Home
Search
Collections
Journals
About
Contact us
My IOPscience
Interfacial magnetism and exchange coupling in BiFeO3–CuO nanocomposite
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2013 Nanotechnology 24 505711 (http://iopscience.iop.org/0957-4484/24/50/505711) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 137.207.120.173 This content was downloaded on 25/05/2014 at 10:49
Please note that terms and conditions apply.
IOP PUBLISHING
NANOTECHNOLOGY
Nanotechnology 24 (2013) 505711 (11pp)
doi:10.1088/0957-4484/24/50/505711
Interfacial magnetism and exchange coupling in BiFeO3–CuO nanocomposite Kaushik Chakrabarti1 , Babusona Sarkar1 , Vishal Dev Ashok1 , Kajari Das1 , Sheli Sinha Chaudhuri2 and S K De1 1
Department of Materials Science, Indian Association for the Cultivation of Science, Jadavpur, Kolkata-700032, India 2 Department of Electronics and Telecommunication Engineering, Jadavpur University, Jadavpur, Kolkata-700032, India E-mail: [email protected]
Received 15 July 2013, in final form 8 October 2013 Published 27 November 2013 Online at stacks.iop.org/Nano/24/505711 Abstract Ferromagnetic BiFeO3 nanocrystals of average size 9 nm were used to form a composite with antiferromagnetic CuO nanosheets, with the composition (x)BiFeO3 /(100 − x)CuO, x = 0, 20, 40, 50, 60, 80 and 100. The dispersion of BiFeO3 nanocrystals into the CuO matrix was confirmed by x-ray diffraction and transmission electron microscopy. The ferromagnetic ordering as observed in pure BiFeO3 occurs mainly due to the reduction in the particle size as compared to the wavelength (62 nm) of the spiral modulated spin structure of the bulk BiFeO3 . Surface spin disorder of BiFeO3 nanocrystals gives rise to an exponential behavior of magnetization with temperature. Strong magnetic exchange coupling between the BiFeO3 nanocrystal and the CuO matrix induces an interfacial superparamagnetic phase with a blocking temperature of about 80 K. Zero field and field cooled magnetizations are analyzed by a ferromagnetic core and disordered spin shell model. The temperature dependence of the calculated saturation magnetization exhibits three magnetic contributions in three temperature regimes. The BiFeO3 /CuO nanocomposites reveal an exchange bias effect below 170 K. The maximum exchange bias field HEB is 1841 Oe for x = 50 at 5 K under field cooling of 50 kOe. The exchange bias coupling results in an increase of coercivity of 1934 Oe at 5 K. Blocked spins within an interfacial region give rise to a remarkable exchange bias effect in the nanocomposite due to strong magnetic exchange coupling between the BiFeO3 nanocrystals and the CuO nanosheets. S Online supplementary data available from stacks.iop.org/Nano/24/505711/mmedia (Some figures may appear in colour only in the online journal)
1. Introduction
fabrication of various spin based devices such as spin valve and magnetic tunneling junctions [5, 6]. EB is also considered to overcome the problem of the superparamagnetic limit in the world of magnetic storage media to attain higher storage capacity [7]. The EB effect occurs when the FM/AFM system is cooled through the Neel temperature (TN ) of the AFM materials in the presence of an external magnetic field. The AFM spins couple to the FM spins below TN in order to minimize the interface exchange interaction. Due to this exchange interaction a single stable configuration for FM spins is induced, resulting in the appearance of a
The exchange bias (EB) phenomenon at the ferromagnetic/antiferromagnetic (FM/AFM) interface has attracted considerable attention both theoretically as well as experimentally since its discovery in Co/CoO by Meiklejohn and Bean in 1956 [1]. The novel EB phenomenon has stimulated many researchers to carry out extensive studies to understand the basic origin of the complicated interfacial magnetic exchange interaction [2–4]. Technologically, the EB effect has become the cornerstone for the design and 0957-4484/13/505711+11$33.00
1
c 2013 IOP Publishing Ltd Printed in the UK & the USA
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
This triggered us to synthesis BFO nanoparticles in order to make BFO a FM material and combine it with an AFM material for a proper and significant EB effect. The most intriguing and attractive features of both BFO and CuO are multiferroicity and an exchange bias effect in pure phases due to the spiral spin structure. This inspired us to synthesis BFO/CuO nanocomposites with BFO as the FM and CuO as the AFM material to find a synergetic effect of both BFO and CuO components. In this work, we present a detailed study on the temperature and magnetic field dependence of the magnetization and EB effect by varying the FM and AFM ratio in the nanocomposite. The synthesis and incorporation of ultrasmall BFO nanoparticles of average size 9 nm in the nanocomposite gives a high EB value of 1841 Oe at 5 K in 50 kOe field cooling.
unidirectional anisotropy called the exchange anisotropy. This exchange anisotropy manifests in the form of a shift in the hysteresis loop depending on the direction and magnitude of the cooling magnetic field. This shift in the hysteresis loop quantifies the estimation of the exchange field [8]. The EB depends on several factors, such as the volume (thickness) of the ferromagnet, the anisotropy energy of the AFM and the strength of the interface interaction. The interfacial spin structure plays an important role in EB systems consisting of more than one magnetic phase. Magnetic interfaces can be generated in different ways, such as an epitaxial layer, a core–shell structure and the dispersion of one magnetic material into different kinds of host magnetic system. Nanostructured magnetic phases are more interesting due to enhanced proximity effects across the interfaces [9, 10]. Dispersion of nanosized FM(AFM) into the AFM(FM) matrix provides high density of the interface and is the best suited for a large EB effect. Multiple random FM/AFM interfaces and different regions of FM and AFM with competing interactions lead to an interesting magnetic property. The strength of the interfacial interaction depends on the type of magnetic structure of the AFM component and the origin of ferromagnetism in the nanoscale FM component. Among the various AFM and FM materials, cupric oxide (CuO) and bismuth ferrite (BiFeO3 ) are very fascinating magnetic materials due to a rich variety of novel magnetic and electric properties. CuO is an AFM semiconductor with bulk TN = 230 K. The monoclinic crystalline structure of CuO due to Jahn–Teller distortion arising from the d9 orbital makes it more attractive among transition metal monoxides. The complicated AFM spin structure of CuO consists of two-dimensional ferromagnetic layers coupled by a one-dimensional antiferromagnetic chain [11]. Recently, it has been predicted that a weakly frustrated inter-sublattice interaction stabilizes a multiferroic phase of CuO with higher ordering temperatures [12, 13]. CuO nanostructures reveal an EB effect below 40 K due to the large surface to volume ratio [14, 15]. The EB effects of nanostructured CuO are commonly interpreted in terms of a magnetic core–shell structure in which the core is AFM and the shell is a spin glass originating from uncompensated surface spins. The multiferroic oxide BiFeO3 (BFO) is a basic G-type antiferromagnetic material with TN = 643 K. A canted spin structure produces spiral modulation with a wavelength of 62 nm [16–20]. A long-range modulated spiral spin structure does not provide macroscopic magnetization in bulk BFO [21]. Recent studies on nanowires and nanoparticles of BFO reveal a weak ferromagnetism owing to the uncompensated surface spins and partial destruction of the spin spiral [20–25]. Exchange bias has been observed in thin films of BFO comprising ferromagnetic grain boundaries of AFM BFO grain [26]. An unusual EB effect is observed in pure and doped BFO nanocrystals with a spin glass transition of about 50 K and has been explained in terms of a two-dimensional dilute AFM shell and an AFM core [27, 28]. Although there are several reports on the EB effect considering BFO as an AFM material, no report is available for the EB effect based on the ferromagnetic property of BFO.
2. Experiment All the solvents and chemicals were of analytical grade and used without further purification. Composites (x)BiFeO3 /(100 − x)CuO with x = 0, 20, 40, 50, 60, 80, 100 were synthesized via a two-step process: (i) synthesis of the BiFeO3 nanoparticles and (ii) dispersion into the CuO matrix. The precursor solution for synthesizing pure BiFeO3 nanoparticles was prepared by mixing appropriate amounts of Bi(NO3 )3 ·5H2 O and Fe(NO3 )3 ·9H2 O (molar ratio of Bi:Fe = 1:1.2) in a mixed solvent of deionized (DI) water and ethylene diamine (4:1) under constant magnetic stirring for 3 h. An aqueous solution of 0.3 N NaOH was added to the above precursor solution. After that the precursor solution was further stirred for another 2 h and then transferred to a Teflon lined steel chamber of volume 57.5 ml filled to 80% of its volume. The chamber was then closed and placed inside a pre-heated box furnace at 180 ◦ C for 16 h. After the reaction was over, the crystalline brownish powders were collected by centrifugation and thoroughly washed with DI water and ethanol. Finally the sample was vacuum dried at 60 ◦ C for 24 h. Synthesis of the BiFeO3 /CuO composites (x)BiFeO3 / (100 − x)CuO was finally carried out using the synthesized BiFeO3 nanoparticles and Cu(CH3 COO)2 ·H2 O as the starting materials through a hydrothermal method. Appropriate amounts of Cu(CH3 COO)2 ·H2 O and BiFeO3 nanoparticles were first dispersed in DI water based on the molar percentage of the composite. After dispersing for 1.5 h, an aqueous solution of 0.17 N NaOH was added slowly to the above solution. After stirring the solution for another 1 h it was transferred to a 57.5 ml Teflon lined steel chamber filled to 80% of its volume. The chamber was then sealed and kept inside a pre-heated box furnace at 160 ◦ C for 16 h. At the completion of the reaction the crystalline blackish-brown powders were harvested by centrifugation and thorough washing with DI water and ethanol. Finally the washed samples were vacuum dried at 60 ◦ C for 24 h. The crystalline phases of the samples were determined by a high-resolution XPert PRO Panalytical x-ray diffractometer ˚ The microstructural with Cu Kα radiation (λ = 1.54 A). properties of the samples were investigated by transmission 2
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 1. (a) Zero field cooled (ZFC) curve for x = 0 (H = 1 kOe) and (b) ZFC and field cooled (FC) curve for x = 100 (H = 1 kOe). Solid lines are exponentially fitted curves. Inset shows ln(χ − χ0 ) as a function of temperature for the field cooled magnetization.
electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM) using JEOL 2010 instrument. Temperature-dependent magnetization–magnetic field (M–H) measurements for all samples were carried out using a vibrating sample magnetometer (Cryogenic Ltd, UK).
also found in iron oxide and ferrite nanoparticles of diameter less than 10 nm [30–32]. In magnetic nanoparticles, surface spin and surface anisotropy determine magnetic properties of an extremely small size. Monte Carlo simulation suggests that the exponential variation arises purely from surface disordered spin [33]. At low temperature, disordered surface spins align along a local anisotropy axis normal to the surface and lead to an increase in net magnetic moment. The temperature dependent susceptibility χ = M H can be represented by χ (T) = D exp(− TT0 ) + χ0 . The inset of figure 1(b) shows ln(χ − χ0 ) as a function of temperature for field cooled magnetization, indicating a linear behavior. This plot confirms the exponential temperature dependent susceptibility, which has a similarity to the model developed for disorder in canonical spin glass systems [34, 35]. Figure 2 shows the low-temperature (T = 5 K) 50 kOe field cooled magnetization–magnetic field (M–H) curves for x = 0 and 100 respectively. The hysteresis curve (50 kOe field cooled) for x = 0 (pure CuO) in figure 2(a) shows an antiferromagnetic behavior as the loops are not saturated even in an applied magnetic field of 50 kOe. A shift in the negative field axis of around 1365 Oe has been observed at a temperature T = 5 K. The hysteresis loop shift is defined here as exchange bias (EB) and mathematically the EB field is defined as HEB = −(HC1 + HC2 )/2 and the coercivity (HC ) is quantized as (HC1 − HC2 )/2, HC1 and HC2 are the points where the hysteresis loop intersects with the magnetic field axis. The HEB of 1365 Oe may be attributed to the exchange coupling between the uncompensated surface spins at the CuO nanoparticles and spins of the AFM core [36]. The hysteresis curve for x = 100 as in figure 2(b) shows ferromagnetic behavior without any loop shift along the magnetic field axis. This ferromagnetic ordering, as observed, occurs mainly due to the reduction in the particle size as compared to the spiral modulated spin structure of the bulk BFO (62 nm). The saturation magnetization (MS ) is 3.5 emu g−1 at T = 5 K and H = 50 kOe. Earlier systematic studies on the size dependence of magnetization predicted that MS increases with a decrease in the size of BFO nanocrystals [23]. A higher value of MS for 9 nm is consistent with previous reports. The monotonous increase of ZFC and
3. Results and discussion The formation of the nanocomposites (x)BiFeO3 /(100 − x) CuO was confirmed by the room-temperature x-ray diffraction (figure S1 of supplementary information available at stacks.iop.org/Nano/24/505711/mmedia) and TEM (figure S2 available at stacks.iop.org/Nano/24/505711/mmedia) studies. No impurity was found in the XRD patterns of any of the composites. The TEM images of the composites also reveal prominent interfaces between BiFeO3 and CuO. The formation mechanism of these nanocomposites is described in the supplementary information (available at stacks.iop.org/ Nano/24/505711/mmedia). Figures 1(a) and (b) show the variation of magnetization as a function of temperature in the presence of a 1 kOe magnetic field for the samples x = 0 and x = 100 respectively. Figure 1(a) shows the ZFC curve for the x = 0 (pure CuO) sample. The thermal variation of magnetization is found to be similar to that for pure CuO [14, 15]. The increase in magnetization (ZFC) below a temperature of 50 K may be attributed to the presence of paramagnetic centers which arise due to defects and surface layers [29]. The ZFC and FC magnetization curves for x = 100 (BFO) clearly show a bifurcation at 300 K. This bifurcation in the ZFC and FC curves reveals that the ferromagnetic Curie temperature (TC ) of BFO nanocrystals is beyond room temperature. BFO nanocrystals constitute an ideal FM phase for the EB system since their TC is higher than TN of the AFM CuO phase and they fulfil the requirement for achieving a strong EB effect. FC and ZFC magnetizations for x = 100 (BFO) do not attain a saturation value at low temperatures down to 5 K, in contradiction to conventional ferromagnetic magnetization features at low temperature. Both FC and ZFC increase exponentially with decreasing temperature, as indicated by the solid lines in figure 1(b). Such a type of behavior is 3
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 2. Field cooled (50 kOe) M–H curves at T = 5 K for (a) x = 0 and (b) x = 100. The inset shows the magnified M–H curve.
The discontinuity in the crystal structure and the presence of competing magnetic interactions results in a disordered magnetic state at low temperature. The ratio of Tirr to TB is 2.38, which suggests a broad distribution of the blocking temperature [42]. This arises due to a large variation of the interfacial width in the nanocomposite. The magnetic moment of BFO is about one order of magnitude larger than CuO, as evident from figure 1. The magnetic properties of the nanocomposite may be dominated by the FM phase of BFO being perturbed by the presence of CuO on the surface, leading to a system having both bound moments behaving like bulk and free moments behaving like the surface. This may be approximated by a model consisting of the interfacial spins surrounding the FM phase of BFO leading to blocking, resulting in a behavior like SPM. The ZFC and FC curves are analyzed with a view to studying the change in the micro-magnetic structure of the nanocomposite with the variation of BFO concentration. The temperature dependence of ZFC and FC can be modeled as a ferromagnetic core with disordered spins constituting the shell. The shape of the ZFC curves indicates that TB varies with a log normal distribution. Considering both ZFC and FC resulting from the same saturation magnetization MS (T) and an anisotropic constant Ka being independent of temperature, the values of ZFC and FC were fitted to a core–shell model [43] Z MS2 (T)B 1 vm (T) 2 v f (v) dv MZFC (B, T) = 3kB T v 0 Z MS2 (T)B 1 ∞ + vf (v) dv (1) Ka v vm (T) Z M 2 (T)B 1 vm (T) 2 MFC (B, T) = S v f (v) dv 3kB T v 0 Z 30MS2 (T)B 1 ∞ vf (v) dv (2) + Ka v vm (T)
FC magnetization with decreasing temperature, finite value of coercive field, almost saturation of magnetization and enhancement of MS support that BFO nanoparticles possess a pure ferromagnetic phase. Size destroys the spin spiral structure and favors a perfect ferromagnetic alignment of spins. Cation and oxygen vacancies also play important roles in inducing ferromagnetism in BFO nanocrystals [37, 38]. The hysteresis loop is purely symmetric about the origin (H = 0.0) for BFO nanocrystals, as shown in the enlarged scale in the inset of figure 2(b). Magnetic nanoparticles exhibit an EB effect due to the co-existence of different magnetic phases of surface and core spins. A single FM phase in BFO nanoparticles rules out the possibility of the EB phenomenon. Figure 3 shows magnetization as a function of temperature (both ZFC and FC processes) with a cooling field of 1 kOe for x = 20, 40, 50, 60 and 80 samples. All the ZFC and FC magnetization curves show an irreversible behavior at a temperature of Tirr ∼ 190 K, which is much less than TN (230 K) of bulk CuO and TC of BFO nanocrystal. The ZFC curves on the other hand show a broad peak around 80 K, commonly known as the blocking temperature (TB ). Quite different behavior in ZFC and FC in comparison with pure CuO and BFO suggests that the magnetization originates from interfacial interactions. A peak in ZFC is generally observed in spin glass and superparamagnetic (SPM) systems. A differentiation between spin glass and SPM may be done based on the temperature dependence of the FC magnetization. In the case of a spin glass, the FC magnetization tends to saturate or decreases below TB with decreasing temperature [39, 40]. The FC magnetization increases continuously with decreasing temperature for a superparamagnetic phase [39, 41]. Hence the monotonous increase of FC magnetization with decreasing temperature confirms the SPM state of the nanocomposite. In order to further confirm SPM behavior, M has been plotted as a function of H/T for x = 50 (figure 4). All plots are superimposed on a single curve, which suggests the existence of SPM. As the thermal energy is less than the anisotropy energy, the interfacial spins are blocked along the easy axis randomly and disordered spin structures are formed.
where the first term is the ferromagnetic contribution and the second term is the blocking term. vm is the maximum volume contributing to the ferromagnetism given by 30kKBa T . The log 4
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 3. ZFC (lower curve) and FC (upper curve) for (a) x = 20, (b) x = 40, (c) x = 50, (d) x = 60 and (e) x = 80.
normal distribution function f (d) is given as 1 [ln(d) − ln(d0 )]2 exp − f (d) = √ . 2σ 2 2π dσ
Table 1. Parameters of the size distribution (σ ), anisotropy constant (Ka ) and diameter (d0 ) obtained from fitting the ZFC–FC curves of different compositions (x) using equations (1) and (2).
