Unexpected sensitization efficiency of the near-infrared Nd3+, Er3+, and Yb3+ emission by fluorescein compared to eosin and erythrosin

Article abstract of DOI:10.1021/jp0260090

Near-infrared emissive lanthanide complexes were synthesized with covalently attached sensitizers that absorb in the visible. This functionalization was designed such that the sensitizer is in close proximity to the lanthanide ion, which is a prerequisite for efficient energy transfer from the excited sensitizer to the lanthanide ion. The sensitizers used were fluorescein, eosin, and erythrosin, which were linked via a β-alanine spacer to the polydentate chelate. The sensitizers were chosen because they absorb visible light and are structurally very similar, but the intrinsic intersystem crossing quantum yields of the sensitizers vary significantly, because of the presence of the heavy atoms (bromine in eosin and iodine in erythrosin). It was expected that an intrinsic high intersystem crossing would be beneficial in the sensitization process, because energy transfer occurs through the triplet state of sensitizers. However, because of the enhanced intersystem crossing of the sensitizers by the nearby heavy and paramagnetic lanthanide ions, these intrinsic differences were largely diminished. It was even found that fluorescein acts as a more efficient sensitizer for the NIR emission than eosin and erythrosin. The donating triplet state of fluorescein is higher in energy than that of eosin and erythrosin, resulting in less energy back transfer and therefore in a higher efficiency of sensitized emission. This and considerations of selection rules for energy transfer to the lanthanide ions made it possible to distinguish the ⁴F_9/2 level of Nd³⁺ as the main acceptor channel for energy transfer. In the Er³⁺ complexes, the enhancement in intersystem crossing was lower in the eosin and erythrosin complexes than in the fluorescein complex, which was concluded from the remaining complex fluorescence. Furthermore, it is tentatively concluded that additional pathways other than those allowed in Dexter energy transfer play a role in the sensitization of Er³⁺. In the Yb^3- complexes, the higher efficiency of sensitization by fluorescein is due to the enhanced intersystem crossing that is larger in the fluorescein complex than in the eosin or erythrosin complex.

Full text of DOI:10.1021/jp0260090

J. Phys. Chem. A 2003, 107, 2483-2491
2483
Unexpected Sensitization Efficiency of the Near-Infrared Nd3+, Er3+, and Yb3+ Emission by
Fluorescein Compared to Eosin and Erythrosin
Gerald A. Hebbink, Lennart Grave, Le´on A. Woldering, David N. Reinhoudt, and
Frank C. J. M. van Veggel*,†
Laboratories of Supramolecular Chemistry and Technology & MESA+ Research Institute,
UniVersity of Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands
ReceiVed: April 24, 2002; In Final Form: February 4, 2003
Near-infrared emissive lanthanide complexes were synthesized with covalently attached sensitizers that absorb
in the visible. This functionalization was designed such that the sensitizer is in close proximity to the lanthanide
ion, which is a prerequisite for efficient energy transfer from the excited sensitizer to the lanthanide ion. The
sensitizers used were fluorescein, eosin, and erythrosin, which were linked via a â-alanine spacer to the
polydentate chelate. The sensitizers were chosen because they absorb visible light and are structurally very
similar, but the intrinsic intersystem crossing quantum yields of the sensitizers vary significantly, because of
the presence of the heavy atoms (bromine in eosin and iodine in erythrosin). It was expected that an intrinsic
high intersystem crossing would be beneficial in the sensitization process, because energy transfer occurs
through the triplet state of sensitizers. However, because of the enhanced intersystem crossing of the sensitizers
by the nearby heavy and paramagnetic lanthanide ions, these intrinsic differences were largely diminished. It
was even found that fluorescein acts as a more efficient sensitizer for the NIR emission than eosin and
erythrosin. The donating triplet state of fluorescein is higher in energy than that of eosin and erythrosin,
resulting in less energy back transfer and therefore in a higher efficiency of sensitized emission. This and
considerations of selection rules for energy transfer to the lanthanide ions made it possible to distinguish the
4F9/2 level of Nd3+ as the main acceptor channel for energy transfer. In the Er3+ complexes, the enhancement
in intersystem crossing was lower in the eosin and erythrosin complexes than in the fluorescein complex,
which was concluded from the remaining complex fluorescence. Furthermore, it is tentatively concluded that
additional pathways other than those allowed in Dexter energy transfer play a role in the sensitization of
Er3+. In the Yb3+ complexes, the higher efficiency of sensitization by fluorescein is due to the enhanced
intersystem crossing that is larger in the fluorescein complex than in the eosin or erythrosin complex.
Introduction
The luminescence of the lanthanide ions originates from
transitions within the partially filled 4f orbitals, which are in
principle spin-forbidden.15 Lanthanide ions with completely
filled (Lu3+) or unfilled (Y3+ and La3+) 4f orbitals are not lumi-
nescent. The luminescence of the other ions ranges from the
UV to the NIR. Characteristics of this luminescence are the
line-like emissions, low absorption coefficients, and long intrin-
sic luminescent lifetimes, up to milliseconds. The energy levels
of the emission are hardly effected by the environment of the
ions, as the filled 5s and 5p orbitals shield the electrons in the
4f levels; in other words, the effect of the crystal field is very
small.
Sensitized lanthanide ion emission attracts a large interest,1
because it is an efficient way to circumvent the low absorption
coefficients of these ions.2 The main interest of this research
has been focused on the visible (VIS) emission of europium
and terbium,3,4 which can be sensitized with UV absorbing
sensitizers. Only in recent years has the interest shifted to the
near-infrared (NIR) emissive ions, like neodymium, ytterbium,
and erbium.5-11 The use of the NIR emissive ions has a number
of advantages compared to the VIS emitting ions. For instance,
the emission in the NIR is ideally applicable to biology,12
because biological tissue is relative transparent to NIR light,
and to telecommunication where the ions act as active material
in optical amplifiers of the NIR signals.13 Furthermore, sensitiz-
ers that absorb in the visible region can be used here to sensitize
the NIR emitting lanthanide ions.14 This gives the advantages
of cheap excitation sources such as diode lasers and the
possibility to use glass instead of quartz for substrate handling
in biological tests. The sensitization of these ions is still not
fully understood and the search for the most efficient sensitizer-
lanthanide complex combination is still ongoing.
A simplified Jablonski diagram for the sensitized emission
of lanthanide ions is depicted in Figure 1. After excitation of
the antenna and subsequent intersystem crossing, the energy is
transferred to the lanthanide ion. In general, it is accepted that
this sensitization occurs from the triplet state via a Dexter mech-
anism,2 although energy transfer from the singlet state cannot
be ruled out completely.16-18 Taking the sensitization via the
triplet state, the overall quantum yield of sensitized lanthanide
ion emission (φSE) is determined by three individual steps: the
intersystem crossing quantum yield (φISC), the energy transfer
efficiency (ηET), and the lanthanide ion quantum yield (φLn)19,20
* To whom correspondence should be addressed. Phone: +1-250-721-
7184. Fax: +1-250-472-5193. E-mail: fvv@uvic.ca.
† Current address: University of Victoria, Department of Chemistry, PO
Box 3065, Victoria, BC, Canada V8W 3V6.
φSE ) φISCηETφLn
(1)
10.1021/jp0260090 CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/13/2003

2484 J. Phys. Chem. A, Vol. 107, No. 14, 2003
Hebbink et al.
