Lanthanide podates with programmed intermolecular interactions: Luminescence enhancement through association with cyclodextrins and unusually large relaxivity of the gadolinium self-aggregates

Article abstract of DOI:10.1021/ja001089b

The synthesis of the phenyl anchored podand H₄L¹ fitted with four 3-carboxylate pyrazole arms and programmed for intermolecular interactions is reported, and its protonation constants are determined. Interaction with Ln³⁺ ions (Ln = La, Eu, Lu) in dilute aqueous solutions leads to complexes with 1:1 and 1:2 metal-ligand stoichiometry. The stability constants are in the range log β¹¹⁰ = 12.7-13.5 and log β¹²⁰ = 22.5-23.8 (pLn values in the range 9-10). The podates display a fair sensitization of the metal-centered luminescence with an absolute quantum yield of 5% in case of Tb^III. The average numbers of water molecules coordinated to the Ln^III ion amount to 3.8 and 4.9 for the 1:1 Eu and Tb podates, respectively, as determined by lifetime measurements. The addition of β- and γ-cyclodextrins to the 1:1 podates results in a strong association with the host, mainly through the phenyl anchor (Ln = Tb, log K₁₁ = 5-6 depending on the cyclodextrin) and, for the Tb complex, to a large increase in luminescence intensity at physiological pH. Self-aggregation of the podates occurs at concentration larger than 3 x 10^-5 M. This process is characterized by luminescence, using a pyrene probe, light-scattering measurements, and transmission electron microscopy. The novelty of the reported systems is their ability to self-aggregate into nanometric, rigid, and spherical particles in a controlled way (mean diameter: 10, 60, and ～300 nm), opening large perspectives for various applications. In particular, a record-high relaxivity (r₁ = 53 mM^-1 s^-1 at 20 MHz, T = 25°C) is observed for the aggregates of 1:2 Gd podate, which is not modified upon addition of an equimolar quantity of Zn^II.

Full text of DOI:10.1021/ja001089b

10810
J. Am. Chem. Soc. 2000, 122, 10810-10820
Lanthanide Podates with Programmed Intermolecular Interactions:
Luminescence Enhancement through Association with Cyclodextrins
and Unusually Large Relaxivity of the Gadolinium Self-Aggregates
Nicolas Fatin-Rouge,† EÄ va To´th,† Didier Perret,† Robin H. Backer,‡
Andre´ E. Merbach,† and Jean-Claude G. Bu1nzli*,†
Contribution from the Institute of Inorganic and Analytical Chemistry, UniVersity of Lausanne,
BCH, CH-1015 Lausanne, Switzerland, and the Institute of Physical Chemistry, Swiss Federal Institute of
Technology, CH-1015 Lausanne, Switzerland
ReceiVed March 28, 2000
Abstract: The synthesis of the phenyl anchored podand H4L1 fitted with four 3-carboxylate pyrazole arms
and programmed for intermolecular interactions is reported, and its protonation constants are determined.
Interaction with Ln3+ ions (Ln ) La, Eu, Lu) in dilute aqueous solutions leads to complexes with 1:1 and 1:2
metal-ligand stoichiometry. The stability constants are in the range log â110 ) 12.7-13.5 and log â120
)
22.5-23.8 (pLn values in the range 9-10). The podates display a fair sensitization of the metal-centered
luminescence with an absolute quantum yield of 5% in case of TbIII. The average numbers of water molecules
coordinated to the LnIII ion amount to 3.8 and 4.9 for the 1:1 Eu and Tb podates, respectively, as determined
by lifetime measurements. The addition of â- and γ-cyclodextrins to the 1:1 podates results in a strong association
with the host, mainly through the phenyl anchor (Ln ) Tb, log K11 ) 5-6 depending on the cyclodextrin)
and, for the Tb complex, to a large increase in luminescence intensity at physiological pH. Self-aggregation
of the podates occurs at concentration larger than 3 × 10-5 M. This process is characterized by luminescence,
using a pyrene probe, light-scattering measurements, and transmission electron microscopy. The novelty of
the reported systems is their ability to self-aggregate into nanometric, rigid, and spherical particles in a controlled
way (mean diameter: 10, 60, and ∼300 nm), opening large perspectives for various applications. In particular,
a record-high relaxivity (r1 ) 53 mM-1 s-1 at 20 MHz, T ) 25 °C) is observed for the aggregates of 1:2 Gd
podate, which is not modified upon addition of an equimolar quantity of ZnII.
Introduction
molecules increases approximately linearly with molecular
weight, researchers are devoting major efforts to the incorpora-
tion of gadolinium chelates in large structures such as polymers,5
dendrimers,6 proteins,7 and micelles.8 Several of these ap-
proaches are based on supramolecular chemistry9,10 and self-
organization.8 For instance, poly-â-cyclodextrins have been used
to bind Gd poly(aminocarboxylate) chelates bearing a lipophilic
phenyl tail9 and polymetallic Gd-containing edifices are self-
assembled through secondary recognition involving octahedral
transition metals.10
Developing strategies to synthesize supramolecular functional
nanostructures is a major goal in modern lanthanide chemistry.11
Association of monomers into reversible supramolecular struc-
tures can be realized by exploiting short-distance interactions
through molecular recognition based upon complementary size,
shape, and chemical functionalities.12 Hydrogen bonds, aromatic
π-stacking, and van der Waals interactions are valuable tools
Lanthanide(III) poly(aminocarboxylates) are widely used as
luminescent probes in fluoroimmunoassays1 and as contrast
agents for magnetic resonance imaging (MRI).2 Luminescent
stains have specific requirements regarding the environment of
the metal ion and the energy of the singlet and triplet states of
the ligating groups.3 Indeed, sensitization of the metal ion
luminescence is usually achieved through energy transfer from
the ligands since the oscillator strengths of the forbidden f-f
transitions are very faint. A considerable luminescence enhance-
ment can be obtained if the emitting ion is inserted in polymeric
structures containing nonluminescent ions, e.g., YIII.4 On the
other hand, the relaxivity of low molecular weight GdIII-
containing contrast agents is mainly limited by their fast
rotational motion, and since the relaxivity of sphere-shaped
* To whom correspondence should be addressed. E-mail:
jean-claude.bunzli@icma.unil.ch.
(5) Desser, T.; Rubin, D.; Muller, H.; Qing, F.; Khodor, S.; Zannazzi,
S.; Young, S.; Ladd, D.; Wellons, J.; Kellar, K.; Toner, J.; Snow, R. J.
Magn. Reson. Imaging 1994, 4, 467.
† University of Lausanne.
‡ Swiss Federal Institute of Technology.
(1) Bioanalytical Applications of Labeling Technologies, 2nd ed.; Hem-
mila¨, I., Stålberg, T., Mottram, P., Eds.; Wallac-Oy: Turku, 1995,
(2) Caravan, P.; Ellison, J. J.; McMurry, T. J.; Lauffer, R. B. Chem.
ReV. 1999, 99, 2293.
(3) Bu¨nzli, J.-C. G.; Ihringer, F.; Besanc¸on, F. In Calixarenes for
Separations; Lumetta, G., Rogers, R. D., Gopalan, A. C., Eds.; ACS
Symposium Series; American Chemical Society, Washington, DC, 2000;
Chapter 14.
(6) Tacke, J.; Adam, G.; Claben, H.; Muhler, A.; Prescher, A.; Gunther,
R. W. J. Magn. Reson. Imaging 1997, 7, 678.
(7) Lauffer, R. B.; Brady, T. J. J. Magn. Reson. Imaging 1985, 3, 11.
(8) Andre´, J. P.; To´th, E.; Fischer, H.; Seelig, A.; Maecke, H. R.;
Merbach, A. E. Chem. Eur. J. 1999, 5, 2977.
(9) Aime, S.; Botta, M.; Frullano, L.; Geninatti Crich, S.; Giovenzana,
G. B.; Pagliarin, R.; Palmisano, G.; Sisti M. Chem. Eur. J. 1999, 5, 1253.
(10) Jacques, V.; Hermann, M.; Humblet, V.; Sauvage, C.; Desreux, J.
F. Coord. Chem. ReV. 1999, 185, 451.
(4) Latva, M.; Takalo, H.; Simberg, K.; Kankare, J. J. Chem. Soc., Perkin
Trans. 2 1995, 995.
(11) Piguet, C.; Bu¨nzli, J.-C. G. Chem. Soc. ReV. 1999, 28, 347.
10.1021/ja001089b CCC: $19.00 © 2000 American Chemical Society
Published on Web 10/21/2000

Lanthanide Podates
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10811
to avoid Ln(OH)3 precipitation. The concentrations were determined
by complexometric titrations using a standardized Na2H2EDTA solution
in urotropine buffered medium and xylene orange as indicator. The
concentrations of stock solutions of cyclodextrins (CDs) were calculated
from the optical rotation (deg) of sodium light at 25.0 °C: permethylated
â-CD (âpm-CD) ) 160 ( 5, â-CD ) 162.5 ( 0.5, and γ-CD ) 177.5
( 0.5.20 Elemental analyses were performed by Dr. Eder, Microchemical
Laboratory, University of Geneva.
Chart 1
1D and 2D NMR data were collected on DPX-400 or AM-360
Bruker spectrometers. 1H NMR spectra were recorded in D2O or CH3-
OD and 13C spectra in D2O with acetonitrile as external reference.21
Assignment of H and 13C spectra was based on H coupling and 2D
1H-13C correlation spectra. Assignment of the signals for C11 and C12
was made using computation of electric charge density (AM1 package,
Hyperquem22). The mass spectra were recorded in electrospray ioniza-
tion mode on a VG BioQ quadrupole mass spectrometer (Tmatrix ) 100
°C). The instrument was calibrated using the horse myoglobin standard.
