Content uploaded by Benjamín Portillo-Rodríguez
Author content
All content in this area was uploaded by Benjamín Portillo-Rodríguez on Jul 27, 2023
Content may be subject to copyright.
Modifying the native aluminum oxide layer by simple
methods for fabricating write-once-read-many memory
devices
B. Portillo-Rodrı
´guez
1,2
,J.D.Sa
´nchez-Va
´squez
1,2
, M. Reyes-Reyes
2
, and R. Lo
´pez-Sandoval
1,
*
1
Advanced Materials Department, IPICYT, Camino a la Presa San José 2055, Col. Lomas 4a sección, San Luis Potosí 78216, México
2
Instituto de Investigación en Comunicación Óptica, Universidad Autónoma de San Luis Potosí, Álvaro Obregón 64,
San Luis Potosí 78000, México
Received: 19 January 2023
Accepted: 7 May 2023
ÓThe Author(s), under
exclusive licence to Springer
Science+Business Media, LLC,
part of Springer Nature 2023
ABSTRACT
In this work, we have fabricated write-once-read-many (WORM) memory devices
based on the modification of the native aluminum oxide (AlOx) layer of the alu-
minum (Al) bottom electrode using different approaches, such as irradiation of a
UV-ozone (UVO) cleaner for 5 min (AlOx-UVO), the exposure for a few seconds of
the native AlOx to a 1 M solution of oxalic acid (AlOx-OA) and a combination of
treatment using the UVO cleaner and the exposure to an oxalic acid solution
(AlOx-UVO-OA). We found that with the different treatments of the native AlOx,
the physical, chemical, and electrical properties of the insulating layer of the Al/
AlOx/Al WORM devices are different. WORM devices based on the AlOx-UVO
and AlOx-OA layers showed an ON/OFF ratio of *1910
2
, while those based on
the AlOx-UVO-OA layer showed an ON/OFF ratio of *1910
4
. Furthermore,
despite presenting a thin insulating layer (B4 nm), the devices did not reveal the
presence of physical deformations on the top electrode. The absence of physical
deformations indicates that the electrical switching mechanism is not based on an
anti-fuse mechanism, suggesting that the mechanism corresponds to a soft
dielectric breakdown related to the creation of oxygen vacancies, with the con-
sequent conductive filamentary paths in the AlOx layer.
1 Introduction
Resistive random-access memories (ReRAMs) devi-
ces based on metal oxides consist of a resistive layer,
the metal oxide, between two metal electrodes, thus
forming a metal-insulator-metal (MIM) structure
[1–4]. ReRAM devices generally require an electro-
forming process, which consists of a soft dielectric
breakdown of the resistive layer. After the electro-
forming process, when a voltage sweep is applied to
the device, the oxygen ions of the insulating layer
Address correspondence to E-mail: sandov@ipicyt.edu.mx
https://doi.org/10.1007/s10854-023-10597-2
J Mater Sci: Mater Electron (2023) 34:1150 (0123456789().,-volV)(0123456789().,-volV)
migrate from the vicinity of one of its electrodes,
leaving vacancies in the active layer and modifying
its conductivity, to the area of the other electrode. The
removal of oxygen ions in the resistive layer leaves
vacancies in the lattice sites of the structure, resulting
in an electrochemical reduction process, which cre-
ates conductive filamentary paths where the resistive
state is decreased (ON state). Then, the oxygen ions
are driven back into the vacancies by applying a
reverse voltage sweep, dissolving the conductive
paths, and switching the devices back to a high-re-
sistance state (OFF state). However, some memory
devices remain ON despite applying a reverse volt-
age to switch them to their OFF state. Such devices
are called write-once-read-many (WORM) data stor-
age [5–7], which implies that the information was
permanently stored in the memory device. WORM
devices are suitable for high-speed permanent file
storage applications for video, images, and other
non-editable databases and low-cost storage devices
such as identification tags [5–8].
