Materials Science and Engineering B 177 (2012) 14–18

Contents lists available at SciVerse ScienceDirect

Materials Science and Engineering B

journal homepage: www.elsevier.com/locate/mseb

Giant magnetostriction in magnetite nanoparticles

G. Balaji a,b , R. Aravinda Narayanan c , A. Weber a , F. Mohammad a,d , C.S.S.R. Kumar a,∗
a
Center for Advanced Microstructures & Devices, Louisiana State University, 6980 Jefferson Highway, Baton Rouge, LA 70806, USA
b
Department of Chemistry, BITS-Pilani Hyderabad Campus, Hyderabad 500078, India
c
Department of Physics, BITS-Pilani Hyderabad Campus, Hyderabad 500078, India
d
Environmental Toxicology, Southern University and A&M College, Baton Rouge, LA 70813, USA

a r t i c l e i n f o a b s t r a c t

Article history: Typically the value of the magnetostrictive coefficient () observed for bulk magnetic materials such
Received 18 April 2011 as cubic ferrites is 10−6 . However, giant magnetostriction ( ≤ 10−3 ) is only observed in a few bulk
Received in revised form 30 June 2011 intermetallic materials based on alloys of rare earth and iron such as TbFe, TbFe2 , DyFe2 and Terefenol-
Accepted 4 September 2011
D. While giant magnetostriction is known in nanostructured films, we show for the first time, this
Available online 17 September 2011
phenomenon occurs in magnetic nanoparticles. By using in-field small angle X-ray scattering (SAXS)
as a tool, we demonstrate that a 4% relative change in dimension of the particle can be observed
Keywords:
in 5.0 nm Fe3 O4 nanoparticles at room temperature with 1 kG magnetic field. Also, we propose that
Giant magnetostriction
Magnetite nanoparticles
the observed values are due to interaction effects and magnetoelastic coupling of particle magnetic
Magnetoelastic coupling moments and external magnetic field.
SAXS © 2011 Elsevier B.V. All rights reserved.

1. Introduction nanoparticles is the dipolar interactions between the particles [9].

Dipole–dipole interactions are basically long range considering that
The dimensions of magnetic materials change in response to the particles are tiny magnetic dipoles. There are several evidence
the external applied magnetic field which is called “magnetostric- cited in the literature that show the influence of dipole–dipole type
tion”. It is the fractional change (l/l) in the length of a magnetic interactions on thermal relaxation of magnetization [10]. This can
single crystal when a field is applied along the easy axis of mag- also lead to spin glass like ordering due to random anisotropy [11].
netization. This is designated in terms of “” and is usually in the Apart from these interactions, a variety of interesting phenomenon
order of 10−6 . An improved magnetostrictive behavior is observed have been proposed to occur in interacting magnetic nanoparticles
in rare earth elements containing iron alloys such as Terefenol, [12].
Terefenol-D, DyFe2 and SmFe2 . These materials are known as giant The question that needs to be answered is: how is the phe-
magnetostrictive materials and the value of  observed for these nomenon of magnetostriction affected in magnetic nanoparticles.
materials is of the order of 10−3 [1]. Magnetostrictive materials Towards this knowledge of the particle size is required in order
have applications in areas such as sound generators, magneto- to predict the magnetic behavior and vice versa. Previous char-
acoustic transformers, and actuators for opto-electronic systems, acterization techniques utilized to study magnetostriction are
devices for non-destructive control and remote detection and rang- strain gauge methods, dilatometry method, optical interferom-
ing. As rare earth elements are expensive, the applications of giant etry and ferromagnetic resonance techniques [13]. Small angle
magnetostrictive materials have been restricted [2]. This has led X-ray scattering (SAXS) is a useful and potent tool to study the
to the studies and applications of currently known giant mag- size and structures at the nanoscale [14]. Although other tech-
netistrictive materials in thin film form [2,3]. There is a necessity niques can be used to characterize the nanomaterials, SAXS is a
for suitable giant magnetostrictive material which is efficient in its direct tool that can resolve structures in nanometer scale and sup-
performance and is relatively inexpensive. plement the magnetic data. This feature allows for establishing
In nanoscale regime, the particles are single domain and their a structure property relationship in nanomaterials. In this study,
properties are chiefly governed by particle size [4,5], size distribu- we report the dimensional change in superparamagnetic Fe3 O4
tion [6], shape [7] and surface effects [8]. Yet another important nanoparticles covered with oleic acid–oleylamine surfactants using
factor that affects the static and dynamic properties of magnetic in-field SAXS technique. SAXS has never been utilized to study
magnetostriction previously. Furthermore, we bring out, the impli-
cations of such dimensional change on the magnetic properties
∗ Corresponding author. Tel.: +1 225 578 9320; fax: +1 225 578 6954. through SQUID magnetometer measurements. We demonstrate
E-mail address: ckumar1@lsu.edu (C.S.S.R. Kumar). that dimensional change is reversible and the origin stems

