d
and 7d (Figure 1). As we show below, the oxidation poten-
tials of 7d and 5d differ significantly as well.
Scheme 3. Compounds under study
Because one of the amino groups of the phenylenediam-
ine unit is part of the bridge, the structure of the bridge can
have a significant effect on the photophysical and electro-
chemical properties of the phenylenediamine unit. For this
Fluorescence of the Model Donor Compounds
The fluorescence spectra of the model donor compounds
reason, the choice of model compounds used to determine 5dϪ7d were recorded in cyclohexane and diethyl ether. All
3106
Eur. J. Org. Chem. 2001, 3105Ϫ3118

Conformational Dynamics of Charge-Transfer States in DonorϪBridgeϪAcceptor Systems
FULL PAPER
Scheme 4. Synthesis of the piperazine-bridged donorϪacceptor systems
Scheme 5. Synthesis of the piperidine-bridged donorϪacceptor sys-
tems
Table 1. Absorption maxima [nm] and molar absorption coeffi-
1
1
cients [Ϫ1 cmϪ1] (in parentheses) of the La and Lb bands of the
phenylenediamine model donor compounds in cyclohexane
Figure 1. UV-absorption spectra of 5d and 7d in cyclohexane
Compound
1La transition
1Lb transition
TMPD
5d
264 (15970)
264 (18040)
265 (16700)
267 (18610)
327 (2410)
319 (2260)
319 (2360)
334 (2570)
Table 2. Quantum yields, decay times [ns], and rate constants [106
6d
s
Ϫ1] of fluorescence (kf), and nonradiative decay (kd) of the refer-
7d
ence donor compounds in cyclohexane and diethyl ether
Cyclohexane
Diethyl ether
Compound
Φf[a]
τ[b]
kf
kd
Φf
τ
kf
kd
compounds show characteristic fluorescence around
390 nm upon excitation at 308 nm. A slight solvent depend-
ency in the position of this band is observed. The quantum
yields and decay times of fluorescence of the four com-
pounds in cyclohexane and diethyl ether are listed in
Table 2.
TMPD and compound 7d show similar radiative and
nonradiative rate constants in the two solvents. Compounds
5d and 6d also show almost identical behavior. For the lat-
TMPD
5d
0.18
0.12
0.11
0.24
4.3
3.4
3.5
5.3
42
35
31
45
191
259
254
143
0.20
0.11
0.11
0.24
5.1
3.9
4.5
6.0
39
28
24
40
157
228
198
127
6d
7d
[a]
Quinine bisulfate in 1  H2SO4 (Φ ϭ 0.546) or anthracene in
cyclohexane (Φ ϭ 0.18)[46] used as standards. Ϫ Determined by
[b]
means of time-correlated single photon counting; decays were
measured at three different wavelengths and could be fitted very
ter two, this is not surprising as the phenylenediamine chro- well with a single exponential function.
Eur. J. Org. Chem. 2001, 3105Ϫ3118
3107

A. M. Brouwer, J. W. Verhoeven et al.
FULL PAPER
mophores are expected to be geometrically very similar.
The largest difference in oxidation potentials is found be-
However, this also implies that the trialkyl-substituted ni- tween 5d and 7d. This difference of circa 120 mV is surpris-
trogen atom in 5d does not influence the photophysical ingly large, as the tetraalkylphenylenediamine units appear
properties of the phenylenediamine part. The similar values very similar. In the next section we show that small but
for TMPD and 7d also point to comparable phenylenedia- significant geometrical differences are responsible for these
mine chromophore conformations. The crystal structure of observations.
compound 7c has been resolved and is presented below,
where we again address this issue.
Quantum Chemical Calculations of the Ionization
Potentials and Excitation Energies
Oxidation Potentials of the Model Donor Compounds
The simplest computational approach to address the
question of why the oxidation potentials of 5d and 7d differ
so much is to calculate the difference between the relaxed
(adiabatic) ionization potentials of the two molecules
[Equation (1)].
With three of the compounds, oxidation potentials were
measured by cyclic voltammetry. Ferrocene was used as an
internal standard.[23,24] The oxidation potentials are refer-
enced to SCE, taking a value of 0.45 V versus SCE for
ferrocene in acetonitrile.[25] TMPD shows two reversible ox-
idation waves at 0.17 and 0.76 V versus SCE in acetonitrile.
Compounds 5d and 7d, however, each show only one re-
versible oxidation wave, at 0.26 and at 0.14 V versus SCE,
respectively. The presence of the saturated trialkyl-substi-
tuted nitrogen atom probably causes the irreversibility of
the second oxidation step, since Ϫ for example Ϫ this step
is reversible in compound 6b, showing up at 0.75 V versus
SCE. The oxidation potential of 6d was not measured, but
might be expected to be close to that of 6b (0.22 V versus
SCE). The cyclic voltammograms, showing the first oxida-
tion waves of three model donor compounds, are displayed
in Figure 2. The oxidation potential(s), including those of
two donorϪacceptor systems, are listed in Table 3.
(1)
If differences are caused by purely geometrical effects, the
second amino group in the ring probably does not play a
role. We therefore took N-phenylpiperidine and N-phenyl-
2-azabicyclo[2.2.1]heptane as models for 5d and 7d, respect-
ively. In principle, the latter compound can exist in two con-
formations, because the amino group is nonplanar. Only the
most stable one (endo) will be considered. To perform the
comparison represented in Equation (1), all calculations
have to be done at the same level of theory. Following previ-
ous favorable experiences,[26,27] we chose to use the B3LYP
hybrid HartreeϪFock/Density Functional method[28] with
the 6-31G* basis set[29] to determine the vertical and re-
laxed ionization potentials (see Figure 3). The obtained re-
sults are given in Table 4.
Figure 2. Cyclic voltammograms of TMPD, 5d, and 7d (circa 2 m
solutions in acetonitrile, 0.1  tetrabutylammonium hexafluoro-
phosphate as electrolyte)
Table 3. Halfwave oxidation potentials in acetonitrile [V] (versus
SCE) of three model donor compounds and two donorϪacceptor
compounds
Figure 3. Schematic potential energy curves of the neutral and cat-
ion state, showing the reorganization energies and vertical and re-
laxed ionization potentials
Table 4. B3LYP/6-31G*-calculated ionization potentials and lowest
excitation energies (eV) and oscillator strengths for N-phenylpiper-
idine and N-phenyl-2-azabicyclo[2.2.1]heptane
Compound
E1/2 (ox1)
E1/2 (ox2)
TMPD
5d
0.17
0.26
0.14
0.27
0.22
0.76
[a]
[a]
[a]
Compound
IPvertical IPrelaxed Eexc f
7d
5b
6b
0.75
N-phenylpiperidine
N-phenyl-2-azabicyclo[2.2.1]heptane 6.36
6.70
6.33
6.20
4.68 0.0035
4.58 0.029
[a]
Irreversible.
