materials
Article
Enhanced Unipolar Resistive Switching
Characteristics of Hf0.5Zr0.5O2 Thin Films with High
ON/OFF Ratio
Zhipeng Wu and Jun Zhu *
State Key Laboratory of Electronic Thin Films and Integrated Devices,
University of Electronics Science and Technology of China, Chengdu 610054, China; [email protected]
* Correspondence: [email protected]; Tel.: +86-28-8320-3873
Academic Editor: Douglas Ivey
Received: 16 January 2017; Accepted: 13 March 2017; Published: 22 March 2017
Abstract: A metal–insulator–metal structure resistive switching device based on H0.5 Z0.5 O2 (HZO)
thin film deposited by pulse laser deposition (PLD) has been investigated for resistive random access
memory (RRAM) applications. The devices demonstrated bistable and reproducible unipolar resistive
switching (RS) behaviors with an extremely high OFF/ON ratio over 5400. The retention property
had no degradation at 6 × 104 s. The current–voltage characteristics of the HZO samples showed
a Schottky emission conduction in the high voltage region (Vreset < V < Vset ), while at the low voltage
region (V < Vreset ), the ohmic contact and space charge limited conduction (SCLC) are suggested to
be responsible for the low and high resistance states, respectively. Combined with the conductance
mechanism, the RS behaviors are attributed to joule heating and redox reactions in the HZO thin film
induced by the external electron injection.
Keywords: Hf0.5 Zr0.5 O2 thin films; resistive switching; RRAM; PLD
1. Introduction
Resistive switching (RS) devices based on various binary transition metal oxides, such as
TiO2 [1–3], NiOx [4,5], ZnO [6,7], HfO2 [8–12], ZrO2 [13–16], etc., have been extensively studied for
their great potential as excellent substitutes in non-volatile memory applications. Such kinds of devices
are commonly referred to as resistive random access memory (RRAM). The devices usually show
bistable resistance state, and one can switch from a conductive ON state (low resistance state, LRS) to
a less conductive OFF state (high resistance state, HRS) by applying electrical bias to the device, and
vice versa. The switching operation is called unipolar when the switching procedure does not depend
on the polarity of the write voltage. In contrast, the operation is called bipolar when the set to ON state
occurs at one voltage polarity and the reset to OFF state on reversed voltage polarity [17,18]. The simple
metal/insulator/metal (MIM) structure of these devices makes them quite easy to fabricate and
integrate with other processes. In recent years, HfO2 - and ZrO2 -based thin films have been extensively
used as promising “high-k” substitutes for integrated oxide gate dielectrics driven by the tendency of
down-scale in complementary metal–oxide–semiconductor (CMOS) applications [19]. This makes it
possible to integrate the HfO2 - and ZrO2 -based RS devices with CMOS technology. However, most
recent studies show that both HfO2 and ZrO2 have some disadvantages in resistive switching memory
applications. HfO2 -based thin films usually demonstrate a bipolar resistive switching behavior and
need a negative polarity power supply support, while ZrO2 -based thin films resistive switching devices
typically have a poor uniformity behavior, which remains an obstacle for their application. Moreover,
both RS behaviors operate with a relatively small OFF/ON ratio of the HfO2 or ZrO2 thin films. Wu [8]
studied the HfO2 /ITO/Invar structure with conductive atomic force microscopy, and a ratio of about
Materials 2017, 10, 322; doi:10.3390/ma10030322
www.mdpi.com/journal/materials
Materials 2017, 10, 322
2 of 7
100 was revealed with this unipolar RS structure. A Ti/ZrO2 /Pt structure was fabricated by Zhang [16],
and an OFF/ON ratio of about 24 was reported. Sharath [9] investigated the RS behaviors in engineered
oxygen-deficient HfO2-x thin films with various thicknesses. It was suggested that both stoichiometric
HfO2 and nonstoichiometric HfO2-x show bipolar RS behaviors with a relatively small OFF/ON ratio
of approximately 5. Ag/HfO2 /ITO and TiN/HfO2 /ITO structures were studied by Ramadoss [11] and
Ye [12], respectively. Both structures presented a bipolar RS behavior with OFF/ON ratios not over 10.
Meanwhile, the essential negative power in bipolar RS devices has also increased the complexity of
the integrated system [20]. To improve the applications in RRAM and integrated systems, a promising
unipolar RS material with high OFF/ON ratio is particularly important. Several groups [20–22] tried
to use the HfO2 /ZrO2 bilayer structure to improve the OFF/ON ratio and stability of the device,
and better results were observed. This makes us believe that the HfO2 -ZrO2 solution might provide
a better performance.
