Introduction

The magnetoelectric coupling between the magnetic and electric orders in multiferroics brings about novel physical mechanism and offers the opportunity for attractive applications, such as non-volatile memory technology, four-state logic device and magnetoelectric sensors1,2,3,4,5,6,7. Of particular interest is electric control of magnetism, which has a promising potential for reducing power consumption and realizing device miniaturization in spintronic devices8,9. However, such applications have been limited by weak magnetoelectric effects and low operating temperature in most discovered single-phase multiferroics. Very recently, giant magnetoelectric effects have been reported in some hexaferrites, in which the electric polarization (P) can be induced by their complex magnetic orders10,11,12,13,14.

Hexaferrites, which have long been utilized in permanent magnets and microwave devices, can be classified into different types such as M, W, X, Y, U and Z15. Their crystal structures can be described as stacked sequences of the basic blocks: S, R and T15. Recently, some significant results about the magnetically induced ferroelectricity and resultant magnetoelectric effects were observed in some Y-, Z- and M-type hexaferrites due to the modifications of spiral magnetic structures by an external magnetic field10,11,12,13,14. This phenomena can be understood by the spin current or the inverse Dzyaloshinskii-Moriya models described by the equation

( denotes spins at neighboring sites i and j, denotes the propagation vector of a spiral structure, A denotes the coupling coefficient determined by the exchange interaction and the spin-orbit interaction.), in which the interaction between noncollinear spins can produce the local P through the spin-orbit interaction16,17,18.

Recently, Z-type hexaferrite Sr3Co2Fe24O41 was reported to show an excellent room-temperature magnetoelectric property at low magnetic field (H), which would shed light on the practical applications of magnetoelectric multiferroics13. Since the electric resistivity in most of these hexaferrites is not high enough to support an electric field, the manipulation of magnetization by an electric field has rarely been reported, especially at room temperature8. Here, we report the discovery of electric control of magnetism in an M-type hexaferrite, SrCo2Ti2Fe8O19 (SCTFO), at room temperature. Furthermore, without magnetic bias field, large effect of electric field-controlled magnetization is observed at room temperature, which would revolutionize the field of low-power magnetoelectronic devices.

Results

The M-type hexaferrites, AFe12O19 (A = Ba, Sr, Pb, etc.), crystallize in the magnetoplumbite-type structure (the space group P63/mmc) with Fe3+ ions in octahedral (12k, 4f and 2a), tetrahedral (4f) and bipyramidal (2b) sites, respectively. This lattice structure could be symbolically described as an alternate stacking of 4 building blocks, namely S, S*, R and R*, which is schematically shown elsewhere13,15. The large A2+ ion, which causes a slight perturbation in the lattice around itself, is responsible for the large magnetic uniaxial anisotropy in these hexaferrites, where the crystallographic c-axis is the magnetic easy axis15. The substituted M-type hexaferrites have been extensively investigated for the application point of view15,19. Among them, the substitution of Fe3+ cations with Co2+-Ti4+ pair is reported to alter the magnetocrystalline anisotropy and give rise to the noncollinear magnetic structures20,21. In this case, Co2+ cations prefer to occupy the octahedrally-coordinated 4f and 12k Wyckoff positions and the magnetic anisotropy changes gradually from an axial anisotropy (along the hexagonal c axis) to a planar anisotropy (a, b plane of the hexagonal cell) with increasing concentration of Co2+15,21. On the other hand, nonmagnetic Ti4+ cations substituted on the 12k would interrupt the magnetic interactions in a way that the two blocks RS and R*S* become more or less magnetically decoupled20. As a result, an intermediate conical magnetic structure was obtained in the M-type hexaferrite with critical Co-Ti concentration, which can be verified by neutron diffraction and Mössbauer spectroscopy20,22.

Figure 1 shows the room-temperature powder X-ray diffraction pattern and Rietveld refinement of polycrystalline SCTFO. All peaks can be indexed as the magnetoplumbite-type cell in the space group P63/mmc (a = 5.881(1) Å, b = 5.881(1) Å, c = 23.102(8) Å, α = 90.00°, β = 90.00°, γ = 120.00°). Table I summarizes the fractional coordinates and occupation factors.

Table 1 The fractional coordinates and occupation factors for polycrystalline SCTFO.
Full size table
Figure 1
figure 1

The room-temperature powder X-ray diffraction pattern and Rietveld refinement for polycrystalline SCTFO.

