Systematic study of the photophysical processes in polydentate triphenylene-functionalized Eu3+, Tb3+, Nd3+, Yb3+, and Er3+ complexes

Article abstract of DOI:10.1021/jp994286+

m-Terphenyl-based lanthanide complexes functionalized with a triphenylene antenna chromophore ((Ln)1) exhibit sensitized visible and near-infrared emission upon photoexcitation of the triphenylene antenna at 310 nm. Luminescence lifetime measurements of the (Eu)1 and (Tb)1 complexes in methanol-h₁ and methanol-d₁ revealed that one methanol molecule is coordinated to the lanthanide ion, indicating that all eight donor atoms provided by the ligand are involved in the encapsulation of the lanthanide ion. The luminescence lifetimes of the near-IR-emitting complexes (Er)1, (Nd)1, and (Yb)1 in DMSO-h₆ and DMSO-d₆ are in the microsecond range, and are dominated by nonradiative deactivation of the luminescent state. The processes preceding the lanthanide luminescence in the sensitization process have been studied in detail. The complexed lanthanide ion reduces the antenna fluorescence and increases the intersystem crossing rate via an external heavy atom effect. The subsequent energy-transfer process was found to take place via the antenna triplet state in all complexes. Luminescence quantum yield measurements and transient absorption spectroscopy indicated that in solution two conformational isomers of the complexes exist: one in which no energy transfer takes place, and one in which the energy transfer does take place, resulting in the lanthanide luminescence. The intramolecular energy-transfer rate is higher in the (Eu)1 and (Tb)1 complexes than in the near-infrared-emitting complexes. In methanol the energy-transfer rate is 3.8 × 10⁷ s^-1 for (Eu)1 and (Tb)1. In DMSO-d₆ the intramolecular energy-transfer rate is higher in the (Nd)1 complex (1.3 × 10⁷ s^-1) than in the (Er)1 (3.8 × 10⁶ s^-1) and (Yb)1 (4.9 × 10⁶ s^-1) complexes.

Full text of DOI:10.1021/jp994286+

J. Phys. Chem. A 2000, 104, 5457-5468
5457
A Systematic Study of the Photophysical Processes in Polydentate
Triphenylene-Functionalized Eu3+, Tb3+, Nd3+, Yb3+, and Er3+ Complexes
Stephen I. Klink, Lennart Grave, David N. Reinhoudt, and Frank C. J. M. van Veggel*
Laboratory of Supramolecular Chemistry and Technology and MESA+ Research Institute,
UniVersity of Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands
Martinus H. V. Werts
Laboratory of Organic Chemistry, UniVersity of Amsterdam, Nieuwe Achtergracht 129,
1018 WS Amsterdam, The Netherlands
Frank A. J. Geurts
Department RGL, Akzo Nobel Central Research, P.O. Box 9300, 6800 SB Arnhem, The Netherlands
Johannes W. Hofstraat
Department of Polymers and Organic Chemistry, Philips Research, Prof. Holstlaan 4,
5656 AA EindhoVen, The Netherlands
ReceiVed: December 7, 1999; In Final Form: March 3, 2000
m-Terphenyl-based lanthanide complexes functionalized with a triphenylene antenna chromophore ((Ln)1)
exhibit sensitized visible and near-infrared emission upon photoexcitation of the triphenylene antenna at 310
nm. Luminescence lifetime measurements of the (Eu)1 and (Tb)1 complexes in methanol-h1 and methanol-d1
revealed that one methanol molecule is coordinated to the lanthanide ion, indicating that all eight donor
atoms provided by the ligand are involved in the encapsulation of the lanthanide ion. The luminescence
lifetimes of the near-IR-emitting complexes (Er)1, (Nd)1, and (Yb)1 in DMSO-h6 and DMSO-d6 are in the
microsecond range, and are dominated by nonradiative deactivation of the luminescent state. The processes
preceding the lanthanide luminescence in the sensitization process have been studied in detail. The complexed
lanthanide ion reduces the antenna fluorescence and increases the intersystem crossing rate via an external
heavy atom effect. The subsequent energy-transfer process was found to take place via the antenna triplet
state in all complexes. Luminescence quantum yield measurements and transient absorption spectroscopy
indicated that in solution two conformational isomers of the complexes exist: one in which no energy transfer
takes place, and one in which the energy transfer does take place, resulting in the lanthanide luminescence.
The intramolecular energy-transfer rate is higher in the (Eu)1 and (Tb)1 complexes than in the near-infrared-
emitting complexes. In methanol the energy-transfer rate is 3.8 × 107 s-1 for (Eu)1 and (Tb)1. In DMSO-d6
the intramolecular energy-transfer rate is higher in the (Nd)1 complex (1.3 × 107 s-1) than in the (Er)1 (3.8
× 106 s-1) and (Yb)1 (4.9 × 106 s-1) complexes.
Introduction
1-10 M-1cm-1), and the luminescence lifetimes are relatively
long (ranging from microseconds to milliseconds). Moreover,
The excitation of lanthanide ions by energy transfer from an
organic antenna chromophore to obtain visible and near-infrared
luminescence remains an intriguing concept that can be applied
in many areas varying from fluoroimmunoassays1 to optical
signal amplification.2,3 Provided that most of the excitation
energy is transferred from the antenna chromophore to the
luminescent lanthanide ion, this process is much more effective
than direct excitation, since the absorption coefficients of organic
chromophores are many orders of magnitude larger (typically
3-5) than the intrinsically low molar absorption coefficients
of trivalent lanthanide ions.4 The lanthanide luminescence arises
from intra-4f transitions which are parity forbidden. As a
consequence, the molar absorption coefficients are low (typically
since the 4f orbitals are shielded from the environment by an
outer shell of 5s and 5p orbitals, the emission bands remain
narrow, even in solution or in an organic matrix at room
temperature.
The sensitization pathway in luminescent lanthanide com-
plexes generally consists of excitation of the antenna chro-
mophore into its singlet excited state, subsequent intersystem
crossing of the antenna to its triplet state, and energy transfer
from the triplet to the lanthanide ion.5-8 This simple, yet
adequate, photophysical model is depicted in Figure 1 together
with the lanthanide 4f energy levels responsible for the
luminescence. The overall quantum yield of sensitized emission
(φse) is therefore the product of the triplet quantum yield (φisc),
the energy-transfer quantum yield (φet), and the intrinsic
luminescence quantum yield of the lanthanide ion (φlum), hence
* To whom correspondence should be addressed. Fax: +31-53-4894645.
E-mail: F.C.J.M.vanVeggel@ct.utwente.nl.
10.1021/jp994286+ CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/23/2000

5458 J. Phys. Chem. A, Vol. 104, No. 23, 2000
Klink et al.
Figure 1. Left: Photophysical model describing the main pathways in the sensitization process. Right: Energy diagram of the 4f levels responsible
for the lanthanide luminescence, where a filled circle denotes the lowest luminescent state and an open circle denotes the highest nonluminescent
state (adapted from Stein, G.; Wu¨rzberg, E. J. Chem. Phys. 1975, 62, 208).
tionalized complexes of the near-infrared-emitting lanthanide
ions Er3+, Yb3+, and Nd3+ are very promising not only for
application in fluoroimmunoassays,16 but also for use in laser
systems,17,18 and for amplification of light.2,3
φse ) φiscφetφlum
(1)
For an efficient population of the antenna triplet state upon
excitation into its first singlet excited state, high absorption
coefficients of the antenna at wavelengths in a suitable excitation
window and a high intersystem crossing yield are required. The
complexed lanthanide ion participates in the population of the
triplet state by enhancing the intersystem crossing in the antenna
chromophore via an external heavy atom effect.9 The subsequent
transfer of the excitation energy from the triplet state takes place
via an electron exchange (Dexter) mechanism.10 According to
this theory, the energy-transfer rate is determined by the distance
between the antenna and the lanthanide ion, and by the spectral
overlap of the phosphorescence spectrum of the antenna and
the absorption spectrum of the lanthanide ion. Dexter derived
the following relationship between the distance and the energy-
transfer rate:
Our ongoing research is concerned with the development of
a polymer-based optical amplifier in which organic lanthanide
complexes are incorporated in polymer waveguides. In the
telecommunication network optical signal amplifiers are based
on lanthanide-doped inorganic materials in which an optical
transition of Er3+ is used for amplification of light around 1550
nm,19 and an optical transition of Pr3+ is used for the amplifica-
tion of light around 1300 nm.20 Recently, a polymeric waveguide
doped with neodymium chloride has been shown to amplify
light of 1060 nm.21
Steemers et al. have shown that triphenylene can be incor-
porated into a calix[4]arene-based ionophore, and that the
corresponding Eu3+ and Tb3+ complexes exhibit triphenylene-
sensitized lanthanide luminescence.22 The triphenylene antenna
allows excitation up to 350 nm, and it has a high intersystem
crossing quantum yield (0.89).23 Oude Wolbers et al.
have reported the synthesis and photophysical properties of
m-terphenyl-based lanthanide complexes.13 Recently, we have
developed a synthesis route that allows the incorporation of a
sensitizer into an m-terphenyl-based ionophore.24,25 In this paper,
we report the synthesis of an m-terphenyl-based ligand that has
been functionalized with a triphenylene antenna chromophore
(Chart 1), and a systematic study of the photophysical processes
of the corresponding visible-light-emitting Eu3+ and Tb3+
complexes and near-IR-light emitting Nd3+, Yb3+, and Er3+
complexes. The present ligand is based on the m-terphenyl and
offers eight oxygen donor atoms for complexation and shielding
of the lanthanide ion: three bidentate oxyacetate moieties and
two amide oxygens. The resulting complexes are overall neutral.
The triphenylene antenna chromophore has been functionalized
with an amide carbonyl to position it in close proximity of the
lanthanide ion upon coordination of this functionality.
ket ∼ e-2r/L
(2)
where r is the distance between the donor and acceptor, in this
case the antenna chromophore and lanthanide ion, and L is the
sum of the van der Waals radii of the donor and acceptor. The
energy-transfer process is strongly distance dependent and the
transfer rate diminishes rapidly at distances larger than 5 Å.
The triplet state of the antenna should be matched to the
luminescent state of the receiving lanthanide ion to fulfill the
energetic requirements, but approximately 2000 cm-1 higher
in energy to ensure a fast and irreversible energy transfer.5
A general strategy in the design of luminescent lanthanide
complexes is to synthesize a ligand that comprises lanthanide-
complexing moieties, preferably charged oxygen donor atoms
such as carboxylates,4 and a sensitizer in close proximity to the
bound lanthanide ion. Other examples of reported ligand systems
for lanthanide complexation are cryptand-type ligands contain-
ing 2,2′-bipyridine,11 DOTA aza-crown ether derivatives,12
m-terphenyl-based ionophores,13 and calix[4]arene-based iono-
phores.14 Sensitizer-functionalized Eu3+ and Tb3+ complexes
have already been applied as long-living luminescent probes in
time-resolved fluoroimmunoassays,15 whereas sensitizer-func-
Results and Discussion
Synthesis. The triphenylene-functionalized ligand (H3)1 was
synthesized in four steps starting from bis(amine) 524 and