In this letter, the RS characteristics of a H0.5 Z0.5 O2 (HZO) thin film prepared by pulse laser
deposition (PLD) for RRAM application have been investigated. This HZO thin film reveals a bistable
RS property with extremely high OFF/ON ratio. To understand the RS mechanism in this film,
the current conductive mechanism is also discussed.
2. Materials and Methods
The HZO films were deposited on Pt/Ti/SiO2 /Si substrates by PLD with a ceramic Hf0.5 Zr0.5 O2
target synthesized by solid phase sintering with high purity HfO2 and ZrO2 powder. During deposition,
the energy of the laser beam was fixed at 2 J/cm2 . The thickness of the HZO thin film was about 50 nm.
For electrical measurement, the Ni (40 nm)/Au (150 nm) top electrodes with a diameter of 190 µm
were deposited by electron beam evaporation using a shadow mask to form the metal–insulator–metal
(MIM) sandwich structure. The current–voltage (I-V) characteristics of the MIM structures were
measured using an Agilent 4156C precision semiconductor parameter analyzer (Agilent Technologies,
Santa Clara, CA, USA) at room temperature with the humidity less than 30%. RS measurement was
performed by applying a DC voltage sweep with the shape of a staircase to the top electrode while
the bottom electrode was grounded. For the endurance testing measurement, we used the staircase
sweep measurement mode with a delay time of 1 s. For the retention testing measurement, we read the
current at different times with an applied pulse voltage of 0.2 V on the devices at 85 ◦ C. The current
compliance of 5 mA was used for the forming and set process. Chemical valence states of elements in
the film were analyzed with Al Kα source XSAM800 X-ray photoelectron spectroscopy (XPS, Kratos
Analytical, Manchester, UK).
3. Results and Discussion
Figure 1 shows the typical I-V characteristics of the HZO samples taken from the cycles test.
Before that, the forming process was taken by applying a sweep voltage from 0 to 5 V with compliance
current at 100 µA. The unipolar switching behavior is revealed in this figure. When a positive electrical
stress was applied to the HRS sample, the current increased to the compliance current abruptly, while
the device switched to LRS. This is defined as the set process, and the Vset main range is from 1.8 to
2.2 V. On the other hand, the reset process was also achieved by sweeping the voltage from zero to
the Vreset , which mainly ranges from 0.7 to 0.9 V to switch the device from LRS to HRS. The current
switching phenomenon is also observed at negative voltage sweeping, confirming the unipolar RS in
the HZO samples [17].
Materials 2017, 10, 322
3 of 7
Materials 2017, 10, 322
3 of 7
Materials 2017, 10, 322
Materials 2017, 10, 322
3 of 7
3 of 7
Figure 1. Typical unipolar resistive switching (RS) current–voltage curves for Au/Ni/HZO/Pt device.
Figure 1. Typical unipolar resistive switching (RS) current–voltage curves for Au/Ni/HZO/Pt device.
The inset shows the scheme of metal/insulator/metal (MIM) structure. HRS: high resistance state; LRS:
The inset shows the scheme of metal/insulator/metal (MIM) structure. HRS: high resistance state; LRS:
Figure
1. Typical
unipolar resistive switching (RS) current–voltage curves for Au/Ni/HZO/Pt device.
low resistance
state.
Figure 1. Typical
low resistance
state. unipolar resistive switching (RS) current–voltage curves for Au/Ni/HZO/Pt device.
The inset shows the scheme of metal/insulator/metal (MIM) structure. HRS: high resistance state; LRS:
The inset shows the scheme of metal/insulator/metal (MIM) structure. HRS: high resistance state; LRS:
Figure
2a shows
low
resistance
state.the endurance characteristic of the MIM structure with the resistance of HRS
low resistance state.
and LRS
a function
of the number
of switching
measured
at the
reading
voltage ofof0.2
V. and
Figure
2aasshows
the endurance
characteristic
of cycles
the MIM
structure
with
the resistance
HRS
Figure
2a
shows
the
endurance
characteristic
of
the
MIM
structure
with
the
resistance
of
HRS
During
the
progress
of
successive
RS
up
to
450
cycles,
the
OFF/ON
ratio
was
at
least
5400
with
no
LRS as a function
the number
of switching
cycles measured
at structure
the reading
voltage
of 0.2 V.ofDuring
Figure 2aofshows
the endurance
characteristic
of the MIM
with
the resistance
HRS the
and
LRS degradation
as a functionobserved—an
of the number
of switching
cycles
atseparation
the reading
of 0.2 V.
obvious
OFF/ON
ratio
of
10 ismeasured
a ratio
sufficient
forvoltage
the
nonvolatile
progress
of
successive
RS
up
to
450
cycles,
the
OFF/ON
was
at
least
5400
with
no
and LRS as a function of the number of switching cycles measured at the reading voltage of 0.2obvious
V.