Full size image

The magnetic hysteresis loops for SCTFO was measured at 300 K. As shown in Fig. 2(a), the sample is ferromagnetic at room temperature with the magnetization of about 30 emu/g with H = 15 kOe. Comparing with SrFe12O19, the coercivity of SCTFO decrease remarkably (see in the inset of Fig. 2(a)), which is resulted from the reduction in anisotropy field with the substitution of Co2+-Ti4+ for Fe3+19. The temperature dependence of magnetization (M-T) from 100–800 K for SCTFO is shown in Fig. 2(b). The Curie temperature of 740 K for this ferrite is less than that of SrFe12O1923, which would be ascribed to the substitution of Fe3+ with Co2+-Ti4+ pair19. Based on the results of neutron diffraction and magnetic measurements, the magnetization anomaly at around 420 K is likely to correspond to the transition from collinear ferrimagnetic phase to the phase with a conical magnetic structure20,22,24.

Figure 2
figure 2

(a) The magnetic hysteresis loop for SCTFO at 300 K; the inset of (a) shows the enlarged magnetic hysteresis loop; (b) The variation of M for the sample as a function of T.

Full size image

Figure 3(a) shows the temperature dependence of P at zero magnetic field, which is obtained from integrating the magnetoelectric current by time with 400 K as the initial point. As we know, the reported Y- or Z-type multiferroics hexaferrites, such as Ba2Mg2Fe12O22 or Sr3Co2Fe24O41, show no spontaneous P at zero magnetic field. When an external magnetic field is applied, a transverse conical spin structure is induced, which could produce P through the spin current and the inverse Dzyaloshinskii-Moriya mechanism11,13,18. However, in the case of SCTFO, it shows spontaneous P at zero magnetic field, indicating that its spin structure has a transverse cycloidal component, which should be further investigated by neutron diffraction measurements. In addition, a sudden decrease of P almost coincides with the anomaly in the M-T curve, suggesting the coupling between the ferromagnetism and ferroelectricity13,14. Figure 3(b) shows the H dependence of P for the sample at 300 K. With increasing H, the P is considerably suppressed, demonstrating a strong magnetoelectric coupling in SCTFO. As shown in Fig. 2(b), P does not vanish even H = 16 T, indicating that the conical spin structure is robust. Accordingly, there is no anomaly in M-H curve (Figure 3(b)), which is consistent with the behavior of P.

Figure 3
figure 3

(a) The temperature dependence of P (red) and M (black) for SCTFO. (b) The magnetic initial curve (black) and H dependence of P (red) at 300 K.

Full size image

Comparing with magnetic field, it is easier to generate an electric field in devices and working with a voltage, rather than an electric current, gives rise to less energy consuming in technological applications5,8,9. Therefore, the electric-field controlled ferromagnetism, which is usually defined as converse magnetoelectric effect, is of great importance from the application point of view. However, the investigation of electric control of magnetism has been hampered in many single-phase multiferroics due to their low resistivity8,9,25. In the case of SCTFO, the resistivity reaches up to 4.36×1010 Ω.cm, which is high enough to support an electric field and ensures the following measurement. We first measured the room-temperature magnetic loops of SCTFO with DC electric field (Edc) parallel and perpendicular to the H, respectively. By applying an Edc, an obvious reduction of M is observed and typical loops at zero and 22 kv/cm field are shown in Fig. 4(a). Similar to the effect of H on P, M decreases with increasing Edc and the change of M with Edc = 22 kv/cm reaches up to 6.3% and 2.8%, for Edc⊥H and Edc∥H, respectively, as shown in Fig. 4(b).

Figure 4
figure 4

(a) The room temperature magnetic hysteresis loop under different Edc (zero and 22 kV/cm); (b) The change rate of M under various Edc with magnetic bias field ( H = 460 Oe) at room temperature: E⊥H (Red); E∥H (Black).

Full size image

The variation of M of SCTFO as a function of Edc was measured without magnetic bias field. Before this measurement, the sample was pre-magnetized in a field of 5 kOe. A considerable reduction of M is observed in Fig. 5(a) despite of positive or negative Edc. This result is of interest since no H is utilized in electric control of magnetism. Figure 5(b) shows the time dependence of M under a square wave electric field of 18.5 kV/cm. M decreases or increases with applying or removing Edc, respectively, indicating a stable response to the electric field, which demonstrates the modulation of E on M in SCTFO.

Figure 5
figure 5

(a) The change of M as a function of Edc at room temperature without magnetic bias field; (b) The time dependence of M under a square wave electric field.