Figure 2. Terphenyl complexes functionalized with fluorescein, eosin,
and erythrosin.
distance between sensitizer and lanthanide ion and by the
normalized spectral overlap between the sensitizer and the ion
(vide infra).34
Figure 1. Top: simplified Jablonski diagram for sensitized emission;
kflu, rate of fluorescence; kISC, intersystem crossing rate; kphos, phos-
phorescence rate; kq, triplet quenching rate; kET, energy transfer rate;
and kLn,rad, radiative decay rate. Bottom: diagram with the energy
levels25 of Yb3+, Nd3+, and Er3+ together with the singlet and triplet
state energies of the three dyes (in methanol, this work51), “flu” stands
for fluorescein, “eo” for eosin, and “ery” for erythrosin.
Here, we report the sensitization properties of NIR emitting
lanthanide ions by a series of dyes based on the xanthene moiety
that are structurally very similar. Complexes were made of a
ligand functionalized with fluorescein, eosin, or erythrosin
(Figure 2). The major difference in these dyes is the intrinsic
intersystem crossing (ISC) quantum yield.35 This difference is
the result of the presence of heavy atoms in the dyes: bromine
in eosin and iodine in erythrosin. This results in ISC quantum
yields in water of 2% for fluorescein, 18% for eosin, and 82%
for erythrosin. Furthermore, efficient energy transfer can be
achieved by reducing the distance between the sensitizer and
the lanthanide ion by exploiting the carboxylate group of these
dyes. Because the complexes used are structurally similar,
conclusions can be drawn on the effect of this inherent ISC on
the sensitization properties. So far, no systematic studies have
been done on the NIR emitting complexes with a series of
related sensitizers. From eq 1, it would be expected that the
higher the intrinsic ISC is, the better the sensitization efficiency
is. However, an external heaVy atom effect of the lanthanide
ions on the intersystem crossing can be operative, thus flawing
this difference.
In the sensitization of europium, it was found that the gap
between the sensitizer triplet state and the energy accepting state
of Eu3+ has to be about 1000-2000 cm-1 to avoid possible
energy back transfer.21,22 This limits the sensitization to UV or
violet light.23,24 When this rule is applied to the NIR emitting
lanthanide ions, it can be seen that there is in principle no
limitation in the use of sensitizers that absorb in the visible.
The lower limit in the triplet state energy, which will still allow
for sensitized emission, will of course depend on the individual
lanthanide ion (see Figure 1 for the energy levels of Yb3+, Nd3+
,
and Er3+).25 Recent work by Horrocks et al. has shown that
Yb3+ might be excited via a photon-induced charge-transfer
mechanism as an alternative pathway.26
A number of NIR emitting lanthanide ion complexes have
been reported that have visible light absorbing antennas.27 Some
examples of these sensitizers that absorb in the visible region
of the electromagnetic spectrum are fluorescein,28,29 lissamine,30
porphyrins,31,32 ferrocene,33 and ruthenium complexes.33 How-
ever, the energy transfer pathways are still not fully understood
for the sensitization of the NIR emitting lanthanide ions by these
sensitizers, although it is expected that they will not differ to a
large extent from those operative in the complexes with Eu3+
This article is structured as follows. After the Experimental
Section, the design of the complexes will be discussed with the
use of molecular modeling simulations, followed by their
synthesis. Then the luminescent properties in the NIR and the
VIS region of these complexes will be presented. Finally, this
luminescence and the energy transfer mechanism from sensitizer
to ion will be discussed and conclusions will be drawn on the
efficiency of sensitization of the three sensitizers.
or Tb3+
.
For efficient lanthanide emission, complexes should have a
high absorption coefficient, a high intersystem crossing quantum
yield, an efficient energy transfer, and an efficient lanthanide
luminescence. The absorption coefficient is not a real problem
as many organic dyes and chromophores absorb strongly.
Ideally, the intersystem crossing quantum yield of a sensitizer
should be 100%. The intersystem crossing is enhanced by the
introduction of heavy atoms such as bromine or iodine in the
dye system. The energy transfer is mainly determined by the
Experimental Section
Synthesis. For a detailed description of the synthesis, see
the Supporting Information to this article.
Luminescence. All measurements, except temperature-de-
pendent measurements, were performed at room temperature
or at 77 K, cooled by liquid nitrogen. The temperature-dependent
measurements were performed in a temperature-controlled cell

Unexpected Sensitization Efficiency
J. Phys. Chem. A, Vol. 107, No. 14, 2003 2485
varying between 0 and 50 °C. Absorption spectra were recorded
on a HP8452A diode array spectrometer. Emission spectra in
the visible and the NIR and decay traces in the visible were
measured on an Edinburgh Analytical Instruments FL/FS 900
fluorimeter, equipped with a 450 W Xe lamp as steady-state
excitation source and nano- and microflashlamps for time-
resolved experiments. The excitation light was fed to a mono-
chromator and focused on a square quartz cuvette (1.00 × 1.00
cm2) containing the sample. The samples were prepared by
dissolving the free ligand, 1R4.H3, in methanol with subsequent
addition of 1 equiv of the appropriate lanthanide ion salt (in
methanol solution) and an excess of triethylamine (about 10
equiv). The emitted visible light was fed to a second mono-
chromator and collected on a Hamamatsu R955 photomultiplier
tube. Near-infrared light was fed to the same monochromator,
but equipped with a 600 lines/mm grating, and was collected
after optical chopping (89 Hz) by an Edinburgh Analytical
Instruments liquid-nitrogen-cooled ultra-sensitive germanium
detector using standard lock-in techniques. All excitation and
emission spectra were corrected for the instrument spectral
response. The spectral resolution was better than 1 nm for the
excitation and emission spectra in the visible region. The spectral
resolution of the emission spectra in the NIR was 8 nm for
spectra of Yb3+ and Nd3+ and 16 nm for the spectra of Er3+
,
Figure 3. Molecular structures of Bz.Ln46 (top) and fluorexon (Fx)37
and the spectral resolution on the excitation side was better than
4 nm. Lifetimes in the nanosecond regime were fitted expo-
nentially with Edinburgh software, implemented on the instru-
ment software, with deconvolution of the decay traces for the
instrument response. This instrument response was measured
on a highly scattering and nonfluorescent sample (Ludox).
Lifetimes in the NIR were measured on an Edinburgh Analytical
Instruments LP900 system, with a pulsed N2 laser as the
excitation source (0.5 ns pulse of a LTB MSG 400 nitrogen
laser, operating at 337.1 nm at 10 Hz, pulse energy: 20 µJ).
The emitted signal was collected with a North Coast liquid-
nitrogen-cooled germanium detector. The decay traces were
averaged by an oscilloscope (Tektronix) and fed to a computer.
Fitting of the lifetimes in the microsecond region was performed
with Edinburgh Analytical Instruments LP900 software, with
deconvolution of the traces using the instrument response
measured with IR140 (Aldrich) in methanol (τ < 1 ns, much
faster than the detector response, which is about 200 ns). Some
typical fits are presented in the Supporting Information (i.e., a
decay trace for Nd3+, Er3+, and Yb3+). Measurements were
performed on solutions with absorption below 0.3 and for
quantum yield determinations absorption below 0.15. For
quantum yield determinations in the visible region, Ru(bpy)3Cl2
(Aldrich) in deoxygenated water (φ ) 0.042) was used as a
standard.36 In the NIR, fluorexon (Molecular Probes, Leiden,
The Netherlands) complexes were used (Fx.Ln, see Figure 3)
as standard.37
(bottom).
