The electric field effect on the magnetism in a MgO/Pd/Co system, in which a magnetic moment is induced in the Pd layer owing to the ferromagnetic proximity effect, has been investigated using various experimental methods. An electric field was applied to the surface of the Pd layer through a solid-state HfO2/MgO dielectric bilayer by applying a gate voltage with a back-gating configuration. Changes in the magnetic properties of the system as a result of gate voltage application were detected using magnetization and polar-Kerr effect measurements as well as X-ray absorption and X-ray magnetic circular dichroism (XMCD) spectroscopies. A systematic change in the magnetic moment of the system by the application of a gate voltage is observed. The magnetic hysteresis loops obtained by the polar-Kerr effect measurement and the element-specific XMCD signal at the Co L3-edge clearly show a reproducible change in the coercivity that is dependent on the gate voltage.

The electric field effect on the magnetism in ferromagnetic materials1 with a capacitor structure including a magnetic tunnel junction2,3 has been intensively studied because of its potential to dramatically reduce the power consumption of magnetic memory devices when writing information. Enhancing the efficiency of the electric field effect on magnetic anisotropy4–8 is one of the most important issues for the application of electric field-induced magnetization switching.2,3 The effect has attracted substantial research interest, not only in applied science but also in fundamental studies. For example, an electric field can result in the switching of ferromagnetism itself, i.e., the Curie temperature is tunable by the application of an electric field.9–13 In experiments using metallic Co,11,12,14–24 the electric field has been often applied to the Co surface through an insulating cap layer. Recently, in a similar structure with a 1- to 2-nm-thick Pd insertion layer between Co and MgO cap layers, the change in the total magnetic moment induced by the electric field has been reported by forming an electric double layer (EDL) through the MgO layer using an ionic liquid.25,26 In a layered system with a Pd/Co stack, a magnetic moment is induced in the Pd layer due to the ferromagnetic proximity effect.27 The electric field effect on the magnetism is expected to be induced by modulation of the electronic structure of the ferromagnetic Pd surface adjacent to the MgO layer. Moreover, this system shows the peculiar temperature dependence of the electric field-induced change in magnetic anisotropy,28,29 in which an increase in modulation efficiency by one or two orders magnitude relative to previous results has been reported using 3d transition metal capacitors.7,8,18,22 However, redox reactions have often been discussed when electrolytes are used for the formation of an EDL.23,30–32 In this paper, we investigate the electric field effect on magnetism in a MgO/Pd/Co system using a solid-state capacitor structure, not an EDL system, in which the charging effect must be dominant.33,34 To detect changes in the magnetic properties of this system, magnetization and polar-Kerr effect measurements, and X-ray absorption and X-ray magnetic circular dichroism (XMCD) spectroscopies were used under the application of electric fields. The modulated moment for a Pd atom per electron was determined using the linear dependence of the total magnetic moment of the system on the gate electric field. In addition, an electric-field modulation of the coercivity was observed in the magnetization curves.

A schematic image of the solid-state capacitor fabricated for the experiments is shown in Fig. 1(a). A high-κ dielectric layer was used in addition to a MgO layer as a gate insulator using atomic layer deposition (ALD). A back-gating configuration, as explained in the following description, was adopted to minimize intermixing between Pd and Co caused by heating during the ALD process. First, the Ta(8.3)/Pt(2.0) layers were directly deposited on a semi-insulating GaAs(001) substrate as a gate electrode from the bottom side by rf sputtering at room temperature. The number in parentheses is the thickness of the layer in nm. Subsequently, a 50-nm-thick layer of HfO2 was deposited onto the Pt surface in an ALD chamber at 150°C. A photoresist was then spun-coated, and an ∼1×1 or ∼2×2 mm2-sized square with an additional area for electric contact was removed from the coated resist using a conventional photolithography technique. Finally, the MgO(2.0)/Pd(tPd)/Co(0.6)/Pt(2.0) layers were deposited by rf sputtering at room temperature and removed. The induced magnetic moment in the Pd layer is expected to penetrate to the layer adjacent to MgO with the Pd thickness of the present samples (tPd = 0.9 or 1.0 nm).26,35 A gate voltage VG was applied between the bottom gate electrode and the top metal layers (top electrode). In our definition, a positive (negative) VG increases (reduces) electron density at the bottom surface of the Pd layer. To apply VG, Au wires were connected to the bottom gate and top electrodes using a conductive epoxy, respectively. Since the bottom gate electrode was covered by the HfO2 gate insulator, the area near the top electrode was scratched, and the conductive epoxy was applied such that it covered the crack to improve the contact.

