Surface magnetism, Morin transition, and magnetic dynamics in antiferromagnetic -Fe2O3 (hematite) nanograins R. N. Bhowmik and A. Saravanan Citation: J. Appl. Phys. 107, 053916 (2010); doi: 10.1063/1.3327433 View online: http://dx.doi.org/10.1063/1.3327433 View Table of Contents: http://jap.aip.org/resource/1/JAPIAU/v107/i5 Published by the American Institute of Physics.
Additional information on J. Appl. Phys. Journal Homepage: http://jap.aip.org/ Journal Information: http://jap.aip.org/about/about_the_journal Top downloads: http://jap.aip.org/features/most_downloaded Information for Authors: http://jap.aip.org/authors
Downloaded 10 Nov 2011 to 210.212.230.196. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
JOURNAL OF APPLIED PHYSICS 107, 053916 共2010兲
Surface magnetism, Morin transition, and magnetic dynamics in antiferromagnetic ␣-Fe2O3 „hematite… nanograins R. N. Bhowmika兲 and A. Saravanan Department of Physics, Pondicherry University, R. Venkataraman Nagar, Kalapet, Pondicherry 605014, India
共Received 4 November 2009; accepted 24 January 2010; published online 15 March 2010兲 The grain size of ␣-Fe2O3 decreases to ⬃20 nm by 64 h mechanical milling of the bulk sample. X-ray diffraction pattern suggested identical crystal structure in bulk and mechanical milled samples. Magnetic study 共at temperatures of 100–900 K and fields of 0 – ⫾ 15 kOe兲 showed many interesting features during the decrease in grain size in antiferromagnetic ␣-Fe2O3, e.g., suppression of Morin transition, enhancement in low temperature magnetization, magnetic blocking at high temperature, exchange bias effect, and unusual relaxation of magnetic spin moment. We understand the results in terms of core-shell spin structure of nanograins, where the core part essentially retained the magnetic structure of the bulk sample and the magnetic structure of the shell part is modified due to grain size reduction and surface modification during mechanical milling. Core-shell structure also plays an important role in exhibiting the increasing soft ferromagnetic character in the present hematite samples. The in field magnetic relaxation at room temperature revealed some interesting properties of the magnetic spin ordering in hematite system. © 2010 American Institute of Physics. 关doi:10.1063/1.3327433兴 I. INTRODUCTION
Recently, iron 共Fe兲 and its oxides have drawn remarkable research attention due to their potential applications in nanotechnology.1–6 Iron 共␣-Fe兲 is a typical metallic ferromagnet with paramagnetic to ferromagnetic transition temperature TC ⬃ 770 ° C 共1043 K兲. On the other hand, Fe oxides,1,2 e.g., ␣-Fe2O3 共hematite兲, ␥-Fe2O3 共maghemite兲, and Fe3O4 共magnetite兲, have shown a variety of crystal structure and physical properties. Maghemite and magnetite are stabilized in a typical cubic spinel structure, where Fe spins are distributed in two different lattice sites known as tetrahedral 共A兲 and octahedral 共B兲 sites. Both are ferrimagnet with spin ordering below TN ⬃ 863– 945 K for maghemite2 and below TN ⬃ 850 K for magnetite.3 In contrast, hematite crystallizes in rhombohedral structure and exhibits the properties of an antiferromagnet with TN ⬃ 950 K. It is the unique property of hematite that the rhombohedral 共111兲 planes form the layers of Fe3+ ions in the temperature range of 950–260 K. The planes are separated by layers of oxygen 共O2−兲 ions. The spins of Fe3+ ions in any 共111兲 plane remained parallel, i.e., ferromagnet, below 300 K, but adjacent planes are antiparallel, i.e., antiferromagnet. The canting between 共111兲 plane produces uncompensated magnetic moments of Fe3+ spins between adjacent planes. This is the root for exhibiting weak ferromagnetism 共or canted ferromagnetism兲 in hematite, and the magnetic properties 共e.g., hysteresis loop兲 are many ways resembled to a ferromagnet. As the temperature lowered below 300 K, there is a gradual change in the spin direction from in-共111兲 plane to out of plane. The spin moments are completely turned perpendicular to the 共111兲 plane below a typical temperature TM ⬃ 260 K, known as Morin Electronic mail:
[email protected]. Tel.: ⫹91-9944064547. FAX: ⫹91-413-2655734.
a兲
0021-8979/2010/107共5兲/053916/10/$30.00
temperature, and ␣-Fe2O3 becomes a normal antiferromagnet.7,8 The orderings of Fe3+ spins in-共111兲 plane and out of 共111兲 plane are mainly controlled by the antiferromagnetic 共Fe3+ – O – Fe3+兲 superexchange interactions and anisotropy energy constants.9,10 The study of antiferromagnetic hematite nanoparticles is important for searching a nanomaterial, which shows less interparticle interactions, better stability in atmospheric conditions, sufficiently strong magnetic moment at room temperature 共5B兲, and possible magnetic enhancement by reducing spin canting or antiparallel spin structure of bulk material.11 Zysler et al.12 reported the magnetic properties of chemical routed 5 nm ␣-Fe2O3 nanoparticles dispersed in a polymer matrix. The samples showed high coercivity, high irreversibility field, and shifted hysteresis loops. These experimental results indicated that the magnetic behavior of hematite grains is determined mainly by surface effects, viz., surface anisotropy, exchange anisotropy, core-shell spin structure, and coupling between disordered surface spins. The magnetization loss in bulk hematite below the Morin transition TM ⬃ 260 K occurs due to the magnetic compensation of sublattices. Recently, experiments have shown the decrease in TM with particle size reduction. In chemical routed samples, Morin transition is almost suppressed below the particle size of ⬃20 nm, and ferromagnetic ordering with spontaneous magnetization has been found down to 5 K.10,11,13,14 The decrease in TM is similar to the decrease in superparamagnetic blocking temperature 共TB兲 with the decrease in particle size.15,16 The significant difference is that superparamagnetic blocking occurs due to the relaxation of magnetic particles along local anisotropic axes, whereas Morin transition occurs due to the compensation of two antiferromagnetic sublattices. It may be mentioned that the chemical routed particles are usually covered by chemical
107, 053916-1
© 2010 American Institute of Physics
Downloaded 10 Nov 2011 to 210.212.230.196. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
053916-2
J. Appl. Phys. 107, 053916 共2010兲
R. N. Bhowmik and A. Saravanan
