Phase transition from the ferromagnetic to superparamagnetic with a loop shift in 5-nm nickel particles

Article abstract of DOI:10.1016/j.jmmm.2006.02.091

Monodispersed 5-nm nickel particles have been successfully prepared by the microwave-assisted polyol method and confirmed by TEM and HRTEM analysis. Magnetic hysteresis loops, measured at different temperatures from 4.2 to 300 K after the sample was cooled from 300 K with a 20 kOe field, have found to occur at the temperature of a phase transition, T_pt = 100 K, i.e., the temperature at which the ferromagnetic to superparamagnetic transition was observed. It is unexpected for magnetic nanoparticles that the loop shift both zero-field cooling and field cooling occurs above T_pt and it is perfectly symmetrical below the temperature. The large shift observed may suggest the existence of the disfigurement and small cavities of nanoparticles, which cause surface spin disorder or canting above T_pt.

Full text of DOI:10.1016/j.jmmm.2006.02.091

ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 305 (2006) 504–508
www.elsevier.com/locate/jmmm
Phase transition from the ferromagnetic to superparamagnetic with a
loop shift in 5-nm nickel particles
a,b,
Ã
a
a
b
, Kai Huang , Shengqiang Zhou , Peng Zhao , Udi Meridor ,
c
Xiansong Liu
c
Aviad Frydman , Aharon Gedanken
b
a
Anhui Key Laboratory of Information Materials & Devices, School of Physics and Materials Science, Anhui University, Hefei 230093, P R China
b
Department of Chemistry and Kanbar Laboratory for Nanomaterials at the Bar-Ilan University Center for Advanced Materials and Nanotechnology,
Bar-Ilan University, Ramat-Gan 52900, Israel
c
Physics Department and the Minerva Center, Bar-Ilan University, Ramat-Gan 52900, Israel
Received 22 September 2005; received in revised form 16 January 2006
Available online 6 March 2006
Abstract
Monodispersed 5-nm nickel particles have been successfully prepared by the microwave-assisted polyol method and confirmed by
TEM and HRTEM analysis. Magnetic hysteresis loops, measured at different temperatures from 4.2 to 300 K after the sample was
cooled from 300 K with a 20 kOe field, have found to occur at the temperature of a phase transition, Tpt ¼ 100 K, i.e., the temperature at
which the ferromagnetic to superparamagnetic transition was observed. It is unexpected for magnetic nanoparticles that the loop shift
both zero-field cooling and field cooling occurs above Tpt and it is perfectly symmetrical below the temperature. The large shift observed
may suggest the existence of the disfigurement and small cavities of nanoparticles, which cause surface spin disorder or canting above Tpt.
r 2006 Elsevier B.V. All rights reserved.
PACS: 75.20.ꢀg; 75.75.+a; 75.60.ꢀd
Keywords: Magnetic hysteresis loop; Superparamagnetic; Loop shift
1
. Introduction
researches focus on the magnetic nanoparticles related to
the application of ultimate density of magnetic information
storage [6,7]. These nanoparticles rely on the stability of
these units, which own single-domain nanoparticles used
to store a bit. In general, the magnetic moments of all
atoms in every magnetic nanoparticle are ferromagnetically
aligned to form the overall magnetic moment of the
particle. However, with decreasing particle size, the
magnetic anisotropy energy per particle responsible for
holding the magnetic moment along certain directions
becomes comparable to the thermal energy. When this
happens, the single-domain particles become non-magnetic
(superparamagnetic) at the blocking temperature [8,9]. So,
a fundamental question for nanoscale magnetic particles is
to research the phase transition from the ferromagnetic
to the superparamagnetic. For example, as hard drives
become capable of storing more information and accessing
it at faster speeds, their data become more susceptible
to corruption. This data-density barrier is known as the
The synthesis of nanoparticles with controlled size and
composition is of technological and applicable interest,
mainly because these systems provide the basic magnetic
architecture required to create spintronics, such as mag-
netic random access memory devices and spin transistors
[
1,2]. Additionally, many new phenomena and properties
are expected to appear in nanoparticles of the size less
than 10 nm by reason that the particle size is smaller
than the dimension of atomic or ionic diffusion length,
electronic elastic and inelastic mean free path length, and
the magnetic domain size [3–5]. Recently, a few more
Ã
Corresponding author. Anhui Key Laboratory of Information
Materials & Devices, School of Physics & Materials Science,
Anhui University, Hefei 230093, P. R. China. Tel.: +865515107237;
fax: +865515107237.
