Lanthanide chelates based on diethylenetriamine fitted with O-benzoic acid pendant arms

Article abstract of DOI:10.1002/ejic.200390173

A new polycarboxylate ligand H₅L has been synthesized by the attachment of five benzoate subunits onto a diethylenetriamine framework. Seven pK_α values have been determined by potentiometry, spectrophotometry and NMR spectroscopy as 1.9(2), 2.8(2), 3.87(5), 4.58(6), 4.87(6), 9.19(6) and 11.68(5), the first four corresponding to the carboxylic functions and the last three to amine sites. The interaction between H₅L and Ln^III ions in dilute aqueous solution has been examined by UV/Vis absorption and emission spectrometries, and has been found to result in monometallic complexes that are moderately stable in the pH range 3.7-7.5. Conditional stability constants at pH 5.3 are logK₁₁ = 5.3(2), 6.6(1), 6.5(1) and 7.2(3) for La, Eu, Tb and Lu, respectively. In the case of Tb^III, the stability constants for [Tb(HL)]^- and [Tb(H₂L)] are logβ₁₁₁ = 22.0(2) and logβ₁₂₁ = 29.8(1), giving a pTb of 10.0. In the pH range 4-7, more than 90% of the Tb^III ions are in the form of the neutral species [Tb(H₂L)]. Lifetime determinations of the Eu(⁵D₀) and Tb(⁵D₄) excited levels in both H₂O and D₂O at pH 5.3 indicate 4.8 ± 0.5 (Eu) and 4.5 ± 0.5 (Tb) water molecules being bound in the inner coordination sphere of the Ln^III. The triplet state of the ligand in water lies at around 26000 cm^-1, resulting in a sizeable sensitisation of the Tb-centred luminescence (absolute quantum yield: φ_abs = 10.3%), while the luminescence of Eu^III is only poorly sensitised (φ_abs = 1.5%). Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2003.

Full text of DOI:10.1002/ejic.200390173

FULL PAPER
Lanthanide Chelates Based on Diethylenetriamine Fitted with O-Benzoic Acid
Pendant Arms
Daniel Imbert,[a] Nicolas Fatin-Rouge,[a] and Jean-Claude G. Bünzli*[a]
Keywords: Carboxylate ligands / Lanthanides / Luminescence / Podand / Stability constant
A new polycarboxylate ligand H5L has been synthesized by
the attachment of five benzoate subunits onto a diethyl-
logβ111 = 22.0(2) and logβ121 = 29.8(1), giving a pTb of 10.0.
In the pH range 4−7, more than 90% of the TbIII ions are in
enetriamine framework. Seven pKa values have been deter- the form of the neutral species [Tb(H2L)]. Lifetime deter-
mined by potentiometry, spectrophotometry and NMR spec-
troscopy as 1.9(2), 2.8(2), 3.87(5), 4.58(6), 4.87(6), 9.19(6) and
11.68(5), the first four corresponding to the carboxylic func-
minations of the Eu(5D0) and Tb(5D4) excited levels in both
H2O and D2O at pH 5.3 indicate 4.8 0.5 (Eu) and 4.5 0.5
(Tb) water molecules being bound in the inner coordination
tions and the last three to amine sites. The interaction be- sphere of the LnIII. The triplet state of the ligand in water lies
tween H5L and LnIII ions in dilute aqueous solution has been
at around 26000 cm−1, resulting in a sizeable sensitisation of
examined by UV/Vis absorption and emission spectrometries, the Tb-centred luminescence (absolute quantum yield: ϕabs
=
and has been found to result in monometallic complexes that
are moderately stable in the pH range 3.7−7.5. Conditional
stability constants at pH 5.3 are logK11 = 5.3(2), 6.6(1), 6.5(1)
and 7.2(3) for La, Eu, Tb and Lu, respectively. In the case of
TbIII, the stability constants for [Tb(HL)]− and [Tb(H2L)] are
10.3%), while the luminescence of EuIII is only poorly sen-
sitised (ϕabs = 1.5%).
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2003)
Introduction
organised (macrocyclic)[10] or predisposed receptors,[11,12]
leading to induced cavities; these receptors can also be ob-
tained by self-assembly processes[13] or by the use of flexible
podands.[14] These latter are readily available, cheap and are
easy to derivatize. In this paper we attempt to combine the
stability of LnIII complexes with ligands derived from poly-
aminocarboxylates, such as dtpa,[15] with the large
Stable chelates of trivalent lanthanide ions are of interest
as contrast enhancement agents in magnetic resonance im-
aging,[1] as catalysts for the cleavage of phosphodiester
bonds in DNA and RNA,[2] as precursors for various func-
tional materials,[3] including doped polymers for optical
amplifiers,[4] or as luminescent probes for biomedical
analysis,[5Ϫ7] for responsive analytical sensors[8] or for im-
aging of cancerous cells.[9] The main difficulty associated
with the design of luminescent probes and sensors lies in
overcoming the very low molar absorption coefficients of
the LnIII ions, due to the forbidden nature of the f-f trans-
itions. To sensitise the metal ion efficiently, there are specific
requirements for ligand molecules i.e. that they should be
strong complexation agents, have large absorption coeffi-
cients and offer a series of photophysical properties al-
lowing efficient energy transfer from the ligand onto the
metal ion and minimization of quenching processes.[5]
Several synthetic strategies intended to meet all these re-
quirements have been examined,[3] including the use of pre-
[a]
Swiss Federal Institute of Technology Lausanne, Institute of
Molecular and Biological Chemistry
1015 Lausanne, Switzerland
Fax: (internat.) ϩ41-21/693-9825
E-mail: jean-claude.bunzli@epfl.ch
Supporting information for this article is available on the
WWW under http://www.eurjic.org or from the author.
Scheme 1
1332  2003 WILEY-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim 1434Ϫ1948/03/0407Ϫ1332 $ 20.00ϩ.50/0 Eur. J. Inorg. Chem. 2003, 1332Ϫ1339

Lanthanide Chelates Based on Diethylenetriamine
FULL PAPER
sensitisation of TbIII ion luminescence (and also other LnIII
ions) achieved by incorporation of the benzoate chromo-
phore,[5,16] absolute quantum yields of 100 and 20% having
been reported for TbIII and EuIII, respectively, in the solid
state.[17] We have therefore synthesized the polydentate li-
gand H5L [1,4,7-heptanetriamine-N,N,NЈ,NЈ,NЈЈ-pentakis-
(2-methylbenzoic acid), Scheme 1] and investigated its
complexation and sensitisation properties towards trivalent
lanthanide ions.
