Magneto-dielectric and magnetoelectric anisotropies of CoFe₂O₄/Bi₅Ti₃FeO₁₅ bilayer composite heterostructural films

Abstract

Bilayer composite heterostructural films consisting of magnetic CoFe₂O₄ and multiferroic Bi₅Ti₃FeO₁₅ films were prepared by the chemical solution deposition method. Morphological, ferroelectric, piezoelectric, magnetic, magneto-dielectric and magnetoelectric properties were investigated for CoFe₂O₄/Bi₅Ti₃FeO₁₅ composite films. The magneto-dielectric and magnetoelectric effects are observed for the composite films, which is attributed to the enhanced interface coupling and mechanical strain transfer due to the small lattice mismatch between CoFe₂O₄ and Bi₅Ti₃FeO₁₅ film layers. Magneto-dielectric and magnetoelectric anisotropies were also observed for the composite films. Correspondingly, the in-plane and out-of-plane magnetoelectric voltage coefficients reach 57 mV cm⁻¹ Oe⁻¹ and 78.9 mV cm⁻¹ Oe⁻¹, respectively. Thus magneto-dielectric and magnetoelectric anisotropies originate from the magnetic anisotropy of CoFe₂O₄/Bi₅Ti₃FeO₁₅ composite films. The present work provides promising candidates for applications in magnetoelectric devices.

---

Introduction

A strong direct magnetoelectric coupling effect (magnetic field induced polarization or electric field induced magnetization) has been demonstrated at room temperature via magneto–mechanical–electric coupling in ferromagnetic/piezoelectric composites, which have potential application in several smart device designs with high sensitivity.^1–3 According to the connectivity, 2–2 type multilayer composites consisting of two-dimensional piezomagnetic and piezoelectric phases are formed by laminating them together. Among them, layered ferromagnetic/piezoelectric composite structures are the most attractive due to their large magnetoelectric voltage coefficient.⁴ With regard to the applications in micro-electro-mechanical systems, many efforts have been paid to magnetoelectric composite films. For these composite structures, the perovskite ferroelectric oxides such as Pb(Zr_0.52Ti_0.48)O₃ (PZT) are usually used as the piezoelectric phase due to their large piezoelectric coefficient (d₃₃ = 330 pm V⁻¹).⁵ CoFe₂O₄ with spinal structure is used as piezomagnetic phase due to their strong magnetostrictive effect.⁶ These composite films present large magnetoelectric coupling effect by magnetic-mechanical-electric transfer across the interface. However, these kinds of lead based magnetoelectric composite films are not favorable to the requirement of the environmental protection. So it is urgent to develop a kind of environmentally friendly materials with higher piezoelectric coefficient to replace PZT films. Recently, single-phase multiferroics, Bi₅Ti₃FeO₁₅ (BTFO) materials with a four-layered perovskite unit of (Bi₃Ti₃FeO₁₃)²⁻ sandwiched by two (Bi₂O₂)²⁺ layers along the c axis, have attracted attention due to their excellent piezoelectric, ferroelectric and magnetic properties.⁷ Therefore, as a kind of environmental friendly lead-free ferroelectric films, Bi₅Ti₃FeO₁₅ (BTFO) films have potential application in magnetoelectric composite films as piezoelectric phases. Besides, the autologous coupling between spins and orbits in the single-phase BTFO multiferroic film layer is also attributed to the magnetoelectric effect.

This work aims to gain strong magnetoelectric effect in composite films by using single-phase multiferroic material BTFO as the piezoelectric phase while the CoFe₂O₄ as the ferromagnetic phase. Thus a new multiferroic heterostructural films consisting of the magnetic film layers and lead-free single-phase multiferroic film layers were successfully prepared by the chemical solution deposition method. A strong magnetoelectric coupling effect and its anisotropy were observed. The origins of these effects are discussed in detail.

