Ferroelectric switching dynamics in 0.5Ba(Zr
0.2
Ti
0.8
)O
3
-0.5(Ba
0.7
Ca
0.3
)TiO
3
thin films
J. P. B. Silva,
1,2,a)
K. Kamakshi,
3,a)
R. F. Negrea,
4
C. Ghica,
4
J. Wang,
5
G. Koster,
5
G. Rijnders,
5
F. Figueiras,
2,6
M. Pereira,
1
and M. J. M. Gomes
1
1
Centre of Physics, University of Minho, Campus de Gualtar, 4710-057 Braga, Portugal
2
IFIMUP and IN-Institute of Nanoscience and Nanotechnology, Departamento de F
ısica e Astronomia,
Faculdade de Ci
^
encias da Universidade do Porto, Rua do Campo Alegre 687, 4169-007 Porto, Portugal
3
Department of Physics, Madanapalle Institute of Technology & Science, Madanapalle 517325,
Andhra Pradesh, India
4
National Institute of Materials Physics, 105 bis Atomistilor, 077125 Magurele, Romania
5
Faculty of Science and Technology and MESA
þ
Institute for Nanotechnology, Inorganic Materials Science,
University of Twente, P. O. Box 217, 7500 AE Enschede, The Netherlands
6
Department of Physics and CICECO-AIM, University of Aveiro, 3810-193 Aveiro, Portugal
(Received 14 June 2018; accepted 4 August 2018; published online 21 August 2018)
In this work, the ferroelectric characteristics of 0.5Ba(Zr
0.2
Ti
0.8
)O
3
-0.5(Ba
0.7
Ca
0.3
)TiO
3
(BCZT)
thin films grown on 0.7 wt. % Nb-doped (001)-SrTiO
3
(Nb:STO) single-crystal have been investi-
gated. High-resolution transmission electron microscopy and electron energy loss spectroscopy
revealed a very sharp Nb:STO/BCZT interface, while selected area electron diffraction revealed
the epitaxial growth of the BCZT layer on the Nb:STO substrate. The ferroelectric nature of the
BCZT films have been investigated by piezoresponse force microscopy and hysteresis loops. The
effect of electric field on polarization switching kinetics has been investigated and has been ana-
lyzed by the nucleation limited switching model with a Lorentzian distribution function. The local
field variation was found to decrease with the increase in the electric field, and thus, the switching
process becomes faster. The peak value of the polarization current and the logarithmic characteris-
tic switching time exhibited an exponential dependence on the inverse of electric field. This model
gave an excellent agreement with the experimental polarization reversal transients throughout the
whole time range. Published by AIP Publishing. https://doi.org/10.1063/1.5044623
Ferroelectric thin films have been investigated inten-
sively in recent years, due to their high polarizability, which
can contribute to improve the performance and efficiency of
non-volatile memory devices and solar cells.
1–4
The understanding of the polarization switching mecha-
nism in ferroelectric films is important from the scientific point
of view as well as for its applications. Various theoretical mod-
els have been developed to explain the switching kinetics
in ferroelectrics.
5–7
The Kolmogorov–Avrami–Ishibashi (KAI)
model describes the polarization reversal behavior of many sin-
gle crystals but is not adequate for thin films.
8
Therefore,
Tagantsev et al. proposed the nucleation limited switching
(NLS) model as an altern ative approach to KAI model to
explain the polarization reversal kinetics.
9
This model is based
on the statistics of nucleation and growth of the reversed
domains.
It is known that there is an urgent demand for a material
capable of replacing Pb(Zr,Ti)O
3
(PZT) in a broad range
of applications.
10
Recently, the lead-free ferroelectric
0.5Ba(Zr
0.2
Ti
0.8
)O
3
-0.5(Ba
0.7
Ca
0.3
)TiO
3
(BCZT) ceramic
has been considered as a potential candidate for various
applications due to their exceptional pr operties.
11,12
However, these properties in thin films are still far from the
exceptional values exhibited by the bulk.