During
progress of
successive
RS upthat
to 450
the OFF/ON
wasthe
at least
with no
memorythe
application
[23],
which means
ourcycles,
HZO sample
is far ratio
beyond
basic 5400
requirement.
During the
progress of successive
up of
to 10
450iscycles,
the OFF/ON
ratio was
at least
5400 withmemory
no
degradation
observed—an
OFF/ONRS
ratio
a sufficient
separation
for the
nonvolatile
obvious
degradation
observed—an
OFF/ONofratio
of 10 isvoltage
a sufficient
the nonvolatile
Figure 2b
demonstrates
the distributions
Set/Reset
takenseparation
from the for
endurance
cycles.
obvious[23],
degradation
observed—an
OFF/ON
ratio of 10
is a beyond
sufficientthe
separation
for the nonvolatile
application
which means
thatmeans
our
HZO
sample
issample
far
basic
Figure 2b
memory
application
[23],
which
that our
HZO
is far
beyond
therequirement.
basic requirement.
The stander
deviation
of Set
and Reset voltages
are
0.17
and 0.06,
respectively.
Figure
2c shows the
memory application
[23], which
means that voltage
our HZOtaken
sample
is far
beyond
the basic
requirement.
demonstrates
the
distributions
of
Set/Reset
from
the
endurance
cycles.
The
stander
Figure
2b demonstrates
the distributions
of Set/Reset
voltage
from
the endurance
cycles.
cumulative
probability distribution
of the resistance
at LRS
and taken
HRS. As
shown
in Figure 2b,c,
the
Figure
2b
demonstrates
the
distributions
of
Set/Reset
voltage
taken
from
the
endurance
cycles.
deviation
of Set deviation
anddevice
Reset
voltages
aregood
0.17
and are
0.06,
respectively.
Figure 2cisFigure
shows
cumulative
The
stander
of
Set and Reset
voltages
0.17
andretention
0.06, respectively.
2cthe
shows
the
HZO-based
RS
demonstrates
uniformity.
The
property
another
significant
The stander deviation of Set and Reset voltages are 0.17 and 0.06, respectively. Figure 2c shows the
probability
distribution
of the
resistance
and
HRS.
Asand
shown
Figure
2b,c,
the
cumulative
probability
distribution
of at
theLRS
resistance
at LRS
HRS.inAs
shown
in Figure
2b,c, the
characteristic
of the nonvolatile
memory
device
[17,23],
and
demonstrates
the ability
of HZO-based
a device
to RS
cumulative probability distribution of the resistance at LRS and HRS. As shown in Figure 2b,c, the
HZO-based
RS
device
demonstrates
good
uniformity.
The
retention
property
is
another
significant
retain
information
afteruniformity.
writing theThe
stored
data. Figure
3 shows
the retention
characteristic
of theof the
device
demonstrates
good
retention
property
is another
significant
characteristic
HZO-based RS device demonstrates good uniformity. The retention property is another significant
characteristic
of
the
nonvolatile
memory
device
and
demonstrates
thetoto
ability
a resistance
device to after
HZO sample.
During
the[17,23],
stress progress,
a small[17,23],
voltage
of 0.2
Vof
was
applied
read of
the
nonvolatile
memory
device
and
demonstrates
the
ability
a
device
retain
information
characteristic of the nonvolatile memory device [17,23], and demonstrates the ability of a device to
retain
the of
stored
data.
3 shows
thethe
retention
characteristic
ofover
thestress
state
the sample.
Thewriting
resistance
both
HRS Figure
and
LRS
conditions
showed
no
degradation
writing
theofinformation
stored
data.after
Figure
3 shows
retention
characteristic
HZO sample.
During
retain
information
after
writing
the the
stored
data. Figure
3 showsof
the retention
characteristic
ofthe
the
4
HZO
sample.
During
the
stress
progress,
a
small
voltage
of
0.2
V
was
applied
to
read
the
resistance
6 × 10 small
s. All voltage
of these of
electrical
properties
that
HZO thin
filmofisthe
an excellentThe
RS material
progress,
0.2stress
V was
appliedindicate
read
thethe
resistance
state
resistance of
HZOasample.
During the
progress,
a to
small
voltage
of 0.2 V was
applied sample.
to read the resistance
state
of the
sample.
The for
resistance
of both HRS and LRS conditions
showed no degradation over
that has
a high
potential
RRAM application.
state of
theLRS
sample.
The resistance
HRS and LRS
no degradation
over
both HRS
and
conditions
showedofnoboth
degradation
overconditions
6 × 104 s. showed
All of these
electrical properties
6 × 1044 s. All of these electrical properties indicate that the HZO thin film is an excellent RS material
6
×
10
s.
