Nanoscale
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
COMMUNICATION
Cite this: Nanoscale, 2014, 6, 8515
Received 9th April 2014 Accepted 28th May 2014
View Article Online View Journal | View Issue
Magnetic-field-induced ferroelectric polarization reversal in magnetoelectric composites revealed by piezoresponse force microscopy Hongchen Miao,a Xilong Zhou,a Shuxiang Dong,b Haosu Luoc and Faxin Li*a
DOI: 10.1039/c4nr01910e www.rsc.org/nanoscale
Controlling electric polarization (or magnetization) in multiferroic materials with external magnetic fields (or electric fields) is very important for fundamental physics and spintronic devices. Although there has been some progress on magnetic-field-induced polarization reversal in single-phase multiferroics, such behavior has so far never been realized in composites. Here we show that it is possible to reverse ferroelectric polarization using magnetic fields in a bilayer Terfenol-D/PMN-33%PT composite. We realized this by ferroelectric domain imaging using piezoresponse force microscopy (PFM) under applied magnetic field loading. The internal electric field caused by the magnetoelectric (ME) effect in the PMN-PT crystal is considered as the driving force for the 180 polarization switching, and its existence is verified by switching spectroscopy PFM testing under a series of external magnetic fields. A quantitative method is further suggested to estimate the local ME coefficient based on the switching spectroscopy PFM testing results.
Multiferroic materials have attracted more and more attention in recent years due to their potential applications as multidimensional memory devices, ultrasensitive magnetic eld sensors, etc.1 A strong impetus is that multiferroics may be promising candidates to solve the problem of large energy dissipation due to high current densities in spintronic devices,2,3 in which controlling the magnetization with external electric elds is very important, with some progress having been achieved up to now.4 For example, 180 reversal of magnetization induced by pure electric eld loading has been realized in the CoFe/BFO heterostructure at room temperature.5 Lahtinen et al. have shown the possibility of writing and erasing regular ferromagnetic domain patterns in magnetoelectric (ME) composites using small electric elds.6 In comparison, a
LTCS and Department of Mechanics and Engineering Science, College of Engineering, Peking University, Beijing, 100871, China. E-mail: [email protected]
b
Department of Material Science and Engineering, College of Engineering, Peking University, China
c Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai, 200050, China
This journal is © The Royal Society of Chemistry 2014
controlling the electric polarization with magnetic elds has been realized only in single-phase ME materials and complete reversal of ferroelectric polarization is rarely seen. Hur et al. showed the possibility of electric polarization reversal induced by magnetic elds in the multiferroic TbMn2O5.7 Yamasaki et al.8 and Shen et al.9 realized the reversal of spontaneous polarization by a magnetic eld in a multiferroic spinel oxide and a Ba1.3Sr0.7Co0.9Zn1.1Fe10.8Al1.2O22 single crystal, respectively. In the studies mentioned above,7–9 the need for lowtemperature environments limits the applications in practice. Keeney et al. demonstrated that switching of ferroelectric polarization by magnetic elds was possible at room temperature in multiferroic Aurivillius phase thin lms.10 Evans et al. also realized switching of ferroelectric domains by magnetic elds at room temperature in a multiferroic PZTFT.11 However, the weak ME effect in single-phase multiferroics is always a big challenge for their practical applications. Although ME composites have been proved to provide strong ME coupling, to date ferroelectric polarization reversal induced by magnetic elds has never been observed in this type of material. One may regard a magnetic eld as not being a viable mechanism to invert the ferroelectric polarization in ME composites, since the coupling between ferroelectricity and ferromagnetism is based on the interfacial strain and thus only non-180 domain switching was thought to be induced by elastic deformation. Currently challenges still exist in measuring the ME coefficient at the nanoscale. Bai et al. provided a quantitative estimation of the local ME coefficient based on the phase shi in magnetic force microscopy (MFM) measurements, while it is usually difficult to measure the phase shi accurately.12 Xie and coworkers obtained a quantitative estimation of the local ME coefficient based on piezoresponse force microscopy (PFM) measurements.13 Caruntu et al. suggested another estimation method also based on PFM measurements.14 However, in their methods, one is required to get accurate piezoelectric coefficients in PFM testing, which is rather difficult or almost impossible at present due to the challenges in determining the exact applied electric eld distribution and eliminating the
Nanoscale, 2014, 6, 8515–8520 | 8515
View Article Online
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
Nanoscale
