Content uploaded by Emilio Pardo
Author content
All content in this area was uploaded by Emilio Pardo on Mar 03, 2022
Content may be subject to copyright.
Switching of easy-axis to easy-plane anisotropy in cobalt(II) complexes
Yuewei Wu, Jing Xi, Tongtong Xiao, Jesús Ferrando-Soria, Zhongwen
Ouyang, Zhenxing Wang, * Shuchang Luo, Xiangyu Liu *
and Emilio Pardo *c
Abstract
Under the guidance of in situ microcalorimetry, a tetranuclear cubane-type complex
[Co4(ntfa)4(CH3O)4(CH3OH)4] (1) with a {Co4O4} core, and a mononuclear complex
[Co(ntfa)2(CH3OH)2] (2) have been manipulated by adjusting the ratio of the β-
diketonate and Co(II) ions. Then, the use of three N-donor coligands, 2,2'-bipyridyl
(bpy), 6,6'-dimethyl-2,2'-bipyridyl (6,6-(CH3)2-bpy) and 5,5'-dimethyl-2,2'-bipyridyl
(5,5-(CH3)2-bpy), replaces two coordinated CH3OH molecules in 2, leading to three
new configurations of 3-5, [Co(ntfa)2(bpy)2] (3), [Co(ntfa)2(6,6-(CH3)2-bpy)2] (4) and
[Co(ntfa)2(5,5-(CH3)2-bpy)2] (5). Although X-ray crystallography shows that
complexes 2-5 are mononuclear with distorted octahedral geometries around the CoII
ions, the introduction of different capping coligands fine-tunes the structures involving
changes in both the distortion degree of the coordination geometry and the
intermolecular interactions, which has impact on magnetic properties of these
complexes. Magnetic investigations reveal field-induced single-ion magnet behavior in
all complexes with distinct energy barriers Ueff of 39.06 K (1), 36.65 K (2), 36.32 K (3),
28.26 K (4) and 15.85 K (5). Remarkably, magnetic experiments, HF-EPR
measurements and theoretical calculations demonstrate that 2 features easy-axis
magnetic anisotropy (D = -60.48 cm-1), whereas the easy-plane magnetic anisotropies
are observed in complexes 3-5 with D = +72.85 cm-1 for 3, +35.71 cm-1 for 4, +51.28
cm-1 for 5. To our knowledge, such reversal of anisotropic nature driven by alternative
coligands is unprecedented.
Introduction
Single-molecule magnets (SMMs) have been intensively studied for more than two
decades due to their significant potential applications in high-density spin-based
information storage and spintronics.1 Historically, SMMs were transition metal
containing polynuclear complexes possessing a high spin ground state associated with
magnetic anisotropy leading to an energy barrier to the reorientation to their
magnetization. More recently, research has been turned to mononuclear SMMs, which
are the so-called single-ion magnets (SIMs) of which the magnetic behaviour primarily
arises from the inherent magnetic anisotropy of the metal center.2 Among 3d-SIMs,
Co(II)-based complexes are most interesting because of their non-integer spin ground
state which declines the probability of quantum tunnelling of magnetization (QTM).3
Investigations on existing mononuclear Co(II)-based SIMs verify that the anisotropic
nature and magnitude of Co(II) centers is greatly sensitive to tiny modifications of
ligand field and coordination geometry as well as coordination environment. Contrary
to traditional SMMs whose axial zero-field splitting (ZFS) parameter is certainly
negative, the sign and magnitude of magnetic anisotropy in Co-based SIMs depend on
more varied and complex parameters.4 In this context, as Ruiz and Luis et al. showed
for a range of mononuclear Co(II) complexes, strong magnetic anisotropies with both
a negative ZFS parameter D (D < 0) and an easy-plane anisotropy (D > 0) can be
obtained for mononuclear Co(II) complexes, and thus, field-induced slow magnetic
relaxation can be observed regardless of the sign of the D values.5 Special attention, for
example, has been paid to establishing a magneto-structural correlation based on D for
mononuclear hexa-coordinated d7 complexes, especially those Co(II) SIMs with large
positive or negative D values.6 However, even though a remarkable number of Co(II)-
based SIMs have been reported so far and great efforts have been made towards finding
a solid conceptual explanation for this behavior, the parameters governing magnetic
anisotropy are still poorly understood and the daunting task of having precise control
over the magnetic anisotropy remains a difficult challenge.
It is well-established that metal complexes are prepared using complicated
multicomponents and heterophase systems in which unobservable physico-chemical
processes occur. Selecting suitable ligands is one of the important factors in forming
the desirable ligand fields and impacting the magnetic property for resulting complexes.
As ideal candidate, β-diketonate and its derivatives have been considered to fabricate
monometallic complexes and thus trigger the single-ion magnetic anisotropy resulting
from their intrinsic characteristics of stable bidentate modes chelating to metal ions and
offering proper ligand fields.7 Meanwhile, the introduction of capping N-donor ligands
into the β-diketonate-metal systems is beneficial to obtain novel complexes with Oh
symmetry and modify their SIM properties.8 Besides, the structural formation of a
compound is often perturbed by the intricate synthetic conditions. A crucial problem is
how to predict and manipulate the formation process of targets.
In view of the coordination geometry of hexa-coordinated Co(II) SIMs, we employed
a β-diketonate, namely, 4,4,4-trifluoro-1-(2-naphthyl)-1,3-butanedione (ntfa), to
assemble Co(II) complexes with expectant octahedral geometries. Guided by in situ
microcalorimetry, two closely related complexes, [Co4(ntfa)4(CH3O)4(CH3OH)4] (1)
and [Co(ntfa)2(CH3OH)2] (2) have been successfully synthesized. Subsequently, 2
could be identified as the precursor contributing to the formation of [Co(ntfa)2(bpy)2]
(3), [Co(ntfa)2(6,6-(CH3)2-bpy)2] (4) and [Co(ntfa)2(5,5-(CH3)2-bpy)2] (5), while the
coordinated CH3OH solvents are substituted by different capping N-donor coligands.