13,14
Moreover,
there are still no comprehensive studies on the polarization
switching dynamics of BCZT thin films.
Therefore, in the present letter, the ferroelectricity in the
BCZT films at the nano and macroscale level was investi-
gated using piezoresponse force microscopy (PFM) and fer-
roelectric hysteresis loops (P-E), respectively. Furthermore,
the polarization reversal characteristics of epitaxial BCZT
thin films have been analyzed using Lorentzian distribution
function for characteristic times of domain growth based on
the NLS model.
A BCZT target prepared by conventional solid state
reaction, as described by Silva et al.,
15
was used for the
deposition of the corresponding thin films. BCZT thin films,
with 160 nm thickness, were grown on 0.7 wt. % Nb-doped
TiO
2
terminated (001) SrTiO
3
(Nb:STO) single-crystal sub-
strates using pulsed laser deposition (PLD) as described in
Silva et al.
16
Cross-section transmission electron microscopy
(TEM) specimens have been prepared as described in Silva
et al.
16
The transmission electron microscopy (TEM) and
Scanning transmission electron microscopy (STEM) investi-
gations have been performed on a Cs probe-corrected JEM
ARM 200F analytical electron microscope equipped with
a Gatan Quantum SE Image Filter for Electron Energy
Loss Spectroscopy (EELS) and EELS—Spectrum Image
(EELS—SI) analysis in the STEM mode. Imaging and spec-
tral data processing have been made using specialized rou-
tines under Gatan Digital Micrograph. Piezoresponse force
microscopy (PFM) was carried out using a scanning probe
microscope (NT-MDT Ntegra Aura) equipped with internal
lock-in amplifiers. A commercial NT-MDT doped silicon
probes with Pt coating with a radius of curvature of about
a)
Authors to whom correspondence should be addressed: josesilva@fisica.
uminho.pt and kamakshikoppole@gmail.com
0003-6951/2018/113(8)/082903/5/$30.00 Published by AIP Publishing.113, 082903-1
APPLIED PHYSICS LETTERS 113, 082903 (2018)
20 nm, resonance frequency of 130 kHz, and spring con-
stant k of about 3 N/m have been used in PFM. The topo-
graphic, the piezo-response amplitude, and the phase images
were edited via a WSxM 5.0 develop 8.0 software. All piezor-
esponse force microscopy and spectroscopy studies were
done out-of-resonance (10–30 kHz) in order to decrease elec-
trostatic responses with the correspondent topographic cross-
talk. For the macroscopic ferroelectric measurements, gold
(Au) top circular electrodes, with a diameter of 1 mm each,
were grown by thermal evaporation using metal shadow
mask patterning. The ferroelectric hysteresis loops (P-E)
were measured, by investigating the capacitors of Nb:STO/
BCZT/Au, with a modified Sawyer-Tower circuit using a
sinusoidal signal at a frequency of 1 kHz. The polarization
switching transients have been studied by applying a bipolar
square pulse with different amplitudes. Experimental method
for transient current measurements is Sawyer-Tower circuit. In
this, a 50 X resistor is connected in series with Nb:STO/BCZT/
Au device and the input bipolar square pulse signal is applied
across them. The voltage drop across the resistor is measured
by cathode-ray oscilloscope (CRO) in the x-t mode.
17
The structural quality of the BCZT thin film deposited
on the Nb:STO substrate has been investigated by TEM.
Figure 1(a) shows a low-magnificati on TEM image of the
Nb:STO/BCZT structure. The 160-nm thick BCZT layer has
a compact and continuous epitaxial structure with a low
roughness. The HRTEM image shown in Fig. 1(b) evidences
the high crystallinity of the BCZT thin film and very sharp
Nb:STO/BCZT interface. Moreover, the selected area elec-
tron diffraction (SAED) pattern in Fig. 1(c) reveals the epi-
taxial growth of the BCZT layer on the Nb:STO substrate.