background effect of the scanning probe microscopy (SPM) setup.15,16 In this work, we demonstrate through PFM testing that it is possible to invert ferroelectric polarization via magnetic elds in a bilayer Terfenol-D/PMN-33%PT composite. The internal electric eld caused by ME coupling is thought to be the driving force for the observed 180 polarization switching. Furthermore, we provide a quantitative estimation of the ME coefficient based on switching spectroscopy PFM testing under different applied magnetic elds. The bilayer Terfenol-D/PMN-PT composite specimen consists of one 12 mm 6 mm 1 mm Terfenol-D layer magnetized along its longitudinal directions and one 5 mm 5 mm 0.4 mm PMN-33%PT rhombohedral single-crystal layer. The PMN-33%PT single crystal is oriented along the h111i direction and poled along the thickness direction. The two layers were carefully bonded together using epoxy resin (West System 206/105, USA) and cured for more than 24 h at room temperature. Here it should be noted that the thickness of the PMN-PT layer is very important in realizing the magnetic-eldinduced polarization reversal. Previously we tried a similar Terfenol-D/PMN-PT specimen with a PMN-PT layer thickness of 0.1 mm and did not observe the magnetic-eld-induced polarization reversal. The experimental setup is schematically shown in Fig. 1(a) where the external magnetic eld was provided by a variable eld module (VFM) in an atomic force microscope (Asylum Research MFP-3D). The VFM can provide a dc magnetic eld with a magnitude of 2250 Oe and a resolution of 1 Oe. Before testing, the Terfenol-D layer only was used to examine its magnetostriction caused by the magnetic eld, and the strain parallel to the applied magnetic eld is shown in Fig. 1(b). It can be seen that the magnetic eld provided by the VFM is fairly strong and can cause a longitudinal magnetostrictive strain of 600–700 ppm in the Terfenol-D layer under a eld of 2250 Oe, which is slightly larger than that of 500–600 ppm in Xie et al.'s work under a eld of 2000 Oe.17 In the setup shown in Fig. 1(a), PFM was used to image the ferroelectric domain patterns and
Communication
characterize polarization switching in the PMN-PT crystal under external magnetic elds. During PFM testing, a 3 V ac voltage was applied to the sample through a conductive probe with a spring constant of 3 N m1 (Budget Sensors, Multi75E-G). The dual frequency resonance tracking technique was used to get a high signal-to-noise ratio. We rstly prefabricated an engineered 180 domain pattern (the logo of our lab) on the PMN-PT layer using lithography PFM, as shown in Fig. 2(a)–(c). The approximate 180 phase contrast shown in Fig. 2(b) and (c) indicates that the reversed 180 domains are well written. Here for simplication, we denote the original domains as +c domains and the reversed domains as c domains. When an in-plane magnetic eld of 2250 Oe is applied to the sample along the magnetization direction of the Terfenol-D, the 180 phase contrast disappears, as shown in Fig. 2(e) and (f), which clearly indicates that the prefabricated c domains had experienced polarization switching. As is known, the rhombohedral single-crystal PMN33%PT has eight equivalent spontaneous polarization directions and only 180 , 109 and 71 domain switching can occur. In PFM, a 180 phase contrast means the angle between the two polarization directions is an obtuse angle and a vanishing phase contrast implies that the angle is an acute angle or zero.18 Thus, the angle between the polarization directions in Fig. 2(d) could only be 0 or 71 . If the angle were 71 , the PFM amplitudes of the engineered domains should be considerably smaller than those of the +c domains. However, Fig. 2(d) shows that the amplitudes of the engineered domains are slightly larger than those of the +c domains, which implies that the engineered domains in Fig. 2(d) and (e) are also +c domains. Therefore, when applying the magnetic eld of 2250 Oe, the engineered c domains in Fig. 2(a) and (b) had undergone 180 switching to +c domains in Fig. 2(d) and (e). In other words, magnetic-eldinduced ferroelectric polarization reversal had been realized in the ME composite. As to the PFM amplitude, the average amplitude in Fig. 2(a) is about 150 pm in the absence of the external magnetic eld, and it increases rapidly to about 1 nm when the magnetic eld
Fig. 1 (a) Schematic of the setup for PFM testing of the Terfenol-D/PMN-PT composite under external magnetic field loading. Arrows P and M denote the poling direction and the magnetized direction, respectively. The yellow cuboids are rare-earth permanent magnet in a variable field module (VFM). By rotating the magnet, different amounts of flux can be channeled through the sample. (b) The magnetostrictive response of the Terfenol-D layer under in-plane magnetic fields from the VFM.
8516 | Nanoscale, 2014, 6, 8515–8520
This journal is © The Royal Society of Chemistry 2014
View Article Online
Nanoscale
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
Communication
Ferroelectric domain evolution in the PMN-PT layer under a series of magnetic fields. Images in the left and middle columns are PFM amplitude images and phase images, respectively. Plots in the right column are the sections of phase images, as indicated by the white dotted lines in the middle column. The left arrow from top to bottom indicates the experimental procedure. (a)–(c) PFM images after PFM lithography; (d)–(f) PFM images under 2250 Oe H field; (g)–(i) PFM images after removing the 2250 Oe H field; (j)–(l) PFM images under 2250 Oe H field; (m)–(o) PFM images after removing the 2250 Oe H field.