0921-5107/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.mseb.2011.09.023
 G. Balaji et al. / Materials Science and Engineering B 177 (2012) 14–18 15

Fig. 1. (a: left) TEM micrographs of Fe3 O4 nanoparticles. Inset: HRTEM images showing lattice fringes. (b: right) Particle Size distribution.

from the coupling of particle magnetic moments with external of 0.0066 to 0.164 Å−1 , where Q = 4␲sin /,  is the wavelength
field. of X-rays used and 2 is the scattering angle. The Q calibra-
tion was done using silver behenate, which is known to have a
well-defined structure (d = 58.48 Å). The measurements were per-
2. Experimental formed on powder samples sandwiched between Kapton tapes at
room temperature. The sample is placed at the center between
Magnetite nanoparticles were synthesized as described by two permanent magnets and the distance between the mag-
Wang et al. [15]. Brief procedure for sample preparation is the fol- nets is changed in order to expose the samples to different field
lowing: 0.71 g of iron acetylacetonate was mixed vigorously under strengths. The magnetic field strength was determined by using a
inert atmosphere with 20 mL of phenyl ether containing 2 mL of Gaussmeter. SAXS data were corrected for the Kapton tape back-
oleic acid and 2 mL of oleylamine. 2.58 g of 1,2 hexadecanediol was ground (Kapton tape did not show any structural feature in the
added to the solution. The solution was heated to around 210 ◦ C Q-range measured). The data reduction procedures have been per-
with reflux for 2 h. After cooling to room temperature, ethanol was formed using Irena 2.0 software package [16] running in Igor Pro
added and the black precipitate was separated by centrifugation. 6.0.5.
The black precipitate was then dispersed in hexane. The precipi-
tated product was washed again with ethanol and dried completely
to product in powder form. 3. Results and discussion
TEM analysis was carried out by using a Hitachi H-7600 with a
125 kV accelerating voltage and operated on a JEOL 100CX with X-ray diffraction pattern observed for Fe3 O4 nanoparticles can
80 kV. The samples for TEM were prepared by dropping a hex- be assigned to the peaks corresponding to standard Fe3 O4 mate-
ane solution of Fe3 O4 nanoparticles onto a carbon coated copper rial except that the reflections are broad due to nanosize nature of
grid and evaporating the solvent at room temperature. X-ray pow- the particles. The lattice parameter of 0.89 nm was obtained; which
der diffraction (XRD) patterns were recorded on a Bruker-Siemens is the same as the bulk value. Fig. 1 shows the Transmission Elec-
D5000 automated powder diffractometer using Cu K␣ radiation. tron Microscopy (TEM) images of Fe3 O4 nanoparticles. The mean
The data were collected from 20 to 80◦ in 2 at a scan rate of size is around 5.0 nm and the image shows that the particles have
0.06◦ per step. The magnetic measurements were conducted using a spherical shape. High-resolution TEM (Fig. 1. inset) shows the lat-
Quantum Design MPMS-5S superconducting quantum interference tice fringes indicating that the sample is a collection of many single
device (SQUID) magnetometer by placing the samples (powders), crystals. The particle size distribution obtained with the synthesis
in gel-capsule stuffed with cotton or kim-wipe paper which is employed in this study is displayed in Fig. 1(b).
then placed inside a plastic straw. The magnetization was mea- The magnetic properties of 5.0 nm Fe3 O4 particles were mea-
sured as a function of field and temperature. The temperature sured. At 300 K, the magnetization curve measured as a function
dependence of magnetization was measured in both zero-field- of field did not show any hysteresis and the curve did not saturate
cooled (ZFC) and field-cooled (FC) conditions. The magnetization indicating that Fe3 O4 nanoparticles exhibit superparamagnetism
values are reported based on the mass of Fe3 O4 within the sam- at the measurement temperature. The saturation magnetization
ple which was determined from TGA data. Typical background ( s ) was obtained from 1/H vs M curve and it is determined to be
corrections to the magnetic data have been duly performed. The 39.2 emu/g which is very much less than that of the bulk value
diamagnetic corrections due to the organic part have not been (92 emu/g). At 5 K, the magnetization curve shows both hysteresis
employed. SAXS measurements were carried out at the Center and saturation. The  s was measured to be 55.5 emu/g; still consid-
for Advanced Microstructures and Devices with synchrotron radi- erably less than the bulk value. The reasons for such low observed
ation of wavelength,  = 1.55 Å. The scattering pattern is imaged value for  s are due to the surface spin disorder, cation distribu-
with a Gabriel style multi-wire gas detector with a 200 mm active tion, superparamagnetic fractions and spin canting due to nanosize
diameter and a resolution of 200-250 ␮m FWHM in a 1024 × 1024 nature of the particles [17–19].
array. The sample and detector chamber were kept under vac- Fig. 2 displays the zero-field cooled (ZFC) and field-cooled (FC)
uum during the measurements to minimize scattering from the curves measured as a function of temperature at 20 Oe. The ZFC
air. Scattering curves were monitored in a wave-factor (Q) range curve exhibits a maximum at 18.1 K and this temperature is the
16 G. Balaji et al. / Materials Science and Engineering B 177 (2012) 14–18