The analyses were conducted in both positive and negative modes. The
ion spray voltage was 4,6 kV, and CID voltage was 20-40 V.
Assignment of the species was based on the isotopic distribution of
the peaks. Electronic UV-visible spectra were recorded on a Perkin-
Elmer Lambda 7 spectrometer. Infrared spectra were measured on a
FT-IR Mattson Alpha Centauri spectrometer (4000-400 cm-1, KBr
pellets).
Syntheses. 3-Carboxylic Acid-Pyrazole. 3-Methylpyrazole (9.8
mL, 0.117 mol) was dissolved in 450 mL of water. To this solution,
40.9 g of aqueous KMnO4 (2.2 equi, 0.259 mol) was added slowly at
room temperature and the mixture was refluxed 4 h. The solution was
cooled, filtered, and evaporated to a small volume. Recrystallization
was performed in water at pH ∼2. After 1 h at 4 °C, white crystals
(yield 100%) were collected and washed with cold acidic water: Rf )
0.83 (CHCl3/CH3OH/25% aqueous NH3, 2/4/1); 1H NMR (D2O) δ 7.78
(d, J ) 2.4 Hz, 1H) and 6.89 (d, J ) 2.4 Hz, 1H).
1
1
for the supramolecular chemist, and since the associated
enthalpies are weak, the design of the monomers must be
focused on maximizing the reaction’s entropy.13 That is, the
monomers should be rigid enough to ensure good intermolecular
contact between interacting surfaces, and release of water upon
association of hydrophobic regions of the aggregating molecules
is needed to overcome the unfavorable loss of translational
entropy of the monomers.
Pursuing our efforts to design lanthanide-containing chelating
systems with potential applications as both luminescent stains
and contrast agents14 and taking into account the advantage of
large nanometric over molecular systems, we present here the
design of chelates able to associate through noncovalent
intermolecular interactions with a potential carrier such as a
cyclodextrin and/or to self-aggregate into larger entities. We
have turned our attention to podands built from a benzene anchor
and pyrazole-containing arms. The benzene ring is essentially
meant to protect the metal ion from outside interaction while
the pyrazole groups are known to provide a good antenna
effect.15-17 We have modified the poly(pyrazole) ligands
developed by Hartsorn and Steel18,19 for RuII complexation by
adding carboxylate groups at the 3-position of the pyrazole rings
because these groups are hydrophilic and they complex strongly
trivalent Ln ions. In this paper, we present the synthesis of ligand
H4L1, 1,2,4,5-tetrakis(pyrazol-1-ylmethyl-3-carboxylic acid)-
benzene (Chart 1), a study of its complexation ability with
trivalent Ln ions (Ln ) La, Eu, Lu) and of the luminescent
properties of the Eu and Tb podates alone or as host-guest
complexes with cyclodextrins (Tb). In addition, the formation
of nanometric self-aggregates is investigated by several experi-
mental techniques and its consequence on the relaxivity of the
Gd complexes is presented.
3-Ethyl Ester-Pyrazole. The ester was made under nitrogen flow
by refluxing for 5 h 1.02 g of the dried acid (9.1 × 10-3 mol) in 20
mL of absolute ethanol and 1.5 mL of concentrated sulfuric acid. The
solvent was then evaporated, and a yellow oil was obtained. Water
(50 mL) was added to dissolve the substance, and the solution was
neutralized with NaHCO3. The product was extracted with ethyl acetate,
the solvent was partially evaporated, and addition of a few milliliters
of petroleum ether initiated crystallization of a white material (85%
1
yield): Rf ) 0.53 (CH2Cl2/CH3OH, 95/5), H NMR (CDCl3) δ 7.62
(d, J ) 2.4 Hz, 1H), 6.87 (d, J ) 2.4 Hz, 1H), 4.42 (q, J ) 7.2 Hz,
2H), 1.40 (t, J ) 7.2 Hz, 3H).
1,2,4,5-Tetrakis(3-ethyl ester pyrazole-1-yl-methyl)benzene. 3-Eth-
yl ester-pyrazole (1.40 g, 0.01 mol) and sodium (0.276 g, 0.012 mol)
were reacted in dry ethanol. After dissolution, tetrakis(bromomethyl)-
benzene (1/4 equiv) in 30 mL of dry THF was added dropwise under
reflux; the mixture was further refluxed 1 h and evaporated to dryness.
Yield: 50%. 1H NMR ((CD3)2CO) δ 7.78 (d, J ) 2.4 Hz, 4H), 6.84 (d,
J ) 2.4 Hz, 4H), 6.72 (s, 2H), 5.59 (s, 8H), 4.30 (q, J ) 7.2 Hz, 8H),
1.32 (t, J ) 7.2 Hz, 12H); ESI-MS, [M + Na]+ 709.8 (calcd 709.7),
[M + K]+ 725.9 (calcd 725.8); IR (KBr, cm-1) 1720 (CdO) , 1480
(CdC) , 1378 (C-N), 2895 (C-H).
Experimental Section
Starting Materials and General Procedures. Solvents and chemi-
cals (Fluka AG) were of analytical grade and used without further
purification except tetrahydrofuran and dry ethanol, which were distilled
from CaH2 and CaO, respectively. Solutions were prepared just before
use with freshly boiled, doubly distilled water saturated with N2. Stock
solutions of LnX3‚nH2O (X ) ClO4, Cl, CF3SO3, AcO) were prepared
from lanthanide oxides (99.99%, Glucydur) and the corresponding acid.
They were systematically acidified with HCl to pH ∼4 before titration
Preparation of L1. The tetraester was hydrolyzed overnight with
KOH (1.2 equiv) in 40 mL of water at 50 °C. The yellow solution was
washed with ether (4 × 50 mL), and the aqueous phase was transferred
into centrifugation tubes, adjusted to pH ∼2 (concentrated HCl), and
cooled to 0 °C, and the precipitate was collected by centrifugation.
This procedure was repeated several times. The isolated ligand was
washed successively with cold water (100 mL), ether (80 mL),
methylene dichloride (50 mL), and acetone (50 mL). Yield: 47%. Rf )
(12) Lehn, J. M. Supramolecular Chemistry, Concepts and PerspectiVes;
VCH: Weinheim, 1995.
(13) Whitesides, G. M.; Mathias, J. P.; Seto C. T. Science 1991, 254,
1312.
(14) Zucchi, G.; Scopelliti, R.; Pittet, P.-A.; Bu¨nzli, J.-C. G.; Rogers, R.
D. J. Chem. Soc., Dalton Trans. 1999, 931.
(15) Reeves, Z.; Mann, K.; Jeffery, J.; McCleverty, J. J. Chem. Soc.,
Dalton Trans. 1999, 349.
1
0.77 (CHCl3/CH3OH/25% aqueous NH3, 4/9/1); H NMR (D2O, 20°
C) δ 5.4 (H1), 6.3 (H3, H6), 6.6 (H11), 7.4 (H12); 13C NMR (D2O, 20
°C) δ 52.3 (C7), 107.8 (C3, C6), 128.2 (C11), 132.4 (C12), 134.8 (C1,2
,
(16) Lawrence, R.; Jones, C.; Kresinski, R. Inorg. Chim. Acta 1999, 285,
283.
(17) Sanada, T.; Suzuki, T.; Yoshida, T.; Kaizaki, S. Inorg. Chem. 1998,
37, 4712.
(20) French, D.; Levine, M.; Pazur, J.; Norberg, E. J. Am. Chem. Soc.
1949, 71, 353.
(21) Gottieb, H.; Kotlyar, V.; Nudelman, A. J. Org. Chem. 1997, 62,
7513.
(18) Hartsorn, C.; Steel, P. Aust. J. Chem. 1995, 48, 1587.
(19) Hartsorn, C.; Steel, P. Angew. Chem., Int. Ed. Engl. 1996, 35, 2655.
(22) Dewar, M. J. S.; Zoebisch, E. V.; Healy, E. F.; Stewart J. J. P. J.
Am. Chem. Soc. 1985, 107, 3902.

10812 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Fatin-Rouge et al.
C4,5), 149.1 (C10), 169.7 (C13); ESI-MS. [M + Na]+ 597.8 (calcd 597.5),
[M + K]+ 613.9 (calcd 613.6); IR (KBr, cm-1) 3407 br (aryl COOH),
2533 br (overtone, aryl COOH), 1712 (CdO(OH)), 1509 (CdC), 1379
(C-N), 1060 (C-O). Elemental Anal. Found: C, 53.60; H, 4.68; N,
17.49. Calcd for C26H22N8O8‚(CH3)2CO‚H2O: C, 53.52; H, 4.65; N,
17.23. Acetone was indeed found in the 1H NMR spectrum and
integration accounted for six protons.
Solutions for Mass Spectrometry Analysis (ESI-MS). The Eu
complexes were prepared in pure methanol, mixing Eu(CF3SO3)3 and
the ligand in stoichiometric ratios at 10-4 M. [Eu(L1-H)]2+. m/z ) 364.1
(calcd 364.04); [Eu(L1-4H)2(H2O)]5-, m/z ) 262.7 (calcd 262.8); [Eu-
(L1 - 4H)2(MeOH)]5-, m/z ) 264.9 (calcd 265.0).
Physicochemical Characterization of the Monomeric Podates.