Fig. 2
of 2250 Oe is applied, as seen in Fig. 2(d). A similar phenomenon was also observed by Xie et al.17 When the positive magnetic eld is removed, the 180 phase contrast occurs again, as shown Fig. 2(h) and (i). It should be noted that the phase contrast area here is smaller than that in Fig. 2(b), which indicates the domains switched by the magnetic eld are metastable and part of them may return to c domains due to the large depolarization eld aer the transient polarization switching, which is similar to the case in controlling the magnetization with external electric elds where the reversed magnetic domains will return to their original conguration upon removal of the electric eld.19 The average PFM amplitude
This journal is © The Royal Society of Chemistry 2014
in Fig. 2(g) is about 350 pm, smaller than that in Fig. 2(d) but still larger than that in Fig. 2(a). When an anti-parallel magnetic eld of 2250 Oe is applied, the 180 phase contrast disappears again, as shown in Fig. 2(k) and (l). This indicates that those engineered domains had reversed again to +c domains under the negative magnetic eld. Thus the ferroelectric polarization switching in the ME composite is insensitive to the sign of the external magnetic eld, which is reasonable because the magnetostriction in Terfenol-D is also insensitive to the sign of the magnetic eld. Upon removal of the negative magnetic eld, Fig. 2(n) and (o) show that only very small areas exhibit 180 phase contrast, which indicates that few domains recover to c
Nanoscale, 2014, 6, 8515–8520 | 8517
View Article Online
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
Nanoscale
domains and the reversed +c domains induced by the magnetic eld become more stable. The mechanism of the observed magnetic-eld-induced polarization reversal is illustrated in Fig. 3. When an external magnetic eld is applied, the Terfenol-D layer elongates and the deformation is transferred to the PMN-PT layer through the interface. The transferred elongation deformation in the PMNPT layer is apt to depole the crystal and thus an internal electric eld (Ei) along the original poling direction is generated. Actually, this is the principle of ME coupling in ME composites. The magnitude of Ei can be estimated as follows. The macroscopic ME eld coefficient of the Terfenol-D/PMN-33%PT laminates is typically about 1 V cm1 Oe1. Thus when a magnetic eld of 2250 Oe (or 2250 Oe) is applied, a strong internal eld of 2250 V cm1 will be generated, which is close to the coercive eld (about 2 kV cm1) of the PMN-33%PT single crystal.20 Furthermore, it should be noted that the c domains in PMN-PT written by lithography PFM are metastable because the surrounding domains are all +c domains. They are more active and relatively easier to switch. Therefore, the generated internal eld Ei can activate polarization reversal of the c domains. This polarization switching mechanism is similar to the case in ferroelectric crystals aer heat treatment where the internal electric eld due to the pyroelectric effect could cause back switching of the engineered domains.21 Moreover, the internal electric eld induced by a magnetic eld can enhance the +c domains, and thus the PFM amplitude increases when a large magnetic eld is applied or just removed, as seen in Fig. 2(d) and (g) and also in previous experiments.17 It should be noted that the observed magnetic-eld-induced polarization reversal actually results from the competition between the generated internal electric eld and the elongation deformation. If the PMN-PT layer was very thin (say less than 0.2 mm), the elongation deformation could become dominant and the magneticeld-induced polarization reversal may not occur. The existence of the internal eld Ei due to the ME effect can be proved by local polarization switching characterization under a series of magnetic elds using switching spectroscopy PFM (SS-PFM).22,23 Fig. 4(a) and (b) show the amplitude buttery curves and phase hysteresis loops of the PMN-PT layer under
Illustration of the internal electric field in PMN-PT layer generated by an external magnetic field due to the ME effect. (a) ME composite with engineered domains in the absence of an external magnetic field: P denotes the original polarization, Pe denotes polarization of the engineered domain, M denotes the magnetization direction. (b) When an external magnetic field H is applied, an internal electric field Ei is generated in the PMN-PT layer due to the ME effect. Fig. 3
8518 | Nanoscale, 2014, 6, 8515–8520
Communication
different magnetic elds of 0 Oe, 500 Oe and 1500 Oe. It can be seen that the amplitude buttery curve is symmetric in the absence of the external magnetic eld, i.e., the positive coercive voltage (3.17 V) and the negative coercive voltage (3.24 V) are approximately equal. Moreover, the amplitudes for the positive dc eld and for the negative dc eld are almost the same. However, when an external magnetic eld of 500 Oe is applied to the sample, both the amplitude curve and the phase loop become asymmetric, with a positive coercive voltage of 2 V and a negative coercive voltage of 4.35 V. Meanwhile, the amplitude for the positive electric eld is considerably larger than that for the negative electric eld. On further increasing the magnetic eld to 1500 Oe, the amplitude loops become more strongly asymmetric. The positive coercive voltages decrease with increasing magnetic eld, from 3.17 V to 2 V to 1.19 V, and the negative coercive voltages increase with increasing magnetic eld, from 3.24 V to 4.35 V to 4.75 V. These results thus provide direct evidence for the existence of the internal electric eld generated by the ME effect, which is similar to the case in ferroelectrics in that the asymmetric strain buttery curves and D–E hysteresis loops are caused by the internal bias eld.24 Fig. 4(c) shows how the internal eld Ei inuences the local polarization switching in SS-PFM testing. Bearing in mind that the internal eld Ei is always along the original poling direction, a larger negative external electric eld is required to reverse the positive local polarization, leading to a larger negative coercive eld. On the contrary, a smaller positive external eld can make the negative local polarization reverse, i.e., the positive coercive eld is smaller. Since the internal electric eld caused by external magnetic elds via the ME effect is the key reason for the asymmetric local polarization switching curves, it can then be estimated from the positive and negative coercive elds. Obviously, the local ME coefficient can also be estimated simultaneously. As we know, the ME eld coefficient is dened as aE ¼
DE DH
(1)
where DH is the increment of the external magnetic eld and DE is the increment of the internal electric eld caused by external magnetic elds. Under the applied magnetic eld of 500 Oe, the offset between the positive and negative coercive voltage is 2.35 V. Here it should be noted that the electric eld near the AFM tip is highly concentrated and the actual coercive eld should be much larger than the nominal coercive eld. Taking the macroscopic coercive eld value of 2 kV cm1 for PMN-33%PT and a nanoscale nominal coercive eld of 3.2 V/0.04 cm ¼ 80 V cm1, the average eld concentration factor is thus estimated to be 25. Thus the internal electric eld under a magnetic eld of 500 Oe is estimated to be 2.35 V/2/0.04 cm 25 ¼ 735 V cm1. Then we can obtain the local ME coefficient as 1.47 V cm-1 Oe-1. Similarly, the local ME coefficient under an external magnetic eld of 1500 Oe can be estimated as 0.74 V cm-1 Oe-1. Therefore, the local ME coefficient is not a constant and is strongly dependent on the bias magnetic eld, which is similar to the case at the macroscopic scale.25,26 Furthermore, the estimated local ME coefficient is close to the macroscopically measured
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
Communication
Nanoscale
Fig. 4 Local polarization switching characteristics in the PMN-PT layer under a series of magnetic fields. (a) Amplitude–voltage butterfly curves
in SS-PFM; (b) phase–voltage hysteresis loops in SS-PFM; (c) schematic diagrams of the effect of internal electric field on the local polarization switching behavior: P is the macroscopic polarization, Pl is the local polarization, Ei is the internal electric field caused by the ME effect, Ee is the external electric field applied by an AFM tip, Ec is the apparent coercive field and Ec0 is the original coercive field.