3
12
FC
10
Magnetization (emu/g)

TB (K)
8
2
6

4
0 1 2 3 4 5
3
H x 10 (Oe)
1

ZFC

0
0 40 80 120
T (K)

Fig. 2. Zero-field cooled and field cooled magnetization as a function of tempera-

ture at 20G external field. Inset displays the dependence of TB on external applied
magnetic field.

blocking temperature (TB ) of 5.0 nm Fe3 O4 nanoparticles. For T > TB ,

the thermal energy dominates over magnetic anisotropy energy Fig. 3. In-field SAXS data for Fe3 O4 nanoparticles at various indicated fields. The
solid line is the fit for Percus-Yevick model for interacting spheres as discussed in
and this process can be described by Neel’s relaxation expression
the text. Inset shows the data measured at zero field and the solid line is the fit of
[20]. data with the unified model displaying different structural levels.
 KV 
 = 0 exp (1)
kT
measures the size of the particles were performed at room tem-
where  0 is the prefactor, K is magnetocrystalline anisotropy, V is perature. We believe that our explanation is very pertinent except
the volume of the particle, k is the Boltzmann constant and T is the fact that the magnitudes obtained from SAXS measurements
the temperature. The other factors that contribute to the energy could be even larger at low temperatures.
barrier (KV) are interparticle interaction energy, shape anisotropy Fig. 3 inset shows the SAXS data measured at zero field. The
and surface anisotropy. data in the low Q range between 0.018 and 0.026 Å−1 were fitted
Fig. 2 inset shows the dependence of the blocking tempera- for power law dependence and the factor was determined to be
ture on external applied magnetic field for Fe3 O4 nanoparticles. −2.52. This value which is different from the Porod factor of 4.0
ZFC magnetization was measured by cooling down the sample [24] indicates the fractal nature of the nanomaterial. The power
from high temperature each time under different dc magnetic field law showed little or no dependence on the external magnetic field.
strength from 5000 Oe to 100 Oe regressively. It was shown ear- Hence, the focus was on an interference peak in the high Q region
lier that the field dependence of TB take a functional form of H which depended on the magnetic field. The center of the peak
where  = −2 at low external fields and  = −2/3 at high fields. The is at Q = 0.11767 Å−1 (the error associated with the center of the
crossover of the  value occur at H = Hc [21]. Experimentally, such peak is around 0.3%) which corresponds to an inter-particle spac-
dependence was observed in uncapped 5 nm spherical Fe3 O4 par- ing (d = 2␲/Q) of 5.3 nm. Similar peak was observed in ␥-Fe2 O3
ticles especially, the H−2/3 dependence at higher fields [22]. nanoparticles [25] and its origin was explained based on dipolar
There are two important features in our data that needs to be magnetic interactions between magnetic nanoparticles leading to
discussed in detail. At fields of 5000 Oe, TB is below 5 K. As the a chain like structure.
field is decreased to 2500 Oe, the TB increases to around 8 K and The particle size distribution of Fe3 O4 nanoparticles at each field
