PHYSICAL REVIEW B 108, 014412 (2023)

Cluster spin-glass state with magnetocaloric effect in open framework structure

of the Prussian blue analog molecular magnet K2x/3 Cu[Fe(CN)6 ]2/3 nH2 O

Pramod Bhatt ,1,2,* Nilasha Maiti ,1,2 M. D. Mukadam ,1 Sher Singh Meena ,1 A. Kumar,1,2 and S. M. Yusuf1,2,†
1
Solid State Physics Division, Bhabha Atomic Research Centre, Mumbai 400 085, India
2
Homi Bhabha National Institute, Anushaktinagar, Mumbai 400 094, India

(Received 19 January 2023; revised 4 April 2023; accepted 22 June 2023; published 11 July 2023)

We report on the existence of cluster spin-glass behavior and magnetocaloric properties of potassium-ion-
intercalated copper hexacyanoferrate open framework material K2x/3 Cu[Fe(CN)6 ]2/3 · nH2 O (x = 0 and 0.3)
synthesized using a coprecipitate method. The structural properties are investigated using room-temperature
powder x-ray and neutron diffraction measurements. The results derived from the fitted x-ray and neutron diffrac-
tion data using the Rietveld refinement method reveal that both compounds are in pure single-nanocrystalline
phase with face-centered cubic crystal structure of space group F m-3m. The room-temperature Mössbauer
spectroscopy reveals the presence of a low-spin (LS) state of FeIII for compound x = 0, whereas for compound
x = 0.3, LS states of FeII and FeIII are evident. The chemical and Mössbauer studies reveal stoichiometries
of compounds as CuII [FeIII (CN)6 ]2/3 · 3.75 H2 O (x = 0) compound (1) and K0.2 CuII [FeII0.3 FeIII
0.7 (CN)6 ]2/3 · H2 O
(x = 0.3). The dc magnetization measurement shows a magnetic phase transition from a paramagnetic to a
ferromagneticlike state below the magnetic ordering temperatures of ∼18.6 and ∼16.3 K for the compounds (1)
and (2), respectively. The positive value of Curie constants obtained from the fitting of the Curie-Weiss law in the
paramagnetic region confirms the presence of ferromagnetic interaction in both compounds. The spin-only effec-
tive paramagnetic moments are found to be ∼2.3 and ∼1.9 μB /f.u. for compounds (1), and (2), respectively. The
analysis of ac susceptibility measurements as a function of frequency shows that the spin dynamics follows the
Vogel-Fulcher and the power law. The best fitted values of τo = 4.037 × 10−7 s and 3.5 × 10−6 s for compounds
(1) and (2), respectively, confirm a cluster spin-glass behavior of both compounds. A neutron depolarization
study further validates the cluster spin-glass nature of both compounds. The magnetocaloric effect for both the
compounds is further investigated and found to be higher for compound (1) than compound (2). The maximum
magnetic entropy change—SM under a magnetic field of 50 kOe is found to be ∼6.61 and 2.95 J kg−1 K−1 for
compounds (1) and (2), respectively. The relative cooling power is found to be 78.3 J kg−1 for compound (1) and
40 J kg−1 for compound (2).

DOI: 10.1103/PhysRevB.108.014412

I. INTRODUCTION Therefore, many materials have been studied in recent years

for the observation of a significant MCE over a broad tem-
Research and development activities on materials exhibit-
perature range and their suitability in magnetic refrigeration
ing the magnetocaloric effect (MCE) have generated great
devices. The MCE has mostly been investigated for the in-
interest recently because of their possible technological appli-
termetallic and alloys [3–5]. However, recently, magnetic
cation in magnetic refrigeration, i.e., alternative technology
materials such as metal-organic frameworks [6–9], high-
for refrigeration [1,2]. Magnetic refrigeration technology is
spin molecular clusters [10–12], organic magnetic molecules
preferable over the conventional gas compression technolo-
[13,14], and Prussian blue analogs (PBAs) [15–17] also have
gies as the former is 20–50% more efficient. The temperature
drawn attention as low-temperature magnetic caloric materi-
change of magnetic materials during the application of a mag-
als. In this regard, the PBA molecular magnets have shown
netic field under adiabatic conditions is the measure of MCE.
promising results as multifunctional materials [18,19] to ex-
It is generally associated with a large value of magnetization
hibit a wide range of magnetic ordering temperatures starting
change and depends upon the magnetic entropy change (SM )
from low to high temperature, hence useful for studying MCE.
or adiabatic temperature change (Tad ) and the refrigerant
The general formula for a PBA-type compound is
capacity (RC ) of the materials. Such materials with the capa-
AX [Y (CN)6 ] · zH2 O, where A, X, and Y are alkali and tran-
bility to show magnetic refrigeration (SM and RC ) are in high
sition metal ions. Here, z is the number of water molecules
demand and offer environmentally friendly efficient energy
present in the compound. The transition metal ions (X and
as alternatives to regular gas-based refrigeration technologies.
Y), coordinated to the C ≡ N ligand are in the low-spin (LS)
and high-spin states, respectively. The water molecules re-
side at the interstitial space of the crystal structure. PBAs
*
prabhatt@barc.gov.in have mostly cubic structure, and their magnetic and magne-
†
smyusuf@barc.gov.in tocaloric properties can, therefore, be tailored by varying the

2469-9950/2023/108(1)/014412(11) 014412-1 ©2023 American Physical Society

PRAMOD BHATT et al. PHYSICAL REVIEW B 108, 014412 (2023)