Potentiometric Titrations. Titrations of H4L1 or Ln3+ alone or 1:1
mixtures ([L1]t ) [Ln]t ) 1.0 × 10-3 M) were carried out using 5-mL
sample solutions in a thermostated (20.0 ( 0.1 °C) glass-jacketed vessel
in an atmosphere of Ar. The ionic strength was fixed with KCl (µ )
0.1 M). Reactants were mixed and acidified to pH ∼2 (HCl) at least
30 min before titration. Titrations were carried out with an automatic
Metrohm Titrino 736 GP potentiometer (resolution 0.1 mV, accuracy
0.2 mV) using a constant-volume addition (0.05 mL) program and
linked to an IBM PS/2 computer. An automatic buret (Metrohm
6.3013.210) 10 mL, accuracy 0.03 mL, was used along with a Metrohm
6.0238.000 glass electrode. The standard base (KOH 0.010 M) or acid
(HCl 0.010 M) (µ ) 0.1 M, KCl) was added through a capillary tip
inside the solution attached to the automatic buret. The data (130 points
per curve, drift <1 mV/mn) were mathematically treated by the program
SUPERQUAD23 using a Marquardt algorithm while the distribution
of species was calculated with the program Haltafall.24 Calibration of
the pH-meter and the electrode system was performed prior to each
measurement using a standardized HCl solution (µ ) 0.1 M, KCl) at
20.0 °C. The ion product of water (pKw ) 13.98) and electrode potential
were refined using the program Scientist by Micromath (version 2.0).
The stepwise protonation constants of the deprotonated ligand [L1]4-
are defined as âi ) [HiL1](4-i)-]/[Hi-1L1](5-i)-‚[H+]i, and the overall
equilibrium constants describing all metal-ligand-proton species,
MiLjHk, are equal to âijk ) [MiLjHk]/[M]i[L]j[H+]k.
Preparation of (H3O)[LnL1]‚solv Podates (Ln ) La, Eu, Gd, Tb,
Lu). Freshly titrated LnCl3.nH2O solutions (∼0.025 M, ∼1,2 mL, ∼0.03
mmol) were added slowly, at room temperature, to an aqueous solution
of K3HL1 (∼25 mg, ∼0.03 mmol, 4 mL, pH 7.1 (adjusted with HCl)).
When turbidity appeared, the pH was adjusted to 7.1 with 0.1 M KOH;
the solution was then transferred into a centrifugation flask, cooled at
4 °C for 2 h, and centrifuged. The collected material was washed with
methanol (4 mL) to remove the unreacted ligand and sonicated. The
precipitation procedure was repeated three times. The solid was dried,
washed with 20 mL of acetone on a sinter filter, and further dried for
1.5 days at 65 °C and 0.01 Torr. Yield: ∼60%.
(H3O)[LaL1]: IR (KBr, cm-1) 3411 br (OH), 1580 and 1371 (νas,
νs COO), 1514, 1485, 1420, 1299, 1192, 1060, 770. Elemental Anal.
Found: C, 41.66; H, 4.03; N, 14.20. Calcd for (H3O)[LaL1]‚2CH3OH‚
H2O, C28H31N8O12La: C, 41.48; H, 3.86; N, 13.83.
(H3O)[EuL1]: IR (KBr, cm-1) 3415 br (OH), 1590 and 1371 (νas,
νs COO), 1519, 1486, 1420, 1299, 1192, 1057, 770. Elemental Anal.
Found: C, 41.08; H, 3.57; N, 14.03. Calcd for (H3O)[EuL1]‚CH3OH‚
2H2O, C27H29N8O12Eu: C, 40.00; H, 3.61; N, 13.83.
(H3O)[GdL1]: IR (KBr, cm-1) 3411 br (OH), 1590 and 1371 (νas,
νs COO), 1519, 1486, 1420, 1299, 1192, 1058, 770 cm-1. Elemental
Anal. Found: C, 40.04; H, 3.42; N, 13.68. Calcd for (H3O)[GdL1]‚
3H2O, C26H27N8O12Eu: C, 38.95; H, 3.40; N, 13.98.
(H3O)[TbL1]: IR (KBr, cm-1) 3422 br (OH), 1601 and 1371 (νas,
νs COO)IR νmax, 1520, 1487, 1420, 1300, 1199, 1061, 773. Elemental
Anal. Found: C, 40.95; H, 3.72; N, 14.30. Calcd for (H3O)[TbL1]‚2CH3-
OH, C28H29N8O11Tb: C, 41.37; H, 3.60; N, 13.79.
(H3O)[LuL1]: IR (KBr, cm-1) 3430 br (OH), 1627 and 1369 (νas,
νs COO), 1507, 1486, 1421, 1295, 1192, 1065, 768 cm-1. Elemental
Anal. Found: C, 38.38; H, 3.30; N, 13.30. Calcd for (H3O)[LuL1]‚
3H2O, C26H27N8O12Lu: C, 38.14; H, 3.33; N, 13.69.
Preparation of K5[Ln(L1)2]‚solv (Ln ) La, Eu, Tb, Lu) and
(H3O)5[Gd(L1)2]‚solv. Freshly titrated LnCl3‚nH2O solutions (∼0.025
M, ∼0.6 mL, ∼0.015 mmol) were added slowly at room temperature
to an aqueous solution of K3HL1 (∼25 mg, ∼0.03 mmol, 4 mL, pH
9.1) or H4L1 (15.1 mg, 2.61 × 10-5 mol). The pH was adjusted to 8.0
with 0.1 M KOH (7.2 for Gd); the solution was partially evaporated
with N2 and centrifuged. For Gd, the solution was left 1 day at room
temperature, transferred in a centrifugation flask, methanol (∼1 mL)
was added, and precipitation of the complex was started with a few
drops of ether; a turbidity appeared and the solution was cooled 2 h at
4 °C and centrifuged. The collected materials were washed on a sinter
filter with acetone (∼100 mL) and then methanol (∼10 mL). The solids
were dried for 1.5 day at 65 °C and 0.01 Torr. Yield: ∼70%.
K5[La(L1)2]: IR (KBr, cm-1) 3418 br (OH), 1592 and 1371 (νas, νs
COO), 1514, 1482, 1414, 1304, 1192, 1058, 812, 770 cm-1. Elemental
Anal. Found: C, 38.72; H, 3.55; N, 13.44. Calcd for K5 [La(L1)2]‚2CH3-
OH‚7H2O, C54H58N16O25K5La: C, 38.94; H, 3.51; N, 13.46.
K5[Eu(L1)2]: IR (KBr, cm-1) 3420 br (OH), 1594 and 1371 (νas, νs
COO), 1514, 1478, 1412, 1302, 1192, 1054, 810, 770. Elemental Anal.
Found: C, 37.25; H, 3.31; N, 13.02. Calcd for K5 [Eu(L1)2]‚10H2O,
C52H56N16O26K5Eu: C, 37.41; H, 3.38; N, 13.43.
Spectrophotometric Titrations. Electronic spectra were recorded
with a Perkin-Elmer Lamdba 7 spectrophotometer at 20.0 °C (210-
500 nm, 100 nm mn-1 scan speed, spectral width 1 nm) using 1-cm
Suprasil quartz cells. Titrations of L1 (1.9 × 10-5 M, 15.0 mL, [HCl]tit
) 0.050 M, µ ) 0.1 (KCl) or 3.5 × 10-5 M, 19.7 mL, [LaCl3]tit
)
1.09 × 10-3 M, pH 7.0 (Pipes buffer 10-3 M), µ ) 0.1 M (KCl)) were
performed in a thermostated (20.0 ( 0.1 °C) glass-jacketed vessel in
an atmosphere of Ar. Aliquots of titrant were added using a Socorex
micropipet. The pH values of the titrated solutions were controlled
continuously. Under these conditions, equilibrium was attained within
10 min, as checked by kinetic measurements. After this delay, 3 mL
of solution was transferred into the quartz cell with a Teflon syringe.
No evidence of aggregation, e.g., turbidity, was detected along the
titration. The stability constants (apparent for La3+ complexation or
true for H+ complexation) were computed using the Specfit program.25
Differences between the measured and the computed absorbances were
less than 0.004 at any wavelength.
NMR Titrations. Ligand, 10-3-10-2 M solutions, was prepared in
2 mL of D2O. Sample solutions with different pD values were prepared
by adding dilute KOD/D2O or dilute D2SO4/D2O (Fluka, puriss). The
pH values of the solutions were determined with a Metrohm Titrino
736 GP potentiometer equipped with a calibrated Metrohm 6.0234.100
glass electrode. Measurements of the potential of a borate buffer solution
either in H2O or in D2O allowed us to convert to pD values by the
equation, pD ) pHmeas + 0.17. Ionic strength was not adjusted.
Luminescence Titrations. [LnL1 ](4i-3)- (5 × 10-6 M; V0 ) 10 mL;
i
Ln ) Tb, Gd; i ) 1, pH 7.2; i ) 2, pH 9.4) were titrated with 0.01 M
aqueous solutions of âpm-, â-, and γ-CDs using the same procedure
as described for potentiometric and spectrophotometric titrations.
Equilibrium was reached within 10 min delay as checked by kinetic
measurements. After this delay, 3 mL of solution was transferred into
a quartz cell and degassed for 5 min with Ar. Absorbance at the
K5[Tb(L1)2]: IR (KBr, cm-1) 3438 br (OH), 1622 and 1371 (νas, νs
COO), 1505, 1421, 1299, 1201, 829, 770. Elemental Anal. Found: C,
37.49; H, 3.38; N, 13.00. Calcd for K5 [Tb(L1)2]‚10H2O, C52H56N16O26K5-
Tb: C, 37.27; H, 3.37; N, 13.38.
K5[Lu(L1)2]: IR (KBr, cm-1) 3438 br (OH), 1624 and 1367 (νas, νs
COO), 1569, 1506, 1421, 1299, 1201, 1190, 1070, 836, 772. Elemental
Anal. Found: C, 35.66; H, 3.34; N, 12.58. Calcd for K6[Lu(L1)2(OH)]‚
10H2O, C52H57N16O27K6Lu: C, 35.74; H, 3.29; N, 12.83.