3108
Eur. J. Org. Chem. 2001, 3105Ϫ3118

Conformational Dynamics of Charge-Transfer States in DonorϪBridgeϪAcceptor Systems
FULL PAPER
In N-phenylpiperidine, steric interactions between the
phenyl ring and the α-CH2 groups of the piperidine ring
produce a significant twist angle between the rings and pyr-
amidalization of the nitrogen atom (sum of the CϪNϪC
bond angles ΣN ϭ 334°, twist angle ω ϭ 34°). In N-phenyl-
2-azabicyclo[2.2.1]heptane, the steric interaction is smaller
and conjugation between the lone pair and the benzene ring
is more favored, as indicated by the smaller degrees of pyr-
amidalization (ΣN ϭ 353°) and twist (ω ϭ 6°). In the rad-
ical cations, the nitrogen atoms are in a planar trigonal en-
vironment (ΣN ϭ 360°), but in the case of N-phenylpiperid-
ine there is still a twist angle of 17°, while in N-phenyl-2-
azabicyclo[2.2.1]heptane this is only 4°. The computed dif-
ference between the relaxed ionization potentials of N-
phenylpiperidine and N-phenyl-2-azabicyclo[2.2.1]heptane
is 0.13 eV, which agrees remarkably well with the experi-
mental difference in oxidation potentials between 5d and 7d
of 0.12 V!
The less favorable conjugation in the piperidine system
(and presumably also in the piperazine system) is also likely
to be the reason for the different photophysical properties
noted in the previous sections. Indeed, time-dependent
Figure 4. Crystal structure of 7c; the numbering of the non-hydro-
gen atoms is shown in the wireframe presentation
DFT calculations[30] confirm that the lowest excitation en-
ergy is lower for the more conjugated N-phenyl-2-azabicy-
clo[2.2.1]heptane than for N-phenylpiperidine (Table 4).
the closest resemblance to the crystal structure of 7c. The
lowest energy conformer in which the phenylenediamine
chromophore is in the exo position (G) has a steric energy
circa 15 kJ molϪ1 above the global minimum.
Donor-Acceptor Systems
Crystal Structure of 7c and Structure of 7a in Solution
With compound 7c, it was possible to obtain a crystal
suitable for single-crystal X-ray analysis (Exp. Sect.). Crys-
tals of 5c were obtained as well, but the structure could not
be resolved satisfactorily (R ഠ 0.22) by X-ray analysis.
From the analysis it was nevertheless clear that both the
donor and the acceptor chromophores are in an equatorial
position, as has been observed before for an analogous
compound.[9] A presentation of the crystal structure of 7c,
with the numbering of the non-hydrogen atoms, is given
in Figure 4.
Interestingly, both chromophores are in endo positions.
Recently, low-temperature 13C NMR measurements and
molecular-mechanics calculations on 2-methyl-2-azabicy-
clo[2.2.1]heptane have been reported.[31] Experimentally,
the endo isomer was also found to be about 1.2 kJ molϪ1
more stable than the exo isomer in this compound. Further-
more an inversion barrier of 30 kJ molϪ1 was determined,
significantly lower than that in model acyclic amines. In the
crystal structure of TMPD,[32] the bond angles around the
nitrogen atoms sum up to 353°. In 7c, the angles around
N14 sum up to 349°, whereas those around N4 sum up to
345°, indicating a greater degree of pyramidalization.
Molecular-mechanics (Macromodel v.4.5,[33] MM3*
force field) calculations were performed on compound 7a.
Using the Monte Carlo conformational search method, 6
conformers were found within an energy range of 8 kJ
molϪ1; these are depicted in the first three rows of Figure 5
(A to F). In all these conformers the phenylenediamine
Figure 5. Lowest-energy conformers (AϪF) of 7a (MM3 force
field); species G is the lowest-energy conformer in which the naph-
chromophore is in the endo position. Conformer C shows thylmethyl group adopts the exo orientation
Eur. J. Org. Chem. 2001, 3105Ϫ3118
3109

A. M. Brouwer, J. W. Verhoeven et al.
FULL PAPER
1
Variable-temperature 400-MHz H NMR measurements
were performed on 7a in [D8]toluene. At 225 K, most sig-
nals were separated and the proton peaks could be assigned
with the aid of CϪH HETCOR and NOESY experiments.
Relevant intramolecular protonϪproton distances in the
calculated structures AϪG were subsequently measured
and checked for presence or absence of a corresponding
cross peak in the NOESY spectrum. The predominance of
the endo configuration in the acceptor unit is supported by
the absence of cross peaks between the protons on C7 and
those on C19/26 and C17, respectively. When the acceptor
is situated in the exo position, these distances are always
˚
shorter than 3.5 A. A clear cross peak is observed between
the C26 naphthalene proton and the bridgehead proton on
˚
C6, which are ca. 4.8 A apart in the crystal structure. In the
˚
calculated conformer E, this HϪH distance is 3.1 A and all
other observed NOE effects are also in accordance with the
interproton distances in this structure. This indicates that
the naphthalene acceptor chromophore is at least not ex-
clusively in the same orientation in solution as it is in the
crystal structure.
Figure 6. UV-absorption spectra in cyclohexane of compounds 5c
and 7c, and the difference spectra of 7c and 7d, and 5c and 5d, re-
spectively
UV-Absorption Spectra of the Donor؊Acceptor
Compounds
emission of 6a is largely obscured by the strong local fluor-
escence. The long-wavelength bands near 500 nm in 5a, 6a,
and 7a are attributable to CT fluorescence from conforma-
tions in which donor and acceptor are in close contact: this
emission resembles that of intermolecular exciplexes be-
tween TMPD and naphthalene.[34] A similar band was also
observed when the fluorescence of solutions containing
high concentrations of naphthalene added to donor model
compound 5d was measured. A corresponding CT emission
in the series 5bϪ7b occurs at lower energy (λmax ഠ 520 nm),
because of the stronger acceptor involved.
The UV-absorption spectra of 5c and 7c, two compounds
incorporating the strongest acceptor (cyanonaphthalene),
are shown in Figure 6. The spectra consist of the overlap-
ping absorptions of the phenylenediamine and cyanonaph-
thalene chromophores. This can be convincingly shown by
subtraction of the absorptions of the corresponding model
donor compounds 5d and 7d. The difference spectra, in-
cluded in Figure 6, are identical to the spectrum of 1-cyano-
4-methylnaphthalene. Thus, direct excitation of the (com-
pact) charge-transfer state is not possible even for 7c, which
on the basis of its crystal structure can easily adopt a sand-
wich-like conformation in which donor and acceptor have
π-overlap.
In the other solvents, a decrease in local fluorescence in-
tensity is observed on going to more polar solvents. The
same trend can be observed for the two series containing
either a naphthalene or cyanobenzene acceptor.