Figure 4. Snapshot of 1H4.Eu, derived from a molecular dynamics
simulation in OPLS41 methanol. Hydrogens are omitted for clarity.
was occupied by the solvent, in other cases, this place was
occupied by a sensitizer, creating ternary complexes.39 This
“free” site is used here to bring the xanthene moiety of the
antennas in close proximity of the lanthanide ion via coordina-
tion of its carboxylate group. A â-alanine spacer connected via
an isothiocyanate moiety enables the free carboxylate of the
dye to coordinate to the lanthanide ion in the ligand as was
confirmed by molecular dynamics simulations for 1000 ps (see
Figure 4) in OPLS methanol.40,41 In this snapshot, the xanthene
unit is at the top of the figure and the terphenyl moiety is at the
bottom. The lanthanide ion is complexed by four carboxylates,
two amides, and three ether oxygens leaving no room for any
solvent. The average distance between the lanthanide ion and
the xanthene unit42 was determined from periodically saved sets
of coordinates during the simulations. The distances found were
6.1 ( 0.2 Å for 1H4.Ln and 6.2 ( 0.1 Å for 1I4.Ln, which is
much smaller than the distance of about 10 Å in similar
complexes with a dye moiety attached via a rigid spacer.43 The
Results and Discussion
Modeling. One of the advantages of this terphenyl moiety is
the possibility to functionalize the building block, without
altering the complexing moiety. An example of the structure
of a terphenyl-based complex (Bz.Ln), without a sensitizer, is
depicted in Figure 3. These complexes serve as model or
reference compounds.
We have used the terphenyl moiety previously as building
block for lanthanide complexes, functionalized with sensitiz-
ers.5,30,33,38 In these reports, it was emphasized that these
complexes have eight coordinating groups, leaving room for a
ninth ligand on the lanthanide ion. In most cases, this place

2486 J. Phys. Chem. A, Vol. 107, No. 14, 2003
Hebbink et al.
SCHEME 1: i: (1) C2O2Cl2, Room Temperature; (2) TEA, CH2Cl2, Room Temperature, Yield: 62%; ii: N2H4.H2O,
EtOH, Reflux, Yield: 99%; iii: TEA, CH2Cl2, Dye Isothiocyanate, Room Temperature, Yield: 60%; iv: TFA, Room
Temperature, Yield: Quantitative
larger iodine atoms in 1I4.Ln do not influence the distance
between sensitizer and ion. Furthermore, in 1I4.Ln in about half
of the snapshots, a methanol was found to be in relative close
proximity of the ion, 2.7 ( 0.2 Å, which is the distance between
the lanthanide(III) ion and the methanol oxygen. This distance
is too large for direct binding in the first coordination sphere
of the lanthanide(III) ion but may act as a second sphere
quencher.44,45 The â-alanine spacer thus enables the carboxylate
of the xanthene dyes to coordinate to the lanthanide ion in the
complex. Doing so, the distance between sensitizer and ion is
decreased and no space is left for solvents to coordinate directly
to the ion.
The reaction was quantitative, and the 1H NMR and IR spectra
confirmed complete removal of the phthalimide. The isothio-
cyanate functionalized dyes were coupled to this amine 5 by
stirring overnight a mixture in CH2Cl2 with a slight excess of
dye. Purification was carried out by size separation chroma-
tography (Sephadex), in order to remove the salts (TEA salts)
and the excess of unreacted dye, and by thin-layer chromatog-
raphy. Typical yields of the dye functionalized compounds 6R4
(R ) H, Br, or I) are in the order of 60% of the strongly colored
products (absorption coefficients in the range of 77 000-84 000
L mol-1 cm-1 at the maximum between 500 and 550 nm in
methanol solution). The fluorescein (6H4) and eosin (6Br4)
compounds are highly fluorescent in solution, whereas the
erythrosin compound (6I4) exhibits much less fluorescence. The
tert-butyl ester compounds (6R4) were converted to the triacids
1R4.H3 by mild hydrolyses in pure TFA. Complete hydrolysis
Synthesis. The synthesis of the complexes was carried out
as depicted in Scheme 1 with monoamine 2 as the starting
material,46 and a detailed description of the synthesis can be
found in the Supporting Information.
1
â-Alanine was protected as the phthalimide 3 by heating the
amino acid with phthalic anhydride, yielding 3 in almost
quantitative yield. By refluxing 3 in oxalyl chloride, the acid
function was converted into the acid chloride, which was
subsequently reacted with monoamine 2 to yield the bisamide
4.46 After purification, the formation of 4 was confirmed by
mass spectrometry, where a peak corresponding to 4 was found,
and by the 1H NMR spectrum, where no signals corresponding
to protons next to an amine moiety or a carboxylic acid were
found, but only signals corresponding to the amides. Further-
more, four strong peaks are found in the IR spectrum in the
carbonyl region around 1700 and 1753 cm-1 for the esters, 1644
cm-1 for the amides, and 1717 cm-1 for the phthalimide. The
deprotection of the phthalimide moiety in 4 was carried out in
ethanol with hydrazine monohydrate under reflux conditions.
was confirmed by H NMR spectroscopy and by mass spec-
trometry. In the IR spectra, the resonance around 1730 cm-1
for the carboxylic acid is present. The purity of the compounds
was checked with TLC with 15% methanol in CH2Cl2 as the
eluent. Single spots were found with an rF of about 0.4. The
lanthanide(III) ion complexes 1R4.Ln were made in methanol
by deprotonation of the ligand with 10 equiv TEA and
subsequent addition of 1 equiv of lanthanide(III) nitrate salt.
Complete deprotonation of the complexes, which consist of three
carboxylate moieties on the terphenyl, the carboxylate on the
xanthene moiety, and the phenol group on the xanthene moiety,
was confirmed by UV spectroscopy, where the absorption of
the dianion form of the dyes was found.35 Complexes were made
of the nitrate salts of Nd3+, Yb3+, and Er3+, which exhibit
sensitized NIR emission, and of Eu3+ and Gd3+, which were

Unexpected Sensitization Efficiency
J. Phys. Chem. A, Vol. 107, No. 14, 2003 2487
Figure 5. NIR Emission spectra of 1H4.Nd, 1H4.Yb, and 1H4.Er in
CH3OD (Nd3+ and Yb3+) and in CD3OD (Er3+) solution.
used in addition to determine the deactivation of the singlet
excited state of the dyes. The mass spectrometry characterization
(ESI-TOF) data of the complexes are summarized in Table 1
of the Supporting Information. The molecular mass and the
isotope pattern of the masses found correspond excellently with
the calculated spectra.
Figure 6. (Normalized) absorption (top) and excitation spectra (bottom)
of 1H4.Nd, 1Br4.Nd, and 1I4.Nd in CD3OD solution. The excitation
spectra were collected by measuring the emission at 1064 nm.
Luminescence. In this section, the luminescence of the
sensitizer-functionalized complexes will be described. This is
done according to the simplified Jablonski diagram in Figure 1
starting with the NIR luminescence of the lanthanide ions and
then the fluorescence and the phosphorescence in the visible
region of the sensitizer. Finally, the energy transfer efficiency
and the energy transfer rates will be discussed. The reported
luminescent properties are the emission and excitation spectra,
quantum yields, lifetimes, and the rates of the various transitions.
