Protonation and lanthanide(iii) complexation equilibria of a new tripodal polyaza-polycatechol-amine

Article abstract of DOI:10.1039/b822945g

The sulfonated tripodal polyaza-polycatechol-amine ligand tris(2,3-hydroxy-5-sulfobenzylamine) ethyl)amine (STRENCAT) was synthesized and its protonation constants determined by potentiometry at 25 °C and in 0.1 mol dm^-3 NaClO₄. In t

Full text of DOI:10.1039/b822945g

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Protonation and lanthanide(III) complexation equilibria of a new tripodal
polyaza-polycatechol-amine†
P. Di Bernardo,*a P. L. Zanonato,a A. Bismondo,b A. Melchiorc and M. Tolazzi*c
Received 19th December 2008, Accepted 2nd March 2009
First published as an Advance Article on the web 6th April 2009
DOI: 10.1039/b822945g
The sulfonated tripodal polyaza-polycatechol-amine ligand tris(2,3-hydroxy-5-sulfobenzylamine)
ethyl)◦amine (STRENCAT) was synthesized and its protonation constants determined by potentiometry
at 25 C and in 0.1 mol dm-3 NaClO4. In the adopted experimental conditions (2.5 < p[H+] < 11.5), six
of the virtual thirteen protonation constants of STRENCAT were determined. The deprotonation
sequence of STRENCAT, identified by UV-vis and NMR studies, is characterized by the initial loss of
three catechol protons, between p[H+] 5 and 8, followed by the loss of three secondary ammonium
protons at higher p[H+] (9 < p[H+] < 11.5). No deprotonation of the three, more basic, catechol sites of
STRENCAT was observed even at the highest p[H+] studied. Complexation measurements run on
lanthanide(III)–STRENCAT systems (Ln = La, Gd and Yb), in the same ionic medium and
temperature, show that the ligand (partly protonated or completely deprotonated) is able to form
soluble 1 : 1 and 1 : 2 metal–ligand complexes over the whole p[H+] range investigated. Due to the high
affinity of lanthanide(III) for catecholate units, even at p[H+] near 10 and before ammonium
deprotonation, STRENCAT forms tri-chelate complexes, in which the metal ions are hosted in the cage
formed by the six catechol oxygen atoms, completely deprotonated. Deprotonation of secondary
ammonium groups, although not directly involved in complexation, increases the stability of the 1 : 1
complexes. They prevail in solution at p[H+] > 10, even in excess of ligand, in all the metal–ligand
systems.
donors are the most frequently studied,4,8 due to the known
Introduction
oxophilicity of lanthanide(III) ions. Phenol-based ligands are also
The coordination chemistry of multidentate ligands with lan-
oxygen donors, but have received much less attention, mainly
thanide(III) ions is receiving growing interest because of the
because of the prevailing hydrolysis reactions of lanthanide(III)
ions at p[H+] > 7–8 where the phenol group (pKa ~ 10)9 begins
numerous and interesting applications of the related compounds,
such as innovative materials,1 luminescence probes in biology and
to release protons. For the same reason, complexes of lanthanides
medicine,2 catalysts for DNA and RNA cleavage,3 or relaxation
agents in magnetic resonance imaging (MRI).4,5
with amines and polyamines cannot be studied in water.10 Not even
the increase in the number of hydroxide groups in the benzene ring
of phenol, to give the chelating 1,2-dihydroxybenzene (catechol)
and some parent compounds, seems strongly to increase the ligand
affinities for lanthanide(III) ions.11
Among the lanthanide(III) ions, the gadolinium(III) cation is
particularly well suited as a contrast agent in diagnostic medical
MRI, due to its high magnetic moment and favourable electronic
relaxation rate.4,6,7 To be useful as a contrast agent, Gd(III) com-
plexes must possess very high thermodynamic stability and kinetic
inertness, which can be achieved only with suitably designed
ligands.4,5 Among these, poly-carboxylate and phosponate oxygen
Therefore, simple ligands containing only amine and phenol
or catechol groups are not expected to be effective ligands
for lanthanide chelation. Nevertheless, it has been shown that
the introduction of phenol groups into a well pre-organized
ligand frame, such as the tripodal amino moiety in tris((2-
hydroxy-5-sulfobenzylamino)ethyl)amine (TRNS, in Scheme 1),
allows ligands to bond lanthanides(III) strongly.12,13 In addition,
the substitution of phenols with catechols in the derivatives
of tripodal polyamines tris((2,3-dihydroxybenzylamino)ethyl)-
amine and (cis,cis-1,3,5-tris((2,3-dihydroxybenzylamino)amino-
ethyl)cyclohexane (TRENCAT and TMACHCAT, respectively,
in Scheme 1) results in significant enhancement of the complexing
ability of the ligands.14,15
aDipartimento di Scienze Chimiche, Universita` di Padova, Via Marzolo 1,
35131, Padova, Italy. E-mail: plinio.dibernardo@unipd.it; Fax: +39-049-
8275161; Tel: +39-049-8275213
bIstituto di Chimica Inorganica e delle Superfici del CNR, Corso Stati Uniti
4, 35127, Padova, Italy
cDipartimento di Scienze e Tecnologie Chimiche, Universita` di Udine, Via
del Cotonificio 108, 33100, Udine, Italy. E-mail: marilena.tolazzi@uniud.it;
Fax: +39-0432-558803; Tel: +39-0432-558852
† Electronic supplementary information (ESI) available: Tables reporting
stepwise stability constants for successive deprotonations of LnL3-;
experimental details for potentiometric titrations; figures reporting elec-
tronic UV-vis spectra for the titrations of the ligand alone or in the
presence of Gd(III); potentiometric titration curves and speciation for
all lanthanides(III)/ligand systems in 1 : 1 and 2 : 1 molar ratios; ESI-
MS spectra of solutions with Gd(III)/ligand in 1 : 1 ratio. See DOI:
10.1039/b822945g