transition metal ions involved in the cyanobridging [20,21]. A field emission scanning electron microscope (make
Moreover, their properties are associated with the synthesis Carl Zeiss Auriga, with Gemini) with energy-dispersive
parameters and methods which lead to the porous struc- x-ray analysis is used for surface morphology and
ture. The Y (CN)6 vacancies also exist in the compounds, chemical analysis. The stoichiometries of the com-
but they depend on the stoichiometry of the compound. pounds are found to be Cu[Fe(CN)6 ]2/3 · 3.75 H2 O
0.7 (CN)6 ]2/3 · H2 O for x = 0 and
However, control over the stoichiometries and vacancies is and K0.2 CuII [FeII0.3 FeIII
relatively difficult to optimize during the synthesis. There- 0.3, respectively. The compounds x = 0 and 0.3 are
fore, the MCE in PBAs is not investigated extensively or denoted further as compounds (1) and (2), respectively.
reported in the literature. The MCE in PBAs has only been Room-temperature x-ray diffraction (XRD) patterns of both
reported for a few compounds, particularly the nickel- and compounds are recorded using a Rigaku diffractometer over
chromium-based PBAs. For example, a thin film of the PBA an angular (2θ ) range of 5◦ −70◦ with Cu Kα radiation.
of nickel hexa-cyanochromate deposited by the layer-by-layer The analysis of XRD data is carried out using FULLPROF
method has been studied for the MCE [16]. The maximum software based on the Rietveld refinement method [26].
entropy changes of 7.43 and 7.07 J kg−1 K−1 at 5 T have The room-temperature neutron diffraction patterns for both
been reported for parallel and perpendicular configurations, compounds are recorded (wavelength λ = 1.094 Å) using a
respectively. For CsNi[Cr(CN)6 ] · H2 O, the highest entropy neutron powder diffractometer, at Dhruva reactor, Trombay,
change is recorded near TC and found to be −SM = (5.9 ± India. The neutron depolarization experiments are carried
0.7) J kg−1 K−1 [22]. The vacancy-dependent MCE has also out using the polarized neutron spectrometer [27] at Dhruva
been studied in nickel-based PBA Csx Ni4 [Cr(CN)6 ](8+x)/3 · reactor, Trombay, India.
H2 O. The vacancies have been controlled during the synthesis, Room-temperature Mössbauer spectroscopy in transmis-
and −SM is found to be 6.6 and 6.9 J kg−1 K−1 for CsNiCr sion geometry with the constant acceleration mode is used
and NiCr 2/3 , respectively. Similarly, the MCE has been inves- for determination of the ionic state and local surroundings
tigated in ternary PBA Ni1.5 [Fe(CN)6 ]x [Cr(CN)6 ]1−x · nH2 O of Fe ions in both compounds. The Co–57 source in Rh
(x = 0, 0.25, 0.5, 0.75, and 1). The maximum value of −S matrix of strength 50 mCi is used for this purpose. Fur-
(∼8.04 J kg−1 K−1 ) has been found for the compound with ther, the magnetic properties are investigated in detail using
x = 1 under magnetic field of 50 kOe [15]. The value of the temperature-dependent dc magnetization measurements in
MCE for the magnetic materials is maximum at the magnetic field-cooled (FC) conditions down to 5 K using a Cryogenic
ordering temperature. The entropy change is also dependent Ltd., UK, make commercial magnetic properties measurement
on magnetic field change as S ∝ H n , where n is the critical system. The magnetization data vs field at 5 K are measured
exponent. The present compound, copper hexacyanoferrate over an applied magnetic field of ±50 kOe for both com-
in the family of PBAs, is useful for its applications in the pounds. The ac susceptibility study as a function of frequency
fields of batteries, energy storage, ion removal, and biosensors in the low-temperature region is also carried out to investigate
[23–25]. The thermoelectric properties of copper hexacyano- the spin dynamics of both magnetic compounds. For the mag-
ferrate have also been investigated. However, despite many netocaloric study, the FC magnetization measurements over
technological applications of copper hexacyanoferrate, its the magnetic ordering temperature are measured at various
magnetic and magnetocaloric properties have not yet been external magnetic fields for both compounds.
investigated in detail. Therefore, in this paper, we report
on the magnetic and magnetocaloric properties of K+ -ion-
intercalated Cu-Fe-based PBA molecular magnets and reveal III. RESULT AND DISCUSSION
cluster spin-glass behavior and high entropy change in such
open framework materials. A. X-ray and neutron diffraction study
The Rietveld refined room-temperature x-ray and neutron
powder diffraction patterns of both compounds are shown
in Fig. 1. The refinement shows that both compounds are
II. EXPERIMENTAL DETAILS in a pure single-crystalline phase with a cubic structure of
the space group F m-3m. The important structural parameters,
The room-temperature coprecipitation method is used for derived from the combined Rietveld analysis of x-ray and
the synthesis of K2x/3 Cu[Fe(CN)6 ]2/3 · nH2 O (x = 0 and 0.3) neutron diffraction data, are shown in Table I.
stoichiometric compounds of PBAs. The chemicals, received The lattice parameters are found to be ∼10.128 ± 0.001
from Sigma Aldrich, are of reagent grade and used with- and 9.997 ± 0.004 Å for compounds (1) and (2), respectively.
out further processing. The chemicals CuCl2 · 2H2 O and The observed slight decrease in lattice constant for compound
K3 [Fe(CN)6 ] in a 0.15 and 0.1 M molar ratio, respectively, (2) as compared with compound (1) could be due to smaller
dissolved in aqueous solution and then mixed together to form bond lengths of Cu-N and Fe-C and the presence of more elec-
a compound Cu[Fe(CN)6 ]2/3 · nH2 O (x = 0), whereas for the trons, which are introduced by the reduction of [FeIII (CN)6 ]
x = 0.3 compound, the Fe3+ ions are reduced with 0.1 M to [FeII (CN)6 ]. The schematic crystal structures of both com-
K2 S2 O3 , resulting in the incorporation of K+ ions into the pounds, as obtained from the Rietveld refinement, are shown
structure, which leads to the formation of the mixed-valence in Fig. 2. It is observed that both compounds are crystallized
0.5± (CN)6 ]2/3 · mH2 O compound. The
K0.2 CuII [FeII0.5± FeIII into a cubic crystal structure in which transition metal ions
precipitated compounds are washed many times with distilled CuII and FeIII occupy the 4a (0, 0, 0) and 4b ( 21 , 21 , 21 ) crystal-
water and dried up at room temperature. lographic sites, respectively. The alkali metal ions K+ occupy

014412-2
CLUSTER SPIN-GLASS STATE WITH MAGNETOCALORIC … PHYSICAL REVIEW B 108, 014412 (2023)

FIG. 1. The Rietveld refined room-temperature (a) x-ray diffraction and (b) neutron diffraction patterns of compounds (1) and (2). The
experimentally observed and fitted patterns are shown by the open circles and solid lines, respectively. The difference between experimental
and fitted patterns is shown by the solid lines at the bottom. The positions of allowed Bragg peaks are shown by vertical lines. The (hkl) values
of the Bragg peaks are also marked for both compounds.