(3)
Good fits to both the ZFC and FC curves are obtained based on the above equations. The size of the ferromagnetic contribution (d0 ), the size distribution (σ ) and the effective anisotropy constant Ka are extracted from the fitting, which is summarized in table 1. The diameter contributing to the ferromagnetic order is found to vary between 3.75 and 4.4 nm. The mean size of the FM region is smaller than the geometrical size (9 nm) of the FM nanocrystals. This indicates that a large fraction of spins constituting the interfacial region are blocked. The size distribution is observed to vary from 1.44 to 1.56. A broad size distribution,
x
σ
Ka (J kg−1 )
d0 (nm)
20 40 50 60 80
1.50 ± 0.006 1.44 ± 0.005 1.57 ± 0.006 1.56 ± 0.006 1.49 ± 0.005
3.86 ± 0.021 3.31 ± 0.016 2.93 ± 0.018 3.22 ± 0.019 3.36 ± 0.018
3.87 ± 0.021 4.44 ± 0.022 3.87 ± 0.022 3.76 ± 0.021 4.11 ± 0.021
σ > 1, is found for all compositions due to the complex interfacial magnetic interaction. There is a steady variation in the effective anisotropic constant, with a minimum at 50% composition. 5
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
characteristic freezing temperature of disordered spins and C is a constant. The fitted curve in figure 5(b) for x = 50 indicates that MS (T) obeys the Bloch law at high temperature (T > 100 K). The exponential law is valid at low temperature, as illustrated in the enlarged scale in the inset of figure 5(b). A clear deviation from Bloch and exponential behavior is found in the temperature interval between 100 and 30 K. A gradual decrease in MS from 100 to 30 K indicates ferrimagnetic features. It seems that ordered and interfacial spins are aligned oppositely to reduce the net magnetic moment. All the fitted parameters obtained from equations (4) and (5) are displayed in table 2. The value of α is smaller than 1.5 for a lower content of BFO, suggesting a deviation from the usual Bloch T 3/2 law. The higher concentration of BFO obeys the Bloch law as the exponent α is very close to 1.5, as indicated table 2. The value of Tf lies between 6 and 10 K. So the contribution of freezing spins to MS (T) is negligible above ∼5Tf , as observed in the behavior of MS (T) above 30 K. The room-temperature saturation magnetization (MS ) increases with increasing BFO content, as shown in figure S3 (supplementary information available at stacks.iop.org/Nano/ 24/505711/mmedia). Figure 6 shows the M–H curves for both zero field cooled and 50 kOe field cooled cases for x = 20, 40, 50, 60 and 80 samples at a temperature of 5 K. M–H curves at an enlarged scale in the low magnetic field region are displayed at the inset of figure 6. The M–H curves for the ZFC process of all the samples show a clear central symmetry about the magnetic field axis. Whereas the curves for the FC process exhibit a negative horizontal shift for all samples. This shift along the magnetic field axis, as discussed earlier, is termed as EB. The occurrence of EB in this case may occur due to enhanced unidirectional anisotropy at the interface of FM BFO nanocrystals embedded in the AFM CuO matrix [3]. Pure CuO reveals the EB effect, which is absent in BFO, as shown in figure 2. EB appears below 50 K for nanostructured CuO due to uncompensated surface spins. The observation of the EB effect at higher temperature (>50 K) and larger values of EB indicate that EB primarily originates from the BFO/CuO interface.
Figure 4. Magnetization (M) as a function of H/T for x = 50.
The resulting temperature dependence of saturation magnetization MS (T) on the concentration of the composite is shown in figure 5(a). All the plots show a monotonic increase in MS (T) with increase in the composition. The thermal natures of all the compositions are similar, with a monotonic increase in the magnetization with decreasing temperature, saturating around 100 K. The temperature dependence of the saturation magnetization generally follows the Bloch T 3/2 law below TC . A sharp increase in MS (T) below about 30 K is observed for all compositions. This additional contribution at low temperature can be ascribed to disordered spins aligned ferromagnetically and varies exponentially, as described earlier. Hence MS (T) has been analyzed separately in two temperature regimes by the following equations: MS (T) = MS (0)[1 − BT α ] T Unbound +C MS (T) = A exp − Tf
(4) (5)
where MS (0) is the saturation magnetization as the temperature tends to zero, the Bloch B constant and the exponent α generally depends on the spin structure and spin dynamics at finite temperature. In equation (5) Tf is the
Figure 5. (a) Temperature dependence saturation magnetization (MS ) for x = 20, 40, 50, 60, 80 and (b) Bloch and exponential fitted MS versus T curve for x = 50, with the inset showing the exponential fitted curve. 6
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 6. Zero field cooled and 50 kOe field cooled M–H curves at T = 5 K for (a) x = 20, (b) x = 40, (c) x = 50, (d) x = 60 and (e) x = 80. Insets show the magnified M–H curves in the low magnetic field region for all samples. Table 2. Parameters of the saturation magnetization at 0 K (MS (0)), Bloch constant (B), exponent (α) and characteristic freezing temperature (Tf ), and constant (C) obtained from fitting MS with temperature using equations (4) and (5) respectively for all compositions (x). x
MS (0) (emu g−1 )
B(10−4 )
α
A
Tf (K)
C (emu g−1 )
20 40 50 60 80
1.01 ± 0.016 0.97 ± 0.008 1.09 ± 0.006 1.28 ± 0.007 1.39 ± 0.007
20.5 ± 7.36 4.95 ± 1.27 0.46 ± 0.164 0.47 ± 0.165 0.72 ± 0.168
0.91 ± 0.058 1.14 ± 0.042 1.52 ± 0.058 1.51 ± 0.058 1.48 ± 0.039
0.21 ± 0.006 0.26 ± 0.006 0.23 ± 0.005 0.27 ± 0.005 0.35 ± 0.007
8.68 ± 0.47 10.26 ± 0.585 9.39 ± 0.417 9.42 ± 0.417 6.56 ± 0.205
0.84 ± 0.002 0.86 ± 0.004 0.99 ± 0.002 1.16 ± 0.002 1.28 ± 0.002
The M–H curves also reveal a difference in the coercivity (HC ) between the ZFC and FC process. HC in the FC process is observed to be greater than HC in the ZFC process. This enhancement in HC for the same sample may be attributed
to the large unidirectional anisotropy of the FM clusters present at the FM/AFM interface [2, 3]. Along with the anomaly in HC , an anomaly along the magnetization axis of the M–H curves is also observed. A vertical shift is also 7
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 7. Variation of exchange bias field (HEB ) and coercivity (HC ) with ferromagnetic (FM) ratio (x) at T = 5 K.
Figure 8. Dependence of HEB with cooling field (Hcool ) for x = 50 at T = 5 K.
observed for all the samples in the FC process as compared to the ZFC process, which seem to be symmetric about the M = 0 axis. This vertical anomaly in the FC process of the M–H curves suggests a strong influence of the pinned magnetization to the exchange fields which may occur due to the moment originating from the uncompensated spins of the CuO matrix [44]. To gain more insight into HEB a detailed study of HEB on the BFO(FM) ratio x, cooling field (Hcool ) and temperature is presented in following sections. Figure 7 shows the composition dependence of HEB and HC (both calculated from the FC process of M–H) on the BFO(FM) ratio (x) at a temperature T = 5 K. It has been observed that the value of HEB for (x)BFO/(100 − x)CuO increases with increasing x, showing a maximum value of 1841 Oe at x = 50, then decreases with further increases of x. The increase in HEB with x until x = 50 may be attributed to enhancement in the BFO/CuO interfacial region and the strong interfacial exchange coupling between BFO and CuO. The decrease in HEB beyond x = 50 may be ascribed to the formation of isolated FM clusters of BFO nanocrystals. The formation of the cluster of BFO nanocrystals reduces the interfacial exchange coupling, resulting in a decrease of HEB . It has been well established that HEB is inversely proportional to the thickness of FM materials (HEB ∝ 1/tFM ) [3]. Beyond x = 50 the FM BFO nanocrystals form clusters, resulting in an increase in the effective thickness of FM, which leads to a decrease of HEB . HC also seems to behave similarly to HEB , initially increasing until x = 50 and then decreasing with further increases in x. In the nanocomposite, the relative phase content of the AFM and FM components plays a vital role in determining the magnitude of the EB effect. Figure 8 shows the influence of Hcool on HEB for the sample x = 50. We observe HEB to be increasing with an increase in Hcool . The increase in the magnitude of Hcool induces more uncompensated spins of the AFM CuO matrix to become aligned along the field direction. These spins lead to a change in the magnetic configuration at the FM/AFM interface [45]. Thus, with increasing Hcool more frozen-in spins are created, ultimately leading to the increase in HEB . The rapid increase of HEB up to about 10 kOe is due to a sharp increase of magnetization in the low magnetic field region,
as evident from the hysteresis displayed in figure 6. A larger number of magnetic moments participate in the EB effect with an increase of the magnetic field. A trend of saturation in HEB at higher magnetic fields is due to saturation magnetization in the M–H curve. Figure 9(a) shows the temperature dependence of HEB for x = 20, 40, 50, 60 and 80. It has been observed that the value of HEB decreases drastically with increasing temperature due to weakening of the exchange interactions in the high-temperature region. Interfacial disorder spins result in an exponential behavior of HEB as a function of temperature [46, 47]. The temperature dependence of HEB has been fitted by an exponential law T 0 HEB = HEB exp − (6) T0 0 is the limiting value of H where HEB EB at T = 0 K and T0 is the characteristic temperature shown in figure 9(b) for 0 and T are shown in x = 50. The best fitted values of HEB 0 0 reveals a maximum table 3. The composition variation of HEB similar to HEB for x = 50. The estimated values of T0 are lower than the blocking temperature, TB , and are about 32 TB . A significant fraction of blocked spins become free with increasing temperature, which causes a rapid decrease of HEB . The value of HEB below 5 K deviates from exponential behavior. Below Tf a large number of disordered spins align ferromagnetically along the direction of the applied magnetic field, which effectively enhances the volume of the FM region. This trend of HEB is similar to the increase of FM thickness which reduces the EB effect [3]. The variation of HC with temperature is shown in figure 10. The value of HC decreases with increasing temperature. An increase of HC in the FC process (HC (FC)) is found compared to the ZFC (HC (ZFC)) data. This enhancement is due to an additional unidirectional anisotropy arising from exchange interactions between FM and AFM spins at the interface [8]. The increment in HC in the field cooled process 1HC = HC (FC) − HC (ZFC) becomes zero above Tirr , as above this temperature spins become unblocked. Hence the temperature dependence of 1HC is fitted by the
8
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 9. Dependence of HEB on temperature for x = 20, 40, 50, 60 and 80 and (b) fitted HEB versus temperature curve for x = 50.