In this context, sulfonated tripodal polyaza-polycatechol-amine
tris(2,3-hydroxy-5-sulfobenzylamine)ethyl)amine (STRENCAT,
Scheme 1) has been synthesized, starting from its parent com-
pound TRENCAT. This ligand was synthesized in order to test to
what degree, sulfonation of the cathechol groups of TRENCAT
4236 | Dalton Trans., 2009, 4236–4244
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Scheme 1
Table 1 Overall and stepwise protonation constants of STRENCAT (L6-)
at 25◦ and I = 0.1 mol dm-3 NaClO4
can influence the complex solubility and eventually give access to
speciation study in the basic p[H+] range, never before investigated
with the other similar ligands in Scheme 1.13–15 In addition,
given the known affinity of Fe(III) ion for this class of ligands,
STRENCAT may also be considered a promising sequestering
agent for therapeutic use in iron-overload diseases.16,17
This paper reports the results of a potentiometric study,
undertaken in order to elucidate both the acid/base characteristics
of the ligand and its affinity toward La(III), Gd(III) and Yb(III)
ions, to check the effect of the increase in the metal-ion charge-to-
radius ratio on complex stabilities. UV-vis and NMR experiments
were also carried out to identify the ligand protonation sequence
and groups involved in metal coordination. Complementary
theoretical calculations were run, to obtain structure and energy
information about ligand protonation and its lanthanide com-
plexes.
Reaction
logbn
logKn
H+ + L6- = HL5-
2H+ + L6- = H2L4-
3H+ + L6- = H3L3-
4H+ + L6- = H4L2-
5H+ + L6- = H5L-
6H+ + L6- = H6L
11.42 (0.04)
21.95 (0.03)
31.75 (0.03)
39.57 (0.03)
46.40 (0.03)
52.09 (0.04)
11.42 (0.04)
10.53 (0.05)
9.80 (0.04)
7.82 (0.04)
6.83 (0.04)
5.69 (0.05)
of ligand after neutralization of excess of mineral acid) suggests
that STRENCAT has three, relatively acidic, functional groups,
which dissociate in the p[H+] range 6–9. The curve differentiation
at p[H+] > 9 also shows that the ligand has an unknown
number of less acidic groups which deprotonate above this p[H+].
STRENCAT has 13 potentially ionizable groups, but three of them
are sulfonic acid moieties, which have pKas < 1,9 and the tertiary
amine group, is exceedingly acidic.12,13 Therefore, in preliminary
treatment of the potentiometric data, up to nine protonation
constants were tested to fit the experimental results. Invariably, the
best fit was achieved with six protonation equilibria, characterized
by the logbn (n = 1–6) reported in Table 1. Considering the
poorly ionizable second phenol of catechol and sulfonated parent
molecules of catechol,9,11 only one of the two acidic functions of
catechols in the ligand could probably be deprotonated in our
experimental conditions. Accordingly, in the following discussion,
neutral STRENCAT, shown in Scheme 1, is called H6L and L6-
is the ligand with the three secondary amines and only one OH
function per catechol, deprotonated.
2. Results and discussion
Protonation studies
Although the experimental set-up was designed in order to
follow ligand deprotonation (see Experimental), in this study, the
STRENCAT acid/base behaviour is expressed in terms of proton-
ligand formation constants or protonation constants. Fig. 1 shows
the titration curves of the ligand alone: the coincidence of the
curves for ligand solutions of different concentrations at p[H+] <
9 and RL ≤ 3 (RL is the ratio between moles of base added per mole
The species distribution diagram of STRENCAT as a function
of p[H+] (Fig. 2) indicates that, in acidic solution below p[H+] = 4,
STRENCAT is almost 100% in the fully protonated form, i.e. H6L.
As the p[H+] increases, deprotonation starts, with the formation
of H5L- and H4L2-, which reach their maximum concentrations at
p[H+] about 6.2 and 7.2, respectively. H3L3- is the dominant species
in the p[H+] range 8–9.8 and reaches its maximum concentration at
p[H+] around 9. The other species, H2L4-, HL5- and L6-, form later
with an increase in p[H+]. Only at p[H+] > 11.5 does the completely
deprotonated form of STRENCAT prevail in solution.
Simple evaluation of the protonation constants of STRENCAT
does not allow to identify which specific donor sites are involved in
the deprotonation reactions. However, as the electronic spectrum
of the ligand is very sensitive to the catechol protonation state,
the deprotonation sequence of H6L could be identified through a
series of spectrophotometric measurements. The electronic spectra
(230–350 nm) of the ligand solutions recorded in the p[H+] range
4.0–11.8 are shown in Fig. S1 of the ESI.† The spectra show a
progressive increase in absorbance, and a bathochromic shift from
Fig. 1 Potentiometric titration curves of STRENCAT, (᭺) C◦L0.59 mmol
dm-3; (ꢀ) C◦L2.40 mmol dm-3.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 4236–4244 | 4237
©

View Article Online
Fig. 3 Change in chemical shift (Dd) of alkyl (a ᭺; b ꢀ; c ꢁ) and aromatic
(᭹) protons of ligand as a function of pD.
are lower than that of 4-sulfophenol (pKa = 8.56)19 and catechol
(pKa = 9.13).19 These significant differences may be ascribed to
the formation of strong hydrogen bonds between the protonated
amine and the closest deprotonated phenolate ions. The lower
basicity of the hydroxyl groups of STRENCAT with respect
to TRENCAT14 (average DlogK = 0.7) is certainly due to the
Fig. 2 Speciation of STRENCAT vs. p[H+], together with absorbance
changes at 258 nm (ꢁ) and 296 nm (᭹). Speciation was calculated
with protonation constants in Table 1 and ligand concentration used in
spectrophotometric experiment, i.e., 4.1 ¥ 10-5 mol dm-3.