The crystallographic relation between the substrate and
BCZT thin film is [001]
Nb: STO
jj[001]
BCZT
.
Figure 2 shows the high angle annul ar dark field—
scanning transmission electron microscopy (HAADF-STEM)
image at low-magnification of the Nb:STO/BCZT structure
and the EELS-SI maps (on the right side of Fig. 2)revealing
the distribution of O, Ca, Ti, Sr, Zr, and Ba elements inside
the green rectangle from HAADF-STEM image. The RGB
(Red-Green-Blue) image (down-right) was obtained overlap-
ping the Ba M and Sr L maps and clearly point to a very sharp
interface without any perceptible atomic interdiffusion, there-
fore confirming the high quality of these BCZT thin films
deposited on the Nb:STO substrate.
Figure 3 shows a topographic scan of the BCZT thin
film taken by AFM revealing a uniform and dense micro-
structure with an average roughness of 0.1 nm and with no
evidence of cracking or defects.
The ferroelectricity at the nanoscale level was investi-
gated by PFM. Simultaneous recording of piezoresponse
amplitude and phase scans [Figures 3(b) and 3(c)]enablesa
visualization of the spatial correlation between strong contrast
regions suggesting ferroelectric domains with 200–300 nm
size. The hysteretic dependencies of the amplitude and phase
piezo-signals to the applied bias electric field are shown in
Fig. 3(d) and clearly reveal the local switching of a ferroelec-
tric domain in the BCZT thin film.
18
The piezo-response
phase signal exhibits a well-shaped hysteresis loop with the
180
phase reversing. The asymmetric shape of the hystere-
sis loops is due to the difference in work functions between
the bottom (Nb:STO) and top (Pt) electrodes.
19,20
The satura-
tion is reached above 5 Vdc, and the local coercive voltages
(V
c
) are taken as the minima of amplitude loop
20
and are
found to be approximately 0.9 and þ1.8 V. The coercive
field (E
c
) was estimated by using the relation V
c
¼ E
c
.t,
21
where t is the thickness of the BCZT film and was estimated
to be 81 kV cm
1
.
Figure 4(a) exhibits the room temperature electric field
dependent polarization-electric field (P–E) hysteresis loops
of the BCZT film. The hysteresis loops reveal that as the
voltage increases, the P–E loop saturates. As the electric field
increased from 60 kV cm
1
to 150 kV cm
1
, both saturation
polarization (P
s
) and the remnant polarization (P
r
) increases.
The average P
s
and P
r
values are shown in Table I for the
different applied electric fields. For an electric field of
150 kV cm
1
, a well-saturated hysteresis loops was obtained
with a P
s
of 30.3 lCcm
2
and a P
r
of 21.3 lCcm
2
. The
obtained values for P
r
and P
s
are higher than those reported
in the literature for epitaxial BCZT thin films.
22,23
Moreover,
the coercive field of 60 kV cm
1
is much higher than the
FIG. 1. (a) TEM image at low-
magnification showing BCZT/Nb: STO
structure; (b) HRTEM image of BCZT-
Nb: STO interface; and (c) SAED pat-
tern corresponding to TEM image.
FIG. 2. HAADF-STEM image at a low-magnification of BCZT/Nb: STO
structure and EELS-SI maps showing the elemental distribution inside of a
green rectangle. RGB map (down-right) was obtained overlapping the Ba M
and Sr L maps.
082903-2 Silva et al. Appl. Phys. Lett. 113, 082903 (2018)
value of 1.75 kV cm
1
for the bul k BCZT ceramic, which is
a consequence of much smaller grains in comparison with
the bulk ceramic.
20
However, the measured coercive field is
similar to the one observed in other BCZT films deposited in
STO substrates.
22
In addition, the difference between the
coercive fields obtained from the P-E loops and from the
PFM measurements can be related to the different top electri-
cal boundary conditions (PFM uses a silicon probe with Pt
coating, while P-E loop measu rements uses Au electrodes).