Interestingly, a combination of magnetism, high-frequency electron paramagnetic
resonance (HF-EPR) spectroscopy and ab initio calculation confirms that complex 2
presents easy-axis magnetic anisotropy, whereas easy-plane magnetic anisotropies are
exhibited with other three mononuclear complexes. Moreover, all complexes are
indicative of the field-induced slow magnetic relaxation.
Experimental
Materials and methods
Elemental analysis was recorded on a Perkin-Elmer 2400 CHN analyzer. IR spectra
were implemented on an EQUINOX55 FT-IR spectrophotometer by using KBr pellets,
in the range 400-4000 cm-1. Powder X-ray diffraction (PXRD) measurements were
recorded on a Rigaku RU200 diffractometer at Cu-Kα radiation (λ = 1.5406 Å) with a
step size of 0.02° in 2θ and a scan speed of 5° min−1. The calorimetric experiment was
performed by using a RD496 type microcalorimeter.9 Magnetic experiments were
performed with a Quantum Design MPMS-XL7 SQUID magnetometer on
polycrystalline samples for all complexes (restrained in eicosane to prevent torqueing
at high fields). Alternating current (ac) magnetic susceptibility measurements were also
carried out with a Quantum Design Physical Property Measurement System (PPMS).
Diamagnetic corrections were evaluated from Pascal’s Tables. High frequency/field
electron paramagnetic resonance (HF-EPR) were measured on a locally developed
instruments at the Wuhan National High Magnetic Field Center with pulsed magnetic
fields.10
Synthesis of [Co4(ntfa)4(CH3O)4(CH3OH)4] (1). To a solution of ntfa (79.9 mg, 0.3
mmol) in methanol (15 mL) was added Et3N (0.014 mL, 0.1 mmol). After stirring for
30 min, CoCl2·6H2O (71.4 mg, 0.3 mmol) were added to the solution, which was stirred
for 24 h at room temperature. The filtrate was allowed to stand at room temperature,
orange crystals had formed after five days and were collected by filtration. Yield: 81%.
Anal. calcd for C40H38CoF4N4O6 (M = 805.67). 1: C 59.63, H 4.75%. Found: C 59.60,
H 4.72, N 6.90%. IR data (KBr, cm-1): 3429(w), 3059(w), 2927(w), 1620(m), 1516(m),
1287(m), 1190(m), 1135(s), 1115(s), 801(m), 760(m), 592(w).
Synthesis of [Co(ntfa)2(CH3OH)2] (2). Complex 2 was prepared by the similar way of
1, while the ratio of CoCl2·6H2O/ntfa is altered to be 1:2. Yellow crystals of 2 were
obtained for 10 days (Yield 63%, based on Co). Anal. calcd for C30H24CoF6O6 (M =
653.42): C, 55.14; H, 3.70%. Found: 55.08; H, 3.57%. IR (KBr, cm-1): 3060(w), 1614(s),
1596(s), 1535(s), 1300(s), 1254(m), 1200(m), 1139(m), 795(m), 691(m), 584(w),
521(w).
Synthesis of Co(ntfa)2(X) [X = bpy (3), 6,6-(CH3)2-bpy (4), 5,5-(CH3)2-bpy (5)].
These complexes were prepared as 2, but two CH3OH molecules was replaced by the
capping ligand X (X = bpy, 6,6-(CH3)2-bpy and 5,5-(CH3)2-bpy). Data for 3: yield: 75%
(based on Co); Anal. calcd for C38H24CoF6N2O4 (M = 745.52): 61.22; H, 3.25; N, 8.58%;
Found: C, 61.09; H, 3.12; N, 8.48%; IR (KBr, cm-1): 3432(w), 2925(w), 1605(m),
1298(m), 1128(s), 867(w), 790(w), 689(w), 637(m), 619(s), 471(w). Data for 4: yield:
67% (based on Co); Anal. calcd for C40H28CoF6N2O4 (M = 773.57): C, 62.11; H, 3.65;
N, 8.27%; Found: C, 62.03; H, 3.51; N, 8.22%; IR (KBr, cm-1): 3337(w), 1610(s),
1592(s) 1559(s), 1533(m), 1297(s), 1203(s), 1145(s), 1022(m), 855(m), 797(s), 691(m),
582(m). Data for 5: yield: 65% (based on Co); Anal. calcd for C40H28CoF6N2O4 (M =
773.57): C, 62.11; H, 3.65; N, 8.27%, Found: 62.06; H, 3.53; N, 8.20%; IR (KBr, cm-
1): 1611(s), 1592(s) 1569(m), 1534(m), 1292(s), 1203(s), 1139(s), 1022(m), 797(s),
591(m).
X-ray crystallography
Suitable crystals of both complexes were selected for X-ray measurements. Crystal
structures were collected with a Bruker SMART APEX-CCD-based diffractometer
using graphite monochromated Mo-Kα radiation (λ = 0.71073 Å). Data processing and
absorption corrections were accomplished using SAINT and SADABS.11 The structures
were solved by direct methods and refined against F2 by full-matrix least-squares with
SHELXTL-2014.12 All non-hydrogen atoms were refined with anisotropic thermal
parameters. All hydrogen atoms were placed in calculated positions and refined
isotropically. Crystallographic data for all complexes is provided in Table S1. Selected
bond lengths and angles are listed in Tables S2-S6.
Theoretical methods
The theoretical calculations were performed with the ORCA 4.2.1 computational
package.13 Based on X-ray by comparing geometries, B3LYP DFT functional14 was
used for calculations of the magnetic exchange constants J by comparing the energies
of high-spin (HS) and broken-symmetry spin (BS) states. The polarized triple-ζ quality
basis set def2-TZVP proposed by Ahlrichs and co-workers was used for all atoms.15
The single-ion zero-field splitting parameters and g-factors were calculated using the
state average complete active space self-consistent field (SA-CASSCF)16 wave
functions complemented by the N-electron valence second order perturbation theory
(NEVPT2) with CAS(7,5) active spaces. 17 In the state averaged approach all multiplets
for the given electron configuration were equally weighted, which means 10 quartet
and 40 singlet states. The ZFS parameters, based on dominant spin-orbit coupling
contributions from excited states, were calculated through the quasi-degenerate
perturbation theory (QDPT),18 in which approximations to the Breit-Pauli form of the
spin-orbit coupling operator (SOMF approximation)19 and the effective Hamiltonian
theory were utilized.20 In all calculations, the polarized triple-ζ quality basis set def2-
TZVP(-f) proposed by Ahlrichs and co-workers was used for all atoms.15 The
calculations utilized the RI approximation with the decontracted auxiliary def2-
TZVP/C Coulomb fitting basis set21 and the chain-of-spheres (RIJCOSX)
approximation to exact exchange.22 Increased integration grids (Grid6) and tight SCF
convergence criteria were used in all calculations.