In the case of the piperidine-bridged harpooning com-
pound 1 studied by Wegewijs (see Scheme 1),[3,4,7,35Ϫ37] it
was convincingly shown that the formation of the compact
Steady-State Fluorescence of the Donor؊Acceptor
Compounds
For the six compounds with naphthalene and cyanoben- charge-transfer (CCT) state is preceded by an extended
zene as electron acceptors, 5aϪ7a and 5bϪ7b, the fluores- charge-transfer (ECT) species in which the piperidine ring
cence spectra were measured in various solvents. The fluor- maintains the chair conformation preferred in the ground
escence maxima (with the total fluorescence quantum yields state. For the compounds studied here, we also observed
in parentheses, including the local fluorescence of the fluorescence from a species in which donor and acceptor
donor) are listed in Table 5. For the 5aϪ7a and 5bϪ7b are in close contact. Because of the similarity of 5 and 6 to
series, the spectra in cyclohexane are given in Figure 7. In 1, we anticipated that the CCT species should be preceded
both series, the piperidine-bridged systems (6a and 6b) by an ECT species in these compounds too. The predicted
show substantially less quenching of the local fluorescence fluorescence maxima for the ECT species of compounds
of the phenylenediamine donor (maximum at 390 nm) than 5aϪb and 6aϪb can be estimated by taking the ECT fluor-
the others do. In the 5aϪ7a series (naphthalene as electron escence maximum of 1 (circa 384 nm in alkane solvents) as
acceptor), compound 7a shows the strongest quenching of a basis and correcting for the differences in redox poten-
the local fluorescence, which is in accordance with the lower tials. In this way we calculated the expected ECT fluores-
oxidation potential of the phenylenediamine donor at- cence maxima of 5a and 6a to be around 372 nm and those
tached to the diazabicycloheptane framework and the fact of 5b and 6b around 400 nm. As we show below, the ECT
that the donorϪacceptor distance in the ground state is sig- emission maxima are actually found at substantially lower
nificantly smaller than that in systems 5a and 6a. The CT energies, which may be due to differences in the charge dis-
3110
Eur. J. Org. Chem. 2001, 3105Ϫ3118

Conformational Dynamics of Charge-Transfer States in DonorϪBridgeϪAcceptor Systems
FULL PAPER
Table 5. CT fluorescence maxima [nm] of the 5a؊7a and 5b؊7b series in various solvents (the total fluorescence quantum yields are
given in parentheses, including the local fluorescence of the donor; excitation wavelength 308 nm)
Solvent
∆f[a]
5a
6a
7a
5b
6b
7b
[b]
[b]
[b]
n-Hexane
Cyclohexane
0.092
0.100
0.100
0.170
0.171
0.194
0.213
0.237
0.251
0.259
0.116[d]
0.122[d]
480 (0.03)
480 (0.03)
480 (0.03)
503[c]
504 (0.02)
508 (0.01)
518 (0.01)
520 (0.01)
547 (Ͻ 0.01)
570 (Ͻ 0.01)
524 (0.02)
554[c]
(0.09)
(0.12)
(0.11)
484 (0.02)
484 (0.02)
484 (0.02)
510[c]
514 (0.02)
518 (0.02)
526 (0.01)
528 (0.01)
543 (Ͻ 0.01)
568 (Ͻ 0.01)
538 (0.02)
564[c]
522 (0.03)
522 (0.04)
522 (0.04)
508 (0.06)
510 (0.07)
510 (0.06)
532 (0.01)
532 (0.01)
532 (0.01)
Methylcyclohexane
Di-n-hexyl ether
Di-n-pentyl ether
Di-n-butyl ether
Di-n-propyl ether
Diisopropyl ether
Diethyl ether
Pentyl acetate
Benzene
499 (0.04)
500 (0.04)
506 (0.01)
508 (0.01)
532 (0.01)
546 (0.01)
554 (0.01)
560 (Ͻ 0.01)
566 (Ͻ 0.01)
542 (0.01)
543 (Ͻ 0.01)
558 (0.01)
560 (Ͻ 0.01)
550 (Ͻ 0.01)
576 (Ͻ 0.01)
507 (0.04)
528 (0.02)
Dioxane
[a]
[b]
[c]
[d]
See Equation (4). Ϫ Maximum obscured by strong local emission. Ϫ Quantum yield not determined. Ϫ Benzene and dioxane
are known to behave as more polar than expected on the basis of their ∆f value.
Figure 8. Fluorescence spectra of compounds 5b؊7b in Telene, ex-
citation wavelength 308 nm
It is tempting to ascribe this band to fluorescence from an
ECT species, although the difference between the position
of this band and the CCT emission band, observed in cyclo-
Figure 7. Fluorescence spectra of 5a؊7a (left) and 5b؊7b (right)
hexane at 522 nm, is rather small. For 6b, a broad CT-type
fluorescence band also seems to be present, slightly red-
shifted in comparison to the local fluorescence. However,
the maximum of this emission band cannot be resolved
properly. Compound 7b shows a broad emission spectrum.
The band around 530 nm in the alkane solvents, ascribed
to emission from a CCT state, seems to have largely disap-
peared, however.
in cyclohexane, intensities are corrected for the differences in ab-
sorbance at the excitation wavelength (308 nm)
tributions and in internal and solvent reorganization ener-
gies between the different systems.
The fluorescence spectra of all compounds were also
measured in a nonpolar polymer matrix (Telene, see Exp.
Sect.). The polarity of this matrix is comparable to that of
alkane solvents. For compounds 5aϪ7a, a single fluores-
cence band around 390 nm was observed in the polymer
films (spectra not shown). The band around 480 nm, which
Temperature Dependence of Fluorescence
With all compounds, fluorescence in methylcyclohexane
had been observed in liquid alkanes and attributed to the was studied as a function of temperature. On lowering the
CCT species, is absent. Clearly, the species in which donor temperature, the compounds with the naphthalene acceptor
and acceptor are in close contact cannot be formed in the Ϫ 5a, 6a, and 7a Ϫ show an increase in the local emission
rigid polymer matrix.
around 390 nm, concomitantly with a decrease in the long
The spectra of compounds 5bϪ7b, scaled to the same wavelength CT emission situated around 480 nm (spectra
intensity, are shown in Figure 8. Compounds 5b and 6b cle- not shown). At 147 K, the temperature at which methyl-
arly show a band around 390 nm, probably local fluores- cyclohexane forms a glass, the spectra of 5a and 6a are sim-
cence from the phenylenediamine donor chromophore. ilar to the fluorescence spectra of the corresponding model
Compound 5b shows an additional band around 470 nm. donor systems 5d and 6d. This implies that the electron
Eur. J. Org. Chem. 2001, 3105Ϫ3118
3111

A. M. Brouwer, J. W. Verhoeven et al.
FULL PAPER
transfer rate is very slow at low temperatures. In the spec-
It is hard to relate the dependence of fluorescence on
trum of 7a at 147 K, a shoulder on the red side of the mobility and temperature to the conformational processes
390 nm band is clearly observable, indicating that, in this that take place in these compounds. A complicating factor
compound, a charge-transfer state can still be populated. in these systems is the low-lying, locally excited state of the
At still lower temperatures (circa 120 K), phosphorescence phenylenediamine chromophore, which hampers observa-
of the naphthalene chromophore is observed.
tion of the extended charge-transfer species. In the next sec-
The spectra at different temperatures of the compounds tion we present time-resolved fluorescence measurements to
with the stronger cyanobenzene acceptor Ϫ 5b, 6b, and 7b provide more convincing evidence for the involvement of
Ϫare displayed in Figure 9. Compound 5b shows a moder- this intermediate ECT state in the formation of the CCT
ate shift of the CT emission band to shorter wavelength species.