E-mail address: xiansongliu@yahoo.com.cn (X. Liu).
0304-8853/$ - see front matter r 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2006.02.091

ARTICLE IN PRESS
X. Liu et al. / Journal of Magnetism and Magnetic Materials 305 (2006) 504–508
505
superparamagnetic effect. The original magnetic properties
of nanometer-sized particles are due to the distinct
contributions of surface and step atoms. To disentangle
these problems is an ongoing challenge for materials
science [10,11]. The answer is very important to nanomag-
netic material applications in the sensor, ultrahigh-density
recording, and medicine [12,13].
A conventional monochrome CCD camera, with a resolu-
tion of 768 ꢁ 512 pixels, was used to digitize the images.
The digital images were processed with the Digital
Micrograph software package (Gatan Inc., Pleasanton,
CA, USA). The EMS package16 was used for the HRTEM
image calculations and electron diffraction indexing.
Magnetic measurements at room temperature were
conducted using an Oxford Instrument vibrating sample
magnetometer (VSM). A total of 20–25 mg of nickel
nanoparticles were inserted into the gelatin capsule and
plugged with cotton to prevent dispersal of powder.
Temperature-resolved magnetic measurements were con-
ducted using a Quantum Design MPMS SQUID magnet-
ometer. Specifically, measurements were performed by
cooling a sample in zero field down to 4 K when a magnetic
field of 20 kOe was applied. After that, the sample was
slowly warmed up to a high temperature (typically 300 K)
in steps of a few K per minute with stabilization at each
temperature and subsequent measurement of the magnetic
moment (zero-field cooling, ZFC). Then, without turning
the field off, the sample was cooled down to 4 K while
measuring the magnetic moment at each intermediate
temperature (field cooling, FC).
Nickel nanoparticles have been confirmed experimentally
using several different techniques: ultrasound irradiation
[
[
14], evaporation technique [15], ultrasonic spray pyrolysis
16], chemical reduction [17], electrochemical technique [18],
and polyol method [19]. In the current study, the micro-
wave-assisted polyol method is used to obtain monodis-
persed nickel particles with 5-nm diameters. Although the
conventional polyol method was previously used to prepare
metallic nickel particles [20], the particles of the correspond-
ing reaction under microwave radiation were slightly larger
and rather less uniform than ours. Furthermore, we observe
a new phenomenon that loop shift happened in the
magnetic phase transition from the ferromagnetic to super-
paramagnetic in 5-nm nickel particles.
2
. Experimental section
The starting material was a solution of nickel (II) acetate
tetrahydrate (99.998%, Aldrich Chemical Co.) in ethylene
glycol (499%, Bio Lab Ltd.). Following the previously
described process [21], a 100-ml glass flask was placed in a
microwave oven (Spectra, 900 W) and connected to a
condenser. In view of nanosized nickel agglomeration [22],
so-called surfactants or dispersants with 1, 2, 3, 4, and 5 g
of polyvinylpyrrolidone (PVP, average molecular weight
3. Results and discussion
TEM image of the as-prepared sample is shown in Fig. 1.