Results and discussion
Synthesis and thermodynamic behaviors for 1 and 2
In general, the influence of the reaction conditions on the reaction processes is
considerable, which may produce the diversity of structures in a specific system. To
better govern and cognize the synthetic processes of resulting products, a detailed study
on the chemical reaction is of great significance. It is known that in situ
microcalorimetry is a powerful tool for investigation of thermal events during the
reaction processes,23 which would provide essential thermodynamics support to
understand the self-assembly mechanism. Therefore, we online monitored the reaction
processes using in situ microcalorimetry and quantitatively obtained the apparent
energy change for the syntheses.24 When the molar ratio of metal ions with ntfa is
adopted to be 1:1, the reaction system on-time monitored by a microcalorimeter shows
that there is one obvious exothermic peak with Q value of 286.6 mJ. By reducing the
concentration of cobalt ions to 50%, it is observed that the Q value minishes to 92.9 mJ,
less than half of the former. Correspondingly, the response time of the latter process is
95 seconds that is distinctly shorter than that of the former process (175 s). From this
point of view, the observation suggests that the reaction system implies two completely
different assembly behaviours, portending the formations of two different products,
respectively. As described above, the reaction of CoCl2·6H2O with ntfa in a 1:1 molar
ratio affords single crystals of 1, whereas complex 2 is prepared in a 1:2 metal to ntfa
molar ratio. By means of varying the ratios of cobalt(II) and ntfa, the formation of
complexes allowed us to obtain tetranuclear (1) and mononuclear (2) cobalt(II) species.
Fig. 1 The black and red lines represent the Heat-flow curve of the mixture of CoCl2·6H2O and ntfa
in 1:1 and 1:2 molar ratio, respectively.
Crystal Structure
X-ray structural analysis reveals that complex 1 is tetranuclear structure that crystallize
in the monoclinic space group P21/c (Table S1). The molecule has a distorted cubane-
type {Co4O4} core with four six-coordinated CoII atoms occupying four alternating
corners of the cube and the other corners are occupied by oxygen atoms of MeOH
molecules (Fig. 2a). The overall arrangement has an approximate S4 symmetry. Similar
to the DPM analogue, 25 the cuboidal core is distorted, with all the O-Co-O angles are
smaller than 90°, while all Co-O-Co angles are greater than 90° (Fig. 2b). Moreover,
the Co···Co distances in cuboidal core ranging from 3.079 to 3.177 Å (Table S2).
Among them, each distorted octahedral CoII ion is coordinated to three μ3-OMe
molecules, chelated by a bidentate 4,4,4-trifluoro-1-(2-naphthyl)-1,3-butanedione (ntfa)
ligand and one methanol molecule.
Fig. 2 The asymmetric unit of complex 1 (a) and emphasis of the cubane-like {Co4O4} fragment of
2(b). H atoms are omitted for clarity.
X-ray crystallography showed that complexes 2-5 are mononuclear with distorted
octahedral geometries around the CoII ions. Complexes 2-4 belong to the triclinic P-1
space group, whereas complex 5 crystallize in the monoclinic system with P21/c space
group (Table S1). The CoII ion in 2 is hexa-coordinated by six O atoms from two ntfa
ligands and two methanol molecules (Fig. 3a). The axial Co-O bond distances (2.130
Å) are longer than the equatorial Co-O bond distances (2.035 Å and 2.048 Å), leading
to stretched octahedral geometries (Table S3). Evidently, two coordinated CH3OH
molecules in 2 are replaced by 2,2'-bipyridyl (bpy), 6,6'-dimethyl-2,2'-bipyridyl (6,6-
(CH3)2-bpy) and 5,5'-dimethyl-2,2'-bipyridyl (5,5-(CH3)2-bpy) ligands, forming the
configurations of 3-5. Consequently, the CoII centers in 3-5 link with six donor atoms
consisting of four O atoms from two ntfa ligands and two N atoms from neutral
coligands (Fig. 3b-3d). For 3-5, the average Co-O bond lengths are 2.06, 2.07 and 2.06
Å, and the average Co-N bond lengths are 2.12, 2.15 and 2.10 Å, respectively (Tables
S4-S6). Thus, almost identical Co-O bond lengths are observed for 3-5. In turn, the Co-
N distance for 4 is appreciably greater than that in other two complexes. It is worth
noting that the smallest intermetallic distance in complex 2 is 5.238 Å, which illustrates
non-negligible intermetallic interaction. In contrast, the metal centers are well-
separated with the shortest Co···Co distance of 7.504, 7.189 and 7.531 Å for 3-5,
respectively, thus excluding potential intermolecular dipole-dipole interactions.
Fig. 3 Crystal structures of 2(a), 3(b), 4(c) and 5(d) of the Co(II) ions. Hydrogen atoms are omitted
for clarity.
The geometric configurations of the CoII cations in 1-5 were calculated by using the
SHAPE 2.1 program (Tables S7 and S8).26 The calculated parameter from the software
is zero for the ideal structure, and larger values indicate greater deviation from the ideal
polyhedron. The investigation of the precise configurations demonstrated that CoII
cations in 1-5 fit well with the hexa-coordinate octahedral polyhedron with continuous
shape measures (CShMs) of 0.445 (Co1), 0.448 (Co2), 0.533 (Co3) and 0.417 (Co4)
(1), 0.057 (2), 1.030 (3), 0.857 (4) and 0.698 (5).