value of about 1 V cm-1 Oe-1, which may indicate the validity of this suggested method. In summary, we have realized magnetic-eld-induced ferroelectric polarization reversal in a Terfenol-D/PMN-PT bilayer composite. The internal electric eld caused by the external magnetic eld due to the ME effect is considered as the driving force for the observed 180 polarization switching. By SS-PFM testing under a series of magnetic elds, the existence of the internal electric eld is veried and a quantitative method is suggested for estimating the local ME coefficient. The results presented in this work may open up a new way to control the electric polarization in ME composites and devices.
Acknowledgements HCM is grateful to Mr Guoxi Liu (College of Engineering, Peking University, China) for the help in material fabrication. Financial support from the National Natural Science Foundation of China under Grant no. 11090331 is also acknowledged.
References 1 J. Ma, J. M. Hu, Z. Li and C. W. Nan, Adv. Mater., 2011, 23, 1062–1087. 2 N. A. Spaldin and M. Fiebig, Science, 2005, 309, 391–392. 3 W. Eerenstein, N. D. Mathur and J. F. Scott, Nature, 2006, 442, 759–765. 4 Y. Tokunaga, Y. Taguchi, T. Arima and Y. Tokura, Nat. Phys., 2012, 8, 838–844. 5 J. T. Heron, M. Trassin, K. Ashraf, M. Gajek, Q. He, S. Y. Yang, D. E. Nikonov, Y. H. Chu, S. Salahuddin and R. Ramesh, Phys. Rev. Lett., 2011, 107, 217202. This journal is © The Royal Society of Chemistry 2014
6 T. H. E. Lahtinen, K. J. A. Franke and S. van Dijken, Sci. Rep., 2012, 2, 258. 7 N. Hur, S. Park, P. A. Sharma, J. S. Ahn, S. Guha and S. W. Cheong, Nature, 2004, 429, 392–395. 8 Y. Yamasaki, S. Miyasaka, Y. Kaneko, J. P. He, T. Arima and Y. Tokura, Phys. Rev. Lett., 2006, 96, 207204. 9 S. P. Shen, L. Q. Yan, Y. S. Chai, J. Z. Cong and Y. Sun, Appl. Phys. Lett., 2014, 104, 032905. 10 L. Keeney, T. Maity, M. Schmidt, A. Amann, N. Deepak, N. Petkov, S. Roy, M. E. Pemble and R. W. Whatmore, J. Am. Ceram. Soc., 2013, 96, 2339–2357. 11 D. M. Evans, A. Schilling, A. Kumar, D. Sanchez, N. Ortega, M. Arredondo, R. S. Katiyar, J. M. Gregg and J. F. Scott, Nat. Commun., 2013, 4, 1038. 12 F. M. Bai, H. W. Zhang, J. F. Li and D. Viehland, J. Phys. D: Appl. Phys., 2010, 43, 285002. 13 S. H. Xie, F. Y. Ma, Y. M. Liu and J. Y. Li, Nanoscale, 2011, 3, 3152–3158. 14 G. Caruntu, A. Yourdkhani, M. Vopsaroiu and G. Srinivasan, Nanoscale, 2012, 4, 3218–3227. 15 T. Jungk, A. Hoffmann and E. Soergel, Appl. Phys. Lett., 2007, 91, 253511. 16 K. Pan, Y. Y. Liu, S. H. Xie, Y. M. Liu and J. Y. Li, Acta Mater., 2013, 61, 7020–7033. 17 S. H. Xie, Y. M. Liu, X. Y. Liu, Q. F. Zhou, K. K. Shung, Y. C. Zhou and J. Y. Li, J. Appl. Phys., 2010, 108, 054108. 18 E. Soergel, J. Phys. D: Appl. Phys., 2011, 44, 464003. 19 T. K. Chung, G. P. Carman and K. P. Mohanchandra, Appl. Phys. Lett., 2008, 92, 112509. 20 Z. Y. Feng, D. Lin, H. S. Luo, S. Li and D. N. Fang, J. Appl. Phys., 2005, 97, 024103.
Nanoscale, 2014, 6, 8515–8520 | 8519
View Article Online
Nanoscale
24 F. X. Li, D. N. Fang and A. K. Soh, Smart Mater. Struct., 2004, 13, 668–675. 25 M. Fiebig, J. Phys. D: Appl. Phys., 2005, 38, R123–R152. 26 G. Srinivasan, Annu. Rev. Mater. Res., 2010, 40, 153–178.
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
21 X. Y. Liu, K. Kitamura and K. Terabe, Appl. Phys. Lett., 2006, 89, 142906. 22 S. Jesse, H. N. Lee and S. V. Kalinin, Rev. Sci. Instrum., 2006, 77, 073702. 23 S. Jesse, A. P. Baddorf and S. V. Kalinin, Appl. Phys. Lett., 2006, 88, 062908.