The luminescence properties of Bz.Ln (Figure 3)46 are reported
in Table 2, to have a comparison. The emission in the visible
region is also reported of the Gd3+ and Eu3+ complexes. These
complexes (1R4.Gd and 1R4.Eu) are a good model for the
sensitizer fluorescence and phosphorescence as these ions have
similar sizes as the emissive ions, and they are paramagnetic
as well. However, they cannot accept energy, because the lowest
excited states are too high in energy to be populated by the
dyes (32 200 cm-1 for Gd3+ and 17 300 cm-1 for Eu3+).
NIR Luminescence. Spectra of the Nd3+, Yb3+, and Er3+
TABLE 1: NIR Emission Sensitization Properties 1R4.Ln
(with R ) H, Br, and I and Ln3+ ) Yb3+, Nd3+, and Er3+ a
)
τ
τ
(µs)d (µs)e
φLn
φISCηET
b
c
f
g
complex φSE, rel
φSE
1H4.Yb
0.52 2.3 × 10-3 11.1
9.8 5.5 × 10-3
0.42
0.25
0.29
0.19
0.09
0.08
1Br4.Yb 0.32 1.4 × 10-3 11.2 11.6 5.6 × 10-3
1I4.Yb
1H4.Nd
0.37 1.7 × 10-3 11.5 10.2 5.8 × 10-3
0.78 3.0 × 10-4 0.41 0.24 1.6 × 10-3
1Br4.Nd 0.37 1.4 × 10-4 0.41 0.36 1.6 × 10-3
1I4.Nd
1H4.Er
0.31 1.2 × 10-4 0.37 0.25 1.5 × 10-3
1.00h
0.91 nd
6.5 × 10-5
1Br4.Er 0.51h
ndi
nd
nd
1I4.Er
0.43h
ndi
a Properties determined in deuterated methanol, unless stated oth-
erwise. b Relative quantum yield, with φFx.Ln )1.0; error (from triplos),
(5%. c Absolute quantum yield, calculated from absolute quantum yield
of fx.Ln: φSE ) φSE,relφfx.Ln, φfx.Yb ) 4.5 × 10-3; φFx.Nd) 3.8 × 10-4
;
estimated error, (5%.37 d Luminescent lifetime of the lanthanide(III)
ions in methanol-d1; error, (5%. e Luminescent lifetime of the lan-
f
thanide(III) ions in methanol; error, (5%. Calculated from φLn) τ/τrad
,
Complexes. In the emission spectra (Figure 5), the typical
where τrad is the pure radiative lifetime of Ln (τrad,Nd ) 0.25 ms, τrad,Yb
4
emission lines for Nd3+ at 890 (4F3/2 f I9/2), 1064 (4F3/2
f
) 2.0 ms, τrad,Er ) 14 ms).6,13,49 g Efficiency of the sensitization
4I11/2), and 1330 nm (4F3/2 f 4I13/2),47 for Er3+ at 1536 nm (4I13/2
process: (φISCη
ET) ) φSE/φLn; error, (10%. h Relative quantum yield,
f I15/2), and for Yb3+ at 980 nm (2F5/2 f I7/2) are present.
The emission of the 1H4.Ln complexes is shown here as
representative for the emission of the other complexes (1Br4.Ln
and 1I4.Ln). The absorption and excitation spectra of the Nd3+
complexes are shown in Figure 6 as representative for all
absorption and excitation spectra. The absorption maxima of
these dyes are 505 nm for fluorescein, 535 nm for eosin, and
545 nm for erythrosin (measured in methanol solution), which
are red shifted by about 5-10 nm compared to the absorption
maxima of these dyes in water.29 The resemblance of the
excitation spectra to the absorption spectra proofs the sensitized
emission by the dyes.
4
2
φrel(1H4.Er), is set to 1, no accurate standard available. i nd: not
determined. No accurate decay trace could be obtained.
The luminescence 1R4.Er was only measured relative to 1H4.Er
because no comparison with Fx.Er standard was available. The
relative quantum yields (φSE, rel) show that for all three lanthanide
ions the quantum yields are somewhat lower than that for Fx.Ln.
A striking observation is that the complexes with fluorescein
as the sensitizer have the highest quantum yields. At first glance
this is in contrast with the fact that fluorescein has the lowest
intrinsic intersystem crossing quantum yield, and the reason for
this is presented in the discussion section. Despite the deuterated
solvent, the φSE’s are rather low, which is mainly caused by
vibrational modes within the complex itself. These quench the
excited state of the lanthanide(III) ions, and this is commonly
observed in organic complexes of these NIR emitting ions. In
OVerall Quantum Yield (φSE). Quantum yields of sensitized
emission in deuterated methanol (φSE) were determined by using
Fx.Nd and Fx.Yb (Figure 3)37 as a standard for the Nd3+ and
Yb3+ emission, respectively, and they are tabulated in Table 1.

2488 J. Phys. Chem. A, Vol. 107, No. 14, 2003
Hebbink et al.
TABLE 2: Lanthanide Luminescent Lifetimes of Bz.Ln
(Figure 3),46 in µs; Error, (5 %
a
Ln3+
CH3OH
CD3OD
kq,solv
Yb3+
Nd3+
Er3+
2.47
0.27
nd
10.9
0.73
1.53
3.1 × 105 s-1
2.3 × 106 s-1
-
a Determined according to kq,
) 1/τCD3OD - 1/τCH3OH.
solv
fact, the highest quantum yield reported for Nd3+ in organic
solution is only 0.03, measured in a completely fluorinated
ligand and in a perdeuterated solvent.7
Quantum Yield of NIR Lanthanide(III) Emission (φLn). The
quantum yields of lanthanide emission were calculated from
the ratio of the measured lifetimes and the radiative lifetimes
of the ions. Luminescent lifetimes of the lanthanide(III) ion
complexes in the NIR were determined in normal and deuterated
methanol and these are also reported in Table 1. For comparison,
the luminescent lifetimes of Bz.Ln (Ln3+ ) Yb3+, Nd3+, and
Er3+, Figure 3) are reported in Table 2.46 The lifetimes of the
ytterbium complexes 1H4.Yb, 1Br4.Yb, and 1I4.Yb are about
11 µs in both deuterated methanol and normal methanol. A rate
constant of quenching by the solvent of these complexes (kq,
solv ) 1/τCH3OH - 1/τCD3OD) is e3 × 104 s-1, whereas the
quenching by the solvent in Bz.Yb is in the order of 3 × 105
s-1. This means that the ligand and coordinating sensitizer shield
the ion very well from the solvent, which is in agreement with
the molecular modeling studies performed on these complexes.
The lifetimes of 1R4.Nd are comparable to each other, about
0.4 µs in deuterated methanol and about 0.3 µs in methanol.
