hite>
Current-induced effects in La
Physica B 354 (2004) 1619
Current-induced effects in La
5=8 y
Pr
y
Ca
3=8
MnO
3
ð
y ¼ 0:35Þ
single crystals
G. Garbarino
a
, M. Monteverde
a,1
, C. Acha
a,
Ã
,2
, P. Levy
b,3
, M. Quintero
b
,
T.Y. Koo
c
, S.-W. Cheong
c
a
Laboratorio de Bajas Temperaturas, Departamento de F´sica, FCEyN, Universidad de Buenos Aires, Ciudad Universitaria,
(C1428EHA) Buenos Aires, Argentina
b
Departamento de F´sica, Comisio´n Nacional de Energ´a Ato´mica, Gral Paz 1499, 1650 San Mart´n, Buenos Aires, Argentina
c
Department of Physics and Astronomy, Rutgers University, NJ, USA
Abstract
DC electrical current-dependent resistance and pulsed currentvoltage characteristics as a function of temperature of
mixed valent Mn oxide-based La
5=8 y
Pr
y
Ca
0:375
MnO
3
ð
y ¼ 0:35Þ single crystals are reported. We nd that the low-
temperature regime of this material is strongly current dependent. For small current densities (
10 mA=cm
2
Þ
; the
metalinsulator transition related to the low-temperature enlargement of the ferromagnetic fraction is not observed
down to 10 K. Higher current densities causes a large decrease of the resistance, which is temperature dependent and
exhibits memory effects. Our results are interpreted within a scenario of strong competition between charge and
ferromagnetic ordering.
r
2004 Published by Elsevier B.V.
PACS: 71.30.þh; 72.15.Eb; 72.20.Ht; 75.47.Lx
Keywords: Non-linear effects; Manganites; Percolation; Phase separation
1. Introduction
The magnetic phase separation scenario that
presents some manganites can be very useful to
produce
samples
with
a
controlled
mixture
of
charge-delocalized
ferromagnetic
(CD-F)
and charge-ordered, antiferromagnetic (CO-AF)
phases. It will be technically interesting to ex-
ternally control the proportions of these highly
conducting and insulating mixtures as a change in
the resistivity of several orders of magnitude can
be obtained as a consequence of the percolating
nature of the problem
[1]
. This kind of framework
was already observed for hole-doped manganites
ARTICLE IN PRESS
www.elsevier.com/locate/physb
0921-4526/$ - see front matter r 2004 Published by Elsevier B.V.
doi:10.1016/j.physb.2004.09.011
ÃCorresponding author. Fax: +054 11 45763357.
E-mail address: acha@df.uba.ar (C. Acha).
1
Also at CRTBT (CNRS) Grenoble, France.
2
Also fellow of CONICET of Argentina.
3
Also fellow of CONICET of Argentina. [24]
, particularly for the ðLa; Pr; CaÞMnO
3
com-
pound
[5,6]
, where the conducting phase can be
tuned by temperature, magnetic or electric eld,
time or grain size. Indeed, it was previously shown
that this prototypical manganite is magnetically
phase separated into the mixture already men-
tioned, as a consequence of the intricate interplay
between metallic F and the AF interactions
[1]
. In
this paper we present measurements of the
electrical eld dependence (DC and pulsed) of
the resistance as a function of temperature in
La
5=8 y
Pr
y
Ca
3=8
MnO
3
ð
y ¼ 0:35Þ single crystals.
Our results can be interpreted within a percolation
scenario where the proportion of the CD-F phase
over the CO-AF is favored by the application of
an electric eld.
2. Experimental
Single
crystals
of
nominal
composition
La
5=8 y
Pr
y
Ca
3=8
MnO
3
(LPCMO, y ¼ 0:35) were
synthesized and characterized as was described
previously
[7]
. The resistivity was measured as a
function of temperature ð4 K
pTp300 KÞ and
electric eld, following different congurations: a
four terminal (4 W) standard conguration for
constant current measurements and a two wire
(2 W) conguration for high resistivities (up to
100 GOusing a Keithley 2400 SourceMeter) or
for a constant voltage measurement. Pulsed
measurements were performed, depending on the
magnitude of their period, by generating a single
square pulse of increasing voltages (up to 10 V) for
20 ms to 2 s (Agilent 33250A 80 MHz Function/
Arbitrary Waveform Generator) and determining
the current by measuring the voltage in a
calibrated resistance using an oscilloscope or
directly with the SourceMeter for longer pulses.
Temperature was measured by a small diode
thermometer thermally anchored directly to the
sample.