FIG. 1.

(a) Left panel shows the schematic image of the capacitor structure with a back-gating configuration. Right panel indicates the cross-sectional structure of the capacitor. (b) The gate current IG at VG = 0 V as a function of the VG sweeping rate r. The inset shows the VG dependence of IG for r = 100, 250, and 500 V/s.

FIG. 1.

(a) Left panel shows the schematic image of the capacitor structure with a back-gating configuration. Right panel indicates the cross-sectional structure of the capacitor. (b) The gate current IG at VG = 0 V as a function of the VG sweeping rate r. The inset shows the VG dependence of IG for r = 100, 250, and 500 V/s.

Close modal

To determine the capacitance C of the device, triangular VG waves with various sweep rates r were applied to the capacitor. The VG dependences of the gate current IG for r = 100, 250, and 500 V/s are shown in the inset of Fig. 1(b). The square hysteresis in the curves represents the charging current Icharge, whereas the very small non-linear component is probably attributed to the leakage current Ileak. We note that |Ileak| under the dc VG application of ±10 V for our typical device is 100-200 pA/mm2. |IG| at VG = 0 V, which is expected to be the pure Icharge, is plotted in the main panel of Fig. 1(b) as a function of r. The slope of the linear fitting corresponds to C, from which C/S is determined to be 0.32 μF/cm2, where S is the total area of the capacitance. Thus, the difference in the sheet electron density Δn/S (=CΔVG/eS,36 where e is elementary charge) for ΔVG = 20 V (VG=±10 V) has been calculated to be 4.0×1013 cm-2 in the present capacitor.

Figure 2(a) shows the magnetic hysteresis curves for VG = ±10 V, which were obtained by sweeping a perpendicular magnetic field μ0H at room temperature using the polar-Kerr effect. The vertical axis is the polar-Kerr signal normalized by the value at which the magnetization is saturated. A square hysteresis loop is observed, as seen in the whole loop (inset of Fig. 2(a)), indicating that the sample shows perpendicular magnetic anisotropy. The squareness ratio is unity, regardless of VG values. The main panel of Fig. 2(a) shows a magnified portion of the loop around the coercivity μ0Hc. A VG-dependent value of μ0Hc is observed, i.e., μ0Hc is lower (higher) at positive (negative) VG. μ0Hc was measured several times by alternately changing the sign of VG (+10 and -10 V), as shown in Fig. 2(b). The result shows that the upper and lower values of μ0Hc were obtained reproducibly depending on the sign of VG. The sign of the change is consistent with our previous result using EDL capacitors.26,28

FIG. 2.

(a) Magnetization hysteresis loops for VG = ±10 V at room temperature, measured using the polar-Kerr effect. The sample with tPd = 0.9 nm was used for the measurements. The vertical axis indicates the normalized polar-Kerr signal. The main panel shows a magnified portion of the loop near the coercivity μ0Hc; the whole loop is shown in the inset. (b) Values of μ0Hc determined from the hysteresis loop. The horizontal axis corresponds to the number of measurements, where VG is changed in the order of +10→−10→+10→−10 V.

FIG. 2.

(a) Magnetization hysteresis loops for VG = ±10 V at room temperature, measured using the polar-Kerr effect. The sample with tPd = 0.9 nm was used for the measurements. The vertical axis indicates the normalized polar-Kerr signal. The main panel shows a magnified portion of the loop near the coercivity μ0Hc; the whole loop is shown in the inset. (b) Values of μ0Hc determined from the hysteresis loop. The horizontal axis corresponds to the number of measurements, where VG is changed in the order of +10→−10→+10→−10 V.