upon cooling, but no local fluorescence appears even at the
lowest temperature, which indicates that electron transfer
has a very low activation barrier in this system. The intens- Fluorescence Kinetics of the Donor؊Acceptor Compounds:
ity of the emission band increases at first, but decreases 5a at Room Temperature in Cyclohexane
again from 173 K to 147 K. At 147 K, a maximum at
around 480 nm is observed, close to the value found in the
polymer matrix. We tentatively attribute this emission to an
ECT species. For 6b, it is more difficult to identify the
The fluorescence decays of the donorϪbridgeϪacceptor
compounds 5a, 6a, and 7a were measured in cyclohexane
at multiple wavelengths, using a streak camera system and
the single-photon counting technique (SPC). We describe
changes that are taking place. A slight increase in the local
the streak camera results obtained for 5a in detail. The SPC
emission around 390 nm is observed upon cooling. The
results are in good agreement with these, and are given in
change in the shape of the spectrum below 173 K suggests
the Supporting Information.
that an additional band is appearing in the 400Ϫ450-nm
Three decay times were necessary in order to obtain satis-
factory fits. In Figure 10 we show two representative decay
traces, together with the global fit. In general, the uncer-
tainties in the estimated rate constants are 10% or less.
region. For 7b, which Ϫ like 5b Ϫ is devoid of local fluores-
cence, the appearance of a new fluorescence band at ca.
460 nm can clearly be seen.
Figure 10. Representative traces (A: 395 nm; B: 507 nm) of fluores-
cence intensity versus time and global fits from streak camera
measurements for 5a in cyclohexane; insets show residuals of a fit
with three monoexponential decays; the sum of the contributions
is represented by a solid line through the measured points. Time
constants 0.29, 1.33, and 7.9 ns (solid, dotted, and dashed lines,
respectively); the time base is linear around the maximum of instru-
ment response and logarithmic from 1 to 100 ns
The fitted decay times of 0.29, 1.33, and 7.9 ns are attrib-
uted to the involvement of the locally excited (LE) state of
the phenylenediamine electron donor, an ECT species, and
the CCT species, respectively. This is the first clear experi-
mental evidence that an ECT intermediate is indeed in-
volved in the formation of the CCT species in 5a. The de-
cay-associated spectra (DAS) are shown in Figure 11.
Figure 9. Fluorescence spectra of 5b (A), 6b (B), and 7b (C) in
methylcyclohexane at different temperatures, excitation wave-
length 308 nm
3112
Eur. J. Org. Chem. 2001, 3105Ϫ3118

Conformational Dynamics of Charge-Transfer States in DonorϪBridgeϪAcceptor Systems
FULL PAPER
Figure 13. Equilibrium model to fit the decay traces of 5a in cyclo-
hexane; decay of the LE state and of the ECT state by fluorescence
or nonradiative decay (other than electron transfer) may occur, but
this is not distinguishable in the kinetic experiment
Figure 11. Decay-associated spectra (DAS) of 5a in cyclohexane;
time constants 0.29, 1.33, and 7.9 ns (solid, dotted, and dashed
lines, respectively); to facilitate comparison, the DAS of the fastest
decaying component has been divided by five
The negative amplitude of the LE state at longer wave-
lengths indicates that it is a precursor for the other species.
We subsequently assumed a sequential model with increas-
ing lifetimes for the kinetics Ϫ i.e., LE Ǟ ECT Ǟ CCT Ϫ
to estimate the species-associated spectra (SAS). The first
SAS [solid in Figure 12 and Figure S2 (Supporting In-
formation)] represents the emission spectrum at time zero
and is attributed to the LE state. After 0.3 ns, this first SAS
is replaced by the second SAS (dotted), which possesses a
shoulder near 470 nm that we attribute to the ECT species.
This second SAS is in turn replaced after 1.3 ns by the third
SAS (dashed), which displays a second peak, attributable to
the CCT species, at 500 nm. The fact that the second and
third SASs still possess peaks near 390 nm indicates that
they also contain emissions from the LE state. Thus, the
SASs obtained in this way indicate that equilibria exist be-
tween the three species.
Figure 14. Species-Associated Spectra (SAS) of 5a in cyclohexane
Ǟ
produced by an analysis using an equilibrium model: LE ECT
Ǟǟ CCT (respectively: solid, dotted, and dashed); in order tǟo estim-
ate the rate constants, the ECT and CCT components are assumed
not to emit below 390 nm, and thus their SAS are zero below
390 nm; scaling of SAS assumes that the ensemble LE/ECT/CCT
decays exclusively by CCT; to facilitate comparison, the SAS of the
LE component has been divided by ten
The observed spectra corroborate our assignment of the
decay times to the different species. The locally excited state
of the phenylenediamine electron donor fluoresces around
390 nm and is largely quenched because of electron transfer.
From the locally excited state, a CT species with a max-
imum at ca. 470 nm is formed. This is probably the ECT
state, which is in equilibrium with the locally excited state.
Around 500 nm we observe the CCT species, which in turn
is in equilibrium with the ECT species. The ECT species
has its maximum at rather long wavelength. For compound
5b, however, we observed a fluorescent species at ca. 470 nm
both in a rigid polymer matrix and in methylcyclohexane at
low temperature, media in which the formation of a CCT
state is severely hindered.
Figure 12. Species-associated spectra (SAS) of 5a in cyclohexane;
resulting from an analysis with a sequential model with increasing
time constants of 0.29, 1.33, and 7.9 ns (solid, dotted, and dashed,
respectively); scaling of SAS assumes that the ensemble LE/ECT/
CCT decays exclusively by CCT; to facilitate comparison, the SAS
of the LE component has been divided by ten
Solvent Dependence of the Fluorescence of 5a, 6a, and 7a
The fluorescence of 5a, 6a, and 7a was measured with
the streak camera system in a series of solvents. Three com-
ponents were needed to model the decay traces. For com-
An equilibrium model to fit the data was therefore ap- pounds 6a and 7a in cyclohexane, SPC data were also ob-
plied (Figure 13). In order to estimate the rate constants, tained; four spectrally and temporally distinct components
the ECT and CCT components are assumed not to emit are needed to describe these. This indicates that, although
below 390 nm, and thus their SAS are zero below 390 nm. the streak camera data give a satisfactory picture of the
The shape of the decay below 390 nm is described by three photophysical behavior of the compounds in terms of three
lifetimes and two relative amplitudes and contains all in- species, the situation may in fact be more complex, due, for
formation needed to estimate the five forward and back- example, to the involvement of slightly different conforma-
ward rate constants. When we now estimate the SAS, the tions of the LE or ECT species. For 5a in all solvents, the
spectra shown in Figure 14 are obtained.