Many insulating materials and metal oxides have
been used as active layers in resistive memory devi-
ces. However, metal oxides are very interesting due
to their various characteristics, such as easy fabrica-
tion, low cost, and physical and chemical properties
that can be modified by doping with different ele-
ments [9–17]. In the present work, we focus on one of
the most common metal oxides present in nature,
aluminum oxide [18,19]. The main reason is that
aluminum is one of the most common elements on
the earths surface, and its physical capabilities are
widely used in manufacturing electrical cables and
thin sheets [20,21]. Aluminum (Al), like other metals,
when in contact with oxygen in the environment,
forms a small native oxide film consisting of various
oxidized forms of Al, such as Al hydroxides found in
the outermost layers of the Al film, and alumina
(Al
2
O
3
). Compared with other metals, Al
2
O
3
film
formation is faster and more stable. In addition, Al
oxide has excellent corrosion resistance. Therefore,
Al
2
O
3
is not only a low-cost material for fabricating
memory devices but has played an essential role in
the resistive memory effect [8,22]. The alumina
sometimes contains oxygen vacancies, and the atomic
ratio of Al atoms to oxygen atoms is larger than 2/3.
For this reason, in this work, we use AlOx to refer to
the Al oxide layer on our memory devices. When Al
is used as a metal electrode, this thin layer of native
AlOx acts as an insulating layer that contains defects,
and some electrons can be trapped in this active
layer. Other electrons contribute to the electro-re-
duction process of this insulating layer, generating
the conductive paths responsible for the memory
effect. In a previous work [8], we show that WORM
devices can be fabricated by irradiating the Al bottom
electrode using a UV-ozone (UVO) cleaner for 30 min
to create an insulating AlOx layer of *4 nm. After
electro-formation, physical deformations were
observed on the top electrode, indicating that the
electrical switching mechanism is related to the anti-
fuse mechanisms. These deformations imply that the
electric fields in the device were so strong and the
AlOx layer so thin that the device suffered a dielectric
breakdown. Consequently, the two Al electrodes
fuse, forming an overlap between them, constituting
the conductive paths. Therefore, it is crucial to control
the thickness and roughness of the AlOx film because
if its thickness is small, the electrons can tunnel the
insulator barrier, and the memory effect is affected. In
constrast, if the thickness of the dielectric barrier is
too large, the soft dielectric breakdown becomes very
complicated, and the increased voltage can induce a
catastrophic dielectric breakdown that destroys the
memory device [8,22]. Furthermore, this device
breakage could also be generated by surface rough-
ness, since an active layer with an irregular surface, at
some points between the Al/active layer, can present
a more significant charge accumulation and, there-
fore, induce a dielectric breakdown in the device.
It has been reported that the use of ozone (O
3
) for
the oxidation of Al films generates AlOx films with
small pore sizes, and higher density, and their elec-
trical impedances are ten times higher than growth
films using molecular oxygen (O
2
)[23]. These dif-
ferences in physical and chemical properties between
both types of films using two different forms of
oxygen are because O
3
induces healing of sites with
oxygen vacancy (V
O
) defects since O
3
is much more
energetic than O
2
. On the other hand, it has been
established that for a soft dielectric breakdown to be
possible in AlOx films, they must contain, due to
their manufacturing process, a good number of V
O
s.