Full size image

In order to further investigate the effect of E on M, the scanning probe microscopy (SPM) experiment was performed using lift mode. Figure 6(a) is the topography of a selected area, while Fig. 6(b)–6(d) show the magnetic force microscopy (MFM) images for the same scanning area, which reveal magnetic contrast between the surface magnetization domains. During the MFM studies, various electric fields were applied on the sample. In this case, the bias electric field would make the magnetic force signal shift uniformly, but would not affect the contrast of the signal. From Fig. 6, no change in domain structure can be detected, but the magnetic contrast fades out as electric voltage increasing from 0 to 10 V, indicating the effect of E on the magnetic domain. This experimental result is significant since it gives a direct and powerful evidence of electric control of magnetism at room temperature.

Figure 6
figure 6

Topography (a) and MFM images of SCTFO at a selected area under different poling voltage: (b) V = 0V; (c) V = 6V; (d) V = 10V.

Full size image

Discussion

In the single-phase SCTFO, the doping of Co2+ is propitious to change the magnetic anisotropy from axial to planar, while the substitution of Ti4+ interrupts the magnetic interactions between the two blocks RS and R*S*15,20,21. As a result, the magnetic frustration at the boundary stabilizes the nonlinear spiral spin structure, inducing the spontaneous ferroelectric polarization at room temperature, which is attributed to the spin current or the inverse Dzyaloshinskii-Moriya models. In present sample, the substitution of a Co2+-Ti4+ pair can maintain the average 3+ valence state and then suppress the hopping of electrons between Fe2+ and Fe3+ in SCTFO. In addition, sintering in oxygen is also helpful to increase the resistivity. As a result, SCTFO has a high enough resistivity which ensures the magnetoelectric measurements. As shown in Fig. 3(b), the values of P decrease monotonously with the increasing magnetic field. According to Eq. (1), the spin cone in SCTFO tends to close with a magnetic field, leading to the decrease of P correspondingly14. As for the effect of E on M, the values of M decreases with increasing E, which is obviously shown in Fig. 4, 5 and 6. When an electric field is applied on the sample, the electric polarization would intrinsically increase. According to Eq. 1, the value of would increase and lead to the enlargement of the spin cone angle correspondingly, which results in the reduction of magnetization in our sample. Since the electric control of magnetism is of significance, the effect of E on M without magnetic bias field is emphasized in present work. As shown in Fig. 5, the remanent magnetization of SCTFO can be reversibly and reproducibly controlled by an electric field, suggesting the potential application in the non-volatileswelling RAM with multiferroic hexaferrites.

Thus we have prepared a single-phase multiferroics SCTFO. The substitution of Co2+-Ti4+ pair for two Fe3+ cations produces a spiral conical magnetic structure and high resistivity in SCTFO. This special spin structure can induce spontaneous P via spin-orbit coupling, suggesting the strong magnetoelectric coupling in SCTFO. Electric control of magnetism is confirmed by magnetization and MFM measurements. The excellent magnetoelectric performance in SCTFO gives a valuable contribution to the study of magnetoelectric effect and enables the realization of electric control magnetization in magnetoelectric devices.

Methods

Polycrystalline sample of SrCo2Ti2Fe8O19 was prepared through conventional solid-state reaction. The high-purity powders of SrCO3, Fe2O3, Co2O3 and TiO2 were mixed stoichiometrically and sintered at 1,200°C for 10 hours for twice in the oxygen atmosphere. The powder X-ray diffraction measurement was carried out to identify the sample as single phase. For the electric polarization measurements, a specimen was cut with a dimension of 3×3×1 mm3 and gold electrodes were painted on both sides as a capacitance. The temperature or magnetic field dependence of P was obtained by integrating the magnetoelectric current, where temperature and applied H were controlled by the physical property measurement system (PPMS, Quantum Design) and magnetoelectric current was gathered by an electrometer (model 6514, Keithley). The magnetic properties of SrCo2Ti2Fe8O19 were measured with a commercial vibrating sample magnetometer (VSM, model 7407, Lakeshore). The effect of Edc on M was measured with a VSM under zero magnetic field at room temperature. During measurement, Edc was applied on the capacitance and the induced M was detected by pick up coils. The MFM measurements were carried out on the polished sample by scanning probe microscopy (SPM) (Dimension V, Veeco). In order to investigate the effect of E on the magnetic domain, various voltages were applied between the probe tip and the sample chuck. During the MFM measuring processes, two different modes, “grounding the tip” and “applying the voltage on the tip”, are used.