Experimental procedure

CFO/BTFO composite films were prepared using chemical solution deposition. For the preparation of BTFO solutions, Bi(NO₃)₃·5H₂O, Fe(NO₃)₃·9H₂O and butyl titanate [CH₃(CH₂)₃O]₄Ti were dissolved in glycol with excess 18% Bi for compensating the loss during annealing. The precursor solutions of BTFO were achieved at the concentration of 0.1 mol L⁻¹ while for CFO solutions, raw materials Fe(NO₃)₃·9H₂O and Co(CH₃COO)₂·4H₂O were solved in mixture solution (citric acid : ethylene glycol ratio 1 : 4) to form the precursor solutions at concentration of 0.2 mol L⁻¹. Then the solutions were stirred until the solution was clear transparent. For the preparation of the composite films, since CFO films are easier to deposit on the Pt/Ti/SiO₂/Si(100) wafer than BTFO films, CFO precursor solutions first were deposited on the wafer by spin coating, with 400 rpm for 18 seconds and 4000 rpm for 50 seconds. Then the wafers were annealed by the rapid thermal processor (RTP-500) at 800 °C in air about 200 seconds. These processes were repeated several times to obtain the desired thickness of CFO films. Next, in order to confirm the piezoelectric coefficient d₃₃ by PFM, BTFO precursor solutions were deposited on CFO films, with similar process and annealed 700 °C in air about 200 seconds. Because the theoretical calculation demonstrated that the optimized volume fraction of CFO magnetic phase in the 2–2 laminate composites is about 0.4 to obtain the strongest magnetoelectric voltage coefficient,⁸ the composite films are prepared with BTFO film thickness of 650 nm and CFO film thickness of 350 nm.

The crystalline phases were characterized by X-ray diffraction (XRD, Panalytical Empyrean with Cu K_α radiation λ = 0.15406 nm). The surface morphologies of the films were investigated by a scanning electron microscope (SEM, Hitachi S-3500). And the ferroelectric, leakage fatigue properties were studied by a multiferroic tester system (Multi-Ferroic100V, Radient Technology, USA). The piezoelectric coefficients d₃₃ of the composite films were investigated by a piezoresponse force microscopy (PFM, Asylum Research Cypher™). The magnetic measurement was performed using a physical property measurement system (PPMS, Quantum Design). The response of the dielectric constant and dielectric loss to the applied bias magnetic field were studied by a precision impedance analyzer (Agilent E4990A) associated with a bias magnetic field at a frequency of 1 kHz. For the magnetoelectric coefficient (α_E) and magnetoelectric fatigue measurement, a magnetic bias field H_bias together with a small alternating magnetic field H_ac = 7.35 Oe and frequency f = 1 kHz was applied parallel and vertical to the film plane. The induced magnetoelectric voltage V_ME was recorded by a lock-in amplifier (SR830, SRS Inc.). All measurements above were performed at room temperature.

Results and discussion

Fig. 1 shows the XRD patterns of CFO/BTFO composite heterostructural films. It is shown that the patterns of the composite films are indexed to two sets of diffraction peaks coming from CFO film layer and BTFO film layer. For BFTO film layer, it belongs to A2₁am space group and exhibits a single Aurivillius structure containing four perovskite layers, which is identified by indexing all the diffraction peaks on the basis of an orthorhombic cell (Joint Committee for Powder Diffraction Standard (JCPDS) no. 38-1257). For CFO film layer, it belongs to cubic spinel structure with space group Fd3̄m according to (JCPDS, no. 22-1086) profiles.⁹ And no other intermediate and impurity phases (e.g. Fe₂O₃, Bi₂O₃, or Bi₂Ti₂O₇) are detected for CFO/BTFO composite films, indicating the absence of any chemical reaction between the two phases during the annealing process. Interestingly, the overlap of (119) diffraction peak for BTFO films and (220) diffraction peak for CFO films suggests a well lattice matching in the composite films since the lattice parameter is very close. Thus well lattice matching is beneficial for the interface coupling and stress transfer. The insert of Fig. 1 presents the cross-sectional SEM image of the composite heterostructural films. It can be seen that the thickness of CFO film layers and BTFO film layers is about 350 nm and 650 nm, respectively. Clear boundary can be obtained without any phase or atomic inter-diffusion between CFO film layers and BTFO film layers and the substrate. The perfect coupling between CFO and BTFO film layers will enhance magnetoelectric effect and drive a large induced voltage output because the chemical interaction between the ferroelectric and magnetic phases may lead to the degrades of their ferroelectric or magnetic properties.