Figure 10. Dependence of (a) HC on temperature for x = 20, 40, 50, 60 and 80 and (b) 1HC versus temperature for x = 50. 0 Table 3. Parameters HEB , T0 and Kex and Tirr obtained from fitting HEB and 1HC with temperature using equations (6) and (7) respectively for all compositions (x).
x
0 HEB (Oe)
T0 (K)
Kex (10−2 ) (J kg−1 )
Tirr (K)
20 40 50 60 80
1801.87 ± 12.513 1965.68 ± 15.496 2323.32 ± 10.869 2162.22 ± 12.883 1819.32 ± 11.234
51.72 ± 1.619 58.21 ± 0.751 47.72 ± 1.316 46.85 ± 0.438 47.16 ± 1.915
3.57 ± 0.065 3.92 ± 0.086 5.01 ± 0.057 4.78 ± 0.108 4.73 ± 0.106
170.01 ± 5.327 177.69 ± 7.001 174.03 ± 5.716 173.86 ± 6.629 170.82 ± 5.26
following formula [48] Kex 1HC = 1− MS (T)
s
T Tirr
responsible for the coercive field is equal to the number of moments participating in the exchange bias effect. A comprehensive temperature dependence of the ZFC and FC investigations indicates that spins are blocked below Tirr . The observation of HEB below Tirr manifests that the blocked spins are responsible for the EB effect. Moreover, theoretical analysis of ZFC and FC shows that the size of the ordered FM region is almost half the actual size of BFO nanocrystals for all compositions. Hence a large fraction, i.e. about 87 ths, of FM spins are involved in the interfacial magnetism. Two distinct magnetic interactions are involved in the BFO/CuO nanocomposite. The bulk AFM of CuO is retained by nearest neighbor superexchange interactions between Cu ions through oxygen atoms. The Cu–O–Cu bond distance and bond angle determine the strength of the exchange interaction. The random orientation
! (7)
where Kex is the exchange anisotropy constant and MS (T) is derived from equations (1) and (2) and shown in figure 10(b) (for x = 50). The calculated parameters Kex and Tirr are listed in table 3. The maximum value of Kex for x = 50 is consistent with the highest value of HEB . The estimated values of Tirr are very close to the experimental values for all compositions, as exhibited in figure 3. The fitted data for Kex suggests that Kex mainly controls the value of HEB . A linear behavior between HEB and HC , as depicted in figure S4 (supplementary information available at stacks.iop.org/ Nano/24/505711/mmedia), implies the number of moments 9
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
(CSIR), Government of India, for providing fellowships. This work is funded by the Council of Scientific and Industrial Research, Government of India, Scheme No: 03(1210)/12/EMR-II.
of surface anisotropy gives rise to disordered spins in BFO nanocrystals. In the presence of an external magnetic field and at a low temperature, disordered spins are aligned to induce a ferromagnetic phase in BFO nanocrystals. Competition between magnetic interactions in the AFM and FM regions and breaking of symmetry at the interface result in the blocking of spins. The interfacial blocked spins exert a microscopic torque on the FM spins to maintain the original alignment and cause a shift in the hysteresis loop to produce the EB effect. AFM may terminate with spin compensated or uncompensated layers due to the complicated spin structure of AFM CuO. In the compensated interface, the net magnetization is zero and does not yield the EB effect. The majority of uncompensated Cu spins take part in the disordered phase. The spins of both the AFM and FM regions are blocked due to the strong interfacial coupling. The uncompensated spins of CuO mediate magnetic coupling between the AFM and FM phases. The magnitude of HEB depends on the number of uncompensated interfacial spins that are pinned to the FM region [49]. Different degrees of coupling among the uncompensated spins of AFM, the disordered spins of FM and bond breaking at the AFM/FM interface give rise to complex magnetism and a large EB effect in the BFO/CuO nanocomposite.
References [1] Meiklejohn W H and Bean C P 1956 Phys. Rev. 102 1413 [2] Berkowitz A E and Takano K 1999 J. Magn. Magn. Mater. 200 552 [3] Nogues J and Schuller I K 1999 J. Magn. Magn. Mater. 192 203 [4] Kiwi M 2001 J. Magn. Magn. Mater. 234 584 [5] Chappert C, Fert A and Van Dau F N 2007 Nature Mater. 6 813 [6] Dieny B, Speriosu V S, Parkin S S P, Gurney B A, Wilhoit D R and Mauri D 1991 Phys. Rev. B 43 1297 [7] Skumryev V, Stoyanov S, Zhang Y, Hadjipanayis G, Givord D and Nogues J 2003 Nature 423 850 [8] Del Bianco L, Fiorani D, Testa A M, Bonetti E and Signorini L 2004 Phys. Rev. B 70 052401 [9] Nogues J, Sort J, Langlais V, Skumryev V, Surinach S, Munoz J S and Barob M D 2005 Phys. Rep. 422 65 [10] Iglesias O, Labrta A and Batlle X 2008 J. Nanosci. Nanotechnol. 8 2761 [11] Filippetti A and Fiorentini V 2005 Phys. Rev. Lett. 95 086405 [12] Kimura T, Sekio Y, Nakamura H, Siegrist T and Ramirez A P 2008 Nature 7 291 [13] Jin G, Cao K, Guo G C and He L 2012 Phys. Rev. Lett. 108 187205 [14] D´ıaz-Guerra C, Vila M and Piqueras J 2010 Appl. Phys. Lett. 96 193105 [15] Punnoose A, Magnone H, Seehra M S and Bonevich J 2001 Phys. Rev. B 64 174420 [16] Eerenstein W, Mathur N D and Scott J F 2006 Nature 442 759 [17] Cheong S W and Mostovoy M 2007 Nature Mater. 6 13 [18] Ramesh R and Spaldin N A 2007 Nature Mater. 6 21 [19] Singh A, Pandey V, Kotnala R K and Pandey D 2008 Phys. Rev. Lett. 101 247602 [20] Wang J et al 2003 Science 299 1719 [21] Ederer C and Spaldin N A 2005 Phys. Rev. B 71 060401 [22] Gao F, Yuan Y, Wang K F, Chen X Y, Chen F, Liu J M and Ren Z F 2006 Appl. Phys. Lett. 89 102506 [23] Park T J, Papaefthymiou G C, Viescas A J, Moodenbaugh A R and Wong S S 2007 Nano Lett. 7 766 [24] Chakrabarti K, Das K, Sarkar B and De S K 2011 J. Appl. Phys. 110 103905 [25] Chakrabarti K, Das K, Sarkar B, De S K, Sinha G and Lahtinen J 2012 Appl. Phys. Lett. 101 042401 [26] Sung K D, Park Y A, Seo M S, Jo Y, Hur N and Jung J H 2012 J. Appl. Phys. 111 033915 [27] Manna P K, Yusuf S M, Shukla R and Tyagi A K 2011 Phys. Rev. B 83 184412 [28] Dong S, Yao Y, Hou Y, Liu Y, Tang Y and Li X 2011 Nanotechnology 22 385701 [29] Chandrasekhar T V and Sahni V C 1994 J. Phys.: Condens. Matter 6 L423 [30] Kasyutich O, Desautels R D, Southern B W and van Lierop J 2010 Phys. Rev. Lett. 104 127205 [31] Shendruk T N, Desautels R D, Southern B W and van Lierop J 2007 Nanotechnology 18 455704 [32] Aquino R, Depeyrot J, Sousa M H, Tourinho F A, Dubois E and Perzynski R 2005 Phys. Rev. B 72 184435 [33] Kachkachi H, Ezzir A, Nogues M and Tronc E 2000 Eur. Phys. J. B 14 681 [34] Jonsson P E, Mathieu R, Wernsdorfer W, Tkachuk A M and Barbara B 2007 Phys. Rev. Lett. 98 256403
4. Conclusion The temperature dependence of the magnetization of pure BFO nanocrystals of size 9 nm follows an exponential decay with increasing temperature due to free spins. Upon dispersion of BFO into the CuO matrix, a significant change in the magnetic behavior is seen in comparison with the pure BFO and CuO phases. Strong interfacial exchange coupling and structural mismatch lead to spin order, spin disorder and a ferrimagnetic state. The radius for the effective FM ordered region is less than the actual particle size determined using TEM analysis, which certainly establishes that a large fraction of FM spins are involved in the interfacial magnetism. The broad distribution of blocking temperature indicates the presence of several types of FM/AFM interfacial regions with different anisotropy due to the large size distribution of the FM region. The temperature dependence of saturation magnetization follows the Bloch law at high temperature and the exponential law at low temperature. Maximum values of the exchange field of 1841 Oe and the coercivity of 1934 Oe are obtained for an equal content of BFO and CuO in the nanocomposite. Strong exchange coupling between the uncompensated surface spins of CuO and the disordered spins of BFO nanocrystals, as well as the discontinuity in crystal structure at the interface, result in interesting magnetic properties and a significant exchange bias field in the BFO/CuO nanocomposite. The large exchange bias effect and multiferroicity in BFO and CuO make the nanocomposite more attractive for multifunctional devices.
Acknowledgments Kaushik Chakrabarti, Babusona Sarkar and Kajari Das are grateful to the Council of Scientific and Industrial Research 10
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
[35] Barbara B 2007 Phys. Rev. Lett. 99 177201 [36] Liu K L, Yuan S L, Duan H N, Zheng X F, Yin S Y, Tian Z M, Wang C H and Huo S X 2010 J. Appl. Phys. 107 023911 [37] Paudel T R, Jaswal S S and Tsymbal E Y 2012 Phys. Rev. B 85 104409 [38] Ederer C and Spaldin N A 2005 Phys. Rev. B 71 224103 [39] Gruyters M 2005 Phys. Rev. Lett. 95 077204 [40] Alonso J, Fdez-Gubieda M L, Barandiaran J M, Svalov A, Fern´andez Barqu´ın L, Alba Venero D and Orue I 2010 Phys. Rev. B 82 054406 [41] Batlle X and Labarta A 2002 J. Phys. D: Appl. Phys. 35 R12 [42] Domingo N, Fiorani D, Testa A M, Binns C, Baker S and Tejada J 2008 J. Phys. D: Appl. Phys. 41 134009 [43] Respaud M et al 1998 Phys. Rev. B 57 2925
[44] Yin S Y, Yuan S L, Tian Z M, Liu L, Wang C H, Zheng X F, Duan H N and Huo S X 2010 J. Appl. Phys. 107 043909 [45] Salabas E L, Rumplecker A, Kleitz F, Radu F and Schuth F 2006 Nano Lett. 6 2977 [46] Ding J F, Lebedev O I, Turner S, Tian Y F, Hu W J, Seo J W, Panagopoulos C, Prellier W, Van Tendeloo G and Wu T 2013 Phys. Rev. B 87 054428 [47] Huang X H, Ding J F, Zhang G Q, Hou Y, Yao Y P and Li X G 2008 Phys. Rev. B 78 224408 [48] Jacobs I S and Bean C P 1963 Magnetism vol 3, ed G T Rado and H Suhl (New York: Academic) p 273 [49] Ohldag H, Scholl A, Nolting F, Arenholz E, Maat S, Young A T, Carey M and Stohr J 2003 Phys. Rev. Lett. 91 017203
11
Search
Collections
Journals
About
Contact us
My IOPscience
Interfacial magnetism and exchange coupling in BiFeO3–CuO nanocomposite
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2013 Nanotechnology 24 505711 (http://iopscience.iop.org/0957-4484/24/50/505711) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 137.207.120.173 This content was downloaded on 25/05/2014 at 10:49
Please note that terms and conditions apply.