(H3O)5[Gd(L1)2]; IR (KBr, cm-1) 3422 br (OH), 1599 and 1371
(νas, νs COO), 1521, 1487, 1421, 1299, 1201, 829, 770. Elemental Anal.
Found: C, 44.02; H, 4.22; N, 14.34. Calcd for (H3O)5[Gd(L1)2]‚(CH3)2-
CO‚4H2O, C55H65N16O26Gd: C, 43.33; H, 4.30; N, 14.71.
excitation wavelength was measured (A < 0.05), and emission (λexc
)
260 nm) and excitation spectra were collected. Each titration was made
twice. Ionic strength was not adjusted.
(23) Gans, P.; Sabatini, A.; Vacca, A. J. Chem. Soc., Dalton Trans. 1985,
1195.
(24) Ingri, N.; Kakolowicz, W.; Sillen, L. G.; Warnqvist, B. Talanta 1967,
14, 1261-1286.
(25) Gampp, H.; Maeder, M.; Meyer, C.; Zuberbu¨hler, A. Talanta 1986,
33, 943.

Lanthanide Podates
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10813
Luminescence Measurements. Low-resolution luminescence mea-
surements (spectra and lifetimes) were recorded on a Perkin-Elmer LS-
50B spectrofluorometer. Phosphorescence lifetimes (τ) were measured
with the instrument in time-resolved mode, on frozen ethanol solutions
put into a quartz capillary or a 1-cm Suprasil cell at room temperature.
They are the average of at least three independent measurements which
were made by monitoring the decay at the maximums of the emission
spectra after a 0.03-0.04-ms delay time. The signals were usually
analyzed as single-exponential decays using the program FLDM
(Perkin-Elmer). Solutions of [Ln(L1)i](4i-3)- were prepared in water at
pH 7.0, using stock solutions of H4L1 (2.0 × 10-3 M) and Ln(ClO4)3
(Ln ) La, Eu, Gd, Tb, Lu) (∼4.5 × 10-3 M). Frozen solutions in
ethanol (0.5-1) × 10-5 M were made using dried solutions of
complexes prepared previously in water and then dissolved in ethanol
(fluorescence quality). Quantum yields of [L1]4- (5 × 10-6 M, pH 13.5)
and [Ln(L1)i](4i-3)- (Ln ) La, Gd, Lu, (0.5-1) × 10-5 M in H2O, pH
7) were determined in degassed water relative to quinine sulfate in
0.05 M aqueous H2SO4 (absolute quantum yield 0.54626); estimated
error, 10%. Metal-centered luminescence of [Ln(L1)i](4i-3)- complexes
(Ln ) Eu, Tb; (0.5-1) × 10-5 M in H2O, pH 7) was determined relative
to degassed solutions of [Eu(terpy)3](ClO4)3 or [Tb(terpy)3](ClO4)3 10-3
M in acetonitrile (terpy ) 2,2′:6′,2′′-terpyridine, absolute quantum yields
1.3.27 and 4.8%,28 respectively); the estimated error is 30%. The number
of coordinated water molecules (q) for the Eu podates was calculated
istics of the particles and aggregates were determined on gray level
images (AnalySIS, Soft Imaging Systems). To ensure consistent results,
the scanned areas (0.64 µm2) were divided into subframes (0.04-0.16
µm2) and the background (collodion/carbon film) was subtracted from
each individual entity. The average net gray value of the objects was
converted into mean particle thickness by calibration with 15-30-nm
spherical colloids. By this method, the maximum local thickness <5%
of the analyzed entities was underestimated (some region of the entity
obviously fell outside the maximum black level). The mechanism of
aggregation was determined by fitting the volume (V
)
calc agg of the entities
f
as a function of ((Rg)agg/(a)unit
) with a least-squares Levenberg-
Marquardt algorithm (Origin; Microcal), where (Rg)agg is the radius of
gyration of the entity, (a)unit is the radius of its constituting elementary
colloids, and f is the fractal dimension.31,32 The error on the final result
was estimated by maximizing the inaccuracies at each step of the
measurement.
Electrophoretic Mobility (µe). The value of µe and the zeta-potential
(ú ) for the {Gd(L1)2}n
self-aggregated particles were determined
5n-
using a Zetasizer IV (Malvern Instruments) electrophoretic light-
scattering apparatus, previously calibrated with anionic latexes. Mea-
surements were made at 20.0 ( 0.1 °C with an aqueous solution of
complex ([L1]t ) 1.1 × 10-2 M and [GdCl3]t ) 5 × 10-3 M at pH 7.4)
sufficient to produce a photon counting rate of at least 2 × 106 counts/
s; µe and ú data reported are averages of five independent analyses.
Calculation of ú from measured µe was obtained from the built-in
software supplied with the instrument. The model used in the calculation
was the Smoluchowski limit (κa . 1) for which the general relationship
between ú and µe reduces to the simple expression: µe ) ꢀú/η, where
ꢀ is the dielectric constant of the medium, η is the dynamic viscosity,
κ the inverse double-layer thickness and a is the hydrated particle radius.
The surface charge density (σ0) of the aggregates was evaluated from
the dimension-less surface potential (Ψ0), the potential Ψ at the distance
x from the surface and ú, according to the following equations:33
from q ) n(τ-1 - τ-1D), where τH and τD are the lifetimes in H2O
H
and D2O, respectively, and n ) 1.05.29
Physicochemical Characterization of the Self-aggregated Par-
ticles. Determination of Critical Aggregation Concentrations (cac).
A stock solution of [L1]t ) 3.6 × 10-3 M and [GdCl3]t ) 1.5 × 10-3
M at pH 9.2 was added to known volumes of water at the same pH.
The solutions were allowed to equilibrate at room temperature for 20
min, and the pH was adjusted with 0.1 M KOH and HCl solutions.
For fluorescence measurements, the initial volume of water was
saturated with pyrene (∼10-7 M) and the samples were degassed with
argon. The emission spectra of pyrene were recorded from 350 to 500
nm (λexc ) 337 nm, excitation and emission band-path 5 nm, scan rate
50 nm min-1). Two scans were accumulated for each run and were
corrected before the ratio of intensities of the first (363 nm) and third
(382 nm) vibrational peaks, I1/I3, was calculated. Dynamic light-
scattering measurements were performed at 22.0 ( 0.1 °C in a toluene
bath at a scattering angle of 90°, using a He-Ne laser light-scattering
instrument operating at 632.8 nm. The intensity of the light scattered
by toluene was used as reference. Diffusion coefficients of particles in
H2O were determined from an exponential fit to the correlation curve.
Hydrodynamic radii were calculated from measured diffusion co-
efficients by means of the Stokes-Einstein equation.
Transmission Electron Microscopy (TEM) of the Gd Self-
Aggregates. Collodion-covered (<10 nm), carbon-coated (<10 nm)
TEM gold grids were used as specimen holders. Collection of particles
was achieved in two ways: (a) by ultracentrifugation30 of a 1.1 × 10-4
M solution of the 1:2 GdIII complex at pH 9.2. After centrifugation,
the supernatant liquid was carefully withdrawn and the grids were
immediately rinsed with a drop of ultrapure water to avoid formation
of KCl microcrystals. (b) By depositing 2 µL of solution of the 1:2
(1.1 × 10-4 M, pH 9.2) or 1:1 (4 × 10-5 M, pH 7.0) Gd podate on the
gold grid, washed with a drop of water. The grids were observed in a
Philips CM 12 electron microscope using a 40-µm aperture at an
acceleration of 80 keV. Spectral analysis was made by energy dispersion
spectroscopy (EDS) with an EDAX DX-4 probe. The aggregation
regime of {Gd(L1)25-}n into particles was assessed by image-fractal
analysis. Electron micrographs were digitized (2000 lpi, 8-bits depth;
intensity and contrast kept constant), and the morphometric character-
σ ) 2 2ꢀn k T sin(Ψ /2)
x
0
i
B
0
e(Ψ/2)-1
(1 + eΨ/2)e-κx
úez
kBT
Ψ0 ) 4 arctan
,
with Ψ )
(
)
κ-1
)
ꢀkBT/e2 zini
∑
x
i
where ꢀ is the dielectric constant, ni the number of ions per cubic meter,
zi their charge, kB Boltzman’s constant, and T the absolute temperature.
Proton Relaxivity Measurements. The nuclear magnetic relaxation
dispersion (NMRD) profiles have been measured at 5, 25. and 75 °C
for [Gd(L1 )]5- (1.5 × 10-3 M, pH 7.3) on a Stelar Spinmaster FFC
2
fast-field cycling NMR relaxometer and on a Bruker electromagnet,
covering a continuum of magnetic fields from 7 × 10-4 to 1.41 T
(corresponding to a proton Larmor frequency range 0.03-60 MHz).
For [GdL1]- (4 × 10-5 M; pH 7.3), the water relaxivity has been
measured at 25 °C and 20 MHz only. The transverse electron spin
relaxation rate has been obtained from the measured peak-to-peak EPR
line widths of the derivative EPR spectrum, recorded on a Bruker ESP
300E spectrometer (X-band; 0.34 T).
Results and Discussion
Protonation Constants of [L1]4-. Protonation constants were
determined by spectrophotometry (Figure F1, Supporting In-
formation) and were checked with pH potentiometry. The first
protonation constant was easily obtained because substantial
changes occur between pH 8 and 13 in the carboxylate
(26) Meech, S. R.; Philips, D. J. J. Photochem. 1983, 23, 193.
(27) Martin, N.; Bu¨nzli, J.-C. G.; McKee, V.; Piguet, C.; Hopfgartner,
G. Inorg. Chem. 1998, 37, 577.
(28) Charbonnie`re, L. J.; Balsiger, C.; Schenk, K. L.; Bu¨nzli, J. C. B. J.
Chem. Soc., Dalton Trans. 1998, 505.