some of the interstitial sites of uncoordinated water molecules Bragg peak, the average crystallite sizes are estimated to be
in compound (2), as shown in Fig. 2. ∼350 ± 15 and ∼220 ± 10 Å for compounds (1) and (2), re-
It has been observed that totals of ∼3.75 and ∼1 water spectively. The refinement results, obtained from the fitting of
molecules are present in compounds (1) and (2), respectively. neutron diffraction data, are consistent with the XRD results.
The water molecules, which occupy the interstitial sites of 8c The background in the neutron diffraction patterns of both
and 32 f in the crystal structure, are known as uncoordinated compounds is due to the hydrogen of the compounds. The
water, whereas the water molecules which are connected to N PBAs exhibit structural disorder which arises due to vacan-
atoms (situated at 24e sites) are known as coordinated water. cies and clustering of uncoordinated water molecules in the
The crystallite size in the compound is derived using compounds evident from the neutron diffraction results [28].
Kλ
τ= ,
t cos θ B. Mössbauer spectroscopy study
where τ is the mean size of the crystalline domains, K is Mössbauer spectroscopy is carried out to find out the spin
a dimensionless shape factor with a value close to unity states and local atomic environment of Fe atoms in both
(0.9), λ is the x-ray wavelength, and t is the line broaden- compounds. Figure 3 shows fitted Mössbauer spectra at room
ing at half the maximum intensity. Using the most intense temperature, recorded at a velocity of 4.0 mm/s, of both

TABLE I. Rietveld refined structural parameters for compounds (1) and (2). The fractional coordinates are denoted as x  , y , and z .

Atom Wyckoff site x y z Occupancy

Compound (1) Cu 4a 0 0 0 1
F m-3m Fe 4b 0.5 0.5 0.5 0.67
a = 10.128 ± 0.001 C 24e 0.322(3) 0 0 0.67
b = 10.128 ± 0.001 N 24e 0.197(7) 0 0 0.67
c = 10.128 ± 0.001 O1 24e 0.237(2) 0 0 3.09(2)
O2 32 f 0.309(5) 0.309(5) 0.309(5) 0.51(4)
O3 8c 0.25 0.25 0.25 0.42(4)
Compound (2)
F m-3m Cu 4a 0 0 0 1
a = 9.997 ± 0.004 Fe 4b 0.5 0.5 0.5 0.67
b = 9.997 ± 0.004 C 24e 0.306(3) 0 0 0.67
c = 9.997 ± 0.004 N 24e 0.189(5) 0 0 0.67
O1 24e 0.207(2) 0 0 1.11(5)
O2 32 f 0.333(7) 0.333(7) 0.333(7) 0.06(4)
K 8c 0.25 0.25 0.25 0.19(7)

014412-3
PRAMOD BHATT et al. PHYSICAL REVIEW B 108, 014412 (2023)

FIG. 2. Schematic crystal structure of compounds (1) and (2).

The positions of Fe(CN)6 vacancies and coordinated and uncoordi- FIG. 3. Room-temperature Mössbauer spectra of compounds (1)
nated water molecules are shown in the cubic crystal structure. The and (2). The recorded experimental and least-squares fitted data are
crystallographic positions of some of the K+ ions in compound (2) shown by the open circles and thick solid lines, respectively.
are marked at interstitial sites.

nected to Cu, which is ∼64%, and (ii) the Fe atoms connected

compounds. The absence of the magnetic six-line hyperfine to N atoms attached with coordinated molecules water, which
pattern confirms that both compounds are in paramagnetic is ∼36%. The Mössbauer spectra of compound (2) are fitted
ground state at room temperature. The fitted Mössbauer pa- with two symmetric doublets. The δ values of both doublets
rameters, such as isomer shift (δ), quadrupole splitting (EQ ), are found to be −0.098 mm/s (doublet A) and 0.156 mm/s
and line width (), for both compounds are listed in Table II. (doublet B), respectively. The observed δ values indicate that
The fitting of the data for compound (1) indicates the presence doublet A is due to LS Fe+2 ions, whereas doublet B belongs
of two paramagnetic doublets (A and B) with isomer shifts (δ) to LS Fe+3 ions. The areas under these two doublets A and
of ∼−0.026 and −0.188 for doublets A and B, respectively. B (Table II) are found to be ∼29.9% (LS Fe+2 ) and ∼70.1%
The values of isomer shift (δ) and quadrupole splitting (EQ ) (LS Fe+3 ), respectively. The reduction of Fe+3 may lead to
suggest that Fe ions are in +3 LS state in compound (1). Fur- the formation of some of the Fe+2 spin state in compound (2).
ther, these two doublets suggest two different environments of The higher value of EQ for doublet B than doublet A in com-
Fe in compound (1). The different δ values of these doublets pound (1) indicates the presence of a relatively larger crystal
are due to two different environments of Fe, which is evident field gradient at the Fe site, connected with Cu atoms. The
from the schematic diagram (Fig. 2) of the crystal structure. higher value of EQ for doublet B confirms more distortion
The Fe has two different environments: (i) the Fe atoms con- at these Fe sites, whereas for compound (2), the EQ value for

TABLE II. Hyperfine parameters for compound (1) and compound (2).

Fe-site and Isomer shift (δ), Quadrupole splitting Line width (), Relative area Goodness of
Sample ionic state (mm/s) (EQ ), (mm/s) (mm/s) (%) fit (χ 2 )
Compound (1) Doublet A −0.026 ± 0.008 0.443 ± 0.005 0.332 ± 0.005 35.9 ± 0.5 0.87
(Fe3+ LS)
Doublet B −0.188 ± 0.002 0.475 ± 0.004 0.278 ± 0.003 64.1 ± 0.5
(Fe3+ LS)
Compound (2) Doublet A −0.098 ± 0.005 0.200 ± 0.005 0.322 ± 0.04 29.9 ± 0.5 0.92
(Fe2+ LS)
Doublet B −0.156 ± 0.005 0.518 ± 0.006 0.371 ± 0.004 70.1 ± 0.5
(Fe3+ LS)

014412-4
CLUSTER SPIN-GLASS STATE WITH MAGNETOCALORIC … PHYSICAL REVIEW B 108, 014412 (2023)