Communication
8520 | Nanoscale, 2014, 6, 8515–8520
This journal is © The Royal Society of Chemistry 2014
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
COMMUNICATION
Cite this: Nanoscale, 2014, 6, 8515
Received 9th April 2014 Accepted 28th May 2014
View Article Online View Journal | View Issue
Magnetic-field-induced ferroelectric polarization reversal in magnetoelectric composites revealed by piezoresponse force microscopy Hongchen Miao,a Xilong Zhou,a Shuxiang Dong,b Haosu Luoc and Faxin Li*a
DOI: 10.1039/c4nr01910e www.rsc.org/nanoscale
Controlling electric polarization (or magnetization) in multiferroic materials with external magnetic fields (or electric fields) is very important for fundamental physics and spintronic devices. Although there has been some progress on magnetic-field-induced polarization reversal in single-phase multiferroics, such behavior has so far never been realized in composites. Here we show that it is possible to reverse ferroelectric polarization using magnetic fields in a bilayer Terfenol-D/PMN-33%PT composite. We realized this by ferroelectric domain imaging using piezoresponse force microscopy (PFM) under applied magnetic field loading. The internal electric field caused by the magnetoelectric (ME) effect in the PMN-PT crystal is considered as the driving force for the 180 polarization switching, and its existence is verified by switching spectroscopy PFM testing under a series of external magnetic fields. A quantitative method is further suggested to estimate the local ME coefficient based on the switching spectroscopy PFM testing results.
Multiferroic materials have attracted more and more attention in recent years due to their potential applications as multidimensional memory devices, ultrasensitive magnetic eld sensors, etc.1 A strong impetus is that multiferroics may be promising candidates to solve the problem of large energy dissipation due to high current densities in spintronic devices,2,3 in which controlling the magnetization with external electric elds is very important, with some progress having been achieved up to now.4 For example, 180 reversal of magnetization induced by pure electric eld loading has been realized in the CoFe/BFO heterostructure at room temperature.5 Lahtinen et al. have shown the possibility of writing and erasing regular ferromagnetic domain patterns in magnetoelectric (ME) composites using small electric elds.6 In comparison, a
LTCS and Department of Mechanics and Engineering Science, College of Engineering, Peking University, Beijing, 100871, China. E-mail: [email protected]
b
Department of Material Science and Engineering, College of Engineering, Peking University, China
c Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai, 200050, China
This journal is © The Royal Society of Chemistry 2014
controlling the electric polarization with magnetic elds has been realized only in single-phase ME materials and complete reversal of ferroelectric polarization is rarely seen. Hur et al. showed the possibility of electric polarization reversal induced by magnetic elds in the multiferroic TbMn2O5.7 Yamasaki et al.8 and Shen et al.9 realized the reversal of spontaneous polarization by a magnetic eld in a multiferroic spinel oxide and a Ba1.3Sr0.7Co0.9Zn1.1Fe10.8Al1.2O22 single crystal, respectively. In the studies mentioned above,7–9 the need for lowtemperature environments limits the applications in practice. Keeney et al. demonstrated that switching of ferroelectric polarization by magnetic elds was possible at room temperature in multiferroic Aurivillius phase thin lms.10 Evans et al. also realized switching of ferroelectric domains by magnetic elds at room temperature in a multiferroic PZTFT.11 However, the weak ME effect in single-phase multiferroics is always a big challenge for their practical applications. Although ME composites have been proved to provide strong ME coupling, to date ferroelectric polarization reversal induced by magnetic elds has never been observed in this type of material. One may regard a magnetic eld as not being a viable mechanism to invert the ferroelectric polarization in ME composites, since the coupling between ferroelectricity and ferromagnetism is based on the interfacial strain and thus only non-180 domain switching was thought to be induced by elastic deformation. Currently challenges still exist in measuring the ME coefficient at the nanoscale. Bai et al. provided a quantitative estimation of the local ME coefficient based on the phase shi in magnetic force microscopy (MFM) measurements, while it is usually difficult to measure the phase shi accurately.12 Xie and coworkers obtained a quantitative estimation of the local ME coefficient based on piezoresponse force microscopy (PFM) measurements.13 Caruntu et al. suggested another estimation method also based on PFM measurements.14 However, in their methods, one is required to get accurate piezoelectric coefficients in PFM testing, which is rather difficult or almost impossible at present due to the challenges in determining the exact applied electric eld distribution and eliminating the
Nanoscale, 2014, 6, 8515–8520 | 8515
View Article Online
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
Nanoscale