Static (dc) magnetic properties and HF-EPR measurements
Magnetic studies were performed on polycrystalline samples of complexes 1-5, and
the phase purity of the bulk materials was confirmed by powder XRD (Fig. S1).
Variable-temperature dc magnetic susceptibility data of these complexes were
measured in the 2-300 K temperature range with an applied field of 1000 Oe. For
complex 1, as shown in Fig. 4a, χMT rises slowly from 13.98 cm3 K mol-1 with
decreasing temperature, then increases sharply to a maximum of 17.14 cm3 K mol-1 at
10 K, finally rapidly drops to 15.04 cm3 K mol-1 at 2 K. From the value of the peak at
10 K, it appears that all four Co2+ ions are ferromagnetically coupled to give an S = 6
ground state. The decrease in χMT below 10 K is probably due to zero-field splitting,27
and also due to the more common complications arising from spin-orbital interaction,28
which is a frequent source of difficulty in the interpretation of magnetic data for Co2+
complexes.
Fig. 4 Plots of χMT versus T for complex 1(a). Inset: The experimental plots of M versus H at
different temperatures. Plots of M vs H/T for 1(b) at different temperatures.
At room-temperature, the χMT values of 2-5 are 3.52 cm3 K mol-1, 2.53 cm3 K mol-1,
3.11 cm3 K mol-1 and 2.89 cm3 K mol-1, respectively. These values are clearly larger
than the spin-only value (1.875 cm3 K mol-1) for a magnetically isolated Co(II) cation
(S = 3/2 and g = 2.0), indicating a significant orbital contribution to the magnetic
moment. Upon cooling, the χMT values decreases monotonously to 2 K, reaching 2.37,
1.45, 1.88 and 1.66 cm3 K mol-1 for 2-5, respectively. The field-dependent
magnetizations in the form of the M vs. H plots for 2-5 are represented in the inset of
Fig. 5. For these complexes, the M vs. H/T plots at different temperatures are not
superimposable (Fig. S2), clearly confirming the presence of a significant magnetic
anisotropy, which is derived from the well-known strong spin-orbital coupling (SOC)
of the Co(II) ion.29 To gain insight into the magnetic anisotropy of complexes 2-5, the
reliable ZFS parameters were obtained by simultaneously fitting the experimental M vs.
H curves using the PHI program30 based on the following spin Hamiltonian of eq (1):
(1)
where µB is the Bohr magneton, D is the axial ZFS parameter, E is the rhombic or
transverse ZFS parameter, S is the spin operator, and B is the magnetic field vector,
respectively. The parameters D, E, and g were selected to correlate the data. The best
fit values are summarized with parameters: D = -60.48 cm-1, |E| = 15.27 cm-1, gx,y =
3.56, gz = 2.20 for 2, D = +72.85 cm-1, |E| = 11.57 cm-1, gx = 2.34, gy = 2.28, gz = 2.12,
for 3, D = +35.71 cm-1, |E| = 4.81 cm-1, gx,y = 2.14, gz = 2.12 for 4 and D = +51.28 cm-
1, |E| = 13.41 cm-1, gx,y = 2.57, gz = 2.56 for 5, respectively. Note that the negative sign
of the D values for complex 2 illustrates easy axial magnetic anisotropy, whereas the
strong easy-plane magnetic anisotropy observed in complexes 3-5.
Fig. 5 Plots of χMT versus T for complexes 2-5. Inset: The experimental plots of M versus H at
different temperatures. The solid lines show the best-fitting curves to the experimental data.
HF-EPR measurements were recorded on polycrystalline samples of 2-5 at
frequencies of up to 170 GHz in order to further confirm the ZFS parameters. As shown
in Fig 6, we can easily get the EPR spectra containing three main modes, which is the
typical feature of a high spin cobalt(II) ion compound with S = 3/2. As expected, like
the magnitudes of the D values are out of the frequency range in our measurements, no
transitions between Kramers doublets MS = ±1/2 and MS = ±3/2 were observed (Fig. 6).
All the EPR signals can be interpreted as from the intra-Kramers transitions within the
lowest doublet MS = ±1/2 multiplet with ΔMS = ±1. The relationship of the resonance
fields and its corresponding various microwave frequencies curve were shown in Fig.
7. The resonance fields were simulated using the |D| value of 60.48 cm-1, 72.85 cm-1,
51.28 cm-1 and 35.71 cm-1 from SQUID measurements while adjusting E (transverse
zero-field splitting parameter) and g values to get the well-estimated data.
Fig. 6 HF-EPR spectrum of 2(a), 3(b), 4(c) and 5(d) at 4.2 K and its simulations (blue trace: positive
D; red trace: negative D) at 120 GHz.
For complex 2, two simulations were done with different signs of D, showing that
the negative D value are well in accord to the experimental data. But beyond that, there
are only two peaks in the spectra, which is typical for high-spin 3/2 Co(II) systems with
large negative D values due to the limit of magnetic field, indicating the easy-axis
magnetic anisotropy in 2.31 The situation above were confirmed by [Co(hfac)2(H2O)2]32
and [Co(acac)2(H2O)2].33 In contrast, the simulation with D > 0 fit better to the
experiment data than did those with D < 0 for 3-5, determining that the D values are
positive. The positive D values of 3-5 might derive from the spin-orbital coupling of
the ground state and excited state electrons, which further demonstrates the easy-plane
magnetic anisotropies.
Fig. 7 Resonance field vs. microwave frequency (quantum energy) for EPR transitions for 2(a), 3(b),
4(c) and 5(d). Simulations were done using the Hamiltonian parameters taken from Fig. 6 The solid
lines show the (x, y, z) transitions as labeled. The vertical dashed lines represent the frequency (120
GHz) used in Fig. 6 at which the spectra were recorded or simulated.
Dynamic (ac) magnetic properties
For the purpose of probing the spin dynamics, alternating current (ac) magnetic
susceptibility measurements were conducted at zero dc field at a frequency of 1000 Hz.