3. Results and discussion
The resistivity of a LPCMO ðy ¼ 0:35Þ single
crystal as a function of temperature for different
constant voltages is displayed in
Fig. 1
. For small
voltages (V
p10 VÞ; the sample remains insulating
down to low temperatures. When higher voltages
are applied ð10 V
pVp100 VÞ; a metalinsulator-
like transition can be observed, reaching more
than four orders of magnitude of resistivity drop
for the highest voltages. Characteristic tempera-
ture hysteresis of this system can also be observed.
The voltage dependence of the resistivity (rðV Þ), at
a xed temperature for a sample cooled in zero
applied voltage, is shown in
Fig. 2
. When voltage
is increased, we measured a very high resistance
(HR) that remains nearly constant until a tem-
perature-dependent
critical
voltage
(V
c
)
is
reached, where a drop of more than four orders
of magnitude is observed. In these conditions, a
time evolution of the resistivity can also be
observed, a reduction of the resistivity can be
noticed even for a constant applied voltage. When
the voltage is decreased, the sample shows memory
effects, remaining in a different and more con-
ducting low resistanceregime (LR). If the
voltage is increased again, a nonlinear dependence
is observed, even for voltages lower than V
c
:
Considering that the decrease of the resistance
produces an increase of the Joule dissipation in the
sample, the highest voltage parts of these rðV Þ
curves could be modied by overheating. To rule
out this possibility, we performed pulsed RðV Þ
ARTICLE IN PRESS
Fig. 1. Resistivity (2 W, voltage controlled) of LPCMOas a
function of temperature for various applied voltages. The
temperature evolution is indicated by arrows. The inset shows
the 4 W, current controlled resistivity.
G. Garbarino et al. / Physica B 354 (2004) 1619
17 measurements in the LR regime. As can be
observed in
Fig. 3
, no time dependence can be
observed for pulsed and DC voltages V
o10 V:
These results can be analyzed considering the
general effective medium (GEM) equations devel-
oped by Mclachlan
[8]
to describe the electric or
thermal conductivity of a binary mixture of
conducting and insulator materials
f ðs
1=t
M
s
1=t
E
Þ
ð
s
1=t
M
þ
As
1=t
E
Þ
þ ð
1
f Þ ðs
1=t
I
s
1=t
E
Þ
ð
s
1=t
I
þ
As
1=t
E
Þ
¼
0;
(1)
where f is the volume fraction of the CDF
domains, s
M
and s
I
the conductivities of the
metallic and insulating phases, respectively. s
E
is
the effective conductivity that we measure, t a
critical exponent, and A ¼ ð1
f
c
=f
c
Þ
; where f
c
is
the percolation threshold.
Assuming a 3D percolation scenario ðt ¼ 2; f
c
¼
0:17Þ and that s
M
ð
TÞ ¼ sðT Þ of x ¼ 0 and s
I
ð
T Þ ¼
sðTÞ of x ¼ 0:625; the f ðV Þ and f ðtÞ values can be
obtained, tting our data with the GEM equation.
The obtained f ðtÞ and f ðV Þ dependencies are
shown in
Figs. 4
and
5
, respectively. In some
cases, not shown here, we obtained that f ðtÞ;
f ðV Þ40:17; with a corresponding metallic-like
temperature-dependent resistivity for temperatures
in the 10 K
oTo20 K range.
The nonlinearities observed in the LR regime
(I
aV þ bV
7=3
þ
cV
7=2
; as shown in
Fig. 6
)
indicate the presence of a different conducting
mechanism, which can be associated with tunnel-
ing processes
[9]
described, particularly, by the
GlazmanMatveev (GM) theory
[10]
. Thus, these
results point out a rich scenario where the voltage
dependence of the resistivity of LPCMOsingle
crystals can be related to different processes: a
classical percolating framework in a conductin-
ginsulating phase mixture and, when the condi-
tions are favorable, tunneling processes between
conducting paths separated by small insulating
barriers.
ARTICLE IN PRESS
Fig. 2. Resistivity of LPCMOas a function of the applied
voltage at T ¼ 15 K and T ¼ 19 K: The system evolves from a
HR to a LR regime, after applying a voltage V XV
c
:
5
9
1
5
9
10
7
10
6 (M cm)
V (V) = 20ms = 200ms = 2s
V (V)

(

cm) = 20ms
DC
T = 7K
0.3
3
7
7
Fig. 3. Resistivity of LPCMOas a function of voltage
measured applying single square pulses of t ¼ 20 ms compared
to DC measurements. The inset shows the comparison of the
RðV Þ curves obtained with t ¼ 20 ms; 200 ms, and 2 s.
Fig. 4. Time evolution of the resistivity and of t