attains a maximum value of 11.8 K at 1000 Oe. When the field is strength was determined by fitting the SAXS data. The size distri-
further decreased, TB decreases to 8 K at 100 Oe. Similar anoma- bution did not show any appreciable trend in response to field. A
lous magnetic data was previously reported for oleic acid coated representative SAXS data with fitting curve (solid line) is displayed
Fe3 O4 nanoparticles [11] and oleic acid–oleylamine capped 2.6 nm in Fig. 3 inset. We used unified fit analysis which is ideal for resolv-
size L10 FePt nanoparticles [23]. In the former, it was attributed to ing hierarchical structures [26]. The philosophy of this method is
interparticle interaction effects while in the latter it was attributed that scattered intensity from each level of the hierarchy is repre-
to the anisotropic field distributions arising due to size distribution sented by a combination of an exponential and a power-law region.
and also slow decrease of high field magnetization above the block- The exponential fit (Guiner region) provides information about the
ing temperature. Based on Eq. (1) it may be argued that any change size of the particle through radius of gyration (Rgi ) and the power-
in TB at moderate fields would be result of a change in KV which law fit reveals the dimensionality of the particle or the correlated
in turn may be due to change in interparticle interaction energy, arrangement of the particles depending on the exponent Pi ; Gi and
magnetic anisotropy or change in size or shape. Surprisingly, this Bi denote the pre-factors and  represents the flat background
interesting trend has been observed only in ligand coated magnetic scattering intensity. The picture that emerges from the fit is that
nanoparticles and not in agglomerated particles. Probably, it may there are three structural levels (Table 1). The size of the primary
be due to the fact that in dense or agglomerated particles the inter- particle is approximately 7.0 nm which interacts to form a three
action energy is so large already that this change in effective KV as dimensional mass fractal of aggregate size 34.0 nm. A third struc-
explained above could be insignificant. tural level which is unconnected with the first and second levels
In order to probe this observation further, we performed in- constitutes contribution from a very small population of particles
field SAXS experiments on Fe3 O4 nanoparticles. At this point, it of size approximately 78.0 nm. The aggregate size of the second
should be noted- that while the magnetic data was collected at low structural level varies significantly with the applied magnetic field.
temperature (5–50 K) because the particles are superparamagnetic In the unified fit analysis we modelled the interference peak about
at room temperature, the SAXS measurements which primarily 0.12 Å−1 as due to weak interactions such as that exists between gas
 G. Balaji et al. / Materials Science and Engineering B 177 (2012) 14–18 17

Table 1
Parameters obtained from in SAXS data measured at various indicated fields using unified model.