No out-of-phase (χ″M) signals were observed for all complexes until the temperature
dropped to 2 K (Fig. S3), signifying a fast quantum tunneling of the magnetization
(QTM) at low temperature. In order to find a suitable applied magnetic field to suppress
the QTM, the χ″M susceptibilities for all complexes at 2.0 K and 1000 Hz were recorded
under different magnetic fields. The χ″M signals with significant peak values at around
2000 Oe dc field suggest that field-induced slow magnetic relaxation and slow
relaxation operate in five complexes. Thus, 2000 Oe was used as a suitable applied field
for 1-5, and in-phase and out-of-phase ac susceptibilities were clearly observed. The
temperature dependence of in-phase (χ′) and out-of-phase (χ″) products were measured
in the temperature ranges 2-6 K for complex 1 and 2-10 K for 2-5, respectively (Fig.
S4 and S5). The downturn in both the χ′ and χ″ susceptibilities in the low temperature
range and the appearance of obvious peaks for all complexes indicate that the relaxation
probability via the quantum pathway has been obviously weakened or suppressed. At
the selected frequency, all complexes go through a maximum and maxima shift to high
temperature with increasing frequency, which is characteristic of a superparamagnet.
Fig. 8 Frequency dependence of the out-of-phase ac susceptibility signals for complexes 1-5 under
a 2000 Oe dc field.
In addition, the frequency dependencies of the ac susceptibility were measured under
an applied dc field of 2000 Oe (Fig. 8 and S6). Both the χ′ and χ″ signals of 1-5 appear
to be frequency dependent. As the temperature increases, the peaks of χ″ in 1-5
gradually shift from low frequency to high frequency. Fitting the high-temperature data
using the Arrhenius law τ = τ0exp(Ueff/kBT) affords the effective spin-reversal energy
barrier (Ueff) and pre-exponential factor τ0 (Fig. 9): 1, Ueff = 30.28 K, τ0 = 5.4 × 10-9 s;
2, Ueff = 29.29 K, τ0 = 1.21 × 10-7 s; 3, Ueff = 29.74 K, τ0 = 8.1 × 10-8 s; 4, Ueff = 25.81
K, τ0 = 7.52 × 10-7 s; 5, Ueff = 13.03 K, τ0 = 1.64 × 10-6 s. Furthermore, the curvature
that emerged in the Arrhenius plots of five complexes also implies non-negligible direct
and/or Raman processes in determining the relaxation rate. Thereby, a model including
three possible relaxation processes, i.e. direct, Raman and Orbach mechanisms,34 was
employed to analyze the contribution to the relaxation in 1-5 using eqn (1):
(1)
where the terms in eqn (1) represent the contributions of multiple mechanisms. For the
second term, n = 7 is expected for the Raman process in non-Kramers ions and n = 9
for Kramers ions, while n = 1-6 can occur for the optical acoustic Raman-like process.
As depicted in Fig. 9, the fitting reproduces the experimental data very well, resulting
in the parameters 1, A = 611 K-1 s-1, C = 83.25 K-n s-1, n = 4.88, τ0 =1.95 × 10-9 s, ΔE/kB
= 39.06 K; 2, A = 275 K-1 s-1, C = 1.65 K-n s-1, n = 4.92, τ0 = 3.61 × 10-8 s, ΔE/kB =
36.65 K; 3, A = 186 K-1 s-1, C = 2.48 K-n s-1, n = 5.56, τ0 = 5.85 × 10-8 s, ΔE/kB = 36.32
K; 4, A = 55.49 K-1 s-1, C = 3.42 K-n s-1, n = 4.41, τ0 = 9.77 × 10-7 s, ΔE/kB = 28.26 K;
5, A = 832 K-1 s-1, C = 43.18 K-n s-1, n = 3.87, τ0 = 1.73 × 10-6 s, ΔE/kB = 15.85 K. For
1-3 and 5, the low temperature region is probably dominated by a direct process,
whereas the relaxation process at high temperature can be mainly attributed to an optical
acoustic Raman-like mechanism. In the case of 4, it was observed that the contributions
of the Orbach and direct processes are small compared with the optical acoustic Raman
process.
Fig. 9 Magnetization relaxation time, lnτ vs. T-1 plot under a 2000 Oe dc field for 1-5. The red and
green lines represent the Arrhenius fit and multiple relaxation processes, respectively.
Cole-Cole diagrams for 1-5 were also obtained (Fig. 10). As shown, individual
semicircular shapes can be evidently noticed in the motifs, which could be well-fitted
using the generalized Debye model.35 As listed in Tables S9-S13, the α parameters are
0.10-0.18 for 1, 0.006-0.014 for 3 and 0.016-0.12 for 4, suggesting narrow distributions
of the relaxation processes. In contrast, the α values are 0.012-0.34 for 2 and 0.013-0.26
5, demonstrating a wide distribution of relaxation times.
Fig. 10 Cole-Cole plots under 2000 Oe for 1-5. The solid lines show the best fitting according to
the generalized Debye model.
Theoretical calculations
Before the analysis of the experimental magnetic data of five complexes, the predictive
role of ab initio calculations in magnetochemistry was utilized with the aim of
estimating the principal parameters describing the exchange coupling and the zero-field
splitting in these molecular systems.36 Therefore, Density Functional Theory (DFT)
was applied to calculate the parameters of the isotropic exchange J between the
paramagnetic ions, and the multireference method based on Complete Active Space
Self Consistent Field (CASSCF) was utilized to derive information about the single-ion
zero-field splitting tensor parameters D and E. Having such information at our disposal,
the trustworthy spin Hamiltonians can be postulated and the calculated parameters can
be used as a starting point for fitting of the experimental magnetic data. All theoretical
calculations were performed with a freely available computational package ORCA.13a
All the calculations were done for molecular geometries extracted from X-ray data and
also for molecular geometries optimized with the BP86 functional (see the
Experimental section for more details).