From this weak solvent effect, it is concluded that no solvent
is coordinating to the ion. The quenching rate by the solvent
for 1R4.Nd is e1 × 106 s-1, caused by the quenching of solvent
vibrations in the second coordination sphere. In comparison,
the quenching of Bz.Nd, with first sphere solvent molecules, is
about 2.3 × 106 s-1. The gap between the lowest excited state
and the highest ground state of Yb3+ (∼10 200 cm-1) is much
larger than the gap in Nd3+ (∼5000 cm-1), which makes Yb3+
less sensitive to high energy vibrations of the organic environ-
Figure 7. Excitation, fluorescence (in CH3OH), and phosphorescence
spectra of 1R4.Gd. The phosphorescence was measured at 77 K in
MeOH/EtOH (1:4) glass with a delay between excitation and emission
to distinguish from fluorescence.
for energy transfer could not be calculated for the Er3+
complexes, it is clear from the relative quantum yields (φSE,rel
)
that also here fluorescein is a better sensitizer for erbium
luminescence than the other two sensitizers. However, the effect
of the more efficient sensitization by fluorescein is the strongest
in the case of the neodymium acceptor.
VIS Luminescence. Fluorescence and Phosphorescence
Spectra. The excitation, fluorescence, and phosphorescence
spectra of 1H4.Gd, 1Br4.Gd, and 1I4.Gd are depicted in Figure
7. The excitation spectra of the three dyes are similar to the
absorption spectra (shown in Figure 6). The maxima of the
fluorescence are around 510 nm (1H4.Gd), 540 nm (1Br4.Gd),
and 550 nm (1Br4.Gd), red shifted by about 10-15 nm (200-
400 cm-1) compared to the spectra in water.29,51 The fluores-
cence spectra of all of the other complexes 1R4.Ln and of 6R4
have the same shape as the 1R4.Gd3+ spectra, but they vary in
intensities. This shows that the ions do not effect the transitions
energetically. Fluorescence lifetimes were determined and are
4.3 ns for 6H4, 2.5 ns for 6Br4, and 0.4 ns for 6I4. The
fluorescent lifetimes of the complexes 1R4.Ln were all shorter
than the detector response: shorter than about 0.3 ns. Applying
the approximation that this is completely due to the enhancement
of the intersystem crossing, it follows that kISC ≈ 1/τflu,complex
-1/τflu,freedye, thus kISC g 3 × 109 s-1. The phosphorescence of
1R4.Gd3+ was detected at 77 K in an ethanol-methanol glass,
with a delay between excitation and emission.52 The peak of
the phosphorescence is located at 640 nm for 1H4.Gd and at
660 for 1Br4.Gd and 1I4.Gd, which is also red shifted by about
10-15 nm compared to the phosphorescence of the dyes in
water. The lifetimes of the phosphorescence of the 1R4.Gd
complexes at 77 K were determined to be 1.6 ms for 1H4.Gd,
2.0 ms for 1Br4.Gd, and 1.5 ms for 1I4.Gd. Of all the other
ment than Nd3+ 8,48
Furthermore, second-sphere coordination
.
solvents can be more important in the Nd3+ complexes, because
Nd3+ is a larger ion than Yb3+ allowing more space for second
sphere quenchers. The lifetime of 1H4.Er in deuterated methanol
is about 0.9 µs, which is comparable with the Er3+ lifetimes in
other reports.8,29,46,38 The lifetimes of the other complexes could
not be determined, because of the low intensity of the collected
traces. From these measured lifetimes and the radiative lifetimes
of the ions, the quantum yields of lanthanide emission were
determined. The radiative lifetimes were taken from the
literature,6,13 where they were calculated to be 2000 µs for Yb3+
and 14 000 µs for Er3+. The radiative lifetime for Nd3+ was
estimated to be 250 µs.49 From this, it follows that Yb3+ has
the largest intrinsic quantum yield (5.5 × 10-3), then Nd3+ (1.6
× 10-3), and Er3+ the lowest (6.5 × 10-5). This significantly
lower quantum yield of Er3+ was observed previously in organic
complexes and is caused by the efficient quenching by C-H
vibrations of the long-lived excited state of erbium50 and
probably also by inefficient sensitization (vide infra).
complexes, phosphorescence was only observed for 1R4.Eu3+
.
The Gd3+ and the Eu3+ complexes exhibit phosphorescence,
because there are no energy levels of these ions that can accept
energy from the triplet states of the dyes. The absence of
phosphorescence in the other complexes is due to energy transfer
from the triplet states to the Nd3+, Yb3+, and Er3+ ions. At room
temperature, no phosphorescence of any complex could be
detected. The decrease in 1R4.Gd fluorescence is accompanied
by an increase in 1R4.Gd phosphorescence, i.e., because of an
external heavy atom effect. To get an estimate of the triplet
state depopulation rate ferrocene, a triplet quencher was added
to the solutions. Upon addition of ferrocene the NIR lumines-
cence was quenched indeed, and a Stern-Volmer plot (data
The efficiency of sensitization (as the product of intersystem
crossing and energy transfer) is calculated from φSE/φLn (Table
1). The efficiency of Yb3+ sensitization is higher than the
efficiency of Nd3+ sensitization. Furthermore, fluorescein is a
more efficient sensitizer for Nd3+ and for Yb3+ luminescence
than eosin and erythrosin, despite the lower intrinsic intersystem
crossing quantum yield of fluorescein. Although the efficiencies

Unexpected Sensitization Efficiency
J. Phys. Chem. A, Vol. 107, No. 14, 2003 2489
TABLE 3: Fluorescence Quantum Yields of the Lanthanide
Complexes (in %); Error, (5 %a
R
6R4
1R4.Gd
1R4.Yb
1R4.Nd
1R4.Er
1R4.Eu
H
Br
I
83
76
8.4
8.8
33
6.3
6.8
30
7.4
0.3
2.6
1.2
2.6
20
5.6
1.4
23
4.3
a The fluorescence quantum yields were determined in methanol with
Ru(bpy)3 in deoxygenated water as a standard.
not shown) revealed a lower estimate of the energy transfer rate
of about 108 s-1
.
Quantum Yield of Dye Fluorescence. Fluorescence quantum
yields of the sensitizers were determined for the NIR emitting
lanthanide complexes and 6R4, and they are presented in Table
3. These tert-butyl protected compounds were used as repre-
sentative of the inherent dye fluorescence. The fluorescence of
fluorescein is reduced from about 85% in 6H4 to 1-5% in the
complexes. The same behavior is found for the eosin (6Br4)
and erythrosin (6I4) compounds, albeit to a smaller extent. The
decreases in fluorescence upon complexation are from 76% in
6Br4 to 2.6% in 1Br4.Nd, 30% in 1Br4.Yb, and 20% in 1Br4.Er
and from 8.4% in 6I4 to 7.4% in 1I4.Yb, 1.2% in 1I4.Nd, and
5.6% in 1I4.Er. The decrease in the fluorescence of the dyes is
due to the enhancement in intersystem crossing, caused by the
nearby paramagnetic lanthanide ions.5,30,53 This conclusion is
further corroborated by the fact that the dye fluorescence is
reduced to the same extent in the Gd3+ and the Eu3+ complex
and because of the presence of phosphorescence in these
complexes. The Gd3+ and Eu3+ ions cannot accept energy from
both the singlet and triplet excited states of the dyes, because
the lowest excited states of the lanthanide ions are too high in
energy. The small differences in the fluorescence quantum yields
with varying lanthanide ion can be explained by variation in
atomic weight and magnetic moment of the ions.53,54 Although
the fluorescence in the 1R4.Nd complexes is quenched almost
completely, the 1H4.Yb and 1H4.Er complexes are quenched
stronger than 1Br4.Yb, 1Br4.Er, 1I4.Yb, and 1I4.Er. This
difference is of importance in the sensitization efficiency as
discussed below.
