Synthesis and properties of a family of unsymmetric dinuclear complexes of LnIII (Ln = Eu, Gd, Tb)

Article abstract of DOI:10.1021/ic1008285

A new ligand has been synthesized with the aim of favoring distinct coordination environments within lanthanide polynuclear complexes. It has led to the formation of three unsymmetrical [Ln^III₂] (Ln = Gd, Tb, Eu) complexes, exhibiting weak antiferromagnetic coupling and, for Eu and Tb, high single-ion magnetic anisotropy. All of these attributes are necessary for these clusters to behave as possible 2qubit quantum gates.

Full text of DOI:10.1021/ic1008285

6784 Inorg. Chem. 2010, 49, 6784–6786
DOI: 10.1021/ic1008285
Synthesis and Properties of a Family of Unsymmetric Dinuclear Complexes
of LnIII (Ln = Eu, Gd, Tb)
†
‡
‡
,‡
ꢀ †
ꢁ
David Aguila, Leonı A. Barrios, Fernando Luis, Ana Repolles, Olivier Roubeau,* Simon J. Teat,§ and
´
,†
´
Guillem Aromı*
†
´
ꢀ
Departament de Quımica Inorganica, Universitat de Barcelona, Diagonal 647, 08028 Barcelona, Spain,
‡
ꢁ
Instituto de Ciencia de Materiales de Aragon, CSIC and Universidad de Zaragoza, Plaza San Francisco s/n,
§
50009 Zaragoza, Spain, and Advanced Light Source, Berkeley Laboratory, 1 Cyclotron Road, Berkeley,
California 94720
Received April 27, 2010
A new ligand has been synthesized with the aim of favoring distinct
coordination environments within lanthanide polynuclear complexes. It
operations, in particular, relatively long decoherence times
and a good definition of qubit states.7,8 In this context, it can
be envisaged that individual lanthanide ions with strong
anisotropy are potential candidates for embodying the qubits
of scalable quantum gates (QGs).9 A universal quantum
computer (QC) can be built through a combination of single
qubits and 2qubit C-NOT QGs. The operation of such a QG
requires dissimilar qubits exhibiting weak interaction.10 These
conditions are necessary for the selective preparation of the
quantum states required for the gate operations. The most
advanced current attempts to prepare spin-based molecular
QGs involve the synthesis of single molecules containing two
connected coordination clusters, each in the form of a hetero-
metallic wheel of the type [Cr7M] (M = various divalent
metals).11 We have also been attempting to prepare candidates
of 2qubit QGs through the synthesis of poly(β-diketone)
ligands, designedtocause the assembly of transition metalsas
pairs of magnetic clusters within molecules.12-15 We have now
turned our attention to lanthanides as possible spin carriers
within such molecular models. In order to introduce the
necessary inequivalence between both halves of the targeted
assemblies, the synthesis of a suitable ligand [6-3-oxo-3-(2-
hydroxyphenyl)propionyl)-2-pyridinecarboxylic acid, H3L;
Chart 1, left] has been devised and carried out successfully.
has led to the formation of three unsymmetrical [LnIII ] (Ln=Gd, Tb,
2
Eu) complexes, exhibiting weak antiferromagnetic coupling and, for
Eu and Tb, high single-ion magnetic anisotropy. All of these attri-
butes are necessary for these clusters to behave as possible 2qubit
quantum gates.
Because of their unique spectroscopic and electronic prop-
erties, lanthanides have for a long time been at the forefront
of many frontier research fields. The presence of this group
of metals in the area of molecular magnetism dates back to
pioneering reports describing the nature of the magnetic
exchange between CuII and GdIII ions.1 This interest has ex-
perienced a renewed impetus since slow relaxation of mag-
netization in single molecules was observed for mononuclear
TbIII and DyIII complexes, resulting from splitting of their
J multiplets into various |Jz| sublevels.2,3 Following this dis-
covery, many other lanthanide complexes of varying nuclearity
and exhibiting single-molecule-magnet4,5 or single-chain-
magnet6 behavior have been made. This raised new expecta-
tions with respect to the goal of using single molecules as bits
for magnetic memories. In addition, magnetic clusters with
a ground-state doublet (effective spin 1/2) fulfill some of the
basic requirements needed to implement quantum logical
(7) Affronte, M. J. Mater. Chem. 2009, 19, 1731–1737.
(8) Ardavan, A.; Rival, O.; Morton, J. J. L.; Blundell, S. J.; Tyryshkin,
A. M.; Timco, G. A.; Winpenny, R. E. P. Phys. Rev. Lett. 2007, 98, 4.
(9) Bertaina, S.; Gambarelli, S.; Tkachuk, A.; Kurkin, I. N.; Malkin, B.;
Stepanov, A.; Barbara, B. Nat. Nanotechnol. 2007, 2, 39–42.