Table 1. Acidity constants of H7L2ϩ (Ϯ σ) and dtpa (see text for
the definition of Kai)
L[a]
L[b]
L[c]
dtpa[d]
pKa1
pKa2
pKa1 ϩ pKa2
pKa3
pKa4
pKa5
pKa6
pKa7
Ϫ
Ϫ
4.20(8)
3.87(5)
4.58(6)
4.87(6)
9.19(6)
11.68(5)
1.9(2)
1.7(4)
1.45
1.75
3.20
2.06
2.73
4.28
8.65
10.59
2.8(2)
4.7(4)
Ϫ
4.5(2)
Ϫ
9.4(3)
11.6(3)
2.9(3)
4.6(7)
Ϫ
4.7(1)
Ϫ
9.0(4)
11.6(4)
Results and Discussion
[a]
Potentiometric data. [L]0 ϭ 1.0·10Ϫ3 ; I ϭ 0.1  (KCl); T ϭ
[b]
20.0 °C.
Spectrophotometric data. [L]0 ϭ 2.0·10Ϫ5 ; I ϭ 1 
(KCl); T ϭ 20.0 °C. [c] 1H NMR spectroscopic data. [L]0 ϭ 1.0·10Ϫ3
; T ϭ 20 °C. [d] 1H NMR spectroscopic data. I ϭ 0.1  (KNO3);
T ϭ 25.0 °C.[34]
Acidity Constants of the Ligand
The ligand H5L was prepared in three steps (Scheme 1)
in an overall yield of 31%. Its acidity constants were deter-
mined separately by potentiometry (Figure 1) and spectro-
1
photometry (Figure 2) and further confirmed by H NMR
(H8ϪiL)(3Ϫi)ϩ o (H7ϪiL)(2Ϫi)ϩ ϩ Hϩ
spectroscopy. The extracted pKa values, defined by the fol-
lowing equations and obtained by titration in the 2Ϫ12 pH
range, are summarized in Table 1.
(1)
pKai (i ϭ 0Ϫ4, carboxylic functions)
(H8ϪiL)(3Ϫi)ϩ o (H7ϪiL)(2Ϫi)ϩ ϩ Hϩ
(2)
pKai (i ϭ 5Ϫ7, amine functions)
The agreement between the values from various experi-
mental techniques is quite good, and the dissociation con-
stants display behaviour similar to that reported for dtpa:
five pKa values are in the 2 to 5 range while the last two
are between 9 and 12. A fit of the data with eight pKa values
did not converge, indicating that pKa0, corresponding to the
deprotonation of a carboxylic group (see below, NMR spec-
tra), probably lies below pH 2. However, the protolytic
groups of H5L are less acidic than those of dtpa. The first is
only slightly less acidic (∆logKa ϭ 0.35), but the difference
increases up to ∆logKa ϭ 1.9 for pKa4. The electron-donor
ability of the benzene moieties and the larger number of
bonds separating the nitrogen atoms from the carboxylate
groups explain this trend.
Figure 1. Potentiometric titration of H7L2ϩ with OHϪ in H2O/
CH3OH (98:2 v/v); T ϭ 20.0 Ϯ 0.2 °C; µ ϭ 0.1  (KCl)
1H NMR spectra were measured from pH 12.2 to 1.1 in
D2O/CD3OD (95:5) solutions (Figure 3), and the reson-
ances were attributed on the basis of their chemical shifts,
signal intensities and 1H-1H COSY experiments. In the aro-
matic region, because of superimposed signals, only the
protons labelled (i) and (h) could be unambiguously as-
signed. The eight protons labelled (c) and the two protons
labelled (d) appear to be equivalent, pointing to an averaged
structure of the molecule on the NMR timescale. The (a)
and (b) protons are nonequivalent and appear as two broad
peaks because of spin coupling and conformational mo-
tions.
The variation of the chemical shifts with pH is shown at
the bottom of Figure 3. At pH 12.2, corresponding to a
solution containing mainly the totally deprotonated ligand
L5Ϫ, the resonances for the (c) and (d) protons have very
similar chemical shifts, and this is also true for protons (a)
and (b). A decrease in the pH to 10.4 by the addition of
one equivalent of acid causes only slight variations in the
Figure 2. UV/Visible absorption spectra of H7L2ϩ as a function of
pH in H2O/CH3OH (98:2 v/v); T ϭ 20.0 Ϯ 0.2 °C; µ ϭ 1  (KCl):
curve a (pH ϭ 1.37), curve b (pH ϭ 12.04)
Eur. J. Inorg. Chem. 2003, 1332Ϫ1339
1333

D. Imbert, N. Fatin-Rouge, J.-C. G. Bünzli
FULL PAPER
As a result, the preferred protonation sites and the extent
of protonation can be estimated from a simple model taking
into account only the two closest protonation sites, where f
is the fractional population of a given protonation site, C is
the chemical shift increment caused by the presence of the
proton in this site, and the indices α and β refer to the
position of the protonated nitrogen atom relative to the res-
onant proton.
δa ϭ δa0 ϩ f1 ϫ CNα ϩ f2 ϫ CNβ
δb ϭ δ0b ϩ f2 ϫ CNα ϩ f1 ϫ CNβ
δc ϭ δ0c ϩ f2 ϫ CNα ϩ f3 ϫ CO3
δd ϭ δ0d ϩ f1 ϫ CNα ϩ f4 ϫ CO4
(3)
(4)
(5)
(6)
We have set CO3, CO4, CO to be equal. The δ0i values were
taken as equal to the chemical shifts measured at pH 12.2.