288 nm, the initial absorption band assigned to p–p transitions
typical of catechol, to 296 nm. In parallel with these changes, a
new band appears at 258 nm, reaches a maximum of absorbance
at p[H+] ~ 9, and then decreases smoothly.
The changes in the electronic spectra during titration are clearly
shown in Fig. 2, which highlights the change in absorbances at 296
and 258 nm (A296 and A258) as a function of p[H+], superimposed to
the speciation plot. The values of A296 and A258 increase until H4L2-
is completely deprotonated at p[H+] ~ 9 and, after this p[H+], they
remain almost constant (A258) or slightly decrease (A296). Since the
catechol/catecholate groups are the only UV-vis chromophores
in STRENCAT, the changes observed between p[H+] ~ 4 and
p[H+] ~ 9, and 0≤ RL ≤ 3, may be safely attributed to the first
deprotonation of the three catechols of the ligand.
The modest absorbance changes at p[H+] > 9 are in agreement
with the deprotonation of groups, which do not influence the
chromophores, i.e., the amino groups. These results support the
idea that the remaining three OH protons of the catechols are
strongly bound to catecholate oxygen atoms by intramolecular
hydrogen bonds and are therefore too weakly acidic to be displaced
in the p[H+] range investigated.
The moderate decrease in A258 observed for p[H+] > 9 may reflect
a reduction in the electronic density on the chromophore, which
may be connected with a change in the extent of the intramolecular
hydrogen bonding with the decreased net charge on the amino
groups, due to their deprotonation.
The NMR titration agrees with the deprotonation sequence
suggested by spectrophotometric measurements. Fig. 3 shows the
displacement of the protons signals of the ligand (Dd) with respect
to the initial values, as a function of increasing pD (Dd = d(pD =
3.31) - d(pDmeas)). Consistently with the given description, the
position of the aromatic signal (initially at d = 7.36 ppm) shifts
towards higher fields between p[H+] = 6 and 9, while the signals
of alkyl protons, multiplets initially centered at d = 4.29, 3.60 and
3.31 ppm, are significantly shifted only at p[H+] > 9.
The pKas of the ligand amino groups (11.42, 10.53, 9.80,
Table 1) are higher than those found for tris(2-aminoethyl)amine
(TREN, 10.12, 9.43, 8.41),9,18 the scaffold amine used to prepare
STRENCAT, and the pKas of the ligand phenols (7.83, 6.83, 5.69)
-
electron-withdrawing effect of the SO3 group, which has been
shown to influence the dissociation processes of the phenol protons
in 4-sulfophenol (pKa = 8.56)19 with respect to phenol (pKa=
9.80)20 and in catechol-4-sulfonate (pKa = 8.26)19 with respect to
catechol (pKa = 9.13).19
The formation of the hydrogen-bond network in partly depro-
tonated STRENCAT, previously indicated as responsible for the
differences between STRENCAT and its parent compounds, is
well reproduced by quantum chemical calculations on one arm of
the podand ligand taken as a simplified model for STRENCAT.
Fig. 4 shows the optimized geometries of this arm when a single
proton is initially removed from the meta (a) or ortho (b) hydroxide.
Intramolecular hydrogen bonds stabilize both structures but are
stronger in the (b) case. Indeed, the interaction between the
closest phenolate oxygen and the amine proton in (b) is so strong
that it favours proton migration from nitrogen to oxygen during
molecular geometry optimization. The UV-vis spectra show, in the
real case, that migration does not occur in solution, probably due
to the solvation of the ammonium protons, not considered here.
The energy difference between the isomers, calculated with
the selected B3LYP method, was DEB3LYP = E(b) - E(a) =
-43.5 kJ mol-1. With the simpler computational AM1 approach,
a qualitatively similar result was obtained (DEAM1 = E(b) -
E(a) = -77.4 kJ mol-1) and, in that case, no proton migration
was evidenced. In conclusion, the results of both theoretical
calculations, although with some differences, strongly indicate
that deprotonation in these molecules occurs preferentially at the
ortho catechol oxygen and is favoured by the formation of strong
intramolecular H-bonds.
Lanthanide(III) complexation
Some examples of the titration curves obtained for 1 : 1 Ln(III)–
ligand systems are shown in Fig. 5. The complete set of ex-
perimental results is given in Fig. S2–S4 of the ESI.† Unlike
the Ln(III)/TRENCAT systems,14 where precipitation occurred
near p[H+] ~ 6.0–6.5, the presence of a sulfonate group in the
catechols of the ligand gives higher solubility to the complex and
4238 | Dalton Trans., 2009, 4236–4244
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Fig. 4 Optimized geometries of two forms of mono-deprotonated single arm of STRENCAT: meta position (a) and ortho position (b) deprotonation.