(10) Ardavan, A.; Blundell, S. J. J. Mater. Chem. 2009, 19, 1754–1760.
(11) Candini, A.; Lorusso, G.; Troiani, F.; Ghirri, A.; Carretta, S.;
Santini, P.; Amoretti, G.; Muryn, C.; Tuna, F.; Timco, G.; McInnes,
E. J. L.; Winpenny, R. E. P.; Wernsdorfer, W.; Affronte, M. Phys. Rev.
Lett. 2010, 104, 037203.
*To whom correspondence should be addressed. E-mail: roubeau@
unizar.es (O.R.), guillem.aromi@qi.ub.es (G.A.).
(1) Bencini, A.; Benelli, C.; Caneschi, A.; Carlin, R. L.; Dei, A.; Gatteschi,
D. J. Am. Chem. Soc. 1985, 107, 8128–8136.
(2) Ishikawa, N.; Sugita, M.; Ishikawa, T.; Koshihara, S.; Kaizu, Y.
J. Am. Chem. Soc. 2003, 125, 8694–8695.
~
(3) Ishikawa, N.; Sugita, M.; Ishikawa, T.; Koshihara, S.; Kaizu, Y.
J. Phys. Chem. B 2004, 108, 11265–11271.
(12) Sanudo, E. C.; Cauchy, T.; Ruiz, E.; Laye, R. H.; Roubeau, O.; Teat,
S. J.; Aromı, G. Inorg. Chem. 2007, 46, 9045–9047.
(13) Barrios, L. A.; Aguila, D.; Roubeau, O.; Gamez, P.; Ribas-Arino, J.;
Teat, S. J.; Aromı, G. Chem.;Eur. J. 2009, 15, 11235–11243.
(14) Barrios, L. A.; Aguila, D.; Mellat, S.; Roubeau, O.; Teat, S. J.;
Gamez, P.; Aromı
´
ꢁ
ꢀ
~
(4) Lin, P. H.; Burchell, T. J.; Clerac, R.; Murugesu, M. Angew. Chem.,
Int. Ed. 2008, 47, 8848–8851.
´
ꢁ
ꢀ
(5) Xu, G. F.; Wang, Q. L.; Gamez, P.; Ma, Y.; Clerac, R.; Tang, J. K.;
Yan, S. P.; Cheng, P.; Liao, D. Z. Chem. Commun. 2010, 46, 1506–1508.
(6) Bogani, L.; Sangregorio, C.; Sessoli, R.; Gatteschi, D. Angew. Chem.,
Int. Ed. 2005, 44, 5817–5821.
´
, G. C. R. Chim. 2008, 11, 1117–1120.
, G.; Gamez, P.; Reedijk, J. Coord. Chem. Rev. 2008, 252, 964–
(15) Aromı
´
989.
r
pubs.acs.org/IC
Published on Web 07/06/2010
2010 American Chemical Society

Communication
Inorganic Chemistry, Vol. 49, No. 15, 2010 6785
Chart 1. Ligand H3L and the Coordination Mode of H2L- in 1-3 (the
Mode for HL2- Would Be Analogous)
It features a rich variety of donor atoms distributed asymme-
trically, which has allowed the production of an analogous
series of dinuclear LnIII complexes with formulas [Gd2(HL)2-
(H2L)Cl(py)(H2O)] (1), [Tb2(HL)2(H2L)Cl(py)2] (2), and
[Eu2(HL)2(H2L)(NO3)(py)(H2O)] (3). Each complex displays
two metals with distinct coordination environments.
While dinuclear Ln2 coordination compounds are very
numerous in the literature (104, 113, and 99 hits for Gd, Tb,
and Eu, respectively, on the CCDC, version 5.31, Feb 2010),
examples involving two dissimilar metal sites are compara-
tively extremely scarce. More precisely, only four, two, and
six examples (see the Supporting Information for references)
respectively have been found for the cases of Gd, Tb, and Eu.
Especially interesting is a gadolinium(III) semiquinonato
complex,16 which is the only asymmetric system that has seen
its magnetic properties investigated. These were interpreted
using a giant spin approach, and, therefore, no attempt was
made to put into evidence the difference between both ions
from a magnetic point of view. This would be easier if, con-
trary to GdIII, the metal centers exhibit significant anisotropy,
as would be likely for EuIII and TbIII.
