Molecular nanoscale magnetic refrigerants: A ferrimagnetic {Cu II15GdIII7} cagelike cluster from the use of pyridine-2,6-dimethanol

Article abstract of DOI:10.1021/ic4020064

The employment of pyridine-2,6-dimethanol in 3d/4f metal cluster chemistry has afforded a new {Cu^II₁₅Gd^III₇} cagelike molecule with a beautiful structure built by fused triangular subunits; the compound exhibits

Full text of DOI:10.1021/ic4020064

Communication
pubs.acs.org/IC
Molecular Nanoscale Magnetic Refrigerants: A Ferrimagnetic
{CuII GdIII } Cagelike Cluster from the Use of Pyridine-2,6-dimethanol
15
7
Despina Dermitzaki,† Giulia Lorusso,‡ Catherine P. Raptopoulou,§ Vassilis Psycharis,§ Albert Escuer,*,⊥
Marco Evangelisti,*,‡ Spyros P. Perlepes,*,† and Theocharis C. Stamatatos*,#
†Department of Chemistry, University of Patras, 265 04 Patras, Greece
‡Departamento de Física de la Materia Condensada, Instituto de Ciencia de Materiales de Aragon
́
(ICMA), CSIC-Universidad de
Zaragoza, 50009 Zaragoza, Spain
⊥Departament de Quimica Inorganica, Universitat de Barcelona, Diagonal 645, 08028 Barcelona, Spain
§Department of Materials Science, Institute of Advanced Materials, Physicochemical Processes, Nanotechnology and Microsystems,
NCSR “Demokritos”, 153 10 Agia Paraskevi Attikis, Greece
#Department of Chemistry, Brock University, L2S 3A1 St. Catharines, Ontario, Canada
S
* Supporting Information
into account, Cu2+/Gd3+ complexes seem to be very good
candidates for molecular refrigerants because the Gd3+ ion has an
ABSTRACT: The employment of pyridine-2,6-dimetha-
isotropic f7 configuration and the Cu2+/Gd3+ magnetic
nol in 3d/4f metal cluster chemistry has afforded a new
{CuII15GdIII7} cagelike molecule with a beautiful structure
interactions are generally ferromagnetic in nature, leading to
ground states with increased and large-spin multiplicity.5
built by fused triangular subunits; the compound exhibits
Further, the Cu2+/Gd3+ magnetic exchange interactions are
an overall ferrimagnetic behavior with an appreciable
very weak because of the very efficient shielding of the Gd3+ 4f
ground-state spin value and shows promise as a low-
orbitals by the fully occupied 5s and 5p orbitals,5 and this
temperature magnetic refrigerant.
generates multiple low-lying excited and field-accessible states,
each of which can contribute to the magnetic entropy of the
system. The magnetic entropy is related to the spin by Sm = R
ln(2S + 1), where R is the gas constant. To date, clusters with very
large MCE include homometallic {Gd2}6a and {Gd24}1b and
heterometallic {Ni12Gd36}6b and {Mo4Gd12}6c complexes with
−ΔSm values of ∼40, 46.1, 36.3, and 35.3 J kg−1 K−1, respectively,
for liquid-helium temperatures and ΔB0 = 7 T.
ne of the most fascinating challenges in modern
coordination chemistry is undoubtedly the combination
O
of an aesthetically pleasing structure with an interesting physical
property or occasionally with more than one property within the
same molecular species. Polynuclear heterometallic 3d/4f metal
compounds are clearly excellent candidates to satisfy such
expectations.1 This is due to the pronounced ability of these
metal ions to form and stabilize high-nuclearity and structurally
complicated metal−oxido/hydroxido inorganic cores sur-
rounded in the periphery by organic bridging/chelating ligands.
