ARTICLE IN PRESS

Journal of Magnetism and Magnetic Materials 307 (2006) 227–231

www.elsevier.com/locate/jmmm

Magnetocaloric effect in ferrite nanoparticles

P. Poddara,1, J. Gassa, D.J. Rebara, S. Srinatha, H. Srikantha,,
S.A. Morrisonb, E.E. Carpenterb
a
Department of Physics, University of South Florida, Tampa, FL 33620, USA
b
Department of Chemistry, Virginia Commonwealth University, Richmond, VA 23284, USA
Received 5 December 2005; received in revised form 27 March 2006
Available online 22 May 2006

Abstract

A comparative study of the magnetocaloric effect (MCE) is reported in two different types of chemically synthesized magnetic
nanoparticle systems—cobalt ferrite and manganese zinc ferrite with mean size around 5 and 15 nm, respectively. While CoFe2O4
nanoparticles were synthesized using co-precipitation, the Mn0.68Zn0.25Fe2.07O4 (MZFO) nanoparticles were prepared by reverse micelle
technique using AOT as surfactant. Our results indicate that the change in entropy with the change in applied magnetic field (dS/dH) is
reasonably large for this class of nanoparticles and has a wide distribution over a broad temperature range covering the region above and
below the blocking temperature. The maximum entropy change is influenced by the particle size, overall distribution in anisotropy and
magnetic moments.
r 2006 Elsevier B.V. All rights reserved.

Keywords: Magnetocaloric effect; Magnetization; Nanoparticles

1. Introduction room temperature applications require magnetic fields as

high as 2–10 T for efficient operation. With the advent of
Magnetic refrigeration is rapidly becoming competitive new micro- and nanoscale electronic devices, miniaturiza-
with conventional gas compression technology because it tion of the cooling techniques (cooler on a chip) for Micro
offers considerable operating cost savings, mechanical Electro Mechanical Systems (MEMS) and Nano Electro
stability, light weight and better performance [1]. Refrig- Mechanical Systems (NEMS) applications are required.
eration techniques have been in high demand for residen- From a materials perspective, it is also extremely difficult
tial cooling, space and defense applications, consumer to make rare-earth binary and ternary alloys that exhibit
goods, medical applications, etc. Development of low-cost, giant MCE in thin film form while maintaining their
long-lasting and highly reliable coolers has been an active stoichiometry and desirable properties. This is a major
field of research from the materials science as well as drawback as coolers for MEMS and NEMS devices require
engineering perspectives. Magnetocaloric materials like materials in compatible thin film geometries. Additionally,
diluted paramagnetic salts and intermetallic compounds most of the current bulk MCE materials rely on the
have also been employed as refrigerants [2,3]. For many entropy change associated with the paramagnetic–ferro-
years magnetocaloric effect (MCE) has been successfully magnetic transition at the Curie temperature. The MCE
used to reach ultra-low temperatures in a research shows a maximum around the Curie temperature and
environment. Gd5Si1.5Ge2.5 is the best known example of drops rapidly on either side [5]. All these factors indicate
a material that exhibits giant MCE [4]. However, all the that there is a critical need for further materials research in
existing materials that are promising for intermediate to potential magnetic refrigerants wherein these limitations
Corresponding author. Tel.: +1 8139742467; fax: +1 8139745813. can be overcome.
E-mail address: [email protected] (H. Srikanth).
Magnetic nanoparticles provide an attractive alternative
1
Permanent address: Materials Chemistry Division, National Chemical to conventional bulk MCE materials because of their ease
Laboratory, Pune 411008, India. of assembly in thin film form and other desirable features