surfactants. Thus, exchange interactions are minimized in chemical routed nanoparticles. This causes most of the surface atoms to be magnetically inactive, and magnetic features of such nanoparticles are manifested mainly at low temperatures 共below 300 K兲. In contrast, surface spins are relatively active in mechanical milled nanoparticles.17 The active role of surface spin configuration in determining the magnetic properties are also realized in mechanical milled ␣-Fe2O3 nanoparticles.18–23 However, the magnetic mechanism in mechanical milled ␣-Fe2O3 nanograins is far from complete understanding, in particular, the effect of core-shell magnetic structure and features of magnetic domains at higher temperatures. In the present work, we prepared nanograined hematite by mechanical milling route and studied the grain size dependent Morin transition, effect of surface magnetism, and magnetism in canted ferromagnetic state. We also incorporated core-shell model for a better understanding of the mechanism of magnetic phenomena in ␣-Fe2O3 nanomaterial. II. EXPERIMENTAL A. Sample preparation
We used a high purity 共⬎99.99%兲 bulk ␣-Fe2O3 material of few micron size grains and 6 g quantity as the starting material for mechanical milling. Milling of the material was carried out up to 64 h in atmospheric condition inside a silicon nitride 共Si3N4兲 container using Fritsch Planetary Mono Mill “Pulverisette 7.” The ball 共mixture of 10 mm agate balls and 5 mm Si3N4 balls兲 to material 共powder form of bulk ␣-Fe2O3兲 mass ratio was maintained at 7:1. The milling process was continued with rotational speed at 300 rpm. The milling process was stopped at 6 h interval for proper mixing and better chemical homogeneity of the milled powder. A small portion of the milled powder was taken out after 14, 24, 36, 48, and 64 h to check the structural phase evolution and grain size reduction of the milled powder. The samples taken out after specific milling hours 共X兲 have been denoted as mhX, i.e., mh0 共bulk sample兲, mh14 共X = 14 h milling兲, mh24 共X = 24 h milling兲, mh36 共X = 36 h milling兲, mh48 共X = 48 h milling兲, and mh64 共X = 64 h milling兲. Each sample was made into a pellet before proceeding for detailed experimental studies. B. Sample characterization and measurements
The x-ray diffraction 共XRD兲 spectrum of ␣-Fe2O3 samples was recorded at room temperature 共300 K兲 using Cu K␣ radiation 共 = 1.540 56 Å兲 from x-ray diffractometer 共model: X-pert Panalytical兲. The spectrum was recorded with 2 ranging from 10° to 90° and a step size of 0.01°. Elemental composition of the samples was determined by energy dispersive analysis of x-ray 共EDX兲 spectrum 共Norton System Six, Thermo Electron Corporation, Instrument Super DRY II, USA兲. Magnetic measurements of the samples were carried out using vibrating sample magnetometer 共VSM兲 共Model: 7404 Lake Shore, USA兲 attached with low temperature and high temperature cryostat and oven, respectively. The temperature dependence magnetization was carried out at 1 kOe
magnetic field and temperature range of 100–900 K. The low temperature cryostat was used for the temperature range of 100–400 K, and a high temperature oven was used for the temperature range of 300–900 K. Different portions of the same pellet of a sample were used for the low and high temperature measurements, respectively. The experimental data were reproducible using two different options in the common temperature range of 300–400 K. In low temperature option, the sample is cooled from 300 to 100 K in the absence of external magnetic field. Then, 1 kOe field was applied and remained constant during the magnetization measurement with temperatures from 100 to 400 K at 5 K interval. We define it zero field cooled 共ZFC兲 mode of magnetization measurement. In the high temperature option, the ZFC magnetic measurement was carried out using another portion of the pellet sample in the presence of 1 kOe 共constant兲 magnetic field when the temperature increases from 300 to 900 K. The magnetization was also recorded by reversing back the temperature from 900 to 300 K. We define this sequence of measurement as field cooled 共FC兲 mode. The field dependence of magnetization was measured at 300 K in the field range of ⫾15 kOe. The time dependence of dc magnetization of the samples was investigated at 300 K in the presence of different constant magnetic fields, ranging from 1 to 8 kOe for 6000 s. Before time dependent measurement, the samples were not used for any other magnetic measurement, and the VSM was degaussed and calibrated. III. RESULTS AND DISCUSSION A. Structural properties
Figure 1 shows the room temperature XRD spectrum of selected samples. The XRD spectrum of milled samples is similar to the pattern of bulk sample before mechanical milling. Structural phase destabilization 共i.e., extra XRD peak兲 was seen in some of the earlier reports on mechanical milled ␣-Fe2O3 samples.21–23 However, the present milled samples are free from any significant additional XRD peaks. The lattice parameters 共a , b , c兲 and unit cell volume 共V兲 of the samples, as shown in Table I, were calculated by Rietveld profile fit using FULLPROF program. The XRD spectrum of crystalline samples was fitted to rhombohedral structure with R3C space group.24 The fitted spectrum is also shown in Fig. 1. The lattice parameter a and c increased with the increase in milling time. Overall, the cell volume expands in the material as an effect of mechanical milling. This could be related to the surface disorder of nanosized grains.18,19 The typical grain size of the bulk sample is a few m. Considering a significant strain induced effect on the XRD peak broadening of mechanical milled samples, we have used the Williamson–Hall equation25—L = 0.89 / 共具d典cos C兲 and G2 = 8共tan2 C兲共rms兲2—to calculate the grain size 具d典 and root mean square lattice strain rms, respectively. Here, is the wavelength of x-ray radiation 共1.540 56 Å兲. L and G are the Lorentzian and Gaussian components of integral width hkl 共defined as the peak area divided by peak height兲 of 共hkl兲 peak 共in degrees兲 with center at 2C 共in degrees兲. The hkl was obtained for six to seven prominent XRD peaks by fitting each peak profile to a pseudo-Voigt function, con-
Downloaded 10 Nov 2011 to 210.212.230.196. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
053916-3
300
J. Appl. Phys. 107, 053916 共2010兲
R. N. Bhowmik and A. Saravanan
(a) bulk
Experimental data Fit data difference between Expt and fit data Bragg position
200 100 0 -100
Intensity (arb. unit)
80
(b) mh24
60 40 20 0 -20 -40 80
(c) mh64
60 40 20 0 -20 -40 10
20
30
40
50 60 2θ (degrees)
70
80
90
FIG. 1. 共Color online兲 XRD spectrum of bulk 共a兲, mh24 共b兲, and mh64 共c兲 samples with profile fitted data.