8, which indicates that the three metal ions, although with different
affinities for the ligand, form complexes of equal stoichiometry.
For RM > 6 (where, virtually, all the acidic protons of H6L are
neutralized), the metal titration curves are still under the ligand
titration curve, and this suggests an additional release of acid
protons, from the ligand itself or due to hydrolysis. The different
curve paths for RM < 5 reflect the different affinity of metal ions
for the ligand, which increase with the charge density of the metal
ions in the order La(III) < Gd(III) < Yb(III).
Curves concerning 1 : 2 metal–ligand titrations are less indica-
tive about the speciation in solution. They are shown in Fig. S2–
S4 of the ESI,† together with the species distribution calculated
with the stability constants of Table 2. Given the metal–ligand
1 : 2 stoichiometric ratio and the presence of a buffer region that
extends to RM ~ 8, the curve paths suggest the formation in solution
of 1 : 2 metal–ligand complexes according to eqn (1) (n = 6–9,
according to Table 2):
Ln3+ + 2H6L ꢀ LnH(12-n)L2
+ nH+
(1)
(n-3)-
Fig. 5 Potentiometric titration curves of STRENCAT in the presence of
metal ions (ꢀ) La(III), (ꢀ) Gd(III) and (᭺) Yb(III) in a 1 : 1 ligand–metal
molar ratio. R = RM for the metal ions and R = RL for the solid bold
line (ligand alone, calculated with the stability constants of Table 1 for an
initial H6L concentration of 1.6 mmol dm-3).
On the basis of these indications, several speciation models were
hypothesized for the three metal–ligand systems. The final results
of the stability constant refinement procedure are listed in Table 2.
The good fit between the solid lines, calculated with stability
constants given in the Tables 1 and 2, and the experimental results
shown in Fig. S2–S4 of the ESI,† is a measure of the reliability of
the minimization procedures.
no solid phase was therefore observed over the whole p[H+] range
investigated (~3–11.5).
The stability constants in Table 2 refer to the general eqn (2):
Fig. 5 shows that the presence of metal ions in solution brings
about large changes in the deprotonation profile of the ligand. In
particular, it is evident that, already at values of RM near to 0 (RM
is ratio between moles of base added per mole of metal ion after
neutralization of excess of mineral acid) the solutions containing
the metal ions are much more acidic than those containing the
ligand alone. For all metal ions, the titration curves show a sharp
inflection for RM ~ 5 and p[H+] ~ 7 and a coalescence above p[H+] ~
6-
(3p+q-6r)
pLn3+ + qH+ + rL ꢀ LnpHqLr
(3p+q-6r)
(2)
[LnpHqLr
[Ln3+ ]p [H+ ]q[L ]r
]
bp,q,r
=
6-
which leads to two series of complexes, one with p = 1, q = 0–4,
r = 1 and one with p = 1, q = 3–6 and r = 2.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 4236–4244 | 4239
©

View Article Online
Table 2 Stability constants for formation of Ln(III)–STRENCAT (L6-) complexes of La(III), Gd(III) and Yb(III) at 25◦ and I = 0.1 mol dm-3 NaClO4
Reaction
logbp,q,r
pLn3+ + qH+ + rL6- = LnpHqLr(3p+q-6r)
p
q
r
Ln = La
Ln = Gd
Ln = Yb
Ln3+ + L6- = LnL3-
1
1
1
1
1
1
1
1
1
1
1
1
0
1
2
3
4
3
4
5
6
1
1
1
1
1
2
2
2
2
1
1
1
17.57(0.02)
26.20(0.02)
33.08(0.03)
39.64(0.02)
—
53.60(0.05)
62.84(0.04)
70.58(0.05)
77.52(0.04)
7.13(0.03)
-4.33(0.05)
—
19.98(0.02)
29.02(0.02)
35.84(0.01)
40.73(0.02)
44.98(0.09)
55.31(0.07)
65.27(0.03)
73.46(0.03)
79.68(0.05)
9.40(0.03)
22.33(0.03)
31.11(0.02)
37.12(0.02)
41.46(0.04)
—
Ln3+ + H+ + L6-= LnHL2-
Ln3+ + 2H+ + L6- = LnH2L-
Ln3+ + 3H+ + L6- = LnH3L
Ln3+ + 4H+ + L6- = LnH4L+
Ln3+ + 3H+ + 2L6- = LnH3L2
57.7(0.1)
6-
Ln3+ + 4H+ + 2L6-= LnH4L2
67.87(0.04)
76.16(0.04)
81.57(0.09)
11.97(0.04)
0.80(0.08)
-11.6(0.2)
5-
Ln3+ + 5H+ + 2L6- = LnH5L2
4-
Ln3+ + 6H+ + 2L6- = LnH6L2
3-
Ln3+ + L6- = LnH-1L4- + H+
Ln3+ + L6- = LnH-2L5- + 2H+
Ln3+ + L6- = LnH-3L6- + 3H+
-1
-2
-3
-1.76(0.08)
-13.5(0.15)
The species LnH-qL(3-q-6) (q = 1–3) form at p[H+] > 8 according
to eqn (3):
Ln3+ + L6- ꢀ LnH-qL(3-q-6) + qH+
(3)
which may represent either the hydrolysis of the metal complex
or the formation of a complex in which the catechol units lose
residual protons, the acidity of which was not determined in this
study (see above).