All
of
these
electrical
properties
indicate
that
the
HZO
thin
film
is
an
excellent
RS
material
indicate
HZO
thin film
an excellent
RS material that has a high potential for RRAM application.
thatthat
has the
a high
potential
for is
RRAM
application.
that has a high potential for RRAM application.
Figure 2. (a) Endurance characteristics of the H0.5Z0.5O2 (HZO) device; the currents are read at 0.2 V;
(b) Set and Reset voltage distribution taken from the endurance cycles; (c) Distribution of the
Figure
2. (a)2.Endurance
H
Z0.50.5Ocycles.
O
device;
currents
are read
Figure
(a)
characteristics
ofthe
the
H0.5
0.5Z
2 (HZO)
device;
the the
currents
are read
at 0.2 at
V;0.2 V;
resistance
at Endurance
HRS andcharacteristics
LRS
taken fromof
the
endurance
2 (HZO)
Figure 2. (a) Endurance characteristics of the H0.5Z0.5O2 (HZO) device; the currents are read at 0.2 V;
(b) Set
voltage
distribution
takentaken
from from
the endurance
cycles;
(c) Distribution
of theofresistance
(b)and
SetReset
and Reset
voltage
distribution
the endurance
cycles;
(c) Distribution
the
(b) Set and Reset voltage distribution taken from the endurance cycles; (c) Distribution of the
resistance
at HRS
LRSthe
taken
from the cycles.
endurance cycles.
at HRS
and LRS
takenand
from
endurance
resistance at HRS and LRS taken from the endurance cycles.
Figure 3. Retention characteristics of the HZO device; the currents are read at 0.2 V.
Figure 3. Retention characteristics of the HZO device; the currents are read at 0.2 V.
Figure
3. Retention
characteristicsof
of the
the HZO
currents
areare
read
at 0.2
Figure
3. Retention
characteristics
HZOdevice;
device;the
the
currents
read
at V.
0.2 V.
Materials 2017, 10, 322
Materials 2017, 10, 322
4 of 7
4 of 7
XPS results are shown in Figure 4 to investigate the chemical states of Hf, Zr, and O in the initial
results
shown
in Figure
4 to investigate
the without
chemical top
states
of Hf, Zr,was
andpre
O insputter-etched
the initial
HZO film.XPS
Before
theare
XPS
analysis,
the original
thin film
electrode
HZO
film.
Before
the
XPS
analysis,
the
original
thin
film
without
top
electrode
was
pre
sputter-etched
with Ar ions to remove the surface contaminants. All spectra were calibrated with a C 1s binding energy
with Ar ions to remove the surface contaminants. All spectra were calibrated with a C 1s binding
located
at 248.6 eV. Shirley backgrounds were removed from all spectra, and the oxide-related peaks
energy located at 248.6 eV. Shirley backgrounds were removed from all spectra, and the oxide-related
were fitted with Gaussian–Lorentzian peak functions. The XPS results displayed in Figure 4a show that
peaks were fitted with Gaussian–Lorentzian peak functions. The XPS results displayed in Figure 4a
the 4fshow
5/2 and
4f 7/2 peaks of Hf are with binding energies of 17.85 and 16.21 eV. The intensity ratio
that the 4f 5/2 and 4f 7/2 peaks of Hf are with binding energies of 17.85 and 16.21 eV.
of theThe
components
of the
Hf components
4f doublet and
theHfspin–orbit
are 0.75 and
1.64 are
eV, 0.75
respectively.
intensity ratio
of the
of the
4f doublet splitting
and the spin–orbit
splitting
and
4+
4+ oxidation
The sample
be fittedThe
with
a single
peak
the Hfpeakoxidation
[9]Hf
with
a full width
1.64 eV,could
respectively.
sample
could
be attributed
fitted with to
a single
attributedstate
to the
at halfstate
maximum
0.95 eV. Figure
4bequal
shows
Zr 3d5/2
and
Zr 3d3/2
[9] with a(FWHM)
full widthequal
at halftomaximum
(FWHM)
to that
0.95 eV.
Figure 4b
shows
that Zrpeaks
3d5/2have
andenergies
Zr 3d3/2 of
peaks
haveand
binding
energies
of 181.57 andThe
183.94
eV, respectively.
intervalthese
binding
181.57
183.94
eV, respectively.
energy
interval of The
2.37energy
eV between
of 2.37represents
eV betweenthe
these
two peaks
represents
theThe
oxide
state of Zr4+ [14].oxygen
The non-symmetric
oxygen
two peaks
oxide
state of
Zr4+ [14].