Photophysical Processes in Polydentate Ln Complexes
J. Phys. Chem. A, Vol. 104, No. 23, 2000 5459
CHART 1
absorption spectrum of the antenna, and proves that the
lanthanide ion is excited via the triphenylene moiety.
The relative intensities and splitting of the (Eu)1 emission
bands are influenced by the symmetry of the ligand, and in
particular by the symmetry of the first coordination sphere.
Experimental data on a variety of Eu3+ complexes established
that the emission band centered around 590 nm corresponding
to the 5D0 f 7F1 transition, which is a magnetic dipole transition,
is relatively strong and largely independent of the coordination
sphere, i.e., the ligand field. The emission band centered around
615 nm, corresponding to the 5D0 f 7F2 transition, which is an
electric dipole transition, is extremely sensitive to the symmetry
of the coordination sphere and is therefore called hypersensitiVe.
It has been established that the intensity ratio of the 5D0 f 7F2
5
7
transition and the D0 f F1 transition is a measure for the
symmetry of the coordination sphere.28 In a centrosymmetric
environment the magnetic dipole 5D0 f 7F1 transition of Eu3+
is dominating, whereas distortion of the symmetry around the
triphenylene aldehyde 222 (see Scheme 1). Reaction of bis-
(amine) 5 with 1.3 equiv of benzoyl chloride gave the mono-
(amide) 6 in 20-30% isolated yield. Triphenylene carboxylic
acid 3 was synthesized in 80% yield by mild oxidation of
triphenylene aldehyde 2. Triphenylene carboxylic acid chloride
4 was prepared in situ with thionyl chloride (SOCl2). After
removal of the excess SOCl2, the triphenylene carboxylic acid
chloride 4 was reacted with mono(amide) 6 in dichloromethane
in the presence of Et3N, giving 7 in 60% yield. After quantitative
hydrolysis of the tert-butyl esters of 7 with trifluoroacetic acid
(TFA), the corresponding complexes were readily formed upon
addition of the lanthanide nitrate salts to methanol solutions of
the ligand in the presence of Et3N as a base. FAB mass spectra
indicate that the complexes have a 1:1 stoichiometry. The IR
spectra confirm the presence of carboxylates.
5
ion causes an intensity enhancement of the hypersensitiVe D0
7
f F2 transition. Complexes with an asymmetric coordination
sphere such as lanthanide-tris(â-diketonate) complexes have
7F2/7F1 intensity ratios ranging from 8 to 12, whereas an intensity
ratio of 0.67 has been reported for the centrosymmetric Eu-
7
tris(oxydiacetate) complex.28 In the present case the F2/7F1
intensity ratio for (Eu)1 is 4. This value is significantly higher
than 0.67, which is not surprising, since the structures obtained
from molecular modeling simulations show that the first
coordination spheres of the (Ln)1 complexes have a (time-
averaged) Cs symmetry.29
The (Eu)1 spectrum at room temperature shows a splitting
5
7
5
7
of the D0 f F1 emission and the D0 f F2 emission, and a
single peak at 580 nm corresponding to the 5D0 f 7F0 transition.
Lowering the temperature to 77 K did not alter the shape and
splitting of the emission bands (not shown). The values of the
ligand field splittings are small, approximately 200 cm-1 for
the 7F1 state and 130 cm-1 for the 7F2 state, which is due to the
shielding of the 4f orbitals from the environment by an outer
shell of 5s and 5p electrons.4 The Eu3+ 7F0 state is nondegenerate
and cannot be split by the ligand field; therefore, the single peak
at 580 nm indicates that there is only one (time-averaged)
luminescent Eu3+ species in solution.
Molecular Dynamics Studies. The structure of the (Ln)1
complexes (Ln ) Nd3+, Eu3+, Er3+, Yb3+) was minimized in
the gas phase and subjected to molecular dynamics simulations
(500 ps) in a cubic OPLS26 box of methanol or DMSO using
the CHARMM force field.27 The simulations show that the
structures of the complexes are comparable in both solvent
boxes: all eight donor atoms from the ligand (three phenol ether
oxygens (PhO), three carboxylate oxygens (OCO), and two
amide oxygens (NCO)) are coordinated to the lanthanide ion,
and one methanol or DMSO molecule is coordinated to the
lanthanide ion. The PhO-Ln (typically 2.5 Å), OCO-Ln
(typically 2.4 Å), and NCO-Ln (typically 2.3 Å) distances are
comparable in this series of (Ln)1 complexes, despite the
differences in ionic radii of the various lanthanides. In the
present complexes the coordination geometry is controlled not
only by steric hindrance and Coulombic interactions, but also
by the conformational restraints of the ligand system. In the
(Ln)1 complexes, the average distance from the center of the
lanthanide ion to the center of the sensitizer is approximately 8
Å. The coordination of the amide carbonyl to the lanthanide
ion positions the bottom phenyl ring of the triphenylene
sensitizer at approximately 5 Å from the lanthanide ion.
Sensitized Lanthanide Emission. a. (Eu)1 and (Tb)1. Upon
excitation at 310 nm, the emission spectra of the (Eu)1 complex
in methanol show the typical narrow bands corresponding to
the Eu3+-centered 5D0 f 7Fj transitions (solid line in Figure 2),
with the strongest emission located around 615 nm originating
Photoexcitation of the antenna at 310 nm in the (Tb)1
complex gives rise to the characteristic green Tb3+ emission
corresponding to the 5D4 f 7Fj transitions (solid line in Figure
3). The strongest emission is centered around 545 nm and
corresponds to the hypersensitive 5D4 f 7F5 transition. Monitor-
ing the intensity of the 545 nm emission band, the excitation
spectrum of (Tb)1 proves the photosensitization via the antenna.
The Tb3+ spectrum shows some fine structure within the
emission bands, but they do not provide a basis for a reliable
diagnostic probe of the symmetry of the complex, as is the case
for the Eu3+ spectra.30
The time-resolved luminescence spectra show monoexpo-
nential decays with a luminescence lifetime of 0.86 ms for (Eu)1
and 1.74 ms for (Tb)1 in methanol (see Table 1). These lifetimes
increase substantially in methanol-d1, caused by the well-known
sensitivity of the Eu3+ and Tb3+ luminescence toward solvent
hydroxyl groups.31 An empirical relationship has been estab-
lished that estimates the number of coordinated methanol
molecules (n):32,33
5
7
from the D0 f F2 transition. Scanning the excitation wave-
length while monitoring the intensity of the Eu3+ emission at
615 nm shows which transitions from the ground state, directly
or indirectly, lead to population of the Eu3+ 5D0 luminescent
state. The resulting excitation spectrum (dashed line in Figure
2) shows a maximum around 300 nm, closely resembles the
n ) q(1/τh - 1/τd - kcorr
)
(3)
where q is 2.1 for Eu3+ and 8.4 for Tb3+, τh is the luminescence
lifetime of the complex in methanol-h1, τd is the luminescence