From the values of the overall formation constants given in
Table 2, it is not possible to identify which specific donor of
the ligand is involved in the complexation reaction. So, in this
case too, information about this point was obtained by a series
of spectrophotometric measurements carried out on the Gd–
STRENCAT = 1 : 1 system taken as reference.
Fig. 6 Absorbance changes at 258 nm (ꢁ) and 302 nm (᭹), together with
speciation of Gd3+–STRENCAT system (L = STRENCAT, CL = 4.0 10-5
mol dm-3, CL/CGd = 1).
The electronic spectra (Fig. S5 of ESI†) were recorded from
p[H+] = 4, where complexation does not yet occur, to p[H+] =
10.8. The spectra show that the solution absorbance increases
with increasing p[H+], with a concomitant bathochromic shift of
the original bands, at 235 and 288 nm, toward the final values
of 258 and 302 nm, respectively. The intensity of the band near
258 nm first increases until p[H+] about 6, and then decreases,
and shifts to higher frequencies until it reaches the value of the
final spectrum. Fig. 6 shows the absorbances at 258 and 302 nm
as a function of the solution p[H+], together with the calculated
distribution curves for the Gd(III) system. The marked increase
in the absorbances in the p[H+] range 4–6 correspond to the
formation of GdH4L+, GdH3L, GdH2L- complexes. Recalling that
spectral changes similar to those shown in Fig. 6 were already
associated to the deprotonation of catechol, the data in that figure
may be taken as strong proof of the involvement of both catechol
oxygen atoms in the complexation reactions occurring at acidic
p[H+]. It is also of interest to note that changes in the bands of
ligand chromophore occur until p[H+] ~9, when six moles of base
per mole of metal ion (see Fig. 5) are introduced in the solutions.
This clearly indicates deprotonation of all the catechol hydroxy-
groups of the ligand at that p[H+]. Also in this case, the decrease
in absorbance at 258 nm observed between p[H+] = 6 and p[H+] =
9 may be attributed to a reduction in the electronic density on
the chromophores, likely due to the hexadentate coordination of
metal ion. The last three deprotonation reactions, which occur
after p[H+] = 9 without spectral changes, may be attributed to the
spectrophotometrically silent deprotonations of the ligand amine
groups.
On the basis of the above discussion, Scheme 2 shows the
suggested coordination modes of Gd3+, and lanthanides in general,
to STRENCAT in the complexes which prevail in solution between
p[H+] = 6 and 10, i.e., GdH2L-, GdHL2-, GdL3- (Fig. 6). In
Scheme 2, it is assumed that, due to electrostatic repulsion
between Gd3+ and ammonium groups, the metal ion first interacts
with the oxygen atoms of the lower rim of the catechols, in
GdH3L, and then, step by step as the p[H+] increases, it is
collocated inside the cage formed by the six oxygen atoms of
catechols, which at p[H+]~10, are completely deprotonated. The
presence of protonated amino groups close to phenolato oxygens
in GdH-nL-(3+n) (n = 0–3) complexes may also be at the origin of
the formation of the intra-strand hydrogen ion bonds13 depicted
in Scheme 2.
Indirect proof of intra-strand hydrogen bond formation in the
Gd(III) complexes comes from observation of the acidity of the
secondary ammonium of LnH-nL(3+n)- (eqn (4), n = 0–2), which is
always lower than that of the same group in the free ligand (eqn
(5), see also Table S1 in the ESI).†
LnH-nL(3+n)- ꢀ LnH-(n+1)L(3+n+1)- + H+ pK¢a
This journal is The Royal Society of Chemistry 2009
(4)
4240 | Dalton Trans., 2009, 4236–4244
©

View Article Online
Scheme 2
(5)
H(3-n)L(3+n)- ꢀ H(3-n-1)L(3+n+1)- + H+ pKa
1 : 1 complexes with the general formula LnH-qL(3-q-6) are the only
species in solution for all the Ln(III)–STRENCAT systems.
The calculated geometry of La(H3L)(H2O)6 in a vacuum
agrees with the structure shown in Scheme 2(a). Semi-empirical
calculations were run, starting from two hypotheses: the La(III)
ion coordinated to the three meta catecholate groups (Fig. 7a)
or the three ortho groups (Fig. 7b). As evidenced by the dotted
lines in Fig. 7, the optimized structures confirm the formation
of many intramolecular hydrogen bonds, consistent with the
spectrophotometric results. The energy difference between the
isomers, DE (DE = E(b) - E(a) = +185.1 kJ mol-1), indicates
that the coordination mode of Fig. 7(a), proposed for GdH3L in
Scheme 2, is the preferred one. This result is not in line with the
results recently obtained by Sahoo et al.15 by means of molecular
mechanics calculations (MM3 force field) for a similar tripodal
ligand (TMACHCAT): the semi empirical method21 employed in
the present study can give more reliable structures than those
obtainable by a molecular mechanics force field method, not
specifically developed to model coordination compounds.15 In any
case, the presence of the cyclohexane moiety in TMACHCAT may
also be responsible for the suggested inner coordination mode.
As regards to the selectivity for lanthanide ions of a series of
catechol based compounds in aqueous solution, it is noteworthy
that simple analysis of the complex stability constants does not
lead to correct predictions. For example, comparison of the
The latter observation suggests a second, perhaps more im-
portant, consideration concerning the nature of the complexes
in solution: the lower acidity of secondary ammonium groups
indicates that they are not directly bound to metal ion in
LnH-nL(3+n)- complexes; indeed, if the ammonium groups were
involved in the binding, their acidity would be higher, not lower,
than that of the free ligand.