sisting of Gaussian and Lorentzian components. In the Williamson–Hall method, the Gaussian component was attributed to the contribution of lattice strain rms 共the lattice strain is assumed to be nonuniform in mechanical milled material兲, and the Lorentzian component was attributed to the size 具d典 of grain 共crystallite size兲. The effective Lorentzian 共L兲 and Gaussian 共G兲 components of hkl were obtained from L = hkl − 0 and G = 共hkl2 − 02兲1/2, respectively. 0 is the integral width of standard silicon powder 共 ⬃0.123°兲 taken as the measurement of instrumental broadening. We noted that grain size calculation is not completely isolated from the strain induced effect because L cos C is not totally independent of cos C. However, lattice strain has been calculated in a more accurate manner. The best values of grain size and lattice strain are shown in Table I. We noted
that the grain size decreases monotonically with the increase in milling time, e.g., 具d典 ⬃ 40, 32, 25, 22, and 20 nm for mh14, mh24, mh36, mh48, and mh64 samples, respectively. On the other hand, lattice strain 共rms ⬃ 1.1, 1.6, 2.2, and 2.6 in 10−2 unit for mh14, mh24, mh36, and mh48 samples, respectively兲 increases with milling time up to 48 h and showed a small decrease 共⬃1.9⫻ 10−2兲 for mh64 sample. The decrease in lattice strain at higher milling time was also noted in other work,20 where the decrease in grain size continued with milling time. This means that the microstrain developed at the lattices during mechanical milling may affect the XRD peak broadening up to a certain extent. However, grain size reduction played a major role in determining the peak broadening of the milled samples. The observed phenomenon 共Fig. 1兲 of decreasing XRD peak intensity in the material with the increase in milling time is associated with the decreasing crystalline nature of the material. For example, the intensity height of 104 peak with respect to the bulk sample 共before milling兲 is 31% and 26% for mh24 and mh64 samples, respectively. On the other hand, a systematic XRD peak width broadening with milling time is associated with decreasing grain size, as we noted from calculation. Since the crystalline lattice structure 共rhombohedral with space group R3C兲 of the bulk sample is retained in the mh64 sample where crystallite domain 共grain size兲 is in nanometer scale, the milled samples can be defined in the class of nanocrystalline material. The increasing lattice disorder with milling time can be realized using the concept of core and shell morphology of nanograins. Core part is the coherent crystallite domain 共grain兲 where lattice structure is essentially identical to the bulk sample. The crystalline structure is perturbed in the shell or grain boundary region due to ions vacancy and breaking of coherent crystalline structure.26 In macroscopic scale the elemental composition of iron 共Fe兲 and oxygen 共O兲 with atomic ratio of 2:3 was confirmed from EDX spectrum irrespective of bulk and mechanical milled ␣-Fe2O3 共hematite兲 samples. Viewing the absence of any significant elemental contamination from the container and balls during milling process and any additional peaks from XRD spectrum, we exclude the effect of any impurity phases on the magnetic properties of the present mechanical milled samples. B. Magnetic properties
We have shown the ZFC magnetization 共MZFC兲 in the low temperature 共100–400 K兲 and high temperature 共300–
TABLE I. Grain size 具d典 was calculated from XRD peaks using Williamson–Hall plot. Cell parameters 共a, b, c, and V兲 of the samples were calculated by full profile fit of XRD spectrum using FULLPROF program. Magnetic parameters 共remanent magnetization, MR, and coercive field, HC兲 were calculated from M-H loops at 300 K.
Sample Bulk 共mh0兲 mh14 mh24 mh36 mh48 mh64
具d典 共nm兲
a=b 共Å兲
c 共Å兲
V 共Å兲3
MR 共emu/g兲
HC 共Oe兲
few m 40 32 25 22 20
5.033 49 5.033 64 5.033 68 5.035 94 5.036 02 5.036 40
13.746 28 13.750 21 13.750 70 13.754 80 13.753 80 13.762 00
301.61 301.72 301.73 302.09 302.08 302.31
0.1714 0.0315 0.0389 0.0404 0.0504 0.0475
1615 2244 2023 1593 1489 900
Downloaded 10 Nov 2011 to 210.212.230.196. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
8
/
TM (300 K)
2 kOe
0.10
bulk mh14 mh24 mh36 mh48 mh64
1 kOe
M(100 K) 0.0083 emu/g (1 kOe) 0.0308 emu/g (2 kOe) 0.1026 emu/g (5 kOe) 0.1588 emu/g (8 kOe)
(b) mh24
(a) MZFC at 1 kOe
0.15
0.10
8 kOe
5 kOe 0.05
2 kOe 1 kOe
4
0.00 1.5 (c) mh64
M(emu)
M (T)/M(100 K)
(a) Bulk sample
6
J. Appl. Phys. 107, 053916 共2010兲
R. N. Bhowmik and A. Saravanan
M(emu)
053916-4
5 kOe
TM (270 K)
2
M (emu/g)
8 kOe 0.05
8 kOe
1.0
5 kOe 0.5
100
150
200
250
300
350
T (K)
0.00 100
2.5
150
200
250
T (K)
300
350
2 kOe
0.0 1 koe 100 150
200 250 T (K)
300
350
M(T)/M(120 K)
(b) mh14 sample
2.0
FIG. 3. 共Color online兲 MZFC共T兲 data at 1 kOe for different samples and 共a兲 at 1–8 kOe for the mh24 sample 共b兲 and mh64 sample 共c兲.
1 kOe
M(120 K) 0.0049 emu/g (1 kOe) 0.0391 emu/g (5 kOe) 0.0611 emu/g (8 koe)
5 kOe 8 kOe /
TM (330 K)
1.5 TM (270 K)
1.0 150
200
250 T (K)
300
350
400
FIG. 2. 共Color online兲 MZFC共T兲 data at 1–8 kOe for bulk sample 共a兲 and mh14 共b兲 sample. The MZFC is normalized by MZFC at 120 K. The normalization values M共120 K兲 are shown in the inset of the figures.
900 K兲 regions separately for better representation of the magnetic features. Figure 2共a兲 represents the temperature 共T: 100–350 K兲 dependence of MZFC, normalized by MZFC at 100 K, for the bulk sample 共before milling兲. The MZFC at 100 K 共i.e., in antiferromagnetic state of bulk sample兲 is very small and showed an appreciable increase with the increase in applied magnetic field from 1 to 8 kOe 关inset of Fig. 2共a兲兴. However, MZFC curves of the bulk sample after normalization showed field independence in the antiferromagnetic state from 100 K up to the Morin transition TM ⬃ 270 K. This is followed by a steep increase in MZFC up to TM/ ⬃ 300 K in the canted ferromagnetic 共antiferromagnetic兲 state of the sample and then nearly temperature independent up to 350 K. By increasing the applied magnetic field, the step height 共roughly the difference of magnetization between canted ferromagnetic and antiferromagnetic states兲 decreases. The separation among normalization curves also decreased above 300 K. We shall see later that this feature is consistent with a typical field dependent magnetization growth in the canted ferromagnetic state, where magnetization rapidly increases in the low field region and a slow increase in magnetization at higher field. The magnetic feature of bulk sample at about Morin transition is also continued in the mh14 sample 关Fig. 2共b兲兴. As an effect of decreasing grain size by mechanical milling 共approximately 40 nm兲 in the mh14 sample, one could see that the increase in magnetization above TM ⬃ 270 K becomes sloppy and increases up to TM/ ⬃ 330 K. The widening between TM and TM/ in the mh14 sample can
be treated as an effect of surface disorder in nanosized grains. The position of TM and TM/ does not show any appreciable field dependent behavior in bulk and also in mh14 samples. This means that the 共canted兲 antiferromagnetic order of the bulk sample is still sufficiently strong in the nanograined mh14 sample. Similar grain size effect on the Morin transition was previously reported.13,18 As an effect of further decreasing the grain size using mechanical milling, a different type MZFC共T兲 feature at 1 kOe field is observed 关Fig. 3共a兲兴 for mh24, mh36, mh48, and mh64 samples. For example, Morin transition is not seen down to 100 K. It may be noted that the magnetization of the mh14 sample above 300 K is not much different compared to other samples, but its magnetization below 300 K is lower than any other samples. We noted a systematic increase in magnetization below 300 K on further decreasing the grain size 共mh24, mh36, mh48, and mh64 samples兲. The low temperature increase in magnetization in the present samples with smaller grain size suggests the increasing superparamagnetic/ paramagnetic contribution of uncompensated surface spins in antiferromagnetic grains.27 The magnetization curve of mh24, mh36, mh48, and mh64 samples clearly shows that the low temperature magnetic enhancement in fact occurred in the temperature range that corresponded to the antiferromagnetic state 共below 270 K兲 of the bulk sample 共before mechanical milling兲 and the magnetization decreases rapidly above 270 K that corresponded to the canted ferromagnetic state of bulk sample. Such interesting magnetic evolution with grain size reduction is attributed to the modified magnetic spin ordering at shell or core-shell interface of antiferromagnetic and ferromagnetic nanoparticles.26,27 This means that the core-shell model is well applicable in the present nanomaterial, where magnetic enhancement due to shell spin disorder in the antiferromagnetic state is expected to be opposite to the magnetic reduction in canted ferromagnetic state. The temperature dependent magnetization curves at different fields for the mh24 sample 关Fig. 3共b兲兴 and the mh64 sample 关Fig. 3共c兲兴 give further evidence of magnetic refinement at core-shell interface of nanograined samples, particularly in the antiferromagnetic state of bulk sample below Morin transition. For the mh24 sample 共grains⬃ 32 nm兲, the
Downloaded 10 Nov 2011 to 210.212.230.196. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
053916-5
J. Appl. Phys. 107, 053916 共2010兲
R. N. Bhowmik and A. Saravanan
6 1.5
(a) bulk
5 4
2 1
MZFC
(b) mh24
4 3
1.0
0.5
0.0
2
5
1 4
0
(c) mh64
4 3 2
3 2 1
1 0 300
MFC (emu/g)
Magnetization (emu/g)
0
MZFC (emu/g)
MFC
3
Tm (810 K)
(d) ZFC at 1 kOe bulk mh14 mh24 mh36 mh64
400
500 600 T(K)
700
800
900
0 300
(e) FC at 1 kOe bulk mh14 mh24 mh36 mh64 400
500
600 T(K)
700
800
900
FIG. 4. 共Color online兲 共a兲–共c兲 show the ZFC 共MZFC兲 and FC magnetizations 共MFC兲 at 1 kOe. The arrows indicate the direction of temperature change. The MZFC and MFC for all the samples at 1 kOe are shown in 共d兲 and 共e兲, respectively.