(29) Lanthanide Probes in Life, Chemical and Earth Sciences; Bu¨nzli,
J. C. G., Choppin G. R., Eds.; Elsevier Science Publ. B. V.: Amsterdam,
1989.
(31) The Fractal Geometry of Nature; Mandelbrot, B. B., Ed.; Free-
man: San Francisco, 1982.
(32) Lin, M. Y.; Lindsay, H. M.; Weitz, D. A.; Ball, R. C.; Klein, R.;
Meakin, P. Nature 1989, 339, 360.
(30) Lienemann, C. P.; Heissenberger, A.; Leppard, G. G.; Perret, D.
Aquat. Microb. Ecol. 1998, 14, 205.
(33) James, A. M. Electrophoresis of Particles in Suspensions; Plenum
Press: New York, 1979; Vol. 11, p 121ff.

10814 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Fatin-Rouge et al.
Figure 1. 1H NMR spectrum of [L1]4- at pD ) 12.9 and variations of
δ vs pD; T ) 20 °C.
chromophore zone of the UV-spectra around 210 nm. Two other
bands were monitored during the titration, one at 230 nm (ꢀ ∼
30 000 M-1cm-1), assigned to a π f π* transition in the
pyrazole chromophore,15 and the other at 305 nm (ꢀ ≈ 3300
M-1cm-1), attributed to a n f π* transition in view of its low
intensity. The extracted pKa values amount to 2.0(3), 3.2(2),
4.6(1), 12.3(1) (µ ) 0.1 M, KCl; 20 °C). The titration was also
Figure 2. Computed distribution curves for the Eu/H4L1 system: (- - -)
[EuL1]-; (- -) [Eu(L1)2]5-; (1) EuIIIaq; (2) {L1}4-; (3) (HL1)3-; (4)
(H2L1)2-; (5) (H3L1)-; (6) H4L1; (7) EuOH2+; (8) Eu(OH)3. Top: [L1]t
) [EuIII]t ) 1.0 × 10-3 M. Bottom: [L1]t ) 2[EuIII]t ) 2.0 × 10-3 M.
T ) 20.0 °C; µ ) 0.1 M (KCl).
Table 1. Stability Constants log âijk ((2 σ) Obtained for
Complexes LniLjHk from pH Potentiometry. [L1]t ) [LnCl3]t )1.0
× 10-3 M; µ ) 0.1 M (KCl); T ) 20.0˚ C
1
followed by H NMR to get more insight into the protonation
mechanism (Figure 1). No variation in chemical shifts is detected
for the first proton association while substantial variations (up
to 0.5 ppm) occur for phenyl protons upon the second and third
protonation processes. Interestingly, both the 1H and 13C NMR
spectra display only one set of signals during the titration,
somewhat broadened in acidic media, and reflecting symmetric
structures and/or the presence of fast equilibria between dif-
ferently protonated species. For methylene protons directly
attached to carboxylate groups, a 0.2 ppm shift is usually
expected while protonation of linear aminocarboxylates leads
to a chemical shift 5-fold larger when the nitrogen atom is
protonated.34 In the case of (L1)4-, the nearest protons are
located far from the carboxylate groups. Comparing the pKa
values obtained with those of 1-methylpyrazole (pKa ) 3.635),
benzoic acid (pKa ) 4.2), pyridine (pKa ∼ 5.4), and dipicolinic
acid (pKa1 ) 1.4 and pKa2 ) 5.3) allows us to draw the following
conclusions. Since spectrophotometric data indicate that the first
protonation occurs on carboxylate groups, the unexpectedly large
binding constant obtained for this first proton association (pKa4
) 12.33) can only be explained by invoking synergetic
interactions in the ligand involving more than one arm in the
proton binding. Moreover, the pKa values expected for pyrazole
nitrogen and carboxylate functions being on the same order of
magnitude and large 1H NMR chemicals shifts being associated
with the second and third protonation, we assume that nitrogen
atoms are involved in these protonation steps. The close
proximity of the pyrazole and carboxylic functions on each arm
probably induces a fast proton exchange between these two
functions (dentate effect), as observed in bipyridyl.36
i
j
k
La3+
Eu3+
Lu3+
1
1
1
1
1
1
0
0
1
1
1
2
-1
-3
0
1
2
-7.90(4)
-24.92(4)
12.7(2)
18.5(2)
22.4(2)
-7.47(4)
-22.96(4)
13.4(2)
19.3(2)
23.1(2)
-7.35(5)
-21.78(4)
12.8(2)
18.9(2)
22.7(2)
0
23.2(2)
23.8(2)
22.5(2)
mixtures and of the corresponding Ln3+ ion alone (Table 1,
Figure F2, Supporting Information). The model used to fit
the titration data was fairly complex, with six different
[LniHkLj](3i-4j+k)+ species: two hydroxide moieties (i ) 1, j )
0, k ) 1, 3), one unprotonated (k ) 0) and two protonated 1:1
complexes (i ) 1, j ) 1, k ) 1, 2), and one unprotonated 1:2
complex (i ) 1, j ) 2, k ) 0). The speciation diagram for Eu
is shown on Figure 2. In the case of La3+, a spectrophotometric
titration (Figure F3, Supporting Information) using a noncom-
plexing buffer ([Pipes] ) 1.0 × 10-3 M)37 was conducted at
fixed pH ) 7.00 ( 0.02, leading to two conditional stability
constants: Log â110 ) 7.3(2); Log â120 ) 12.5(4). Transforma-
tion of these apparent stability constants into true constants by
using the ligand pKa’s values and assuming the presence of
nonprotonated complexes only resulted in a good agreement
with potentiometric data. Ligand-metal stoichiometry used in
the modeling of pH potentiometric data were suggested from
spectrophotometric titrations and ESI-MS data on the Eu system.
The podates are therefore quite stable, as demonstrated by the
pLn values calculated with [Ln]t ) 10-6 M, [H4L1]t ) 10-5 M,
and pH ) 7.4: 9.3 for La, 10.0 for Eu, and 9.0 for Lu, as
compared to 15.4 for [Eu(edta)]- 38 for instance. Monitoring
the titration of L1 with Ln3+ ions in D2O (pD ) 7) by 1H NMR
revealed that the largest chemical shift variations occur when
Interaction of HiL1 with Trivalent Ln Ions in Dilute
Solution. The stability constants of the complexes formed
between Ln3+ ions (Ln ) La, Eu, Lu) and HiL1 have been
determined by potentiometric titration of 1:1 ligand-metal
(34) Letkeman, P.; Martell, A. Inorg. Chem. 1979, 18, 1284.
(35) Barszcz, B.; Lenarcik, B. Pol. J. Chem. 1989, 63, 371.
(36) Linnell, R. H.; Kaczmarczyk, A. J. Phys. Chem. 1961, 65, 1196.
(37) Yu, Q.; Kandegedara, A.; Xu, Y.; Rorabacher, D. Anal. Biochem.
1997, 253, 50.
(38) Galaktionov, Y.; Astaklov, K. Zh. Neorg. Khim. 1963, 8, 460.

Lanthanide Podates
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10815
1
3
Table 2. Energy E of the ππ* and ππ* States and Lifetime (τ)
Table 3. Relative Quantum Yields (Qrel). [L1]4- (pH 13) and 1:1,
1:2 Complexes ((0.5-1) × 10-5 M) with Ln ) La, Eu, Gd, Tb, and
Lu in H2O at pH 7
of the ππ* State of Ligand L1 Alone or in the Ln 1:1 and 1:2
Podates, As Measured from Frozen Solutions in Ethanol (77 K)
3
E/cm-1 a
compound
Qrel
λ
exc/nm
compound
Qrel
λ
exc/nm
[Lu(L1)2]5-
[Eu(terpy)3]3+
[Eu(L1)]-
0.054
1
3.75
0.50
1
1.04
0.62
0.67
312
437
230
230
409
230
230
254
compound
[L1]4-
τ /ms (2σ)
1ππ*
3ππ*
QSO4
1
0.11
0.029
0.046
0.019
317
317
317
317
317
317
312
[L1]4-
[LaL1]-
0.10 (2)
33 715, 29 720
27 770, 27 095, 26 295,
25 545, 24 865,
24 270, 23 450
25 865
[LaL1 ]5-
[Eu(L1)2]5-
[Tb(terpy)3]3+
[TbL1]-
2
[LaL1]-
0.128 (4)
0.104 (2)
0.45 (6)
33 610
33 610
[GdL1]-
[La(L1)2]5-
[GdL1]-
25 865
[Gd(L1)2]5- 0.041
33 645, 30 795, 28 985 27 625, 25 905,
[LuL1]-
0.035
[Tb(L1)2]5-
25 240, 24 650
[Gd(L1)2]5-
[LuL1]-
0.15 (4)
1.1 (4)
33 645, 28 210
33 785, 30 635, 28 945 27 640, 25 760, 25 070
33 785, 30 635, 28 390 24 020
33 510, 30 940, 29 795 24 840
33 510, 30 940, 28 555 25 445
23 665
[Lu(L1)2]5-
[EuL1]-
0.44 (8)
0.50 (6)
0.71 (4)
0.63 (2)
0.84 (4)
intensity of which decreases with increasing protonation of the
ligand. No triplet-state emission is observed. Upon complex-
ation, a 1000-cm-1 shift to lower energies occurs with increasing
atomic number of the ion (Figure F5, Supporting Information).
The luminescence intensity is pH dependent; for [GdL1]- for
instance, it goes through two maximums at pH 4 and 7. Ligand-
centered luminescence quantum yields are reported in Table 3.
They are respectively 2 and 3 times smaller in 1:2 and 1:1
podates compared to the value for the free ligand.