Close modal

Another important point in the present system is that both μ0Hc and the perpendicular magnetic moment m are affected by VG. Figure 3(a)–(c) show the histograms of m/S at H = 0 for three different values of VG when the sample is in the positively and negatively saturated states. m were measured using a superconducting quantum interference device magnetometer. To obtain these data, the positive or negative finite μ0H was first applied to saturate the magnetization, and then the remanent magnetization m was repeatedly measured 42 times. Note that the squareness ratio of the magnetization hysteresis curve is confirmed to be unity (see the inset of Fig. 3(b)) as in the case of the result obtained using the polar-Kerr effect measurement (see the inset of Fig. 2(a)), indicating that the remanent m equals the saturated value. This result statistically demonstrates that the magnitude of m/S increases (decreases) with the positive (negative) VG application. The direction of the change in m/S in the present system is the same as that in our previous result using a similar system in a top-gating configuration with an EDL structure.25,26 Figure 3(d) shows the median of the Gaussian fitting to the histogram of m/S (solid curves in Fig. 3(a)–(c)) as a function of VG. The magnitudes of m/S for both the positively and negatively saturated states almost linearly increase with VG. From the slope of the linear fitting (dotted lines in Fig. 3(d)), the efficiency of the electric-field modulation of m/S can be determined to be 14.7±1.4 μA/(V/nm). Here, we note that the electric field EG is simply determined from VG divided by the total thickness of the HfO2 and MgO solid dielectric layers (52 nm).

FIG. 3.

The histograms of m/S for (a) VG = −10, (b) 0, and (c) +10 V at H = 0 for the positively and negatively saturated states. Solid curves show the Gaussian fitting. The inset of panel (b) shows the hysteresis loop of m/S. All the data were measured at 250 K. (d) VG dependence of the Gaussian fitting median of the positively and negatively saturated m/S histograms. We confirmed that the standard error of each data point is smaller than the size of symbols. The dotted lines show the linear fits to the data. The sample with tPd = 0.9 nm was measured for the measurements.

FIG. 3.

The histograms of m/S for (a) VG = −10, (b) 0, and (c) +10 V at H = 0 for the positively and negatively saturated states. Solid curves show the Gaussian fitting. The inset of panel (b) shows the hysteresis loop of m/S. All the data were measured at 250 K. (d) VG dependence of the Gaussian fitting median of the positively and negatively saturated m/S histograms. We confirmed that the standard error of each data point is smaller than the size of symbols. The dotted lines show the linear fits to the data. The sample with tPd = 0.9 nm was measured for the measurements.

Close modal

From the experimentally-determined change in m/S [Δm/S (= 5.7±0.6×10-6 A)] and Δn/S (= 4.0×1013 cm-2) for ΔVG = 20 V, the change in the magnetic moment per electron (Δm/Δn) can be determined to be 1.5±0.2 μB, which is almost in agreement with the Slater-Pauling relation as well as our previous result obtained with the EDL capacitor.25 Assuming that the Pd layer is in an fcc (111) structure29 with the same lattice constant in a bulk state, the sheet atom density in the Pd layer (nPd/S) is 5.8×1015 cm-2. If the change in the magnetic moment observed is fully attributed to the electric field effect in the Pd layer, the average change in the magnetic moment per Pd atom Δm/nPd for ΔVG = 20 V is 0.010±0.001 μB. Because of the screening effect, however, the electric field effect is expected to dominate in the Pd surface monolayer (ML) adjacent to the MgO layer. In this case, the change in the magnetic moment per Pd atom in the surface ML becomes 0.040±0.004μB using a sheet Pd atom density of 1.5×1015 cm-2.