We carried out DFT calculations on structures 1 to help rationalise the experimental
information on molecular magnetism. Since the DFT deals with single Slater
determinants to describe open-shell electronic configurations instead of spin-adapted
states, we applied the broken-symmetry approach to compute the energies of the open-
shell configurations, namely those with MS ≤ SHS. By inspection of the structural
parameters, we considered a two-J model based on the presence of four shorter (J1) and
two longer (J2) Co···Co distances to determine the sign and magnitude of the magnetic
interactions between the local spin moments of Co(II) ions. As deduced from the χMT
values at high T shown in Fig. 5, the Co(II) ions in 1 are characterised by the local spins
of 3/2. Thus, to extract the J1 and J2 parameters we computed the three situations
depicted in Fig. 11: the high-spin (|αααα>, SHS = 6), the intermediate-spin (|βααα>, SBS1
= 3) and the low-spin (|βααβ>, SBS2 = 0) configurations, where each arrow represents
spin 3/2 for Co(II). (Fig. 11 and Table S14)37 The magnetic coupling constant between
magnetic centers was modeled using the Heisenberg approach:
1 2 3 3 4 1 4 1 2 2 1 3 2 4
ˆ( ) ( )H J S S S S S S S S J S S S S
where each arrow represents spin 3/2 for Co. From this model, the energy spectrum is:
21 2
9
9JJEHS
0
1
BS
E
212 2
9
9JJEBS
2111 2
9
9JJEE HSBS
a
122 18JEE HSBS
a
The calculation details are listed in Table S14 and the spin densities of the HS spin
states are shown in Fig. S7 for 1. The partial spin delocalization from the metal atoms
to the respective donor atoms is clearly visible. Next, the J-parameters calculated
according to eqn (2) adopted the values J1 = 5.663 cm-1 and J2 = 10.121 cm-1. Evidently,
there are dominant antiferromagnetic interactions in 1, but there is no simple correlation
of the calculated J-parameters with either the Co-O-Co angle or the Co···Co distance.
Fig. 11 Three possible orientations of local spin moments (arrows: spin 3/2 for Co) and total (S,
sum of arrows) spin quantum numbers for the cubane-type Co4 complex. The coupling constants J1
and J2 represent the strength of the interactions between the Co ions bridged by μ3-O ligands.
{d23=3.0791, d34=3.0797, d12=3.0899, d14=3.0812(J1); d13=3.1563, d24=3.1774(J2)}
Furthermore, in order to better understand the impact of the auxiliary ligands on the
electronic structure of the six-coordinate complexes under study, especially on their
magnetic anisotropy, post-Hartree-Fock CASSCF calculations were performed. The
approach done with ORCA resulted in large D-values (-29.30 cm-1, 59.18 cm-1, 51.82
cm-1 and 45.73 cm-1 for 2-5, respectively) (Table 1). Although the calculated values
show a certain deviation from the fitting values, the sign of the D constant matches well
with the experimental values, which definitely confirms the easy-axis magnetic
anisotropy of the complex 2 and easy-plane magnetic anisotropy of 3-5. It is difficult
to estimate the accuracy of the theoretical calculations, but the work is only qualitative.
This may result from the fact that the real complexes are not made up of scattered
entities as they have been modelled, but are very complicated across the whole
structures. The calculated effective gz tensors are 2.365 (gx,y = 2.034, 2.268), 2.669 (gx,y
= 2.054, 2.426), 2.666 (gx,y = 2.044, 2.392) and 2.539 (gx,y = 2.071, 2.428) for
complexes 2-5, respectively. The energy levels and the contributions of the excited spin
states to the D-tensors are listed in Tables S15-S22. The results manifest that the first
two excited quartets have a significant contribution to the D parameters for all
complexes.
Table 1 ORCA/CASSCF+NEVPT2 computed D, E, and g value for complex 2-5
Complex
D (cm-1)
(experiment)
D (cm-1)
(calculation)
E (cm-1)
giso
gx
gy
gz
2
-60.48
-29.304
-9.559
2.222
2.034
2.268
2.365
3
72.85
59.179
9.898
2.383
2.054
2.426
2.669
4
35.71
51.822
13.470
2.367
2.044
2.392
2.666
5
51.28
45.731
6.386
2.346
2.071
2.428
2.539
Magneto-structural correlations
To understand the origin of the magnetization dynamics, it is necessary to provide a
structural comparison of the coordination spheres for the four mononuclear complexes.
Structurally, complex 2 exhibits a CoO6 chromophore with slightly distortion (CShMs
= 0.057) from the ideal octahedron, whereas the Co(II) ions in 3-5 present CoN2O4
chromophore with obvious distortion (CShMs = 1.030, 0.857 and 0.658 for 3-5,
respectively) from the corresponding ideal geometry. Consequently, the energy barrier
for 2 is larger than that in other three complexes. For 3-5, the discrepancies of magnetic
anisotropy and energy barrier dominantly depends on the different terminal substituents
of capping N-donor ligands. Among them, 2,2'-bipyridyl coligand in complex 3 is
conducive to promoting the single-ion anisotropy and thus the slow relaxation of the
magnetization, achieving enhanced energy barrier, suggesting that the electron-
donating -CH3 groups in 4 and 5 produce negative effects on the dynamic magnetic
properties. Complexes 4 and 5 containing two isomeric capping N-donor coligands
where two -CH3 groups are located on the different positions display distinctly different
energy barriers, confirming the significance of the substituent effect on the single-ion
behaviour. To further acquire the deep understanding for the variously anisotropic
nature of complexes 2-5, the charge distributions of Co(II) ions and coordination atoms
are considered concurrently (Table S23). Evidently, the charge distributions derived
from the changes of substituents lead to different bond lengths, which further has a
significant effect to alter the distortion degree from the ideal octahedral geometry of the
complexes. As a result, the reversals of the anisotropic sign have been observed.