kinetics were fitted to the model of Figure 13, in which
Eur. J. Org. Chem. 2001, 3105Ϫ3118
3113

A. M. Brouwer, J. W. Verhoeven et al.
FULL PAPER
Table 6. Estimated emission maxima of the different species of 5a, 6a, and 7a in various solvents, obtained from analysis of the streak
images (excitation wavelength 337 nm)
5a[a]
ECT
6a[a]
ECT
7a[a]
CT1
Solvent
∆f
LE
CCT
LE
CCT
LE
CT2
Methylcyclohexane
Cyclohexane
trans-Decalin
Di-n-pentyl ether
Benzene
Diisopropyl ether
Diethyl ether
0.100
0.100
0.110
0.171
Ϫ
0.237
0.251
25.2
25.0
25.0
24.9
24.1
25.0
24.9
21.6
21.4
21.1
19.2
18.6
18.8
18.2
20.2
19.7
19.7
19.1
18.4
18.5
18.0
24.8
25.0
24.7
24.8
Ϫ
24.8
24.7
24.7
25.0
24.0
20.4
Ϫ
19.3
18.6
20.0
19.6
19.8
19.1
Ϫ
18.8
18.4
24.8
24.9
25.1
24.9
Ϫ
25.0
24.9
20.0
19.9
20.0
19.0
Ϫ
18.5
18.1
19.9
19.7
19.9
18.9
Ϫ
18.4
18.1
[a]
Emission maximum νmax (103 cmϪ1) as defined in Equation (2).
Table 7. Estimated decay times (ns; estimated uncertainties Ͻ 10%) for 5aϪ7a in various solvents, obtained from analysis of the streak im-
ages
5a
τ2
6a
τ2
7a
τ2
Solvent
∆f
τ1
τ3
τ1
τ3
τ1
τ3
Methylcyclohexane
Cyclohexane
trans-Decalin
Di-n-pentyl ether
Benzene
Diisopropyl ether
Diethyl ether
0.100
0.100
0.110
0.171
Ϫ
0.237
0.251
0.35
0.34
0.32
0.16
0.16
0.14
0.08
2.0
1.7
2.5
3.2
4.8
3.8
15
11
9.2
11
16
22
28
33
0.37
1.1
0.90
0.21
Ϫ
0.14
0.21
3.1
2.8
3.0
4.4
Ϫ
4.9
11
31
31
13
40
Ϫ
51
52
0.30
0.30
0.28
0.17
Ϫ
0.17
0.22
2.6
7.6
2.1
4.6
Ϫ
5.0
5.9
17
20
17
25
Ϫ
27
26
equilibria are invoked both between the locally excited state
As can be seen for 5a, the emission maximum of the loc-
and the ECT state and between the ECT state and the CCT ally excited state hardly changes upon increasing the solvent
state. For 6a and 7a, equilibria were also necessary in most polarity, but a significant decrease in its decay time is ob-
cases in order to obtain reasonable SAS. This yields the served, due to the electron transfer process yielding the
emission maxima given in Table 6 and the decay times as ECT species. On increasing the solvent polarity, this ECT
presented in Table 7. The experimental parameters to fit the species shows a red shift, together with an increase in the
emission spectra and the rate constants describing the equi- decay time. The red shift, concomitant with an increase in
librium model for 5a are given in Tables 8 and 9, respect- the CCT decay time, is observed for the CCT species as
ively. The fluorescence spectra were fitted with the model well, with increasing solvent polarity (Table 7).
function given in Equation (2).
If we study the rate constants and the free energy differ-
ences calculated from them (see Table 8) (under the assump-
tion that the LE/ECT/CCT ensemble decays exclusively
through CCT), the following conclusions can be drawn. On
increasing the solvent polarity, the free energy difference
(2)
Note that, with the skewness parameter b Ǟ 0, the expo- between the locally excited state and the ECT state in-
nent in Equation (2) reduces to a Gaussian. The shape de- creases as expected. This is also evident from the increasing
scribed by this model function is indistinguishable from the difference between the emission maxima of the LE and
spectral shape predicted for CT-type emission bands.[38]
ECT species on going to more polar solvents. In the calcu-
Table 8. Estimated rate constants [109
s
Ϫ1] (estimated uncertainties Ͻ 10%) for 5a in various solvents at room temperature and free
energy differences [kJ molϪ1], calculated using the kinetic scheme of Figure 13
Solvent
k1
k2
k3
k4
k5
∆GLEϪECT
∆GECTϪCCT
Methylcyclohexane
Cyclohexane
trans-Decalin
Di-n-pentyl ether
Benzene
Diisopropyl ether
Diethyl ether
0.372
0.374
0.275
0.087
0.089
0.029
0.005
2.4
2.5
2.8
6.1
6.0
7.1
13
0.034
0.053
0.025
0.041
0.017
0.018
0.002
0.56
0.65
0.42
0.26
0.19
0.25
0.064
0.098
0.12
0.10
0.077
0.050
0.039
0.032
Ϫ4.6
Ϫ4.7
Ϫ5.8
Ϫ10.5
Ϫ10.4
Ϫ13.6
Ϫ19.5
Ϫ7.0
Ϫ6.2
Ϫ6.9
Ϫ4.6
Ϫ6.0
Ϫ6.5
Ϫ9.0
3114
Eur. J. Org. Chem. 2001, 3105Ϫ3118

Conformational Dynamics of Charge-Transfer States in DonorϪBridgeϪAcceptor Systems
FULL PAPER
˚
lated free energy difference between the ECT and CCT spe- ity radius of 5.4 A is assumed). This holds equally for both
cies, no clear trend can be discerned. It would be expected CT species of 7a. The slope for the ECT species of 5a is
that this difference would decrease, due to the relatively rather small compared to previously studied piperidine-
greater stabilization of the ECT species with increasing bridged harpooning compounds 1 and compound 6a. It
solvent polarity, relative to that of the CCT species. This seems likely that the negative charge on the acceptor plays
shows up in, for example, the decreasing difference between a role in polarizing the ECT state in 5a in such a way that,
the energies of the ECT and CCT emission maxima even in the presence of the very strong phenylenediamine
(Table 6). Perhaps the assumption that the ECTǞCCT con- donor, the positive charge remains partly on the trialkyl-
version takes place quantitatively is incorrect, and the ECT substituted nitrogen atom. As observed previously, weaker
species also decays by other pathways, such as intersystem donors than the phenylenediamine unit result in a situation
crossing to a local triplet state. This becomes more signific- in which the positive charge is mainly localized on that tri-
ant in more polar solvent as the folding rate k4 decreases.
alkyl-substituted nitrogen atom.[8] In these cases, har-
For 6a and 7a, the emission maximum of the shortest pooning is prevented because the electrostatic driving force
living component also lies at approximately the same wave- is absent, but in 5a there is sufficient charge on the aromatic
length in all solvents, and is attributable to emission from donor to allow the folding process. In the CCT species, the
the locally excited state. For 6a, the second component is positive charge is definitely on the phenylenediamine unit.
attributable to an ECT species. In alkane solvents its fluor-
escence maximum is approximately the same as that of the
LE species, as can be concluded from Table 6. We had al-
ready anticipated that the maximum of the ECT species
should be fairly close to that of the locally excited state of
the phenylenediamine electron donor. The position of the
CCT species is comparable to that of compound 5a. Com-
pound 7a also shows two CT species, but at approximately
the same wavenumber. Given the crystal structure of 7a and
the molecular mechanics calculations discussed earlier, we
cannot really speak of ECT and CCT species in this com-
pound. The similar maxima indicate that no large differ-
ences in the donorϪacceptor distance exist in the two CT
species.