doublet B is higher (∼0.518 mm/s) than doublet A, suggesting

that the distortion increases at LS Fe+3 sites due to inclusion
of K in the structure. The Mössbauer results confirm that Fe is
in Fe+3 LS state with two different environments arising due
to Fe connected to C ≡ N without and with water molecules,
whereas in compound (2), both LS states of Fe+2 and Fe+3
exist with relative areas of 30 and 70%, respectively, due to
inclusion of K ions in the crystal structure. The identification
of spin states and their doublet for Fe ions is analyzed based
on the δ values. The δ value for LS Fe+2 is in the range of
−0.2 to 0.5, whereas it is −0.5 to 0.4 for LS Fe+3 [29,30].
Based on Mössbauer spectroscopy, the stoichiometries of the
compounds are found to be Cu[Fe(CN)6 ]2/3 · 3.75 H2 O and
0.7 (CN)6 ]2/3 · H2 O for x = 0 [compound (1)]
K0.2 CuII [FeII0.3 FeIII
and 0.3 [compound (2)], respectively.
Both compounds have a three-dimensional (3D) network of
Cu−N ≡ C−Fe−C ≡ N−Cu, in which Cu and Fe atoms are
octahedrally surrounded by −N ≡ C− and −C ≡ N− units,
which could show Jahn-Teller distortion due to the presence
of Cu+2 (d 9 ) ions and consequently affect the magnetic prop-
erties of these compounds. The distortion and its influence
on the structural properties of PBAs have been reported in
the literature. For example, a cubic-to-tetragonal structural
phase transition in rubidium manganese hexacyanoferrate [31]
and a reduction of crystal symmetry to tetragonal caused
by the strong cooperative Jahn-Teller effect of Cu2+ ions in
CsCu[Cr(CN)6 ] · 2.1 H2 O have been reported [32]. For the
present compounds, with nearly an ideal octahedral geometry,
lowering of the crystal symmetry has not been observed as a
function of temperature. However, a 3D network of octahe-
dral coordination units of CuN4 (H2 O)2 is expected to induce
Jahn-Teller distortion in these compounds. The Jahn-Teller FIG. 4. (a) and (b) show the field-cooled magnetization vs tem-
distortion can lead to the change in the strength of magnetic perature curves for compounds (1) and (2), respectively, under an
exchange coupling constants. In fact, the quadrupole split- external applied magnetic field of 200 Oe. The inset shows the
ting (Table II), obtained from the Mössbauer study at room magnetic ordering transition in the magnetified range of temperature
temperature, is clear evidence of the presence of Jahn-Teller 5–50 K.
distortion in these compounds. The Jahn-Teller distortion is
expected to be slightly higher in compound (1) than com-
region of the susceptibility is fitted using the Curie-Weiss law:
pound (2) due to presence of more CuN4 (H2 O)2 octahedral
units in compound (1). C
χ= ,
T − θp
C. DC magnetization study where θp is the paramagnetic Curie temperature, and C is
Figure 4 shows the FC magnetization (M) vs temper- the Curie constant. The positive values of θp (18 and 16 K)
ature (T) curves under an external applied magnetic field for compounds (1) and (2), respectively, confirm the pres-
of 200 Oe for both compounds. The compounds show a ence of the ferromagnetic interaction in both compounds.
paramagnetic-to-ferromagneticlike transition with decreasing The spin-only effective paramagnetic moments μeff are also
temperature. Magnetic transitions at ∼18.6 and ∼16.3 K have calculated using (3CkB /NA )1/2 μB ∼ (8C)1/2 μB , where NA is
been observed from the magnetization data and could be fer- Avogadro’s number and kB is the Boltzmann constant, which
romagnetic or ferrimagnetic in nature. The magnetic ordering is found to be ∼2.3 and ∼1.9 μB /f.u. The theoretically ex-
temperature has been estimated from the first derivative of M pected (spin-only) values of μeff are calculated by using the
vs T data. formula (μeff )2 = [g2 {nS(S + 1)}]μ2B , where g is gyromag-
The magnetic susceptibility (χ ) is measured as a function netic ratio (∼2), n is the number of magnetic ions with spin
of temperature. Figure 5 shows χ T vs T curves at 200 Oe for S in formula units, and summation runs over all magnetic
both compounds. The χ T vs T curves rise continuously with ions in formula units. The theoretically calculated values of
the lowering of temperature until the transition temperature is spin-only μeff are found to be 2.45 μB /f.u. for compound
reached for both compounds, which is a characteristic behav- (1), assuming Fe+3 in their LS states (S = 21 ), whereas Cu is
ior of ferromagnetic materials. in high-spin state (S = 21 ). However, for compound (2), ∼70%
The inverse of the susceptibility (χ −1 ) as a function of Fe+3 ions are in LS state (S = 21 ), while ∼30% are in Fe+2
temperature is shown in the inset of Fig. 5. The paramagnetic LS state (S = 0), as obtained from the Mössbauer study. The

014412-5
PRAMOD BHATT et al. PHYSICAL REVIEW B 108, 014412 (2023)

FIG. 6. Magnetization (M) vs field (H) curves at 2 K for both

compounds up to 50 kOe. The inset shows the corresponding M vs H
curve in Bohr magnetron (μB ) per f.u.

D. AC susceptibility study
Figure 7 shows ac magnetic susceptibility data for both
compounds. The ac magnetic field amplitude of 2 Oe is ap-
plied in the temperature range of 5–25 K. Both the in-phase
(χ  ) and out-of-phase (χ  ) susceptibilities exhibit peaks at
around TP ∼ 16.8 K. The in-phase susceptibility (χ  ) changes
with respect to frequency. The peak maxima of the in-
phase (χ  ) susceptibility are frequency dependent for both

FIG. 5. (a) and (b) show χ T vs temperature (T) curves for

compounds (1) and (2), respectively. The inverse susceptibility vs
temperature curves, fitted by using the Curie-Weiss law, are shown
in the inset.

theoretically calculated spin-only value of μeff for compound

(2) is found to be 2.24 μB /f.u. The experimentally observed
values of μeff are consistent with the theoretically calculated
spin-only values of both compounds.
Figure 6 shows the magnetization (M) as a function of the
applied magnetic field (H) for both compounds at ±5 Tesla.
The coercivity values of ∼2 and 3 kOe have been found for
compounds (1) and (2), respectively, whereas the maximum
saturation magnetization values of ∼1.5 and 0.92 μB /f.u.
have been observed for compounds (1) and (2), respectively.
The shape of the M-H loops indicates a ferromagnetic nature
of the compound. The theoretically expected (spin-only) value
of the ordered magnetic moment μs per formula unit is calcu-
lated by using the formula μs = (gnS)μB , where g is the
gyromagnetic ratio (∼2), n is the number of magnetic ions,
and S is the spin of the magnetic ions. The μs per formula
unit (spin-only) values are found to be 1.67 and 1.46 μB /f.u.
for compounds (1) and (2), respectively. The Mössbauer study
reveals that ∼30% of Fe is in LS Fe+2 state, whereas ∼70% is
in the Fe+3 LS state. The difference in the experimentally ob-
served and theoretically calculated ordered moments is due to FIG. 7. The frequency dependence in-phase (χ  ) magnetic ac
the nonsaturation of the hysteresis curves. The magnetization susceptibility for both compounds as a function of temperature. The
study confirms a ferromagnetic nature of both compounds. inset shows the magnified region of the ac susceptibility data.