There is, however, other possibilities. The first one is the diffused Co atoms in the Pd layer can be the source of the electric field effect on the total magnetic moment. The second possibility is that the electric-field modulation of the magnetic moment in Pd atoms results in the slight change in the moment in Co atoms through the orbital hybridization. X-ray absorption spectra (XAS) and the XMCD spectra for Co L-edges were measured under application of VG to investigate these possibilities. The measurements were performed using a soft X-ray beamline, BL25SU at SPring-8, at room temperature. The diameter of the X-ray beam spot used was ∼100 μm. A partial fluorescence yield method was used to measure the XAS intensity for positive (μ+) and negative (μ-) helicities. The experimental setup is shown in Fig. 4(a). An external magnetic field μ0H of ±1.9 T was applied and tilted by 30° from the normal direction to the plane. The polarization-averaged XAS [(μ+ + μ-)/2] for VG = ±8 V at room temperature are shown in the top panel of Fig. 4(b). The XAS intensity IXAS in the graph was normalized by the value at the edge jump (the average XAS intensity in the range of the photon energy between 803 and 819 eV). The normalized XMCD intensity IXMCD shown in the top panel of Fig. 4(c) is determined from the difference in IXAS between two helicities [μ+ - μ-]. The spectra were obtained by changing VG = ±8 V at each photon energy. Importantly, there are almost no differences in the spectra between VG = ±8 V, indicating that at least no apparent redox changes occur in the Co layer. For confirmation, the out-of-plane orbital magnetic moment morb and the effective spin magnetic moment mspineff. per Co atom have been calculated using the magneto-optical sum rules: morb = 0.19(0.15) μB and mspineff. = 1.69(1.68) μB for VG = +8(−8) V by assuming the 3d hole number of 2.45.37 Although, these values for two VGs are slightly different, the difference in IXAS (ΔIXAS) and IXMCD (ΔIXMCD) between VG = ±8 V (see the bottom panels in Figs. 4(b) and 4(c), respectively) shows no clear signal around the Co L-edges, suggesting that the difference in the calculated values is not statistically significant. Nevertheless, the electric-field induced change in the magnetic moments in Co atoms is negligibly smaller than morb or mspineff. itself within the experimental accuracy.

FIG. 4.

(a) Schematic illustration of the X-ray spectroscopy experimental setup. (b) The top panel shows the XAS near the Co L-edges measured under the application of VG = ±8 V. The bottom panel shows the difference in IXAS (ΔIXAS) between VG = ±8 V. (c) The top panel shows the XMCD spectra near the Co L-edges measured under the applications of VG = ±8 V. The bottom panel shows the difference in IXMCD (ΔIXMCD) between VG = ±8 V. The XAS and XMCD spectra were measured using for the sample with tPd = 1.0 nm. (d) The element-specific magnetization curve for VG = ±8 V measured using the XMCD intensity at the Co L3-edge. The vertical axis is IXMCD divided by its saturated value IsXMCD. In this hysteresis measurement, VG and the photon energy were fixed during the magnetic field sweep. The sample with tPd = 0.9 nm was used for the hysteresis measurements.

FIG. 4.

(a) Schematic illustration of the X-ray spectroscopy experimental setup. (b) The top panel shows the XAS near the Co L-edges measured under the application of VG = ±8 V. The bottom panel shows the difference in IXAS (ΔIXAS) between VG = ±8 V. (c) The top panel shows the XMCD spectra near the Co L-edges measured under the applications of VG = ±8 V. The bottom panel shows the difference in IXMCD (ΔIXMCD) between VG = ±8 V. The XAS and XMCD spectra were measured using for the sample with tPd = 1.0 nm. (d) The element-specific magnetization curve for VG = ±8 V measured using the XMCD intensity at the Co L3-edge. The vertical axis is IXMCD divided by its saturated value IsXMCD. In this hysteresis measurement, VG and the photon energy were fixed during the magnetic field sweep. The sample with tPd = 0.9 nm was used for the hysteresis measurements.

Close modal

Although the difference in the XAS and XMCD spectra for VG = ±8 V was not clear, the element-specific magnetization curve observed from the XMCD intensity at the Co L3-edge shows a clear change in the coercivity, as displayed in the inset of Fig. 4(d), in which the modulation direction is consistent with the results from the polar-Kerr effect measurement (see Fig. 2).