The type and nature of the latter groups play a significant role in
the self-assembly synthesis of new polynuclear compounds, and
the choice of a small, flexible, and polydentate ligand (or ligand
“blend”) seems to be an attractive route. Polynuclear 3d/4f metal
complexes (or coordination clusters) have shown a remarkable
ability to act either as single-molecule magnets2 mainly when the
f-block ions are highly anisotropic and possess a significant spin
(i.e., DyIII, TbIII) or as magnetic refrigerants3 when the molecules
are isotropic and high-spin, conditions which are fulfilled by the
copresence of, for instance, GdIII and CuII metal ions.
Magnetic refrigeration is based on the magnetocaloric effect
(MCE). MCE deals with a change of the magnetic entropy upon
application of a magnetic field and can be used for cooling
purposes via adiabatic demagnetization.4 In the last 6 years, the
systematic exploration of the MCE in isotropic molecular
magnetic compounds has demonstrated enormous values, larger
than those observed on lanthanide alloys and magnetic
nanoparticles.3,4
In simple words, to construct polynuclear metal complexes
behaving as magnetic coolers, we need a metal−ligand “blend”,
which will guarantee aggregation of the metal ions into a large-
nuclearity motif and will produce a ferro- or ferrimagnetic system
with negligible anisotropy. Toward this end, we have decided to
employ the tridentate (N, O, O) ligand pyridine-2,6-dimethanol
(pdmH2) in heterometallic copper(2+)/gadolinium(3+) carbox-
ylate chemistry as a means of satisfying the above requirements.
Thus, the reaction of Cu(O2CPh)2·2H2O, Gd(NO3)3·6H2O,
pdmH2, and NEt3 in a 1:1:1:1 molar ratio in MeCN led to a blue
suspension, which upon filtration gave a pale-blue filtrate. The
latter was allowed to slowly evaporate at room temperature,
yielding after 10 days blue crystals of [Cu15Gd7(OH)6(CO3)4-
(O2CPh)19(pdm)9(pdmH2)3(H2O)2] (1) in 20% yield.7 The
formula of 1 is based on metric parameters, charge-balance
considerations, and bond-valence-sum calculations on the O
atoms. The CO32− ions are presumably derived from the fixation
8
−
of atmospheric CO2. Identification of the CO32− ions (vs NO3 )
in 1 was determined via careful consideration of the X-ray
diffraction data and observation of carbonate-related IR
absorption bands at ∼1400 and 845 cm−1.8 Although rare, the
The MCE is greatly enhanced in molecules containing
Received: August 3, 2013
isotropic magnetic ions with large total spin S.3 Taking this
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ic4020064 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Communication
copresence of fully deprotonated and neutral forms of an
alkoxide-based ligand, as the pdm2−/pdmH2 combination in 1, is
with precedence in 3d/4f metal cluster chemistry.9
fields of 0.3 T (300−30 K) and 0.02 T (30−2 K) to avoid
saturation effects and are plotted as χMT vs T in Figure 2. The
The structure of 1 consists of a {CuII15GdIII7} cagelike cluster
(Figure 1, left) with C3 crystallographic symmetry (the 3-fold axis
Figure 2. χMT vs T plot of 1. Inset: Field-dependent magnetization plot
at the indicated temperatures. Lines are guides to the eye.