For example, the atomic layer deposition of AlOx thin
films with controlled V
O
s has been performed using
water and O
3
as oxidant sources [24]. The ReRAM
devices fabricated using these AlOx films showed
reliable resistive switching behavior, which was
attributed to the formation of stable conductive fila-
ments in the AlOx (water) layer and the partial
1150 Page 2 of 15 J Mater Sci: Mater Electron (2023) 34:1150
connection/interruption of a conductive filament
located in the AlOx (ozone) layer. Additionally,
porous AlOx films made by anodizing Al in oxalic
acid (OA) with high anodizing voltages have been
used to fabricate ReRAM devices [25]. Due to the
manufacturing process of the anodizing AlOx films,
many oxygen vacancies were introduced at the time
of anodizing. In these devices, the resistive switching
mechanism was shown to be based on the injection/
extraction of electrons in the localized V
O
s and not on
the ionic migration. The injection/extraction of elec-
trons is advantageous due to more writing, erasing
cycles, and low electricity consumption. In the pre-
sent work, we manufacture WORM memory by
modifying the AlOx native layer of the bottom elec-
trode either using UVO irradiation for 5 min, expo-
sure of the native AlOx layer to the oxalic acid
solutions for 20 s, or a combination of UVO irradia-
tion and exposure to the oxalic acid solution. We
observed that the devices that presented a larger
ON/OFF ratio first received UVO radiation on the Al
bottom electrode and then were exposed to an oxalic
acid solution for a few seconds.
2 Experimental details
Oxalic acid (H
2
C
2
O
4
C98%) was purchased from
Sigma Aldrich and used without further purification.
Before the Al bottom electrode deposition, 2.5 cm
2
glass substrates were cleaned by washing in acetone,
methanol, and isopropanol in ultrasonic baths for
20 min and then dried for 20 min at 60 °C. The sub-
strates were exposed to UV-ozone radiation for
25 min with a Hybralign series 200 lamp to eliminate
the remains of the organic matter. After that,
approximately 80 nm of Al bottom electrode was
deposited by thermal evaporation in a vacuum
atmosphere at 110
-6
Torr and patterned using a
shadow mask. Four types of Al oxide active layers
were manufactured: native Al oxide (AlOx), Al oxide
by UVO irradiation (AlOx-UVO), Al oxide by UVO
irradiation with deposition of an oxalic acid film
(AlOx-UVO-OA), and Al oxide by deposition of an
oxalic acid film, immediately after Al electrode
evaporation (AlOx-OA). The AlOx layer was syn-
thesized by exposing the Al bottom electrode to the
oxygen from the environment after the Al evapora-
tion process. The AlOx-UVO active layer was syn-
thesized by irradiating the Al bottom electrode with
the UVO lamp for 5 min, after the natural oxidation
by the environment exposition of Al electrodes. In the
Al-UVO-AO active layer, the Al bottom electrode
was irradiated for 5 min with the UVO lamp.
Immediately, 1 M of oxalic acid was deposited by
spin coating (2500 rpm for 20 s) and dried at 60 °C
for 10 min. Furthermore, for the AlOx-OA active
layer, after the exposition of the Al bottom electrode
to the environment, 1 M of oxalic acid was deposited
by spin coating (2500 rpm for 20 s) and dried at 60 °C
for 10 min. Finally, for all devices, a thermally
evaporated *80 nm top electrode of Al in a vacuum
at 110
-6
Torr was deposited onto the Al oxide active
layer and patterned using a shadow mask. A
scheme of the memory device manufacturing process
is shown in Fig. 1a. The active area of the devices
corresponds to the overlapping area between the
bottom and top electrodes, which was *6mm
2
.
Current-voltage (I–V) measurements were performed
using a programmable Keithley 236 source meter at
room conditions. The bias voltage was applied to the
top electrode with the grounded bottom electrode in
all I–Vmeasurements. All the devices were I–
Vcharacterized using voltage sweeps from 0 to 4 V, 4
to -4V,and -4 to 0 V. At least 98% of the fabri-
cated memory devices showed good reproducibility.