0304-8853/$ - see front matter r 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2006.04.007
 ARTICLE IN PRESS
228 P. Poddar et al. / Journal of Magnetism and Magnetic Materials 307 (2006) 227–231

such as control over the entropy change across the

superparamagnetic-blocking transition usually determined
by the particle size. It has been theoretically shown that by
reducing the average particle size close to the single
magnetic domain, magnetic entropy change increases by
several orders of magnitude as compared to the entropy
change in bulk materials [6]. In addition, the large surface
area in nanostructured materials has the potential to
provide better heat exchange with the surrounding
materials. By careful design of core-shell structures, it
would be possible to control the heat exchange between the
magnetic nanoparticles and the surrounding matrix. Our
contention is that by using a multilayer structure with
varying particle sizes, a broader range of cooling (from
room temperature and above to sub-Kelvin temperatures)
can potentially be achieved. Theoretical studies have
suggested that nanostructured crystalline superparamag-
netic materials are attractive candidates for MCE [7].
To date, not too many studies of the MCE in
nanocomposites have been reported [8–12]. Even though
these reports indicate an enhanced MCE due to the
reduction of grain size to the nanoscale dimensions, Fig. 1. Field-cooled (FC) and zero-field cooled (ZFC) magnetization
the overall effect is quite low due to the fact that the measured at a field of 100 Oe. Insets show TEM images for CoFe2O4, (top
panel) and MZFO (bottom panel).
nanoparticles have to be dispersed in a matrix in order to
prevent the agglomeration associated with the surface
energy minimization. In a nanocomposite material, the
concentration of nanoparticles remains less than 50% and Monodisperse manganese zinc ferrite (MZFO) nanopar-
the non-magnetic inactive matrix acts as a heat sink. In this ticles were synthesized using the reverse micelle method
work, instead of having the nanocomposite with thermally using bis-(2-ethylhexl) sodium sulfosuccinate (AOT) as the
conducting Ag or Au gold matrices, we have used the surfactant medium. These particles were extensively
surfactant-coated particles thereby increasing the active characterized for their size and crystalline phase using
magnetocaloric content and also decreasing the effective EXAFS and XRD. The details are presented elsewhere
mass of the nanocomposite. While for practical refrigera- [14]. The average size of the particles in this case is around
tion purposes this configuration is unlikely to be efficient, 15 nm with a much narrower standard deviation than in the
our goal is to try to obtain a fundamental understanding of case of the cobalt ferrite system. A representative TEM
the cooperative magnetism in such nanoparticle assemblies image is also shown in the inset of the bottom panel of
and its influence on the MCE. In this paper, we have Fig. 1 where the image is blurred due to the presence of
presented our results on the entropy change in cobalt excess surfactant. Removal of the surfactant and subse-
ferrite and manganese zinc ferrite nanoparticles. quent annealing of nanoparticles are known to cause
clustering and fusion of particles, which usually results in
2. Synthesis of ferrite nanoparticles broadening of the superparamagnetic-blocking transition
in susceptibility. Since one of our goals was to retain the
To synthesize CoFe2O4 nanoparticles, we used the monodisperse nature and the associated sharp transition in
method developed by Song and Zhang [13]. A mixture of this case, we did not attempt to remove the surfactant.
cobalt (II) acetylacetonate, phenyl ether, 1,2-hexadecane-
diol, oleic acid and oleylamine were mixed and heated to 3. Magnetism and MCE measurements
156 1C. At this point iron (III) acetylacetonate mixed in
phenyl ether was added and heated to reflux (260 1C) for Magnetic characterization of the samples was done using
1 h. The particles precipitated upon addition of excess a commercial Physical Property Measurement System
ethanol and centrifugation. They were then washed several (PPMS) from Quantum Design. Fig. 1 shows the zero-
times using deionized water and magnetic decantation. field cooled (ZFC) and field-cooled (FC) magnetization for
After that, particles were dried and used in powder form CoFe2O4 (top panel) and MZFO (bottom panel) samples.
for magnetic measurements or suspended in hexane for Insets in the panels show the TEM images of the
further size selection. Analysis of transmission electron corresponding nanoparticle samples. While the overall
micrograph (TEM) images such as the one shown in the trend in ZFC and FC magnetization curves are the same,
inset of Fig. 1 established that the average size of the the variation in the blocking-superparamagnetic transition
particles is
5 nm with a standard deviation of 3 nm. peak behavior is strikingly different for both the samples as
 ARTICLE IN PRESS
P. Poddar et al. / Journal of Magnetism and Magnetic Materials 307 (2006) 227–231 229