Sample Level 1a Level 2 Level 3

G Rg B P G Rg B P G Rg B P

0 kG 5.4e−04 25.2 3.2e−07 4 8.8e−03 132 4.3e−08 2.52 0.4 306 1.7e−10 4
0.38 kG 5.0e−04 27.4 2.7e−07 4 0.05 172 6.9e−08 2.68 0.7 321 5.9e−11 4
0.51 kG 8.7e−04 26.3 4.1e−07 4 0.01 126 5.6e−08 2.54 0.6 295 1.9e−10 4
0.70 kG 5.1e−04 25.2 3.0e−07 4 6.3e−03 146 3.1e−08 2.43 0.4 305 1.7e−10 4
1.06 kG 1.1e−03 25.5 5.7e−07 4 12.9e−03 144 2.0e−07 2.08 0.9 316 4.2e−10 4
1.63 kG 7.7e−04 26.5 4.3e−07 4 54.4e−03 172 2.2e−07 2.38 0.8 323 1.1e−10 4
a
= 534; pack = 2.0.

particles. The poor fit at high Q (Fig. 3 inset) shows that the inter- relative change in dimension () which we found to be reversible,
action is stronger. Therefore, the peak was modelled using the is about 4% at 1.0 kG; which is one of the larger values reported so
Percus-Yevick approximation for hard sphere fluids (dark lines in far. At higher magnetic field strength (>1.0 kG), the nanoparticles
Fig. 3) whose structure factor is respond to the external field in an expected manner as confirmed
by low temperature magnetic studies. We propose that these
1
S(Q ) = observations are the result of magnetoelastic coupling between
1 + 24 G(2QRHS , )/2QRHS particle moments in the presence of an externally applied field.
G is a trignometric function, RHS is the hard-sphere radius and However, more experiments are needed to determine the role of
is the volume fraction of the hard spheres. The hard-sphere radius inter-particle interactions (in low external fields) on the magne-
which we found to be approximately 5.2 nm is not sensitive to the tostriction at surface thereby separating the effects of interaction
magnetic field. Two parameters from the fit, interaction radius - the field strength and external field strength. The consequences of
distance between the ‘spherical’ particles, and the volume fraction these processes needed to be taken into account in order to under-
of the hard spheres (Fig. 4) are sensitive to the applied magnetic stand the static and dynamic properties of magnetic nanoparticles.
field. The interaction radius varies as much as 8% from the zero- More detailed studies that could throw light on our present under-
field values. The change in the interaction radius imply a change in standing of the role of the core part of nanoparticle and whether
the size of the aggregates i.e. a 4% change in distance between the giant magnetostriction can be observed in macroscopic scale are in
nanoparticles with external applied magnetic field. progress.
As mentioned earlier, the interparticle interaction effects con-
trol the magnetic properties of nanoparticles in a major way Acknowledgements
including the existence of spin-glass type ordering. At low field
strengths, the interaction effect dominate over external mag- We thank Louisiana Board of Regents for an equipment grant
netic field leading to a magnetoelastic coupling between particle to purchase SQUID magnetometer (LEQSF (2008-10)-ENH-TR-07).
moments. As a result, the ligand coated aggregates of nanoparti- FM thanks CAMD for the opportunity to carry out his thesis work in
cles respond to a magnetostrictive effect resulting in a change in CAMD’s nanomaterials and nanofabrication facility. Partial funding
the size of the aggregates and the interaction distance between the for this work, utilized for synthesis and characterization of iron