The solvent dependence of charge-transfer fluorescence
can be analyzed on the basis of dielectric continuum theory,
using the well known LippertϪMataga Equation (3).[39,40]
In this equation, the solvent dependence of the charge-
transfer emission wavenumber νCT correlates with the solv-
ent polarity parameter ∆f [Equation (4)], where νCT(0)
[cmϪ1] is the emission maximum in the gas phase, µCT [De-
bye] is the dipole moment of the charge-transfer state, h is
˚
Planck’s constant, c is the velocity of light, ρ [A] is the ef-
fective radius of the solvent cavity occupied by the molec-
ule, εs is the solvent dielectric constant, and n is its refract-
ive index.
(3)
Figure 15. LippertϪMataga plots for 5a, 6a, and 7a
(4)
Table 9. Experimental parameters [103 cmϪ1] for 5aϪ7a obtained
from LippertϪMataga analysis [Equations (3) and (4)]
If we plot the emission maxima of the various CT species
versus the corresponding solvent polarity parameters ∆f,
the LippertϪMataga plots shown in Figure 15 are ob-
tained. As can be seen, the CT maxima can be fitted reason-
ably well by linear relationships. The values obtained from
this analysis are given in Table 9. For 5a and 6a, the slopes
obtained for the CCT species are indicative of a species with
a relatively small dipole moment (around 13 Debye if a cav-
Compound
Species
2µ2/hcρ3
hνCT(0)
5a
6a
7a
ECT
CCT
ECT
CCT
CT1
CT2
20.8 Ϯ 2.5
11.2 Ϯ 1.5
40.3 Ϯ 4.4
8.8 Ϯ 1.2
11.7 Ϯ 1.1
11.1 Ϯ 1.1
23.4 Ϯ 0.4
21.0 Ϯ 0.3
28.5 Ϯ 0.8
20.7 Ϯ 0.2
21.1 Ϯ 0.2
20.9 Ϯ 0.2
Eur. J. Org. Chem. 2001, 3105Ϫ3118
3115

A. M. Brouwer, J. W. Verhoeven et al.
FULL PAPER
total of 3525 unique reflections were measured within the range
0 Յ h Յ 11, 0 Յ k Յ 12, 0 Յ l Յ 42. Of these, 1625 were above the
significance level of 2.5 σ(I). The maximum value of sin(θ)/λ was
Conclusion
Study of the different phenylenediamine donor com-
pounds demonstrates the importance of choosing appropri-
ate model chromophores. It has been shown that the re-
markable differences between these compounds can largely
be traced back to differences in molecular geometries of the
phenylenediamine chromophores.
With the available data on the model donor compounds,
it is possible to perform a proper comparison of the differ-
ent semiflexibly and rigidly bridged systems. No evidence
for ground-state interaction between the donor and ac-
ceptor was found, implying that the charge-transfer states
formed in these compounds are populated by excitation to
the locally excited state of the donor and/or acceptor. The
0.70 AϪ1. Two reference reflections (032, 308) were measured
˚
hourly and showed no decrease during the 49 h of collecting time.
In addition, some 60 ‘‘Friedel’’ reflections were measured and used
in the determination of the absolute configuration of the com-
pound. Unit cell parameters were refined by a least-squares fitting
procedure, using 23 reflections with 40° Ͻ 2θ Ͻ 44°. Corrections
for Lorentz and polarization effects were applied. The structure
was solved by the CRUNCH program package.[41] The hydrogen
atoms were calculated. Full-matrix, least-squares refinement on F,
anisotropic for the non-hydrogen atoms and isotropic for the hy-
drogen atoms restraining the latter in such a way that the distance
˚
to their carrier remained constant at approximately 1.0 A, con-
verged to R ϭ 0.052, Rw ϭ 0.046, (∆/σ)max ϭ 0.11, S ϭ 0.279(6).
A weighting scheme w ϭ {10.5 ϩ 0.011·[σ(Fobs)]2 ϩ 0.0002/
[σ(Fobs)]}Ϫ1 was used. The secondary isotropic extinction coeffi-
cient[33] refined to Ext ϭ 0.42(7). The absolute structure para-
meter[42] refined to Xabs ϭ 0, thus confirming the correct enanti-
omer. A final difference Fourier map revealed a residual electron
charge-transfer
fluorescence
observed
for
the
donorϪacceptor systems in nonpolar solvents is attribut-
able to species in which donor and acceptor are in close
contact (CCT). In rigid or very viscous media (polymer
matrix or methylcyclohexane at low temperature), the emis-
sion attributed to the CCT species is not present. Under
these conditions, emission bands that must stem from ECT-
type species are observed for the compounds with the cyan-
ophenyl acceptor Ϫ 5b, 6b, and 7b Ϫ but the steady-state
spectra do not allow accurate determination of band max-
ima.
For the piperidine and piperazine compounds 5a and 6a,
time-resolved fluorescence measurements provide clear
evidence for the involvement of ECT species. For 7a, the
presence of two similar CT species, with relatively small
charge separation distances, can be inferred.
The spectrotemporal analysis of the fluorescence of 5a
demonstrates the presence of equilibria between the differ-
ent excited state species, and has allowed the free energy
differences between them to be estimated. With increasing
solvent polarity, the ECT state is strongly stabilized relative
to the LE state, but the difference between ECT and CCT
does not appear to decrease as expected. Although this may
in part be due to the assumptions made in our analysis, it
should be noted that the analysis of the solvatochromic
shifts of the fluorescence maxima indicates that the ECT
species in compound 5a still has some charge density on
the trialkyl-substituted nitrogen atom, which reduces the
electrostatic interaction that triggers the harpooning pro-
cess. The tendency to delocalize positive charge over the
two nitrogen atoms in a piperazine ring is very strong.
density between Ϫ0.1 and 0.1 eAϪ3. Scattering factors were taken
˚
from Cromer and Mann.[43,44] All calculations were performed with
XTAL,[45] unless stated otherwise. The structure is shown in Fig-
ure 4. Bond lengths and bond angles are reported in Tables S1 and
S2 (Supporting Information).
Electrochemistry: Cyclic voltammetry measurements were per-
formed using an EG&G Model 273 potentiostat at a scan rate of 50
mV sϪ1. Solutions of circa 2 m in acetonitrile (Acros p.A. grade,
distilled from calcium hydride) were used, 0.1  tetrabutylammo-
nium hexafluorophosphate (Fluka, electrochemical grade) being
present as supporting electrolyte. Redox potentials were measured
relative to the Ag/AgNO3 (0.01 ) system and were referenced to
ϩ·
SCE by measuring the oxidation potential of the FeCp2/FeCp2
couple. Since, for the FeCp2/FeCp2ϩ· couple, we measured E 0.05
ox
1/2
V versus Ag/AgNO3 and a value of 0.45 V versus SCE has been
reported in the literature[25] for this couple, a value of 0.40 V has
been added to our values to reference to SCE.