Conclusion
In summary, we report here two completely different Co(II)-containing complexes,
[Co4(ntfa)4(CH3O)4(CH3OH)4] (1) and [Co(ntfa)2(CH3OH)2] (2), that are successfully
guided via in situ microcalorimetry. Three new mononuclear complexes,
[Co(ntfa)2(bpy)2] (3), [Co(ntfa)2(6,6-(CH3)2-bpy)2] (4) and [Co(ntfa)2(5,5-(CH3)2-bpy)2]
(5), are evolved from 2 through combining the corresponding capping N-donor
coligands. Co(II) ions in all cases are encompassed by the octahedral coordination
geometries with variously architectural distortions. The coligand effect on the magnetic
dynamics and magnetic anisotropy of the octahedral Co(II) centers is systematically
investigated. All complexes are characteristic of field-induced slow magnetic relaxation,
with energy barriers Ueff of 39.06 K (1), 36.65 K (2), 36.32 K (3), 28.26 K (4) and 15.85
K (5). Excitedly, it is first noted that the coligand-induced structural perturbation
reverses the sign of the single-ion anisotropy from easy-axis type for 2 to easy-plane
type for 3-5. The outcomes illustrated in this work would offer new possibility for the
accurate assessment of controllable preparation by in situ microcalorimetry, and
provide solid evidence of the effects of coligand variations on magnetic anisotropy and
magnetization dynamics in Co(II)-based SIMs, the ultimate goal of which is to advance
the deliberate tailoring of SIMs. Further studies following this guideline are actually
ongoing.
Acknowledgements
We gratefully acknowledge the financial support from the National Natural Science
Foundation of China (21863009, U20A2073), the Natural Science Foundation of
Ningxia Province (2020AAC02005), the Discipline Project of Ningxia
(NXYLXK2017A04), and the Guizhou Education Department Youth Science and
Technology Talents Growth Project (No. KY [2020]158). Thanks are also extended to
the 2019 Post-doctoral Junior Leader-Retaining Fellowship, la Caixa Foundation
(ID100010434 and fellowship code LCF/BQ/PR19/11700011), the “Generalitat
Valenciana” (SEJI/2020/034) and the “Ramón y Cajal” program (J. F.-S.). E. P.
acknowledges the financial support from the European Research Council under the
European Union’s Horizon 2020 research and innovation programme/ ERC Grant
Agreement No. 814804, MOF-reactors.
Notes and references
1. (a) E. Moreno-Pineda, C. Godfrin, F. Balestro, W. Wernsdorfer and M. Ruben,
Chem. Soc. Rev., 2018, 47, 501-513; (b) M. Shiddiq, D. Komijani, Y. Duan, A.
Gaita-Ariño, E. Coronado and S. Hill, Nature, 2016, 531, 348-351; (c) J. Ferrando-
Soria, J. Vallejo, M. Castellano, J. Martínez-Lillo, E. Pardo, J. Cano, I. Castro, F.
Lloret, R. Ruiz-García and M. Julve, Coord. Chem. Rev. 2017, 339, 17–103.
2. (a) X.-Y. Liu, L. Sun, H.-L. Zhou, P.-P. Cen, X.-Y. Jin, G. Xie, S.-P. Chen and Q.-
L. Hu, Inorg. Chem., 2015, 54, 8884-8886; (b) X.-Y. Liu, X.-F. Ma, W.-Z. Yuan,
P.-P. Cen, Y.-Q. Zhang, J. FerrandoSoria, S.-P. Chen and E. Pardo, Inorg. Chem.,
2018, 57, 14843-14851.
3. (a) J. M. Zadrozny, D. J. Xiao, M. Atanasov, G. J. Long, F. Grandjean, F. Neese
and J. R. Long, Nat. Chem., 2013, 5, 577-581; (b) Y.-W. Wu, D.-N. Tian, J.
Ferrando-Soria, J. Cano, L. Yin, Z.-W. Ouyang, Z.-X. Wang, S.-C. Luo, X.-Y. Liu
and E. Pardo, Inorg. Chem. Front., 2019, 6, 848-856; (c) J. M. Zadrozny, M.
Atanasov, A. M. Bryan, C.-Y. Lin, B. D. Rekken, P. P. Power, F. Neese and J. R.
Long, Chem. Sci., 2013, 4, 125-138.
4. S. Vaidya, S. Tewary, S. K. Singh, S. K. Langley, K. S. Murray, Y. Lan, W.
Wernsdorfer, G. Rajaraman and M. Shanmugam, Inorg. Chem., 2016, 55, 9564-
9578.
5. S. Gomez-Coca, A. Urtizberea, E. Cremades, P. J. Alonso, A. Camon, E. Ruiz and
F. Luis, Nat. Commun., 2014, 5, 4300-4307.
6. Z.-W. Chen, L. Yin, X.-N. Mi, S.-N. Wang, F. Cao, Z.-X. Wang, Y.-W. Li, J. Lu
and J.-M. Dou, Inorg. Chem. Front., 2018, 5, 2314-2320.
7. (a) K. Bernot, J. Luzon, L. Bogani, M. Etienne, E. C. Sangregorio, M. Shanmugam,
A. Caneschi, R. Sessoli and D. Gatteschi, J. Am. Chem. Soc., 2009, 131, 5573-5579;
(b) Y. Ma, G.-F. Xu, X. Yang, L.-C. Li, J.-K. Tang, S.-P. Yan, P. Cheng and D.-Z.
Liao, Chem. Commun., 2009, 46, 8264-8266.
8. T. T. da Cunha, J. Jung, M. E. Boulon, G. Campo, F. Pointillart, C. L. M. Pereira,
B. L. Guennic, O. Cador, K. Bernot, F. Pineider, S. Golhen and L. Ouahab, J. Am.
Chem. Soc., 2013, 135, 16332-16335.
9. M. Ji, M.-Y. Liu and S.-L. Gao, Instrum. Sci. Technol., 2001, 29, 53-57.
10. (a) S.-L. Wang, L. Li, Z.-W. Ouyang, Z.-C. Xia, N.-M. Xia, T. Peng and K.-B.
Zhang, Acta Phys. Sin., 2012, 61, 107601; (b) H. Nojiri and Z.-W. Ouyang,
Terahertz Sci. Technol., 2012, 5, 1.
11. G. M. Sheldrick, SADABS, Program for Empirical Absorption Correction,
University of Göttingen, Göttingen, Germany, 1996.
12. G. M. Sheldrick, SHELXS-2014 and SHELXL-2014, Program for Crystal Structure
Determination, University of Göttingen, Göttingen, Germany, 2014.
13. (a) F. Neese, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2012, 2, 73-78; (b) F.