Figure 8. Plot of ln(I0/I) versus (1/T) of the NIR luminescence of
1H4.Nd and 1Br4.Nd, where I is the integrated emission intensity of
the F3/2 f I11/2 transition around 1066 nm and excitation at 505 nm
(1H4.Nd) or 530 nm (1Br4.Nd) in CD3OD.
4
4
Figure 9. Phosphorescence spectra of 1R4.Gd at 77 K in a MeOH/
EtOH (1:4) glass and part of the absorption spectrum of a 0.2 M
Nd(NO3)3 solution in MeOH.
Energy Transfer. The acceptor levels of the lanthanide ions
are important in determining the energy transfer mechanisms
from dye to lanthanide ion. A choice of potential accepting levels
can be made based on the energy difference between donor
(sensitizer) and acceptor level (Ln3+), which should not be too
large for sufficient spectral overlap,22 and on the selection rules
for energy transfer. Malta et al. performed theoretical analysis
of Eu3+ and Sm3+ luminescence.2,55 The selection rules for
energy transfer from sensitizer to the Ln3+ (2S+1)ΓJ 4f levels are
|∆J| ) 2, 4, 6 for the dipolar and multipolar energy transfer
(Fo¨rster type) and |∆J| ) 0, 1 (but forbidden for J ) J′ ) 0)
for the electron exchange mechanism (Dexter type). It should
be noted that these rules are (somewhat) relaxed by mixing or
“stealing” from allowed transitions (e.g., 4f-5d transitions).56
In this work, energy transfer takes place from the triplet excited
state, because the singlet state of the dyes are depopulated very
fast to the triplet state and the Dexter type of transfer is in
general the dominant mechanism.2,55
TABLE 4: Spectral Overlap Integral J of Fluorescein,
4
Eosin, and Erythrosin Phosphorescence with the F9/2 and
4F7/2 Nd3+ Absorptions
4F9/2
4F7/2
4F9/2/F7/2
fluorescein
eosin
erythrosin
0.47 × 10-2
0.97 × 10-2
1.0 × 10-2
0.08 × 10-2
0.17 × 10-2
0.15 × 10-2
5.9
5.7
6.7
(J) of the dye phosphorescence and the absorption of Nd3+ (see
Figure 9) of the 4F7/2 and the 4F9/2 levels is calculated according
to J ) ∫ꢀ(λ) F(λ) dλ,34 with the absorption of each accepting
transition (∫ꢀ(λ) dλ) and the phosphorescence (∫F(λ) dλ)
normalized to 1. The results of these calculations are sum-
marized in Table 4. These calculations show that the spectral
4
overlap of the dyes is about six times higher with the F9/2
4
absorption than the spectral overlap with the F7/2 absorption.
From this, it is concluded that the 4F9/2-energy level of Nd3+ is
the main channel for energy transfer from these three dyes.58 It
4
Energy Transfer to Nd3+. Applying these considerations to
fluorescein, eosin, and erythrosin and Nd3+, the levels that are
should be noted that for both transitions (4F9/2 and F7/2) the
spectral overlap with fluorescein is a factor 2 smaller than with
eosin and erythrosin. An increase of the temperature of the
solutions gave a substantial decrease in the luminescence
intensities of Nd3+. The reason for this could be an energy-
back transfer mechanism, which is supported by the fact of the
relative small energy-gaps between the triplet states and the
4
expected to accept energy are the F7/2 (at 13 400 cm-1) and
4F9/2 (at 15 000 cm-1) levels for all three dyes (Figure 1).57 The
unlikeness of the 4F7/2 level as acceptor level can be illustrated
with a calculation of the spectral overlap of the Nd3+ absorption
and the sensitizers phosphorescence spectra. The spectral overlap

2490 J. Phys. Chem. A, Vol. 107, No. 14, 2003
Hebbink et al.
to the other dyes is somewhat smaller than in the Nd3+ and
Er3+ complexes. The fluorescence of eosin and erythrosin is
less quenched than the fluorescence of fluorescein in 1R4.Yb
(Table 3). The receiving energy level of Yb3+ can only be the
2F5/2 at 10 200 cm-1, as this is the only excited state of this
ion. A Dexter type mechanism to this level is allowed by the
selection rules, because |∆J| ) 1. The energy gap between the
triplet states of the sensitizers and this receiving level is about
4000 cm-1, a gap that is not too large for energy transfer but is
too large for energy back transfer. A photon-induced electron
transfer26 could be a pathway in the sensitization process,
because Yb3+ is easily reduced to Yb2+. However, this is not
likely here because there is no different quenching effect on
the fluorescence of the dyes in the Yb3+ complexes compared
to other complexes (see above under fluorescence quantum
yields and Table 3). Furthermore, the calculated driving force
for charge transfer, according to Rehm and Weller,61 is very
small: -0.18 eV for transfer from the singlet excited state of
fluorescein.62 An additional argument is that in the Eu3+
Figure 10. Jablonski diagram of the sensitization of Nd3+ by
fluorescein (1H4.Nd), with the rate constants and efficiencies of the
various transitions.
accepting level of Nd3+ (700 cm-1 for fluorescein and 400 cm-1
for eosin and erythrosin). The results of variations in the
temperature on the luminescence intensities are plotted in Figure
8 for 1H4.Nd and 1Br4.Nd. The luminescence of the 1H4.Nd
complex decreased about 40% going from 0 to 50° C. The
1Br4.Nd also showed a decrease in intensity as well, about 20%
going from 15 to 45° C. The slope in a plot of ln(I0/I) versus
1/T is the energy gap (∆E/k) between the donor state (from
Nd3+) and the acceptor state (the triplet state of the dyes)
because I0/I ∼ kBT ) Ae(-∆E/kT) where I0/I is the reciprocal of
the normalized luminescence intensity, kBT is the back transfer
rate, A is a constant, ∆E is the energy gap (in J), k is the
Bolzmann constant, and T is the absolute temperature.59 The
slopes (Figure 8) correspond to an energy gap of 700 cm-1 in
the 1H4.Nd complex and 400 cm-1 in the 1Br4.Nd complex.
This difference between the two complexes (i.e., 300 cm-1) is
about the same as the difference in the position of the
phosphorescence maxima, which shows that the main accepting
level of Nd3+ is the same for both complexes.60 This is
corroborated by the calculation of the spectral overlap.
complex, which is reduced more easily than Yb3+ 63
, phospho-
rescence is observed. This leads to the conclusion that in the
current work the sensitization of Yb3+ occurs by a Dexter
mechanism from the triplet excited state of the dyes.
Conclusions
NIR emitting lanthanide complexes were designed and
synthesized that bring the sensitizer in close proximity to the
ion and that shield the ion completely from the solvent. The
sensitizers coupled to lanthanide complexes are fluorescein,
eosin, and erythrosin, which enable visible light (green)
sensitization of the NIR lanthanide emission. The fluorescence
of the sensitizers was strongly reduced, because of the enhance-
ment in intersystem crossing by the nearby paramagnetic
lanthanide ions. The sensitization of the NIR emitting ions is
solely determined by the energy transfer step, because of the
enhanced intersystem crossing quantum yields. However, the
overall luminescence is in all cases largely determined by the
low intrinsic quantum yield of the lanthanide ions. Surprisingly,
the efficiency of sensitization was found the highest for the
fluorescein complexes 1H4.Ln. In the 1R4.Nd complexes, it was
found that the main channel for energy transfer is the 4F9/2 Nd3+
level. Fluorescein is a more efficient sensitizers than eosin or
erythrosin, because of the larger gap between the triplet of
fluorescein with this receiving Nd3+ level which leads to less
energy back transfer. In the Yb3+ complexes, the more efficient
excitation is most likely due to the less efficient enhancement
of the intersystem crossing of eosin and erythrosin. The reason
of the more efficient sensitization by fluorescein in the Er3+
complexes is that an additional level of Er3+ is probably capable
of accepting energy from fluorescein but not from eosin or
erythrosin. As long as these sensitizers, and especially fluores-
cein, are in close proximity (<6 Å) to the energy accepting
ions, the energy transfer is determined by the energy levels
involved.