particles. In turn, this results in the increase of interparticle interac- oxide nanoparticles, from DOE Energy Frontier Research Center
tion energy and this was the reason for increase of TB (Fig. 2 inset) (Grant no. DE-SC0001058) is gratefully acknowledged.
below 1.0 kG based on Eq. (1) and thus far not accounted in the
literature. Above 1.0 kG, the effect of external field may dominate References
resulting in decrease of TB (inset of Fig. 2). However, the restoration
[1] O. Soderberg, A. Sozinov, Y. Ge, S.-P. Hannula, V.K. Lindroos, Handb. Magn.
of interparticle distance (from SAXS data) can be explained either
Mater. 16 (2006) 1.
by a negative magnetostriction or a complex magnetoelastic cou- [2] M. Hirschner, F. Fischer, T. Reininger, Mod. Trends Magnetostriction 5 (2000)
pling mechanism. Further, investigation is required to elucidate the 307.
[3] S. Zhao, F. Bi, J.G. Wan, M. Han, F. Song, J.M. Liu, G. Wang, Nanotechnology 18
role of interface structures and address the question it whether a
(2007) 265705.
continuous network (agglomerated) of nanoparticles is a requisite [4] J.P. Chen, C.M. Sorenson, K.J. Klabunde, G.C. Hadjipanayis, J. Appl. Phys. 76
for a macroscopic observation. (1994) 6316.
In summary, we demonstrate, for the first time, giant magne- [5] M. Respaud, M. Goiran, J.M. Broto, F. Lionti, L. Thomas, B. Barbara, T. Ould Ely,
C. Amiens, B. Chaudret, Europhys. Lett. 47 (1999) 122.
tostriction in 5.0 nm Fe3 O4 nanoparticles at room temperature. The [6] J.I. Gittleman, B. Abeles, S. Bozoeski, Phys. Rev. B 9 (1974) 3891.
[7] S. Peng, J. Kim, S. Sun, Ann. Rev. Nano Res. 3 (2010) 275.
[8] O. Iglesias, A. Labarta, Phys. Rev. B 63 (2001), 184416/1.
[9] S. Morup, F. Bodker, P.V. Hendricksen, S. Linderoth, Phys. Rev. B 52 (1995) 287.
[10] M. El-Hilo, K. O‘Grady, R.W. Chantrell, J. Magn. Magn. Mater. 114 (1992) 295.
[11] W. Luo, S.R. Nagel, T.F. Rosenbaum, R.E. Rosensweig, Phys. Rev. Lett. 67 (1991)
2721.
[12] W. Kleemann, O. Petracic, Ch. Binek, G.N. Kakazei, Yu.G. Pogorelov, J.B. Sousa,
S. Cardoso, P.P. Freitas, Phys. Rev. B 63 (2001), 134423/1.
[13] N.B. Ekreem, A.G. Olabi, T. Prescott, A. Rafferty, M.S.J. Hashmi, J. Mater. Process.
Technol. 191 (2007) 96.
[14] D.W. Schaefer, J.E. Martin, P. Wiltzius, D.S. Cannell, Phys. Rev. Lett. 52 (1984)
2371–2374.
[15] L. Wang, J. Luo, Q. Fan, M. Suzuki, I. Suzuki, M.H. Engelhard, Y. Lin, N. Kim, J.Q.
Wang, C.J. Zhong, J. Phys. Chem. B 109 (2005) 21593.
[16] http://usaxs.xor.aps.anl.gov/staff/ilavsky/irena.html.
[17] J.M.D Coey, Phys. Rev. Lett. 27 (1971) 1140.
[18] R.H. Kodama, A.E. Berkowitz, E.J. McNiff Jr., S. Foner, Phys. Rev. Lett. 77 (1996)
394.
[19] J.P. Chen, C.M. Sorenson, K.J. Klabunde, G.C. Hadjipanayis, E. Devlin, A. Kostikas,
Fig. 4. The field dependence of interaction radius of Fe3 O4 nanoparticles. Phys. Rev. B 54 (1996) 9288.
18 G. Balaji et al. / Materials Science and Engineering B 177 (2012) 14–18

[20] L. Neel, Ann. Geophys. 5 (1949) 49. [24] A. Guinier, G. Fournet, Small-Angle Scattering of X-rays, John Wiley & Sons,
[21] M. El-Hilo, K. O‘Grady, R.W. Chantrell, J. Magn. Magn. Mater. 114 (1992) 307. New York, 1955.
[22] G.F. Goya, M.P. Morales, J. Metastable Nanocryst. Mater. 20-21 (2004) 673. [25] M. Bonini, E. Fratini, P. Baglioni, Mater. Sci. Eng. C 27 (2007) 1377.
[23] R.K. Zheng, H. Gu, B. Xu, X.X. Zhang, J. Phys. Condens. Matter 18 (2006) 5905. [26] G. Beaucage, J. Appl. Cryst. 28 (1995) 717.