014412-6
CLUSTER SPIN-GLASS STATE WITH MAGNETOCALORIC … PHYSICAL REVIEW B 108, 014412 (2023)

compounds. The characteristic peak temperatures (TP ) are

found to be ∼17.3 and ∼16.4 K at low frequency of
47 Hz for compounds (1) and (2), respectively. The peak
amplitude decreases, and peak maxima systematically shift
to higher temperatures and are found to be at ∼17.8 and
∼16.8 K for higher frequency of 3987 Hz for compounds (1)
and (2), respectively. This is expected for a compound with
ferromagnetic or ferrimagnetic ordering having long-range
magnetic correlations. The frequency-dependent ac suscep-
tibility reveals the spin behavior and nature of the magnetic
ordering in the compounds. It is known that the anomaly in ac
susceptibility depends on the excitation frequency. When the
position of peak maxima shifts to a higher temperature with
the increasing frequency, it indicates the characteristic feature
of superparamagnets and spin-glass-type ordering [33,34].
The ac susceptibility data are further analyzed in terms
of relaxation time τ , experimentally measured characteristic
time which describes the dynamical fluctuation over obser-
vation time t. Since the characteristic relaxation time at the
in-phase susceptibility maximum of the system corresponds to
the observation time, we can write t = 1/ f . The relaxation of
spin-glass and superparamagnetic clusters follows the Néel-
Arrhenius law. The relaxation time τ is calculated from the
maximum of the susceptibility curve using the Arrhenius law:
 
Ea
τ = τ0 exp , (1)
kB T
where τ is the relaxation time, τ0 is the pre-exponential factor,
and Ea is the energy barrier. Figure 8 shows ln( f ) vs 1/TP
fitting of experimental data using the Néel-Arrhenius law. The
fitted values of τ0 and Ea /kB are found to be ∼2.1 × 10−171 s FIG. 8. The bottom figures show the ln( f ) vs 1/TP plot for both
and 6841 K for compound (1), which are totally unphysical compounds. The data are fitted by a solid line using least-squares
and unrealistic. However, the frequency dependence of freez- fitting by considering the Néel-Arrhenius equation. The figures in
ing temperature is fitted well to the empirical Vogel-Fulcher the middle show ln( f ) vs 1/(TP − T0 ) plot for both compounds, fitted
law, which considers intercluster interaction: using the Vogal-Fulcher law. The top panel shows a log-log plot of
  the frequency vs (TP − T0 )/T0 , fitted by using the power law.
Ea
τ = τ0 exp , (2)
kB (TP − T0 )
parameters, it can be revealed that both compounds exhibit
where TP is the peak temperature in Fig. 7, and the values of T0 cluster spin-glass behaviors [38].
and Ea are obtained from the fitting. Here, kB is the Boltzmann Though both the Néel-Arrhenius and Vogel-Fulcher laws
constant. The interactions between the superparamagnetic are used to confirm the nature of magnetic ordering in
clusters are governed by the term (TP − T0 ). The fitting of spin-glass, cluster spin-glass, and superparamagnetic systems,
Eq. (2) gives the characteristic frequency (1/τ0 ), the activation there is a further important criterion (as used below), based
energy barrier (Ea ), and the temperature T0 = 17.38(1) for on a power law, to check the magnetic ordering nature of
compound (1). Thus, Ea /kB and T0 can be obtained from the the compound. We have analyzed our frequency-dependent ac
slope and intercept of the ln( f ) vs TP plot, as shown in Fig. 8. susceptibility data using the power law described below:
Figure 8 shows the fitted curve using the Vogel-Fulcher law,
 −zv
and the fitted parameters are found to be Ea /kB = 1.5 K and TP
τ0 = 4.037 × 10−7 s for compound (1). Similarly, the best fit τ (TP ) = τ0 −1 , (3)
Tg
values are found to be Ea /kB = 0.2 K and T0 = 16.6(3) K for
compound (2). The value of relaxation time τ0 is found to where Tg is freezing temperature, τ0 is related to the relaxation
be 3.5 × 10−6 s for compound (2). The larger value of τ0 is of the individual cluster magnetic moment, and zν is a criti-
expected for interacting magnetic spin clusters (τ0 lies in the cal exponent. The best fitted values of τ0 = 2.04 × 10−12 s
range ∼10−6 to 10−10 s), which is mainly due to slow dynam- and 2.54 × 10−9 s are found for compounds (1), and (2),
ics, and matches well with the previously reported values of respectively. The values of critical exponent zν are found to
several other cluster spin-glass-type compounds [35–37]. For be 6.01 and 4.82 for compounds (1) and (2), respectively. The
canonical spin-glass compounds, the value of τ0 (∼10−13 to critical exponent value lies between 4 and 12 for spin-glass
10−14 ) is several orders of magnitude smaller than that of the compounds. The values derived from the power law are dif-
cluster spin glasses. Thus, based on the ac susceptibility fitted ferent from the values obtained using the Vogel-Fulcher law;

014412-7
PRAMOD BHATT et al. PHYSICAL REVIEW B 108, 014412 (2023)

FIG. 9. Neutron depolarization curves at 10 Oe for compounds

(1) and (2).

however, they are acceptable for a cluster spin-glass com-

pound. The larger values of τ0 only reflect nanosized spin
clusters instead of single atomic spins in these compounds.
The analysis of ac susceptibility data confirms that both com-
pounds exhibit cluster spin-glass-like behavior with freezing
temperatures of 17.3 and 16.5 K. The observation of spin-
glass-type behavior is also reported in other PBA compounds
[33,39–43].

E. Neutron depolarization study

Figure 9 shows the neutron depolarization data for
compounds (1) and (2). In a neutron depolarization
study, polarized neutrons, while passing through ferromag-
netic/ferrimagnetic samples, get depolarized owing to the FIG. 10. The temperature-dependent magnetization at different
Larmor precession of neutron spins in the magnetic fields applied magnetic fields for both compounds.
of randomly oriented magnetic domains inside the sample
and provide information about size/magnetization of domains Maxwell relation, i.e., the change in M with respect to T:
[21,44–46]. Both compounds show finite depolarization of    
the polarized neutron beam < ∼20 K, indicating the onset ∂S ∂M
of ferromagnetic/ferrimagnetic-type correlations. Compound = . (4)
∂H T ∂T H
(2) has a slightly lower TC than that of compound (1) and
is consistent with our dc magnetization study. At the lowest The magnetic entropy change (SM ) is calculated using
temperature of 3 K, compound (2) shows slightly larger de-  H 
polarization, and this may be related to the slower relaxation ∂M(H, T )
(SM )H, T = dH. (5)
rate of magnetic spin clusters in compound (2) than com- 0 ∂T H
pound (1). The neutron beam polarization curve for compound
The dc magnetization data as a function of temperature for
(2) shows a continuous decrease in polarization value with
both compounds at various applied magnetic fields up to 50
decreasing temperature. This is in contrast with nearly con-
kOe are shown in Fig. 10. The measurements are carried out in
stant depolarization just below TC for compound (1) probably
zero-FC condition to avoid the memory effect due to cooling
due to relatively slower spin dynamics in compound (2) than
under field. The magnetization curve rises slowly with applied
compound (1). The neutron depolarization data are consistent
fields. With decreasing temperature, an enhancement of mag-
with the ac susceptibility study, confirming cluster spin-glass
netization and a paramagnetic-to-ferromagneticlike transition
behavior.
is observed for both compounds. Moreover, the magnetic or-
dering temperature is found to shift to higher temperature with
increasing applied magnetic field.
F. Magnetocaloric study The value of −SM is calculated using Eq. (5) as a func-
The magnetocaloric study is carried out in terms of the tion of temperature with respect to various applied magnetic
entropy change (−SM ) for both compounds as a function fields for each temperature. The maximum magnetic entropy
of applied magnetic field (H). It can be expressed using the change is expected near the magnetic transition or ordering