non-symmetric
O 1s peak in
Figure 4c
O
1s
peak
in
Figure
4c
can
been
fitted
into
two
peaks
with
the
binding
energies
of
529.5
and
530.5
eV. with
can been fitted into two peaks with the binding energies of 529.5 and 530.5 eV. The higher peak
The higher peak with a FWHM of 1.13 eV is for 4+
O 1s bonded to the Hf4+, while the lower one with a
a FWHM of 1.13 eV is for O 1s bonded to the Hf , while the lower one with a FWHM of 2.73 eV is
FWHM of 2.73 eV is bonded to the Zr4+. The concentration ratio of Zr/Hf is 0.86, comparable to the
bonded to the Zr4+ . The concentration ratio of Zr/Hf is 0.86, comparable to the original ratio of 1.
original ratio of 1. This means that these are only hafnia and zirconia two oxide in the initial HZO
This means
that these are only hafnia and zirconia two oxide in the initial HZO sample. Hence, the
sample. Hence, the XPS results demonstrate that there is no elemental state Hf or Zr in the sample
XPS results
demonstrate
that elements
there is no
Hf or Zr
in oxide
the sample
and
both
transition
and both
transition metal
areelemental
bonded tostate
the oxygen.
The
forms of
these
metallic
metalconstituents
elements are
oxygen.
The
oxide
forms
of these
are those of
arebonded
those to
of the
HZOs,
which
have
been
reported
to metallic
present constituents
bipolar RS behavior
HZOs,
which have [8,9,13].
been reported to present bipolar RS behavior independently [8,9,13].
independently
Figure
4. X-ray
photoelectron
spectroscopy(XPS)
(XPS) spectra
spectra of
Zr;Zr;
and
(c) (c)
O in
original
Figure
4. X-ray
photoelectron
spectroscopy
of(a)
(a)Hf;
Hf;(b)(b)
and
Othe
in the
original
sample.
HZO HZO
sample.
The typical current–voltage curves of the Au/Ni/HZO/Pt MIM structure in the low voltage
The
typical
curves of the Au/Ni/HZO/Pt MIM structure in the low voltage
region
(V < Vcurrent–voltage
reset) are plotted in log vs. log scale in Figure 5a. The current–voltage curves fitting for
region
(V LRS
< Vand
are are
plotted
in lines,
log vs.
loga scale
Figure
5a. at
The
fitting
reset )HRS
both
straight
while
slope =in0.9
is present
LRScurrent–voltage
and slope = 1.79 iscurves
found at
for both
HRS are
straight
lines, while
a slope considered
= 0.9 is present
at of
LRS
slope =
HRS.LRS
Thisand
indicates
an ohmic
conduction
mechanism
because
theand
formation
of 1.79
a is
conductive
filament
during
the
set
procedure.
However,
for
the
HRS,
the
slope
is
approximately
2,
found at HRS. This indicates an ohmic conduction mechanism considered because of the formation of
which suggests
thatduring
the conduction
at HRS is presumably
dominated
by thethe
trap
control
space charge
a conductive
filament
the set procedure.
However,
for the HRS,
slope
is approximately
2,
limited
conduction
(SCLC)
mechanism
[7].
which suggests that the conduction at HRS is presumably dominated by the trap control space charge
In the high voltage regions (Vreset < V < Vset), the current–voltage curves are plotted in log vs.
limited conduction
(SCLC) mechanism [7].
square root scale, as shown in Figure 5b. It is natural to suggest that the Ni/HZO/Pt structure has a
In the high voltage regions (Vreset < V < Vset ), the current–voltage curves are plotted in log
Schottky nature because the Schottky equation is widely used in the leakage current analysis of the
vs. square
root scale, as shown in Figure 5b. It is natural to suggest that the Ni/HZO/Pt structure
metal/semiconductor structures [14]. Note that the current–voltage curves have a linear relationship
has a between
Schottkylog
nature
because
the Schottky
equation
is widely
in the leakage
analysis
current
and square
root voltage,
which
indicatesused
a Schottky
emissioncurrent
conductive
of themechanism
metal/semiconductor
structures
[14].
Note
that
the
current–voltage
curves
have
a linear
in this voltage region, while the voltage dependence of this mechanism is
relationship between log current and square
root voltage, which indicates a Schottky emission
 / kT )exp  1(V/ d)1/2 
J  region,
AT 2 exp(while
(1)is
conductive mechanism in this voltage
the voltage dependence of this mechanism
i Hence, it appears that the
where J is current density, V is applied voltage, and This temperature.
2
J = of
ATthe
exp
(− ϕ/kT
exp are
β 1 (governed
V/d)1/2 by the Schottky conduction (1)
current–voltage characteristics
HZO
thin )film
mechanism in the high voltage region Vreset < V < Vset.
where J isAccording
current to
density,
V isresults
applied
voltage,
and
T is temperature.
Hence,
appears
the analysis
given
above, the
switching
of the initial HZO
film toitLRS
is due that
to the forming process,
which isof
accompanied
by the
the Schottky
reset process,
the
the current–voltage
characteristics
the HZO thin
filmsoft
arebreakdown.
governed For
by the
conduction
resistance
increased
suddenly,
which
indicates
that
the
filamentary
conducting
paths
might
have
mechanism in the high voltage region Vreset < V < Vset .