5460 J. Phys. Chem. A, Vol. 104, No. 23, 2000
Klink et al.
SCHEME 1a
a Reagents and conditions: (i) H2NSO3H, NaClO2, CHCl3/acetone/water, 3 h, room temperature (80%); (ii) SOCl2, reflux 4 h (100%); (iii)
benzoyl chloride, Et3N, CH2Cl2, 12 h, rt (20%); (iv) triphenylene acid chloride, Et3N, CH2Cl2, 12 h, rt (70%); (v) TFA, 12 h, rt (100%).
Figure 2. Excitation spectrum (dashed line, λem ) 615 nm) and
emission spectrum (λex ) 310 nm) of a 10-4 M solution of (Eu)1 in
methanol. The emission bands correspond to 5D0 f 7Fj transitions and
were recorded with a 1 nm emission bandwidth.
Figure 3. Excitation spectrum (dashed line, λem ) 545 nm) and
emission spectrum (λex. ) 310 nm) of a 10-4 M solution of (Tb)1 in
methanol. The emission bands correspond to 5D4 f 7Fj transitions and
were recorded with a 1 nm emission bandwidth. The inset shows a
magnification of the emission bands between 625 and 700 nm.
lifetime of the complex in methanol-d1, and kcorr is a term to
TABLE 1: Luminescence Lifetimes of (Eu)1 and (Tb)1
Measured in Methanol-h1 (τh), Methanol-d1 (τd), and
DMSO-h6 (τh), Number of Coordinated Methanol-h1
Molecules (n) Determined from Eq 1, and Overall Quantum
Yield of Sensitized Emission (OSe)a
correct for closely diffusing second-sphere methanol molecules,
which is 0.125 ms-1 for Eu3+ and 0.03 ms-1 for Tb3+ 33
The
.
calculations show that approximately one methanol molecule
is coordinated to the lanthanide ion (see Table 1). The
coordination number of Eu3+ and Tb3+ complexes in solution
is usually 9,34 which implies that all eight donor atoms of the
ligand are coordinated to the lanthanide ion. This is in excellent
agreement with the results obtained from the molecular dynam-
ics studies.
complex
τh (ms)
τd (ms)
n
φse
(Eu)1
(Tb)1
(Eu)1
(Tb)1
0.86b
1.74b
1.89d
2.21d
2.21c
2.78c
nd
1.2 ( 0.5
1.6 ( 1.5
na
0.03b
0.15b
0.02d
0.03d
nd
na
The efficiency of the overall sensitization process was
established by determining the quantum yield of the sensitized
emission, which is 0.03 for (Eu)1 and 0.15 for (Tb)1 in
methanol. The emission spectra of these complexes in DMSO-
h6, a strongly coordinating aprotic solvent, closely resemble the
spectra in methanol, but surprisingly the overall quantum yields
are much lower: 0.02 for (Eu)1 and 0.03 for (Tb)1. Time-
resolved measurements showed that the luminescence lifetimes
a Excitation at 310 nm, 10-4 M solutions. b In methanol-h1. c In
methanol-d1. d In DMSO-h6.
are longer in DMSO than in methanol-h1 (see Table 1);
therefore, the sensitization process of the (Eu)1 and (Tb)1
complexes in DMSO must be much less efficient than in
methanol. The sensitization process will be discussed in more
detail below.

Photophysical Processes in Polydentate Ln Complexes
J. Phys. Chem. A, Vol. 104, No. 23, 2000 5461
TABLE 2: Luminescence Lifetimes of the Near-IR-Emitting
(Ln)1 Complexes in DMSO-h6 (τh ) 1/kh) and DMSO-d6 (τd
) 1/kd), as Well as the Natural Lifetimes (τ0 ) 1/k0)a
complex τh (µs) kh (s-1
)
τd (µs)
kd (s-1
)
τ0 (ms) k0 (s-1
)
(Yb)1
(Er)1
(Nd)1
9.4
2.4
1.4
1.1 × 105 18.6 0.54 × 105
2.0
14.0
0.25
500
71
4000
4.2 × 105
7.1 × 105
3.4 2.9 × 105
2.5 4.0 × 105
a Excitation at 337 nm, 10-3 M solutions.
TABLE 3: Quantum Yields of the Triphenylene Antenna
Fluorescence in Methanola
complex
φantenna
complex
φantenna
7
0.019
(Yb)1
(Er)1
(Nd)1
0.012
0.0095
0.0095
Figure 4. Emission spectra of 10-3 M solutions of the near-infrared-
emitting complexes (Yb)1, (Nd)1, and (Er)1 in DMSO-d6 upon
excitation at 320 nm recorded with a 15 nm emission bandwidth.
(Eu)1
(Tb)1
(Gd)1
0.0054
0.0086
0.015
a Excitation at 310 nm, 10-4 M solutions.
The luminescence decay curves obtained from time-resolved
luminescence experiments could be fitted monoexponentially
with time constants in the range of microseconds (see Table
2). The luminescence lifetime of (Yb)1 in DMSO-h6 is the
longest (9.6 µs), followed by that of (Er)1 (2.4 µs), and that of
(Nd)1 (1.5 µs). The sensitivity of the near-IR luminescent
complexes toward quenching by the C-H vibrations becomes
apparent when the solvent is changed to DMSO-d6. In DMSO-
d6 the luminescence lifetimes are significantly longer due to
the fact that the C-D vibrations of the coordinated DMSO-d6
molecule are less efficient quenchers of the lanthanide excited
state than the C-H vibrations of a DMSO-h6 molecule.
According to the energy gap law39 the smaller the number of
vibrational quanta that are required to match the energy gap
between the lowest luminescent state and the highest nonlumi-
nescent state of the lanthanide ion, the more effective the
vibronic quenching will be. For the C-H vibration with
vibrational quanta of 2950 cm-1, the number of harmonics (nh)
necessary to match the energy gap is largest for Yb3+ (nh )
3-4), followed by Er3+ (nh ) 2-3), and smallest for Nd3+ (nh
) 1-2). As a consequence, the quenching rate constant by the
solvent C-H vibrations is higher for (Nd)1 than for (Yb)1. For
comparison, the energy gap of Eu3+ (with nh ) 6-7) and Tb3+
(with nh ) 8-9) is much larger than the energy gap of the near-
IR-emitting lanthanide ions, which renders them far less
sensitive toward C-H quenching. The radiative rate constants
(k0), i.e., the rate of spontaneous emission, of Er3+ and Yb3+ in
these type of complexes in DMSO have been added to Table
2. A typical radiative rate constant of Nd3+ has also been added
to Table 2.40 Since the observed rate constant k is the sum of
the natural radiative rate constant k0 and the nonradiative rate
constant knonr, it is obvious from Table 2 that for the near-IR-
emitting ions k is dominated by nonradiative deactivation of
the luminescent state. This competition between nonradiative
and radiative decay is most dramatic for Er3+, because this ion
Figure 5. Absorption spectrum of a 10-2 M solution of (Yb)1 in
DMSO (dashed line) and emission spectrum (solid line) of a 10-3
M
solution of (Yb)1 in DMSO-d6 upon laser excitation at 350 nm (Ar
laser) recorded with a 6 nm emission bandwidth.
b. (Nd)1, (Er)1, and (Yb)1. The near-IR-emitting lanthanide
complexes in DMSO-d6 exhibit the typical line-like lanthanide
emission upon excitation of the triphenylene antenna chro-
mophore, as is depicted in Figure 4. Emission bands at 880
(not shown), 1060, and 1330 nm (4F3/2 f 4I9/2, 4I11/2, and 4I13/2
transitions, respectively) are observed for (Nd)1. The strongest
emission is observed at 1060 nm, whereas the emissions at 880
and 1330 nm are weaker. The shape of the emission bands and
the relative intensities are in agreement with previously reported
spectra of organic Nd3+ complexes in solution.13,16,35 A single
emission band centered at 1540 nm (4I13/2 f 4I15/2 transition) is
observed for (Er)1. No emission was observed from higher
excited states, such as green emission at 545 nm corresponding
4
4
to the S3/2 f I15/2 transition, despite the fact that these states
are populated via the antenna triplet state. Apparently, the higher
excited states relax nonradiatively to the 4I13/2 state, from which
radiative decay is observed. Upon excitation of (Yb)1 at 310
nm, sensitized emission was observed at 980 nm (2F5/2 f 2F7/2
transition). A shoulder seems to be present at the lower energy
side of the 980 nm emission band, which is indicative of splitting
of the 2F7/2 manifold by the ligand field. Using laser excitation,
an emission spectrum of (Yb)1 was obtained (solid line in Figure
5) with a higher resolution (6 nm) in which a sharp peak at 977
nm (10 235 cm-1) is observed with a broad shoulder at 1005
nm (9950 cm-1). The same sharp peak at 977 nm is observed
in the absorption spectrum (dotted line in Figure 5) together
with a broad shoulder at 940 nm (10 600 cm-1), which means
that also the 2F5/2 manifold is split by the ligand field. The shapes
of the luminescence and absorption spectra are similar to those
of the reported spectra of Yb3+ in both organic and inorganic
matrixes.36,37,38
has a very low k0 of 71 s-1
.
Antenna Fluorescence and Phosphorescence. The first step
of the sensitization pathway is the population of the triphenylene
triplet state (3ππ*) via its singlet excited state (1ππ*). In this
step fluorescence and radiationless deactivation from the tri-
phenylene 1ππ* state compete with the spin-forbidden conver-
sion to the 3ππ* state. The shapes of the UV absorption spectra
of the (Ln)1 complexes in methanol are similar to that of ligand
7, indicating that the lanthanide ion does not significantly
1
1
influence the energy of the triphenylene ππ* state. The ππ*
state energy of 29 000 cm-1 (345 nm) was determined from
the 0-0 transition, which is clearly discernible in the absorption