X-ray spectroscopy conducted here shows that the magnetization process in the Co layer is clearly affected by the application of VG, but the change in magnetic moment is smaller than the experimental accuracy. Since the magnetic moment of the Pd layer is induced by the adjacent Co layer, the magnetization process of the Co layer must be linked with that of the Pd layer. Thus, if the magnetic anisotropy at the MgO/Pd interface is modulated by the application of VG, the coercivity of the system (ferromagnetic Pd and Co) should be changed in the same way.

The problem here, however, is that the potentially-existing difference in the spectra near the Co L-edges may be hidden by noise; i.e., the experimental accuracy in the X-ray spectra may not be sufficient to conclude that the modulation of the magnetic moment in the Co atoms is smaller than expected. The extreme case is that all of the change in magnetic moment observed in the magnetization measurement (Fig. 3(d)) arises from the Co atoms. In this case, the average change in the magnetic moment per Co atom Δm/nCo for ΔVG = 16 V is calculated to be 0.009±0.001 μB. Even in this “extreme” case, the expected amount of the modulation is comparable with the difference in the calculated moments for VG = ±8 V. However, since the signal-to-noise level of IXAS (at the edge jump) is ∼1-2%, the experimental accuracy for the X-ray spectroscopies is not sufficient to detect this slight change. Thus, a higher accuracy of the moment determination in the X-ray spectroscopic measurements is necessary to further understand the origin of the magnetic moment change in the present system.

In summary, we have investigated the electric field effect on magnetism in an MgO/Pd/Co system with a solid-state capacitor structure. The clear VG-dependent change in coercivity is observed in the hysteresis curves obtained using the polar-Kerr effect and XMCD measurements at the Co L3-edge. The direct magnetization measurement shows that the magnetic moment of the system is systematically changed depending on VG. Although future work will be directed towards obtaining an element-specific measurement, not only for Co but also for Pd, with higher accuracy using harder X-rays, we believe that the present results include important insights for understanding the electric field effect on proximity-induced magnetism.25,26,29,38

The authors thank K. Yamada, S. Ono and A. Tsukazaki for their technical help. This work was partly supported by JSPS KAKENHI (Grant Nos. 25220604 and 15H05702) and Spintronics Research Network of Japan. Part of the work was performed using facilities at the Cryogenic Research Center, University of Tokyo, and at the Collaborative Research Program of the Institute for Chemical Research, Kyoto University. X-ray spectroscopy was performed at SPring-8 with the approval of the Japan Synchrotron Radiation Research Institute (Proposal Nos. 2016A0902, 2016B0902, 2017A0902, 2017A1869, and 2017B0902).