Speciation plots calculated for the three metal/ligand systems
(Fig. S2–S4 in the ESI†) show that STRENCAT is able to form
with metal cations several complexes (which in some cases are only
minor species) whose presence in solution may result from a com-
pensation of errors in the minimization procedures, e.g. GdH4L+ in
Gd/STRENCAT systems. In order to ascertain the results of the
minimization processes with another technique, Gd/STRENCAT
solutions were investigated by ESI-MS spectroscopy, which in
some cases can give such information. In effect, in this case,
the ESI-MS spectra confirmed speciation given in Table 2 (see
Fig. S6(a) and S6(b) in the ESI†) revealing presence in solution of
both GdH4L+ (positive mode, assigned peak at m/z = 908.10) and
GdH2L-, GdHL2- and GdL3- complexes (negative mode at m/z =
906.03, 452.33 and 301.30, respectively). The speciation plots in
Fig. S2–S4† also show that, starting from moderately basic p[H+]
(p[H+] ~ 10) and even in the presence of excess of STRENCAT,
Fig. 7 Optimized structures of La(H3L)(H2O)6: (a) meta catecholate coordination; (b) ortho catecholate coordination. Hydrogen atoms bound to carbon
atoms are omitted. Calculated hydrogen bonds are represented as dotted lines.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 4236–4244 | 4241
©

View Article Online
stability constants for the formation of MH3L (charge omitted)
complexes of Gd(III) with the ligands shown in Scheme 1,
A final remark on pM values for STRENCAT: they are
systematically higher than those calculated for TMACHCAT
(Fig. 8), indicating that the stabilizing inductive effect of SO3
upon metal complexation for STRENCAT overcomes the already
claimed15 favourable size effect due to the introduction of the
cyclohexane ring as spacer group in TMACHCAT.
even though it reflects the correct trend (logbTRNS = 38.0613
<
-
logbTRENCAT = 40.0814 < logbTMACHCAT = 40.5315 < logbSTRENCAT
=
40.73), indicates that the complex stabilities are quite comparable
with each other. Instead, the pM (M = Gd3+) plot (Fig. 8) for
the same complexes shows that STRENCAT, probably due to
the lower basicity of its catechols is much more competitive for
Gd3+ than the parent compounds at any p[H+]. In fact, the pM
values (calculated, for example, at pH = 7.4 and [L]total = 1 ¥
10-5, [Gd3+]total = 10-6 mol dm-3) are 12.6, 10.7, 10.2 and 8.2 for
STRENCAT, TMACHCAT, TRENCAT and TRNS, respectively.
In spite of the values of pM for STRENCAT at the physiological
pH are higher than those for TMACHCAT, TRENCAT, TRNS,
they are lower than those calculated for 1,4,7,10-N,N¢,N¢¢,N¢¢¢-
tetraacetic acid (DOTA) and diethylenetriaminepentacetic acid
(DTPA) (Fig. 8). However, the lower coordination number of
the present complex with respect to that of DOTA or DTPA
complex may favour the increase of the number of coordinated
water molecules and concomitant increase of the relaxation, a
determinant parameter for signal enhancement in MRI.
Conclusions
Thanks to the increase in solubility due to sulfonation of the
catechol ring, STRENCAT is demonstrated to be able to form
strong complexes with lanthanide(III) ions over a large range
of p[H+] (3–11.5), partly unexplored in studies with the parent
compound TRENCAT. In that p[H+] range, STRENCAT is able
to form several 1 : 1 partially protonated complexes of general
formula LnHqL(3+q-6) (q = 1–4) with metal cations. Semi-empirical
calculations on the neutral complex LaH3L show that, in this
complex, the metal ion is bonded to the catecholate oxygen atoms
in the lower ligand rim and then, step by step, as p[H+] and
ligand deprotonation increase, it is collocated inside the cage
formed by the six catechol oxygen atoms, which at p[H+] near
10, are completely deprotonated in [LnL]3-. Deprotonation of
secondary ammonium groups of [LnL]3- increases the stability of
the complexes so much that the highly negative LnH-nL-(3+n) (n =
1–3) complexes are the prevalent species in solution at p[H+] > 10,
even in excess of ligand.
The pM (M = Ln3+) values calculated for STRENCAT are
higher than those of other tripodal poly-catecholate ligands,
making it the strongest currently available Ln(III)-sequestering
agent of this class.
Experimental
Chemicals
Typically, 5.0 g of TRENCAT, prepared and purified as already
reported22 were dissolved in cold 20% SO3 oleum (~100 cm3) and
stirred overnight at room temperature. The solution was then
poured onto ice and left to reach room temperature.
The resulting solution was neutralized with NaHCO3 and
concentrated until a precipitate was obtained. The mixture was
filtered and the crude STRENCAT was recovered by addition
of ethanol to the solid, filtration, and successive precipitation
with diethyl ether. The light brown precipitate was re-crystallized
several times by dissolution in ethanol and precipitation with ether
to obtain the final white product (yield 30%). Elemental analysis,
calculated for C27H36N4O15S3·H2O: C 42.07%, H 4.97%, N 7.27%.
Found: C 41.97%, H 5.02%, N 7.41%. Mass spectrum m/z = 753.2
[M + H]+, 775.2 [M + Na]+.
Solutions of lanthanide(III) perchlorates were prepared by
dissolving the appropriate oxide (Aldrich> 99.9%) in a small
excess of perchloric acid and diluted to the desired concentration.