Figure 1. PovRay representation of 1. The smallest gray atoms are hydro-
gen; the other gray atoms are carbon. Only hydrogen atoms riding on oxy-
gen atoms are represented. The structures of 2 and 3 are analogous, with one
py ligand instead of H2O in the former and one monodentate NO3- ligand
instead of Cl- in the latter. Selected bond distances (A) and angles (deg) in a
1/2/3 format: Ln-Cl, 2.7758(9)/2.727(3)/-; Ln-Onitrate, -/-/2.458(9);
Ln-Owater, 2.460(3)/-/2.462(9); Ln-Npy, 2.700(3)/2.644(9)-2.652(10)/
2.703(12); Ln-N (range), 2.486(3)-2.622(3)/2.505(10)-2.547(9)/2.511(11)-
2.652(10); Ln-O (range), 2.353(2)-2.597(2)/2.320(6)-2.537(6)/2.368(8)-
2.647(8); Ln Ln, 3.8127(4)/3.7853(13)/3.8192(10); Ln-O-Ln (range),
3 3 3
98.86(8)-103.36(9)/99.9(3)-103.7(2)/97.4(3)-103.6(3).
The ligand H3L was prepared as was previously done for a
similar asymmetric β-diketone,17 from the Claisen condensa-
tion between 2-hydroxyacetophenone and the appropriate
ester (see the Supporting Information). Complexes 1-3 were
obtained as crystals from layers with either of the reac-
tion systems formed by H3L and the corresponding LnX3
salt (X- = Cl- or NO3-; see the Supporting Information).
They exhibit virtually the same IR spectrum (the only diffe-
rence is caused by the anion), and their identity could be
established by single-crystal X-ray diffraction.
Figure 2. Labeled representation of the coordination environment around
the metals of complex 1, emphasizing their disparity and the reference
polyhedra lying closest to their geometry. According to CShMs,18 Gd1 (left)
is not close enough to any reference shape, with the closest one being the so-
called “irregular muffin shape”. Gd2 (right) lies close to a “capped square
antiprism”. For complexes 2 and 3, the geometries are analogous.
two pockets. Nonacoordination around Ln1 (Figure 2) is
completed with one terminal Cl- (1 and 2) or NO3- (3) ligand,
whereas the same coordination number in Ln2 is reached
through binding of either one H2O and one py ligand (1 and
3) or two py groups (2). The stereochemical disparity between
both metals within each complex has been quantified by
means of continuous-shape measures (CShMs).18 These have
established that the coordination geometry of the sites Ln1 is
very irregular and does not sufficiently approach any of the
known reference polyhedra proposed for coordination num-
ber 9.18 By contrast, these calculations show that the shape of
metals Ln2 is best described as a distorted capped square
antiprism (see Table S5 in the Supporting Information for
CShMs on all of the metals with respect to eight reference
polyhedra for 1-3). In general, the bond distances to the
metals (see the caption of Figure 1 or the Supporting Infor-
mation for a complete list) follow the trend expected from an
increase of the atomic number within the period.
The three compounds exhibit very similar molecular
structures (Figure 1 and Figures S1-S3 and Tables S1-S4
in the Supporting Information) with one dinuclear molecule
per asymmetric unit and pyridine (2 and 3) or pyridine/water
(1) lattice solvent molecules. The LnIII metals are bridged and
chelated by two HL2- ligands disposed along the molecular
axis in a head-to-tail manner and by one H2L- donor. This
renders both metallic sites nonequivalent. The ligands H3L,
therefore, lose upon coordination either the carboxylic and
β-diketonic protons or just the latter. The carboxylic proton
of H2L- was found crystallographically for 1 and 3 (see
Figure 1) and forms a strong hydrogen bond with a solvate py
molecule (see Figure S4 in the Supporting Information). In
complex 2, the proton is likely to be distributed over two
carboxylate oxygen atoms, forming a triangular hydrogen
bond with a py molecule oriented accordingly (see Figure S5
in the Supporting Information). In all complexes, these
ligands coordinate two metals through their β-diketonate-
and dipicolinate-like coordination pockets, respectively, thus
linking both centers via the common oxygen atom of these
For their relevance in QC, the presence of weak antiferro-
magnetic (AF) coupling within complexes 1-3 and of mag-
netic anisotropy inthe metals was evaluated by means of bulk
magnetic measurements on polycrystalline samples. Figure 3
(16) Dei, A.; Gatteschi, D.; Massa, C. A.; Pardi, L. A.; Poussereau, S.;
Sorace, L. Chem.;Eur. J. 2000, 6, 4580–4586.
(17) Aromı
´
, G.; Gamez, P.; Roubeau, O.; Berzal, P. C.; Kooijman, H.;
(18) Ruiz-Martinez, A.; Casanova, D.; Alvarez, S. Chem.;Eur. J. 2008,
14, 1291–1303.