According to the analysis results given above, the switching of the initial HZO film to LRS is
due to the forming process, which is accompanied by the soft breakdown. For the reset process, the
Materials 2017, 10, 322
5 of 7
resistance increased suddenly, which indicates that the filamentary conducting paths might have
been ruptured. A joule heating effect caused by the external current is considered for the rupture of
filaments. By increasing the voltage to Vreset , the high current flow through many filaments heated up
the film, which induced a simultaneous rupture of the filaments, and the HRS was achieved. For the
set process,
because
Materials
2017, 10, the
322 sample is composed of stoichiometric ZrO2 and HfO2 , it is impossible
5 of 7 to form
conducting filaments by oxygen vacancies (VO S). This suggests that some reactions might take place
been ruptured.
joule
effect
by the external
current
considered
for the rupture
of can be
in the sample
during A
the
setheating
process.
It iscaused
reasonable
to suggest
thatisan
oxygen-deficient
region
filaments. By increasing the voltage to Vreset, the high current flow through many filaments heated up
formed during the electroforming process. Transition metal cations accommodate this deficiency by
the film, which induced a simultaneous rupture of the filaments, and the HRS was achieved. For the
trapping
injected
the
[17].
In the case
of2 HZO,
setelectrons
process, because
thefrom
sample
is cathodes
composed of
stoichiometric
ZrO
and HfO2, it is impossible to form
conducting filaments by oxygen vacancies (VOS). This suggests that some reactions might take place
−
(4−n)+
M4+ → to
Msuggest
(2)
in the sample during the set process.neIt is+
reasonable
that an oxygen-deficient region can
be formed during the electroforming process. Transition metal cations accommodate this deficiency
where M
Hf and
Zr transition
metal elements.
anode, the oxidation reaction may
bystands
trappingfor
electrons
injected
from the cathodes
[17]. In the At
casethe
of HZO,
lead to the evolution of oxygen gas, according
to
ne  M 4  M ( 4n)
(2)
00
where M stands for Hf and Zr transition
At the, anode, the oxidation reaction may
OO → metal
Vo +elements.
2e− + 1/2O
2
lead to the evolution of oxygen gas, according to
00
(3)
where Vo denotes oxygen vacanciy with a double
with respect to the regular
→ " + 2positive
+ 1/2 charge
,
(3) lattice
and OO represents an oxygen ion on a regular site. The consequent joule heating effect significantly
where " denotes oxygen vacanciy with a double positive charge with respect to the regular lattice
increases
the local generation of VO S at the anodic interface and their drift and diffusion toward the
and O represents an oxygen ion on a regular site. The consequent joule heating effect significantly
cathode. The Ogenerated abundant VO S move to the location where electron injection occurs due to the
increases the local generation of VOS at the anodic interface and their drift and diffusion toward the
electrostatic
force and then accumulate there, leading to an M O2n-1 filament.
The forming process
cathode. The generated abundant VOS move to the location wherenelectron
injection occurs due to the
is complete
after
the
filament
connects
the
cathode
and
anode,
and
the
LRS
is
achieved.
During
the
electrostatic force and then accumulate there, leading to an MnO2n-1 filament. The forming process
is
set process,
the after
incomplete
filament
that
atanode,
the reset
process
modified
by joule
complete
the filament
connects
the ruptured
cathode and
and the
LRS isis
achieved.
During
the setheating.
process,
the incomplete
filament
at the
reset process
is modified
joule
heating.
The external
current
flow warms
upthat
theruptured
thin film,
abundant
generated
VO Sbyare
moved
to The
the virtual
external
current
flow
warms
up
the
thin
film,
abundant
generated
V
OS are moved to the virtual
electrode (remaining incomplete filament), driven by electrostatic force, until the filament is completed
electrode (remaining incomplete filament), driven by electrostatic force, until the filament is
as in the forming process. Because of the compliance current, a conductive filament with a controlled
completed as in the forming process. Because of the compliance current, a conductive filament with
resistance
is formed
composed
of thecomposed
sub-oxides
[24,25].
Considering
the detailstheofdetails
the Set
a controlled
resistance
is formed
of the
sub-oxides
[24,25]. Considering
of procedure,
the
the ruptured
filament
forms
incompletely,
and
takes
the
role
of
virtual
electrode.