Metal-centered luminescence of dilute aqueous solutions is
also pH dependent: 1:1 podates (Ln ) Eu, Tb) display emission
between pH 3 and 8.5, with maximum intensity occurring again
at pH 4 and 7. In aqueous solution, EuIII and TbIII are equally
sensitized, as shown by the absolute quantum yields of the 1:1
podates found to be equal to 4.9 and 5.0%, respectively. This
points to a significantly improved energy transfer from the ligand
to the Eu(5D0) level in water over that found in frozen ethanol
solutions (Figure F5, Supporting Information). We also note
that the quantum yields of the 1:2 podates are about half the
values found for the 1:1 complexes. Using the empirical equation
proposed by Horrocks et al.41 and the experimental lifetimes at
room temperature for solutions of [LnL1]- in H2O (pH ) 7.6)
and D2O (pD ) 8.0), we determine the following average
numbers of water molecules bonded in the inner coordination
sphere of the metal ion: q ) 3.8 ( 1.0 and 4.9 ( 0.4 for the
monomeric 1:1 podates with Eu and Tb, respectively.
[Eu(L1)2]5-
[TbL1]-
33 510, 30 720, 27 880
33 510, 30 720, 29 700
b
b
[Tb(L1)2]5-
a The most intense component is italicized. b Not detected due to L1
f Tb3+ energy transfer.
the stoichiometry changes from 1:2 to 1:1, especially for benzyl
and phenyl protons; this indicates that the metal ion is lying
closer to the anchor in the 1:1 podates. The stability constants
do not vary much from one Ln ion to the other. We have
modeled [GdL1]- in a vacuum with the program Cerius 2. The
minimized structure points to π interactions between two
neighboring pyrazole rings, which affect their orientation and
consequently the binding of the N atoms to the metal ion. Single
crystals of adequate quality could not be obtained for structural
characterization.
Photophysical Properties of the Monomeric Podates. The
photophysical properties of H4L1 and of its 1:1 and 1:2
monomeric podates with La, Eu, Gd, Tb, and Lu in dilute frozen
ethanolic solutions (77 K) are summarized in Table 2 (see Figure
F4, Supporting Information). UV excitation in the π f π* and
n f π* absorption bands of the pyrazole units results in ligand-
centered emission from L1 displaying two structured and
overlapping bands assigned to emission from the 1ππ* and 3ππ*
states. The 0-phonon component of the singlet-state emission
is located at 33 750 cm-1 while the corresponding transition
from the triplet state occurs at 27 740 cm-1. The overlap
Luminescence Enhancement of the 1:1 Monomeric Po-
dates through Association With Cyclodextrins. Molecular
modeling showed that partial inclusion of the 1:1 podates should
occur in â-, âpm-, and γ-cyclodextrins. In the case of â-CD,
1H NMR spectra of aqueous solutions demonstrate a deshielding
of the aromatic anchor protons only, meaning that solely the
anchor is included in the macrocyclic cavity. Moreover, the
NMR data suggest a 1:1 association between the 1:1 podates
and the cyclodextrin. Metal-centered emission increases con-
siderably upon addition of CD, and the most spectacular
luminescence enhancement is obtained with Tb (Figure F6,
Supporting Information). Titrations conducted in both phos-
phorescence and fluorescence modes (Figure 3 and Figure F6,
Supporting Information) allowed us to estimate the apparent
association constants of the Tb 1:1 podate with CDs in the
following way.
1
3
between the ligand ππ* and ππ* states induces an efficient
3
intersystem crossing process. The ππ* transition is not much
altered in the 1:1 and 1:2 La podates, but complexation to Gd3+
and Lu3+ induces a broadening of the 3ππ* emission band which
has a relatively short (0.1 ms) lifetime. For the Eu and Tb
podates, metal-centered emission is detected upon excitation in
the π f π* absorption band of the pyrazole units, but not when
the samples are excited through the n f π* state. As expected
3
from the relatively high energy of the ππ* 0-phonon compo-
nent, Eu3+ is only weakly sensitized, particularly in the 1:1
podate, whereas good energy transfer occurs between the ligand
3ππ* and Tb(5D4) states. This is in agreement with what has
been observed for Tb poly(aminocarboxylates) for which the
more efficient antenna effect was observed for 3ππ* states with
energy around 26 000 cm-1 39 and with the golden rule that the
The metal-centered phosphorescence intensity of [TbL1]-
without CD can be written as
energy gap ∆E(3ππ* - D4) should be larger than 3500 cm-1
5
for efficient and irreversible energy transfer.40
The room-temperature emission spectrum of (L1)4- in aque-
ous solution displays two bands centered at 26 350 and 23 050
cm-1 assigned as arising from the 1nπ* and 1ππ* states,
respectively, which are progressively blue-shifted and the
I0(λ) ) C100[TbL1]
0
where C10 is a constant that incorporates the quantum yield
and the absorptivity of the complex in the absence of CD. In
(39) Latva, M.; Takalo, H.; Mukkala, V. M.; Matachescu, C.; Rodriguez-
Ubis, J. C.; Kankare, J. J. Lumin. 1997, 75, 149.
(40) Steemers, F. J.; Verboom, W.; Reinhoudt, D. N.; van der Tool, E.
B.; Verhoeven, J. W. J. Am. Chem. Soc. 1995, 117, 9408.
the presence of CD, the formation of a host-guest adduct leads
(41) Horrocks, W.; De, W., Jr.; Sudnick, D. R., Jr. J. Am. Chem. Soc.
1979, 101, 334.

10816 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Fatin-Rouge et al.
Figure 3. Titration of [TbL1]- 5 × 10-6 M in H2O by â-CD. Top:
emission spectra at [â-CD] ) (0, 1.2, 2.5, 6.2, and 8.7) × 10-6 M.
Bottom: increase of the phosphorescence intensity at 550 nm vs [â-CD].
λexc ) 260 nm; pH ) 7.30 ( 0.03.
Figure 4. Top: I1/I3 ratio of pyrene fluorescence vs [Gd(L1)2]5-
computed concentration (arrows point to the observed breaks). Bot-
tom: computed distributions of the various L1 species, pH ) 9.2.
enhancement ratio C11/C10 decreases with the polarity of the
CD; e.g., â-CD < γ-CD.
Table 4. Association Constants (K11) for 1:1 Complexes between
[L1Tb]- and CDs and Phosphorescence Enhancement Ratio
(C11/C010) Obtained from Luminescence Titrations. [L1]t ) [TbCl3]t
) 5 × 10-6 M; pH 7.30 ( 0.03; T ) 20.0 ˚C
Self-Aggregation of the 1:2 Podates. As expected from the
design of the encapsulating podand, turbidity due to self-
aggregation appears at relatively low concentration, ∼10-4 M.
The aggregation behavior of the [La(L1)2]5- podate was
investigated by 1H NMR spectroscopy in D2O at pD ) 7.5. All
the protons become shielded upon increasing the concentration
(Figure F7 Supporting Information), the largest chemical shifts
being obtained for the benzene protons, which can be attributed
to an intermolecular aromatic ring current taking place in the
aggregates.44 The shielding of the benzene protons as the
concentration increases points to a stacking between hydropho-
bic anchors contributing to the aggregation process. We believe
that the formation of self-aggregates originates from weak
intermolecular interactions and not from strong metal-ligand
bonds inducing polymerization, mainly because the stacking
interaction observed between phenylene anchors would not be
possible in a polymeric network. Moreover, no oligomers were
detected by ESI-MS in methanolic solutions of the Eu podates.
On the other hand, the smaller chemical shift experienced by
the pyrazole protons means that these moieties participate less
than expected in the association process.
host
Log K11 ((2σ)
C11/C0
10
â-CD
âpm-CD
γ-CD
6.0 (2)
5.1 (2)
5.3 (1)
4.2 (4)
4.2 (2)
2.7 (1)
to a global phosphorescence intensity I(λ) at a given wavelength:
I(λ)) C10[TbL1] + C11[TbL1 ⊂ CD]
0
For each titration, we were able to show that C10 ) C10, so
that combining the previous expressions leads to
1 + (C11/C010)K11[CD]
I(λ)/I0(λ) )
,
with
K11
1 + K11[CD]
[TbL1 ⊂ CD]
)
[TbL1][CD]
The aggregates were characterized by three different experi-
mental techniques. First, the intensity ratio between the first
and third vibrational components of the pyrene fluorescence
band at 363 and 382 nm, I1/I3, was used as polarity probe.45
Upon addition of [Gd(L1)2]5- to pyrene-saturated water, two
breaks are observed in the I1/I3 ratio at [Gd(L1)2]5- ) (8.0 (
0.3) × 10-7 and (3.2 ( 0.2) × 10-5 M (Figure 4). The first
one reflects the rise in the concentration of the 1:2 species as
shown by the speciation curves computed from the stability
constants. The second break points to the beginning of the
aggregation process. Opalescence of the solution starts at ∼9
× 10-5 M and reflects the formation of higher aggregates.
The association constants K11 and the phosphorescence en-
hancement ratios (C11/C10) upon CD complexation were intro-
duced as fitting parameters in the nonlinear regression of the
experimental I(λ)/I0(λ) values as a function of CD concentration.
The binding constants reported in Table 4 are ∼3 orders of
magnitude larger compared to association constants between
â-CD and transition metal complexes with macrocycles42 or
cyclopentadienyl ligands.43 The magnitude of K11 increases with
the hydrophobic character of CDs, meaning that hydrophobic
interactions and the concomitant release of solvent molecules
represent important favorable factors in the driving force leading
to the association. We note indeed that the phosphorescence
(44) Haigh, C. W.; Mallion, R. B. Prog. Nucl. Magn. Reson. Spectrosc.
1980, 13, 303.
(42) Raj, C. R.; Ramaraj, R. Electrochemica Acta 1999, 44, 2685.