IOP PUBLISHING
NANOTECHNOLOGY
Nanotechnology 24 (2013) 505711 (11pp)
doi:10.1088/0957-4484/24/50/505711
Interfacial magnetism and exchange coupling in BiFeO3–CuO nanocomposite Kaushik Chakrabarti1 , Babusona Sarkar1 , Vishal Dev Ashok1 , Kajari Das1 , Sheli Sinha Chaudhuri2 and S K De1 1
Department of Materials Science, Indian Association for the Cultivation of Science, Jadavpur, Kolkata-700032, India 2 Department of Electronics and Telecommunication Engineering, Jadavpur University, Jadavpur, Kolkata-700032, India E-mail: [email protected]
Received 15 July 2013, in final form 8 October 2013 Published 27 November 2013 Online at stacks.iop.org/Nano/24/505711 Abstract Ferromagnetic BiFeO3 nanocrystals of average size 9 nm were used to form a composite with antiferromagnetic CuO nanosheets, with the composition (x)BiFeO3 /(100 − x)CuO, x = 0, 20, 40, 50, 60, 80 and 100. The dispersion of BiFeO3 nanocrystals into the CuO matrix was confirmed by x-ray diffraction and transmission electron microscopy. The ferromagnetic ordering as observed in pure BiFeO3 occurs mainly due to the reduction in the particle size as compared to the wavelength (62 nm) of the spiral modulated spin structure of the bulk BiFeO3 . Surface spin disorder of BiFeO3 nanocrystals gives rise to an exponential behavior of magnetization with temperature. Strong magnetic exchange coupling between the BiFeO3 nanocrystal and the CuO matrix induces an interfacial superparamagnetic phase with a blocking temperature of about 80 K. Zero field and field cooled magnetizations are analyzed by a ferromagnetic core and disordered spin shell model. The temperature dependence of the calculated saturation magnetization exhibits three magnetic contributions in three temperature regimes. The BiFeO3 /CuO nanocomposites reveal an exchange bias effect below 170 K. The maximum exchange bias field HEB is 1841 Oe for x = 50 at 5 K under field cooling of 50 kOe. The exchange bias coupling results in an increase of coercivity of 1934 Oe at 5 K. Blocked spins within an interfacial region give rise to a remarkable exchange bias effect in the nanocomposite due to strong magnetic exchange coupling between the BiFeO3 nanocrystals and the CuO nanosheets. S Online supplementary data available from stacks.iop.org/Nano/24/505711/mmedia (Some figures may appear in colour only in the online journal)
1. Introduction
fabrication of various spin based devices such as spin valve and magnetic tunneling junctions [5, 6]. EB is also considered to overcome the problem of the superparamagnetic limit in the world of magnetic storage media to attain higher storage capacity [7]. The EB effect occurs when the FM/AFM system is cooled through the Neel temperature (TN ) of the AFM materials in the presence of an external magnetic field. The AFM spins couple to the FM spins below TN in order to minimize the interface exchange interaction. Due to this exchange interaction a single stable configuration for FM spins is induced, resulting in the appearance of a
The exchange bias (EB) phenomenon at the ferromagnetic/antiferromagnetic (FM/AFM) interface has attracted considerable attention both theoretically as well as experimentally since its discovery in Co/CoO by Meiklejohn and Bean in 1956 [1]. The novel EB phenomenon has stimulated many researchers to carry out extensive studies to understand the basic origin of the complicated interfacial magnetic exchange interaction [2–4]. Technologically, the EB effect has become the cornerstone for the design and 0957-4484/13/505711+11$33.00
1
c 2013 IOP Publishing Ltd Printed in the UK & the USA
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
This triggered us to synthesis BFO nanoparticles in order to make BFO a FM material and combine it with an AFM material for a proper and significant EB effect. The most intriguing and attractive features of both BFO and CuO are multiferroicity and an exchange bias effect in pure phases due to the spiral spin structure. This inspired us to synthesis BFO/CuO nanocomposites with BFO as the FM and CuO as the AFM material to find a synergetic effect of both BFO and CuO components. In this work, we present a detailed study on the temperature and magnetic field dependence of the magnetization and EB effect by varying the FM and AFM ratio in the nanocomposite. The synthesis and incorporation of ultrasmall BFO nanoparticles of average size 9 nm in the nanocomposite gives a high EB value of 1841 Oe at 5 K in 50 kOe field cooling.
unidirectional anisotropy called the exchange anisotropy. This exchange anisotropy manifests in the form of a shift in the hysteresis loop depending on the direction and magnitude of the cooling magnetic field. This shift in the hysteresis loop quantifies the estimation of the exchange field [8]. The EB depends on several factors, such as the volume (thickness) of the ferromagnet, the anisotropy energy of the AFM and the strength of the interface interaction. The interfacial spin structure plays an important role in EB systems consisting of more than one magnetic phase. Magnetic interfaces can be generated in different ways, such as an epitaxial layer, a core–shell structure and the dispersion of one magnetic material into different kinds of host magnetic system. Nanostructured magnetic phases are more interesting due to enhanced proximity effects across the interfaces [9, 10]. Dispersion of nanosized FM(AFM) into the AFM(FM) matrix provides high density of the interface and is the best suited for a large EB effect. Multiple random FM/AFM interfaces and different regions of FM and AFM with competing interactions lead to an interesting magnetic property. The strength of the interfacial interaction depends on the type of magnetic structure of the AFM component and the origin of ferromagnetism in the nanoscale FM component. Among the various AFM and FM materials, cupric oxide (CuO) and bismuth ferrite (BiFeO3 ) are very fascinating magnetic materials due to a rich variety of novel magnetic and electric properties. CuO is an AFM semiconductor with bulk TN = 230 K. The monoclinic crystalline structure of CuO due to Jahn–Teller distortion arising from the d9 orbital makes it more attractive among transition metal monoxides. The complicated AFM spin structure of CuO consists of two-dimensional ferromagnetic layers coupled by a one-dimensional antiferromagnetic chain [11]. Recently, it has been predicted that a weakly frustrated inter-sublattice interaction stabilizes a multiferroic phase of CuO with higher ordering temperatures [12, 13]. CuO nanostructures reveal an EB effect below 40 K due to the large surface to volume ratio [14, 15]. The EB effects of nanostructured CuO are commonly interpreted in terms of a magnetic core–shell structure in which the core is AFM and the shell is a spin glass originating from uncompensated surface spins. The multiferroic oxide BiFeO3 (BFO) is a basic G-type antiferromagnetic material with TN = 643 K. A canted spin structure produces spiral modulation with a wavelength of 62 nm [16–20]. A long-range modulated spiral spin structure does not provide macroscopic magnetization in bulk BFO [21]. Recent studies on nanowires and nanoparticles of BFO reveal a weak ferromagnetism owing to the uncompensated surface spins and partial destruction of the spin spiral [20–25]. Exchange bias has been observed in thin films of BFO comprising ferromagnetic grain boundaries of AFM BFO grain [26]. An unusual EB effect is observed in pure and doped BFO nanocrystals with a spin glass transition of about 50 K and has been explained in terms of a two-dimensional dilute AFM shell and an AFM core [27, 28]. Although there are several reports on the EB effect considering BFO as an AFM material, no report is available for the EB effect based on the ferromagnetic property of BFO.
2. Experiment All the solvents and chemicals were of analytical grade and used without further purification. Composites (x)BiFeO3 /(100 − x)CuO with x = 0, 20, 40, 50, 60, 80, 100 were synthesized via a two-step process: (i) synthesis of the BiFeO3 nanoparticles and (ii) dispersion into the CuO matrix. The precursor solution for synthesizing pure BiFeO3 nanoparticles was prepared by mixing appropriate amounts of Bi(NO3 )3 ·5H2 O and Fe(NO3 )3 ·9H2 O (molar ratio of Bi:Fe = 1:1.2) in a mixed solvent of deionized (DI) water and ethylene diamine (4:1) under constant magnetic stirring for 3 h. An aqueous solution of 0.3 N NaOH was added to the above precursor solution. After that the precursor solution was further stirred for another 2 h and then transferred to a Teflon lined steel chamber of volume 57.5 ml filled to 80% of its volume. The chamber was then closed and placed inside a pre-heated box furnace at 180 ◦ C for 16 h. After the reaction was over, the crystalline brownish powders were collected by centrifugation and thoroughly washed with DI water and ethanol. Finally the sample was vacuum dried at 60 ◦ C for 24 h. Synthesis of the BiFeO3 /CuO composites (x)BiFeO3 / (100 − x)CuO was finally carried out using the synthesized BiFeO3 nanoparticles and Cu(CH3 COO)2 ·H2 O as the starting materials through a hydrothermal method. Appropriate amounts of Cu(CH3 COO)2 ·H2 O and BiFeO3 nanoparticles were first dispersed in DI water based on the molar percentage of the composite. After dispersing for 1.5 h, an aqueous solution of 0.17 N NaOH was added slowly to the above solution. After stirring the solution for another 1 h it was transferred to a 57.5 ml Teflon lined steel chamber filled to 80% of its volume. The chamber was then sealed and kept inside a pre-heated box furnace at 160 ◦ C for 16 h. At the completion of the reaction the crystalline blackish-brown powders were harvested by centrifugation and thorough washing with DI water and ethanol. Finally the washed samples were vacuum dried at 60 ◦ C for 24 h. The crystalline phases of the samples were determined by a high-resolution XPert PRO Panalytical x-ray diffractometer ˚ The microstructural with Cu Kα radiation (λ = 1.54 A). properties of the samples were investigated by transmission 2
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 1. (a) Zero field cooled (ZFC) curve for x = 0 (H = 1 kOe) and (b) ZFC and field cooled (FC) curve for x = 100 (H = 1 kOe). Solid lines are exponentially fitted curves. Inset shows ln(χ − χ0 ) as a function of temperature for the field cooled magnetization.
electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM) using JEOL 2010 instrument. Temperature-dependent magnetization–magnetic field (M–H) measurements for all samples were carried out using a vibrating sample magnetometer (Cryogenic Ltd, UK).
also found in iron oxide and ferrite nanoparticles of diameter less than 10 nm [30–32]. In magnetic nanoparticles, surface spin and surface anisotropy determine magnetic properties of an extremely small size. Monte Carlo simulation suggests that the exponential variation arises purely from surface disordered spin [33]. At low temperature, disordered surface spins align along a local anisotropy axis normal to the surface and lead to an increase in net magnetic moment. The temperature dependent susceptibility χ = M H can be represented by χ (T) = D exp(− TT0 ) + χ0 . The inset of figure 1(b) shows ln(χ − χ0 ) as a function of temperature for field cooled magnetization, indicating a linear behavior. This plot confirms the exponential temperature dependent susceptibility, which has a similarity to the model developed for disorder in canonical spin glass systems [34, 35]. Figure 2 shows the low-temperature (T = 5 K) 50 kOe field cooled magnetization–magnetic field (M–H) curves for x = 0 and 100 respectively. The hysteresis curve (50 kOe field cooled) for x = 0 (pure CuO) in figure 2(a) shows an antiferromagnetic behavior as the loops are not saturated even in an applied magnetic field of 50 kOe. A shift in the negative field axis of around 1365 Oe has been observed at a temperature T = 5 K. The hysteresis loop shift is defined here as exchange bias (EB) and mathematically the EB field is defined as HEB = −(HC1 + HC2 )/2 and the coercivity (HC ) is quantized as (HC1 − HC2 )/2, HC1 and HC2 are the points where the hysteresis loop intersects with the magnetic field axis. The HEB of 1365 Oe may be attributed to the exchange coupling between the uncompensated surface spins at the CuO nanoparticles and spins of the AFM core [36]. The hysteresis curve for x = 100 as in figure 2(b) shows ferromagnetic behavior without any loop shift along the magnetic field axis. This ferromagnetic ordering, as observed, occurs mainly due to the reduction in the particle size as compared to the spiral modulated spin structure of the bulk BFO (62 nm). The saturation magnetization (MS ) is 3.5 emu g−1 at T = 5 K and H = 50 kOe. Earlier systematic studies on the size dependence of magnetization predicted that MS increases with a decrease in the size of BFO nanocrystals [23]. A higher value of MS for 9 nm is consistent with previous reports. The monotonous increase of ZFC and
3. Results and discussion The formation of the nanocomposites (x)BiFeO3 /(100 − x) CuO was confirmed by the room-temperature x-ray diffraction (figure S1 of supplementary information available at stacks.iop.org/Nano/24/505711/mmedia) and TEM (figure S2 available at stacks.iop.org/Nano/24/505711/mmedia) studies. No impurity was found in the XRD patterns of any of the composites. The TEM images of the composites also reveal prominent interfaces between BiFeO3 and CuO. The formation mechanism of these nanocomposites is described in the supplementary information (available at stacks.iop.org/ Nano/24/505711/mmedia). Figures 1(a) and (b) show the variation of magnetization as a function of temperature in the presence of a 1 kOe magnetic field for the samples x = 0 and x = 100 respectively. Figure 1(a) shows the ZFC curve for the x = 0 (pure CuO) sample. The thermal variation of magnetization is found to be similar to that for pure CuO [14, 15]. The increase in magnetization (ZFC) below a temperature of 50 K may be attributed to the presence of paramagnetic centers which arise due to defects and surface layers [29]. The ZFC and FC magnetization curves for x = 100 (BFO) clearly show a bifurcation at 300 K. This bifurcation in the ZFC and FC curves reveals that the ferromagnetic Curie temperature (TC ) of BFO nanocrystals is beyond room temperature. BFO nanocrystals constitute an ideal FM phase for the EB system since their TC is higher than TN of the AFM CuO phase and they fulfil the requirement for achieving a strong EB effect. FC and ZFC magnetizations for x = 100 (BFO) do not attain a saturation value at low temperatures down to 5 K, in contradiction to conventional ferromagnetic magnetization features at low temperature. Both FC and ZFC increase exponentially with decreasing temperature, as indicated by the solid lines in figure 1(b). Such a type of behavior is 3
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 2. Field cooled (50 kOe) M–H curves at T = 5 K for (a) x = 0 and (b) x = 100. The inset shows the magnified M–H curve.
The discontinuity in the crystal structure and the presence of competing magnetic interactions results in a disordered magnetic state at low temperature. The ratio of Tirr to TB is 2.38, which suggests a broad distribution of the blocking temperature [42]. This arises due to a large variation of the interfacial width in the nanocomposite. The magnetic moment of BFO is about one order of magnitude larger than CuO, as evident from figure 1. The magnetic properties of the nanocomposite may be dominated by the FM phase of BFO being perturbed by the presence of CuO on the surface, leading to a system having both bound moments behaving like bulk and free moments behaving like the surface. This may be approximated by a model consisting of the interfacial spins surrounding the FM phase of BFO leading to blocking, resulting in a behavior like SPM. The ZFC and FC curves are analyzed with a view to studying the change in the micro-magnetic structure of the nanocomposite with the variation of BFO concentration. The temperature dependence of ZFC and FC can be modeled as a ferromagnetic core with disordered spins constituting the shell. The shape of the ZFC curves indicates that TB varies with a log normal distribution. Considering both ZFC and FC resulting from the same saturation magnetization MS (T) and an anisotropic constant Ka being independent of temperature, the values of ZFC and FC were fitted to a core–shell model [43] Z MS2 (T)B 1 vm (T) 2 v f (v) dv MZFC (B, T) = 3kB T v 0 Z MS2 (T)B 1 ∞ + vf (v) dv (1) Ka v vm (T) Z M 2 (T)B 1 vm (T) 2 MFC (B, T) = S v f (v) dv 3kB T v 0 Z 30MS2 (T)B 1 ∞ vf (v) dv (2) + Ka v vm (T)
FC magnetization with decreasing temperature, finite value of coercive field, almost saturation of magnetization and enhancement of MS support that BFO nanoparticles possess a pure ferromagnetic phase. Size destroys the spin spiral structure and favors a perfect ferromagnetic alignment of spins. Cation and oxygen vacancies also play important roles in inducing ferromagnetism in BFO nanocrystals [37, 38]. The hysteresis loop is purely symmetric about the origin (H = 0.0) for BFO nanocrystals, as shown in the enlarged scale in the inset of figure 2(b). Magnetic nanoparticles exhibit an EB effect due to the co-existence of different magnetic phases of surface and core spins. A single FM phase in BFO nanoparticles rules out the possibility of the EB phenomenon. Figure 3 shows magnetization as a function of temperature (both ZFC and FC processes) with a cooling field of 1 kOe for x = 20, 40, 50, 60 and 80 samples. All the ZFC and FC magnetization curves show an irreversible behavior at a temperature of Tirr ∼ 190 K, which is much less than TN (230 K) of bulk CuO and TC of BFO nanocrystal. The ZFC curves on the other hand show a broad peak around 80 K, commonly known as the blocking temperature (TB ). Quite different behavior in ZFC and FC in comparison with pure CuO and BFO suggests that the magnetization originates from interfacial interactions. A peak in ZFC is generally observed in spin glass and superparamagnetic (SPM) systems. A differentiation between spin glass and SPM may be done based on the temperature dependence of the FC magnetization. In the case of a spin glass, the FC magnetization tends to saturate or decreases below TB with decreasing temperature [39, 40]. The FC magnetization increases continuously with decreasing temperature for a superparamagnetic phase [39, 41]. Hence the monotonous increase of FC magnetization with decreasing temperature confirms the SPM state of the nanocomposite. In order to further confirm SPM behavior, M has been plotted as a function of H/T for x = 50 (figure 4). All plots are superimposed on a single curve, which suggests the existence of SPM. As the thermal energy is less than the anisotropy energy, the interfacial spins are blocked along the easy axis randomly and disordered spin structures are formed.
where the first term is the ferromagnetic contribution and the second term is the blocking term. vm is the maximum volume contributing to the ferromagnetism given by 30kKBa T . The log 4
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 3. ZFC (lower curve) and FC (upper curve) for (a) x = 20, (b) x = 40, (c) x = 50, (d) x = 60 and (e) x = 80.
normal distribution function f (d) is given as 1 [ln(d) − ln(d0 )]2 exp − f (d) = √ . 2σ 2 2π dσ
Table 1. Parameters of the size distribution (σ ), anisotropy constant (Ka ) and diameter (d0 ) obtained from fitting the ZFC–FC curves of different compositions (x) using equations (1) and (2).