background effect of the scanning probe microscopy (SPM) setup.15,16 In this work, we demonstrate through PFM testing that it is possible to invert ferroelectric polarization via magnetic elds in a bilayer Terfenol-D/PMN-33%PT composite. The internal electric eld caused by ME coupling is thought to be the driving force for the observed 180 polarization switching. Furthermore, we provide a quantitative estimation of the ME coefficient based on switching spectroscopy PFM testing under different applied magnetic elds. The bilayer Terfenol-D/PMN-PT composite specimen consists of one 12 mm 6 mm 1 mm Terfenol-D layer magnetized along its longitudinal directions and one 5 mm 5 mm 0.4 mm PMN-33%PT rhombohedral single-crystal layer. The PMN-33%PT single crystal is oriented along the h111i direction and poled along the thickness direction. The two layers were carefully bonded together using epoxy resin (West System 206/105, USA) and cured for more than 24 h at room temperature. Here it should be noted that the thickness of the PMN-PT layer is very important in realizing the magnetic-eldinduced polarization reversal. Previously we tried a similar Terfenol-D/PMN-PT specimen with a PMN-PT layer thickness of 0.1 mm and did not observe the magnetic-eld-induced polarization reversal. The experimental setup is schematically shown in Fig. 1(a) where the external magnetic eld was provided by a variable eld module (VFM) in an atomic force microscope (Asylum Research MFP-3D). The VFM can provide a dc magnetic eld with a magnitude of 2250 Oe and a resolution of 1 Oe. Before testing, the Terfenol-D layer only was used to examine its magnetostriction caused by the magnetic eld, and the strain parallel to the applied magnetic eld is shown in Fig. 1(b). It can be seen that the magnetic eld provided by the VFM is fairly strong and can cause a longitudinal magnetostrictive strain of 600–700 ppm in the Terfenol-D layer under a eld of 2250 Oe, which is slightly larger than that of 500–600 ppm in Xie et al.'s work under a eld of 2000 Oe.17 In the setup shown in Fig. 1(a), PFM was used to image the ferroelectric domain patterns and
Communication
characterize polarization switching in the PMN-PT crystal under external magnetic elds. During PFM testing, a 3 V ac voltage was applied to the sample through a conductive probe with a spring constant of 3 N m1 (Budget Sensors, Multi75E-G). The dual frequency resonance tracking technique was used to get a high signal-to-noise ratio. We rstly prefabricated an engineered 180 domain pattern (the logo of our lab) on the PMN-PT layer using lithography PFM, as shown in Fig. 2(a)–(c). The approximate 180 phase contrast shown in Fig. 2(b) and (c) indicates that the reversed 180 domains are well written. Here for simplication, we denote the original domains as +c domains and the reversed domains as c domains. When an in-plane magnetic eld of 2250 Oe is applied to the sample along the magnetization direction of the Terfenol-D, the 180 phase contrast disappears, as shown in Fig. 2(e) and (f), which clearly indicates that the prefabricated c domains had experienced polarization switching. As is known, the rhombohedral single-crystal PMN33%PT has eight equivalent spontaneous polarization directions and only 180 , 109 and 71 domain switching can occur. In PFM, a 180 phase contrast means the angle between the two polarization directions is an obtuse angle and a vanishing phase contrast implies that the angle is an acute angle or zero.18 Thus, the angle between the polarization directions in Fig. 2(d) could only be 0 or 71 . If the angle were 71 , the PFM amplitudes of the engineered domains should be considerably smaller than those of the +c domains. However, Fig. 2(d) shows that the amplitudes of the engineered domains are slightly larger than those of the +c domains, which implies that the engineered domains in Fig. 2(d) and (e) are also +c domains. Therefore, when applying the magnetic eld of 2250 Oe, the engineered c domains in Fig. 2(a) and (b) had undergone 180 switching to +c domains in Fig. 2(d) and (e). In other words, magnetic-eldinduced ferroelectric polarization reversal had been realized in the ME composite. As to the PFM amplitude, the average amplitude in Fig. 2(a) is about 150 pm in the absence of the external magnetic eld, and it increases rapidly to about 1 nm when the magnetic eld
Fig. 1 (a) Schematic of the setup for PFM testing of the Terfenol-D/PMN-PT composite under external magnetic field loading. Arrows P and M denote the poling direction and the magnetized direction, respectively. The yellow cuboids are rare-earth permanent magnet in a variable field module (VFM). By rotating the magnet, different amounts of flux can be channeled through the sample. (b) The magnetostrictive response of the Terfenol-D layer under in-plane magnetic fields from the VFM.
8516 | Nanoscale, 2014, 6, 8515–8520
This journal is © The Royal Society of Chemistry 2014
View Article Online
Nanoscale
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
Communication
Ferroelectric domain evolution in the PMN-PT layer under a series of magnetic fields. Images in the left and middle columns are PFM amplitude images and phase images, respectively. Plots in the right column are the sections of phase images, as indicated by the white dotted lines in the middle column. The left arrow from top to bottom indicates the experimental procedure. (a)–(c) PFM images after PFM lithography; (d)–(f) PFM images under 2250 Oe H field; (g)–(i) PFM images after removing the 2250 Oe H field; (j)–(l) PFM images under 2250 Oe H field; (m)–(o) PFM images after removing the 2250 Oe H field.