Additional
filaments
Set procedure, the ruptured filament forms incompletely, and takes the role of virtual electrode.
filaments
could
be successively
formedby
near
the first
byand
localelectron
joule heating
and due to
could beAdditional
successively
formed
near
the first filament
local
joulefilament
heating
injection
electron density
injection [4].
due Normally,
to a high current
density
[4]. Normally,
multiple
conductive
could regions
a high current
multiple
conductive
filaments
could
lead tofilaments
several HRS
lead to several HRS regions by applying different voltages, which is in accordance with the
by applying
different voltages, which is in accordance with the experimental data (not shown here).
experimental data (not shown here). Therefore, the observed enhanced bistable resistance switching
Therefore,
the observed enhanced bistable resistance switching in HZO samples can be attributed to
in HZO samples can be attributed to the joule heat and the redox reactions induced by the external
the jouleelectron
heat and
the redox reactions induced by the external electron injection.
injection.
(a)
(b)
Figure 5. Current–voltage curve fitting of the HZO device in (a) log vs. log scale and (b) log vs. square
Figure 5.
Current–voltage
curve fitting of the HZO device in (a) log vs. log scale and (b) log vs. square
root
scale.
root scale.
Materials 2017, 10, 322
6 of 7
4. Conclusions
In conclusion, high OFF/ON ratio RS properties of the HZO thin film deposited by PLD have
been investigated for RRAM applications. Each MIM cell demonstrated bistable and reproducible
unipolar RS behaviors. The high ratio of OFF/ON states over 5400 was verified with 450 reversal
cycles. The current–voltage characteristics of the HZO samples showed a Schottky emission conduction
mechanism in the high voltage region (Vreset < V < Vset ), while at the low voltage region (V < Vreset )
the ohmic contact and SCLC are suggested to be responsible for the LRS and HRS, respectively.
Combined with the conductance mechanism, the RS behavior is attributed to the joule heating and the
redox reaction induced by the external electron injection.
Acknowledgments: This work was supported by The National Science Foundation of China (No. 51372030) and
The National Key Research and Development Program of China (No. 2016YFB0700201).
Author Contributions: J. Zhu and Z.P. Wu conviceived and designed the experiments; Z.P. Wu performed the
experiments; J. Zhu and Z.P. Wu analyzed the data; J. Zhu contributed reagents/materials/analysis tools; Z.P. Wu
wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Verrelli, E.; Tsoukalas, D.; Normand, P.; Kean, A.H.; Boukos, N. Forming-free resistive switching memories
based on titanium-oxide nanoparticles fabricated at room temperature. Appl. Phys. Lett. 2013, 102. [CrossRef]
Noman, M.; Sharma, A.A.; Lu, Y.M.; Skowronski, M.; Paul, A.; Salvador, P.A.; Bain, J.A. Transient
characterization of the electroforming process in TiO2 based resistive switching devices. Appl. Phys. Lett.
2013, 102. [CrossRef]
Zou, L.; Hu, W.; Xie, W.; Chen, R.Q.; Qin, N.; Li, B.J.; Bao, D.H. Excellent resistive switching property and
physical mechanism of amorphous TiO2 thin films fabricated by a low-temperature photochemical solution
deposition method. Appl. Surf. Sci. 2014, 311, 697–702. [CrossRef]
Kim, D.C.; Seo, S.S.; Ahn, E.; Suh, D.S.; Lee, M.J. Electrical observations of filamentary conductions for the
resistive memory switching in NiO films. Appl. Phys. Lett. 2006, 88. [CrossRef]
Yu, Q.; Lim, W.M.; Hu, S.G.; Chen, T.P.; Deng, L.J.; Liu, Y. Flexible Nanoscale Memory Device Based on
Resistive Switching in Nickel Oxide Thin Film. Nanosci. Nanotechnol. Lett. 2012, 4, 940–943. [CrossRef]
Wang, X.L.; Shao, Q.; Leung, C.W.; Lortz, R.; Ruotolo, A. Non-volatile, electric control of magnetism in
Mn-substituted ZnO. Appl. Phys. Lett. 2014, 104. [CrossRef]
Zhang, J.; Yang, H.; Zhang, Q.L.; Shurong, D.; Luo, J.K. Bipolar resistive switching characteristics of low
temperature grown ZnO thin films by plasma-enhanced atomic layer deposition. Appl. Phys. Lett. 2013, 102.
[CrossRef]
Wu, Y.L.; Liao, C.W.; Ling, J.J. Effect of current compliance and voltage sweep rate on the resistive switching
of HfO2 /ITO/Invar structure as measured by conductive atomic force microscopy. Appl. Phys. Lett. 2014,
104. [CrossRef]
Sharath, S.U.; Bertaud, T.; Kurian, J.; Hildebrandt, E.; Walczyk, C.; Calka, P.; Zaumseil, P.; Sowinska, M.;
Walczyk, D.; Gloskovskii, A.; et al. Towards forming-free resistive switching in oxygen engineered HfO2-x .