The optical images of the Al/modified AlOx/Al
devices were captured using the optical microscope
of a WiTec Raman confocal Alpha-30. For the char-
acterization of the different AlOx active layers, 2.5
cm
2
glass slices were cleaned in acetone, methanol,
and isopropanol ultrasonic baths. Then, the glasses
were exposed to UVO radiation for 25 min. After
that, approximately *80 nm Al film was deposited
by thermal evaporation on the glass substrates in a
vacuum atmosphere at 110
-6
Torr without a pat-
terned shadow mask. Then, the four types of active
layers were designed in the previously described
way. Note that for the characterization of the differ-
ent types of AlOx active layers, the top Al electrode
was not deposited, i.e., the setup was glass/Al bot-
tom electrode/AlOx modified with the different
oxidative treatments. Fourier-transform infrared
spectroscopy (FTIR) measurements of glass/Al bot-
tom electrode/modified AlOx layers were recorded
using a NICOLET iS10 Thermo-Scientific Instrument
in attenuated total reflectance (ATR) mode. X-ray
diffraction (XRD) and grazing incidence X-ray
diffraction (GIXRD) analysis of all the films (Al bot-
tom electrode ?modified AlOx) was carried out
J Mater Sci: Mater Electron (2023) 34:1150 Page 3 of 15 1150
with a SmartLab RIGAKU diffractometer from
2h=10 to 90°, with grazing angle, i.e., the angle
between the film surface and the incident beam,
a=1°. The detector scans in a plane defined by the
incident beam and the normal to the film surface, and
the scanning was performed with a speed of 2°per
min. The diffractometer uses a copper tube for the
X-rays generation (k= 1.54 A
˚) and a NaI scintillation
detector. The morphological characterizations of the
Al bottom electrode ?modified AlOx layers were
Fig. 1 aScheme of the
fabrication of the memory
devices. bI–
VCharacterization of the
memory devices with the
different AlOx active layers:
b
1
Al/AlOx/Al, b
2
Al/AlOx-
UVO/Al, b
3
Al/AlOx-UVO-
OA/Al and b
4
Al/AlOx-OA/Al
1150 Page 4 of 15 J Mater Sci: Mater Electron (2023) 34:1150
performed by atomic force microscopy (AFM) using
an SPM Multimode brand Digital Instruments model
AFM-4 in the contact mode. AFM images were pro-
cessed using WSXM software. X-ray photoelectron
spectroscopy (XPS) measurements of glass/Al
BE/modified AlOx layers were performed using a
monochromatic Al-KaX-ray beam with the JEOL JPS-
9030 photoelectron spectrometer.
3 Results and discussion
I–Vmeasurements were carried out to characterize
the electrical behavior of the different AlOx active
layers. Figure 1b shows the I–Vmeasurements for all
memory devices: (b
1
) Al/AlOx/Al, (b
2
) Al/AlOx-
UVO/Al, (b
3
) Al/AlOx-UVO-OA/Al, and (b
4
) Al/
AlOx-OA/Al. The sweep voltage was performed for
all devices from 0 to 4 V, 4 to -4 V, and -4to0V
without previous electroforming. From Fig. 1b
1
,we
observed that the Al/AlOx/Al devices with the
native AlOx layer presented an ohmic electrical
behavior, i.e., the devices do not exhibit memory-like
behavior. The ohmic behavior indicates that the
thickness of the AlOx layer in these devices is negli-
gible, which allows electrons to flow through the
insulator layer due to the quantum tunneling. Hence,
the electrons jump from the bottom electrode to the
top electrode during the voltage sweep, despite the
presence of the AlOx native layer. In a previous
paper [22], the thickness of this native AlOx layer was
calculated using XPS spectroscopy and found to be
*2.8 nm, although some authors reported it at
*4 nm. In the literature, direct tunneling currents
are observed for Al
2
O
3
film thicknesses B4 nm,
while films with thicknesses larger than C5 nm did
not show direct tunneling [26]. For this reason, the
AlOx native layer on the bottom electrode was
increased using various oxidation processes to avoid
direct tunneling. Memory devices, whose layer
received a post-process treatment to increase the
native Al oxide layer, started in an OFF state at var-
ious resistive levels (I*10
-4
A at 1 V Al/AlOx-