Finally, the fi values were normalised by the following rela-
tionship, where n is the number of equivalents of acid added
to the fully deprotonated ligand L5Ϫ
.
f1 ϩ 2f2 ϩ 4f3 ϩ f4 ϭ n
(7)
Values of f1, f2 and CNα, CNβ were obtained only for n ϭ
1 and 2, because the data become too imprecise beyond this
amount of acid. For n ϭ 0Ϫ2, f3 and f4 were set as equal
to zero, which is a reasonable assumption. The best results
for n ϭ 1 were obtained at pH 1/2(pKa6 ϩ pKa7), and Equa-
tions (3), (4), (6) and (7) give f1 ϭ 0.22, f2 ϭ 0.39, CNα
ϭ
0.59 ppm and CNβ ϭ 0.05 ppm. Generally speaking, the C
values have large error margins because of the fairly small
CD3OD (95:5 v/v) at T ϭ 20 °C, and (bottom) variation of the H chemical shift changes. The calculated CNα value is in good
Figure 3. (Top) 1H NMR spectra of HnL(5Ϫn)Ϫ vs. pH in D2O/
1
NMR chemical shifts vs. pH
agreement with literature data for dtpa (0.75), but the value
found for CNβ is much lower than the reported value of
about 0.35 ppm.[18] At pH 1/2(pKa4 ϩ pKa3), different
shifts of the methylene protons, which become more shi-
elded, except for protons (d), which become deshielded.
Similarly, the aromatic protons undergo small positive
shifts, except for protons (h), which become deshielded.
Further addition of acid again results in relatively small
changes until around pH 9, except that the shift tendency
inverts for protons (d) and (i), which become more shielded.
In the pH range 9 to 3, dramatic changes occur for the
methylene protons and for protons (h); in particular, the
resonances of protons (a) and (b) on one hand and (c) and
(d) on the other become clearly separated. After a discon-
tinuity at pH 3, all the protons, particularly protons (h),
become more shielded upon further addition of acid.
The observed shift of the methylene protons is in fact an
average value reflecting the mean time that each proton
spends in a given chemical environment.[18] Quantitative
analysis of the pH dependence of these chemical shifts can
therefore be performed by assuming that the contributions
to the shift of a given proton i (δi) are additive and linearly
values of n were again tested and n ϭ 2 gave the best result,
with Equations (3)Ϫ(5) and Equation (7) yielding f1 ϭ 0.38,
f2 ϭ 0.81, CNα ϭ 0.82 ppm, and CNβ ϭ 0.77 ppm. These
calculations offer evidence of the drastic change in the
charge repartition on the three nitrogen atoms on going
from n ϭ 1 to n ϭ 2, consistent with the data for dtpa.[18]
At n ϭ 2, for instance, the protons are mainly bound to the
end nitrogen atoms N2, as a result of electrostatic repulsion.
Further attempts to determine quantitatively the extent of
protonation of the benzoate groups with pH failed. Never-
theless, the similar values of the three largest pKa values for
dtpa and for H5L point to the protonation of L5Ϫ occurring
first on the three nitrogen atoms. However, distribution dia-
grams show that protonation of the benzoate sites starts
while full protonation of the central nitrogen atom is not
yet complete.
Interaction with Trivalent Lanthanide Ions
The interaction between LnIII ions (Ln ϭ La, Eu, Tb,
related to the fraction of time spent in each environment. Lu) and H5L was first monitored in dilute aqueous solu-
1334
Eur. J. Inorg. Chem. 2003, 1332Ϫ1339

Lanthanide Chelates Based on Diethylenetriamine
FULL PAPER
tions by UV/Vis spectrophotometry at pH 5.3. The vari-
ation of the absorbance versus R ϭ [LnIII
]tot/[H5L]tot (Fig-
ure 3) points to the presence of a single 1:1 complex species
in the stoichiometry range investigated (0 Ͻ R Ͻ 5). This
is in agreement with the ESI-MS spectra. The conditional
stability constants extracted from these data amount to
logK11 ϭ 5.3(2), 6.6(1), 6.5(1) and 7.2(3), respectively (Fig-
ure S1, Supporting Information, see also the footnote on
the first page of this article). The linear relationship be-
tween the magnitude of the binding constants and the
charge density of the cation is a clear indication that there
is no steric effect in the coordination of the metal ions. To
determine the influence of the pH on the speciation, we
titrated the ligand in the presence of one equivalent of TbIII
and monitored the experiment both by spectrophotometry
(Figure 4) and by luminescence spectroscopy (Figure 5). Po-
tentiometry could not be used because of precipitation of
the metal complexes at concentrations greater than 10Ϫ3 .
The best model used to fit the titration data was fairly com-
plex, with ten different [Lni(OH)jHkLl](3iϪjϩkϪ5l)ϩ species:
one hydroxide moiety (i ϭ 1; j ϭ 3; k ϭ l ϭ 0), six pro-
tonated forms of the ligand (i ϭ j ϭ 0) determined inde-
pendently from the potentiometric titration described
above, and two protonated 1:1 complexes (i ϭ 1, j ϭ 0, k ϭ
1 and 2, l ϭ 1). The absorption maxima of the different
species are given in Table 2. The two proton-dependent
stability constants refined amount to logβ111 ϭ 22.0(2)
[Equation (8)] and logβ121 ϭ 29.8(1) [Equation (9)].
Figure 5. Emission spectra of H7L2ϩ 2·10Ϫ5  in H2O/CH3OH
(98:2 v/v), µ ϭ 0.1  (KCl) in the presence of one equivalent of
TbIII, at T ϭ 20.0 Ϯ 0.2 °C as a function of pH; inset: phosphores-
cence intensity at 545 nm vs. pH
Table 2. Absorption maxima of the ligand species at T ϭ 20 °C
Species λmax
εmax
Species
[H5L]
λmax
εmax
(nm)[a] (Ϫ1·cmϪ1 [b]
)
(nm)[a] (Ϫ1·cmϪ1 [b]
)
[L]5Ϫ
268 sh 3300
230
278
33800
7200
6700
[HL]4Ϫ 268 sh 3600
[H7L]2ϩ 277
[H2L]3Ϫ 272
4300
30000
5600
33900
6600
[TbHL]Ϫ 272 sh 3600
[H3L]2Ϫ 228sh
[TbH2L] 272
3700
275
229
278
Ϫ
[H4L]
[a]
[b]
Ϯ 2 nm. Ϯ 10%.
which explains the difficulty in solubilising the metal com-
plexes after their isolation as solid-state samples. At lower
pH, competition with protons dissociates the complex,
while very stable hydroxides are the major species at pH
values higher than 8.5. Altogether, the metal complexes are
moderately stable, as shown by the pTb value of 10.0,[19]
calculated with [Tb]tot ϭ 10Ϫ6 , [H5L]tot ϭ 10Ϫ5  and pH
7.4, in comparison with 19.2 for [Tb(dtpa)]Ϫ [20] and with
6.7 for terbium benzoate. This result is a clear indication
that the nitrogen atoms of the H5L ligand are most prob-
ably nonbonding and remain protonated in the metal com-
plexes, as indicated by the unchanged values of the two
highest apparent pKa values upon addition of LnIII ions.