(43) Wang, Y.; Mendoza, S.; Kaifer, A. E. Inorg. Chem. 1998, 37, 317.
(45) Winnik, F. M. Regismond, S. T. A. Colloid Surf. A: Physicochem.
Eng. Aspects 1996, 118, 1.

Lanthanide Podates
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10817
Figure 7. Pseudo-3D transmission electron micrograph of the
{[Gd(L1)2]5-}n particles and aggregates collected on the Au grid after
deposition of 2 µL of the 1.1 × 10-4 M solution on a gold grid and
washing with a drop of pure water. The pseudo-3D effect was obtained
by applying an asymmetric 3 × 3 matrix filter emphasizing the
topographic contrast of the original negative. The arrows point to typical
classes of particles: 20-30 (A), 70-90 (B), and >100 nm (C).
Figure 5. Plots of the correlation curve baseline values as a function
of [Gd(L1)2]5-, at 22 °C, pH 9.2 for the determination of critical
aggregation concentrations. Top: first part of the curve. Bottom: entire
curve.
where n0 and n stand for the refractive indices of the solvent
and the solution, respectively, NA is the Avogadro number, λ is
the wavelength of the incident light, and ∆I90 is the Rayleigh
diffusion ratio of the solution in excess to that of a solution
with a concentration equal to cac-1; dn/dc was calculated from
refractive index measurements and found to be 0.39. Extrapola-
tion of the curve to c ) 0 yields the inverse of the aggregate
molecular weight Magg from which the average aggregation
number Nagg can be estimated. In the case of [Gd(L)2]5-, this
number corresponds to ∼1050 monomers (Magg ≈ 1.8 × 106
g‚mol-1).
Finally, information on the shape and size of the aggregated
particles was obtained from transmission electron microscopy
of centrifuged or deposited particles. The particles revealed to
be spherical (Figure 7 and Figure F8, Supporting Information)
and two mean diameters could be directly determined by this
technique, 20 and 60 nm. The particles are amorphous, as
revealed by diffraction experiments, and they have a large
electron density reflecting a high gadolinium content but are
potassium free, as shown by energy dispersive spectroscopy
(Figure 8 and Figure F9, Supporting Information). The very
similar elemental composition of the two kinds of spherical
particles can probably be traced back to a second nucleation
phenomenon leading to larger aggregates from the 20-nm ones.
There are also a few much larger aggregates with a size of ∼300
nm, similar to those observed in solution. Counting the particles
Figure 6. Determination of the average mass of the {[Gd(L1)2]5-
aggregates (Magg) in H2O at pH 9.2, using a plot of Kc/∆I90 (see text)
vs the monomer molality c.
}
n
Second, the critical aggregation concentrations were determined
by measuring the concentration dependence of the light-
scattering ratio Isol/Itoluene. A plot of the correlation curve baseline
value, M(base), computed from these data (Figure 5) presents
two discontinuities at 3.5 × 10-5 M (cac-1), in good agreement
with fluorescence measurements, and at 10-4 M (cac-2). The
presence of a second cac indicates a change in the nucleation
process of monomeric [Gd(L1)2]5-. Analysis of the correlation
curves reveals the presence of three types of particles with the
following mean hydrodynamic radii: (i) 11 ( 1 nm, which
account for all the diffusion in the concentration range between
cac-1 and cac-2, (ii) 57 ( 15 nm, which become the major
particles in solution at concentrations larger than cac-2, and (iii)
280 ( 80 nm corresponding to the coagulation of the previous
particles; the proportion of these large particles increases with
increasing concentration until 5 × 10-4 M, where they are the
more numerous particles in solution and stay constant up to
1.55 × 10-3 M, the highest concentration investigated.
lying on a 14-µm2 surface, after deposition of a 1.2 × 10-4
M
solution of [Gd(L1)2]5- on a gold grid and evaporation of the
solvent, led to the following statistics: 68% of 15-45-nm
particles, 25% of 60-nm particles, and 7% of larger particles.
Smaller aggregates could not be accurately discriminated from
their background (collodion-carbon film) under our experi-
mental conditions, although light-scattering measurements in-
dicate the presence of 10-nm entities. The larger share of the
bigger particles can be explained by the experimental conditions
favoring their deposition and making them more resistant to
washing of the grid with water.
The average aggregation number Nagg of the 1:2 podates was
estimated from a plot of Kc/∆I90 as a function of the solute
concentration c (Figure 6). In this expression, K is defined as
follows for the vertically polarized incident light:46
(46) Taboada, P.; Attwood, D.; Ruso, J. M.; Sarmiento, F.; Mosquera,
V. Langmuir 1999, 15, 2022.
K ) 4π2n0 (dn/dc)2/ΝAλ4
2

10818 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Fatin-Rouge et al.
Figure 9. Plot of the volume of {[Gd(L1)2]5-]n particles and aggregates
vs the ratio (radius of gyration of aggregate)/(radius of their elementary
units), as determined by image analysis on electron micrographs (128
entities). The inset shows the raw data; the power-law fitting was
calculated on the Log-Log representation to minimize the statistical
weight of the large (Rg)agg/(a)unit on the results.
Figure 8. Energy dispersion spectrum of 20-nm spherical particles of
{[Gd(L1)2]5-}. The Au signal originates from the gold supporting grid.
Cu and Zn signals originate from pieces of the microscope.
Statistical fractal analysis of the deposited larger self-
aggregates was used to determine the main kinetic regime of
their formation from the smaller particles. The kinetics of colloid
aggregation is related to the structure and morphology of the
particles and clusters formed.31,47,48 Namely, the mass magg or
volume Vagg of the aggregates can be described in function of
their morphological features:
f
magg Vagg ((Rg)agg/(a)unit
)
with the exponent f representing the fractal dimension of the
aggregates, i.e., the self-similarity of their structure.31 When
elementary colloids with radius (a)unit collide to produce an
aggregate with a radius of gyration (Rg)agg, their aggregation
rate may be controlled either by Brownian motion (diffusion-
limited colloid aggregation, DLCA) or by electrostatic repulsive
forces (reaction-limited colloid aggregation, RLCA). The DLCA
mechanism is induced by attractive van der Waals forces and
is usually an irreversible process, while RLCA requires that
elementary colloids overcome their repulsive barrier to stay in
close contact. The fractal dimension allows one to differentiate
between these two regimes: f ≈ 1.8 for DCLA and f ≈ 2.1 for
RLCA. While the aggregates can be analyzed by light-scattering
techniques at a bulk level, electron microscopy provides an
insight into the characteristics of the individual entities. Figure
9 shows the results of the fractal analysis performed on TEM
images of individual self-aggregates. A Log-Log plot of the
equation above yields f ) 2.21(4), which is in good agreement
with values reported for reaction-limited colloid aggregation.32,48
The radius of the elementary particles within the aggregates
was taken as (a)unit ) 8.25 nm (Øunit ) 16.5 nm). The fractal
dimension was checked to be insensitive to the size of the
elementary particles for (a)unit ) 5 (light scattering), 8.25 (TEM),
and 15 nm (arbitrary size). Moreover, to test the statistical
relevance of the calculated fractal dimension of our system, a
fit was enforced with f ) 1.8 (representative of DLCA), but
this slope produced a significantly larger ø2 value. Finally, fits
under boundary limits taking into account the propagation of
Figure 10. NMRD profiles of {[GdL1 ]}nn5- at 5 (2), 25 (9), and 75
°C (b); pH 7.2 (pipes).
2
{Gd(L1)25-}n entities can be explained in terms of the structure
of the constitutive podates, which is highly charged while
bearing a hydrophobic moiety, the aromatic anchor. Stacking
of the podates via their anchor, as suggested by the NMR shifts,
allows the formed entities to minimize their electrostatic
repulsion and the total entropy of the growing self-aggregates.
The ill-defined morphology of the aggregates analyzed by TEM
confirms the absence of crystallinity of the species and suggests
that stacking must proceed by an isotropic scheme.
The formation of spherical {Gd(L1)25-}n self-aggregates with
10-nm diameter and their subsequent coagulation into larger
structures suggests that the particles are perhaps less negatively
charged than the formal -5n value, because electric repulsion
could otherwise be larger than the forces responsible for
coagulation. In an effort to estimate the effective charge of the
particles, we have measured their electrophoretic mobility µe
and extracted the ú potential. The resulting ú curves obtained
for {Gd(L1)2}5n- particles in water at pH 7.4 display a peak,
centered at -31.9 (5) mV, indicating negatively charged
aggregates and good colloidal stabilization because of electric
repulsion. The surface charge density (σ0) was evaluated to be
6.14 × 10-3 C/m2. This estimation is 10-20 times smaller than
expected. Charge compensation therefore occurs, which could
be due to the presence of cations such as H3O+ or K+ in the
mother liquor during the synthesis. This is supported by the
microanalyses which effectively indicate the presence of these
counterions (see Experimental Section).
experimental errors on (Vcalc agg and (Rg)agg yielded fractal
)
dimensions comprised between 2.13(4) (Vmax (Rg/a)fmin) and
2.34(6) (Vmin (Rg/a)fmax), which defines the variability range
of the actual f value. The reaction-limited aggregation of the
Water Proton Relaxation of the Self-Aggregated 1:2 Gd
Podate. The proton relaxivity of a solution of Gd 1:2 podate is
reported on Figure 10. The high-field peak in the range 20-60
MHz is an indication of a slow rotational correlation time and
of a high rigidity of the system. This can only be explained by
(47) Aggregation and Fractal Aggregates; Jullien, R., Botet, R., Eds.;
World Scientific: Singapore, 1987.
(48) Julien, R. New J. Chem. 1990, 14, 239.