χMT at 300 K is 66.79 cm3 K mol−1, essentially equal to the
calculated value of 66.98 cm3 K mol−1 for 15 CuII (S = 1/2) and 7
GdIII (S = 7/2, L = 0) noninteracting ions, assuming an average g
value of 2.10. It slowly decreases with decreasing temperature to
a value of 62.53 cm3 K mol−1 at 20 K and then rapidly increases to
79.93 cm3 K mol−1 at 4.0 K before it drops to 64.78 cm3 K mol−1
at 2.0 K. The shape of the χMT vs T plot indicates that both
antiferro- and ferromagnetic exchange interactions dominate at
different T regions. It is well established that alkoxido-bridged
Cu2+ ions, with Cu−OR−Cu bond angles larger than 95°, are
antiferromagnetically coupled, and thus the former types of
interactions in 1 can be reasonably assigned to Cu2+···Cu2+ and
the latter to Cu2+···Gd3+ or Gd3+···Gd3+. The ground-state spin
value of this ferrimagnetic system cannot be accurately
determined because of its complicated structure with many
fused triangular subunits. The magnetization measurements
(inset of Figure 2) show a trend for saturation at a value of 58.3
NμB at the highest fields and lowest temperatures employed,
suggesting a net spin state of S = 58/2 = 29. This can be tentatively
rationalized assuming all of the CuII···GdIII interactions to be
ferromagnetic (12 × 1/2 + 7 × 7/2 = 61/2) and the interactions
between the three extrinsic CuII ions (Cu1, Cu1a, and Cu1b)
with their directly bridged CuII ions (Cu2/Cu3, Cu2a/Cu3a, and
Cu2b/Cu3b) to be antiferromagnetic, thus forcing those three S
= 1/2 spins to align antiparallel to all other and resulting in a field-
stabilized spin state of S = 61/2 − 3/2 = 58/2. This is likely to occur
because the average Cu1−O−Cu2/3 bond angle is ∼120°, which
will be antiferromagnetic and stronger than the Cu2−O−Cu3
interaction (average angle of ∼101°).11a
The large value for the magnetization makes 1 a possible
candidate for low-temperature magnetic cooling, and we have
thus evaluated its MCE, which includes calculation of the
magnetic entropy change ΔSm and adiabatic temperature change
ΔTad for selected applied field changes ΔB0, from the measured
heat capacity (Figure S5 in the SI) and magnetization (Figure 2).
As for the former, ΔSm can be obtained from the T and field
dependencies of the entropy. The results are shown in Figure 3.
We report an appreciable value of −ΔSm, which reaches ∼22.2 J
kg−1 K−1 at T = 2.5 K for ΔB0 = 7 T. This is what we could expect
considering the large net magnetic moment and negligible
anisotropy of the molecule, as well as the weak intracluster
interactions involving GdIII ions that lead to low-lying excited
Figure 1. Structure of 1 (left) and the three different types of constituent
subunits of its core (right). H atoms are omitted for clarity. The circles
and arrows indicate the connection “points” between the subunits.
Color scheme: CuII, cyan; GdIII, yellow; O, red; N, green; C, gray.
passing from the central, unique Gd2 atom) and irregular
topology resulting from many fused triangular subunits. The
metal ions within the core of 1 (Figure S1 in the Supporting
Information, SI) are bridged by six μ3-OH− ions, three
η2:η2:η2:μ5 and an η2:η2:η2:μ6 CO3 ions, and the alkoxido
2−
arms of nine η2:η1:η3:μ4 pdm2− and three η1:η1:η2:μ pdmH2
groups (Figure S2 in the SI). Peripheral ligation about the core is
provided by the chelating part of the pdm2−/pdmH2 groups, two
terminal H2O molecules, and 19 PhCO2− ligands. The latter are
arranged into four classes: seven are monodentate, with the
dangling O atoms hydrogen-bonded to the OH groups of the
pdmH2 ligands, six are bidentate-chelating, three are η1:η1:μ, and
three η1:η2:μ3. All Cu ions are six-coordinate with distorted
octahedral geometries except Cu4 and Cu5 (and their symmetry-
related partners), which are five-coordinate with almost perfect
square-pyramidal geometries (τ = 0.03 and 0.06, respectively).
The Gd ions are nine-coordinate (Gd1 and Gd2) and eight-
coordinate (Gd3) with very distorted geometries. The
complicated core of 1 can be conveniently described as
consisting of a central, nonplanar, μ6-carbonato-bridged
{Cu3Gd3} unit (A), each Cu···Gd edge of which is additionally
linked to three Gd3+ ions by a μ5-CO32− (B). The resulting inner
{Cu3Gd7(CO3)4}19+ subcore (A + B, Figure 1, right) is further
bridged to three extrinsic {Cu4} subunits (C, Figure 1, right)
through the μ3-hydroxido and alkoxido groups. Finally, regarding
the topology of the seven GdIII atoms, this could be roughly
described as distorted, capped trigonal antiprismatic, with Gd2
occupying the capping site (Figure S3 in the SI). The space-filling
representation of 1 reveals its large nanometer-sized structure
with an average diameter of ∼2.6 nm, as defined by the longest
H···H distance (Figure S4 in the SI). Complex 1 is the fourth
largest Cu/Gd cluster isolated to date after the {Cu24Gd8},10a
{Cu24Gd6}10b and {Cu36Gd24}10c cages, while it is one of the
highest-nuclearity heterometallic compounds10c showing MCE
(vide infra).