Fig. 1 XRD patterns of CFO/BFTO composite films, insert shows the cross-sectional SEM images.

Fig. 2(a) shows the ferroelectric hysteresis loops of CFO/BTFO composite films. Clearly, the composite films exhibit a well-shaped ferroelectric polarization hysteresis loop (P–E loop), indicating that they possess good ferroelectric properties at the applied external electric field from 300 kV cm⁻¹ to 650 kV cm⁻¹. The saturation polarization P_s, remnant polarization P_r and coercive field E_c of composite films at different electric field are summarized in Table 1.

Fig. 2 (a) The polarization (P)-electric field (E) hysteresis loops of CFO/BFTO composite films; (b) the leakage current density (J) versus electric field (E), insert plots ln E vs. ln J.

Table 1 The comparison of P_s, P_r and E_c for CFO/BTFO composite films at different electric fields E

E (kV cm^−1)	300	320	390	450	520	580	650
Ps (μC cm^−2)	40	45.8	57.4	66.1	73.5	80.7	87.3
Pr (μC cm^−2)	23.6	28.0	33.8	38.9	42.6	46.2	50.6
+Ec (kV cm^−1)	128	152	177	193	218	233	242

---

As summarized in Table 1, the saturation polarization P_s and remanent polarization P_r reach 87.3 μC cm⁻² and 50.6 μC cm⁻², respectively, at a maximum applied electric field of 640 kV cm⁻¹. The polarization of the present films is comparable ever larger than that of other reported BTFO composite films and single-phase BFTO films.^10,11 The results suggest that the ferroelectric polarization of BTFO film layers is not depressed. That is to say, no interface diffusion between CFO and BTFO film layers, which agrees well with the SEM images. The values of P_s and P_r increase with further increasing applied external electric field, implying good ferroelectric properties and without being broken down. And the E_c values of the composite films increase with the further increase of the applied electric field. These results originate from the increase of threshold field at the interfacial layer because of the appearance of the insulating CFO layer, which leads to the enhanced E_c value. Thus improved ferroelectric properties of CFO/BTFO composite films are attributed to the outstanding ferroelectric properties of BTFO film layers. On one hand, BFTO exhibited well ferroelectric polarization due to a four-layered perovskite consisted two (Bi₂O₂)²⁺ layers. On the other hand, mobile free charges including oxygen vacancies and other ion defects can move following the applied electric field and contribute to the total polarization.¹² In addition, leakage current also directly affects the ferroelectric polarization, as discussed as follows.

Fig. 2(b) plots the leakage current density (J) as a function of the applied electric field (E) for CFO/BTFO composite films. The leakage current density of CFO/BTFO composite films exhibits regular increasing over the entire region of the applied electric field. It is shown that the leakage current density of CFO/BTFO composite films is 1.35 × 10⁻³ A cm⁻² at an applied electric field of 650 kV cm⁻¹, which is lower than that of the pure BFTO films.⁷ The lower leakage is attributed to CFO layers that can stabilize the charge fluctuation caused by the mixed-valence state of Fe²⁺ and Fe³⁺ due to the introduction of oxygen vacancies. Additionally, CFO film layers can prevent the flow of charge carriers moving to the bottom electrodes due to its high resistance.¹³ The insert plots the fitted curves of ln E vs. ln J to further understand the leakage mechanism at an applied electric field. It is shown that the curves of ln E vs. ln J are linear and the slope is 1.93 around 2 for CFO/BTFO composite films in the entire electric field, which suggests that the films are dominated by the space-charge-limited current conduction (SCLC).¹⁴ Thus mechanism originates from the decreased Fe³⁺ ions on the octahedral sites due to the interface interaction.¹⁵ Actually, the electron traps exist and capture the injected electrons. As the capturing progress is completed, the injected electrons would participate in the conduction with further increasing electric fields, resulting in the increase of the current.¹⁶ For CFO/BTFO composite films the conduction only occurs one mechanism, so the entire region belongs to capturing process.