(3)
Good fits to both the ZFC and FC curves are obtained based on the above equations. The size of the ferromagnetic contribution (d0 ), the size distribution (σ ) and the effective anisotropy constant Ka are extracted from the fitting, which is summarized in table 1. The diameter contributing to the ferromagnetic order is found to vary between 3.75 and 4.4 nm. The mean size of the FM region is smaller than the geometrical size (9 nm) of the FM nanocrystals. This indicates that a large fraction of spins constituting the interfacial region are blocked. The size distribution is observed to vary from 1.44 to 1.56. A broad size distribution,
x
σ
Ka (J kg−1 )
d0 (nm)
20 40 50 60 80
1.50 ± 0.006 1.44 ± 0.005 1.57 ± 0.006 1.56 ± 0.006 1.49 ± 0.005
3.86 ± 0.021 3.31 ± 0.016 2.93 ± 0.018 3.22 ± 0.019 3.36 ± 0.018
3.87 ± 0.021 4.44 ± 0.022 3.87 ± 0.022 3.76 ± 0.021 4.11 ± 0.021
σ > 1, is found for all compositions due to the complex interfacial magnetic interaction. There is a steady variation in the effective anisotropic constant, with a minimum at 50% composition. 5
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
characteristic freezing temperature of disordered spins and C is a constant. The fitted curve in figure 5(b) for x = 50 indicates that MS (T) obeys the Bloch law at high temperature (T > 100 K). The exponential law is valid at low temperature, as illustrated in the enlarged scale in the inset of figure 5(b). A clear deviation from Bloch and exponential behavior is found in the temperature interval between 100 and 30 K. A gradual decrease in MS from 100 to 30 K indicates ferrimagnetic features. It seems that ordered and interfacial spins are aligned oppositely to reduce the net magnetic moment. All the fitted parameters obtained from equations (4) and (5) are displayed in table 2. The value of α is smaller than 1.5 for a lower content of BFO, suggesting a deviation from the usual Bloch T 3/2 law. The higher concentration of BFO obeys the Bloch law as the exponent α is very close to 1.5, as indicated table 2. The value of Tf lies between 6 and 10 K. So the contribution of freezing spins to MS (T) is negligible above ∼5Tf , as observed in the behavior of MS (T) above 30 K. The room-temperature saturation magnetization (MS ) increases with increasing BFO content, as shown in figure S3 (supplementary information available at stacks.iop.org/Nano/ 24/505711/mmedia). Figure 6 shows the M–H curves for both zero field cooled and 50 kOe field cooled cases for x = 20, 40, 50, 60 and 80 samples at a temperature of 5 K. M–H curves at an enlarged scale in the low magnetic field region are displayed at the inset of figure 6. The M–H curves for the ZFC process of all the samples show a clear central symmetry about the magnetic field axis. Whereas the curves for the FC process exhibit a negative horizontal shift for all samples. This shift along the magnetic field axis, as discussed earlier, is termed as EB. The occurrence of EB in this case may occur due to enhanced unidirectional anisotropy at the interface of FM BFO nanocrystals embedded in the AFM CuO matrix [3]. Pure CuO reveals the EB effect, which is absent in BFO, as shown in figure 2. EB appears below 50 K for nanostructured CuO due to uncompensated surface spins. The observation of the EB effect at higher temperature (>50 K) and larger values of EB indicate that EB primarily originates from the BFO/CuO interface.
Figure 4. Magnetization (M) as a function of H/T for x = 50.
The resulting temperature dependence of saturation magnetization MS (T) on the concentration of the composite is shown in figure 5(a). All the plots show a monotonic increase in MS (T) with increase in the composition. The thermal natures of all the compositions are similar, with a monotonic increase in the magnetization with decreasing temperature, saturating around 100 K. The temperature dependence of the saturation magnetization generally follows the Bloch T 3/2 law below TC . A sharp increase in MS (T) below about 30 K is observed for all compositions. This additional contribution at low temperature can be ascribed to disordered spins aligned ferromagnetically and varies exponentially, as described earlier. Hence MS (T) has been analyzed separately in two temperature regimes by the following equations: MS (T) = MS (0)[1 − BT α ] T Unbound +C MS (T) = A exp − Tf
(4) (5)
where MS (0) is the saturation magnetization as the temperature tends to zero, the Bloch B constant and the exponent α generally depends on the spin structure and spin dynamics at finite temperature. In equation (5) Tf is the
Figure 5. (a) Temperature dependence saturation magnetization (MS ) for x = 20, 40, 50, 60, 80 and (b) Bloch and exponential fitted MS versus T curve for x = 50, with the inset showing the exponential fitted curve. 6
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 6. Zero field cooled and 50 kOe field cooled M–H curves at T = 5 K for (a) x = 20, (b) x = 40, (c) x = 50, (d) x = 60 and (e) x = 80. Insets show the magnified M–H curves in the low magnetic field region for all samples. Table 2. Parameters of the saturation magnetization at 0 K (MS (0)), Bloch constant (B), exponent (α) and characteristic freezing temperature (Tf ), and constant (C) obtained from fitting MS with temperature using equations (4) and (5) respectively for all compositions (x). x
MS (0) (emu g−1 )
B(10−4 )
α
A
Tf (K)
C (emu g−1 )
20 40 50 60 80
1.01 ± 0.016 0.97 ± 0.008 1.09 ± 0.006 1.28 ± 0.007 1.39 ± 0.007
20.5 ± 7.36 4.95 ± 1.27 0.46 ± 0.164 0.47 ± 0.165 0.72 ± 0.168
0.91 ± 0.058 1.14 ± 0.042 1.52 ± 0.058 1.51 ± 0.058 1.48 ± 0.039
0.21 ± 0.006 0.26 ± 0.006 0.23 ± 0.005 0.27 ± 0.005 0.35 ± 0.007
8.68 ± 0.47 10.26 ± 0.585 9.39 ± 0.417 9.42 ± 0.417 6.56 ± 0.205
0.84 ± 0.002 0.86 ± 0.004 0.99 ± 0.002 1.16 ± 0.002 1.28 ± 0.002
The M–H curves also reveal a difference in the coercivity (HC ) between the ZFC and FC process. HC in the FC process is observed to be greater than HC in the ZFC process. This enhancement in HC for the same sample may be attributed
to the large unidirectional anisotropy of the FM clusters present at the FM/AFM interface [2, 3]. Along with the anomaly in HC , an anomaly along the magnetization axis of the M–H curves is also observed. A vertical shift is also 7
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 7. Variation of exchange bias field (HEB ) and coercivity (HC ) with ferromagnetic (FM) ratio (x) at T = 5 K.
Figure 8. Dependence of HEB with cooling field (Hcool ) for x = 50 at T = 5 K.
observed for all the samples in the FC process as compared to the ZFC process, which seem to be symmetric about the M = 0 axis. This vertical anomaly in the FC process of the M–H curves suggests a strong influence of the pinned magnetization to the exchange fields which may occur due to the moment originating from the uncompensated spins of the CuO matrix [44]. To gain more insight into HEB a detailed study of HEB on the BFO(FM) ratio x, cooling field (Hcool ) and temperature is presented in following sections. Figure 7 shows the composition dependence of HEB and HC (both calculated from the FC process of M–H) on the BFO(FM) ratio (x) at a temperature T = 5 K. It has been observed that the value of HEB for (x)BFO/(100 − x)CuO increases with increasing x, showing a maximum value of 1841 Oe at x = 50, then decreases with further increases of x. The increase in HEB with x until x = 50 may be attributed to enhancement in the BFO/CuO interfacial region and the strong interfacial exchange coupling between BFO and CuO. The decrease in HEB beyond x = 50 may be ascribed to the formation of isolated FM clusters of BFO nanocrystals. The formation of the cluster of BFO nanocrystals reduces the interfacial exchange coupling, resulting in a decrease of HEB . It has been well established that HEB is inversely proportional to the thickness of FM materials (HEB ∝ 1/tFM ) [3]. Beyond x = 50 the FM BFO nanocrystals form clusters, resulting in an increase in the effective thickness of FM, which leads to a decrease of HEB . HC also seems to behave similarly to HEB , initially increasing until x = 50 and then decreasing with further increases in x. In the nanocomposite, the relative phase content of the AFM and FM components plays a vital role in determining the magnitude of the EB effect. Figure 8 shows the influence of Hcool on HEB for the sample x = 50. We observe HEB to be increasing with an increase in Hcool . The increase in the magnitude of Hcool induces more uncompensated spins of the AFM CuO matrix to become aligned along the field direction. These spins lead to a change in the magnetic configuration at the FM/AFM interface [45]. Thus, with increasing Hcool more frozen-in spins are created, ultimately leading to the increase in HEB . The rapid increase of HEB up to about 10 kOe is due to a sharp increase of magnetization in the low magnetic field region,
as evident from the hysteresis displayed in figure 6. A larger number of magnetic moments participate in the EB effect with an increase of the magnetic field. A trend of saturation in HEB at higher magnetic fields is due to saturation magnetization in the M–H curve. Figure 9(a) shows the temperature dependence of HEB for x = 20, 40, 50, 60 and 80. It has been observed that the value of HEB decreases drastically with increasing temperature due to weakening of the exchange interactions in the high-temperature region. Interfacial disorder spins result in an exponential behavior of HEB as a function of temperature [46, 47]. The temperature dependence of HEB has been fitted by an exponential law T 0 HEB = HEB exp − (6) T0 0 is the limiting value of H where HEB EB at T = 0 K and T0 is the characteristic temperature shown in figure 9(b) for 0 and T are shown in x = 50. The best fitted values of HEB 0 0 reveals a maximum table 3. The composition variation of HEB similar to HEB for x = 50. The estimated values of T0 are lower than the blocking temperature, TB , and are about 32 TB . A significant fraction of blocked spins become free with increasing temperature, which causes a rapid decrease of HEB . The value of HEB below 5 K deviates from exponential behavior. Below Tf a large number of disordered spins align ferromagnetically along the direction of the applied magnetic field, which effectively enhances the volume of the FM region. This trend of HEB is similar to the increase of FM thickness which reduces the EB effect [3]. The variation of HC with temperature is shown in figure 10. The value of HC decreases with increasing temperature. An increase of HC in the FC process (HC (FC)) is found compared to the ZFC (HC (ZFC)) data. This enhancement is due to an additional unidirectional anisotropy arising from exchange interactions between FM and AFM spins at the interface [8]. The increment in HC in the field cooled process 1HC = HC (FC) − HC (ZFC) becomes zero above Tirr , as above this temperature spins become unblocked. Hence the temperature dependence of 1HC is fitted by the
8
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
Figure 9. Dependence of HEB on temperature for x = 20, 40, 50, 60 and 80 and (b) fitted HEB versus temperature curve for x = 50.