All rate constants and efficiencies of the sensitization of Nd3+
by fluorescein (in the 1H4.Nd complex) are summarized in the
Jablonski diagram in Figure 10. After excitation, the singlet level
is depopulated very fast to the triplet state. From this triplet
4
state, the energy is transferred to the F9/2 level of Nd3+. This
level deactivates nonradiative to the emissive 4F3/2 level (internal
conversion) or by energy-back transfer to the fluorescein triplet
4
state. From the F3/2 level, the neodymium luminescence is
observed at 890, 1066, and 1330 nm.
Er3+ and Yb3+ Sensitization. Applying the same selection
rules to the sensitization of Er3+, only one level can be assigned
that allows for excitation via an exchange mechanism, i.e., the
4I13/2 at 6500 cm-1. This implies that the energy gap between
the triplets of the sensitizers and the accepting level is very
large, which would result in a very small kET. Additional
pathways, because of relaxation of the selection rules, may be
the reason for sensitization of this ion. An additional pathway
in the case of fluorescein sensitization may be one of the reasons
of the more efficient sensitization of Er3+ by fluorescein. The
4F9/2 level of Er3+ (at 15 500 cm-1) could be an additional
pathway in the fluorescein complex. Moreover, the decrease in
fluorescence of fluorescein is stronger than in the eosin and
erythrosin fluorescence deactivation (Table 3), which is a second
reason for the higher sensitization efficiency by fluorescein.
For Yb3+, fluorescein is found to be a more efficient sensitizer
as well. The difference in sensitization of fluorescein compared
Acknowledgment. This research is supported by the Tech-
nology Foundation STW, applied science division of NWO and
the technology program of the Ministry of Economic Affairs.
Akzo Nobel is gratefully acknowledged for technical support.
Hans Hofstraat, Philips research and University of Amsterdam,
is acknowledged for providing help in the lifetime measurements
in the near-infrared.
Supporting Information Available: A detailed description
of the synthesis and the mass spectrometry characterization (ESI-

Unexpected Sensitization Efficiency
J. Phys. Chem. A, Vol. 107, No. 14, 2003 2491
(35) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of photochem-
istry, 2nd ed.; Marcel Dekker: New York, 1993.
(36) van Houten, J.; Watts R. J. J. Am. Chem. Soc. 1976, 98, 4853.
(37) Werts, M. H. V.; Verhoeven, J. W.; Hofstraat, J. W. J. Chem. Soc.,
Perkin Trans. 2 2000, 433.
TOF) data of the complexes (Table 1) are presented. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(38) Hebbink, G. A.; Reinhoudt, D. N.; van Veggel, F. C. J. M. Eur. J.
Org. Chem. 2001, 4101.
(39) Klink, S. I.; Hebbink, G. A.; Oude Alink, P. G. B.; Grave, L.; van
Veggel, F. C. J. M.; Werts, M. H. V. J. Phys. Chem. A 2002, 106, 3681.
(b) Hebbink, G. A.; Klink, S. I.; Oude Alink, P. G. B.; van Veggel, F. C.
J. M. Inorg. Chim. Acta 2001, 317, 114.
(40) For the procedure, see: van Veggel, F. C. J. M.; Oude Wolbers,
M. P.; Reinhoudt, D. N. J. Phys. Chem. A 1998, 102, 3060. For the Lennard
Jones parameters for the lanthanide ions, see: van Veggel, F. C. J. M.;
Reinhoudt, D. N. Chem. Eur. J. 1999, 5, 90.
(41) Jorgensen, W. L. J. Phys. Chem. 1986, 90, 1276 (OPLS stands for
optimized potentials for liquid simulations).
(42) The central carbon atom of the xanthene unit, at the para position
to the oxygen in the central ring, was taken as a measure of the distance.
(43) Slooff, L. H.; Polman, A.; Klink, S. I.; Hebbink, G. A.; Grave, L.;
van Veggel, F. C. J. M.; Reinhoudt, D. N.; Hofstraat, J. W. Opt. Mater.
2000, 14, 101. Slooff, L. H.; Polman, A.; Hebbink, G. A.; van Veggel, F.
C. J. M.; Reinhoudt, D. N.; Cacialli, F.; Friend, R. H. Appl. Phys. Lett.
2001, 78, 2122.
(44) Horrocks, W. DeW.; Sudnick, D. R. Acc. Chem. Res. 1981, 14,
384.
(45) Beeby, A.; Clarkson, I. 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.
References and Notes
(1) For a review, see: Parker, D.; Williams, J. A. G. J. Chem. Soc.,
Dalton Trans. 1996, 3613.
(2) De Sa, G. F.; Malta, O. L.; de Mello Donega, C.; Simas, A. M.;
Longo, R. L.; Santa-Cruz, P. A.; da Silva, E. F., Jr. Coord. Chem. ReV.
2000, 196, 165.
(3) Steemers, F. J.; Verboom, W.; Reinhoudt, D. N.; van der Tol, E.
B.; Verhoeven, J. W. J. Am. Chem. Soc. 1995, 117, 9408.
(4) Hemmila¨, I. K. Applications of Fluorescence in Immunoassays;
Wiley and Sons: New York, 1991.
(5) Klink, S. I.; Grave, L.; Reinhoudt, D. N.; van Veggel, F. C. J. M.;
Werts, M. H. V.; Geurts, F. A. J.; Hofstraat, J. W. J. Phys. Chem. A 2000,
104, 5457.
(6) 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.
(7) Hasegawa, Y.; Ohkubo, T.; Sogabe, K.; Kawamura, Y.; Wada, Y.;
Nakashima, N.; Yanagida, S. Angew. Chem., Int. Ed. 2000, 39, 357.
(8) Beeby, A.; Faulkner, S. Chem. Phys. Lett. 1997, 266, 116.
(9) Maupin, C. L.; Dickins, R. S.; Govenlock, L. G.; Mathieu, C. E.;
Parker, D.; Williams, J. A. G.; Riehl, J. P. J. Phys. Chem. A 2000, 104,
6709.
(10) Voloshin, A. I.; Shavaleev, N. M.; Kazakov, V. P. J. Lumin. 2001,
93, 115.
(11) Slooff, L. H.; Polman, A.; Oude Wolbers, M. P.; van Veggel, F.
C. J. M.; Reinhoudt, D. N.; Hofstraat, J. W. J. Appl. Phys. 1998, 83,
497.
(46) 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. 2000,
192.
(47) The low-energy side of the 4F3/2 f 4I13/2 transition of Nd3+ at 1330
nm is distorted by absorption around 1400 nm, which is the first overtone
of the 2950 cm-1 C-H vibration of the solvent (CH3OD).