ferromagnetic plateau below 300 K appeared as an effect of field induced magnetic ordering of core-shell interfacial spins and suppression of superparamagnetic type low temperature spin ordering at higher measurement field. The interfacial magnetic refinement between shell and core spins and increasing interactions between grains effectively produce larger magnetic volume. For smaller grains the disorder effect also affects the spin ordering of the core part of the grains. The increasing intergrained interactions in samples with smaller grain size 共but larger effective moment兲 cause the blocking of magnetic particles at lower temperatures,28 and signature is noted for the mh64 sample 共grain size of ⬃20 nm兲. In this case the superparamagnetic type surface effect of magnetic grains is suppressed by the field induced magnetic effects and intergrain interactions.29 It is worthy to mention that we used the pellet form of the nanograined samples 共applying minimum pressure of 1.5 ton/ cm2兲 for the XRD and magnetic measurements. Hence, the magnetic grains or particles we are dealing with are not isolated from each other. Rather, the magnetic grains are in contact with each other, which may increase the intergrain interactions and also affect the magnetic properties up to a certain extent.29 The present results show that grain size dependent magnetic properties have definitely dominated over the minor effects of the contact between grains, and the nanograined samples nowhere showing the signature of regaining the properties of the sample before milling. In fact, the effect of contact between two grains can be taken into account of the modified magnetic interactions at core-shell interface of antiferromagnetic nanograins.26–29 Figure 4 shows the temperature 共300–900 K兲 dependence of ZFC and FC magnetizations at a measurement field of 1 kOe. The magnetization curves for bulk, mh24, and mh64 samples are shown in Figs. 4共a兲–4共c兲, respectively. The interesting feature is that there is a clear splitting between MZFC and MFC curves for all samples. Another common feature is that the plateau type MZFC共T兲 behavior that
was seen from the low temperature measurement extends up to nearly 600 K. The features of MZFC共T兲 curve noted in the present nanograined samples is similar to the characteristic nature observed in hematite single crystal.30 In our samples, MZFC共T兲 curve rapidly increases above 600 K to show a peak at about Tm ⬃ 810 K and then decreases to indicate weak temperature dependence above 870 K. A comparative plot 关Fig. 4共d兲兴 of MZFC共T兲 curves showed no significant change in the MZFC peak position 共Tm ⬃ 810 K兲 with the decrease in grain size by mechanical milling, except the decrease in peak magnetization values. Any general conclusion about the high temperature magnetic properties of hematite nanoparticles cannot be drawn from the literature data because the high temperature range is either limited or the number of reports is very few on hematite particles.24,30–33 The present observations are significantly different in comparison with the high temperature magnetic changes noted in a recent work, where antiferromagnetic ordering temperature decreases in hematite nanowire33 and magnetic moment continuously increases with the decrease in temperature down to 300 K. In the ␣-Fe2O3 nanowire work,33 the measurement was performed at a higher field of 5 kOe. In the nanowire sample, a signature of second transition, which may not be a peak like character at Tm in our samples at relatively lower field of 1 kOe, was noted at a temperature 共⬃738 K兲 below the Néel temperature 共⬃852 K兲 of the sample. The continuous increase in magnetization below Néel temperature and the suppression of peaklike feature in nanowire sample may be the effect of higher applied field 共5 kOe兲. On the other hand, the high temperature MZFC共T兲 character of chemical routed nanoparticles24,32 is well reflected in our mechanical milled samples. A small difference in the shape of magnetization curve can be noted and the Tm value 共⬃940 and ⬃845 K for applied fields of 100 and 200 Oe, respectively兲 is slightly higher in chemical routed samples, which is measured at low magnetic fields. The reported Tm 共⬃940 K at 100 Oe兲 in the chemical routed sample is still lower than the expected paramagnetic to canted ferromagnetic ordering temperature, TN ⬃ 950– 960 K, of the bulk hematite sample.33 This means that the existence of MZFC maximum at Tm is supportive of the reported work and is highly sensitive to the magnitude of measurement field. This fact was verified in the Cr doped ␣-Fe2O3 sample.31 On the other hand, FC magnetization 共MFC兲 curve overlapped on the MZFC curve while temperature was decreasing from 900 K in the presence of a measurement field of 1 kOe. Another unique feature of the samples is that the MFC curve at 1 kOe separates out from MZFC 共i.e., onset of magnetic irreversibility兲 below the irreversibility temperature Tirr ⬃ 860 K. The separation between MFC and MZFC continued to increase to 300 K. It is interesting to note that MFC curve remained lower than MZFC curve between Tirr and Tm and crossed over MZFC below Tm. The MFC curve for the mh24 sample 关Fig. 4共b兲兴, while field cooling measurement was reversed in the temperature direction by heating the sample from 300 to 900 K, again merges to the MZFC curve at Tirr 共⬃860 K兲 and shows thermal hysteresis loop in the canted ferromagnetic state 共300–860 K兲. Such thermal hysteresis behavior was earlier noted in material exhibited exchange
Downloaded 10 Nov 2011 to 210.212.230.196. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
J. Appl. Phys. 107, 053916 共2010兲
R. N. Bhowmik and A. Saravanan 0.6
bulk mh24 mh48
0.6
bulk
0.4
M (emu/g)