Figure 10. Dependence of (a) HC on temperature for x = 20, 40, 50, 60 and 80 and (b) 1HC versus temperature for x = 50. 0 Table 3. Parameters HEB , T0 and Kex and Tirr obtained from fitting HEB and 1HC with temperature using equations (6) and (7) respectively for all compositions (x).
x
0 HEB (Oe)
T0 (K)
Kex (10−2 ) (J kg−1 )
Tirr (K)
20 40 50 60 80
1801.87 ± 12.513 1965.68 ± 15.496 2323.32 ± 10.869 2162.22 ± 12.883 1819.32 ± 11.234
51.72 ± 1.619 58.21 ± 0.751 47.72 ± 1.316 46.85 ± 0.438 47.16 ± 1.915
3.57 ± 0.065 3.92 ± 0.086 5.01 ± 0.057 4.78 ± 0.108 4.73 ± 0.106
170.01 ± 5.327 177.69 ± 7.001 174.03 ± 5.716 173.86 ± 6.629 170.82 ± 5.26
following formula [48] Kex 1HC = 1− MS (T)
s
T Tirr
responsible for the coercive field is equal to the number of moments participating in the exchange bias effect. A comprehensive temperature dependence of the ZFC and FC investigations indicates that spins are blocked below Tirr . The observation of HEB below Tirr manifests that the blocked spins are responsible for the EB effect. Moreover, theoretical analysis of ZFC and FC shows that the size of the ordered FM region is almost half the actual size of BFO nanocrystals for all compositions. Hence a large fraction, i.e. about 87 ths, of FM spins are involved in the interfacial magnetism. Two distinct magnetic interactions are involved in the BFO/CuO nanocomposite. The bulk AFM of CuO is retained by nearest neighbor superexchange interactions between Cu ions through oxygen atoms. The Cu–O–Cu bond distance and bond angle determine the strength of the exchange interaction. The random orientation
! (7)
where Kex is the exchange anisotropy constant and MS (T) is derived from equations (1) and (2) and shown in figure 10(b) (for x = 50). The calculated parameters Kex and Tirr are listed in table 3. The maximum value of Kex for x = 50 is consistent with the highest value of HEB . The estimated values of Tirr are very close to the experimental values for all compositions, as exhibited in figure 3. The fitted data for Kex suggests that Kex mainly controls the value of HEB . A linear behavior between HEB and HC , as depicted in figure S4 (supplementary information available at stacks.iop.org/ Nano/24/505711/mmedia), implies the number of moments 9
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
(CSIR), Government of India, for providing fellowships. This work is funded by the Council of Scientific and Industrial Research, Government of India, Scheme No: 03(1210)/12/EMR-II.
of surface anisotropy gives rise to disordered spins in BFO nanocrystals. In the presence of an external magnetic field and at a low temperature, disordered spins are aligned to induce a ferromagnetic phase in BFO nanocrystals. Competition between magnetic interactions in the AFM and FM regions and breaking of symmetry at the interface result in the blocking of spins. The interfacial blocked spins exert a microscopic torque on the FM spins to maintain the original alignment and cause a shift in the hysteresis loop to produce the EB effect. AFM may terminate with spin compensated or uncompensated layers due to the complicated spin structure of AFM CuO. In the compensated interface, the net magnetization is zero and does not yield the EB effect. The majority of uncompensated Cu spins take part in the disordered phase. The spins of both the AFM and FM regions are blocked due to the strong interfacial coupling. The uncompensated spins of CuO mediate magnetic coupling between the AFM and FM phases. The magnitude of HEB depends on the number of uncompensated interfacial spins that are pinned to the FM region [49]. Different degrees of coupling among the uncompensated spins of AFM, the disordered spins of FM and bond breaking at the AFM/FM interface give rise to complex magnetism and a large EB effect in the BFO/CuO nanocomposite.
References [1] Meiklejohn W H and Bean C P 1956 Phys. Rev. 102 1413 [2] Berkowitz A E and Takano K 1999 J. Magn. Magn. Mater. 200 552 [3] Nogues J and Schuller I K 1999 J. Magn. Magn. Mater. 192 203 [4] Kiwi M 2001 J. Magn. Magn. Mater. 234 584 [5] Chappert C, Fert A and Van Dau F N 2007 Nature Mater. 6 813 [6] Dieny B, Speriosu V S, Parkin S S P, Gurney B A, Wilhoit D R and Mauri D 1991 Phys. Rev. B 43 1297 [7] Skumryev V, Stoyanov S, Zhang Y, Hadjipanayis G, Givord D and Nogues J 2003 Nature 423 850 [8] Del Bianco L, Fiorani D, Testa A M, Bonetti E and Signorini L 2004 Phys. Rev. B 70 052401 [9] Nogues J, Sort J, Langlais V, Skumryev V, Surinach S, Munoz J S and Barob M D 2005 Phys. Rep. 422 65 [10] Iglesias O, Labrta A and Batlle X 2008 J. Nanosci. Nanotechnol. 8 2761 [11] Filippetti A and Fiorentini V 2005 Phys. Rev. Lett. 95 086405 [12] Kimura T, Sekio Y, Nakamura H, Siegrist T and Ramirez A P 2008 Nature 7 291 [13] Jin G, Cao K, Guo G C and He L 2012 Phys. Rev. Lett. 108 187205 [14] D´ıaz-Guerra C, Vila M and Piqueras J 2010 Appl. Phys. Lett. 96 193105 [15] Punnoose A, Magnone H, Seehra M S and Bonevich J 2001 Phys. Rev. B 64 174420 [16] Eerenstein W, Mathur N D and Scott J F 2006 Nature 442 759 [17] Cheong S W and Mostovoy M 2007 Nature Mater. 6 13 [18] Ramesh R and Spaldin N A 2007 Nature Mater. 6 21 [19] Singh A, Pandey V, Kotnala R K and Pandey D 2008 Phys. Rev. Lett. 101 247602 [20] Wang J et al 2003 Science 299 1719 [21] Ederer C and Spaldin N A 2005 Phys. Rev. B 71 060401 [22] Gao F, Yuan Y, Wang K F, Chen X Y, Chen F, Liu J M and Ren Z F 2006 Appl. Phys. Lett. 89 102506 [23] Park T J, Papaefthymiou G C, Viescas A J, Moodenbaugh A R and Wong S S 2007 Nano Lett. 7 766 [24] Chakrabarti K, Das K, Sarkar B and De S K 2011 J. Appl. Phys. 110 103905 [25] Chakrabarti K, Das K, Sarkar B, De S K, Sinha G and Lahtinen J 2012 Appl. Phys. Lett. 101 042401 [26] Sung K D, Park Y A, Seo M S, Jo Y, Hur N and Jung J H 2012 J. Appl. Phys. 111 033915 [27] Manna P K, Yusuf S M, Shukla R and Tyagi A K 2011 Phys. Rev. B 83 184412 [28] Dong S, Yao Y, Hou Y, Liu Y, Tang Y and Li X 2011 Nanotechnology 22 385701 [29] Chandrasekhar T V and Sahni V C 1994 J. Phys.: Condens. Matter 6 L423 [30] Kasyutich O, Desautels R D, Southern B W and van Lierop J 2010 Phys. Rev. Lett. 104 127205 [31] Shendruk T N, Desautels R D, Southern B W and van Lierop J 2007 Nanotechnology 18 455704 [32] Aquino R, Depeyrot J, Sousa M H, Tourinho F A, Dubois E and Perzynski R 2005 Phys. Rev. B 72 184435 [33] Kachkachi H, Ezzir A, Nogues M and Tronc E 2000 Eur. Phys. J. B 14 681 [34] Jonsson P E, Mathieu R, Wernsdorfer W, Tkachuk A M and Barbara B 2007 Phys. Rev. Lett. 98 256403
4. Conclusion The temperature dependence of the magnetization of pure BFO nanocrystals of size 9 nm follows an exponential decay with increasing temperature due to free spins. Upon dispersion of BFO into the CuO matrix, a significant change in the magnetic behavior is seen in comparison with the pure BFO and CuO phases. Strong interfacial exchange coupling and structural mismatch lead to spin order, spin disorder and a ferrimagnetic state. The radius for the effective FM ordered region is less than the actual particle size determined using TEM analysis, which certainly establishes that a large fraction of FM spins are involved in the interfacial magnetism. The broad distribution of blocking temperature indicates the presence of several types of FM/AFM interfacial regions with different anisotropy due to the large size distribution of the FM region. The temperature dependence of saturation magnetization follows the Bloch law at high temperature and the exponential law at low temperature. Maximum values of the exchange field of 1841 Oe and the coercivity of 1934 Oe are obtained for an equal content of BFO and CuO in the nanocomposite. Strong exchange coupling between the uncompensated surface spins of CuO and the disordered spins of BFO nanocrystals, as well as the discontinuity in crystal structure at the interface, result in interesting magnetic properties and a significant exchange bias field in the BFO/CuO nanocomposite. The large exchange bias effect and multiferroicity in BFO and CuO make the nanocomposite more attractive for multifunctional devices.
Acknowledgments Kaushik Chakrabarti, Babusona Sarkar and Kajari Das are grateful to the Council of Scientific and Industrial Research 10
Nanotechnology 24 (2013) 505711
K Chakrabarti et al
[35] Barbara B 2007 Phys. Rev. Lett. 99 177201 [36] Liu K L, Yuan S L, Duan H N, Zheng X F, Yin S Y, Tian Z M, Wang C H and Huo S X 2010 J. Appl. Phys. 107 023911 [37] Paudel T R, Jaswal S S and Tsymbal E Y 2012 Phys. Rev. B 85 104409 [38] Ederer C and Spaldin N A 2005 Phys. Rev. B 71 224103 [39] Gruyters M 2005 Phys. Rev. Lett. 95 077204 [40] Alonso J, Fdez-Gubieda M L, Barandiaran J M, Svalov A, Fern´andez Barqu´ın L, Alba Venero D and Orue I 2010 Phys. Rev. B 82 054406 [41] Batlle X and Labarta A 2002 J. Phys. D: Appl. Phys. 35 R12 [42] Domingo N, Fiorani D, Testa A M, Binns C, Baker S and Tejada J 2008 J. Phys. D: Appl. Phys. 41 134009 [43] Respaud M et al 1998 Phys. Rev. B 57 2925
[44] Yin S Y, Yuan S L, Tian Z M, Liu L, Wang C H, Zheng X F, Duan H N and Huo S X 2010 J. Appl. Phys. 107 043909 [45] Salabas E L, Rumplecker A, Kleitz F, Radu F and Schuth F 2006 Nano Lett. 6 2977 [46] Ding J F, Lebedev O I, Turner S, Tian Y F, Hu W J, Seo J W, Panagopoulos C, Prellier W, Van Tendeloo G and Wu T 2013 Phys. Rev. B 87 054428 [47] Huang X H, Ding J F, Zhang G Q, Hou Y, Yao Y P and Li X G 2008 Phys. Rev. B 78 224408 [48] Jacobs I S and Bean C P 1963 Magnetism vol 3, ed G T Rado and H Suhl (New York: Academic) p 273 [49] Ohldag H, Scholl A, Nolting F, Arenholz E, Maat S, Young A T, Carey M and Stohr J 2003 Phys. Rev. Lett. 91 017203
11