expected. For the CoFe2O4 nanoparticles, a broad peak is or 2 T [18]. In systems exhibiting metamagnetic transitions
seen in the ZFC curve located at around 223 K, which is and close coupling between the magnetic and structural
characteristic of the blocking transition followed by a degrees of freedom (such as the manganites), giant MCE
gradual decrease in magnetization at low temperature. This with one or two orders of magnitude bigger entropy change
broad variation is consistent with the polydisperse nature have been reported in the vicinity of the Curie temperature
of this system with a large distribution in size. MZFO [19]. However these changes do not persist over a broad
particles are relatively monodisperse and as expected, the temperature range. As indicated before, it is somewhat
blocking temperature is quite sharp and located around cumbersome to systematically tune the ferromagnetic Curie
46 K [15]. The blocking temperature is a function of the temperature in materials through solid-state chemistry and
experimental measurement time scale (which is the same in doping. It is precisely for this reason that magnetic
both cases), external magnetic field (100 Oe for both the nanoparticles may have a unique advantage if the MCE
samples), particle volume, anisotropy constant and inter- could be enhanced in them. Unlike the Curie temperature,
particle interactions [16]. the blocking temperature TB can be easily varied by
We would like to briefly comment on the correlation adjusting the particle size, shape, dispersion and inter-
between the average particle size and the blocking particle interactions.
temperatures in the two ferrite systems. Normally one The top panel of Fig. 2 shows the M–H curves for
expects a lower blocking temperature for smaller size CoFe2O4 particles over a field range of 0.5–3 T taken at
particles and a higher TB for larger size particles. However, 10 K temperature intervals from 10 to 300 K, which spans a
this simple correlation holds good only for monodisperse, sizeable region covering the blocking peak. The bottom
non-interacting nanoparticles. Intrinsic material properties
such as magnetocrystalline and surface anisotropies along
with inter-particle interactions would play dominant roles
in increasing the blocking temperature and often tend to 10 K
completely mask the size-dependence. The MZFO nano- 35 CoFe2O4
particles are synthesized by reverse micelle method, which
is known to yield highly monodisperse nanoparticles with T
individual nanoparticles completely encapsulated by fatty
M (emu/g)

acid surfactants thus minimizing chances of particle

30
agglomeration and interactions. So the sharper transition
300 K
and lower blocking temperature are consistent with that. In
the case of Co ferrite nanoparticles, the chemical route
does not guarantee perfect coating of particles with
surfactants; so we suspect that when the samples in powder 25
form are prepared for magnetization measurements, there
would be clustering and associated strong dipolar or even
exchange interactions that would broaden the blocking 5 10 15 20 25 30 35
transition and push it to higher temperatures. The much H (kOe)
larger TB in Co ferrite nanoparticles is consistent with
other reports in the literature [17].
As monodisperse nanoparticles exhibit a larger change in 30 kOe
the magnetic entropy in comparison to the polydisperse 30
samples, in this study we have deliberately chosen to
compare the MCE between two such representative
systems. The reasoning behind this is based on the simple
M (emu/g)

conjecture that if the magnetic entropy change (DS) is 20

maximized in the vicinity of the blocking transition, then
having a broad transition would retain a nearly constant H
DS over a broader temperature range. However, it should
10 CoFe2O4
be noted that for two samples with same saturation
magnetization, the total area under the DS vs. T curve
remains constant, therefore the broader DS vs. T curve will 0.1 kOe
reduce the maximum entropy change [6]. 0
Following the thermodynamical Maxwell’s equations, 100 200 300
the magnetic entropy change can be calculated from T (K)
measurements of M (H, T) i.e. field- and temperature- Fig. 2. The top panel shows isothermal curves of magnetization vs. field
dependent magnetization curves. Typical values of the (M–H) for CoFe2O4 and the bottom panel shows the magnetization vs.
entropy change range from 0.1 to 1 J/kg K over a field of 1 temperature curves obtained from the M–H curves.
 ARTICLE IN PRESS
230 P. Poddar et al. / Journal of Magnetism and Magnetic Materials 307 (2006) 227–231

panel shows the M–T curves as obtained from the top all the particles in its direction, whereas just above the
panel. The entropy change in a magnetic system can be blocking temperature, the nanoparticles are superpara-
calculated from the thermodynamic Maxwell relation as magnetic and are free to re-orient even at low field
follows: strengths.