It can be seen that nickel particle size and morphology in
Fig. 1(a) are different from that in Fig. 1(b). Although we
can see that the size of the particles in Fig. 1(a) is not more
than 10 nm, its uniformity is not as good as that in
Fig. 1(b). We have now found that it is possible for a
suitable amount of PVP (2 g PVP added in our case) to
stabilize the nanosized cluster produced during the micro-
wave-assisted reduction, in controlling the particle growth,
stabilizing the particles, preventing the agglomeration, and
permitting the isolation of stable nanoparticles. Our
process can narrow the size distribution, in diameter about
5 nanoparticles of nickel, through adding PVP. The nickel
particle can be stable as nickel suspended particles in
ethylene glycol solution exist in a form of a binary electrical
tension colloid. However, the nickel particles experience
strong Van der Waals attractions and magnetic dipole
interactions, which create a challenge to stabilize the
system. A few of particles shown in Fig. 1(b) are connected
together.
As revealed by HRTEM, the nickel particles have a
roughly spherical about 5 nm in diameter, and a face-
centered cubic (FCC) structure (Fig. 2). The image of the
individual nanocrystallite depicts the well-resolved lattice
planes as illustrated in the inset of Fig. 2. The lattice fringe
spacing was measured to be 2.033 A, which is in good
˚
agreement with the reported value of 2.034 A for the
reflecting plane (1 0 0) of the FCC structure (JCPDS:
4-850). The product is therefore identified as FCC crystal-
line nickel.
4
reaction, 50 ml of a solution of 0.100 M Ni(Ac) in ethylene
0 000, Sigma Chemicals) have to be added. Before the
2
glycol was purged by argon for 20 min, then the microwave
oven was turned on at a power level of 50% with a
continued flow of the gas. The reaction was stopped as
soon as the black suspension appeared, and immediately
cooled in iced water. The resulting solid product was
washed thoroughly with ethanol and centrifuged. All these
processes were repeated five times. The nickel particles
were dried under vacuum and kept in a glove box.
The powder X-ray diffraction (XRD) patterns were
collected on a Bruker AXS D* Advance Powder X-ray
diffractometer (using Cu-Ka radiation l ¼ 0:15418 nm).
The elemental composition of the material was analyzed by
energy-dispersive X-ray analysis (EDAX; JEOL-JSM 840
scanning electron microscope). The particle size, morphol-
ogy, and nature of Ni particles were studied by transmis-
sion electron microscopy (TEM) employing a JEOL-JEM
1
00 SX microscope working at 100 kV. After a sonication
bath for 20 min in absolute ethanol, samples for TEM were
prepared by placing a drop of the sample suspension on a
copper grid (400 meshes, Electron Microscopy Sciences)
coated with carbon film. The grid was then air dried.
High-resolution TEM (HRTEM) images were taken
using a JEOL 3010 with 200 kV accelerating voltage.
˚

ARTICLE IN PRESS
X. Liu et al. / Journal of Magnetism and Magnetic Materials 305 (2006) 504–508
506
The XRD patterns of the as-prepared powder showed
the diffraction peaks of FCC crystalline nickel. The
product is therefore identified as nickel particles. The
elemental composition of the as-prepared samples was also
analyzed by EDAX. A quantitative measurement of a
series of the samples was from 92.13 to 95.17 at% of nickel.
Although it is well known that EDAX is not a good
quantitative method for light elements such as oxygen, the
numbers obtained, between 4.83 and 2.03 at%, show the
expected trend of a little amount oxygen on the particle’s
surface.
Hysteresis loops were measured by VSM with 16 kOe on
samples of different particle size. The loops at 300 K are
displayed in Fig. 3. It is worth recalling that the loops shift
and the shift values are different with different sizes
although saturation magnetizations are very small; it
suggests that the materials at 300 K are superparamagnetic.
We define the shift, which is a measure of the shift
anisotropy field as, Hpt ¼ ꢀðHright þ HleftÞ=2, where the
meaning of H
and H is the X value of the loops on
left
right
Y ¼ 0. For the bigger particles of 10 nm (Fig. 1(a)) the loop
shifted less towards the applied field direction, while in the
small particles (Fig. 1(b)) the loop shifted more. The values
of H directed in Fig. 3 are 300, 900, and 2100 Oe for 10, 8,
pt
and 5 nm, respectively. There is a great contrast between
the 5-nm loop and 10-nm loops. The former is nearly an
oblique line and the latter can seem to be thought a
hysteresis loop.