Fig. 2
of 2250 Oe is applied, as seen in Fig. 2(d). A similar phenomenon was also observed by Xie et al.17 When the positive magnetic eld is removed, the 180 phase contrast occurs again, as shown Fig. 2(h) and (i). It should be noted that the phase contrast area here is smaller than that in Fig. 2(b), which indicates the domains switched by the magnetic eld are metastable and part of them may return to c domains due to the large depolarization eld aer the transient polarization switching, which is similar to the case in controlling the magnetization with external electric elds where the reversed magnetic domains will return to their original conguration upon removal of the electric eld.19 The average PFM amplitude
This journal is © The Royal Society of Chemistry 2014
in Fig. 2(g) is about 350 pm, smaller than that in Fig. 2(d) but still larger than that in Fig. 2(a). When an anti-parallel magnetic eld of 2250 Oe is applied, the 180 phase contrast disappears again, as shown in Fig. 2(k) and (l). This indicates that those engineered domains had reversed again to +c domains under the negative magnetic eld. Thus the ferroelectric polarization switching in the ME composite is insensitive to the sign of the external magnetic eld, which is reasonable because the magnetostriction in Terfenol-D is also insensitive to the sign of the magnetic eld. Upon removal of the negative magnetic eld, Fig. 2(n) and (o) show that only very small areas exhibit 180 phase contrast, which indicates that few domains recover to c
Nanoscale, 2014, 6, 8515–8520 | 8517
View Article Online
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
Nanoscale
domains and the reversed +c domains induced by the magnetic eld become more stable. The mechanism of the observed magnetic-eld-induced polarization reversal is illustrated in Fig. 3. When an external magnetic eld is applied, the Terfenol-D layer elongates and the deformation is transferred to the PMN-PT layer through the interface. The transferred elongation deformation in the PMNPT layer is apt to depole the crystal and thus an internal electric eld (Ei) along the original poling direction is generated. Actually, this is the principle of ME coupling in ME composites. The magnitude of Ei can be estimated as follows. The macroscopic ME eld coefficient of the Terfenol-D/PMN-33%PT laminates is typically about 1 V cm1 Oe1. Thus when a magnetic eld of 2250 Oe (or 2250 Oe) is applied, a strong internal eld of 2250 V cm1 will be generated, which is close to the coercive eld (about 2 kV cm1) of the PMN-33%PT single crystal.20 Furthermore, it should be noted that the c domains in PMN-PT written by lithography PFM are metastable because the surrounding domains are all +c domains. They are more active and relatively easier to switch. Therefore, the generated internal eld Ei can activate polarization reversal of the c domains. This polarization switching mechanism is similar to the case in ferroelectric crystals aer heat treatment where the internal electric eld due to the pyroelectric effect could cause back switching of the engineered domains.21 Moreover, the internal electric eld induced by a magnetic eld can enhance the +c domains, and thus the PFM amplitude increases when a large magnetic eld is applied or just removed, as seen in Fig. 2(d) and (g) and also in previous experiments.17 It should be noted that the observed magnetic-eld-induced polarization reversal actually results from the competition between the generated internal electric eld and the elongation deformation. If the PMN-PT layer was very thin (say less than 0.2 mm), the elongation deformation could become dominant and the magneticeld-induced polarization reversal may not occur. The existence of the internal eld Ei due to the ME effect can be proved by local polarization switching characterization under a series of magnetic elds using switching spectroscopy PFM (SS-PFM).22,23 Fig. 4(a) and (b) show the amplitude buttery curves and phase hysteresis loops of the PMN-PT layer under
Illustration of the internal electric field in PMN-PT layer generated by an external magnetic field due to the ME effect. (a) ME composite with engineered domains in the absence of an external magnetic field: P denotes the original polarization, Pe denotes polarization of the engineered domain, M denotes the magnetization direction. (b) When an external magnetic field H is applied, an internal electric field Ei is generated in the PMN-PT layer due to the ME effect. Fig. 3
8518 | Nanoscale, 2014, 6, 8515–8520
Communication
different magnetic elds of 0 Oe, 500 Oe and 1500 Oe. It can be seen that the amplitude buttery curve is symmetric in the absence of the external magnetic eld, i.e., the positive coercive voltage (3.17 V) and the negative coercive voltage (3.24 V) are approximately equal. Moreover, the amplitudes for the positive dc eld and for the negative dc eld are almost the same. However, when an external magnetic eld of 500 Oe is applied to the sample, both the amplitude curve and the phase loop become asymmetric, with a positive coercive voltage of 2 V and a negative coercive voltage of 4.35 V. Meanwhile, the amplitude for the positive electric eld is considerably larger than that for the negative electric eld. On further increasing the magnetic eld to 1500 Oe, the amplitude loops become more strongly asymmetric. The positive coercive voltages decrease with increasing magnetic eld, from 3.17 V to 2 V to 1.19 V, and the negative coercive voltages increase with increasing magnetic eld, from 3.24 V to 4.35 V to 4.75 V. These results thus provide direct evidence for the existence of the internal electric eld generated by the ME effect, which is similar to the case in ferroelectrics in that the asymmetric strain buttery curves and D–E hysteresis loops are caused by the internal bias eld.24 Fig. 4(c) shows how the internal eld Ei inuences the local polarization switching in SS-PFM testing. Bearing in mind that the internal eld Ei is always along the original poling direction, a larger negative external electric eld is required to reverse the positive local polarization, leading to a larger negative coercive eld. On the contrary, a smaller positive external eld can make the negative local polarization reverse, i.e., the positive coercive eld is smaller. Since the internal electric eld caused by external magnetic elds via the ME effect is the key reason for the asymmetric local polarization switching curves, it can then be estimated from the positive and negative coercive elds. Obviously, the local ME coefficient can also be estimated simultaneously. As we know, the ME eld coefficient is dened as aE ¼
DE DH
(1)
where DH is the increment of the external magnetic eld and DE is the increment of the internal electric eld caused by external magnetic elds. Under the applied magnetic eld of 500 Oe, the offset between the positive and negative coercive voltage is 2.35 V. Here it should be noted that the electric eld near the AFM tip is highly concentrated and the actual coercive eld should be much larger than the nominal coercive eld. Taking the macroscopic coercive eld value of 2 kV cm1 for PMN-33%PT and a nanoscale nominal coercive eld of 3.2 V/0.04 cm ¼ 80 V cm1, the average eld concentration factor is thus estimated to be 25. Thus the internal electric eld under a magnetic eld of 500 Oe is estimated to be 2.35 V/2/0.04 cm 25 ¼ 735 V cm1. Then we can obtain the local ME coefficient as 1.47 V cm-1 Oe-1. Similarly, the local ME coefficient under an external magnetic eld of 1500 Oe can be estimated as 0.74 V cm-1 Oe-1. Therefore, the local ME coefficient is not a constant and is strongly dependent on the bias magnetic eld, which is similar to the case at the macroscopic scale.25,26 Furthermore, the estimated local ME coefficient is close to the macroscopically measured
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
Communication
Nanoscale
Fig. 4 Local polarization switching characteristics in the PMN-PT layer under a series of magnetic fields. (a) Amplitude–voltage butterfly curves
in SS-PFM; (b) phase–voltage hysteresis loops in SS-PFM; (c) schematic diagrams of the effect of internal electric field on the local polarization switching behavior: P is the macroscopic polarization, Pl is the local polarization, Ei is the internal electric field caused by the ME effect, Ee is the external electric field applied by an AFM tip, Ec is the apparent coercive field and Ec0 is the original coercive field.