bias effect34 共where surface spin disorder played a significant role in determining the properties of magnetic nanoparticles兲. Considering the higher value of Tirr 共⬃860 K兲 than Tm 共⬃810 K兲, we suggest that Tm does not represent a true paramagnetic to canted ferromagnetic ordering temperature 共TN兲 of the samples. Rather, Tm can be identified as the temperature below which magnetic domains are blocked along the local anisotropy axes, where anisotropy energy is much higher than the measurement field.31 The difference 共⬃50 K兲 between Tirr 共860 K兲 and Tm 共810 K兲 is correlated with the distribution of exchange interactions; arising from core-shell spin structure, field cooling effect on the magnetic ordering of interfacial spins and grain size distribution in the samples. Figure 4共e兲 shows that FC magnetization 共MFC兲 also decreases with grain size reduction. The results indicate the lowering of effective magnetization of the material in canted ferromagnetic state on reducing the size of grains. However, the magnetization decrease above 300 K in canted ferromagnetic state, irrespective of MZFC or MFC mode, with the decrease in grain size is a different character in comparison with the low temperature enhancement of magnetization below 300 K. This experimental result is well consistent within the framework of the core-shell model27,28,35 that the low temperature magnetic enhancement in smaller antiferromagnetic grains is mainly controlled by surface magnetism. The surface effect is found to be suppressed at higher temperature as well as during FC measurement. The magnetism observed in the present hematite particles is also drastically different from chemical routed samples, and we assign a main cause to the differences in surface magnetism.17 Next, we used fresh samples to investigate the magnetic field 共H兲 dependence 共0–15 kOe兲 of magnetization 共M兲 at room temperature 共300 K兲. The inset of Fig. 5 shows a typical canted ferromagnetic isotherm for bulk sample 关共also expected above Morin transition at TM ⬃ 270 K 共Ref. 10兲兴, characterized by a rapid increase in magnetization within 5 kOe and a slow increase on further increasing measurement field. The M共H兲 curve of the mh14 sample is also dominated by ferromagnetic ordering at 300 K, but magnetization is significantly reduced in comparison with the bulk sample. The magnetic reduction in the mh14 sample was also observed in the temperature dependence of MZFC 关Fig. 3共a兲兴. The magnetization curve of the mh24 sample is increasing almost linearly with magnetic field without showing the tendency of magnetic saturation within the +15 kOe field. On the other hand, the magnetization curve of mh36, mh48, and mh64 samples increases with applied magnetic field with a tendency of up curvature. The nature of showing up the curvature is more prominent for samples with higher milling time 共e.g., mh64 sample兲. This is another experimental observation of dominating surface magnetism on lowering the size of antiferromagnetic grains.15 Although the maximum magnetic field up to 15 kOe is not sufficient to fully saturate the magnetic moments, the field cyclic process 共from 15 to ⫺15 to 15 kOe兲 showed a magnetic hysteresis loop at room temperature 共300 K兲 for all samples 共main panel of Fig. 5兲 and confirming ferromagnetic contribution in the material. The notable feature is that the MH loop is closed for the bulk
mh64 mh48
0.4
mh14 mh36 mh64
mh36
0.2
0.2
mh24 mh14
0.0 0
M (emu/g)
053916-6
5000 10000 15000 H (Oe)
0.0
-0.2
-0.4
-0.6 -15000
-10000
-5000
H(Oe)
0
5000
10000
15000
FIG. 5. 共Color online兲 Hysteresis loop for different samples. Inset shows the field dependence of magnetization by increasing the field from 0 to 15 kOe.
sample, but the field range of ⫾15 kOe is not enough to achieve a saturated loop for the samples with smaller grain size. This could be attributed to the increasing surface spin disorder in milled samples. We have determined remanent magnetization 共MR兲 and coercivity 共HC兲 from the hysteresis loops. The variation in MR and HC for different samples is shown in Table I. We observed that the MR of the bulk sample 共⬃0.1714 emu/ g兲 drastically reduces 共⬃0.0315 emu/ g兲 in the mh14 sample. Thereafter, MR increases with a further increase in milling time 共⬃0.0389, 0.0404, 0.0504, and 0.0475 emu/g for mh24, mh36, mh48, and mh64 samples, respectively兲. On the other hand, HC initially increases from 1615 Oe of the bulk sample to 2244 Oe for the mh14 sample and followed by a continuous decrease with the increase in milling time 共⬃2023, 1593, 1489, and 900 Oe for mh24, mh36, mh48, and mh64 samples, respectively兲. This results in a sharp decrease in energy product per gram 共⬃MR ⫻ HC兲 from ⬃277 emu Oe for the bulk sample to 71 emu Oe for the mh14 sample, and the decreasing trend continued 共with in between fluctuation兲 on further increase in milling time 共i.e., ⬃79, 64, 75, and 43 for mh24, mh36, mh48, and mh64 samples, respectively兲. Although the applied magnetic field 共maximum of 15 kOe兲 in the present work is not sufficient to obtain complete magnetic saturation and magnetic parameters are the properties of minor loop,36 the magnetic observations 共increasing MR and decreasing HC and energy product also兲 indicated increasing soft ferromagnetic character in the material by decreasing the grain size and can be attributed to the magnetic modifications at the core-shell interface.37,29 The signature of the exchange bias effect37,29 in the present nanograined hematite samples, as an
Downloaded 10 Nov 2011 to 210.212.230.196. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
0.3
J. Appl. Phys. 107, 053916 共2010兲
R. N. Bhowmik and A. Saravanan
(a) mh36 sample
M(emu/g)
053916-7
600
0.1
Hc (Oe)
M (emu/g)
0.2
0.0
0.03 400 0.02 200
-0.1
0.01 5 Hm (KOe)10
-0.2 -5000
0
H (Oe) 5000
15
10000
15000
sample was pre heated 300 K to 900 K in the presence of 1 kOe 1 and the field made to zero at 900 K and the sample cooled to 300 K.
150
-2
(b) mh64 sample
-12000
-6000
0
annealed (red) milled (blue)
(113)
100
-1
(110)
0
(104)
M-H loop for as milled sample (with out any thermal effect)
Intensity (arb.unit)
M (emu/g)
2 Before M-H loop at 300 K, the
50 0 30 2θ (deg.) 35
6000
40
12000
H (Oe)
FIG. 6. 共Color online兲 共a兲 Minor hysteresis loops are shown within different measurement field 共Hm兲 ranges for the mh36 sample. Inset of 共a兲 shows the variation in HC and MR for different Hm values. 共b兲 shows the loop for the mh64 samples in two measurement modes. Inset of 共b兲 compares the XRD data 共selected 2 range indicating three prominent peaks兲 of as milled and after FC measurement of the mh64 sample.