UVO/Al device, I*10
-5
A at 1 V for Al/AlOx-UVO-
OA/Al device and I*10
-4
A at 1 V for Al/AlOx-
OA/Al device) until a threshold voltage is reached,
then the current increases by switching the device to
a permanent ON state (I*10
-1
A at 1 V Al/AlOx-
UVO/Al device, I*10
-1
A at 1 V Al/AlOx-UVO-
OA/Al device and I*10
-2
at 1 V Al/AlOx-OA/Al
device). The Al/AlOx-UVO/Al and Al/AlOx-UVO-
OA/Al devices showed well-defined threshold volt-
age at *2 V. However, in the case of the Al/AlOx-
OA/Al device, the change from the OFF to ON states
is continuous, not showing a jump in the electrical
current during the I–Vmeasurements. When a sec-
ond voltage sweep was carried out in all the devices,
it was confirmed that they presented WORM-type
characteristics. The devices with a post-treatment in
the native oxide layer showed an apparent increase in
their initial resistivity, which is an indirect indicator
of the thickness modification and changes in the
physicochemical properties of the AlOx layer. This
effect of resistivity increasing in the devices can also
be observed when calculating the ON/OFF ratios
(Fig. 2) of the different memory devices, which vary
between the absence of the memory behavior for the
Al/AlOx/Al device up to *4 orders of magnitude
for the Al/AlOx-UVO-OA/Al devices, at a reading
voltage of 1 V.
It is necessary to identify the mechanism respon-
sible for electroforming during the first voltage
sweep to understand the differences in memory-like
behavior of the various modified AlOx films. We
have plotted the I–Vmeasurements of the OFF state
for all devices by using different transport mecha-
nisms. In Fig. 3a and b, and 3c, the log (V) vs. log
(I) curves for the OFF state of all the devices are
shown to find out if the responsible transport mech-
anism is the space charge limited current (SCLC)
mechanism [27–29]. In the case of Al/AlOx-UVO/Al
devices (Fig. 3a), we observe that in the low voltage
regime, in the 0.05–0.9 V range, the device slope is
m*1, corresponding to an ohmic transport. In the
intermediate voltage range (0.9–2 V), the curve slope
is m*2.0, related to the transport of charge carriers
in insulating films free of traps or with shallow traps.
As the voltage sweep is continued, the current
increases abruptly at the threshold voltage (*2 V),
and the slope of the I–Vcurve is m*7.0, indicating
the filling of deep traps in the film with the conse-
quent electrical switching from the OFF to the ON
state. Similarly, the transport mechanisms of Al/
AlOx-UVO-OA/Al and Al/AlOx-OA/Al devices
were analyzed using the SCLC model (Fig. 3b and c).
We observe that for both devices in their OFF states,
the SCLC model describes them more or less ade-
quately. However, in many metal oxides, due to the
presence of oxygen vacancies, the charge transport
mechanism is more closely related to the Schottky
J Mater Sci: Mater Electron (2023) 34:1150 Page 5 of 15 1150
emission mechanism [29,30] or the Frenkel Poole
mechanism [29,31]. For this reason, we have ana-
lyzed the OFF curves of all devices using both models
and have found that the Schottky emission model fits
well over the entire voltage range of the OFF-state
curves of all memory devices, contrary to the SCLC
model in which the OFF-state curve is divided into
various voltage ranges for the curve fitting (Fig. 3a
1
,
3b
1
,3c
1
). In the case of the Al/AlOx-OA/Al devices
(Fig. 3c
1
), the agreement between the experimental
results and the fitting curve using the Schottky
emission model is quite good. These results suggest
the possible increase of oxygen vacancies induced by
the Al oxidation process. The Schottky emission
model is generally used to describe phenomena that
occur at the interfaces between two surfaces, in our
case, the Al electrodes, where AlOx is the interfacial
layer separating both electrodes and whose thickness
is\4 nm. Finally, we observe that the ON state of all
devices is well described by an ohmic model, indi-
cating the formation of conductive filaments.