 In Fig. 4, the entropy change for the case of MZFO
qS qM
¼ (1) nanoparticles is plotted based on a similar analysis. The
qH T qT H upper panel represents the low-field data and the bottom
or integrating over the field, panel shows the entropy change as calculated in the higher
magnetic field region. It should be noted from Fig. 4 that
Z H

qM the overall (DS) values are larger in comparison to the
DS ¼ dH. (2)
0 qT H values for CoFe2O4 particles, which is likely due to the
relatively higher monodispersity of the MZFO nanoparti-
In Fig. 3, we have presented the entropy change vs. cles. Another notable aspect is that above
75 K, the
temperature curve for CoFe2O4 and MZFO nanoparticles entropy change vs. temperature profiles (both low-field
numerically calculated from the data sets shown in Fig. 2 region and high-field region) for MZFO nanoparticles
using the Maxwell equation. The results for CoFe2O4 for show a broad maximum with a nearly flat behavior over a
low fields are plotted in Fig 3 (top panel) and for high fields much larger temperature range. This behavior could be
are plotted in bottom panel of Fig. 3. The temperature potentially useful for the operation of cooling devices over
dependence of magnetic entropy change is quite interesting. a broad temperature range. At higher temperatures, the
The low-field (750 Oe–3 kOe) entropy curves for CoFe2O4 entropy change shows a decrease in its value, which is
as shown in the top panel of Fig. 3 clearly shows a peak in expected as the value of qM/qT goes down (lower panel
the vicinity of TB. With further increase in the temperature of Fig. 1).
above
250 K, a sharp drop in the entropy change is It is clear from Figs. 3 and 4 that while the expected
observed, followed by an increase. However, it should be trend of increased DS is observed for higher applied fields
noted that DS vs. T curves for higher field values (as seen in in both systems, the DS increases uniformly as a function of
the lower panel of Fig. 3) do not show the peak. A similar temperature (up to 300 K which is the limit of our PPMS)
observation was reported earlier in the case of the and does not follow the shape of the blocking peak which is
Fe1xHgx system [20]. The appearance of the peak in centered around 220 K for the 5 nm CoFe2O4 particles.
MCE close to the blocking temperature for low fields can This is even more apparent in the case of MZFO which has
be understood as follows. Below the blocking temperature, a much lower blocking temperature (
46 K), but appears
low fields cannot produce a maximum entropy change to exhibit a broad maximum in DS around a temperature
unless the field is strong enough to orient the moments of about 4–5 times larger than the blocking temperature.

2
750 Oe 1 kOe MZFO
1.3 KOe 2 2 kOe
CoFe2O4 3 kOe
2 KOe
3 KOe
1
1
(-∆S (10-2 J/Kg K) )

-∆S 10-2 (J/kg K)

0 0
10 kOe
15 10 KOe 20 kOe
20
20 KOe 30 kOe
30 KOe
10

10
5

0
0
50 100 150 200 250 300 50 100 150 200 250 300
T (K) T (K)

Fig. 3. The change in entropy vs. temperature curves at various magnetic Fig. 4. The change in entropy vs. temperature curves at various magnetic
field values for CoFe2O4 nanoparticles at low fields (top panel) and high field values for MZFO nanoparticles at low fields (top panel) and high
fields (bottom panel). Error bars resulting from the PPMS-based fields (bottom panel). Error bars resulting from the PPMS-based
measurements are shown. measurements are shown.
 ARTICLE IN PRESS
P. Poddar et al. / Journal of Magnetism and Magnetic Materials 307 (2006) 227–231 231