Appl. Phys. Lett. 2014, 104. [CrossRef]
Haemori, M.; Nagata, T.; Chikyow, T. Impact of Cu Electrode on Switching Behavior in a Cu/HfO2 /Pt
Structure and Resultant Cu Ion Diffusion. Appl. Phys. Express 2009, 2. [CrossRef]
Ramadoss, A.; Krishnamoorthy, K.; Kim, S.J. Resistive Switching Behaviors of HfO2 Thin Films by Sol–Gel
Spin Coating for Nonvolatile Memory Applications. Appl. Phys. Express 2012, 5. [CrossRef]
Ye, C.; Zhan, C.; Tsai, T.M.; Chang, K.C.; Chen, M.C.; Chang, T.C.; Deng, T.F.; Wang, H. Low-power bipolar
resistive switching TiN/HfO2 /ITO memory with self-compliance current phenomenon. Appl. Phys. Express
2014, 7. [CrossRef]
Zhang, H.; Gao, B.; Sun, B.; Chen, G.P.; Zeng, L.; Liu, L.F.; Liu, X.Y.; Lu, J.; Han, R.Q.; Kang, J.F.; et al. Ionic
doping effect in ZrO2 resistive switching memory. Appl. Phys. Lett. 2010, 96. [CrossRef]
Wu, X.; Zhou, P.; Li, J.; Chen, L.Y.; Lv, H.B.; Lin, Y.Y.; Tang, T.A. Reproducible unipolar resistance switching
in stoichiometric ZrO2 films. Appl. Phys. Lett. 2007, 90. [CrossRef]
Materials 2017, 10, 322
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
7 of 7
Xu, Z.; Yu, L.; Xu, X.G.; Miao, J.; Jiang, Y. Effect of oxide/oxide interface on polarity dependent resistive
switching behavior in ZnO/ZrO2 heterostructures. Appl. Phys. Lett. 2014, 104. [CrossRef]
Zhang, G.Y.; Lee, D.Y.; Yao, I.C.; Hung, C.J.; Wang, S.Y.; Huang, T.Y.; Wu, J.W.; Tseng, T.Y. Unipolar Resistive
Switching in ZrO2 Thin Films. Jpn. J. Appl. Phys. 2013, 52. [CrossRef]
Waser, R.; Aono, M. Nanoionics-based resistive switching memories. Nat. Mater. 2007, 6, 833–840. [CrossRef]
[PubMed]
Waser, R.; Dittmann, R.; Staikov, G.; Szot, K. Redox-Based Resistive Switching Memories—Nanoionic
Mechanisms, Prospects, and Challenges. Adv. Mater. 2009, 21, 2632–2663. [CrossRef]
Muller, J.; Boscke, T.S.; Schröder, U.; Mueller, S.; Bräuhaus, D.; Böttger, U.; Frey, L.; Mikolajick, T.
Ferroelectricity in Simple Binary ZrO2 and HfO2 . Nano Lett. 2012, 12, 4318–4323. [CrossRef] [PubMed]
Lee, D.Y.; Tsai, T.L.; Tseng, T.Y. Unipolar resistive switching behavior in Pt/HfO2 /TiN device with inserting
ZrO2 layer and its 1 diode–1 resistor characteristics. Appl. Phys. Lett. 2013, 103. [CrossRef]
Lee, J.; Bourim, E.M.; Lee, W.; Park, J.; Jo, M.; Jung, S.; Shin, J.; Hwang, H. Effect of ZrOx /HfOx bilayer
structure on switching uniformity and reliability in nonvolatile memory applications. Appl. Phys. Lett. 2010,
97. [CrossRef]
Huang, C.Y.; Huang, C.Y.; Tsai, T.L.; Lin, C.A.; Tseng, T.Y. Switching mechanism of double forming
process phenomenon in ZrOx /HfOy bilayer resistive switching memory structure with large endurance.
Appl. Phys. Lett. 2014, 104. [CrossRef]
Tulu, B.; Chang, W.Z.; Chu, J.P.; Wang, S.F. Forming-free resistive switching characteristics of 15 nm-thick
multicomponent oxide. Appl. Phys. Lett. 2013, 103. [CrossRef]
Choi, B.J.; Jeong, D.S.; Kim, S.K.; Rohde, C.; Choi, S.; Oh, J.H.; Kim, H.J.; Hwang, C.S.; Szot, K.; Waser, R.; et al.
Resistive switching mechanism of TiO2 thin films grown by atomic-layer deposition. J. Appl. Phys. 2005, 98.
[CrossRef]
Chudnovskii, F.A.; Odynets, L.L.; Pergament, A.L.; Stefanovich, G.B. Electroforming and switching in oxides
of transition metals: The role of metal–insulator transition in the switching mechanism. J. Solid State Chem.
1996, 122, 95–99. [CrossRef]
© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).