(48) Ermolaev, V. L.; Sveshnikova, E. B. Russ. Chem. ReV. 1994, 63,
905.
(12) Werts, M. H. V.; Woudenberg, R. H.; Emmerink, P. G.; van Gassel,
R.; Hofstraat, J. W.; Verhoeven, J. W. Angew. Chem., Int. Ed. 2000, 39,
4542.
(13) Miniscalco, W. J. J. LightwaVe Technol. 1991, 9, 234.
(14) For an early paper, see: Meshkova, S. B.; Rusakova, N. V.; Bolshoi,
D. V. Acta Chim. Hung. 1992, 129, 317.
(49) Weber, M. J. Phys. ReV. 1968, 168, 283.
(50) Polman, A. Physica B 2001, 300, 78.
(15) Gschneidner, K. A.; Eyring, L. R. Handbook on the Physics and
Chemistry of Rare Earths; North-Holland Publishing Company: Amsterdam,
1979.
(16) Lowe, M. P.; Parker, D. Inorg. Chim. Acta 2001, 317, 163.
(17) Vo¨gtle, F.; Gorka, M.; Vicinelli, V.; Ceroni, P.; Maestri, M.;
Balzani, V. Chem. Phys. Chem. 2001, 769.
(18) Hebbink; G. H.; Klink; S. I.; Grave, L.; Oude Alink, P. G. B.; van
Veggel, F. C. J. M. ChemPhysChem. 2002, 3, 1014.
(19) Sabbatini, N.; Guardigli, M.; Lehn, J.-M. Coord. Chem. ReV. 1993,
123, 201.
(20) One of the referees argued that it should be efficiency of
luminescence instead of quantum yield, because this value was not measured
directly. However, as stated it is correct notwithstanding the fact that this
value was indeed measured indirectly.
(51) The energy levels of the three dyes, the 0-0 transitions of the singlet
and the triplet states, were determined from the spectra and are represented
in Figure 1. The 0-0 transitions of the singlet state are estimated as the
crossing of the absorption and fluorescence spectra. The 0-0 of the triplet
state was estimated from crossing of the singlet-triplet excitation spectrum
and the phosphorescence spectrum of 1R4.Gd at 77 K. Although the singlet-
triplet absorption is relative weak, the excitation spectrum could easily be
distinguished from the singlet-singlet excitation of the phosphorescence
(cf. ref 35 p 54).
(52) Delayed detection was accomplished by equipping the spectro-
fluorimeter with a rotating drum. Doing so, a delay between excitation and
emission of about 2 ms was accomplished. Compare ref 37.
(53) Tobita, S.; Arakawa, M.; Tanaka, I. J. Phys. Chem. 1984, 88, 2697.
Tobita, S.; Arakawa, M.; Tanaka, I. J. Phys. Chem. 1985, 89, 5649.
(54) Guldi, D. M.; Mody, T. D.; Gerasimchuk, N. N.; Magda, D.; Sessler,
J. L. J. Am. Chem. Soc. 2000, 122, 8289.
(21) Sato, S.; Wada, M. Bull. Chem. Soc. Jpn. 1970, 43, 1955.
(22) Latva, M.; Takalo, H.; Mukkala, V. M.; Matachescu, C.; Rodriguez-
Ubis, J. C.; Kankare, J. J. Lumin. 1997, 75, 149.
(23) Werts, M. H. V.; Duin, M. A.; Hofstraat, J. W.; Verhoeven, J. W.
Chem. Commun. 1999, 799.
(55) Gonc¸alves e Silva, F. R.; Malta, O. L.; Reinhard, C.; Gu¨del, H.
U.; Piguet, C.; Moser, J. E.; Bu¨nzli, J.-C. G. J. Phys. Chem. A 2002, 106,
1670.
(24) Dadabhoy, A.; Faulkner, S.; Sammes, P. G. J. Chem. Soc., Perkin
Trans. 2 2000, 2359.
(25) Stein, G.; Wu¨rzberg, E. J. Phys. Chem. 1975, 62, 208.
(26) Horrocks, W. DeW.; Bolender, J. P.; Smith, W. D.; Supkowski, R.
M. J. Am. Chem. Soc. 1997, 119, 5972.
(56) Blasse, G.; Grabmaier B. C. Luminescent materials; Springer-
Verlag: Berlin, 1994.
7
(57) The H11/2 level of Nd3+ can be an acceptor level because of the
selection rules. However, this level is higher in energy than the triplet states
of the sensitizers and is therefore not taken in to account.
(27) Korovin, Y.; Rusakova, N. ReV. Inorg. Chem. 2001, 21, 299.
(28) 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.
(29) Werts, M. H. V.; Hofstraat, J. W.; Geurts, F. A. J.; Verhoeven, J.
W. Chem. Phys. Lett. 1997, 276, 196.
(58) Sato, S.; Wada, M. Bull. Chem. Soc. Jpn. 1970, 43, 1955.
(59) The assumption here is that all other energy conversion processes
are not temperature-dependent. This is not completely true, but the effects
are much smaller than the temperature dependence of the back-tranfer rate
(Turro, N. J. Modern Molecular Photochemistry; The Benjamin/Cummings
Publishing Co., Inc.: Menlo Park, CA, 1978).
(30) Klink, S. I.; Oude Alink, P.; Grave, L.; Peters, F. G. A.; Hofstraat,
J. W.; Geurts, F. A. J.; van Veggel, F. C. J. M. J. Chem. Soc., Perkin Trans.
2 2001, 363.
(31) Wong, W. K.; Hou, A.; Guo, J.; He, H.; Zhang, L.; Wong, W. Y.;
Li, K. F.; Cheah, K. W.; Xue, F.; Mak, T. C. W. J. Chem. Soc., Dalton
Trans. 2001, 3092.
(32) Beeby, A.; Dickins, R. S.; FitzGerald, S.; Govenlock, L. J.; Maupin,
C. L.; Parker, D.; Riehl, J. P.; Siligardi, G.; Williams, J. A. G. Chem.
Commun. 2000, 1183.
(33) Klink, S. I.; Keizer, H.; van Veggel, F. C. J. M. Angew. Chem.,
Int. Ed. 2000, 39, 4319.
(60) Payne, S. A.; Bibeau, C. J. Lumin. 1998, 79, 143.
(61) Rehm, D.; Weller, A. Isr. J. Chem. 1970, 8, 259.
(62) -∆G ) -E(dye•+/dye) + Edye* + E(Yb3+/Yb2+); E(dye•+/dye) )
1.14 V vs NHE (Jones, G., III.; Qian, X. J. Photochem. Photobiol. A: Chem.
1998, 113, 125), E(Yb2+/Yb3+) ) 1.05 V vs NHE (Lide, D. R., Ed.;
Handbook of Chemistry and Physics, 76th ed.; CRC Press: Boca Raton,
FL, 1996). The oxidation potential of Yb3+ is expected to be even higher
as the triple charged ion is stabilized by the ligand. This leads to even lower
values of the driving force.
(63) E(Eu3+/Eu2+) ) -0.35 V (vs NHE), E(Yb3+/Yb2+) ) -1.05 V
(vs NHE), Lide, D. R., Ed.; Handbook of Chemistry and Physics, 76th ed.;
CRC Press: Boca Raton, FL, 1996.
(34) Dexter, D. L. J. Chem. Phys. 1953, 21, 836.