Solid-state direct-current magnetic susceptibility (χM) data on
dried 1·2H2O were collected in the 2.0−300 K range in applied
B
dx.doi.org/10.1021/ic4020064 | Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Communication
*E-mail: evange@unizar.es.
*E-mail: perlepes@upatras.gr.
*E-mail: tstamatatos@brocku.ca.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the CICYT (Project CTQ2012-
30662) and Excellence in Research ICREA-Academia Award (to
A.E.), the MINECO (Grant MAT2012-38318-C03 to M.E. and
G.L.), an EU Marie Curie IEF (Grant PIEF-GA-2011-299356 to
G.L.), the ARISTEIA Action (Project code 84, acronym
MAGCLOPT) of the Operational Programme “Education and
Lifelong Learning”, cofunded by ESF and National Resources (to
S.P.P.), and the NSERC Discovery Grant (to T.C.S.).
REFERENCES
■
(1) (a) Zheng, Y.-Z.; Evangelisti, M.; Tuna, F.; Winpenny, R. E. P. J.
Am. Chem. Soc. 2012, 134, 1057. (b) Chang, L.-X.; Xiong, G.; Wang, L.;
Cheng, P.; Zhao, B. Chem. Commun. 2013, 49, 1055.
Figure 3. (top) T dependencies of the magnetic entropy change as
obtained from the heat capacity (C) and magnetization (M) data, for the
indicated applied field changes. (bottom) T dependencies of the
adiabatic temperature change obtained from C data, for the indicated
applied field changes.
̌
́ ́ ́
(2) Holynska, M.; Premuzic, Jeon, I.-R.; Wernsdorfer, W.; Clerac, R.;
Dehnen, S. Chem.Eur. J. 2011, 17, 9605.
(3) Evangelisti, M.; Brechin, E. K. Dalton Trans. 2010, 39, 4672.
(4) Evangelisti, M.; Luis, F.; de Jongh, L. J.; Affronte, M. J. Mater. Chem.
2006, 16, 2534.
spin states.3 However, this value is still lower than the maximum
entropy value per mole (25R ∼33.5 J kg−1 K−1) for 15 CuII and 7
GdII fully decoupled ions. In the 2 K < T < 6 K range, the zero-
field magnetic entropy shows a “plateau” at a considerably lower
value, i.e., ∼17R, indeed (Figure S5 in the SI). This is ascribed to
the overall ferrimagnetism of 1, which results from the relatively
strong antiferromagnetic CuII···CuII exchange, and it is thus clear
that this type of interaction is the least favorable for observing the
maximum MCE.11b Fields higher than 7 T are needed in order to
fully break the antiferromagnetic exchanges (Figure S5 in the SI).
With regards to the adiabatic temperature change, Figure 3
(bottom) shows that ΔB0 = 7 T provides a maximum ΔTad = 7.5
K for the same temperatures at which we observe the −ΔSm
maxima. When ΔB0 was lowered to 3 and 1 T, ΔTad decreases to
4.2 and 1.7 K, respectively. Therefore, the field dependence of
ΔTad increases slightly from nearly ∼1 to ∼1.7 K T−1,
respectively, setting this material as a promising refrigerant for
this low-temperature region.3,4,12 The −ΔSm value of ∼22.2 J
kg−1 K−1 for 1 is the second largest in Cu/Gd cluster chemistry,
after the value of ∼31 J kg−1 K−1 reported for a {Cu5Gd4}
compound.11a
(5) Andruh, M.; Costes, J. P.; Diaz, C.; Gao, S. Inorg. Chem. 2009, 48,
3342.