Fig. 3 presents the typical butterfly-shaped piezoelectric strain loop of CFO/BTFO composite films. The asymmetry is attributed to the self-poling effect at the interface between the film and the bottom electrode.¹⁷ The piezoelectric coefficient d₃₃ can be defined by d₃₃ = δl/V, where δl is the displacement and V is the applied voltage. The effective d₃₃ could be calculated, when a direct current voltage of up to 20 V was applied. It is shown that the films occur the displacement δl about 5 nm at the voltage of 20 V. According to the equation, effective piezoelectric coefficient d₃₃ is about 253 pm V⁻¹, which is obviously larger than that of the single-phase BTFO films.⁷ The high piezoelectric coefficient of CFO/BTFO composite films is dependent on following factors. First, the piezoelectric coefficient d₃₃ can be calculated by the equation of d₃₃ = 2Q_effεP in perovskite structure materials.¹⁰ Where Q_eff is the effective electrostrictive coefficient, P is the spontaneous polarization and ε is permittivity. That is to say, a large ε and P are crucial for the enhancement of d₃₃. For the present CFO/BTFO composite films, the ε and P_r values are large as discussed above. Besides, as shown in the ferroelectric hysteresis loops, the low coercive field E_c allows more ferroelectric domains to switch and thus enhances the unipolar strain.¹⁸ Thus, the larger effective electrostrictive coefficient Q_eff is expected to be obtained, which also attributed to the larger piezoelectric coefficient.

Fig. 3 The piezoelectric coefficient (d₃₃) and surface displacement (d) vs. applied voltage (V) of CFO/BFTO composite films.

Fig. 4 shows the in-plane and out-of-plane magnetic hysteresis (M–H) loops of CFO/BTFO composite films at room temperature. Magnetic measurements do show strong ferromagnetic properties for CFO/BTFO composite films instead of the antiferromagnetic nature. For a two-dimension film material, the magnetic anisotropy generally is described by the difference of magnetism along specific directions including in-plane and out-of-plane direction. Thus, considering the source of magnetic anisotropy, shape anisotropy, stress anisotropy, and interface exchange anisotropy are main factors. The ratios of remanence to saturation magnetization (M_r/M_s) and corresponding coercivity magnetic field H_c values are summarized in Table 2 to characterize their magnetic anisotropy.

Fig. 4 In-plane and out-of-plane magnetic hysteresis loops of CFO/BTFO composite films, and the insets depict corresponding enlarged central parts of the magnetic hysteresis curves.

Table 2 Magnetic properties (H_c and M_r/M_s in out-of-plane and in-plane directions) and anisotropy energies of CFO/BTFO composite films

Hc∥ (Oe)	Hc⊥ (Oe)	(Mr/Ms)⊥	(Mr/Ms)∥	Ks1 (erg cm^−3)	Ks3∥ (erg cm^−3)	Ks3⊥ (erg cm^−3)
1400	2300	52.1%	33.7%	1.51 × 10^6	−4.5 × 10^5	+1.2 × 10^6

---

As shown, the in-plane M_r and M_s are 90 emu cm⁻³ and 267 emu cm⁻³, respectively. Thus in-plane M_r/M_s ratios (M_r/M_s)_∥ is 33.7%. While the out-of-plane M_r and M_s are large, reaching about 229 emu cm⁻³ and 440 emu cc⁻³, respectively. Thus the out-plane M_r/M_s ratios (M_r/M_s)_⊥ reaches 52.1%. As the in-plane M_r/M_s ratios smaller than 50%, so the uneasy axis is well defined. Therefore, a strong out-of-plane magnetic anisotropy is observed with a larger M_r/M_s ratios value of 52.1% in composite films. Besides, the out-of-plane coercive field (H_c⊥) becomes almost 1.7 times more than the in-plane coercive field (H_c∥). In this case, the spin state is more stable along the out-of-plane direction than along the in-plane one. Thus, the remanent magnetization (M_r) along the out-of-plane direction was a little higher than in-plane one.