Spek, A. L.; Driessen, W. L.; Reedijk, J. Inorg. Chem. 2002, 41, 3673–3683.

ꢀ
Aguila et al.
6786 Inorganic Chemistry, Vol. 49, No. 15, 2010
at ca. 23.3 cm3 K mol-1 at 280 K down to 150 K, when a
gradual decrease begins. The shape of theχT vs T curve below
25 K is very similar to that of 1, reaching a value of 17.46 cm3
K mol-1 at 1.8 K. The high-temperature values agree well
with those expected for two uncoupled TbIII ions (7F6, S=3,
L=3, g=3/2, χT=23.62 cm3 K mol-1). By analogy with 1,
the decline at low temperature is probably due to the weak
AF coupling betweenTbIII ions, whilethe higher temperature
behavior can be ascribed to the effects of the strong spin-
orbit coupling of TbIII, thus unveiling the strong magnetic
anisotropy in 2. For compound 3, χT decreases continuously
from 3.45 cm3 K mol-1 at 280 K, slightly above the calculated
value for two uncoupled EuIII ions, which suggests that
excited states with higher J values are populated. At near
20 K, χT is already close to zero, indicating a J = 0 ground
state of the EuIII ions (6F0). The observed χT vs T curve
should thus be attributed to depopulation of the Stark levels
of the EuIII ions, masking the possible weak AF interaction
that could be inferred by analogy with 1 and 2.
Figure 3. χT vs T curves for compounds 1-3. χ is the molar paramag-
netic susceptibility per [Ln2] unit. The full line is the best fit of the experi-
mental data for 1 (see the text).
shows the temperature dependence of χT (with χ being the
molar paramagnetic susceptibility per Ln2 unit) for the
three compounds. Complex 1 presents a stable χT value of
ca. 15.5 cm3 K mol-1 from 280 K down to near 30 K, where a
small decrease is observed down to 13.58 cm3 K mol-1 at 1.8 K.
In the absence of anisotropic effects (as is confirmed by
isothermal magnetization data; see Figure S6 in the Support-
ing Information), this decline can only be ascribed to weak
AF interactions. On the other hand, at 2 K the reduced
magnetization of 1 saturates rapidly, reaching MS=13.84 μB
at 5 T (Figure S6 in the Supporting Information). The
experimental values of MS and χT agree with those expected
for two weakly antiferromagnetically coupled GdIII ions with
g = 2 (χT = 15.75 cm3 K mol-1 at high temperature and
MS=14). Indeed, the χT vs T curve was reproduced satisfac-
torily (Figure 3) with a model based on the isotropic spin
The above results demonstrate a successful synthetic stra-
tegy for locating two LnIII ions within a molecule, exhibiting
two distinct environments and weak AF coupling. In addi-
tion, for compounds 2 and 3, the metals present strong spin-
orbit coupling, resulting in strong magnetic anisotropy and
thus gathering many of the requirements needed for these
species to operate as QGs. Single-crystal, ultralow-temperature
magnetic studies are in progress in order to put into evidence
the magnetic dissymmetry of both metals within complex 2
and therefore evaluate its potential as hardware in QC.
Acknowledgment. G.A. thanks the Generalitat de Cat-
alunya for the prize ICREA Academia 2008. The authors
thank the Spanish MCI through Grants CTQ2009-06959
(to G.A., L.A.B., and D.A.) and MAT2009-13977-C03
(to F.L.) and the CONSOLIDER-INGENIO in Mole-
cular Nanoscience (Grant CSD2007-00010 to F.L.). The
Advanced Light Source (to S.J.T.) is supported by the U.S.
Department of Energy under Contract DE-AC02-
Hamiltonian H=-J(SGd1 SGd2) including a fixed temperature-
3
independent paramagnetism (TIP) term,19 which yielded the
following best parameters; J/kB=-0.041(1) K, g = 1.98(1),
and TIP=6 Â 10-4 cm3 K mol-1. These values are in correct
agreement with the literature data.20 For 2, χT remains stable
05CH11231. The authors thank Jorge Echeverrı
CShM calculations.
´
a for
Supporting Information Available: Synthetic and crystallo-
graphic details, CShMs, and various structural figures and tables.
This material is available free of charge via the Internet at http://
pubs.acs.org.
(19) Panagiotopoulos, A.; Zafiropoulos, T. F.; Perlepes, S. P.; Bakalbassis,
E.; Massonramade, I.; Kahn, O.; Terzis, A.; Raptopoulou, C. P. Inorg. Chem.
1995, 34, 4918–4920.
(20) Roy, L. E.; Hughbanks, T. J. Am. Chem. Soc. 2006, 128, 568–575.