5462 J. Phys. Chem. A, Vol. 104, No. 23, 2000
Klink et al.
crossing yield of the chromophore, which results in a reduction
of the fluorescence intensity and a concomitant increase in the
phosphorescence intensity. This effect has been attributed to
an enhanced spin-orbit coupling of the system which relaxes
the selection rules for electronic transitions, but also to an
exchange interaction of metal-unpaired electrons with the
π-electrons of the organic chromophore. Tobita and co-workers9
have studied the influence of different lanthanide ions on the
photophysical properties of the ligand (the antenna) in lan-
thanide-tris(methylsalicylate) and lanthanide-tris(benzoyltri-
fluoroacetonato) complexes. They found that the decrease of
the antenna fluorescence intensity depended on the magnetic
properties of the lanthanide ion. The paramagnetic Gd3+ ion
was found to lower the antenna triplet-state lifetime more than
the diamagnetic Lu3+ and La3+ ions. If these results were solely
caused by spin-orbit coupling, then the effect should have
increased with increasing atomic number Z. Since this was not
the case, an additional exchange mechanism between the metal-
unpaired electrons and the chromophore electrons was proposed.
The exchange leads to mixing of the chromophore singlet state
and triplet state, ultimately resulting in more allowed singlet-
triplet and triplet-singlet conversions. Also in the present case
the effect, i.e., the reduction of the antenna fluorescence, does
not increase with increasing Z, which implies that a paramag-
netic exchange mechanism may also be predominant.
Figure 6. Fluorescence spectra of the triphenylene antenna in ligand
7 and in the (Ln)1 complexes in methanol upon excitation at 310 nm.
The intensities are relative to the fluorescence intensity of 7.
In the case of (Eu)1, which has the lowest antenna fluores-
cence intensity, a photon-induced electron transfer may also play
a role in the deactivation of the singlet excited state. Instead of
radiative decay to the ground state, or intersystem crossing to
the triplet state, an electron is transferred to the Eu3+ ion upon
excitation of the antenna into its singlet excited state, resulting
Figure 7. Fluorescence and phosphorescence spectra of the tri-
phenylene antenna of 7 and (Gd)1 taken in a methanol-ethanol glass
at 77 K (λexc ) 310 nm).
in the transient formation of an antenna radical cation and Eu2+
.
One of the reasons for the possible occurrence of this competing
process41 is the low reduction potential of Eu3+ in comparison
with other trivalent lanthanide ions.42
spectrum. Upon excitation of the (Ln)1 complexes in methanol
at 310 nm, a structured fluorescence band is observed at 375
nm that corresponds to fluorescence of the triphenylene moiety.
Figure 6 shows the fluorescence intensities of the (Ln)1
complexes relative to the free ligand 7. The fluorescence
quantum yields are summarized in Table 3. The presence of
the lanthanide ion in close proximity of the antenna reduces
the fluorescence intensity, but the effect is not the same for all
the lanthanide ions. The fluorescence quantum yield of ligand
7 is 0.019, whereas the fluorescence quantum yields of the (Ln)1
complexes are even lower, <0.01. The fluorescence quantum
yields are low, which is not surprising, since triphenylene has
an intrinsically high intersystem crossing quantum yield.23
Intramolecular Energy Transfer Process. a. (Eu)1 and
(Tb)1. Since the energy transfer takes place through the
triphenylene triplet state, oxygen may compete with the lan-
thanide ion as the acceptor for the excitation energy, and as a
result the triplet state is quenched. In that case less lanthanide
luminescence will be observed. The effect of oxygen on the
sensitized luminescence intensity gives an indication of the
energy-transfer rate, since the competing oxygen quenching rate
is equal to the product of the diffusion-controlled quenching
rate constant and the oxygen concentration (kdiff[O2]). However,
deoxygenation of the (Eu)1 and (Tb)1 methanol solutions did
not influence the luminescence intensity, which means that ket
The luminescence spectrum of (Gd)1 in a methanol/ethanol
glass at 77 K shows strong phosphorescence bands next to very
weak fluorescence bands. Since Gd3+ has no energy levels
below 32 000 cm-1, Gd3+ cannot accept any energy from the
triphenylene triplet state. The 3ππ* state energy of 22 830 cm-1
of the triphenylene moiety was determined from the 0-0
transition in the phosphorescence spectrum (see Figure 7). This
value is approximately 600 cm-1 lower in energy than the 3ππ*
state of unfunctionalized triphenylene (23 400 cm-1) and is
caused by the electron-withdrawing amide carbonyl at the
2-position.
exceeds 107 s-1 43
.
The energy-transfer process was further studied by probing
the antenna’s triplet state directly with the aid of transient
absorption spectroscopy. To obtain the triplet-triplet absorption
spectra of the complex, samples of (Gd)1 in deoxygenated
DMSO were excited with 335 nm pulses and the changes in
absorbance resulting from the population of the long-lived triplet
state were monitored. As mentioned before, the (Gd)1 complex
enables the study of the photophysical behavior of the antenna
in the presence of a lanthanide ion, but in the absence of energy
transfer. The transient absorption spectrum of (Gd)1 (see Figure
8) shows features at 360 and 440 nm, which are very similar to
the transient absorption spectrum of unfunctionalized tri-
phenylene, but slightly broadened. These absorption bands are
due to triplet-triplet absorptions and can therefore be used to
monitor the lifetime of the triplet state, which was 15.2 µs in
the (Gd)1 complex. This lifetime is the lifetime of the antenna’s
Compared to the phosphorescence spectrum of the free ligand
7, the antenna phosphorescence bands of (Gd)1 have slightly
shifted to the red. Furthermore, the decrease of the antenna
fluorescence intensity of (Gd)1 is accompanied by an increased
phosphorescence intensity. An external heaVy atom effect can
be induced by a heavy (paramagnetic) metal ion in close
proximity of a chromophore, and increases the intersystem

Photophysical Processes in Polydentate Ln Complexes
J. Phys. Chem. A, Vol. 104, No. 23, 2000 5463
Figure 8. Transient absorption spectrum of (Gd)1 in deoxygenated
methanol recorded 100 ns after the laser pulse (340 nm, energy <1
mJ).
Figure 10. (a) Quenching of the 615 nm emission band of Eu3+ by
piperylene. (b) Quenching of the 545 nm emission band of Tb3+ by
piperylene. The arrows denote the effect of the increasing piperylene
concentration in methanol. (c) Enhancement of the 980 nm emission
band of Yb3+ and 1060 nm emission band of Nd3+ upon deaeration.
(d) Enhancement of the 1550 nm emission band of Er3+ upon
deaeration. The arrows denote the effect of deaeration of the DMSO-
d6 solutions.
populations of complexes: one in which energy transfer is
absent and which is responsible for the observed long-lived
triplets, and one in which the energy transfer to the lanthanide
ion is relatively fast (responsible for the fast component),
resulting in the sensitized luminescence. This is in line with
the surprisingly low quantum yields of sensitized luminescence
of the complexes in DMSO. According to our transient
absorption measurements the majority of complexes do not
display efficient energy transfer. Changing to methanol as the
solvent, the antenna triplet-state kinetics show the same biex-
ponential behavior as in DMSO, but the contribution of the fast
component has increased to 65%. The corresponding short-lived
triplet lifetime components are 29 and 26 ns for Tb(1) and Eu(1),
respectively. Obviously, the equilibrium between the dark and
the luminescent forms of the complexes has been shifted toward
the luminescent side. Indeed, we already noted that in methanol
the overall luminescence quantum yields are higher than in
DMSO.
To substantiate the proof for the presence of two distinct
forms of complexes in solution, experiments with an external
triplet quencher were performed. cis-Piperylene has a triplet state
Figure 9. Kinetics of the transient absorption measurements obtained
by singular value decomposition of the resulting datasets. The pump
wavelength was 340 nm (<1 mJ/pulse). The corresponding transient
absorption spectra were similar to the one shown in Figure 8. Top:
(Eu)1 in deoxygenated DMSO and methanol. Bottom: (Tb)1 in
deoxygenated DMSO and methanol.
of 20 070 cm-1 23,44 which is approximately 2800 cm-1 lower
,
than that of triphenylene. Addition of increasing concentrations
of piperylene to methanol solutions of (Eu)1 and (Tb)1 quenched
the sensitized lanthanide luminescence upon excitation of the
antenna at 310 nm (see Figure 10a,b). This effect is more
pronounced for Tb3+ than for Eu3+, which may indicate that
the energy-transfer rate to Eu3+ is faster. With the Stern-Volmer
equation 4 the lifetime of the quenched species, i.e., the
triphenylene triplet state, can be calculated. In eq 4, I0 is the
triplet state in the absence of energy transfer, yielding a kT of
6.6 × 104 s-1
.
The transient absorption spectra of the Tb(1) and Eu(1)
complexes in the submicrosecond domain show interesting
details about the energy-transfer process in DMSO and metha-
nol. In DMSO the same absorption spectra are observed as for
the (Gd)1 complex, but the triplet-state kinetics of the antenna
have become biexponential (see Figure 9): besides a “slow”
component (76%) having the same rate constant as observed in
Gd(1), a second small component (24%) is observed, which
has a much faster decay (26 ns for Tb(1) and Eu(1)). The long
triplet lifetimes would suggest that energy transfer to the
lanthanide ion is extremely slow and an oxygen sensitivity of
the lanthanide luminescence should have been observed.
However, the intensity of the luminescence of both (Eu)1 and
(Tb)1 is not influenced by deoxygenation. We therefore attribute
this behavior to the presence of two distinct (conformational)
I0/I ) 1 + kdiffτT[Q] ) 1 + Ksv[Q]
(4)
lanthanide luminescence intensity without quencher, I is the
lanthanide luminescence intensity, kdiff is the diffusion-controlled
quenching rate constant, τΤ is the lifetime of the triphenylene
triplet state, [Q] is the concentration of the quencher, and Ksv
is the product of kdiff and τT, the so-called Stern-Volmer
constant. If the luminescence intensities are plotted against the
quencher concentration, the slope of the fitted data is equal to
the Stern-Volmer constant Ksv (see Figure 11). When for kdiff
a value of 1010 M-1 s-1 is taken,23 the lifetime of the antenna