24
The understanding of process of polarization reversal is
one of the important characteristics of ferroelectric materials
due to its direct relevance in memory applications. The ferro-
electric polarization reversal behavior of BCZT films was
studied using the square pulse at different pulse amplitudes.
The discrete points in Fig. 4(b) correspond to the experimen-
tally observed values of temporal dependence of switching
current for BCZT films at different electric fields in range of
60–150 kV cm
1
. The polarization reversal phenomenon
arises due to the nucleation of new domains, their propaga-
tion, and coalescence. The peak value of current (i
m
) occurs
at a time (t
m
), and their variation with pulse amplitudes is
shown in Table I. The i
m
values show an exponential depen-
dence of reversal electrical field as per Merz’s law,
25
and the
activation field was found to be 64 kV cm
1
. The t
m
value
was found to be 5.7 10
8
s at a field of 150 kV cm
1
,
which is faster when compared to other ferroelectric thin
films.
8,26,27
As shown in Table I, as the electric field
increases 2.5 times, the i
m
increases 4.6 times, while the t
m
decreases 2 times. This could be explained by the contribu-
tion of new switched regions. The spontaneous polarization
(P
s
) was estimated by calculating the area under the switch-
ing current transients (Q) using the relation Q ¼ 2P
s
A, where
“A” is the area of the electrode.
28
The P
s
values estimated
from the polarization reversal are in good agreement with
those of values obtained from the P-E loop, as shown in
Table I.
Kolmogorov-Avrami-Ishibashi proposed first a theoreti-
cal model known as KAI model to explain the ferroelectric
switching kinetics in single crystals based on classical theory
of statistical nucleation and unrestricted domain growth.
Later, Tagantsev et al.
9
proposed the nucleation limited
switching (NLS) model based on statistics of nucleation and
domain growth rate, and this model has been chosen to ana-
lyze the switching kinetics of present BCZT thin films. The
NLS model assumes that the sample is as ensemble of regions
in which switching takes place independently and simulta-
neously and thus each region may have different switching
characteristic time t
o
. It represents the characteristic time for
FIG. 3. (a) Topography, (b) off-plane
piezo-response amplitude, and (c)
phase scans. (d) Local PFM amplitude
butterfly loops and phase hysteresis
loops of the BCZT thin film.
FIG. 4. (a) Electric field dependent
P–E loops of the BCZT film, (b) polar-
ization reversal transients of BCZT
films at different electric fields (dis-
crete points correspond to experimen-
tal data and solid line corresponds
NLS theory), and (c) Lorentzian distri-
bution function and inset shows varia-
tion of P
s
with the electric field.
TABLE I. The effect of the applied electric field on the average remnant
and spontaneous polarization (P
r
and P
s)
and on the switching parameters
for the BCZT thin film.
Electric
field
(kV cm
1
)
Pr
(lCcm
2
)
(P-E loops)
Ps
(lCcm
2
)
(P-E loops)
Ps
(lCcm
2
)
(switching)
i
m
(mA)
t
m
(s)
60 3.7 5.5 6.1 133 1.1 10
7
90 11.1 14.0 13.7 293 9.9 10
8
120 17.0 22.0 21.5 419 7.6 10
8
150 21.3 30.3 30.2 612 5.7 10
8
082903-3 Silva et al. Appl. Phys. Lett. 113, 082903 (2018)
domain growth and is proportional to average distance
between the nuclei, divided by the domain wall speed. Since
each region have different t
o
values, NLS model assumed a
broad distribution of characteristic switching time t
o
and is
represented by F(log t
o
). The temporal dependence of polari-
zation current has been expressed by considering the
Lorentzian distribution function of logarithmic switching
times in the definite regions of the film as follows:
8
it
ðÞ
¼ 2P
s
A
d
dt
ð
1
1
1 exp
t
t
0
n
"#
F log t
0
ðÞ
d log t
0
ðÞ
()
;
(1)
where F log t
o
ðÞ
is the distribution function for log t
o
as
follows:
8
F log t
0
ðÞ
¼
K
p
x
log t
0
log t
1
ðÞ
2
þ x
2
; (2)
where K is a normalized constant, x is the half-width at half
maximum, and t
1
peak time value of Lorentz distribution func-
tion. In general, either Gaussian or Lorentz distribution func-
tion can be used in statistically independent random process
like switching kinetics, film growth models, etc. However,
Vleck has shown the dipolar broadening of magnetic reso-
nance lines in crystals with randomly distributed dipole impu-
rities obey to Lorentzian distribution.