The temperature (T) dependence of the saturation
magnetization ss under an applied magnetic field
(
Happ ¼ 20 kOe) was measured by a SQUID magnetometer
after ZFC and after FC shown in Fig. 4. The phase
transition temperature Tpt is the temperature at which
Fig. 1. TEM images of nickel nanoparticles. (a) Normal microwave-
assisted polyol method. (b) Prepared nickel nanoparticles stabilized by
PVP.
superparamagnetism sets in. Below T , magnetic phase
pt
transition is observed from the superparamagnetic to
ferromagnetic in both FC and ZFC magnetization curves,
whereas above T , they coincide as the saturation
pt
0
0
0
.4
.2
.0
T = 300 K
5
8
1
nm
nm
0 nm
-
-
0.2
0.4
-
15000 -10000 -5000
0
5000 10000 15000
Applied field H (Oe)
Fig. 2. HRTEM images of as-prepared nickel particles. The inset shows
diffractograms of nanoparticle with [1 0 0]FCC crystallographic orienta-
tion.
Fig. 3. Hysteresis loops measured at T ¼ 300 K on samples with different-
sized nanoparticles.

ARTICLE IN PRESS
X. Liu et al. / Journal of Magnetism and Magnetic Materials 305 (2006) 504–508
507
1
1
1
1
1
8
6
4
2
0
8
6
4
2
0
0
0
0
0
.3
.2
.1
.0
T = 300 K
.2
0
ZFC
FC
10
5
0.0
0.2
-
T = 4.2 K
-
15000
0
15000
(
a)
Applied field H (Oe)
0
0
50 100 150 200 250 300
2
000
1500
000
T (K)
ZFC
FC
-5
Tpt
1
5
00
0
Tpt
50
-10
0
100 200 300
Temperature T (K)
(
b)
0
100
150
200
250
Temperature T (K)
-20000
-10000
0
10000
20000
Applied field H (Oe)
Fig. 4. Saturation magnetization of 5-nm nickel particles. The tempera-
ture dependence of the ZFC (square symbols) and FC (H ¼ 20 kOe)
saturation magnetization of 5-nm nickel particles is shown. Inset:
temperature derivative (ꢀdM/dT) of ZFC and FC magnetization.
Fig. 5. Hysteresis loops at 4.2 K of 5-nm nickel particles. Inset: (a) ZFC
and FC loops of superparamagnetic nickel particles measured at 300 K;
(
C
b) coercivity H and loop shift Hls as functions of temperature.
magnetization and coercivity vanish. We get the phase
transition temperature Tpt as about 100 K. It should be
pointed out that MZFC is higher than MFC over the
temperature although it is difficult to distinguish two
curves in the used coordinate. Apparently, Tpt is different
bias effect. The M–H loops are symmetric, showing no sign
of a shift. This is in contrast to other literature reports
relating to Ni/NiO particles with an average particle size
below 30 nm, which have demonstrated a nonsymmetrical
magnetic hysteresis loop [25–27]. It can be understood that
there is no antiferromagnetic layer on the surface of nickel
nanoparticles, although it seems that there exists a little
possibility for oxygen in the as-prepared nanomaterials
according to the previous results from EDAX.
from Curie temperature of nickel, T ¼ 631 K, above
C
which a ferromagnet loses its ferromagnetic ability to
possess a net (spontaneous) magnetization in the absence
of an external magnetic field. According to our results,
which are similar to Curie temperature, magnetic moments
are partially aligned within magnetic domains in ferro-
magnetic nanoparticles at the temperature below the phase
transition. As the phase transition is approached, thermal
fluctuations increasingly destroy this alignment, until the
net magnetization becomes zero at and above the phase
transition.
In the inset (a) of Fig. 5, both ZFC and FC loops of
superparamagnetic nickel particles are measured at 300 K.