5464 J. Phys. Chem. A, Vol. 104, No. 23, 2000
Klink et al.
The transient absorption measurements, the piperylene quench-
ing experiment, and the luminescence rise time measurements,
all show that the energy-transfer rates in the (Eu)1 and (Tb)1
complexes are on the order of 107-108 s-1. The results of these
measurements support the hypothesis that two distinct confor-
mations of the complexes occur in solution: one in which the
energy transfer to the lanthanide ion does not occur and which
is responsible for the long-lived triplet lifetime observed by
transient absorption spectroscopy, and one in which energy
transfer is fast and which is responsible for the short-lived triplet
lifetime. In the nonluminescent isomer the antenna and lan-
thanide ion are probably remote or the orientation of the
lanthanide ion with respect to the antenna π-plane is unfavorable
for energy transfer to occur.45 In methanol there seems to be
an equal distribution between the luminescent and nonlumines-
cent conformational isomers of the complexes. This explains
the relatively low luminescence quantum yields that were
measured for (Eu)1 and (Tb)1 in methanol: 0.03 and 0.15,
respectively. On the basis of measured luminescence lifetimes,
the ISC yield of the antenna near unity, and an assumed
complete energy transfer, the luminescence quantum yields can
be expected to be as high as 0.29 for (Eu)1 and 0.44 for (Tb)1.46
As mentioned before, the overall luminescence quantum yields
of these complexes in DMSO are even lower, 0.02 for (Eu)1
and 0.03 for (Tb)1. In this solvent the nonluminescent confor-
mational isomer seems to be dominating (according to the
transient absorption measurements), resulting in an overall very
inefficient energy-transfer process. The decrease in the lumi-
nescence quantum yield when the solvent is changed from
methanol to DMSO is less dramatic for (Eu)1 than for (Tb)1,
because the significant increase of the (Eu)1 luminescence
lifetime partly compensates the inefficiency of the energy-
transfer process. For the (Eu)1 complex in methanol and DMSO,
energy may also be lost in the step preceding the energy transfer,
i.e., the population of the triplet state. On the basis of the
observation that the antenna fluorescence intensity of (Eu)1 (φflu
) 0.0054) is lower than the antenna fluorescence intensity of
(Tb)1 (φflu ) 0.0086), a photon-induced electron transfer (a
metal-to-ligand charge transfer (MLCT)) may also compete with
the intersystem crossing to the triplet state in the former
complex. Since, in general, the deactivation of such an MLCT
state does not result in the population of the Eu3+ 5D0 excited
state,41 a photon-induced electron-transfer process will reduce
the overall luminescence quantum yield.47
Figure 11. Stern-Volmer plot of quenching of triphenylene-sensitized
Eu3+ and Tb3+ emission by piperylene: I0/I ) 1 + kdiff
Ksv[Q].
τtriplet[Q] ) 1 +
TABLE 4: Quenching of Sensitized (Ln)1 Luminescence by
Piperylene (pip) or Oxygen (Ox)a
complex
τT (ns)
kET (s-1
)
relative kET (s-1
)
(Eu)1 (pip)
(Tb)1 (pip)
(Nd)1 (ox)
(Er)1 (ox)
(Yb)1 (ox)
11.3
81.1
76.1
263.0
202.2
8.0 × 107
1.0 × 107
1.3 × 107
3.8 × 106
4.9 × 106
100
13
16
5
6
a Tabulated are the triphenylene triplet-state lifetimes (τΤ) using the
Stern-Volmer equation and the energy-transfer rates (ket).
triplet state can be calculated. This resulted in an antenna triplet
lifetime of 11.3 ns for (Eu)1 and 81.1 ns for (Tb)1. Since the
antenna triplet lifetime is 15.2 µs in the absence of energy
transfer (τT of (Gd)1), the deactivation of the antenna triplet is
dominated by the rate of energy transfer to the lanthanide ion
(ket), and thus it follows that
ket ) 1/τT
(5)
The intramolecular energy-transfer rate in (Eu)1 is 8.0 × 107
s-1, and 1.0 × 107 s-1 in (Tb)1 (see Table 4).
Another way of studying the energy transfer is to monitor
the early stages of the lanthanide luminescence using a streak
camera and pulsed laser excitation. During the time that the
energy transfer from the antenna to the lanthanide ion is in
progress, the lanthanide luminescence intensity is expected to
rise with a rate equal to the rate of the energy transfer. In these
measurements it was found that the first 200 ns of the detected
luminescence was completely dominated by the antenna fluo-
rescence. Although the fluorescence has a low quantum yield,
it has a very high radiative rate (i.e., the photons emitted per
unit time) compared to the lanthanide luminescence. Therefore,
on this short time scale, the lanthanide luminescence is relatively
weak compared to the antenna fluorescence. After the antenna
fluorescence signal had disappeared, no rise in the Tb3+ emission
at 545 nm was observed for (Tb)1, signifying that the energy
transfer to this ion is complete within 200 ns. This means that
ket exceeds 2 × 107 s-1 (on the basis of 99% energy transfer
after 200 ns).
The Stern-Volmer plot of the triplet quenching experiments
only provides information on the luminescent conformational
isomer of the complexes. From these experiments it was
concluded that the energy-transfer rate is significantly faster in
the (Eu)1 complex (8.0 × 107 s-1) than in the (Tb)1 complex
(1.0 × 107 s-1). However, on the basis of transient absorption
measurements the energy-transfer rates were approximately the
same for the (Eu)1 and (Tb)1 complexes (3.8 × 107 s-1). The
explanation for this is that the values obtained from the
piperylene quenching experiment are lower limits, because it
was found that piperylene also quenched the luminescent states
of Eu3+ and Tb3+, albeit to different extents.48 Furthermore,
the quenching of Tb3+ is more efficient, because the triplet state
The (Eu)1 complex exhibited an interesting behavior. After
the disappearance of the antenna fluorescence, it shows emission
5
from its D1 state at 530 nm. This state decays with a time
constant of 2.1 µs, being converted nonradiatively into the 5D0
state that is responsible for the typical Eu3+ emission observed
in steady-state measurements (main emission at 615 nm), and
radiatively to the 7Fj manifold. This behavior is known for the
sensitized emission of Eu3+ if the donating triplet level is above
5
of cis-piperylene (20 070 cm-1) is just below the D4 state of
Tb3+ (20 400 cm-1), whereas it is higher in energy than the
5
5D1 and D0 states of Eu3+ (19 000 and 17 500 cm-1, respec-
tively). As a result, the calculated energy-transfer rate for (Tb)1
as obtained from the quenching experiment has been underes-
timated more than the calculated energy-transfer rate of (Eu)1.
5
the D1 as in the present case.6,7 The energy transfer from the
antenna to the Eu3+ ion is faster than 2 × 107 s-1
.