In previous works reported by our research group,
we have shown that the Al electrodes exposed to an
oxidative medium generated by a UVO cleaner
increased the AlOx native layer from *2.8 to
*4.0 nm after 30 min of UVO exposure [8,22]. The
modification of the native oxide layer due to UVO
irradiation for 30 min was used to fabricate WORM
memories. After the electroforming process, physical
deformations on the top electrode were observed due
to the catastrophic dielectric breaking of the AlOx
dielectric layer. However, in the case of the devices
fabricated in this work, after electroforming during
the first voltage sweep, such damage was not
observed on the top electrode (Fig. S1 in the supple-
mentary information). The deformation absence is
due mainly to the short UVO irradiation time on the
bottom electrode of the Al/AlOx-UVO/Al devices,
which was only 5 min. It is noteworthy that using the
1 M oxalic acid solution deposited on the Al elec-
trodes with and without prior UVO treatment, the
Al/AlOx-UVO-OA/Al and the Al/AlOx-OA/Al
devices, respectively, showed comparable results
with the Al/AlOx-UVO/Al devices. Thus, the elec-
troforming effect does not occur abruptly in the
devices. This lack of structural damage after electro-
forming indicates a more homogeneous or thinner
AlOx insulator layer or other physicochemical chan-
ges occurring in the active layers due to the use of
oxalic acid. To better understand the morphology
obtained in our devices, these were characterized
using atomic force microscopy.
Figure 4and S2 (in the supplementary informa-
tion) show the AFM micrographs of the active layers
Fig. 2 Retention time of the ON and OFF states with a reading
voltage of 1 V, for the devices: aAl/AlOx-UVO/Al, bAl/AlOx-
UVO-OA/Al and cAl/AlOx-OA/Al
1150 Page 6 of 15 J Mater Sci: Mater Electron (2023) 34:1150
Fig. 3 Fitting of the I–Vmeasurements using different transport
mechanisms. Log (V) vs. Log (I) curves for both conductivity
states of: aAl/AlOx-UVO/Al, bAl/AlOx-UVO-OA/Al and cAl/
AlOx-OA/Al devices. Ln(I) vs. V
1/2
curves for the OFF states of:
a
1
Al/AlOx-UVO/Al, b
1
Al/AlOx-UVO-OA/Al and c
1
Al/AlOx-
OA/Al devices
J Mater Sci: Mater Electron (2023) 34:1150 Page 7 of 15 1150
of the different devices. The Al bottom electrode with
the AlOx active layer has an average thickness of
*80 nm (measured using a profilometer) with a root
mean square (RMS) roughness of *21 nm (Fig. 4a),
which shows that the Al film with its native Al oxide
layer is not very uniform. The film roughness of the
deposited Al film is due to the used deposition
technique of the Al electrodes. On the other hand,
when the Al electrodes are exposed to the UVO
oxidative environment, they present interesting
physical changes. In the Al electrode with the AlOx-
UVO active layer, exposed to the UVO oxidative
environment for 5 min, the ozone interacts with the
surface of the Al electrode modifying the morphol-
ogy and uniformity of the film, which now has an
RMS roughness of *12.5 nm (Fig. 4b). Another
effect of the UVO radiation was to make the AlOx
active layer much more homogeneous, where the film
roughness decreased by almost 50%. The decrease in
roughness implies that UVO irradiation interacts
with the surface layers of the bottom electrode, which
are made up of different Al oxides and hydroxides,
eliminating some of the oxygenated and hydroxide
groups linked to Al. In this way, the stoichiometry of
the AlOx layer is modified due to UVO irradiation.
Consequently, the electrical behavior of the AlOx-
UVO active layers (Fig. 1b
2
) is entirely different from
that of the AlOx active layer (Fig. 1b
1
).