To elucidate the origins of the out-of-plane magnetic anisotropy, many major sources might be considered, including shape anisotropy, stress anisotropy, and interface exchange anisotropy. The anisotropy of films is determined by the competition between several uniaxial anisotropies. The total uniaxial anisotropy energy E_t in composite films is given by the equation as follows,^19,20

where K_s1 is the stress anisotropy energy, K_s2 is the surface anisotropy energy, and K_s3 is the shape anisotropy energy, t refers to the thickness of the films. As the present composite films are relative thick about 1 μm, the second term can be ignored.

For the shape anisotropy, its anisotropy energy K_s3 is calculated based on a two-dimensional model of K_s3 = ±2πM_s²,²¹ where M_s is the saturated magnetization, the negative and positive signs correspond to in-plane and out-of-plane anisotropy, respectively. Another is anisotropy energy source stems from the stress anisotropy, which originates from the magnetostrictive effect. Since CFO has a significant positive magnetostriction, the magnetoelastic energy will tend to be dominated, favoring the out-of-plane direction to be the easy axis. The stress anisotropy energy (K_s1) is given by the equation.

where λ is the magnetostrictive coefficient of CFO (−350 × 10⁶) and Y is the Young's modulus (∼141.6 GPa), ε denotes compressive strain (∼−1.1%).²² λ of CFO can be used to estimate the K_s1 values of CFO/BTFO composite films. Thus the anisotropy energy is shown in Table 2. The out-of-plane anisotropy is obtained. Finally, the spin dependent coupling and interfacial domain formation spin-flop coupling are considered to also be responsible for the enhanced out-of plane magnetic anisotropy for CFO/BTFO films.^19,23 Therefore, the out-of-plane magnetic anisotropy is reasonable.

To investigate the magnetoelectric response of CFO/BTFO composite films, one then addresses the magneto-dielectric properties as the in-plane and out-of-plane magnetic field are applied to CFO/BTFO composite films. These results are showed in Fig. 5. Fig. 5(a) describes εⁱⁿ_r as the relative dielectric constant with in-plane magnetic field and ε^out_r as that with out-of-plane magnetic field. It is shown that composite films exhibit relative high values ε_r, conforming well dielectric properties. When the applied magnetic field increases from 0 to 8 kOe, the value of ε_rⁱⁿ decreases from 1746 to 1732, while ε_r^out decreases to 1717. Thus results suggest that the magnetic ordering suppresses the dielectric constant. This effect is defined as magneto-dielectric effect, which is expressed by the magneto-dielectric constant MDC and magneto-dielectric loss MDL as following equations,

Fig. 5 (a) The magneto-dielectric constants as the in-plane and out-of-plane magnetic field are applied to CFO/BTFO composite films, (b) magneto-dielectric constant MDC dependant on the in-plane and out-of-plane magnetic field for CFO/BTFO composite films, (c) magneto-dielectric loss MDL dependant on the in-plane and out-of-plane magnetic field for CFO/BTFO composite films.

Fig. 5(b) further illustrates the MDC dependant on the applied bias magnetic field. The magneto-dielectric effect is obtained with 1.6% in perpendicular direction and 0.8% in parallel direction, respectively. The results suggest that the coupling interaction between the magnetic field and dipoles takes place. Thus magneto-dielectric effect is expected to be an intrinsic effect, which also demonstrates that the spin reorientation can affect the interaction between dipoles.²⁴ Besides, the dielectric loss also changes with the external magnetic field, as shown in Fig. 5(c). The dielectric loss increases with the applied magnetic field, reaching the maximum value of 0.7% with the applied magnetic field perpendicular to the film plane and 0.5% with the applied magnetic field parallel to the film plane. Interestingly, the dielectric constant and loss all show anisotropy under the action of the magnetic field. Thus anisotropic magneto-dielectric effect originates from the magnetic anisotropy of CoFe₂O₄/Bi₅Ti₃FeO₁₅ composite films due to the magnetoelectric coupling interaction. To be specific, since the magnetic properties of composite films have obviously anisotropy, CFO layers would possess different magnetoelectriction at in-plane and out-of-plane magnetic field. Then the stresses transferred by interface between CFO lay and BTFO layer are different. The different stresses would result in the different polarization for BTFO piezoelectric film layer. Therefore, the anisotropic dielectric properties are reasonable. These suggest the magnetoelectric coupling interaction between CFO film layer and BTFO film layer.