It can be obviously seen that the ZFC and FC loop are
unsymmetrical loops, and the ZFC Hpt is larger than
FC H . We can note the change in H and HC with
pt
pt
temperature from 4.2 to 300 K in the inset (b) of Fig. 5.
When the temperature is more than Tpt ¼ 100 K, the 5-nm
nickel particles are superparamagnetic. The loops of ZFC
and FC shift and the values of Hpt increase as temperature
increases. Below the temperature, there are no shifts and
the coercivities increase as temperature decreases. The loop
shift of the superparamagnetic has no relation to the
magnetic cooling condition, i.e., no matter what ZFC or
The temperature derivative of the difference in satura-
tion magnetization (DM) has been calculated: representa-
tive curves (ꢀdM/dT) are shown in an inset of Fig. 4. They
are charaterized by the presence of a peak of both M
ZFC
and MFC curves at a low temperature of 20 K, which is also
confirmed in a nanoscale Fe/Fe oxide granular system [23].
The common feature is consistent with the hypothesized
picture of a frozen, disordered magnetic state at low
temperature and implies that the temperature does not
depend significantly on nanophased materials. According
to this phenomenon, one may think the temperature
derivative of the saturation magnetization reflects the
effective distribution of anisotropy energy barries of the
system in the presence of independent relaxation for
noninterating nanoparticles [24].
MZFC was measured for 5-nm nickel particles while
warming from 4 to 300 K, whereas MFC was recorded
during the subsequent cooling. Fig. 5 compares the ZFC
and FC loops at T ¼ 4:2, showing no sign of the exchange
FC is, the values of H are not dependent on it. Inset (b) of
ls
Fig. 5 shows that H of ZFC is larger than that of FC over
ls
all the temperatures from 100 to 300 K. This response
indicates that the loop shift is a magnetic phenomenon.
In all the discussion, we have not well understood what
cause gives rise to a shift of the hysteresis loop of the
superparamagnetic nanomaterials along the field axis.
However, it can be got rid of the exchange coupling of
the uncompensated interface of the ferromagnetic/antifer-
romagnetic [28]. The large shift observed may suggest the
existence of spin-glass configurations. By analogy with
magnetic oxide nanoparticles [29,30], such configurations
possess different structure and magnetic order between the

ARTICLE IN PRESS
X. Liu et al. / Journal of Magnetism and Magnetic Materials 305 (2006) 504–508
508
surface and interior of 5-nm nickel particles. In this case,
the disfigurement and small cavities of nanoparticles, which
cause surface spin disorder or canting, may be expected to
be responsible for the loop shift of the superparamagnetic
nanomaterials.
[7] S. Rusponi, T. Cren, N. Weiss, M. Epple, P. Buluschek, L. Claude,
H. Brune, Nat. Mater. 2 (2003) 546.
[
8] V. Skumryev, S. Stoyanov, Y. Zhang, G. Hadjipanayis, D. Givord,
J. Nogues, Nature 423 (2003) 850.
9] J. Eisenmenger, I.V. Schuller, Nat. Mater. 2 (2003) 437.
´
[
[
[
10] M. Albrecht, C.T. Rettner, A. Moser, M.E. Best, B.D. Terris, Appl.
Phys. Lett. 81 (2002) 2875.
11] P. Gambardella, S. Rusponi, M. Veronese, S.S. Dhesi, C. Grazioli,
A. Dallmeyer, I. Cabria, R. Zeller, P.H. Dederichs, K. Kern,
C. Carbone, H. Brune, Science 300 (2003) 1130.
4
. Conclusions
In conclusion, we have presented a new method of
[
12] J.I. Martin, J. Nogue
Magn. Mater. 256 (2003) 449.
[13] A.S. Lubbe, C. Alexiou, C. Bergemann, J. Surg. Res. 95 (2001) 200.