Spectroscopic Measurements: Electronic absorption measurements
were performed with a HewlettϪPackard 8453 diode array spectro-
meter or a Cary 3E (Varian) spectrophotometer. Molar absorption
coefficients were determined using concentrations of 10Ϫ4Ϫ10Ϫ5 .
Commercially available spectrograde solvents were used. Ϫ Telene
OP 2000x (BFGoodrich Company) was obtained as a generous gift
from G. W. Buning (Philips Research Laboratories). Telene is a
hydrocarbon polymer consisting of [2.2.1]bicycloheptane units. The
polymer films were prepared by dissolving the polymer beads and
a small amount of the donorϪacceptor compound in spectrograde
toluene (Aldrich). After spontaneous evaporation of the solvent (at
ambient temperature and pressure), the polymer films were dried
in a vacuum desiccator. Ϫ Fluorescence spectra were recorded with
a Spex Fluorolog II emission spectrometer, using an RCA-C31034
GaAs photomultiplier as detector. Spectra were corrected for the
wavelength-dependent response of the detection system with the
aid of an in-house correction file. Fluorescence quantum yields
Experimental Section
Single-Crystal X-ray Structure Determination of Compound 7c: Or-
thorhombic crystals, P212121, a ϭ 7.893(2), b ϭ 8.919(2), c ϭ were determined relative to a reference solution [anthracene in
3
30.342(5) A, V ϭ 2136.0(8) A , Z ϭ 4, Dx ϭ 1.13 gcmϪ3, λ(Mo-
cyclohexane (ϕ ϭ 0.18) or quinine bisulfate in 1  sulfuric acid
˚
˚
Kα) ϭ 0.71069 A, µ(Mo-Kα)ϭ0.64 cmϪ1, F(000) ϭ 736, room tem- (ϕ ϭ 0.546)][46] and corrected for the refractive index of the solvent.
perature. Final R ϭ 0.052 for 1625 observed reflections. A crystal The samples were made with an absorbance between 0.1 and 0.2
with dimensions approximately 0.40 ϫ 0.65 ϫ 0.70 mm was used at the excitation wavelength and were deoxygenated by purging
for data collection with an EnrafϪNonius CAD-4 diffractometer with argon for 10Ϫ15 min. Commercially available spectrograde
˚
with graphite-monochromated Mo-Kα radiation and ω-2θ scan. A
solvents were used (Merck, Uvasol), except for di-n-butyl ether
3116
Eur. J. Org. Chem. 2001, 3105Ϫ3118

Conformational Dynamics of Charge-Transfer States in DonorϪBridgeϪAcceptor Systems
FULL PAPER
[2]
M. Polanyi, Atomic Reactions, Williams and Nordgate, Lon-
don, 1932.
(Merck, Ͼ 99%, washed three times with sulfuric acid). When the
purity of the solvent was found to be insufficient, the solvent was
purified by standard procedures. All alkyl ethers were distilled from
CaH2 or LiAlH4 prior to use. All acetates were washed with a sat-
urated sodium carbonate solution and distilled from CaH2. Ϫ
Fluorescence measurements at low temperature were performed us-
ing an Oxford Instruments liquid nitrogen cryostat DN 1704 with
an ITC4 control unit. The samples were degassed using at least four
freeze-pump-thaw cycles. Each sample was allowed to equilibrate
thermally for at least 20 min prior to data collection. Ϫ Fluores-
cence decay curves at single wavelengths were measured by picose-
cond time-correlated single-photon counting (SPC). The experi-
mental set-up has been fully described elsewhere.[47] A mode-locked
argon-ion laser (Coherent 486 AS Mode Locker and Coherent In-
nova 200 laser) was used to pump a DCM dye-laser (Coherent
model 700) synchronously. The output frequency was doubled with
a BBO crystal, resulting in 310Ϫ320-nm pulses. A Hamamatsu
microchannel plate photomultiplier (R3809) was used as detector.
The response function [full width at half maximum (fwhm) ഠ 18
ps] was obtained by monitoring the Raman band of a cell filled
with spectrograde water. The cells were painted black with camera
varnish on two adjacent sides to avoid reflection at the quartz/air
boundary. Ϫ The 2D time-resolved fluorescence measurements
were performed using a Hamamatsu streak camera system, con-
sisting of a Chromex IS250 spectrograph, an M5677 slow-speed
sweep unit, a C4792 trigger unit, a C5680 blanking unit, and a
C4742-95 digital CCD camera. For excitation, an LTB MSG400
nitrogen laser (337 nm, fwhm ഠ 0.5 ns) operating at 50 Hz was used
(fwhm ഠ 0.5 ns). The excitation and emission light were fiber-
coupled to the spectrograph. The images consist of 512 ϫ 512 pix-
els. In the time direction, the resolution is limited by the width of
the excitation pulse (fwhm ഠ 500 ps) and by the pointspread func-
tion of the imaging device. Thus, the Gaussian shaped fwhm of the
instrument response depends upon the time base used, which va-
ried from fwhm ϭ 0.5 to 5 ns. In the wavelength direction, the
resolution is limited by the slit and the grating used and was
fwhm ഠ 6 nm. We reduced the number of data points for the ana-
lysis by summing the response over 11Ϫ13 pixels (depending on
the spectral window used), which is effectively a smoothing over
circa 7Ϫ8 nm. Thus, a data set of 512 ϫ 33 points was routinely
used in the further analysis.
[3]
[4]
[5]
[6]
[7]
B. Wegewijs, R. M. Hermant, J. W. Verhoeven, A. G. M. Kunst,
R. P. H. Rettschnick, Chem. Phys. Lett. 1987, 140, 587.
B. Wegewijs, R. M. Hermant, J. W. Verhoeven, M. P. de Haas,
J. M. Warman, Chem. Phys. Lett. 1990, 168, 185.
I. H. M. Van Stokkum, T. Scherer, A. M. Brouwer, J. W. Verho-
even, J. Phys. Chem. 1994, 98, 852.
T. Scherer, I. H. M. van Stokkum, A. M. Brouwer, J. W. Verho-
even, J. Phys. Chem. 1994, 98, 10539.
B. Wegewijs, J. W. Verhoeven, in: Electron transfer Ϫ from isol-
ated molecules to biomolecules, part 1, vol. 106 (Eds.: J. Jortner,
M. Bixon), Wiley, New York, 1999, pp. 221.
X. Y. Lauteslager, B. Wegewijs, J. W. Verhoeven, A. M.
Brouwer, J. Photochem. Photobiol. A 1996, 98, 121.
G. F. Mes, H. J. van Ramesdonk, J. W. Verhoeven, J. Am.
Chem. Soc. 1984, 106, 1335.
X. Y. Lauteslager, M. J. Bartels, J. J. Piet, J. M. Warman, J. W.
Verhoeven, A. M. Brouwer, Eur. J. Org. Chem. 1998, 2467.
R. J. Willemse, J. J. Piet, J. M. Warman, F. Hartl, J. W. Verho-
even, A. M. Brouwer, J. Am. Chem. Soc. 2000, 122, 3721.