It has been reported that the oxidation of Al by O
3
generates Al oxide films with different properties
than naturally generated Al oxide films [23]. More-
over, it was shown that ozone-grown Al oxide films
differ from films grown using O
2
because they have
*4% higher density and exhibit a 10-fold higher
electrical impedance. The differences in the proper-
ties of these Al oxide films are related to the fact that
the oxidation of Al with ozone reduces the oxygen
vacancies, which results in a contraction of its crys-
talline structure and an increase in its electrical
resistivity. In the case of the Al electrode with the
Fig. 4 2D AFM images of the
active layers: aglass/Al
bottom electrode/AlOx,
bglass/Al bottom electrode/
AlOx-UVO, cglass/Al bottom
electrode/AlOx-UVO-OA and
dglass/Al bottom electrode/
AlOx-OA
1150 Page 8 of 15 J Mater Sci: Mater Electron (2023) 34:1150
AlOx-UVO-OA active layer, similar morphology to
the AlOx-UVO devices was obtained, with a much
more homogeneous layer (Fig. 4c). Thus, oxalic acid
modifies the surface roughness of the AlOx layer [32].
However, post-treatment with oxalic acid of the Al
electrodes previously treated with UVO possibly
generated a thicker AlOx layer since they showed a
more resistive state in the OFF state than those only
treated with UVO. Despite the possible increase in
thickness of the AlOx-UVO-OA layer, these devices
showed no signs of structural damage after the
electroforming process (Fig. S1c). It can also be seen
that the electrical behavior of the AlOx-OA devices
shows the presence of an AlOx layer thick enough to
prevent direct electron tunneling. There is, however,
a clear difference with the other types of active layers
since it is the only one that presented an ON state
with I*110
-2
A, indicating a clear difference
between AlOx-OA devices and the others.
In resistive memory devices, it has been reported
that the creation of V
O
s is responsible for their
resistive switching. High constant voltages are gen-
erally used in synthesizing metal oxides with the
anodization method [25,33]. However, in our work,
only an oxalic acid solution was deposited on the Al
bottom electrode (with or without previous UVO
treatment), and this solution was allowed to interact
with the metal oxide (native or modified by UVO) on
the Al bottom electrode for *20 s before the spin
coating deposition. The objective was to study how
various properties, such as the thickness of the alu-
mina layer, the roughness of the surface, the oxida-
tion state of the various AlOx surfaces, the number of
oxide vacancies, and the electrical properties of the
manufactured devices, were modified. The results of
the electrical characterizations (Fig. 1b) and the
morphology studies using AFM (Fig. 4and S2) sug-
gest that the effect of the UVO treatment is to increase
the thickness of the AlOx layer and, possibly, to
remove oxygen vacancies. Thus, it increases the
resistivity of the devices. On the contrary, with oxalic
acid, an increase in the thickness of the AlOx layer is
generated but without the removal and, probably,
with an increase of the oxygen vacancies present in
the AlOx layer. The different properties of the AlOx
layers could explain the differences in their electrical
behavior.
Additionally, for analyzing the differences between
the memory devices, their active layers were char-
acterized using GIXRD (Fig. S3 in the supplementary
information). This technique allows the identification
of the crystalline structure of thin films deposited on
a substrate whose XRD signal, as far as possible, it is
desired to remove. Fig. S3a shows the X-ray diffrac-
tion patterns obtained for the glass/Al bottom elec-
trode/AlOx active layer of all MIM devices. In these
XRD patterns, the peaks at 2h*38°, (111) plane,
2h*44°, (200) plane, 2h*64°, (220) plane, 2h*78°,
(311) plane, and 2h*83°, (222) plane, stand out. These
peaks correspond to the cubic structure of Al (Crys-
tallographic Card 01-072-3440) [34]. However, the
peaks from the thin films of the Al bottom elec-
trode ?modified AlOx have low intensity because
the films are thin (*80 nm). Consequently, the signal
coming from the glass, which is an amorphous
material, corresponds to the broad peak in the
2h= 15–30°region and turns out to be the most
prominent. Moreover, the XRD peaks of Al are broad,
indicating the polycrystallinity of the thin films.