The lanthanide(III) content in the stock solutions was determined
by titration with EDTA and xylenol orange as indicator;23 free
acid concentrations were determined by Gran’s method.24
The stock solution of perchloric acid was prepared by dilution
of concentrated perchloric acid (70% Aldrich ACS reagent)
with doubly-distilled deionized water and standardized by titra-
tion with tris(hydroxymethyl)aminomethane (Aldrich). Diluted
Fig. 8 Calculated pM for Gd(III) catechol-based ligands of Scheme 1
(CG◦ d = 10-6, CL◦ = 10-5 mol dm-3). Dotted lines: p[H+] values higher
than those corresponding to formation of hydroxo-complexes (TRNS)
or precipitation (TRENCAT, TMACHCAT).
The same calculation, applied to La(III)–Yb(III)/STRENCAT
systems, shows an increase in pM of about 5 orders of magnitude
on going from La to Yb at pH 7.4 (pLa3+ = 9.8, pGd3+ = 12.6,
pYb3+ = 14.8), thus revealing that the electrostatic effects dominate
in this kind of interactions.
As mentioned above, the complexes of Ln(III) with STRENCAT
are soluble over the whole p[H+] range studied in this work (~3–
11.5). This offers the chance to verify that complete deprotonation
of secondary ammonium groups of the ligand, although they are
not involved in complexation, greatly increases the metal–ligand
affinity (see the constant increase in pM after p[H+] = 10 in Fig. 8).
This is not surprising, as deprotonation of amino groups increases
the negative charge of the ligand, which interacts with Ln3+ in a
purely electrostatic manner.
4242 | Dalton Trans., 2009, 4236–4244
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
sodium hydroxide solutions were prepared from a saturated NaOH
solution (50% NaOH, 50% H2O), left standing in a polyethylene
bottle for several days. Standardization of these solutions was
carried out with potassium hydrogen phthalate (Aldrich).23
The ionic strength of all solutions used in this study was adjusted
to 0.1 mol dm-3 by appropriate amounts of 1 mol dm-3 NaClO4.
Solutions of the ligand and metal ions were also prepared in
deoxygenated water.
NMR titrations
1
The H spectra of STRENCAT (1.3 mmol dm-3) were recorded
in D2O (Aldrich 99.95% D) on a Bruker instrument operating
at 200 MHz, with a pre-saturation sequence for suppression of
the water signal. The pD was adjusted to the desired value, with
either NaOD or DCl (Aldrich), measured with a pH-meter. The
relationship pD = pHmeas. + 0.4 was used.27
ESI mass spectrometry
Potentiometry
The mass spectra of solutions containing STRENCAT 4.8 ¥ 10-5
mol dm-3 and Gd(III) chloride in a 1 : 1 ratio were recorded by
direct injection in an Ion Trap LXQ Finnigan instrument in the
absence of ionic medium. This technique is particularly useful
with Gd(III) compounds, because of the typical presence of seven
isotopic signals by which immediate confirmation of the charge of
a given species can be obtained.
The protonation constants of the ligand and the formation con-
stants of its metal complexes were determined by potentiometric
titrations. The temperature in the titration cell was maintained
at 25.0
0.1 ◦C by means of a circulatory bath (HAAKE
K10). For all titrations, the electromotive force (emf) at the
terminals of a Metrohm combined glass electrode (Unitrode,
6.0259.100) was measured by an Amel mod. 338 pH-meter. The
original electrode filling solution (3 mol dm-3 KCl) was replaced
with a 100 mmol dm-3 NaCl solution, to avoid clogging of
the electrode sleeve joint by KClO4. Before each titration, the
glass electrode was calibrated to determine the hydrogen ion
concentration following the reported procedure.25 A computer-
controlled potentiometric apparatus collected emf data after each
titrant addition, at intervals given by the data collection criterion,
i.e. Demf = 0.0 mV for 2 min: this condition was usually satisfied
within 2–4 min in the calibration experiments, whereas longer
time was necessary to reach equilibrium in the experiments
concerning ligand protonation and metal/ligand complexation
(5–10 min).
Potentiometric titrations of the ligand acid, alone or in the
presence of metal ions, were carried out by adding standard NaOH
to solutions of the ligand, containing an excess of mineral acid
in the absence and presence of lanthanide(III) ions. STRENCAT
has basic sites which may be partially or fully protonated in
acidic medium. In the previous discussion, the symbol CH,exc
refers to the calculated concentration of free hydrogen ions in the
titration cell after complete protonation of all the basic sites of the
ligand, except sulfonic groups. Analytical details of potentiometric
titrations are listed in Table S2 (ESI).† All measurements were
carried out under an argon stream, pre-saturated in water. No
solid phase was assessed in the concentration range studied
here. The protonation constants of the ligand were obtained by
the Hyperquad program26 and used as input data to evaluate
the formation constants of the metal complexes with the same
minimization program.
Molecular modelling
Density functional theory calculations were performed to obtain
energy information of the possible isomers and hydrogen bonding
structure within a single arm of the STRENCAT ligand. Optimiza-
tion was carried out with the B3LYP functional, as implemented in
the GAMESS program28 and the double split valence 6–31++G(d,
p) basis set, which contains polarization and diffuse basis functions
for a better description of intramolecular hydrogen bonding. The
final energies were recalculated including PCM29 water to mimic
the bulk solvent effect. Hessian calculations were performed to
confirm that the final structures were local minima. Calculations
were also repeated with the semiempirical method AM1.30
The geometries of La(H3L)(H2O)6 complexes were optimized
with the MOPAC2007 program31 using the PM3 semiempirical
method,32,33 which includes parameters which have been demon-
strated to be able to produce reliable structures for lanthanide
complexes.21
References
1 J. Zhang and S. Petoud, Chem.–Eur. J., 2008, 14, 1264–1272.
2 S. Pandya, J. H. Yu and D. Parker, Dalton Trans., 2006, 2757–2766.
3 C. L. Liu, M. Wang, T. L. Zhang and H. Z. Sun, Coord. Chem. Rev.,
2004, 248, 147–168.