Neese, Wiley Interdiscip. Rev.: Comput. Mol.Sci., 2018, 8, e1327; (c) F. Neese,
ORCA-an Ab Initio, Density Functional and Semiempirical Program Package, 4.2.1,
University of Bonn, Bonn, Germany, 2019.
14. (a) A. D. Becke, Phys. Rev. A, 1988, 38, 3098-3100; (b) C. Lee, W. Yang and R. G.
Parr, Phys. Rev. B, 1988, 37, 785-789; (c) P. J. Stephens, F. J. Devlin, C. F.
Chabalowski and M. J. Frisch, J. Phys. Chem., 1994, 98, 11623-11627.
15. F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005, 7, 3297-3305.
16. P. A. Malmqvist and B. O. Roos, Chem. Phys. Lett., 1989,155, 189-194.
17. (a) C. Angeli, R. Cimiraglia, S. Evangelisti, T. Leininger and J. P. Malrieu, J. Chem.
Phys., 2001, 114, 10252-10264; (b) C. Angeli, R. Cimiraglia and J. P. Malrieu,
Chem. Phys. Lett., 2001, 350, 297-305; (c) C. Angeli, R. Cimiraglia and J. P.
Malrieu, J. Chem. Phys., 2002, 117, 9138-9153; (d) C. Angeli, S. Borini, M. Cestari
and R. Cimiraglia, J. Chem. Phys., 2004, 121, 4043-4049; (e) R. Herchel, I. Nemec,
M. Machata and Z. Trávníček, Dalton Trans., 2016, 45, 18622-18634.
18. D. Ganyushin and F. Neese, J. Chem. Phys., 2006, 125, 024103.
19. F. Neese, J. Chem. Phys., 2005, 122, 034107.
20. (a) R. Maurice, R. Bastardis, C. Graaf, N. Suaud, T. Mallah and N. Guihéry, J. Chem.
Theory Comput., 2009, 5, 2977-2984; (b) I. Nemec, R. Herchel, M. Machata and Z.
TrávnÍček, New J.
Chem., 2017, 41, 11258-11267.
21. F. Weigend, Phys. Chem. Chem. Phys., 2006, 8, 1057-1065.
22. (a) F. Neese, F. Wennmohs, A. Hansen and U. Becker, Chem. Phys., 2009, 356, 98-
109; (b) R. Izsak and F. Neese, J. Chem. Phys., 2011, 135, 144105.
23. (a) J. Wu, S.-P. Chen and S.-L. Gao, Mater. Chem. Phys., 2010, 122, 301-304; (b)
Y. Mi, Z.-Y. Huang, J.-Y. Jiang and Y.-F. Li, Mater. Lett., 2011, 65,1768-1771;(c)
L.-D. Wang, Z. Ma, S.-G. Liu and Z.-Y. Huang, J. Therm. Anal. Calorim., 2014,
115, 201-208; (d) J. Chen, Y. J. Ma, G.-C. Fan, Y.-F. Li, J.-Y. Jiang and Z.-Y. Huang.
Mater. Lett., 2011, 65, 1768-1771.
24. (a) M.-H. Zeng, S.-H. Q. Chen, G. Xie, Q. Shuai, S.-L. Gao and L.-Y. Tang, Inorg.
Chem., 2009, 48, 7070-7079; (b) S. Fischer, G. Krahn and B. Reimer, Thermochim.
Acta, 2006, 445, 160-167.
25. J. F. Berry, F. A. Cotton, C.-Y. Liu, T. Lu, C. A. Murillo, B. S. Tsukerblat, D.
Villagran and X. Wang, J. Am. Chem. Soc., 2005, 127, 4895-4902.
26. M. Llunell, D. Casanova, J. Cirera, P. Alemany and S. Alvarez, SHAPE, v2.1, 2013.
27. R. Boča, Coord. Chem. Rev., 2004, 248, 757-815.
28. (a) A. V. Palii, B. S. Tsukerblat, E. Coronado, J. M. Clemente-Juan, J. J. Borras-
Almenar, Inorg. Chem., 2003, 42, 2455-2458; (b) A. V. Palii, B. S. Tsukerblat, E.
Coronado, J. M. Clemente-Juan, J. J. Borras-Almenar, J. Chem. Phys., 2003, 118,
5566-5581.
29. R. L. Carlin, Magnetochemistry, Springer-Verlag, Berlin, 1986.
30. N. F. Chilton, R. P. Anderson, L. D. Turner, A. Soncini and K. S. Murray, J. Comput.
Chem., 2013, 34, 1164-1175.
31. Y.-J. Zhang, L. Yin, J. Li, Z.-B. Hu, Z.-W. Ouyang, Y. Song and Z. Wang, RSC
Adv., 2020, 10, 12833-12840.
32. D. V. Korchagin, A. V. Palii, E. A. Yureva, A. V. Akimov, E. Y. Misochko, G. V.
Shilov, A. D. Talantsev, R. B. Morgunov, A. A. Shakin, S. M. Aldoshin and B. S.
Tsukerblat, Dalton Trans., 2017, 46, 7540-7548.
33. R. Herchel, L. Váhovská, I. Potočňák and Z. Trávníček, Inorg. Chem., 2014, 53,
5896-5898.
34. R. Orbach, Proc. R. Soc. London, Ser. A, 1961, 264, 458-484.
35. (a) K. S. Cole and R. H. Cole, J. Chem. Phys., 1941, 9, 341-351. (b) K. Suzuki, R.
Sato and N. Mizuno, Chem. Sci., 2013, 4, 596-600.
36. (a) R. Herchel, I. Nemec, M. Machata and Z. Travnicek, Inorg. Chem., 2015, 54,
8625-8638; (b) J.-D. Leng, S.-K. Xing, R. Herchel, J.-L. Liu and M.-L. Tong, Inorg.
Chem., 2014, 53, 5458-5466; (c) M. Machata, I. Nemec, R. Herchel and Z.
Travnicek, RSC Adv., 2017, 7, 25821-25827.
37. A. Ghisolfi, K. Y. Monakhov, R. Pattacini, P. Braunstein, X. López, C. Graaf, M.
Speldrich, J. Leusen, H. Schilder and P. Kögerler, Dalton Trans., 2014, 43, 7847-
7859.