014412-8
CLUSTER SPIN-GLASS STATE WITH MAGNETOCALORIC … PHYSICAL REVIEW B 108, 014412 (2023)

FIG. 12. The magnetic entropy change −SM as a function of

(H/TC )2/3 for both compounds.

compounds are comparable with the values for other similar

PBAs presented in Table III.
Here, −SM is also presented as a function of (H/TC )2/3
for both compounds, as shown in Fig. 12. A linear relationship
between −SM and (H/TC )2/3 is evident for both compounds.
However, the lattice entropy contribution is not considered
in linear relations, indicating the lattice contribution is either
negligible or constant. The linear relation between −SM and
(H/TC )2/3 shows the critical behavior of the spin system and
information related to second-order phase transition. Accord-
ing to mean-field theory, it is already established that, near
the vicinity of a second-order phase transition, the value of
FIG. 11. Magnetic entropy change −SM as a function of tem- (H/TC ) is ∼ 23 [48].
perature for the various values of H for compounds (1) and (2). The relative cooling power (RCP) or refrigerant capacity
has also been estimated for both compounds. It is the param-
eter which provides an estimate of the heat transfer during
temperature. Figure 11 shows the −SM vs T curves for the one ideal refrigeration cycle between the cold and hot heat
various values of H for both compounds. It is observed that sinks. The shaded region in Fig. 13(a) represents the RCP for
the value of −SM increases with an increasing applied mag- both compounds. The RCP is estimated by integrating the area
netic field with a peak near the paramagnetic-to-ferromagnetic under the −SM vs T curves as
phase transition temperature. The maximum value of −SM  T hot
is found to be ∼6.61 and ∼2.95 J kg−1 K−1 for compounds
(1) and (2), respectively. The values of −SM found for both RCP = SM (T )dT.
T cold

TABLE III. Comparison between the maximum magnetic entropy change −SM and the RCP for the present compounds and other similar
PBAs reported in the literature.

−SM RCP
Compounds TC (K) H (kOe) (J kg−1 K−1 ) (J kg−1 ) References
Cu[Fe(CN)6 ]2/3 · 3.75 H2 O 18.6 50 6.61 78.3 This paper
0.7 (CN)6 ]2/3 · 1H2 O
K0.2 CuII [FeII0.3 FeIII 16.3 50 2.95 40.0 This paper
Ni3 [Cr(CN)6 ]2 · 15H2 O 60 50 6.49 273.07 [15]
Ni3 [(Cr(CN)6 )1.5 Fe(CN)6 )0.5 ] · 15 H2 O 52 50 5.78 273.33 [15]
Ni3 [(Cr(CN)6 )1 Fe(CN)6 )1 ] · 15 H2 O 41 50 5.47 235.84 [15]
Ni3 [(Cr(CN)6 )0.5 Fe(CN)6 )1.5 ] · 15 H2 O 33 50 6.06 208.62 [15]
Ni3 [(Fe(CN)6 )]2 · 15H2 O 22 50 8.04 145.35 [15]
Csx Ni4 [(Cr(CN)6 )](8+x)/3 (x = 0) 60 70 6.9 [47]
Csx Ni4 [(Cr(CN)6 )](8+x)/3 (x = 4) 90 70 6.6 [47]
CsNi[(Cr(CN)6 )] · H2 O 90 70 6.6 [22]
Cr 3 [(Cr(CN)6 )]2 · 12 H2 O 230 70 0.93 [22]

014412-9
PRAMOD BHATT et al. PHYSICAL REVIEW B 108, 014412 (2023)

compound (1) and x = 0.3 for compound (2)] are synthesized

using a coprecipitate method. The combined x-ray and neu-
tron diffraction study reveals that both compounds are in a
pure single-crystalline cubic phase (space group F m-3m). The
Mössbauer spectroscopy study reveals LS states of FeIII and
FeII for compound (2). The ferromagnetic ordering temper-
atures are found to be ∼18.6 and ∼16.3 K for compounds
(1) and (2), respectively. The cluster spin-glass behavior has
been established for both compounds which are analyzed
using the frequency-dependent ac susceptibility following
FIG. 13. (a) The shaded area shows the relative cooling power the Vogel-Fulcher and power laws. The relaxation times are
(RCP) for both compounds. (b) The RCP as a function of applied found to be ∼4.03 × 10−7 s and 3.5 × 10−6 s for com-
magnetic field. pounds (1) and (2), respectively. The neutron depolarization
study supports the ac susceptibility analysis of having clus-
ter spin-glass behavior of both compounds. Moreover, these
Here, T cold and T hot are the lower and higher temperatures at compounds exhibit enhanced magnetocaloric properties near
half maximum of the −SM peak in the −SM vs T curve the transition temperature with maximum magnetic entropy
and can be considered the temperatures of the cold and hot change (−SM ) values of ∼6.61 and 2.95 J kg−1 K−1 under
reservoirs, respectively. The highest values of RCP are found a magnetic field of 50 kOe for compounds (1) and (2), re-
to be ∼78.3 and 40 J kg−1 for compounds (1) and (2), respec- spectively. The corresponding RCP is found to be 78.3 J kg−1
tively, for H = 5T, which is comparable with the values that for compound (1) and 40 J kg−1 for compound (2). In this
are reported in the literature for other PBAs listed in Table III. paper, we demonstrate that the open framework structure
The RCP increases linearly with increasing magnetic field, as of copper-hexacyanoferrate-based PBAs exhibits multifunc-
shown in Fig. 13(b). tional behavior.