effect of core-shell magnetic structure, is confirmed from a typical minor loop 共S shape兲 at 300 K. The distorted 共S兲 shapes are shown in Fig. 6共a兲 for the mh36 sample when magnetization was measured within different field ranges of Hm = ⫾ 4, ⫾7, ⫾12, and ⫾15 kOe. The minor loop effects shown in the Fig. 6 are not a clear signature of exchange bias, but they could be an indirect evidence of the presence of exchange bias, and the effect is more pronounced for the mh48 and mh64 samples. The calculated HC and MR from the minor loops of the mh36 sample increase with Hm 关inset of Fig. 6共a兲兴. Although the increase in HC and MR is an effect of minor loop property,36 it indicates the field induced magnetic growth at core-shell interface of nanograins. The nonequilibrium magnetism at the core-shell structure is further verified for the mh64 sample by heating the sample from 300 to 900 K in the presence of 1 kOe. Then, the applied field was made to be zero at 900 K, the sample was ZFC to 300 K, and the field dependence of magnetization was recorded at 300 K in the field range of ⫾15 kOe. The M共H兲 loop of the field heated sample is shown in Fig. 6共b兲. The results are highly interesting in the sense that the soft ferromagnetic character in the sample is drastically enhanced with the enlargement of magnetization 共spontaneous magnetization MS ⬃ 2.221 emu/ g and remanent magnetization MR ⬃ 0.4924 emu/ g兲 and the simultaneous reduction in coerciv-
ity 共HC ⬃ 240 Oe兲 in comparison with the values 共MS ⬃ 0.07 emu/ g, MR ⬃ 0.047 emu/ g, and HC ⬃ 900 Oe兲 for the as milled 共without heat treatment兲 mh64 sample 共shown in the main panel of Fig. 5兲. The spontaneous magnetization 共MS兲 was estimated by extrapolating the high field magnetization data 共in the field sequence of +15 kOe to 0 Oe兲 to magnetization axis. From the experimental data, we noted a remanent magnetization shift of ⬃+0.345 emu/ g, spontaneous magnetization of ⬃+1.151 emu/ g, and coercivity shift of ⬃−660 Oe in the field heated samples, comparing the data of as milled samples. This experiment was performed to inform the effect of spin ordering during field heating process, and one should not be confused with the similar magnetic changes reported previously for the FC hysteresis loop38,39 because our sample is cooled from higher temperature to the measurement temperature in the absence of applied magnetic field. To confirm any thermal induced structural changes in the sample, we heated a fresh batch of 64 samples to 950 ° C and kept them for 10 min and then cooled to 300 K before recording the XRD spectrum 共2 = 10° – 90°兲. The XRD spectrum in the inset of Fig. 6共b兲 共showing the spectrum in the selected 2 range of 30° – 43°兲 confirms that the rhombohedral lattice structure with space group R3C is maintained after annealing the sample and no extra phase appeared. This excludes the possibility of any structural phase transformation to be associated in the modified hysteresis loop in the materials after annealing, as assigned for oxidized Co nanoparticles.39 The improved crystalline nature of the material is seen from the increased relative peak height in the annealed sample. For example, the intensity height of 104 peak with respect to the bulk sample 共before milling兲 is 26% and 47% for mh64 and mh64 共annealed兲 samples, respectively. The estimated grain size after annealing is enlarged 共⬃35 nm兲 in comparison with 20 nm before annealing. Noting the enhanced magnetization in our annealed sample, we suggest that the improvement of crystalline nature of the material affects up to a certain extent the modified magnetic parameters32 because the enhanced magnetization is even larger than the bulk sample and the magnetization of the bulk sample is the largest among the studied samples at room temperature. The main contribution of enhanced magnetization is originating from the ordering of core-shell interfacial spin structure during the field heating process to the temperature above magnetic irreversible temperature 共⬃860 K兲, and such an enhancement of magnetization is possible in canted ferromagnetic system.26 When the sample is cooled from 900 to 300 K 共through peak temperature at 810 K兲 in the absence of applied field, the interfacial spins are ordered in a metastable higher magnetic state. This results in the increase in magnetization and decreases coercivity in the hysteresis loop. We also like to mention that the observed M共H兲 loop measured after cooling from 900 K is nearly symmetric about magnetization and field axis. Hence, the observed change in magnetization and coercivity cannot be associated with the exchange bias effect of the mh64 sample but definitely an effect of modified core-shell spin magnetism on nanograined samples, where shell part is highly sensitive to the mode of magnetization measurements.
Downloaded 10 Nov 2011 to 210.212.230.196. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
053916-8
J. Appl. Phys. 107, 053916 共2010兲
R. N. Bhowmik and A. Saravanan
TABLE II. The fit parameters of the relaxation data at 300 K for applied fields of 1, 3, and 5 kOe. 1 kOe
M1
␣ 共10−3兲
M共0兲
M1
␣ 共10−3兲
M共0兲
M1
␣ 共10−4兲
0.0263 0.024 65 0.042 19 0.039 82
0.0423 0.0311 0.0532 0.0619
2.96 1.22 2.10 3.95
0.2631 0.0435 0.0754 0.1252
0.2889 0.0501 0.0828 0.1369
3.775 1.242 1.150 1.984
0.3450 0.0664 0.1331 0.2010
0.344 61 0.067 73 0.135 03 0.203 95
1.70 2.18 3.67 3.55
dency of either saturation or slow increase with increasing time for fieldⱖ 5 kOe. As shown in Fig. 8, the M共t兲 data of the samples at applied fieldⱕ 5 kOe can be best fitted with the equation M共time兲 = M1 − ␣ ln共time兲. Such equation is generally applied for magnetically viscous materials.43,44 In the above equation ␣ is known as magnetic viscosity parameter. It may be noted that M1 is a little bit higher than M共0兲 values. This can be realized from the fact that M1 is obtained mainly by best fitting the long time scale data, where M共0兲 is the starting value of measurement. In the process of fitting, a small deviation of experimental data from the fitted line is not very significant in understanding the overall time dependent spin dynamics of the samples. The fit parameters 共after correction by normalization magnetization M0兲 of different samples at H = 1, 3, and 5 kOe are shown in Table II. This equation, along with the fit parameters, implied that a fraction of magnetic entity 共mainly spin moments兲 is aligned out of the field direction, i.e., the angle between the magnetic spin and magnetic field 共H兲 is increasing with respect to the angle 共0兲 at zero field. In fact, the complex magnetic dynamics related to the spin ordering out of 共111兲 basal plane is 1.00
(a) Bulk 1kOe 3kOe 5kOe 8kOe
0.9
0.8
M(t)/M(0)
1.0
(c) mh48 1kOe 3kOe 5kOe 8kOe
0.90
0.85
0.7
0.6
0.95
0
2000 4000 Time (sec)
6000
0.80 0
2000 4000 Time (sec)
6000
1.00
8 kOe
5 kOe M(t)/M(0)