4 P. Caravan, J. J. Ellison, T. J. McMurry and R. B. Lauffer, Chem. Rev.,
1999, 99, 2293–2352.
5 A. E. Merbach and E. Toth, The Chemistry of Contrast Agents in
Medical Magnetic Resonance Imaging, Wiley, Chichester, 2001.
6 P. Caravan, Chem. Soc. Rev., 2006, 35, 512–523.
7 E. Toth, L. Helm and A. E. Merbach, Relaxivity of MRI Contrast
Agents, in Topics in Current Chemistry, Springer, Berlin, 2002, pp. 61–
101.
8 S. Aime, S. G. Crich, E. Gianolio, G. B. Giovenzana, L. Tei and E.
Terreno, Coord. Chem. Rev., 2006, 250, 1562–1579.
9 A. E. Martell, R. M. Smith, and R. J. Motekaitis, Critically Selected
Stability Constants of Metal Complexes Database, 2005, v. 5.82.
10 P. Di Bernardo, P. L. Zanonato, A. Melchior, R. Portanova, M. Tolazzi,
G. R. Choppin and Z. Wang, Inorg. Chem., 2008, 47, 1155–1164.
11 D. H. Zhu, D. Kappel and K. Raymond, Inorg. Chim. Acta, 1988, 147,
115–121.
UV-vis measurements
Spectrophotometric titrations were carried out on a Varian Cary
50 instrument equipped with a thin optic-fiber probe (1 cm path
length) dipped in the cup solution, which was maintained at the
selected temperature by an external thermostat. The hydrogen ion
concentration in the titration cell was measured by a combined
Metrohm glass electrode. The p[H+] in the solution cup was
changed by small additions of standard NaOH. The metal ion
and ligand concentrations in the titration cell were 4.1 ¥ 10-5 mol
dm-3.
12 M. P. Lowe, P. Caravan, S. J. Rettig and C. Orvig, Inorg. Chem., 1998,
37, 1637–1647.
13 P. Caravan, T. Hedlund, S. Liu, S. Sjoberg and C. Orvig, J. Am. Chem.
Soc., 1995, 117, 11230–11238.
14 A. Bismondo, P. Di Bernardo, R. Portanova, M. Tolazzi and P. L.
Zanonato, Polyhedron, 2002, 21, 1393–1396.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 4236–4244 | 4243
©

View Article Online
15 S. K. Sahoo, M. Baral and B. K. Kanungo, Polyhedron, 2006, 25, 722–
736.
16 M. A. Santos, Coord. Chem. Rev., 2008, 252, 1213–1224.
17 G. Crisponi and M. Remelli, Coord. Chem. Rev., 2008, 252, 1225–
1240.
24 G. Gran, Analyst, 1952, 77, 661–671.
25 P. Di Bernardo, P. L. Zanonato, S. Tamburini, P. Tomasin and P. A.
Vigato, Dalton Trans., 2006, 4711–4721.
26 P. Gans, A. Sabatini and A. Vacca, Talanta, 1996, 43, 1739–1753.
27 P. K. Glasoe and F. A. Long, J. Phys. Chem., 1960, 64, 188–190.
28 M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert, M. S. Gordon,
J. H. Jensen, S. Koseki, N. Matsunaga, K. A. Nguyen, S. J. Su, T. L.
Windus, M. Dupuis and J. A. Montgomery, J. Comput. Chem., 1993,
14, 1347–1363.
18 Z. W. Mao, G. Liehr and R. van Eldik, J. Chem. Soc., Dalton Trans.,
2001, 1593–1600.
19 Y. Murakami, K. Nakamura and M. Tokunaga, Bull. Chem. Soc. Jpn.,
1963, 36, 669–675.
20 G. Crisponi, M. Casu, V. M. Nurchi, F. Cesare-Marincola, T. Pivetta
and R. Silvagni, Talanta, 2002, 56, 441–449.
29 J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999–
3093.
21 N. B. da Costa, R. O. Freire, A. M. Simas and G. B. Rocha, J. Phys.
Chem. A, 2007, 111, 5015–5018.
22 A. Bismondo, C. Comuzzi, P. Di Bernardo and P. L. Zanonato, Inorg.
Chim. Acta, 1999, 286, 103–107.
23 A. I. Vogel, Vogel’s Textbook of Quantitative Chemical Analysis,
Longman, Harlow, 5th edn., 1989.
30 M. J. S. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Stewart, J. Am.
Chem. Soc., 1985, 107, 3902–3909.
31 J. J. P. Stewart, MOPAC2007, 2007, v. 8.205L.
32 J. J. P. Stewart, J. Comput. Chem., 1989, 10, 209–220.
33 R. O. Freire, N. B. da Costa, G. B. Rocha and A. M. Simas, J. Phys.
Chem. A, 2006, 110, 5897–5900.
4244 | Dalton Trans., 2009, 4236–4244
This journal is
The Royal Society of Chemistry 2009
©