The in-plane and out-of-plane magnetoelectric coupling coefficient α_E31, α_E33 curves versus external magnetic for CFO/BTFO composite films are shown in Fig. 6. As an interfacial magnetoelectric coupling between the CFO film layers and BTFO film layers. By considering the interface strain transfer efficiency k = 1 the in-plane and out-of-plane magnetoelectric voltage coefficient α_E33 and α_E31 can be estimated by the equations,²⁵

where υ = ^pυ(^pυ + ^mυ) and ^pυ and ^mυ denote the volume of piezoelectric phase and magnetostrictive phase, respectively, m and p denotes magnetostrictive and piezoelectric phase. And s, ε, d, q denotes elastic compliance constant, dielectric constant, permeability, piezoelectric coefficient and piezomagnetic coefficient, respectively. As piezoelectric phase and magnetostrictive phase perfect lattice matching, two phases formed well-defined coupling, as demonstrated by the largest α_E31 and α_E33 of 57 mV cm⁻¹ Oe⁻¹ and 80 mV cm⁻¹ Oe⁻¹, which are larger than the other BTFO based composite films.¹⁰

Fig. 6 In-plane and out-of-plane magnetoelectric effect coefficient α_E as a function of bias magnetic field for CFO/BTFO composite films.

It is well known that the magnetoelectric coupling is derived from the interaction between typical ferroelectric and ferromagnetic phases through the magnetic-mechanical-electric transfer across the interface. According to the above equation discussed, the magnetoelectric effect is the product effect by combining with piezoelectric effect of ferroelectric phase and magnetostriction effect of ferromagnetic phase, which directly affects the induced out-put voltage. Therefore, the interface coupling efficiency, piezoelectric coefficient and piezomagnetic coefficient directly decides the magnetoelectric effect. To be specific, for the present composite films, nearly ideal interface strain transfer derived by well lattice matching between CFO and BTFO phase, which was confirmed by the XRD patterns and cross-sectional SEM images. Thus well interface coupling is expected to be obtained. And a larger piezoelectric coefficient d₃₃ have been measured by PFM, meaning the larger d₃₁ since it is considered d₃₃ ≈ 2d₃₁.²⁶ Thus the larger magnetoelectric voltage coefficient proportional to d₃₁ is understandable. At the mean time, single-phase BTFO films exhibit simultaneously magnetoelectric effect derived from spin-driven coupling in multiferroic system,¹⁶ which also attributes to the strong magnetoelectric effect.

Interestingly, the obtained curves present obviously magnetoelectric anisotropy suggested by the different values of α_E31 and α_E33, which originates from the piezomagnetic anisotropy of CFO ferromagnetic layer. To be specific, according to the eqn (5) and (6), there is an approximately linear relationship between the magnetoelectric voltage coefficient α_E and the piezomagnetic coefficient q. While piezomagnetic coefficient q is expressed by q = δλ/δH_bias, where λ is the magnetostrictive coefficient. That is to say, magnetoelectric voltage coefficient can directly be influenced by the q values. For CFO materials, there are obvious differences between the in-plane piezomagnetic coefficient q₁₁ and out-of-plane piezomagnetic coefficient q₁₃. The maximum value of q₁₁ is 17 × 10⁻¹² Oe⁻¹, while the maximum value of q₁₃ is only 5.8 × 10⁻¹² Oe⁻¹.²⁷ The former is almost three times larger than the latter. And it shows obvious anisotropy for piezomagnetic coefficient vs. magnetic field. It is just the obvious magnetic anisotropy in piezomagnetic coefficient that leads to the magnetoelectric anisotropy.