Photophysical Processes in Polydentate Ln Complexes
J. Phys. Chem. A, Vol. 104, No. 23, 2000 5465
transfer. Our photophysical data, i.e., the triphenylene fluores-
cence quantum yield and the oxygen dependence of the
sensitized emission, show that the energy-transfer takes place
via the triplet state. Therefore, the energy transfer mechanism
is most likely an electron-exchange mechanism. The extremely
small spectral overlap of the antenna phosphorescence spectrum
and the Yb3+ absorption spectrum causes the energy transfer
to be slower than, for example, in the (Eu)1 and (Nd)1
complexes. The recently reported sensitized near-infrared Nd3+
,
Yb3+, and Er3+ luminescence by energy transfer from the dye
fluorescein was shown to be oxygen sensitive.16 Especially, the
sensitized Yb3+ luminescence was very sensitive to oxygen,
indicating not only that the energy-transfer process proceeds
via the triplet state, but also that it is very slow.
Figure 12. Schematic representation of the photophysical processes
leading to sensitized luminescence of (Eu)1 together with the rate
constants of the processes.
Conclusion
The photophysical processes and the rate constants in (Eu)1
are summarized in Figure 12.
This study has shown that the (Ln)1 complexes exist in
solution in two conformational isomers: one in which energy
transfer does not take place, and one in which the energy transfer
takes place, resulting in sensitized luminescence. Using quench-
ing experiments to study this luminescent species, it was found
that the energy-transfer process is fast in the (Eu)1 and (Tb)1
complexes (no oxygen dependence), whereas it is slower in the
near-IR-emitting complexes (oxygen dependence). The energy-
transfer rates should be improved by direct coordination of the
antenna chromophore to the lanthanide ion, which would reduce
not only the distance between the donor and acceptor, but also
the conformational freedom of the antenna relative to the
lanthanide ion. An improvement of the spectral overlap of the
near-IR-emitting lanthanide ions (especially Yb3+) and the
antenna should be achieved by incorporating antenna chro-
mophores with lower lying triplet states. We have shown that
dyes such as fluorescein can sensitize Er3+, Yb3+, and Nd3+
emission,3,16 and we are currently investigating the incorporation
of these dyes in m-terphenyl-based ligands.
b. (Nd)1, (Er)1, and (Yb)1. The energy-transfer rates were
determined in the near-IR-emitting complexes by measuring the
influence of oxygen on the sensitized luminescence. The
luminescence intensity is enhanced by 35% for (Nd)1, 120%
for (Er)1, and 90% for (Yb)1 upon deoxygenation of the DMSO-
d6 solutions (see Figure 10c,d), indicating that, contrary to the
(Eu)1 and (Tb)1 complexes, oxygen quenching is competing
with the energy transfer to the lanthanide ion. The Stern-
Volmer equation can be used to estimate the antenna triplet-
state lifetimes and the energy-transfer rates from this oxygen
dependence. The diffusion-controlled quenching constant kdiff
was taken as 10-10 M-1s-1, and the oxygen concentration in
DMSO was taken as 0.47 mM.23 The results are summarized
in Table 4. The energy transfer to Yb3+ and Er3+ is significantly
slower than the energy transfer to Nd3+
.
Changing the solvent to methanol-d4 would shift the equi-
librium between the dark and luminescent conformational
isomers of the complex toward the luminescent conformational
isomer, as has been shown for (Eu)1 and (Tb)1, and would thus
increase the quantum yield of the energy transfer (φet).
Unfortunately, the luminescence lifetimes of (Nd)1, (Yb)1, and
(Er)1 are significantly lower in methanol-d4: for example, the
lifetimes of (Nd)1 and (Er)1 in methanol-d4 are 0.89 and 1.16
µs, respectively. Thus, changing the solvent from DMSO-d6 to
methanol-d4 will not increase the oVerall luminescence quantum
yield (φse).
Experimental Section
General Synthesis. Melting points were determined with a
Reichert melting point apparatus and are uncorrected. Mass
spectra were recorded with a Finnigan MAT 90 spectrometer
using m-NBA (nitrobenzyl alcohol) as a matrix, unless stated
otherwise. IR spectra were obtained using a Biorad 3200 or a
Nicolet 5SXC FT-IR spectrophotometer. 1H NMR and 13C NMR
spectra were recorded with a Bruker AC 250 spectrometer in
CDCl3 using residual solvent peaks as the internal standard,
unless stated otherwise. Preparative column chromatography
separations were performed on Merck silica gel (particle size
0.040-0.063 mm, 230-400 mesh). CH2Cl2, CHCl3, and hexane
(mixed isomers) were distilled from CaCl2 and stored over
molecular sieves (4 Å). Ethyl acetate was distilled from K2-
CO3 and stored over molecular sieves (4 Å). Triethylamine
(Et3N) was distilled in vacuo and stored over KOH. Acetone
and methanol were of analytical grade and were dried over
molecular sieves prior to use (4 and 3 Å, respectively). All
reactions were carried out under an argon atmosphere. Standard
workup means that the organic layers were finally washed with
water, dried over magnesium sulfate (MgSO4), filtered, and
concentrated to dryness in vacuo.
A special case is Yb3+, since it has only one excited state
(2F5/2), which is approximately 10 000 cm-1 lower in energy
than the antenna triplet state. It has been argued that the spectral
overlap is therefore negligible.49 On the basis of the fact that,
like Eu3+, Yb3+ is relatively easily reduced to Yb2+ 50
, an internal
redox mechanism has been proposed for the sensitized Yb3+
luminescence that takes place through the singlet state of the
antenna chromophore.49 The driving force -∆GRET of such a
redox energy transfer can be estimated with the equation ∆GRET
) E(Triph•+/Triph) - ETriph* - E(Yb3+/Yb2+). However, this
mechanism is not operative in our system. In a cyclic voltam-
metry measurement on a reference Yb3+ complex51 similar to
(Yb)1 in DMSO, no reduction was observed up to -1.90 V vs
SCE. Since the oxidation potential of triphenylene is 1.55 V vs
SCE,52 E(Triph•+/Triph) - E(Yb3+/Yb2+) is >3.45 eV, whereas
the energy of the triphenylene singlet excited state (ETriph*) is
3.6 eV. Therefore, there is almost no driving force for the
photoinduced redox reaction in our system. Moreover, if the
redox energy-transfer mechanism is operative, then the fluo-
rescence of the triphenylene moiety of (Yb)1 must be competing
not only with intersystem crossing, but also with the energy
Triphenylene-2-carboxylic Acid 3. To a solution of tri-
phenylene-2-carboxaldehyde 2 (0.60 g, 2.34 mmol) in a mixture
of CHCl3 (30 mL) and acetone (90 mL) were subsequently
added a solution of H2NSO3H (0.50 g, 5.16 mmol) in water (1
mL) and a solution of NaClO2 (0.50 g, 6.50 mmol) in water (1
mL). The resulting mixture was stirred for 3 h at room