́
(6) (a) Evangelisti, M.; Roubeau, O.; Palacios, E.; Camon, A.; Hooper,
T. N.; Brechin, E. K.; Alonso, J. J. Angew. Chem., Int. Ed. 2011, 50, 6606.
(b) Peng, J.-B.; Zhang, Q.-C.; Kong, X.-J.; Ren, Y.-P.; Long, L.-S.;
Huang, R.-B.; Zheng, L.-S.; Zheng, Z. Angew. Chem., Int. Ed. 2011, 50,
10649. (c) Zheng, Y.; Zhang, Q.-C.; Long, L.-S.; Huang, R.-B.; Muller,
̈
A.; Schnack, J.; Zheng, L.-S.; Zheng, Z. Chem. Commun. 2013, 49, 36.
(7) Anal. Calcd (found) for 1·2H2O: C, 41.34 (41.16); H, 3.12 (3.02);
N, 2.62 (2.86). Crystal structure data for 1·4.5MeCN·1.5H2O:
C230H211.5Cu15Gd7N16.5O83.5, Mw = 6596.51, hexagonal, space group
R3 with a = b = 34.3349(6) Å, c = 37.9490(7) Å, V = 38743.8(12) Å3, T =
̅
160(2) K, Z = 6, R1 [I > 2σ(I)] = 0.0506, wR2 = 0.1236 (F2, all data).
(8) (a) Langley, S. K.; Moubaraki, B.; Murray, K. S. Inorg. Chem. 2012,
51, 3947 and references cited therein. (b) Tian, H.; Zhao, L.; Guo, Y.-
N.; Guo, Y.; Tang, J.; Liu, Z. Chem. Commun. 2012, 48, 708.
́
(9) Mereacre, V.; Ako, A. M.; Clerac, R.; Wernsdorfer, W.; Hewitt, I. J.;
Anson, C. E.; Powell, A. K. Chem.Eur. J. 2008, 14, 3577.
(10) (a) Baskar, V.; Gopal, K.; Helliwell, M.; Tuna, F.; Wernsdorfer,
W.; Winpenny, R. E. P. Dalton Trans. 2010, 39, 4747. (b) Zhang, J.-J.;
Hu, S.-M.; Xiang, S.-C.; Sheng, T.; Wu, X.-T.; Li, Y.-M. Inorg. Chem.
2006, 45, 7173. (c) Leng, J.-D.; Liu, J.-L.; Tong, M.-L. Chem. Commun.
2012, 48, 5286.
−
In conclusion, we have shown that the Cu2+/Gd3+/PhCO2 /
(11) (a) Langley, S.; Chilton, N. F.; Moubaraki, B.; Hooper, T.;
Brechin, E. K.; Evangelisti, M.; Murray, K. S. Chem. Sci. 2011, 2, 1166.
(b) Hooper, T. N.; Schnack, J.; Piligkos, S.; Evangelisti, M.; Brechin, E.
K. Angew. Chem., Int. Ed. 2012, 51, 4633.
(12) Colacio, E.; Ruiz, J.; Lorusso, G.; Brechin, E. K.; Evangelisti, M.
Chem. Commun. 2013, 49, 3845.
pdmH2 system is a promising reaction scheme to follow, not only
for the synthesis of new high-nuclearity heterometallic
complexes with beautiful topologies but also for magnetically
interesting species with promising low-T MCE. We are currently
focusing on the effect of the nature of the carboxylate group on
the {Cu15Gd7} identity and the isolation of Cu2+/Gd3+ magnetic
coolers with larger MCE.
ASSOCIATED CONTENT
* Supporting Information
Crystallographic data (CIF format), synthetic details, and various
figures for 1. This material is available free of charge via the
Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Authors
*E-mail: albert.escuer@ub.edu.
■
C
dx.doi.org/10.1021/ic4020064 | Inorg. Chem. XXXX, XXX, XXX−XXX