A. M. Brouwer, J. M. Zwier, C. Svendsen, O. S. Mortensen, F.
W. Langkilde, R. Wilbrandt, J. Am. Chem. Soc. 1998, 120,
3748.
A. M. Brouwer, P. G. Wiering, J. M. Zwier, F. W. Langkilde,
R. Wilbrandt, Acta Chem. Scand. 1997, 51, 217.
J. M. Zwier, Ph.D. thesis, University of Amsterdam, 2000.
J. M. Zwier, A. M. Brouwer, G. Balakrishnan, T. Keszthelyi, J.
F. Offersgaard, R. Wilbrandt, manuscript in preparation.
J. L. de Boer, A. Vos, Acta Crystallogr., Sect. B 1972, 28, 835.
J. L. de Boer, A. Vos, Acta Crystallogr., Sect. B 1972, 28, 839.
L. Meites, P. Zuman, CRC Handbook Series in Organic Electro-
chemistry, CRC Press, Cleveland OH, 1976.
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
E. C. Taylor, Synthesis 1981, 606.
A. Nose, T. Kudo, Chem. Pharm. Bull. 1981, 29, 1159.
A. G. Giumanini, G. Chiavara, M. M. Musiani, P. Rossi, Syn-
thesis 1980, 743.
T. Scherer, W. Hielkema, B. Krijnen, R. M. Hermant, C. Eijck-
elhoff, F. Kerkhof, A. K. F. Ng, R. Verleg, E. B. van der Tol,
A. M. Brouwer, J. W. Verhoeven, Recl. Trav. Chim. Pays-Bas
1993, 112, 535.
[22]
[23]
[24]
G. Gritzner, J. Kuta, Pure Appl. Chem. 1984, 56, 461.
´
R. R. Gagne, C. L. Koval, G. C. Lisensky, Inorg. Chem. 1980,
19, 2855.
[25]
W. C. Barette, Jr., H. W. Johnson, Jr., D. T. Sawyer, Anal.
Chem. 1984, 56, 1890.
A. M. Brouwer, J. Phys. Chem. A 1997, 101, 3626.
J. M. Zwier, J. Wichers Hoeth, A. M. Brouwer, J. Org. Chem.
2001, 66, 466.
A. D. Becke, J. Chem. Phys. 1993, 98, 5648.
M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B. G.
Johnson, M. A. Robb, J. R. Cheeseman, T. Keith, G. A.
Petersson, J. A. Montgomery, K. Raghavachari, M. A. Al-La-
ham, V. G. Zakrzewski, J. V. Ortiz, J. B. Foresman, J. Cioslow-
ski, B. B. Stefanov, A. Nanayakkara, M. Challacombe, C. Y.
Peng, P. Y. Ayala, W. Chen, M. W. Wong, J. L. Andres, E. S.
Replogle, R. Gomperts, R. L. Martin, D. J. Fox, J. S. Binkley,
D. J. Defrees, J. Baker, J. J. P. Stewart, M. Head-Gordon, C.
Gonzalez, J. A. Pople, Gaussian 94, Revision B.2, Pittsburgh
PA, 1995.
R. E. Stratmann, G. E. Scuseria, M. J. Frisch, J. Chem. Phys.
1998, 109, 8218.
D. A. Forsyth, W. Zhang, J. A. Hanley, J. Org. Chem. 1996,
61, 1284.
I. Ikemoto, G. Katagiri, S. Nishimura, K. Yakushi, H. Kurado,
Acta Crystallogr., Sect. B 1979, 35, 2264.
F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp,
M. Lipton, C. Caufield, G. Chang, T. Hendrickson, W. C. Still,
J. Comput. Chem. 1990, 11, 440.
Syntheses: Details of the syntheses and characterizations of the
compounds studied are provided in the Supporting Information.
[26]
[27]
Supporting Information: Details of syntheses and characterization,
X-ray structural data of compound 7c, comparison of SPC and
streak-camera measurements of 5a in cyclohexane, estimated spec-
tral and kinetic parameters of 5a, 6a, and 7a in various solvents,
as derived from the streak-camera experiments (see also footnote
on the first page of this article).
[28]
[29]
Acknowledgments
We thank Bregje van den Berg and Dr. Jurriaan Zwier for the syn-
thesis of and measurements on compound 6d. Prof. Dr. L.W. Jenne-
skens (Utrecht University) is gratefully acknowledged for use of
his cyclic voltammetry apparatus. This work was supported by the
Netherlands Organization for Scientific Research (Nederlandse Or-
ganisatie voor Wetenschappelijk Onderzoek, NWO).
[30]
[31]
[32]
[33]
[1]
J. W. Verhoeven, in: Electron transfer Ϫ from isolated molecules
to biomolecules, part 1, vol. 106 (Eds.: J. Jortner, M. Bixon),
Wiley, New York, 1999, pp. 603.
Eur. J. Org. Chem. 2001, 3105Ϫ3118
3117

A. M. Brouwer, J. W. Verhoeven et al.
FULL PAPER
[34]
[42]
[43]
K. A. Zachariasse, Ph.D. Thesis, Vrije Universiteit Amster-
H. D. Flack, Acta Crystallogr., Sect. A 1983, 46, 688.
D. T. Cromer, J. B. Mann, Acta Crystallogr., Sect. A 1968, 24,
321.
International Tables for X-ray Crystallography, vol. IV, Kynoch
Press, Birmingham, 1974.
dam, 1972.
[35]
B. Wegewijs, A. K. F. Ng, R. P. H. Rettschnick, J. W. Verho-
[44]
even, Chem. Phys. Lett. 1992, 200, 357.
B. Wegewijs, T. Scherer, R. P. H. Rettschnick, J. W. Verhoeven,
[36]
[45]
[46]
Chem. Phys. 1993, 176, 349.
B. Wegewijs, J. W. Verhoeven, S. E. Braslavsky, J. Phys. Chem.
S. R. Hall, H. D. Flack, J. M. Stewart, XTAL 3.2, 1992.
S. L. Murov, I. Carmichael, G. L. Hug, Handbook of Photo-
chemistry, 2nd ed., Marcel Dekker, Inc., New York, 1993.
S. I. Van Dijk, P. G. Wiering, C. P. Groen, A. M. Brouwer, J.
W. Verhoeven, W. Schuddeboom, J. M. Warman, J. Chem. Soc.,
Faraday Trans. 1995, 91, 2107.
[37]
1996, 100, 8890.
[38]
[47]
R. A. Marcus, J. Phys. Chem. 1989, 93, 3078.
[39]
E. Lippert, Z. Naturforsch., Teil A 1955, 10A, 541.
[40]
N. Mataga, Y. Kaifu, M. Koizumi, Bull. Chem. Soc. Jpn. 1956,
29, 465.
Received December 27, 2000
[41]
R. de Gelder, R. A. G. de Graaff, H. Schenk, Acta Crystallogr.,
[O00663]
Sect. A 1993, 49, 287.
3118
Eur. J. Org. Chem. 2001, 3105Ϫ3118