value of about 1 V cm-1 Oe-1, which may indicate the validity of this suggested method. In summary, we have realized magnetic-eld-induced ferroelectric polarization reversal in a Terfenol-D/PMN-PT bilayer composite. The internal electric eld caused by the external magnetic eld due to the ME effect is considered as the driving force for the observed 180 polarization switching. By SS-PFM testing under a series of magnetic elds, the existence of the internal electric eld is veried and a quantitative method is suggested for estimating the local ME coefficient. The results presented in this work may open up a new way to control the electric polarization in ME composites and devices.
Acknowledgements HCM is grateful to Mr Guoxi Liu (College of Engineering, Peking University, China) for the help in material fabrication. Financial support from the National Natural Science Foundation of China under Grant no. 11090331 is also acknowledged.
References 1 J. Ma, J. M. Hu, Z. Li and C. W. Nan, Adv. Mater., 2011, 23, 1062–1087. 2 N. A. Spaldin and M. Fiebig, Science, 2005, 309, 391–392. 3 W. Eerenstein, N. D. Mathur and J. F. Scott, Nature, 2006, 442, 759–765. 4 Y. Tokunaga, Y. Taguchi, T. Arima and Y. Tokura, Nat. Phys., 2012, 8, 838–844. 5 J. T. Heron, M. Trassin, K. Ashraf, M. Gajek, Q. He, S. Y. Yang, D. E. Nikonov, Y. H. Chu, S. Salahuddin and R. Ramesh, Phys. Rev. Lett., 2011, 107, 217202. This journal is © The Royal Society of Chemistry 2014
6 T. H. E. Lahtinen, K. J. A. Franke and S. van Dijken, Sci. Rep., 2012, 2, 258. 7 N. Hur, S. Park, P. A. Sharma, J. S. Ahn, S. Guha and S. W. Cheong, Nature, 2004, 429, 392–395. 8 Y. Yamasaki, S. Miyasaka, Y. Kaneko, J. P. He, T. Arima and Y. Tokura, Phys. Rev. Lett., 2006, 96, 207204. 9 S. P. Shen, L. Q. Yan, Y. S. Chai, J. Z. Cong and Y. Sun, Appl. Phys. Lett., 2014, 104, 032905. 10 L. Keeney, T. Maity, M. Schmidt, A. Amann, N. Deepak, N. Petkov, S. Roy, M. E. Pemble and R. W. Whatmore, J. Am. Ceram. Soc., 2013, 96, 2339–2357. 11 D. M. Evans, A. Schilling, A. Kumar, D. Sanchez, N. Ortega, M. Arredondo, R. S. Katiyar, J. M. Gregg and J. F. Scott, Nat. Commun., 2013, 4, 1038. 12 F. M. Bai, H. W. Zhang, J. F. Li and D. Viehland, J. Phys. D: Appl. Phys., 2010, 43, 285002. 13 S. H. Xie, F. Y. Ma, Y. M. Liu and J. Y. Li, Nanoscale, 2011, 3, 3152–3158. 14 G. Caruntu, A. Yourdkhani, M. Vopsaroiu and G. Srinivasan, Nanoscale, 2012, 4, 3218–3227. 15 T. Jungk, A. Hoffmann and E. Soergel, Appl. Phys. Lett., 2007, 91, 253511. 16 K. Pan, Y. Y. Liu, S. H. Xie, Y. M. Liu and J. Y. Li, Acta Mater., 2013, 61, 7020–7033. 17 S. H. Xie, Y. M. Liu, X. Y. Liu, Q. F. Zhou, K. K. Shung, Y. C. Zhou and J. Y. Li, J. Appl. Phys., 2010, 108, 054108. 18 E. Soergel, J. Phys. D: Appl. Phys., 2011, 44, 464003. 19 T. K. Chung, G. P. Carman and K. P. Mohanchandra, Appl. Phys. Lett., 2008, 92, 112509. 20 Z. Y. Feng, D. Lin, H. S. Luo, S. Li and D. N. Fang, J. Appl. Phys., 2005, 97, 024103.
Nanoscale, 2014, 6, 8515–8520 | 8519
View Article Online
Nanoscale
24 F. X. Li, D. N. Fang and A. K. Soh, Smart Mater. Struct., 2004, 13, 668–675. 25 M. Fiebig, J. Phys. D: Appl. Phys., 2005, 38, R123–R152. 26 G. Srinivasan, Annu. Rev. Mater. Res., 2010, 40, 153–178.
Published on 03 June 2014. Downloaded by Temple University on 27/10/2014 16:08:31.
21 X. Y. Liu, K. Kitamura and K. Terabe, Appl. Phys. Lett., 2006, 89, 142906. 22 S. Jesse, H. N. Lee and S. V. Kalinin, Rev. Sci. Instrum., 2006, 77, 073702. 23 S. Jesse, A. P. Baddorf and S. V. Kalinin, Appl. Phys. Lett., 2006, 88, 062908.
Communication
8520 | Nanoscale, 2014, 6, 8515–8520
This journal is © The Royal Society of Chemistry 2014