Now, we try to understand the magnetic spin dynamics of the samples from isothermal in field magnetic relaxation, i.e., time 共t兲 dependence of magnetization 共M兲 at 300 K 共room temperature兲. In this experiment, a fresh sample was used. Before putting the sample in magnetometer, zero field correction of the magnetometer was checked. To avoid the effect of residual field of the electromagnet, if anything still retained in the magnetic pickup coil after proper zero checking, we preferred to apply measurement fieldⱖ 1 kOe. Before starting the magnetization versus time measurement at a constant magnetic field 共say, 2 kOe兲, the sample was waited for 5 min in the presence of constant applied magnetic field. The measurement process was repeated for different applied fields, which were maintained throughout the measurement. Immediate effect is that the initial value of recorded magnetization 关M共0兲兴 depends on the magnitude of applied magnetic field, which could be expected from the magnetic measurement with increasing field. For better representation of the experimental data, we have normalized M共t兲 data by the initial magnetization 关M共0兲兴, and M共0兲 values at selected measurement fields are shown in Table II. The normalized magnetization 关M共t兲/M共0兲兴 data of selected samples in the applied field range of 1–8 kOe are shown in Fig. 7. The in field magnetization of the samples, including bulk antiferromagnetic sample, showed an unconventional decrease with the increase in time. To verify whether it is due to an artifact or a true property of the samples, we checked the same experiment using Ni 共ferromagnetic兲 sample at a very low field of ⬃5 Oe. The in field magnetization of Ni sample slowly increases with time and becomes nearly constant 共within experimental error兲 at higher times. This is the typical character of a ferromagnetic Ni sample. We also checked the quality of M共t兲 data using antiferromagnetic Co3O4 共TN ⬃ 35 K兲 sample that exhibits paramagnetic properties at 300 K. From the quality of M共t兲 data of the tested samples, we believe that the decrease in in field magnetization with time may be within 4%, but the observation can be treated as an important intrinsic property of the present antiferromagnetic samples beyond the experimental error. The nature of magnetic relaxation in the present antiferromagnetic grains seems to be unusual in the sense that magnetization will either increase with time for magnetic disorder in a material below its TC / TN or remain time independent for a typical paramagnet or long range ordered magnet without disorder. Recently, many nonconventional magnetic relaxations in the presence of magnetic field have been found in various magnetic systems.40–42 We noted that the rate of magnetization decay with time reduces at higher fields and magnetization showed the ten-
M(t)/M(0)
Bulk mh14 mh48 mh64
5 kOe
M共0兲
0.95 M(t)/M(0)
Sample
3 kOe
0.90
3 kOe
1.0
0.8
0.85
(b) mh14
1 kOe 0.6
0
2000 4000 Time (sec)
6000
0
1kOe 5kOe 8kOe
(d) mh64
2000 4000 Time(sec)
6000
FIG. 7. 共Color online兲 Time dependence of normalized magnetization 关M共t兲/ M共0兲兴 in the presence of magnetic fields of 1–8 kOe for selected samples.
Downloaded 10 Nov 2011 to 210.212.230.196. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
053916-9
J. Appl. Phys. 107, 053916 共2010兲
R. N. Bhowmik and A. Saravanan
1.0
decreasing magnetic moment兲 is the natural tendency for a perfect antiferromagnet, and one could expect the decrease in Mnoneq with increasing time in a disordered antiferromagnet 共in our case canted ferromagnetic state兲 to achieve the lower value of local equilibrium state, provided the antiferromagnetic exchange interactions must dominate over ferromagnetic interactions and a fraction of spin systems 共we believe it is from the shell兲 has the tendency to align out of the applied field direction when applied field is below a critical value. At higher magnetic field and also during M-H measurement, the local magnetic change is suppressed by the field induced large change of dc magnetization.
M(t)/M(0)
0.9
0.8
(a) 1 kOe bulk mh14 mh48 mh64
0.7 line guides the fit of M(t) data for bulk sample to equation M(t) = M1-αln(time) 0.6 5
6
7 ln(time in seconds)
8
9
IV. CONCLUSIONS
M(t)/M(0)
1.00
(b) 3 kOe Bulk mh14 mh48 mh64
0.95
line guides the fit of M(t) data for mh14 sample to equation M(t) = M1-αln(time)
0.90 5
6
7 ln(time in seconds)
8
9
FIG. 8. 共Color online兲 Variation in M共t兲/M共0兲 vs ln共time兲 data for different samples in the presence of measurement fields of 1 kOe 共a兲 and 3 kOe 共b兲.
expected in canted ferromagnetic state of bulk hematite7,8 and also previously reported in many hematite nanoparticles.11,14,28 The present experimental results may provide the evidence of spin reorientation perpendicular to the basal plane in microscopic level. A mechanism has been previously proposed40 to understand such magnetic phenomenon in antiferromagnetic nanomaterials. A typical canted ferromagnet or antiferromagnet with multimode relaxations may exhibit the increase in magnetization in the macroscopic scale with the increase in magnetic field 共e.g., M-H data兲 at each field value. The typical time for each measurement point is nearly 1 s. During this measurement time the microscopic change in magnetization of the samples is nonvisible and normally neglected for the description of field dependent magnetization process. However, one could easily visualize the nonequilibrium local magnetic state by measuring the magnetization of the sample at constant magnetic field for long measurement time, as carried out in the present work. The magnetization of a sample at constant field is the superposition of two contributions, viz., initial magnetization 共M0兲 and nonequilibrium magnetization 共Mnoneq兲. The Mnoneq is expected to be zero for a material with perfect para/ferro/ antiferromagnetic order 共i.e., absent of disorder兲; otherwise one could expect a finite value of Mnoneq. For disordered ferromagnetic materials, the equilibrium magnetization Mnoneq 共H, T, and t兲 is higher than M0, and one could expect an increase in Mnoneq with time to achieve the local equilibrium state. On the other hand, magnetic compensation 共i.e.,
We applied the novel technique of mechanical milling to produce nanograined ␣-Fe2O3 samples. The experimental results are interesting to understand the surface magnetism of antiferromagnetic materials, having significant spin canting in bulk state. The magnetic features of nanograined samples below 300 K 共essentially in antiferromagnetic state of the bulk sample兲 are drastically different in comparison with the features above 300 K 共essentially in canted ferromagnetic state of the bulk sample兲. Based on experimental observations, we suggest that the magnetic properties of mechanical milled nanograined samples are mainly controlled by surface magnetism, and the properties are well described within the frame work of core-shell spin structure. The Morin transition of bulk sample is not seen in the milled samples with grain size of ⱕ32 nm. The magnetic ordering of the samples strongly depends on the mode of experiments. In comparison with the grain size dependent magnetic blocking phenomenon below room temperature, as reported for chemical routed hematite nanoparticles, the mechanical milled samples exhibited grain size independent magnetic blocking well above room temperature 共Tm ⬃ 810 K兲 during zero cooled process. FC measurement overcomes the magnetic blocking state and samples remained at higher magnetic state at 300 K. We also noted that field heating process significantly improved the magnetic softness with enhanced magnetization in the material. This shows a new route for preparing application oriented magnetic nanoparticles, where magnetic moment can be controlled by MFC and MZFC modes. We conclude that surface magnetism in mechanical milled samples plays an important role in comparison with the chemical routed samples and assigned as the cause for different type magnetic properties. In field magnetic relaxation experiment shows that local magnetic ordering of spins in antiferromagnetic grains at lower measurement field and longer time scale may differ from the macroscopic observation in field dependence of magnetization measurement over a smaller time scale. ACKNOWLEDGMENTS
The author thanks CIF, Pondicherry University, for providing the experimental facilities. The authors also thank Mr. M. Malaidurai for his support during the experiments and data analysis. The financial support from UGC 关No. 33-5/ 2007 共SR兲兴 is also highly acknowledged.