Figure 4. UV/Visible absorption spectra of H7L2ϩ 2·10Ϫ5  in H2O/
CH3OH (98:2 v/v), µ ϭ 0.1  (KCl) in the presence of one equiva-
lent of TbIII, at T ϭ 20.0 Ϯ 0.2 °C at various pH; curve a, b, and
c: pH ϭ 1.92, 9.22, and 12.0; the inset shows the absorbance vari-
ation at 240 nm
Photophysical Properties of the Ligand and the [LnH2L]
Complexes
(8)
(9)
In water, the ligand displays two main adsorption bands
around 46080 (shoulder at 42370 cmϪ1) and 39370 cmϪ1
.
The high-energy band is slightly blue-shifted (ca. 400Ϫ500
cmϪ1) in the LnIII complexes. The photophysical properties
of the ligand and of its 1:1 complexes with La, Lu, Eu and
The calculated distribution curves from these stability Tb in water (pH 5.3) and in frozen glycerol/water mixtures
constants are presented in Figure 6. In the pH range are summarized in Table 3. UV excitation in the π Ǟ π*
3.5Ϫ7.5, the neutral species [Tb(H2L)] is clearly dominant, and n Ǟ π* absorption bands of the ligand results in a
Eur. J. Inorg. Chem. 2003, 1332Ϫ1339
1335

D. Imbert, N. Fatin-Rouge, J.-C. G. Bünzli
FULL PAPER
Figure 7. Emission spectra of the [LnH2L] complexes (Ln ϭ La,
Eu, Tb, Lu), pH ϭ 5.3; (left) phosphorescence spectra in solutions
of glycerol/water, 10:90% at 77 K; (right) fluorescence spectra at
295 K in water
Figure 6. Corresponding formation curves of proton and terbium
complexes, computed from the stability constants given in the text;
[L]tot ϭ [TbIII
]
ϭ 2.0·10Ϫ5 ; λexc ϭ 274 nm; T ϭ 20 °C
tot
5
by analysis of the maximum on the D0Ǟ7F2 transition at
ligand-centred emission displaying two overlapping bands 10 K, displays a single broad and symmetrical peak at
1
3
assigned to emission from the ππ* (33110 cmϪ1) and ππ* 17246 cmϪ1 with a full width at half height (fwhh) of 26.9
states (Figure 7). The absolute fluorescence quantum yield cmϪ1, which is consistent with the presence of a single site
of the fully deprotonated ligand (L5Ϫ) is small at 0.3%, a for the EuIII ion.
value that increases slightly for the [Ln(H2L)] complexes:
The metal-centred luminescence of diluted aqueous solu-
0.4 and 0.5% for La and Lu, respectively, at pH 5.3. At tions is pH dependent, due to the distribution of the various
77 K, the 3ππ* state emission is structured with a vibra- species present in solution. The Tb complex displays emis-
tional progression of 1350 Ϯ 30 cmϪ1, corresponding to a sion over a large pH range, from pH 3.5 to 10.5 (Figure 4),
ring breathing mode and its 0-phonon transition is located with maximum intensity occurring in the pH range 5Ϫ8, in
3
1
at 26520 cmϪ1. The gap between the ππ* and ππ* states which [TbH2L] is the major species in solution. The average
(ca. 6320 cmϪ1) is favourable for a relatively efficient in- numbers of water molecules bound to the metal ions were
tersystem crossing process.[5] The emission spectra of the determined from the experimental lifetimes at room temper-
ligand and its complexes at room temperature are consistent ature for solutions of [LnH2L] (Ln ϭ Eu, Tb) in H2O (pH
with this finding, since they display emission from both the 5.3) and D2O (pD 5.7).[23,24] For the Eu complex, the meas-
singlet and triplet states, although very weakly for Eu and ured lifetimes (τ ϭ 1.41 Ϯ 0.04 ms in D2O and 0.20 Ϯ 0.01
3
Tb. The ππ* lifetime amounts to 1.88 s, a value close to ms in H2O) give q ϭ 4.8 Ϯ 0.5, while q ϭ 4.5 Ϯ 0.5 is
that obtained for benzoic acid (τ ϭ 2 s).[21] As would be obtained for the Tb podate (τ ϭ 1.47 Ϯ 0.01 ms in D2O
expected from the relatively high energy of the 3ππ* 0- and 0.61 Ϯ 0.01 ms in H2O). The presence of at least eight
phonon component, both the EuIII ions, with ∆E(3ππ* Ϫ OH oscillators in the inner coordination sphere is consistent
5D0) ഠ 9700 cmϪ1, and the TbIII ions, with ∆E(3ππ* Ϫ with the reasoning, based on the stability constants, that at
5D4) ഠ 6500 cmϪ1, are reasonably well sensitised (Figure 7). least four benzoate units are coordinated to the metal ions
This is in agreement with what has been observed for TbIII (in a monodentate fashion), but the nitrogen atoms are not.
poly(aminocarboxylates), for which the best antenna effect The large number of water molecules in the first coordina-
3
was observed for ππ* states around 26000 cmϪ1 [22]
.
tion sphere explains the inability of the ligand to sensitise
The high-resolution excitation spectrum of a solid-state the luminescence of the DyIII, ErIII and TmIII ions. Despite
sample of [EuH2L] in the 5D0ǟ7F0 spectral range, obtained this, the quantum yields of metal-centred luminescence in
Table 3. Ligand-centred absorption and emission properties of the ligand (pH 12) and [LnH2L] podates at pH 5.3
Compound
E (cmϪ1 [a]
)
E (cmϪ1 [a]
)
τ(3ππ*) (ms)[b]
1ππ*[c]
3ππ*[c]
3ππ*[b]
[L]5Ϫ
46080, 42370 sh, 39370
46510, 42370 sh, 39275
46590, 42370 sh, 39525
46619, 42370 sh, 36490
46619, 42370 sh, 36490
33110
33005
33445
33655
33700
25706
26140
26040
25445
25640
26790, 25360, 23980
1879 Ϯ 26
789 Ϯ 12
809 Ϯ 6
1363 Ϯ 14
456 Ϯ 12
[LuH2L]
[LaH2L]
[EuH2L]
[TbH2L]
26920, 25540, 24240, 22910, 21530
26880, 25510, 24210, 22940, 21600
26920, 25610, 24230
26920, 25610, 24270
[a]
[b]
[c]
The most intense component is italicised. Luminescence data and lifetimes for frozen solutions in glycerol/water 10:90% (77 K).