29
Theoretical studies
also revealed that the distribution of any interaction field com-
ponent in the system of dilute aligned dipoles obeys to
Lorentzian distribution.
30
Moreover, it is known that ferroelec-
tric thin films contain dipole defects which hinder the domain
wall motion. Therefore, a Lorentz distribution function is con-
sidered in the NLS model.
Equation (1) was used to fit the experimental data of
polarization reversal curves and the corresponding fitted
curves were shown by solid lines in Fig. 4(b). The corre-
sponding logarithmic distribution function of Eq. (2) is
shown in Fig. 4(c). The excellent agreement of NLS curves
with the experimental data in whole time region suggest that
films consist of different regions in which the polarization
reversal take place independently. The existence of indepen-
dent regions is also corroborated from the linear field depen-
dence of P
s
in the studied electric field range, as shown in
the inset of Fig. 4(c). Therefore, the new switching regions
are contributing to the polarization as electric field increases.
The values of parameters t
1
and x are obtained from the
fittings and then are related to microscopic parameters using
the approximations
8,31
log t
1
/
dm
E
; (3)
x
C /
dm
E
2
; (4)
where /
dm
is the activation field for domain wall motion and
C is the full width at half maximum in the Lorentzian local
field distribution function. F(
E) is related to the concentra-
tion of pinning sites and is described as follows:
8,31
F
E
ðÞ
¼
K
p
C
E
2
þ C
2
; (5)
where
E is the local electric field distribution that exists at
pinning sites. The presence of non-homogenous structural
defects, such as domain walls, dislocations, or grain bound-
aries, acts as pinning sites of polarization reversal and causes
local field vacations between the switching and non-switching
regions.
32
Figures 5(a) and 5(b) show the plot of log t
1
as a
function of E
1
and the plot of w versus
/
dm
E
2
, respectively.
The values of /
dm
and C are found to be 3 and 62 kV
cm
1
, respectively. This activation energy of C can be
related to the threshold energy for the pinned domains and is
in good agreement with values obtained from Merz’s law.
This activation field is higher than the one observed in
BaTiO
3
crystals, but lower than the ones reported for PZT,
BiFeO
3
, and Ba
0.8
Sr
0.2
TiO
3
thin films.
25,27,33–35
In summary, high structural quality epitaxial BCZT thin
films were grown by pulsed laser deposition on single crys-
talline Nb:STO (001) substrates. The ferroelectric nature of
films was examined at the nanoscale level by PFM and at the
macroscopic level by P-E hysteresis loops. The BCZT films
exhibit promising ferroelectric properties with a notable rem-
nant polarization of 21.3 lCcm
2
and a coercive field of
60 kV cm
1
. The domain growth limited switching process
based on NLS model is found to be appropriate to describe
the reversal kinetics and the logarithmic characteristic switch-
ing time obeyed the Lorentzian distribution. Moreover, the
values of P
s
from the switching and P-E loops are in good
agreement. The peak values of polarization current and loga-
rithmic characteristics switching time obey the exponential
dependence on the electric field. The t
m
values are found to
be faster as compared to previous reports. Therefore, this
work provides a comprehensive study about the polarization
reversal behavior on BCZT thin films with enhanced ferro-
electric properties.
This work was supported by (i) Portuguese Foundation for
Science and Technology (FCT) in the framework of the
Strategic Funding No. UID/FIS/04650/2013 and (ii) Project
Norte-070124-FEDER-000070 Nanomateriais Multifuncionais.