´
s, K. Liu, J.L. Vicent, I.K. Schuller, J. Magn.
fabricating monodispersed 5-nm nickel nanoparticles with
a narrow distribution. The present results demonstrate that
the phase transition temperature Tpt from the ferromag-
netic to superparamagnetic is about 100 K. It can be
thought as a magnetic phenomenon that the shifts depend
on the magnetic cooling condition. However, it is obviously
different from the exchange coupling of the uncompen-
sated interface of the ferromagnetic/antiferromagnetic
because the FC loops are symmetrical below Tpt.
[
[
[
14] R.V. Kumar, Y. Koltypin, O. Palchik, A. Gedanken, J. Appl. Polym.
Sci. 86 (2002) 160.
15] Y.D. Yao, Y.Y. Chen, M.F. Tai, D.H. Wang, H.M. Lin, Mater. Sci.
Eng. A 217 (1996) 281.
16] S. Stopic, J. Nedeljkovic, Z. Rakocevic, D. Uskokovic, J. Mater. Res.
14 (1999) 3059.
[
[
[
17] Y.L. Hou, S. Gao, J. Mater. Chem. 13 (2003) 1510.
18] M.P. Zach, R.M. Penner, Adv. Mater. 12 (2000) 878.
19] S.I. Nikitenko, Y. Koltypin, O. Palchik, I. Felner, X.N. Xu,
A. Gedanken, Angew. Chem. Int. Ed. 40 (2001) 4447.
20] O. Palchik, S. Avivi, D. Pinkert, A. Gedanken, Nanostruct. Mater. 11
Acknowledgments
[
(
1999) 415.
21] B. Vaidhyanathan, K.J. Rao, Chem. Mater. 9 (1997) 1196.
[
[
This work was supported by the Key Program of the
Education Department of Anhui Province under Grant no.
22] K.S. Suslick, M.M. Fang, T. Hyeon, J. Am. Chem. Soc. 118 (1996)
1
23] L. Del Bianco, D. Fiorani, A.M. Testa, E. Bonetti, L. Savini,
1960.
2
005KJ033ZD.
[
S. Signoretti, Phys. Rev. B 66 (2002) 174418.
[24] K. O’Grady, R.W. Chantrell, in: J.L. Dormann, D. Fiorani (Eds.),
Magnetic Properties of Fine Particles, North-Holland, Amsterdam,
References
1
992, p. 93.
25] J. Loffler, H. Van Swygenhoven, W. Wagner, J. Meier, B. Doudin,
J.P. Ansermet, Nanostruct. Mater. 9 (1997) 523.
26] J. Loffler, W. Wagner, H. Van Swygenhoven, A. Wiedenmann,
[
[
1] S.H. Sun, C.B. Murray, D. Weller, L. Folks, A. Moser, Science 287
2000) 1989.
[
[
[
¨
(
2] K. Inomata, N. Koike, T. Nozaki, S. Abe, N. Tezuka, Appl. Phys.
Lett. 82 (2003) 2667.
¨
Nanostruct. Mater. 9 (1997) 331.
[
[
[
3] M. Valden, X. Lai, D.W. Goodman, Science 281 (1998) 1647.
4] S. Link, M.A. El-sayed, J. Phys. Chem. B 103 (1999) 8410.
5] H.T. Yang, C.M. Shen, Y.K. Su, T.Z. Yang, H.J. Gao, Y.G. Wang,
Appl. Phys. Lett. 82 (2003) 4729.
27] Y.D. Yao, Y.Y. Chen, C.M. Hsu, H.M. Lin, C.Y. Tung, M.F. Tai,
D.H. Wang, K.T. Wu, C.T. Suo, Nanostruct. Mater. 6 (1995) 933.
28] W.P. Meiklejohn, C.P. Bean, Phys. Rev. 105 (1957) 904.
29] J.M.D. Coey, Phys. Rev. Lett. 27 (1971) 1140.
[
[
[
[
6] M. Bode, O. Pietzsch, A. Kubetzka, R. Wiesendanger, Phys. Rev.
Lett. 92 (2004) 067201.
30] R.H. Kodama, J. Magn. Magn. Mater. 200 (1999) 359.