Electronic spectroscopic data at 295 K in water; energies are given for the maximum of the band envelope (sh: shoulder).
1336
Eur. J. Inorg. Chem. 2003, 1332Ϫ1339

Lanthanide Chelates Based on Diethylenetriamine
FULL PAPER
distribution of the peaks. Electronic UV/Vis spectra were recorded
on a PerkinϪElmer Lambda 900 spectrometer. Infra-red spectra
were measured on a FT-IR Mattson Alpha Centauri spectrometer
(4000Ϫ400 cmϪ1, KBr pellets).
water at pH 5.3 remain sizeable, at 1.5 and 10.3% for the
Eu and Tb podates, respectively (Table 4). This points to
fairly efficient energy transfer from the ligand to the
Tb(5D4) level. It is noteworthy that the quantum yield of
[Tb(H2L)] increases significantly upon addition of HMPA,
which is know to be a good complexation agent for LnIII
ions and which removes the water from the inner coordina-
tion sphere (2 eqs: ϩ45%, 4, 10, 100 and 500 eqs: ϩ105%).
Syntheses (see Scheme 1). Methyl 2-(Bromomethyl)benzoate (1):
This compound was synthesized by a known procedure.[26]
1,4,7-Heptanetriamine-N,N,NЈ,NЈ,NЈЈ-pentakis(methyl
2-methyl-
benzoate) (2): Potassium carbonate (10.04 g, 72.7 mmol) was ad-
Moreover, a graph of the quantum yield of [Tb(H2L)] vs. ded under nitrogen atmosphere to a solution of diethylenetriamine
(500 mg, 4.9 mmol) in freshly distilled acetonitrile (100 mL). The
resulting mixture was heated at reflux for 30 min, and a solution
of 1 (6.08 g, 26.7 mmol) in dry acetonitrile (50 mL) was added over
2 h. The mixture was stirred and heated at reflux for an additional
12 h and filtered while hot. The solvents were then evaporated to
give a black oil, which was dissolved in dichloromethane, washed
with brine and dried with Mg2SO4, and the solvents were evapor-
ated to dryness. The residue was purified by column chromato-
graphy (silica gel; 2% MeOH in CH2Cl2). Yield: 1.76 g (43%), or-
ange foam. IR (KBr): ν˜ ϭ 1711 (CϭO), 1452, 1384 cmϪ1. 1H NMR
(300 MHz, CDCl3, 25 °C): δ ϭ 2.46 (m, 8 H, CH2), 3.72 (s, 8 H,
CH2), 3.76 (s, 2 H, CH2), 3.82 (s, 12 H, CH3), 3.86 (s, 3 H, CH3),
pH (Figure S2, Supporting Information) reflects the distri-
bution curve of this species reasonably well.
Conclusion
The solution studies described in this paper show that
podand H5L gives stable LnIII complexes in water that are
resistant toward hydrolysis and show interesting photophys-
ical properties. Thermodynamically stable 1:1 neutral pod-
ates [Ln(H2L)] form in the pH range 3.7Ϫ8. No nitrogen
atom is involved in the binding, as a result of the pro-
1
2
7.21 (m, 5 H, ArH), 7.27 (dt, J ϭ 7.5, J ϭ 1.3 Hz, 1 H, ArH),
1
2
tonation of the amine functions, preventing the wrapping 7.34 (dt, J ϭ 7.8, J ϭ 1.3 Hz, 4 H, ArH), 7.41 (d, J ϭ 7.5 Hz, 1
1
H, ArH), 7.58 (d, J ϭ 7.8 Hz, 4 H, ArH), 7.68 (dd, J ϭ 7.5, 2J ϭ
of the ligand around the metal ion. The emission lifetimes
1.3 Hz, 1 H, ArH), 7.74 (dd, 1J ϭ 7.8, 2J ϭ 1.3 Hz, 4 H, ArH)
ppm. 13C NMR (CDCl3): δ ϭ 51.7, 51.8 (CH3), 52.5, 52.6, 57.0,
57.6 (CH2), 126.3, 129.3, 129.7, 129.9 (CH), 130.1, 130.4 (Cq),
131.3, 131.6 (CH), 141.4 (Cq), 168.2, 168.4 (CϭO) ppm. ESI-
MS (CH3CN, H2O, CH3CO2H): m/z ϭ 844.3 (calcd. 844.3) [M ϩ
H]ϩ, 422.8 (calcd. 422.7) [M ϩ 2H]2ϩ. C49H53N3O10 (843.97):
calcd. C 69.73, H 6.33, N 4.98; found C 69.72, H 6.31, N 4.96.
of the EuIII and TbIII complexes recorded in D2O and H2O
point to four water molecules completing the coordination
sphere of the metal ions. The luminescence study shows that
the ligand exhibits a good antenna effect with respect to the
TbIII ion, due to efficient intersystem crossing (1ππ*- to-
3ππ*) and ligand-to-metal energy transfer. Despite the large
number of coordinated water molecules, the quantum yields
1,4,7-Heptanetriamine-N,N,NЈ,NЈ,NЈЈ-pentakis(2-methylbenzoic
acid) (H5L): Compound 2 (500 mg, 0.59 mmol) was dissolved in a
solution of methanol (200 mL) and KOH in water (1 , 15 mL).
The resulting mixture was heated at reflux for 4 h, and the solvents
were evaporated. The residue was dissolved in water (400 mL), co-
oled to 0 °C and acidified to pH 2 with aqueous HCl. The yellow
precipitate was collected, washed with water and then diethyl ether,
and dried. Yield: 417 mg (91%), yellow solid. IR (KBr): ν˜ ϭ 3426
br (OH), 2606 br (OH), 1700 (CϭO), 1590, 1552, 1451, 1381, 1295,
of the metal-centred luminescence in the TbIII (and EuIII
)
complexes are encouraging and demonstrate that benzoate
is a potentially interesting moiety for attachment onto a
flexible receptor for LnIII ions.