Conclusions

In summary, CFO/BTFO composite heterostructural films have been prepared by chemical solution deposition method. The phase structure, electric properties, magnetic properties and magnetoelectric coupling effect for CFO/BFTO composite films were systematically investigated. The composite films display ferroelectric, leakage and piezoelectric properties. The existence of the MDC effect proves that it occurs the coupling interaction between the magnetic and electric orders, which confirms the multiferroic behavior of composite films. What's more a strong magnetoelectric effect and its anisotropy are observed for CFO/BFTO composite films. Corresponding in-plane magnetoelectric voltage coefficient reaches 57 mV cm⁻¹ Oe⁻¹ and out-of-plane magnetoelectric voltage coefficient reaches 78.9 mV cm⁻¹ Oe⁻¹. Thus magneto-dielectric and magnetoelectric anisotropies originate from the piezomagnetic anisotropy of CFO layer for the composite films. The present work opens an ideal avenue to enhance the magnetoelectric effect of the piezomagnetic/piezoelectric composite heterostructural films, which facilitates their applications in MEMS devices.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (Grant no. 11264026, 11564028, 51202103), and Inner Mongolia Science Foundation for Distinguished Young Scholars (Grant no. 2014JQ01), the Youth Science and Technology Talents Foundation of Inner Mongolia (Grant no. NJYT-B02).