Downloaded 10 Nov 2011 to 210.212.230.196. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
053916-10
R. Zboril, M. Mashlan, and D. Petridis, Chem. Mater. 14, 969 共2002兲. R. C. Handley, Modern Magnetic Materials: Principles and Application 共Wiley, New York, 2000兲. 3 J. Spałek, A. Kozlowski, Z. Tarnawski, Z. Kakol, Y. Fukami, F. Ono, R. Zach, L. J. Spalek, and J. M. Honig, Phys. Rev. B 78, 100401 共2008兲. 4 C. Lang, D. Schüler, and D. Faivre, Macromol. Biosci. 7, 144 共2007兲. 5 A. M. Testa, S. Foglia, L. Suber, D. Fiorani, L. Casas, R. Roig, E. Molins, J. M. Grenèche, and J. Tejada, J. Appl. Phys. 90, 1534 共2001兲. 6 P. Tartaj, M. del P. Morales, S. Veintemillas-Verdaguer, T. GonzálezCarreño, and C. J. Serna, J. Phys. D: Appl. Phys. 36, R182 共2003兲. 7 C. G. Shull, W. A. Strauser, and E. O. Wollan, Phys. Rev. 83, 333 共1951兲. 8 S. Sun, C. B. Murray, D. Weller, L. Folks, and A. Moser, Science 287, 1989 共2000兲. 9 Y. Ishikawa and S. Akimoto, J. Phys. Soc. Jpn. 12, 1083 共1957兲. 10 A. H. Hill, F. Jiao, P. G. Bruce, A. Harrison, W. Kockelmann, and C. Ritter, Chem. Mater. 20, 4891 共2008兲. 11 S. Mørup, D. E. Madsen, C. Frandsen, C. R. H. Bahl, and M. F. Hansen, J. Phys.: Condens. Matter 19, 213202 共2007兲. 12 R. D. Zysler, M. V. Mansilla, and D. Fiorani, Eur. Phys. J. B 41, 171 共2004兲. 13 N. Amin and S. Arajs, Phys. Rev. B 35, 4810 共1987兲. 14 F. Bødker, M. F. Hansen, C. B. Koch, K. Lefmann, and S. Mørup, Phys. Rev. B 61, 6826 共2000兲. 15 L. Neel, in Low Temperature Physics, edited by C. Dewitt, B. Drefus, and P. D. de Gennes 共Gordon and Beach, New York, 1962兲, p. 413. 16 W. F. Brown, Jr., Phys. Rev. 130, 1677 共1963兲. 17 R. N. Bhowmik, R. Ranganathan, R. Nagarajan, B. Ghosh, and S. Kumar, Phys. Rev. B 72, 094405 共2005兲. 18 S. J. Stewart, R. A. Borzi, E. D. Cabanillas, G. Punte, and R. C. Mercader, J. Magn. Magn. Mater. 260, 447 共2003兲. 19 R. A. Borzi, S. J. Stewart, G. Punte, R. C. Mercader, M. VasquezMansilla, R. D. Zysler, and E. D. Cabanillas, J. Magn. Magn. Mater. 205, 234 共1999兲. 20 O. M. Lemine, Superlattices Microstruct. 45, 576 共2009兲. 21 N. Randrianantoandro, A. M. Mercier, M. Hervieu, and J. M. Greneche, Mater. Lett. 47, 150 共2001兲. 22 M. Zdujić, C. Jovalekic, Lj. Karanovic, M. Mitric, D. Poleti, and D. Skala, Mater. Sci. Eng., A 245, 109 共1998兲. 1 2
J. Appl. Phys. 107, 053916 共2010兲
R. N. Bhowmik and A. Saravanan 23
M. Hofmann, S. J. Campbell, W. A. Kaczmarek, and S. Welzel, J. Alloys Comp. 348, 278 共2003兲. 24 R. D. Zysler, M. Vasquez-Mansilla, C. Arciprete, M. Dimitrijewits, D. Rodriguez-Sierra, and C. Saragovi, J. Magn. Magn. Mater. 224, 39 共2001兲. 25 G. K. Williamson and W. H. Hall, Acta Metall. 1, 22 共1953兲. 26 R. H. Kodama and A. E. Berkowitz, Phys. Rev. B 59, 6321 共1999兲. 27 R. N. Bhowmik, R. Nagarajan, and R. Ranganathan, Phys. Rev. B 69, 054430 共2004兲. 28 Ò. Iglesias, A. Labarta, and X. Batlle, J. Nanosci. Nanotechnol. 8, 2761 共2008兲. 29 J. Nogués, V. Skumryev, J. Sort, S. Stoyanov, and D. Givord, Phys. Rev. Lett. 97, 157203 共2006兲. 30 S. T. Lin, Phys. Rev. 116, 1447 共1959兲. 31 R. N. Bhowmik, N. Murty, and S. Srinadhu, PMC Phys. B 1, 20 共2008兲. 32 M. V. Mansilla, R. Zysler, D. Fiorani, and L. Suber, Physica B 320, 206 共2002兲. 33 C. Díaz-Guerra, L. Pérez, J. Piqueras, and M. F. Chioncel, J. Appl. Phys. 106, 104302 共2009兲. 34 V. Skumryev, S. Stoyanov, Y. Zhang, G. Hadjipanayis, D. Givord, and J. Nogués, Nature 共London兲 423, 850 共2003兲. 35 A. Poddar, R. N. Bhowmik, A. De, and P. Sen, J. Magn. Magn. Mater. 321, 2015 共2009兲. 36 J. Geshev, J. Magn. Magn. Mater. 320, 600 共2008兲. 37 J. Nogues, J. Sort, V. Langlais, V. Skumryev, S. Surinach, J. S. Munoz, and M. D. Baro, Phys. Rep. 422, 65 共2005兲. 38 V. Šepelak, D. Baabe, D. Mienert, D. Schultze, F. Krumeich, F. J. Litterst, and K. D. Becker, J. Magn. Magn. Mater. 257, 377 共2003兲. 39 A. N. Dobrynin, D. N. Ievlev, K. Temst, P. Lievens, J. Margueritat, J. Gonzalo, C. N. Afonso, S. Q. Zhou, A. Vantomme, E. Piscopiello, and G. Van Tendeloo, Appl. Phys. Lett. 87, 012501 共2005兲. 40 R. N. Bhowmik and R. Ranganathan, Phys. Rev. B 75, 012410 共2007兲. 41 S. N. Kaul and A. Semawal, Pramana 61, 1129 共2003兲. 42 B. Li, H. L. Alberts, A. M. Strydom, B. M. Wu, A. R. E. Prinsloo, and Zh. J. Chen, J. Magn. Magn. Mater. 321, 61 共2009兲. 43 A. Ray, R. N. Bhowmik, R. Ranganathan, A. Roy, J. Ghose, and S. Chaudhury, J. Magn. Magn. Mater. 223, 39 共2001兲. 44 J. Tejada, R. F. Ziolo, and X. X. Zhang, Chem. Mater. 8, 1784 共1996兲.
Downloaded 10 Nov 2011 to 210.212.230.196. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