Experimental Section
1
1145, 1085, 752 cmϪ1. H NMR (300 MHz, D2O, 20 °C): δ ϭ 2.8
Starting Materials and General Procedures: Analytical grade solv-
ents and chemicals (Fluka AG) were used without further purifica-
tion, except for acetonitrile, which was distilled from CaH2. Solu-
tions were prepared just before use with freshly boiled, doubly dis-
tilled water saturated with N2. Stock solutions of LnX3·nH2O (X ϭ
ClO4, CF3SO3) were prepared from lanthanide oxides (99.99%,
(m, 4 H, CH2), 3.3 (m, 4 H, CH2), 3.63 (s, 2 H, CH2), 4.44 (s, 8 H,
CH2), 7.2 (m, 1 H, ArH), 7.3Ϫ7.4 (m, 14 H, ArH), 7.74 (m, 1 H,
ArH), 7.93 (m, 4 H, ArH) ppm. 13C NMR (D2O, 20 °C): δ ϭ 51.1,
55.4, 56.7, 57.9, 58.1 (CH2), 127.2, 129.6, 130.6, 130.8 (CH), 131.4,
131.6 (quat. C), 133.1, 134.4 (CH), 137.5 (quat. C), 170.7 (CϭO)
ppm. ESI-MS (CH3CN, H2O, CH3CO2H): m/z ϭ 774.4 (calcd.
774.3) [M ϩ H]ϩ, 387.8 (calcd. 387.7) [M ϩ 2H]2ϩ. H5L·H2O
C44H45N3O11 (791.85): calcd. C 66.74, H 5.73, N 5.31; found C
66.72, H 5.60, N 5.20.
ˆ
Rhone Poulenc) and the corresponding acid. They were systematic-
ally acidified with HCl to a pH of about 4 before titration to avoid
hydroxide precipitation. The concentrations were determined by
complexometric titrations with a standardized Na2H2EDTA solu-
tion in urotropine-buffered medium and with xylenol orange as
indicator.[25] Elemental analyses were performed by Dr. Eder, Mic-
rochemical Laboratory, University of Geneva.
Preparation of [Ln(H2L)]·nH2O Complexes (Ln ؍ La, Eu, Tb, Lu):
Freshly titrated LnClO4.nH2O solutions (10Ϫ3 , 0.039 mmol) were
added over 4 h, at room temperature, to an aqueous solution of
H5L (40 mL, 0.039 mmol, 10Ϫ3 , pH ϭ 5.3, adjusted with HCl).
The mixture was stirred for an additional 2 h and the resulting
white precipitate was filtered, washed successively with water (sev-
eral times) and diethyl ether, and further dried for 3 days at 65 °C
and 0.01 Torr. These complexes being very hygroscopic, variable
numbers of water molecules were found when performing microan-
alyses.
1-D and 2-D NMR spectroscopic data were collected on Bruker
DPX 400 or AM 360 spectrometers. 1H NMR and 13C NMR spec-
tra were recorded in CDCl3, D2O or CH3OD and their assignment
was based on 1H coupling and 2-D 1H-13C correlation spectra. The
mass spectra were recorded in electrospray ionisation mode on a
Finnigan quadrupole mass spectrometer (Tmatrix ϭ 240 °C). The
instrument was calibrated with a horse myoglobin standard. The
analyses were conducted in positive mode. The ion spray voltage [La(H2L)]·5H2O (75%): IR (KBr): ν˜ ϭ 3422 br (OH), 1554 and
was 4.6 kV. Assignment of the species was based on the isotopic
1399 (νas, νs COO), 1606, 1498, 1449, 1294, 1203, 1151, 1120, 1089,
Eur. J. Inorg. Chem. 2003, 1332Ϫ1339
1337

D. Imbert, N. Fatin-Rouge, J.-C. G. Bünzli
753 cmϪ1. C44H50LaN3O15 (999.80): calcd. C 52.86, H 5.04, N 4.20; lower than 3.2. The pH values of the solutions were determined
FULL PAPER
found C 52.27, H 4.50, N 4.27.
with a Metrohm Titrino 736 GP potentiometer equipped with a
calibrated Metrohm 6.0234.100 glass electrode in H2O. Ionic
strength was not adjusted. The pD values were obtained from the
equation pD ϭ pHmeasd. ϩ 0.4.[30]
[Eu(H2L)]·3H2O (89%): IR (KBr): ν˜ ϭ 3413 br (OH), 1556 and
1403 (νas, νs COO), 1607, 1490, 1449, 1294, 1202, 1150, 1120, 1094,
752 cmϪ1. ESI-MS: a stoichiometric mixture of Eu(ClO4)3·xH2O
(x ϭ 4.2) and H5L 10Ϫ4  in pure methanol gave [EuH3L]ϩ m/z ϭ
923.4 (calcd. 923.4). C44H46EuN3O13 (976.81): calcd. C 54.10, H
4.75, N 4.30; found C 53.65, H 4.59, N 4.30.
[Tb(H2L)]·7H2O (84%): IR (KBr): ν˜ ϭ 3416 br (OH), 1556 and
1407 (νa, νs COO), 1608, 1491, 1450, 1295, 1203, 1151, 1120, 1087,
752 cmϪ1. C44H54N3O17Tb (1055.83): calcd. C 50.05, H 5.15, N
3.98; found C 49.16, H 4.78, N 3.72. Complexometry: calcd. Tb
15.05; found Tb 15.28 Ϯ 0.28.
Luminescence Titrations: The solutions {[H5L]0 ϭ 2.0·10Ϫ5 ,
[Tb(ClO4)3]0 ϭ 2.0·10Ϫ5 , 10 mL, pH0 ഠ 2; [KOH]tit ϭ 0.050 ;
µ ϭ 0.1  (KCl)} were identical to those used for the spectrophoto-
metric pH-dependent titration described previously. After a 15 min
delay, 3 mL of solution were transferred into a quartz cell and de-
gassed with Ar for 5 min. Absorbance at the excitation wavelength
was measured (A Ͻ 0.05) and emission (λexc ϭ 285 nm) and excita-
tion spectra were collected.