ACKNOWLEDGMENTS
IV. CONCLUSIONS
P.B. would like to thank A. Jain and V. B. Jayakrish-
Potassium-ion-intercalated copper hexacyanoferrate open nan for their help in the neutron and XRD measurements,
framework materials, K2x/3 Cu[Fe(CN)6 ]2/3 · nH2 O [x = 0 for respectively.

[1] J. K. A. Gschneidner and V. K. Pecharsky, Magnetic refrigera- [9] P. W. Doheny, S. J. Cassidy, and P. J. Saines, Investigations
tion materials (invited), J. Appl. Phys. 85, 5365 (1999). of the magnetocaloric and thermal expansion properties of
[2] J. R. Gomez, R. F. Garcia, A. D. M. Catoira, and M. R. Gomez, the Ln3 (adipate)4.5 (DMF)2 (Ln = Gd−Er) framework series,
Magnetocaloric effect: A review of the thermodynamic cycles Inorg. Chem. 61, 4957 (2022).
in magnetic refrigeration, Renewable Sustainable Energy Rev. [10] S. Haldar and S. Ramasesha, Magnetocaloric effect in molec-
17, 74 (2013). ular spin clusters and their assemblies: Exact and Monte Carlo
[3] S. B. Roy, Chapter two—Magnetocaloric effect in intermetallic studies using exact cluster eigenstates, J. Magn. Magn. Mater.
compounds and alloys, Handb. Magn. Mater. 22, 203 (2014). 500, 166424 (2020).
[4] M. Ovichi, H. Elbidweihy, E. D. Torre, L. H. Bennett, M. [11] M. Gajewski, R. Pelka, M. Fitta, Y. Miyazaki, Y. Nakazawa,
Ghahremani, F. Johnson, and M. Zou, Magnetocaloric effect in M. Balanda, M. Reczynski, B. Nowicka, and B. Sieklucka,
NiMnInSi Heusler alloys, J. Appl. Phys. 117, 17D107 (2015). Magnetocaloric effect of high-spin cluster with Ni9 W6 core,
[5] T. Krenke, E. Duman, M. Acet, E. F. Wassermann, X. Moya, J. Magn. Magn. Mater. 414, 25 (2016).
L. Mañosa, and A. Planes, Inverse magnetocaloric effect in [12] F. Torres, J. M. Hernández, X. Bohigas, and J. Tejada, Giant and
ferromagnetic Ni–Mn–Sn alloys, Nat. Mater. 4, 450 (2005). time-dependent magnetocaloric effect in high-spin molecular
[6] V. Zeleňák, M. Almáši, A. Zeleňáková, P. Hrubovčák, R. magnets, Appl. Phys. Lett. 77, 3248 (2000).
tarasenko, S. Bourelly, and P. Llewellyn, Large and tunable [13] P. Konieczny, D. Czernia, and T. Kajiwara, Rotating
magnetocaloric effect in gadolinium-organic framework: Tun- magnetocaloric effect in highly anisotropic TbIII and
ing by solvent exchange, Sci. Rep. 9, 15572 (2019). DyIII single molecular magnets, Sci. Rep. 12, 16601
[7] G. Lorusso, J. W. Sharples, E. Palacios, O. Roubeau, E. K. (2022).
Brechin, R. Sessoli, A. Rossin, F. Tuna, E. J. L. McInnes, D. [14] Q.-F. Xu, B.-L. Liu, M.-Y. Ye, L.-S. Long, and L.-S. Zheng,
Collison et al., A dense metal-organic framework for enhanced Magnetocaloric effect and thermal conductivity of a 3D coor-
magnetic refrigeration, Adv. Mater. 25, 4653 (2013). dination polymer of [Gd(HCOO)(C2 O4 )]n , Inorg. Chem. 60,
[8] J.-J. Yin, T.-Q. Lu, C. Chen, G.-L. Zhuang, J. Zheng, X.-Y. 9259 (2021).
Zheng, and F. Shao, Magnetocaloric effect and slow magnetic [15] M. Fitta, W. Sas, and T. Korzeniak, Tunable critical temperature
relaxation on two-dimensional layered 3d − 4 f cluster-based and magnetocaloric effect in ternary Prussian blue analogue,
metal-organic frameworks, Cryst. Growth Des. 20, 4005 (2020). J. Magn. Magn. Mater. 465, 640 (2018).

014412-10
CLUSTER SPIN-GLASS STATE WITH MAGNETOCALORIC … PHYSICAL REVIEW B 108, 014412 (2023)