Part of this work was supported by the COST Action MP1308
“Towards Oxide-Based Electronics (TO-BE).” The authors
acknowledge the CERIC-ERIC Consortium for access to
experimental facilities and financial support under Proposal
No. 20157018. The authors J.P.B.S. and F.F. are grateful for
financial support through the FCT Grant Nos. SFRH/BPD/
92896/2013 and SFRH/BPD/80663/2011, respectively. R.F.N.
and C.G. acknowledge the financial support from the
Romanian Ministry of Research and Innovation in the frame of
the Core Program No. PN18-110101. The authors would also
FIG. 5. (a) Plot of log t
1
versus 1/E and (b) plot of w versus
/
dm
E
2
.
082903-4 Silva et al. Appl. Phys. Lett. 113, 082903 (2018)
like to acknowledge P. B. Tavares, Centro de Qu
ımica da
Universidade de Tr
as-os-Montes e Alto Douro, for the supply
of the 0.5BZT-0.5BCT PLD target.
1
A. Ghosh, G. Koster, and G. Rijnders, APL Mater. 4, 066103 (2016).
2
D.-F. Pan, G.-F. Bi, G.-Y. Chen, H. Zhang, J.-M. Liu, G.-H. Wang, and
J.-G. Wan, Sci. Rep. 6, 22948 (2016).
3
J. P. B. Silva, F. L. Faita, K. Kamakshi, K. C. Sekhar, J. A. Moreira, A.
Almeida, M. Pereira, A. A. Pasa, and M. J. M. Gomes, Sci. Rep. 7, 46350
(2017).
4
H. Matsuo, Y. Noguchi, and M. Miyayama, Nat. Commun. 8, 207 (2017).
5
A. Ghosh, G. Koster, and G. Rijnders, Adv. Funct. Mater. 26,5748
(2016).
6
Y.-H. Shin, I. Grinberg, I.-W. Chen, and A. M. Rappe, Nature 449, 881
(2007).
7
A. Gruverman, “Polarization behavior in thin film ferroelectric capacitors
at the nanoscale,” in Scanning Probe Microscopy of Functional Materials,
edited by S. Kalinin and A. Gruverman (Springer, New York, NY, 2010).
8
K. C. Sekhar and R. Nath, Appl. Phys. Lett. 94, 102905 (2009).
9
A. K. Tagantsev, I. Stolichnov, N. Setter, J. S. Cross, and M. Tsukada,
Phys. Rev. B 66, 214109 (2002).
10
J. R
€
odel, K. G. Webber, R. Dittmer, W. Jo, M. Kimura, and D.
Damjanovic, J. Eur. Ceram. Soc. 35, 1659 (2015).
11
W. Liu and X. Ren, Phys. Rev. Lett. 103, 257602 (2009).
12
J. Gao, D. Xue, Y. Wang, D. Wang, L. Zhang, H. Wu, S. Guo, H. Bao, C.
Zhou, W. Liu, S. Hou, G. Xiao, and X. Ren, Appl. Phys. Lett. 99, 092901
(2011).
13
M. Acosta, N. Novak, V. Rojas, S. Patel, R. Vaish, J. Koruza, G. A.
Rossetti, and J. R
€
odel, Appl. Phys. Rev. 4, 041305 (2017).
14
J. P. B. Silva, K. Kamakshi, K. C. Sekhar, E. C. Queir
os, J. A. Moreira, A.
Almeida, M. Pereira, P. B. Tavares, and M. J. M. Gomes, J. Phys. D:
Appl. Phys. 49, 335301 (2016).
15
J. P. B. Silva, E. C. Queir
os, P. B. Tavares, K. C. Sekhar, K. Kamakshi, J.
A. Moreira, A. Almeida, M. Pereira, and M. J. M. Gomes, J. Electroceram.
35, 135 (2015).