5466 J. Phys. Chem. A, Vol. 104, No. 23, 2000
Klink et al.
temperature, after which it was concentrated in vacuo until the
product precipitated. The solid was filtered off, washed thor-
oughly with water, and dried in vacuo. The product was obtained
Et3N in methanol was added 1.1 equiv of the lanthanide nitrate
salt. The resulting solution was stirred for 2 h, after which the
solvent was concentrated to dryness in vacuo. The complex was
redissolved in CHCl3 and washed twice with water, followed
by standard workup. The complexes were obtained as solids in
quantitative yields. The complexes were characterized by FAB
mass spectrometry and IR spectroscopy. FAB-MS data: [(Eu)1]:
m/z ) 1289.3 [(M + H)+, calcd 1289.4]; [(Tb)1] m/z ) 1295.2
[(M + H)+, calcd 1295.4]; [(Gd)1] m/z ) 1294.1 [(M + H)+,
calcd 1294.4]; [(Er)1] m/z ) 1304.5 (M + H)+, calcd 1304.4];
[(Yb)1] m/z ) 1310.4 [(M + H)+, calcd 1310.4]; [(Nd)1]: m/z
) 1280.3 [(M + H)+, calcd 1280.4]. The complexes all gave
similar IR spectra: a peak at 1635-1630 cm-1 (νNCdO) with a
shoulder around 1600-1590 cm-1 (νCOO).
Photophysical Studies. Steady-state luminescence measure-
ments in the visible region were performed with a Photon
Technology International (PTI) Alphascan spectrofluorimeter,
which has a 75 W quartz-tungsten-halogen lamp as the
excitation source and a Hamamatsu R928 photomultiplier. For
steady-state photoluminescence measurements in the near-IR
region, the excitation light beam was modulated with a
mechanical chopper at 40 Hz. The luminescence signal was
detected with a liquid nitrogen cooled Ge detector, using
standard lock-in techniques. Alternatively, the 351.1/363.8 nm
lines of an Ar ion pump laser at a power of 60 mW were used
for excitation. The laser beam was modulated with an acous-
tooptic modulator at a frequency of 40 Hz.
1
as an off-white solid, yield 80%. Mp: >250 °C. H NMR
(DMSO-d6): δ 9.29 (s, 1H), 8.94-8.74 (m, 5H), 8.19 (d, 1H,
J ) 8.0 Hz), 7.80-7.68 (m, 4H). 13C NMR: 168.1, 133.2, 130.7,
130.2-128.2, 125.5-124.2. IR (KBr): 1684 cm-1 (νCOOH).
Mass spectrum (EI): m/z ) 272.0 [M+, calcd 272.1]. Anal.
Calcd. for C19H12O2: C, 83.81, H, 4.44. Found: C, 83.87; H,
4.46.
Mono(amide) Terphenyl (6). To a solution of bis(amine) 5
(2.0 g, 2.11 mmol) and Et3N (0.58 mL, 4.2 mmol) in CH2Cl2
(200 mL) was slowly added a solution of benzoyl chloride (0.41
g, 2.95 mmol) in CH2Cl2 (50 mL). The resulting solution was
stirred overnight at room temperature. Subsequently, CH2Cl2
was added (100 mL), and the reaction mixture was washed twice
with 5% K2CO3, followed by standard workup. The crude
product was purified by flash column chromatography (MeOH/
CH2Cl2, 1:9) to give mono(amide) 6 as a colorless oil, yield
30%. 1H NMR (CDCl3): δ 7.45-7.20 (m, 5H), 7.20-7.00 (m,
6H), 4.93-4.69 (m, 2H), 4.10-4.00 (m, 6H), 3.80-3.70 (m,
2H), 3.60-2.85 (m, 16H), 2.36 (s, 3H), 2.31 (s, 6H), 2.10-
1.95 (m, 4H), 1.55-1.48 (m, 4H), 1.35 (s, 9H), 1.25 (s, 9H),
1.15 (s, 9H), 0.90-0.75 (m, 6H). 13C NMR (CDCl3): δ 172.3,
167.9, 167.1, 151.9, 151.3, 136.5-126.6, 81.5, 80.9, 70.6, 70.4,
69.6, 68.56, 67.6, 48.4, 46.3, 31.9, 27.9, 21.0, 20.7, 19.3, 13.9.
Mass spectrum (FAB): m/z ) 1053.5 [(M + H)+, calcd for
C62H89N2O12 1053.6].
Time-resolved luminescence measurements in the visible
region were performed with an Edinburgh Analytical Instru-
ments FL900 system. Luminescence lifetime measurements
(Edinburgh Analytical Instruments LP900 system) in the near-
IR-region were performed by monitoring the luminescence
decay after excitation with a 0.5 ns pulse of a LTB MSG 400
nitrogen laser (λexc ) 337 nm, pulse energy 20 µJ, 10 Hz
repetition rate). Decay signals were recorded using a liquid
nitrogen cooled Ge detector with a time resolution of 0.3 µs.
The signals were averaged using a digitizing Tektronix oscil-
loscope. All decay curves were analyzed by deconvolution of
the measured detector response and fitting with monoexponential
functions.
Triphenylene-Functionalized Triester (7). A solution of
triphenylene-2-carboxylic acid 3 (0.10 g, 0.37 mmol) in SOCl2
(10 mL) was refluxed for 4 h. Subsequently, the excess SOCl2
was removed in vacuo. The acid chloride 4 was redissolved in
CH2Cl2 (5 mL) and added to a solution of mono(amide) 6 (0.30
g, 0.29 mmol) and Et3N (0.10 g, 1.0 mmol) in CH2Cl2 (100
mL). The resulting solution was stirred overnight at room
temperature. Subsequently CH2Cl2 was added (100 mL) and
the reaction mixture was washed twice with 1 N HCl, followed
by standard workup. The crude product was purified by flash
column chromatography (ethyl acetate/hexane, 2:3) to give 7
as a white solid, yield 60%. Mp: 61-63 °C. 1H NMR
(CDCl3): δ 8.86-8.60 (m, 6H), 8,29 (d, 1H, J ) 8.0 Hz), 7.80-
6.90 (m, 15H), 5.15 (s, 2H), 5.03 (s, 2H), 4.93 (s, 2H), 4.77 (s,
2H), 4.25-3.04 (m, 18H), 2.40-1.80 (m, 17H), 1.68-0.80 (m,
37H). 13C NMR (CDCl3): δ 172.3, 167.9, 167.1, 151.9, 151.3,
136.5-123.4, 81.5, 80.9, 70.6, 70.4, 69.6, 68.56, 67.6, 48.4,
46.3, 31.9, 27.9, 21.0, 20.7, 19.3, 13.9. Mass spectrum (FAB):
m/z ) 1330.3 [(M + Na)+, calcd 1329.7]. Anal. Calcd. for
C81H98N2O13: C, 74.40; H, 7.55; N, 2.14. Found: C, 74.15; H,
7.48; N, 2.20.
The quantum yield of the triphenylene fluorescence and the
quantum yield of the overall sensitized lanthanide emission were
determined relative to a reference solution of quinine sulfate in
1 M H2SO4 (φ ) 0.546), and correctected for the refractive
index of the solvent.53 The absorbance of the solutions was 0.1
at the excitation wavelength (310 nm). The solvents were of
spectroscopic grade. Deaeration of the samples was performed
by purging the solutions thoroughly with Ar(g) for15-20 min.
Transient Absorption Spectroscopy. The 340 nm pump
light was delivered by a frequency-doubled tunable optical
parametric oscillator system (Coherent Infinity XPO). The width
of the pulse was approximately 2 ns (fwhm), and its energy
was less than 1 mJ/pulse. For recording the transient absorption
spectra at a definite time after the laser pulse, the probe light
from a flashlamp (EG&G FX504, 2.5 µs pulse width) went via
the sample through a spectrograph (Acton SpectraPro 150) to a
gated intensified charge-coupled device camera (Princeton
Instruments ICC-576-G/RB-EM). This system was also used
to obtain kinetic data in the microsecond domain by recording
a set of transient spectra at increasing time intervals. Alterna-
tively, the flashlamp was used in combination with the streak
camera system (Hamamatsu) as the detector to simultaneously
probe the wavelength and time dependence of the transient
signals in the submicrosecond domain. The streak images
Triphenylene-Functionalized Triacid (H3)1. A solution of
triester 7 (0.20 g, 0.18 mmol) in TFA (15 mL) was stirred
overnight at room temperature. Subsequently, toluene (15 mL)
was added, and the TFA/toluene mixture was azeotropically
evaporated. The residue was taken up in CH2Cl2 (100 mL) and
washed twice with 1 N HCl, followed by standard workup. The
triacid (H3)1 was obtained as a white solid in quantitative yield.
Mp: 102-104 °C. 1H NMR (MeOD-d4): δ 8.80-8.10 (m, 7H),
7.70-6.80 (m, 15H), 5.10-4.60 (m, 4H), 4.20-4.90 (m, 18H),
2.35-2.15 (m, 9H), 2.10-0.70 (m, 18H). Mass spectrum
(FAB): m/z ) 1161.4 [(M + Na)+, calcd 1161.5]. Anal. Calcd
for C69H74N2O13‚H2O: C, 71.61; H, 6.62; N, 2.42. Found: C,
71.39; H, 6.50; N, 2.40.
General Procedure for the Preparation of the Complexes.
To a solution of 1.0 equiv of triacid (H3)1 and 4.0 equiv of