1.
F.
Matsukura
,
Y.
Tokura
, and
H.
Ohno
,
Nat. Nanotechnol.
10
,
209
(
2015
).
2.
Y.
Shiota
,
T.
Nozaki
,
F.
Bonell
,
S.
Murakami
,
T.
Shinjo
, and
Y.
Suzuki
,
Nat. Mater.
11
,
39
(
2012
).
3.
S.
Kanai
,
M.
Yamanouchi
,
S.
Ikeda
,
Y.
Nakatani
,
F.
Matsukura
, and
H.
Ohno
,
Appl. Phys. Lett.
101
,
122403
(
2012
).
4.
D.
Chiba
,
M.
Yamanouchi
,
F.
Matsukura
, and
H.
Ohno
,
Science
301
,
943
(
2003
).
5.
M.
Weisheit
,
S.
Fähler
,
A.
Marty
,
Y.
Souche
,
C.
Poinsignon
, and
D.
Givord
,
Science
315
,
349
(
2007
).
6.
D.
Chiba
,
M.
Sawicki
,
Y.
Nishitani
,
Y.
Nakatani
,
F.
Matsukura
, and
H.
Ohno
,
Nature (London, UK)
455
,
515
(
2008
).
7.
T.
Maruyama
,
Y.
Shiota
,
T.
Nozaki
,
K.
Ohta
,
N.
Toda
,
M.
Mizuguchi
,
A. A.
Tulapurkar
,
T.
Shinjo
,
M.
Shiraishi
,
S.
Mizukami
,
Y.
Ando
, and
Y.
Suzuki
,
Nat. Nanotechnol.
4
,
158
(
2009
).
8.
M.
Endo
,
S.
Kanai
,
S.
Ikeda
,
F.
Matsukura
, and
H.
Ohno
,
Appl. Phys. Lett.
96
,
212503
(
2010
).
9.
H.
Ohno
,
D.
Chiba
,
F.
Matsukura
,
T.
Omiya
,
E.
Abe
,
T.
Dietl
,
Y.
Ohno
, and
K.
Ohtani
,
Nature (London, UK)
408
,
944
(
2000
).
10.
D.
Chiba
,
F.
Matsukura
, and
H.
Ohno
,
Appl. Phys. Lett.
89
,
162505
(
2006
).
11.
D.
Chiba
,
S.
Fukami
,
K.
Shimamura
,
N.
Ishiwata
,
K.
Kobayashi
, and
T.
Ono
,
Nat. Mater.
10
,
853
(
2011
).
12.
K.
Shimamura
,
D.
Chiba
,
S.
Ono
,
S.
Fukami
,
N.
Ishiwata
,
M.
Kawaguchi
,
K.
Kobayashi
, and
T.
Ono
,
Appl. Phys. Lett.
100
,
122402
(
2012
).
13.
M.
Oba
,
K.
Nakamura
,
T.
Akiyama
,
T.
Ito
,
M.
Weinert
, and
A. J.
Freeman
,
Phys. Rev. Lett.
114
,
107202
(
2015
).
14.
D.
Chiba
,
M.
Kawaguchi
,
S.
Fukami
,
N.
Ishiwata
,
K.
Shimamura
,
K.
Kobayashi
, and
T.
Ono
,
Nat. Commun.
3
,
888
(
2012
).
15.
A. J.
Schellekens
,
A.
van den Brink
,
J. H.
Franken
,
H. J. M.
Swagten
, and
B.
Koopmans
,
Nat. Commun.
3
,
847
(
2012
).
16.
U.
Bauer
,
S.
Emori
, and
G. S. D.
Beach
,
Nat. Nanotechnol.
8
,
411
(
2013
).
17.
A.
Bernand-Mantel
,
L.
Herrera-Diez
,
L.
Ranno
,
S.
Pizzini
,
J.
Vogel
,
D.
Givord
,
S.
Auffret
,
O.
Boulle
,
I. M.
Miron
, and
G.
Gaudin
,
Appl. Phys. Lett.
102
,
122406
(
2013
).
18.
K.
Yamada
,
H.
Kakizakai
,
K.
Shimamura
,
M.
Kawaguchi
,
S.
Fukami
,
N.
Ishiwata
,
D.
Chiba
, and
T.
Ono
,
Appl. Phys. Express
6
,
073004
(
2013
).
19.
T.
Koyama
,
A.
Obinata
,
Y.
Hibino
, and
D.
Chiba
,
Appl. Phys. Express
6
,
123001
(
2013
).
20.
C.
Bi
,
Y.
Liu
,
T.
Newhouse-Illige
,
M.
Xu
,
M.
Rosales
,
J. W.
Freeland
,
O.
Mryasov
,
S.
Zhang
,
S. G. E.
te Velthuis
, and
W. G.
Wang
,
Phys. Rev. Lett.
113
,