[16] M. Fitta, M. Szuwarzyński, P. Czaja, G. Lewińska, and M. analogue with a tetragonal structure, Chem. Lett. 34, 1278
Bałanda, The magnetocaloric effect in the thin film of a (2005).
Prussian blue analogue, Eur. J. Inorg. Chem. 2017, 4817 [33] M. D. Mukadam, A. Kumar, S. M. Yusuf, J. V. Yakhmi, R.
(2017). Tewari, and G. K. Dey, Spin-glass behavior in ferromagnetic
[17] N. Ma, R. Ohtani, H. M. Le, S. S. Sørensen, R. Ishikawa, S. Fe[Fe(CN)6 ] · xH2 O nanoparticles, J. Appl. Phys. 103, 123902
Kawata, S. Bureekaew, S. Kosasang, Y. Kawazoe, K. Ohara (2008).
et al., Exploration of glassy state in Prussian blue analogues, [34] M. D. Mukadam, S. M. Yusuf, P. Sharma, S. K. Kulshreshtha,
Nat. Commun. 13, 4023 (2022). and G. K. Dey, Dynamics of spin clusters in amorphous Fe2 O3 ,
[18] Y. Huang and S. Ren, Multifunctional Prussian blue ana- Phys. Rev. B 72, 174408 (2005).
logue magnets: Emerging opportunities, Appl. Mater. Today 22, [35] V. K. Anand, L. Opherden, J. Xu, D. T. Adroja, A. D. Hillier,
100886 (2021). P. K. Biswas, T. Herrmannsdörfer, M. Uhlarz, J. Hornung,
[19] P. Bhatt, M. K. Sharma, M. D. Mukadam, S. S. Meena, J. Wosnitza et al., Evidence for a dynamical ground state in
and S. M. Yusuf, Electrochemical capacitance and cubic- the frustrated pyrohafnate Tb2 Hf 2 O, Phys. Rev. B 97, 094402
rhombohedral phase transition of sodium intercalated ferri- (2018).
magnetic manganese hexacyanoferrate based open framework [36] N. Hanasaki, K. Watanabe, T. Ohtsuka, I. Kézsmárki, S. Iguchi,
material, Sustainable Mater. Technol. 35, e00532 (2023). S. Miyasaka, and Y. Tokura, Nature of the Transition between a
[20] A. Kumar, S. M. Yusuf, L. Keller, and J. V. Yakhmi, Micro- Ferromagnetic Metal and a Spin-Glass Insulator in Pyrochlore
scopic Understanding of Negative Magnetization in Cu, Mn, Molybdates, Phys. Rev. Lett. 99, 086401 (2007).
and Fe Based Prussian Blue Analogues, Phys. Rev. Lett. 101, [37] L. J. V. Stimpson, J. M. Powell, G. B. G. Stenning, M. Jura, and
207206 (2008). D. C. Arnold, Spin-glass behavior in Kx Ru4−y Niy O8 hollandite
[21] P. Bhatt, A. Kumar, S. S. Meena, M. D. Mukadam, and materials, Phys. Rev. B 98, 174429 (2018).
S. M. Yusuf, Magnetic proximity effect in ferrimagnetic- [38] A. Kumar, A. Senyshyn, and D. Pandey, Evidence for cluster
ferromagnetic core-shell Prussian blue analogues molecular spin glass phase with precursor short-range antiferromag-
magnet, Chem. Phys. Lett. 651, 155 (2016). netic correlations in the B-site disordered Ca(Fe1/2 Nb1/2 )O3
[22] E. Manuel, M. Evangelisti, M. Affronte, M. Okubo, C. perovskite, Phys. Rev. B 99, 214425 (2019).
Train, and M. Verdaguer, Magnetocaloric effect in hexa- [39] W. E. Buschmann, J. Ensling, P. Gütlich, and J. S. Miller, Elec-
cyanochromate Prussian blue analogs, Phys. Rev. B 73, 172406 tron transfer, linkage isomerization, bulk magnetic order, and
(2006). spin-glass behavior in the iron hexacyanomanganate Prussian
[23] P. Bhatt, S. Banerjee, M. Mukadam, P. Jha, M. Navaneethan, blue analogue, Chem. Eur. J 5, 3019 (1999).
and S. M. Yusuf, Enhanced hydrogen adsorption in alkali [40] W. E. Buschmann and J. S. Miller, Magnetic ordering and
metal based copper hexacyanoferrate Prussian blue analogue spin-glass behavior in first-row transition metal hexacyanoman-
nanocubes, J. Power Sources 542, 231816 (2022). ganate(IV) Prussian blue analogues, Inorg. Chem. 39, 2411
[24] S. Liu, G. L. Pan, G. R. Li, and X. P. Gao, Copper hexacyano- (2000).
ferrate nanoparticles as cathode material for aqueous Al-ion [41] B. Gao, J. Yao, and D. Xue, Vacancy-induced spin-glass behav-
batteries, J. Mater. Chem. A 3, 959 (2015). ior of Prussian blue analogue FeII1.1 Cr IIx [Cr III (CN)6 ]0.6−x · nH2 O
[25] C. D. Wessells, R. A. Huggins, and Y. Cui, Copper hexacyano- nanowires, J. Magn. Magn. Mater. 322, 2505 (2010).
ferrate battery electrodes with long cycle life and high power, [42] L. Lartigue, S. Oh, E. Prouzet, Y. Guari, and J. Larionova,
Nat. Commun. 2, 550 (2011). Superspin-glass behavior of Co3 [Fe(CN)6 ]2 Prussian blue
[26] H. M. Rietveld, A profile refinement method for nuclear and nanoparticles confined in mesoporous silica, Mater. Chem.
magnetic structures, J. Appl. Cryst. 2, 65 (1969). Phys. 132, 438 (2012).
[27] S. M. Yusuf and L. M. Rao, Magnetic studies in mesoscopic [43] B. Gao, J. Yao, and D. Xue, Spin-glass behavior of
length scale using polarized neutron spectrometer at Dhruva the polyvinyl pyrrolidone-protected Prussian blue analog
reactor, Trombay, Pramana J. Phys. 47, 171 (1996). K1.14 Mn[Fe(CN)6 ]0.88 nanocubes, Phys. B: Condens. Matter
[28] P. Bhatt, N. Thakur, M. D. Mukadam, S. S. Meena, and S. 406, 2528 (2011).
M. Yusuf, Evidence for the existence of oxygen clustering and [44] M. Ghanathe, A. Kumar, and S. M. Yusuf, Microscopic and
understanding of structural disorder in Prussian blue analogues mesoscopic understanding of magnetization compensation phe-
molecular magnet M1.5 [Cr(CN)6 ] · zH2 O (M = Fe and Co): nomenon in ferrimagnetic Li0.5 FeCr 1.5 O4 spinel, J. Appl. Phys.
Reverse Monte Carlo simulation and neutron diffraction study, 125, 093903 (2019).
J. Phys. Chem. C 117, 2676 (2013). [45] I. Dhiman, A. Das, R. Mittal, Y. Su, A. Kumar, and A.
[29] P. Gütlich, C. Schröder, and V. Schünemann, Mössbauer Radulescu, Diffuse neutron scattering study of magnetic
spectroscopy—An indispensable tool in solid state research, correlations in half-doped La0.5 Ca0.5−x Sr x MnO3 manganites
Spectrosc. Eur. 24, 21 (2012). (x = 0.1, 0.3, and 0.4), Phys. Rev. B 81, 104423 (2010).
[30] P. Gütlich and C. Schröder, Mössbauer spectroscopy, Bunsen- [46] S. M. Yusuf and A. Kumar, Neutron scattering of advanced
Magazin 12, 4 (2010). magnetic materials, Appl. Phys. Rev. 4, 031303 (2017).
[31] H. Tokoro, S.-i. Ohkoshi, T. Matsuda, and K. Hashimoto, A [47] M. Evangelistia, E. Manuel, M. Affrontea, M. Okubo, C.
large thermal hysteresis loop produced by a charge-transfer Trainc, and M. Verdaguer, Vacancy-driven magnetocaloric ef-
phase transition in a rubidium manganese hexacyanoferrate, fect in Prussian blue analogues, J. Magn. Magn. Mater. 316,
Inorg. Chem. 43, 5231 (2004). e569 (2007).
[32] K. Wataru, I. Tohru, Y. Hisashi, T. Yayoi, H. Kazuhito, and [48] H. Oesterreicher and F. T. Parke, Magnetic cooling near Curie
O. Shin-ichi, Synthesis of ferromagnetic CsCuCr Prussian blue temperatures above 300 K, J. Appl. Phys. 55, 4334 (1984).

014412-11