16
J. P. B. Silva, J. Wang, G. Koster, G. Rijnders, R. F. Negrea, C. Ghica, K.
C. Sekhar, J. A. Moreira, and M. J. M. Gomes, ACS Appl. Mater.
Interfaces 10, 15240 (2018).
17
I. K. Yoo, “Testing and characterization of ferroelectric thin film capaci-
tors,” in Nanoscale Phenomena in Ferroelectric Thin Films, edited by S.
Hong (Kluwer Academic, Boston, 2004).
18
Q. R. Lin, D. Y. Wang, B. C. Luo, R. Ding, D. L. Lorenzen, and S. Li,
Appl. Surf. Sci. 331, 477 (2015).
19
A. Jalalian, A. M. Grishin, X. L. Wang, Z. X. Cheng, and S. X. Dou, Appl.
Phys. Lett. 104, 103112 (2014).
20
V. S. Puli, D. K. Pradhan, S. Adireddy, R. Mart
ınez, P. Silwal, J. F. Scott,
C. V. Ramana, D. B. Chrisey, and R. S. Katiyar, J. Phys. D: Appl. Phys.
48, 355502 (2015).
21
S. M. Neumayer, I. N. Ivanov, M. Manzo, A. L. Kholkin, K. Gallo, and B.
J. Rodriguez, J. Appl. Phys. 118, 224101 (2015).
22
B. C. Luo, D. Y. Wang, M. M. Duan, and S. Li, Appl. Phys. Lett. 103,
122903 (2013).
23
M.-J. Choi, J.-R. Lim, J.-S. Choi, J.-H. Eom, B.-J. Park, K.-S. Kim, D.
Kim, and S.-G. Yoon, Scr. Mater. 108, 96 (2015).
24
S. V. Kalinin, B. J. Rodriguez, S. Jesse, Y. H. Chu, T. Zhao, R. Ramesh,
S. Choudhury, L. Q. Chen, E. A. Eliseev, and A. N. Morozovska, Proc.
Natl. Acad. Sci. U. S. A. 104, 20204 (2007).
25
J. Y. Jo, H. S. Han, J.-G. Yoon, T. K. Song, S.-H. Kim, and T. W. Noh,
Phys. Rev. Lett. 99, 267602 (2007).
26
O. Lohse, M. Grossmann, U. Boettger, D. Bolten, and R. Waser, J. Appl.
Phys. 89, 2332 (2001).
27
D. Pantel, Y.-H. Chu, L. W. Martin, R. Ramesh, D. Hesse, and M. Alexe,
J. Appl. Phys. 107, 084111 (2010).
28
N. Dabra, J. S. Hundal, A. Nautiyal, K. C. Sekhar, and R. Nath, J. Appl.
Phys. 108, 024108 (2010).
29
J. H. V. Vleck, Phys. Rev. 74, 1168 (1948).
30
J. R. Klauder and P. W. Anderson, Phys. Rev. 125, 912 (1962).
31
J. P. B. Silva, K. C. Sekhar, S. A. S. Rodrigues, A. Khodorov, J. Mart
ın-
Sanchez, M. Pereira, and M. J. M. Gomes, Curr. Appl. Phys. 12, 1144
(2012).
32
W. Li and M. Alexe, Appl. Phys. Lett. 91, 262903 (2007).
33
H. H. Wieder, J. Appl. Phys. 28, 367 (1957).
34
J. F. Scott, L. Kammerdiner, M. Parris, S. Traynor, V. Ottenbacher, A.
Shawabkeh, and W. F. Oliver, J. Appl. Phys. 64, 787 (1988).
35
J. P. B. Silva, K. C. Sekhar, A. Almeida, J. Agostinho Moreira, J. Mart
ın-
S
anchez, M. Pereira, A. Khodorov, and M. J. M. Gomes, J. Appl. Phys.
112, 044105 (2012).
082903-5 Silva et al. Appl. Phys. Lett. 113, 082903 (2018)