These results imply that a reasonably high DS can be References

maintained at temperatures well above the blocking
transition, which makes it different from conventional [1] V.K. Pecharsky, K.A. Gschneidner Jr., J. Magn. Magn. Mater. 200
(1999) 44.
ferromagnets where DS drops off sharply away from Tc.
[2] V.K. Pecharsky, K.A. Gschneidner Jr., J. Appl. Phys. 86 (1999) 565;
The observation of a peak in MCE close to the blocking V. Pecharsky, K.A. Gschneidner Jr., Phys. Rev. Lett. 78 (1997) 4494.
temperature for low fields in the case of CoFe2O4 and an [3] M. Földeàki, R. Chahine, T.K. Bose, J. Appl. Phys. 77 (1995) 3528.
increase of MCE with increase in temperature for high [4] L.H. Lewis, M.H. Yu, R.J. Gambino, Appl. Phys. Lett. 83 (2003) 515.
fields could be interesting as these two trends might serve [5] K.J. Klabunde (Ed.), Nanoscale Materials in Chemistry, Wiley, New
different purposes for different applications. The origin of York, 2001.
[6] R.D. McMichael, R.D. Shull, L.J. Swartzendruber, L.H. Bennett,
these behaviors of DS needs further investigation in R.E. Watson, J. Magn. Magn. Mater. 111 (1992) 29.
nanoparticle samples where different parameters such as [7] R.D. Shull, IEEE Trans. Magn. 29 (1993) 2614.
the average size, distribution, and interactions as well as [8] T.A. Yamamoto, M. Tanaka, Y. Misaka, T. Nakagawa,
shape are controllably changed and their individual T. Nakayama, K. Niihara, T. Numazawa, Scr. Mater. 46 (2002) 89.
influence on DS monitored. [9] T.A. Yamamoto, M. Tanaka, T. Nakayama, K. Nishimaki,
T. Nakagawa, M. Katsura, K. Niihara, Jpn. J. Appl. Phys. 39
In comparison to the bulk materials that show giant (2000) 4761.
MCE, the values for the entropy change shown in this [10] V. Provenzano, J. Li, T. King, E. Canavan, P. Shirron, M. Dipirro,
work are still quite small but are significantly higher than R.D. Shull, J. Magn. Magn. Mater. 266 (2003) 185.
what has been reported so far for several other iron or iron- [11] L.E. Hueso, P. Sande, D.R. Miguens, J. Rivas, F. Rivadulla,
M.A. Lopez-Quintela, J. Appl. Phys. 91 (2002) 9943.
oxide-based nanoparticle systems. This is associated with
[12] T. Kinoshita, S. Seino, H. Muruyama, Y. Otome, K. Okitsu,
stronger inter-particle interactions in our samples and T. Nakayama, K. Niihara, T. Nakagawa, T.A. Yamamoto, J. Alloys
consistent with models that indicate that introducing some Compounds 365 (2004) 281.
amount of interactions could lead to higher entropy change [13] Q. Song, Z.J. Zhang, J. Am. Chem. Soc. 126 (2004) 6164.
in comparison to strictly non-interacting superparamag- [14] S. Calvin, E.E. Carpenter, V.G. Harris, S.A. Morrison, Appl. Phys.
netic particles [8]. To further enhance the change in the Lett. 81 (2002) 3828;
S.A. Morrison, C.L. Cahill, E.E. Carpenter, S. Calvin, V.G. Harris,
magnetic entropy change for nanoparticles systems, we J. Appl. Phys. 93 (2003) 7489.
need to further control the size dispersion, inter-particle [15] P. Poddar, J.L. Wilson, H. Srikanth, S.A. Morrison, E.E. Carpenter,
interaction and generally explore other materials with Nanotechnology 15 (2004) S570.
higher magnetic moments. [16] P. Poddar, T. Shafir, T. Fried, G. Markovich, Phys. Rev. B 66 (2002)
060403.
[17] C. Liu, A.J. Rondinone, Z.J. Zhang, Pure Appl. Chem. 72 (2000) 37.
[18] X. Bohigas, J. Tejada, E. del Barco, X.X. Zhang, M. Sales, Appl.
Phys. Lett. 73 (1998) 390.
Acknowledgment
[19] Z.B. Guo, Y.W. Du, J.S. Zhu, H. Huang, W.P. Ding, D. Feng, Phys.
Rev. Lett. 78 (1997) 1142.
H.S. thanks the National Science Foundation for [20] M.S. Pedersen, S. Morup, S. Linderoth, C. Johansson, M. Hanson,
supporting this research through grant NSF-CTS-0408933. J. Phys: Condens. Matter 9 (1997) 7173.