Lanthanide Podates
J. Am. Chem. Soc., Vol. 122, No. 44, 2000 10819
Table 5. Relaxivity of the Most Interesting Gadolinium-Containing
dependent, while displaying always a high relaxivity (Figure
10), pointing to aggregation being temperature dependent.
Finally, to test the potentiality of the described system to be
used in in vivo analyses,50 we have added an equimolar quantity
of ZnII (with respect to GdIII) and found that after 1 day the
relaxivity remains unchanged, meaning that the transition metal
ions does not replace GdIII inside the self-aggregated particles.
Proton relaxivity is determined by numerous parameters,
involving correlation times for rotation, water exchange and
electronic relaxation, the number of inner-sphere water mol-
ecules, the 1H-Gd distance (inner-sphere relaxation term), and
the diffusion correlation time (outer-sphere relaxation term). The
Contrast Agents for MRI Compared to That of the L1 Podates
ligand
r1 per Gd/mM-1 s-1
conditions
DTPA
DTPA-BMA
DOTA
linear polymers
dendrimeric compounds
natural polymer-based
macromolecular agents
covalent conjugates
noncovalent protein
adducts
4.3a
25 °C; 20 MHz
25 °C; 20 MHz
25 °C; 20 MHz
37 °C; 20 MHz
37 °C; 20 MHz
37 °C; 10/20 MHz
4.39a
4.2a
6-18a
11-36a
7-14a
up to 40-50a,b
37 °C; 20 MHz
4.5% HAS
25 °C; 20 MHz
25 °C; 20 MHz
5n-
transverse electron spin relaxation rate of the {Gd(L1)2}n
{[Gd(L1)2]}n
53
5-n
system has been estimated from X-band EPR measurements to
be around 9 × 109 s-1 (B ) 0.34 T). Unfortunately, due to
solubility problems, 17O NMR measurements could not be
performed to determined the water exchange rate. Furthermore,
{Gd(L1)2}n5n- self-aggregates are quite complex with differently
sized nanometric particles. As a consequence, reliable fits of
{[GdL1]}n
18c
n-
a From ref 2. b Observed values in 4.5% HSA. c Observed value. The
relaxivity has been determined in a solution that initially contained
87.5% [GdL1]- and 12.5% [Gd(L1)2]5- according to the speciation. The
computed relaxivity for pure {[GdL1]}nn- amounts to r1 ) 13 mM-1s-1
.
the NMRD profiles could not be generated. The lower relaxivity
the presence of nanometric particles in solution obtained through
monomer aggregation, as demonstrated above. Proton relaxivity
measured at 10 MHz slightly decreases (5-10%) upon decreas-
ing the [Gd(L1)2]5- concentration from 1.5 × 10-3 to 1.5 ×
10-4 M, suggesting that the particle size is somewhat affected
by concentration changes, as pointed out by light-scattering
measurements. Characteristic peaks in longitudinal proton
relaxivity profiles have been observed for superparamagnetic
species.49 However, this possibility can be excluded in the
present case since the GdIII ions in these nanometric particles
are located too far from each other to generate magnetic
interactions leading to superparamagnetic effect. Furthermore,
the relaxivity peaks found for superparamagnetic systems are
usually much broader and their maximum is rather at lower
frequencies (around 10 MHz). Consequently, the only explana-
tion for the presence of the high-field peak in the NMRD profiles
of the Gd podate is the slow rotation of the aggregates.
The astounding result is that solutions of the 1:2 Gd podate
display a very large relaxivity, which amounts to r1 ) 53 mM-1
s-1 at 25° C and 20 MHz; that is 12.6 times larger than the
widely used [Gd(DOTA)]- contrast agent and larger than the
relaxivity of adducts between contrast agents and proteins2
(Table 5). In the case of contrast agents obtained from
noncovalent association between proteins and Gd chelates, the
measured values of the water relaxivity are not higher than 40-
50 mM-1 s-1 and the association constants of the adducts are
not always very high; thus the adducts are not fully formed.
This means that the measured relaxivity is lower than the
calculated bound relaxivity. Moreover, because of the high
molecular weight of these proteins, the measured relaxivity of
millimolar solutions of the Gd complexes is often overestimated
by up to 25% because their molality is in fact larger.5 Interest
obtained for the {GdL1}n
aggregates (Table 5) can be
n-
explained partially from the TEM experiments on deposited
aggregates on gold grids which reveal that the size of the
particles is smaller than 130 nm in any case, most of them being
in the 20-40-nm range.
Conclusion
We have synthesized a predisposed podand H4L1 able to
encapsulate trivalent lanthanide ions and to generate secondary
supramolecular structures through hydrophobic and weak mul-
tiple interactions. The resulting 1:1 and 1:2 lanthanide podates
display a good stability under physiological conditions with,
for instance, a computed pEu ) 10. Investigations of photo-
physical properties of the lanthanide complexes reveal that
pyrazole units provide a relatively efficient energy transfer to
Eu(5D0) and Tb(5D4) excited states and act therefore as good
energy harvesters; a quantum yield of 5% is obtained in the
case of the [TbL1]-. Addition of â-CD to solutions of the 1:1
podates leads to strong association and to a more than a 4-fold
increase in the metal-centered luminescence (quantum yield
∼20%). Due to the flexible design of the podand H4L1,
3
improvement of its antenna effect shifting the ππ* state to
somewhat lower energy and by increasing the molar absorbance
of the podand appears to be within reach, for instance by adding
a chromophoric residue on the 5-position of the pyrazole units.
Secondary supramolecular interactions occur in solution and
the [Ln(L1)2]5- podates aggregate into spherical, rigid, but
porous nanoparticles of low dispersivity, with 10- and 60-nm
diameter, which further coagulate into nonsymmetrical larger
structures (∼280-300 nm). Generation of aggregates at ∼10-3
M is a fast process (t1/2 in the range of seconds at 20 °C), and
colloidal solutions are stable over weeks at room temperature,
without showing sedimentation. Nanosized spherical particles
are obtained instead of microcrystals because in the conditions
used the growth of the nanometric supramolecular edifices is
too fast to generate a regular stacking between hydrophobic
surfaces. Fractal analysis of TEM-generated images of the
particles points to a pure reaction-limited regime generating the
larger self-aggregates under the experimental conditions used.
This regime occurs when the repulsive forces between the
particles are the limiting factor, inducing the formation of
compact aggregates, reasonably rigid although porous. It is
in larger particles with contrast agent ability, such as those
generated in {Gd(L1)2}n
solutions, stems from their large
5n-
retention time in blood, a prerequisite for angiographic applica-
tions. Another advantage of {Gd(L1)2}n
over the existing
5n-
contrast agent systems lies in the fact that the rigid, but porous,
nanometric particles it generates display at 25 °C a large and
almost constant relaxivity in the frequency range 30-50 MHz,
which is important for biomedical applications. Many of the
known supramolecular systems have a sharper relaxivity peak
around 20 MHz followed by an abrupt drop at higher frequen-
cies. We also note that the NMRD profiles are temperature
(49) Bulte, J. W. M.; Brooks, R. A.; Moskowitz, B. M.; Bryant, L. H.;
Frank, J. A. Magn. Reson. Med. 1999, 42, 379.
(50) Bu¨nzli, J.-C. G.; Fatin-Rouge, N. United States Provisional Patent
Application USSN 60/170,311, Dec 13, 1999.

10820 J. Am. Chem. Soc., Vol. 122, No. 44, 2000
Fatin-Rouge et al.
known that other kinetic regimes of aggregation can be obtained
by changing the experimental conditions (ionic strength, con-
centration, temperature), which opens the way for the design
of size-controlled nanoparticles with enhanced physicochemical
properties.
analysis or for the luminescent mapping of biomolecules seems
to be quite promising.
Acknowledgment. We thank Dr P. Bowen for helpful
discussion on the aggregation processes. This research is
supported through grants by the Swiss National Science
Foundation. A.E.M. thanks the Swiss Federal Office for
Education and Research for a grant (COST Action D8) and
J.-C.G. B. thanks the Fondation Herbette (Lausanne) for the
gift of spectroscopic equipment.
The porous supramolecular structure of the self-aggregates
allows free circulation of water inside the edifice, which explains
why solutions of the Gd 1:2 podate display the largest relaxivity
ever reported for a gadolinium chelate, more than 1 order of
magnitude larger than the relaxivity of the first-generation
contrast agents. Two problems associated with the supermol-
ecules designed as MRI contrast agents9,10 are the presence of
internal motions reducing their rigidity and leading to a lower
relaxivity than expected and the large amount of nondirectly
active substance needed for a MRI measurement. The system
described here avoids both disadvantages and points to potential
applications in various fields of medical diagnosis, including
angyology and in vivo temperature mapping. Several clues show
that control of the coagulation process of the spherical particles,
that is, of their molecular mass, size, and shape, is possible,32
which leads us to foresee further improvement in the properties
of these systems, for instance with respect to the ability to
control their relaxivity reversibly, depending on the size and
number of the nanoparticles in solution. In conclusion, the new
supramolecular methodology we are developing to synthesize
both contrast agents for MRI and luminescent stains for medical
Supporting Information Available: Figures showing the
spectrophotometric titration of [L1]4- with H3O+ and of [HL1]3-
by La3+, and the potentiometric titration of H4L1 and (H4L1 +
Eu) with KOH. Figures displaying the low-resolution and time-
resolved spectra of the ligand-centered and metal-centered
luminescence in frozen ethanol and water, figure showing the
influence on the Tb-luminescence of â-CD addition to aqueous
[TbL1]- solution, figure displaying the variation of the 1H NMR
shifts upon concentration of [La(L1)2]5-, transmission electron
micrographs of {[Gd(L1)2]5-
} particles, energy disperson
n
spectrum of 60-nm spherical particles of {[Gd(L1)2]5-}n (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA001089B