References

1. T. Nan, Y. Hui, M. Rinaldi and N. X. Sun, Self-Biased 215 MHz Magnetoelectric NEMS Resonator for Ultra-Sensitive DC Magnetic Field Detection, Sci. Rep., 2013, 3, 1985.
2. E. Lage, C. Kirchhof, V. Hrkac, L. Kienle, R. Jahns, R. Knochel, E. Quandt and D. Meyners, Exchange biasing of magnetoelectric composites, Nat. Mater., 2012, 11, 523.
3. Y. Zhang, Z. Li, C. Deng, J. Ma, Y. Lin and C.-W. Nan, Demonstration of magnetoelectric read head of multiferroic heterostructures, Appl. Phys. Lett., 2008, 92, 152510.
4. H. B. Yang, G. Zhang and Y. Lin, Electrical, magnetic and magnetoelectric properties of laminated 0.65 BiFeO₃–0.35 BaTiO₃/BiY₂Fe₅O₁₂ composites, Smart Mater. Struct., 2015, 24, 065028.
5. D. M. Kim and C. B. Eom, Thickness dependence of structural and piezoelectric properties of epitaxial PbZr_0.52Ti_0.48O₃ films on Si and SrTiO₃ substrates, Appl. Phys. Lett., 2006, 88, 142904.
6. S. Roy, R. Chatterjee and S. B. Majumder, Magnetoelectric coupling in sol–gel synthesized dilute magnetostrictive–piezoelectric composite thin films, J. Appl. Phys., 2011, 110, 036101.
7. Y. L. Bai, J. Y. Chen, R. N. Tian and S. F. Zhao, Enhanced multiferroic and magnetoelectric properties of Ho, Mn co-doped Bi₅Ti₃FeO₁₅ films, Mater. Lett., 2016, 164, 618.
8. C. W Nan, M. I. Bichurin, S. X. Dong, D. Viehland and G. Srinivasan, Multiferroic magnetoelectric composites: historical perspective, status, and future directions, J. Appl. Phys., 2008, 103, 031101.
9. N. Adhlakha, K. L. Yadav and R. Singh, BiFeO₃–CoFe₂O₄–PbTiO₃ composites: structural, multiferroic, and optical characteristics, J. Mater. Sci., 2015, 50, 2073.
10. J. Y. Chen, Y. L. Bai, C. H. Nie and S. F. Zhao, Strong magnetoelectric effect of Bi₄Ti₃O₁₂/Bi₅Ti₃FeO₁₅ composite films, J. Alloys Compd., 2016, 663, 480.
11. Y. D. Qiu, S. F. Zhao and Z. P. Wang, Magnetoelectric effect of Dy doped Bi₅Ti₃FeO₁₅, Mater. Lett., 2016, 170, 89.
12. H. Sun, X. M. Lu, J. Su, T. T. Xu, C. C. Ju, F. Z. Huang and J. S. Zhu, Multiferroic behaviour and the magneto-dielectric effect in Bi₅Ti₃FeO₁₅ thin films, J. Phys. D: Appl. Phys., 2012, 45, 385001.
13. B. Nagaraj, S. Aggarwal, T. K. Song, T. Sawhney and R. Ramesh, Leakage current mechanisms in lead-based thin-film ferroelectric capacitors, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 16022.
14. W. Y. Xing, Y. N. N. Ma, Z. Ma, Y. L. Bai, J. Y. Chen and S. F. Zhao, Improved ferroelectric and leakage current properties of Er-doped BiFeO₃ thin films derived from structural transformation, Smart Mater. Struct., 2014, 23, 085030.
15. P. Kumar, S. K. Sharm, M. Knobel and M. Singh, Effect of La³⁺ doping on the electric, dielectric and magnetic properties of cobalt ferrite processed by co-precipitation technique, J. Alloys Compd., 2010, 508, 115.
16. K. X. Jin, B. Yang, Y. Zhang, B. C. Luo, L. Y. Chen and C. L. Chen, Tunable dielectric properties induced by optical fields in barium strontium titanate/manganite heterostructures, Scr. Mater., 2016, 112, 62.
17. H. Y. Zhao, H. Kimura, Z. X. Cheng, M. Osada, J. L. Wang, X. L. Wang, S. X. Dou, Y. Liu, J. D. Yu, T. Matsumoto, T. Tohei, N. Shibata and Y. Ikuhara, Large magnetoelectric coupling in magnetically short-range ordered Bi₅Ti₃FeO₁₅ film, Sci. Rep., 2014, 4, 5255.
18. Q. Wang, J. Chen, L. L. Fan, L. J. Liu, L. Fang and X. R. Xing, Preparation and Electric Properties of Bi_0.5Na_0.5TiO₃–Bi(Mg_0.5Ti_0.5)O₃ Lead-Free Piezoceramics, J. Am. Ceram. Soc., 2013, 96, 1171.
19. W. R. Zhang, J. Jian, A. P. Chen, L. Jiao, F. Khatkhatay, L. G. Li, F. Chu, Q. X. Jia, J. L. MacManus-Driscoll and H. Y. Wang, Strain relaxation and enhanced perpendicular magnetic anisotropy in BiFeO₃:CoFe₂O₄ vertically aligned nanocomposite thin films, Appl. Phys. Lett., 2014, 104, 062402.
20. A. Chen, N. Poudyal, J. Xiong, J. P. Liu and Q. X. Jia, Modification of structure and magnetic anisotropy of epitaxial CoFe₂O₄ films by hydrogen reduction, Appl. Phys. Lett., 2015, 106, 111907.
21. J. Dubowik, Phys. Shape anisotropy of magnetic heterostructures, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 1088.
22. H. Zheng, J. Kreisel, Y.-H. Chu, R. Ramesh and L. Salamanca-Riba, Heteroepitaxially enhanced magnetic anisotropy in BaTiO₃–CoFe₂O₄ nanostructures, Appl. Phys. Lett., 2007, 90, 113113.
23. D. Y. Qiu, K. Ashraf and S. Salahuddin, Nature of magnetic domains in an exchange coupled BiFeO₃/CoFe heterostructure, Appl. Phys. Lett., 2013, 102, 112902.
24. H. Sun, X. M. Lu, T. T Xu, J. Su, Y. M. Jin, C. C. Ju, F. Z. Huang and J. S. Zhu, Study of multiferroic properties in Bi₅Fe_0.5Co_0.5Ti₃O₁₅ thin films, J. Appl. Phys., 2012, 111, 124116.
25. M. I. Bichurin, V. M. Petrov and G. Srinivasan, Theory of low-frequency magnetoelectric effects in ferromagnetic–ferroelectric layered composites, J. Appl. Phys., 2002, 92, 7681.
26. M. Feng, J. J. Wang, J. M. Hu, J. Wang, J. Ma, H. B. Li, Y. Shen, Y. H. Lin, L. Q. Chen and C. W. Nan, Optimizing direct magnetoelectric coupling in Pb(Zr,Ti)O₃/Ni multiferroic film heterostructures, Appl. Phys. Lett., 2015, 106, 072901.
27. R. V. Krishnaiah, A. Srinivas, S. V. Kamat, T. Karthik and S. Asthana, Effect of CoFe₂O₄ mole percentage on multiferroic and magnetoelectric properties of Na_0.5Bi_0.5TiO₃/CoFe₂O₄ particulate composite, Ceram. Int., 2014, 40, 7799.

---