Photophysical Processes in Polydentate Ln Complexes
J. Phys. Chem. A, Vol. 104, No. 23, 2000 5467
(17) Heller, A. J. Am. Chem. Soc. 1967, 89, 167.
containing the transient absorbance data were subjected to
principal component analysis by means of singular value
decomposition. All the datasets contained only one significant
component. This component was used in the reconstruction of
the transient absorption spectrum of this single component in
the top of the excitation pulse and its kinetics.
Electrochemical Measurements. Cyclic voltammetry was
performed in a three-electrode cell containing a platinum
working electrode, a platinum counter electrode, and a Ag/AgCl
reference electrode, with an Autolab PGSTAT10 (ECO-
CHEMIE, Utrecht, The Netherlands). The ferrocene/ferricinium
redox couple was used as an internal standard (E° ) 0.44 V vs
SCE in DMSO).54 The solvent DMSO was of spectroscopic
grade and was deoxygenated by being purged with nitrogen.
The ground electrolyte was Bu4NPF6 (0.1 M). The measure-
ments were performed in a glovebox under a nitrogen atmo-
sphere. The cyclic voltammograms were recorded at a scan rate
of 100 mV/s.
(18) Iwamuro, M.; Hasegawa, Y.; Wada, Y.; Murakoshi, K.; Kitamura,
T.; Nakashima, N.; Yamanaka, T.; Yanagida, S. Chem. Lett. 1997, 1067.
(19) Desurvire, E. Phys. Today 1994, 97, 20.
(20) Oshishi, Y.; Kanamori, T.; Kitagawa, T.; Takashashi, S.; Snitzer,
E.; Sigel Jr., G. H. Opt. Lett. 1991, 16, 1747.
(21) An, D.; Yue, Z.; Chen, R. T. Appl. Phys. Lett. 1998, 72, 2806.
(22) Steemers, F. J.; Verboom, W.; Reinhoudt, D. N.; van der Tol, E.
B.; Verhoeven, J. W. J. Am. Chem. Soc. 1995, 117, 9408.
(23) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochem-
istry, 2nd ed.; Marcel Dekker Inc.: New York, 1993.
(24) Klink, S. I.; Hebbink, G. A.; Grave, L.; Peters, F. G. A.; Van
Veggel, F. C. J. M.; Reinhoudt, D. N.; Hofstraat, J. W. Eur. J. Org. Chem.,
in press.
(25) Klink,S. I.; Hebbink, G. A.; Grave, L.; van Veggel, F. C. J. M.;
Reinhoudt, D. N.; Slooff, L. H.; Polman, A.; Hofstraat, J. W. J. Appl. Phys.
1999, 86, 1181.
(26) Jorgenson, W. L.; Briggs, J. M. Mol. Phys. 1988, 63, 547.
(27) van Veggel, F. J. C. M.; Reinhoudt, D. N. Recl. TraV. Chim. Pays-
Bas 1995, 114, 387. The Lennard-Jones parameters for the Ln3+ ions were
taken from: van Veggel, F. C. J. M.; Reinhoudt, D. N. Chem. Eur. J. 1999,
5, 90.
(28) (a) Kirby, A. F.; Foster, D.; Richardson, F. S. Chem. Phys. Lett.
1983, 95, 507. (b) Kirby, A. F. Richardson, F. S. J. Phys. Chem. 1983, 87,
2544.
Acknowledgment. Akzo Nobel Research is gratefully
acknowledged for its financial and technical support. Frank
Steemers (University of Twente) is gratefully acknowledged for
his contribution to the synthesis of the triphenylene antenna.
Lenneke Slooff and Professor Albert Polman (FOM Institute
for Atomic and Molecular Physics, The Netherlands) are
gratefully acknowledged for their support with the near-infrared
luminescence measurements. We express gratitude to Professor
Jan Verhoeven (University of Amsterdam) for his valuable
comments. This research has been financially supported by the
Council for Chemical Sciences of The Netherlands Organization
for Scientific Research (CW-NWO).
(29) 1H NMR spectra of similar complexes in DMSO-d6 show that these
types of complexes have a (time-averaged) Cs symmetry. See also ref 24.
(30) Richardson, F. S. Chem. ReV. 1982, 82, 541.
(31) (a) Kropp, J. L.; Windsor, M. W. J. Chem. Phys. 1963, 39, 2769.
(b) Kropp, J. L.; Windsor, M. W. J. Chem. Phys. 1965, 42, 1599. (c) Kropp,
J. L.; Windsor, M. W. J. Chem. Phys. 1966, 45, 761.
(32) (a) Horrocks, W. D., Jr.; Sudnick, D. R. Acc. Chem. Res. 1981,
14, 384. (b) Holz, R. C.; Chang, C. A.; Horrocks, W. D., Jr. Inorg. Chem.
1991, 30, 3270.
(33) Beeby, A.; Clarkson, J. M.; Dickins, R. S.; Faulkner, S.; Parker,
D.; Royle, L.; de Sousa, A. S.; Williams, J. A. G.; Woods, M. J. Chem.
Soc., Perkin Trans. 2 1999, 493.
(34) Frey, S. T.; Horrocks,W. D., Jr. Inorg. Chim. Acta 1995, 229, 383
and references therein.
(35) (a) Hasegawa, Y.; Kimura, Y.; Murakoshi, K.; Wada, Y.; Kim,
J.-H.; Nakashima, N.; Yamanaka, T.; Yanagida, S. J. Phys. Chem. 1996,
100, 10201. (b) Hasegawa, Y.; Murakoshi, K.; Wada, Y.; Yanagida, S.;
Kim, J.-H.; Nakashima, N.; Yamanaka, T. Chem. Phys. Lett. 1996, 248, 8.
(36) (a) Carnall, W. T.; Goodman, G. L.; Rajnak, K.; Rana, R. S. J.
Chem. Phys. 1989, 90, 3443. (b) Beeby, A.; Dickins, R.; Faulkner, S.; Parker,
D.; Williams, J. A. G. J. Chem. Soc., Chem. Commun. 1997, 1401.
(37) Steemers, F. J.; Verboom, W.; Hofstraat, J. W.; Geurts, F. A. J.;
Reinhoudt, D. N. Tetrahedron Lett. 1998, 39, 7583.
References and Notes
(1) Mukkala, V.-M.; Helenius, M.; Hemmila¨, I.; Kankare, J.; Takalo,
H. HelV. Chim. Acta 1993, 76, 1361.
(2) Slooff, L. H.; Polman, A.; Oude Wolbers, M. P.; van Veggel, F.
C. J. M.; Reinhoudt, D. N.; Hofstraat, J. W. J. Appl. Phys. 1997, 83, 497.
(3) Oude Wolbers, M. P.; van Veggel, F. C. J. M.; Peters, F. G. A.;
van Beelen, E. S. E.; Hofstraat, J. W.; Geurts, F. A. J.; Reinhoudt, D. N.
Chem. Eur. J. 1998, 4, 772.
(4) (a) Gschneider, K. A.; Eyring, L. R. Handbook on the Physics and
Chemistry of Rare Earths; North-Holland Publishing Co.: Amsterdam, 1979.
(b) Sabbatini, N.; Guardigli, M.; Lehn, J.-M. Coord. Chem. ReV. 1993, 123,
201 and references therein.
(5) Sato, S.; Wada, M. Bull. Chem. Soc. Jpn. 1970, 43, 1955.
(6) Crosby, G. A.; Whan, R. E.; Alire, R. M. J. Chem. Phys. 1961, 34,
743.
(7) Tanaka, M.; Yamagughi, G.; Shiokawa, J.; Yamanaka, C. Bull.
Chem. Soc. Jpn. 1970, 43, 549.
(8) Haynes, A. V.; Drickamer, H. G. J. Chem. Phys. 1982, 76, 114.
(9) (a) Tobita, S.; Arakawa, M.; Tanaka, I. J. Phys. Chem. 1984, 88,
2697. (b) Tobita, S.; Arakawa, M.; Tanaka, I. J. Phys. Chem. 1985, 89,
5649.
(38) Crosby, G. A.; Kasha, M. Spectrochim. Acta 1958, 10, 377.
(39) Stein, G.; Wu¨rzberg, E. J. Chem. Phys. 1975, 62, 208.
(40) Weber, M. J. Phys. ReV. 1968, 171, 283.
(41) (a) Sabbatini, N.; Perathoner, S.; Lattanzi, G.; Dellonte, S.; Balzani,
V. J. Phys. Chem. 1987, 91, 6136. (b) Steemers, F. J.; Meuris, H. G.;
Verboom, W.; Reinhoudt, D. N. J. Org. Chem. 1997, 62, 4229.
(42) The reduction potential of Eu3+ is -0.35 V in water (vs NHE).
Bard, A. J.; Parsons, R.; Jordan, J. Standard Potentials in Aqueous Solution;
Marcel Dekker Inc.: New York, 1985.
(43) kdiff is taken as 1010 s-1; [O2] in methanol is 2.1 mM (ref 23). Since
kox ) kdiff[O2] ) 2 × 107 s-1, and there is no oxygen effect, kET > kox,
which means that kET exceeds 107 s-1
.
(44) Bhaumik, M. L.; El-Sayed. J. Chem. Phys. 1965, 42, 787.
(45) Bhattacharyya, S.; Sousa, L. R.; Ghosh, S. Chem. Phys. Lett. 1998,
(10) Dexter, D. L. J. Chem. Phys. 1953, 21, 836.
297, 154.
(11) Sabbatini, N.; Guardigli, M.; Manet, I.; Ungaro, R.; Casnati, A.;
Ziessel, R.; Ulrich, G.; Asfari, Z.; Lehn, J.-M. Pure Appl. Chem. 1995, 67,
135.
(46) φse ) φiscφetφlum ) φiscφetτ/τ0. Typical values of the natural lifetime
τ0 of Eu3+ and Tb3+ are 3 and 4 ms, respectively. If φisc and φet are taken
as 1, then based on the luminescence lifetimes φse is 0.29 for (Eu)1 and
0.44 for (Tb)1.
(12) Parker, D.; Williams, J. A. G. J. Chem. Soc., Dalton Trans. 1996,
3613, and references therein.
(47) It can be demonstrated that a photon-induced electron-transfer
process may indeed be the additional cause of the low luminescence quantum
yield of (Eu)1. The antenna fluoresence lifetime of (Tb)1 was measured to
(13) (a) Oude Wolbers, M. P.; van Veggel, F. C. J. M.; Snellink-Rue¨l,
B. H. M.; Hofstraat, J. W.; Geurts, F. A. J.; Reinhoudt, D. N. J. Am. Chem.
Soc. 1997, 119, 138. (b) Oude Wolbers, M. P.; van Veggel, F. C. J. M.;
Hofstraat, J. W.; Geurts, F. A. J.; Reinhoudt, D. N. J. Chem. Soc., Perkin
Trans. 2 1997, 2275. (c) Oude Wolbers, M. P.; van Veggel, F. C. J. M.;
Snellink-Rue¨l, B. H. M.; Hofstraat, J. W.; Geurts, F. A. J.; Reinhoudt, D.
N. J. Chem. Soc., Perkin Trans. 2 1998, 2141.
(14) Rudkevich, D. M.; Verboom, W.; van der Tol, E. B.; van Staveren,
C.; Kaspersen, F.; Verhoeven, J. W.; Reinhoudt, D. N. J. Chem. Soc., Perkin
Trans. 2 1995, 131.
be approximately 1 ns. Using the equation φflu ) kflu/(kflu + kisc) ) kfluτflu
,
it follows that kflu ) 0.86 × 107 s-1 and kisc ) 9.9 × 108 s-1. It can
furthermore be calculated that φisc . 0.99. If it is assumed that kflu and kisc
are the same for (Eu)1 as for (Tb)1, then the rate of the photon-induced
electron transfer (kPET) can be calculated via φflu ) kflu/(kflu + kisc + kPET),
resulting in kPET ) 5.9 × 108 s-1. As a result, the intersystem crossing
yield of (Eu)1 is lower than that of (Tb)1 with φisc ≈ 0.63.
(48) Van der Tol, E. B. Design and Photophysics of Luminescent Labels
for Use in Time-ResolVed Biological Assays. Ph.D. Thesis, University of
Amsterdam, 1998.
(15) Hemmila¨, I. K. Applications of Fluorescence in Immunoassays;
Wiley and Sons: New York, 1991.
(16) Werts, M. H. V.; Hofstraat, J. W.; Geurts, F. A. J.; Verhoeven, J.
W. Chem. Phys. Lett. 1997, 276, 196.
(49) Horrocks, W. D., Jr.; Bolender, J. P.; Smith, W. D.; Supkowski,
R. M. J. Am. Chem. Soc. 1997, 119, 5972 and references therein.

5468 J. Phys. Chem. A, Vol. 104, No. 23, 2000
Klink et al.
(50) The reduction potential of Yb3+ in water is -1.05 V (vs NHE).
(52) Pysh, E. S.; Yang, N. C. J. Am. Chem. Soc. 1963, 85, 2124.
See ref 42.
(53) Demas, N. J.; Crosby, G. A. J. Phys. Chem. 1971, 75, 991.
(54) Barrette, W. C., Jr.; Johnson, H. W., Jr.; Sawyer, D. T. Anal. Chem.
1984, 56, 1890.
(51) This complex has a phenyl group instead of a triphenylene moiety.
See ref 24.