[Lu(H2L)]·4H2O (92%): IR (KBr): ν˜ ϭ 3417 br (OH), 1561 and
1410 (νas, νs COO), 1608, 1491, 1450, 1294, 1203, 1152, 1120, 1088,
752 cmϪ1. C44H48N3O14Lu (1017.84): calcd. C 51.92, H 4.75, N
4.13; found C 51.98, H 4.41, N 4.14.
Luminescence Measurements: Low-resolution luminescence meas-
urements (spectra and lifetimes) were recorded on a PerkinϪElmer
LS-50B spectrofluorimeter. Phosphorescence lifetimes (τ) were
measured with the instrument in time-resolved mode, on frozen
glycerol/water (10:90%) solutions in a quartz capillary or a 1-cm
Suprasil cell. They are the average of at least three independent
measurements, made by monitoring the decay at the maxima of the
emission spectra, enforcing a 0.03Ϫ0.04 ms delay. The decays were
mono-exponential and were analysed by use of the FLDM pro-
gram (PerkinϪElmer). Solutions of [Ln(HjL)](3ϩjϪ5)ϩ were pre-
pared in water at pH ϭ 5.3, with stock solutions of H5L (2.0·10Ϫ3
) and Ln(ClO4)3·xH2O (x ഠ 4.5, Ln ϭ La, Eu, Tb, Lu; ഠ 4·10Ϫ3
). Quantum yields of L5Ϫ (10Ϫ5 , pH ϭ 12) and [Ln(H2L)] (Ln ϭ
La, Eu, Tb, Lu, 10Ϫ5  in H2O) were determined in degassed water
relative to quinine sulfate in 0.05  aqueous H2SO4 (absolute
quantum yield: 0.546)[31] and cresol violet (absolute quantum yield:
0.52);[32] estimated error Ϯ10%. The number of coordinated water
molecules (q) for the Eu and Tb complexes were calculated from
q ϭ A(τHϪ1 Ϫ τDϪ1 Ϫ kcorr), where τH and τD are the lifetimes in
Physical Measurements. Potentiometric Titrations: H7L2ϩ was ti-
trated with a 5.1 mL sample ([H7L2ϩ
]
tot
ϭ 3.7·10Ϫ3 , solvent:
H2O/CH3OH 98:2 v/v) in a thermostatted (20.0 Ϯ 0.1 °C) glass-
jacketed vessel under an Ar atmosphere. The ionic strength was
fixed with KCl (µ ϭ 0.1 ). The solution was acidified to a pH of
about 1.8 (HCl) 30 min before titration. Titrations were carried out
with an automatic Metrohm Titrino 736 GP potentiometer linked
to an IBM PS/2 computer (resolution 0.1 mV, accuracy 0.2 mV)
and by use of constant volume addition (0.05 mL). An automatic
burette (Metrohm 6.3013.210, 10 mL, accuracy 0.03 mL) was used,
together with a Metrohm 6.0238.000 glass electrode. The standard
base solution (KOH ϭ 0.050 , µ ϭ 0.1 , KCl) was added inside
the solution through a capillary tip attached to the automatic bur-
ette. The data (200 points, drift Ͻ 1 mV/min) were mathematically
treated by use of the program SUPERQUAD[27] with a Marquardt
algorithm while the distribution of species was calculated with the
program Haltafall.[28] Calibration of the pH meter and the elec-
trode system was performed prior to each measurement with a
standardized HCl solution (µ ϭ 0.1 , KCl) at 20.0 °C: 10 mL of
the latter were titrated with a standardized 0.1  KOH solution at
a ionic strength of 0.1, and the electrode potential readings were
converted to pH. The ion product of solvent (pKw ϭ 13.91) and
electrode potential were refined by use of the program Scientist
by Micromath (Version 2.0).
H2O and D2O respectively, A ϭ 1.2 (Eu) and 5.0 (Tb), and kcorr
0.25 (Eu) and 0.06 (Tb) msϪ1 [24]
High-resolution spectra were re-
corded on a previously described setup.[33]
ϭ
.
Table 4. Absolute quantum yield of the ligand-centred fluorescence
in [L]5Ϫ (pH ϭ 12) and [LnH2L] (Ln ϭ La, Lu) and of the metal-
centred phosphorescence in [LnH2L] (Ln ϭ Eu, Tb) measured in
water at 295 K and pH ϭ 5.3 (values in D2O are italicised)
Compound
λexc (nm)
ϕabs (%)
Spectrophotometric Titrations: Electronic spectra were recorded
with a PerkinϪElmer Lamdba 900 spectrophotometer at 20 °C
(210Ϫ350 nm, 60 nm minϪ1 scan speed, spectral width 2 nm) in 1
cm Suprasil quartz cells. Titrations of 10 mL samples were per-
formed in a thermostatted (20.0 Ϯ 0.1 °C), glass-jacketed vessel
filled with Ar, at µ ϭ 1  (KCl); H7L2ϩ (pH, 2) was titrated with
KOH (0.1 ⁾, H5L (pH, 5.3, HCl) with Ln(ClO4)3 (ഠ 4.0·10Ϫ3 ,
Ln ϭ La, Eu, Tb, Lu) at pH ϭ 5.3, and [Tb(H2L)] 2.0·10Ϫ5  at
pH0 ഠ 2 by KOH (0.050 ⁾. Aliquots of the titrant were added
through a Socorex micropipette. The pH values of the titrated
solutions were monitored continuously. After a 15 min delay, 3 mL
of solution was transferred into the quartz cell with a Teflon syr-
inge. The stability constants were computed by use of the Specfit
program.[29] Differences between the measured and the computed
absorbances were less than 7 ϫ 10Ϫ3 absorbance unit at any wave-
length and exhibited statistical behaviour.
[L]5Ϫ
268
273
273
270
270
0.3
0.4
0.5
1.5, 3.3
10.3, 18.9
[LuH2L]
[LaH2L]
[EuH2L]
[TbH2L]
Acknowledgments
This work is supported through grants from the Swiss National
Science Foundation and the Swiss Federal Office for Science and
Education (COST Action D18). We thank Frederic Gumy for his
help in recording the high-resolution luminescence data.
´ ´
NMR Titration: A solution of H5L (11.6 mg, 7.3·10Ϫ3 ) was pre-
pared in 2 mL of a D2O/CD3OD (95:5, v/v) mixture, by addition
of a concentrated KOD/D2O aliquot (pD0 ഠ 12). Samples with
different pD values were prepared by addition of dilute D2SO4/
D2O (Fluka, puriss). Precipitation of ligand occurred at pD values
[1]
The chemistry of contrast agents in medical magnetic reson-
ance imaging, eds A. E. Merbach and E. Toth, Wiley, Lon-
don, 2001.