267202
(
2014
).
21.
U.
Bauer
,
L.
Yao
,
A. J.
Tan
,
P.
Agrawal
,
S.
Emori
,
H. L.
Tuller
,
S. V.
Dijken
, and
G. S. D.
Beach
,
Nat. Mater.
14
,
174
(
2015
).
22.
J.
Shiogai
,
T.
Ohashi
,
T.
Yang
,
M.
Kohda
,
T.
Seki
,
K.
Takanashi
, and
J.
Nitta
,
J. Phys. D: Appl. Phys.
49
,
03LT01
(
2016
).
23.
T.
Hirai
,
T.
Koyama
,
A.
Obinata
,
Y.
Hibino
,
K.
Miwa
,
S.
Ono
,
M.
Kohda
, and
D.
Chiba
,
Appl. Phys. Express
9
,
063007
(
2016
).
24.
F.
Ando
,
H.
Kakizakai
,
T.
Koyama
,
K.
Yamada
,
M.
Kawaguchi
,
S.
Kim
,
K.-J.
Kim
,
T.
Moriyama
,
D.
Chiba
, and
T.
Ono
,
Appl. Phys. Lett.
109
,
022401
(
2016
).
25.
A.
Obinata
,
Y.
Hibino
,
D.
Hayakawa
,
T.
Koyama
,
K.
Miwa
,
S.
Ono
, and
D.
Chiba
,
Sci. Rep.
5
,
14303
(
2015
).
26.
Y.
Hibino
,
T.
Koyama
,
A.
Obinata
,
K.
Miwa
,
S.
Ono
, and
D.
Chiba
,
Appl. Phys. Express
8
,
113002
(
2015
).
27.
J.
Okabayashi
,
Y.
Miura
, and
H.
Munekata
,
Sci. Rep.
8
,
8303
(
2018
).
28.
Y.
Hibino
,
T.
Koyama
,
A.
Obinata
,
T.
Hirai
,
S.
Ota
,
K.
Miwa
,
S.
Ono
,
F.
Matsukura
,
H.
Ohno
, and
D.
Chiba
,
Appl. Phys. Lett.
109
,
082403
(
2016
).
29.
Y.
Hayashi
,
Y.
Hibino
,
F.
Matsukura
,
K.
Miwa
,
S.
Ono
,
T.
Hirai
,
T.
Koyama
,
H.
Ohno
, and
D.
Chiba
,
Appl. Phys. Express
11
,
013003
(
2018
).
30.
M.
Endo
,
D.
Chiba
,
H.
Shimotani
,
F.
Matsukura
,
Y.
Iwasa
, and
H.
Ohno
,
Appl. Phys. Lett.
96
,
022515
(
2010
).
31.
K.
Leistner
,
N.
Lange
,
J.
Hänisch
,
S.
Oswald
,
F.
Scheiba
,
S.
Fähler
,
H.
Schlörb
, and
L.
Schultz
,
Electrochim. Acta
81
,
330
(
2012
).
32.
L.
Reichel
,
S.
Oswald
,
S.
Fahler
,
L.
Schultz
, and
K.
Leistner
,
J. Appl. Phys.
113
,
143904
(
2013
).
33.
Y.
Wakabayashi
,
H.
Fujii
,
T.
Kimura
,
O.
Sakata
,
H.
Tajiri
,
T.
Koyama
, and
D.
Chiba
,
Z. Phys. Chem.
230
,
569
(
2015
).
34.
S.
Miwa
,
K.
Matsuda
,
K.
Tanaka
,
Y.
Kotani
,
M.
Goto
,
T.
Nakamura
, and
Y.
Suzuki
,
Appl. Phys. Lett.
107
,
162402
(
2015
).
35.
D.-O.
Kim
,
K. M.
Song
,
Y.
Choi
,
B.-C.
Min
,
J.-S.
Kim
,
J. W.
Choi
, and
D. R.
Lee
,
Sci. Rep.
6
,
25391
(
2016
).
36.
D.
Chiba
and
T.
Ono
,
J. Phys. D: Appl. Phys.
46
,
213001
(
2013
).
37.
N.
Nakajima
,
T.
Koide
,
T.
Shidara
,
H.
Miyauchi
,
H.
Fukutani
,
A.
Fujimori
,
K.
Iio
,
T.
Katayama
,
M.
Nývlt
, and
Y.
Suzuki
,
Phys. Rev. Lett.
81
,
5229
(
1998
).
38.
S.
Miwa
,
M.
Suzuki
,
M.
Tsujikawa
,
K.
Matsuda
,
T.
Nozaki
,
K.
Tanaka
,
T.
Tsukahara
,
K.
Nawaoka
,
M.
Goto
,
Y.
Kotani
,
T.
Ohkubo
,
F.
Bonell
,
E.
Tamura
,
K.
Hono
,
T.
Nakamura
,
M.
Shirai
,
S.
Yuasa
, and
Y.
Suzuki
,
Nature Commun.
8
,
15848
(
2017
).