Z. P. Zheng, Chem. Commun. 2001, 2521Ϫ2529.
J.-C. G. Bünzli, C. Piguet, Chem. Rev. 2002, 102, 1897Ϫ1928.
´
[2]
[3]
1338
Eur. J. Inorg. Chem. 2003, 1332Ϫ1339

Lanthanide Chelates Based on Diethylenetriamine
FULL PAPER
J. R. Duffield, F. Marsicano, D. R. Williams, Polyhedron. 1991,
10, 1105Ϫ1111.
[4]
[20]
[21]
[22]
[23]
L. H. Slooff, A. van Blaaderen, A. Polman, G. A. Hebbink, S.
I. Klink, F. C. J. M. Van Veggel, D. N. Reinhoudt, J. W. Hof-
straat, J. Appl. Phys. 2002, 91, 3955Ϫ3980.
J.-C. G. Bünzli, C. Piguet, in Encylopedia of Materials: Science
H. J. Maria, S. P. McGlynn, J. Chem. Phys. 1970, 52,
[5]
3399Ϫ3402.
and Technology (Eds.: K. H. J. Buschow, R. W. Cahn, M. C.
Flemings, B. Ilschner, E. J. Kramer, S. Mahajan), Elsevier Sci-
ence Ltd, Oxford, 2001, 10, Ch. 1. 10. 4, 4465Ϫ76.
M. Latva, H. Takalo, V. M. Mukkala, C. Matachescu, J.-C.
Rodriguez-Ubis, J. Kankare, J. Lumin. 1997, 75, 149Ϫ169.
J.-C. G. Bünzli, in Lanthanide Probes in Life, Chemical and
Earth Sciences. Theory and Practice (Eds.: J.-C. G. Bünzli, G.
R. Choppin), Elsevier Science Publ. B. V., Amsterdam, 1989,
chapter 7, 219Ϫ93.
A. Beeby, I. M. Clarkson, R. S. Dickins, S. Faulkner, D. Parker,
L. Royle, A. S. de Sousa, J. A. G. Williams, M. Woods, J. Chem.
Soc., Perkin Trans. 2 1999, 493Ϫ504.
G. Schwarzenbach, Complexometric Titrations, Chapman &
Hall, London, 1957.
D. A. Quagliato, P. M. Andrae, E. M. Matelan, J. Org. Chem.
2000, 65, 5037Ϫ5042.
P. Gans, A. Sabatini, A. Vacca, J. Chem. Soc., Dalton Trans.
1985, 1195Ϫ1200.
N. Ingri, W. Kakolowicz, L. G. Sillen, B. Warnqvist, Talanta
1967, 14, 1261Ϫ1286.
H. Gampp, M. Maeder, C. J. Meyer, A. D. Zuberbühler, Tal-
anta 1986, 33, 943Ϫ951.
K. Mikkelsen, S. O. Nielsen, J. Phys. Chem. 1960, 64, 632Ϫ637.
S. R. Meech, D. C. Phillips, J. Photochem. 1983, 23, 193Ϫ217.
D. Magde, J. H. Brannon, T. L. Cremers, J. Olmsted, III, J.
Phys. Chem. 1979, 83, 696Ϫ699.
[6]
V. W. W. Ya m , K . K . W. L o, Coord. Chem. Rev. 1999, 184,
157Ϫ240.
[7]
G. Mathis, in Rare Earths (Eds.: R. Saez Puche, P. Caro), Edit-
[24]
orial Complutense, Madrid, 1998, 285Ϫ97.
[8]
D. Parker, Coord. Chem. Rev. 2000, 205, 109Ϫ130.
[9]
D. J. Bornhop, D. S. Hubbard, M. P. Houlne, C. Adair, G. E.
Kiefer, B. C. Pence, D. L. Morgan, Anal. Chem. 1999, 71,
[25]
[26]
[27]
[28]
[29]
2607Ϫ2615.
[10]
G. Zucchi, A.-C. Ferrand, R. Scopelliti, J.-C. G. Bünzli, Inorg.
Chem. 2002, 41, 2459Ϫ2465.
[11]
`
F. M. Ramirez, L. J. Charbonniere, G. Muller, R. Scopelliti, J.-
C. G. Bünzli, J. Chem. Soc., Dalton Trans. 2001, 3205Ϫ3213.
C. Piguet, J.-C. G. Bünzli, Chem. Soc. Rev. 1999, 28, 347Ϫ358.
M. Elhabiri, R. Scopelliti, J.-C. G. Bünzli, C. Piguet, J. Am.
Chem. Soc. 1999, 121, 10747Ϫ10762.
[12]
[13]
[14]
´
N. Fatin-Rouge, E. Toth, D. Perret, R. H. Backer, A. E. Mer-
bach, J.-C. G. Bünzli, J. Am. Chem. Soc. 2000, 122,
[30]
[31]
[32]
10810Ϫ10820.
[15]
[16]
J.-C. G. Bünzli, in Rare Earths (Eds.: R. Saez Puche, P. Caro),
Editorial Complutense, Madrid, 1998, 223Ϫ259.
W. Shrott, H. Büchner, U. E. Steiner, J. Inf. Recording 1996,
23, 79Ϫ84.
[33]
[34]
R. Rodriguez-Cortinas, F. Avecilla, C. Platas-Iglesias, D. Im-
bert, J.-C. G. Bünzli, A. de Blas, T. Rodriguez-Blas, Inorg.
Chem. 2002, 41, 5336Ϫ5349.
P. Letkeman, A. E. Martell, Inorg. Chem. 1979, 18, 1284Ϫ1289.
Received October 4, 2002
[17]
[18]
M. Bredol, U. Kynast, C. Ronda, Adv. Mat. 1991, 3, 361Ϫ367.
J. L. Sudmeier, C. N. Reilley, Anal. Chem. 1964, 36,
1698Ϫ1706.
K. N. Raymond, G. Muller, F. Matzanke, Top. Curr. Chem.
1984, 123, 49Ϫ102.
[19]
[I02550]
Eur. J. Inorg. Chem. 2003, 1332Ϫ1339
1339