Magnifying the XRD patterns in the 2h= 35–50
°
region (Fig. S3b), we observe that all the active layers
have the same X-ray pattern, regardless of the oxi-
dation treatment to which they were subjected. This
absence of changes in X-ray patterns implies that the
XRD peaks related to AlOx cannot be identified using
GIXRD. The absence of XRD peaks from the modified
AlOx layers can be explained as follows. As previ-
ously discussed, the Al bottom electrode has a
thickness of *80 nm. When this electrode is exposed
to air, a small layer of native AlOx of *2.8 nm is
generated [22]. In this work, we have modified this
AlOx layer by UVO irradiation for 5 min, an oxalic
acid solution deposition for a few seconds, or a
combination of them. In the case of the modification
of the native AlOx using UVO irradiation, it has been
reported that the thickness of this layer increases up
to *4 nm after an irradiation time of 30 min, and
this layer is only slightly modified when the irradi-
ation time increases [22]. Therefore, the AlOx-UVO
layer in the present work is in the range of 2.8-4 nm,
while the thickness of the Al bottom electrode is
*76 nm. On the other hand, oxalic acid has been
used to fabricate AlOx films using an anodizing
process. This anodizing process requires high volt-
ages (*40 V), and the thickness of the AlOx-OA
layer will depend on the anodizing time [25]. In this
work, the native AlOx layer was modified by
depositing a 1 M oxalic acid solution for a few sec-
onds, which allowed the oxalic acid to interact with
the native AlOx. After that, this was removed by spin
J Mater Sci: Mater Electron (2023) 34:1150 Page 9 of 15 1150
coating, and the bottom electrode with its AlOx-OA
surface layer was allowed to dry. We can infer from
the electrical properties of the Al/AlOx-UVO/Al,
Al/AlOx-OA/Al, and Al/AlOx-UVO-OA/Al devi-
ces, and the X-ray diffractograms of glass/Al bottom
electrode/AlOx-UVO, glass/Al bottom electrode/
AlOx-UVO-OA, glass/Al bottom electrode/AlOx-
OA that the layer of Al oxide is very thin. Therefore,
the XRD peaks of the AlOx layer are difficult to
determine from the XRD patterns. Note that although
the AlOx layer is thin, it has crucial effects on the
electrical properties of these devices.
FTIR measurements were used for analyzing the Al
bottom electrode ?modified AlOx layers to explain
the modification in the active layers caused by the
different oxidation treatments. Figure 5shows the
FTIR spectra performed on the different active layers
of AlOx. In the case of the native AlOx layer, only
two IR bands are present at *2300 and *940 cm
-1
.
The band at *2300 cm
-1
corresponds to the
symmetric stretching vibrations associated with CO
2
in the environment, and the 940 cm
-1
bands are
reported as the symmetric stretching vibrations of the
Al–O interaction, which is directly related to the
AlOx layer [35,36]. Likewise, as can be observed in
the case of devices with the AlOx-UVO active layer
(red curve in Fig. 5), the intensity of the band at
940 cm
-1
presents an intensity increase, which indi-
cates that in these active layers, the exposure to UVO
treatment generated a thickness increase in the AlOx
layer, without another chemical change occurring in
the active layer. This result is expected where UVO
cleaners are used to remove organic material from
substrates and increase their hydrophilicity to facili-
tate the adhesion of the materials. Analyzing the
AlOx-UVO-OA active layer (blue curve in Fig. 5),
which was deposited with the oxalic acid solution
after the UVO oxidative treatment, we observed the
same IR bands as those observed in the AlOx-UVO
active layer. In addition, a low-intensity band is
observed at 3620 cm
-1
, which is related to –OH
groups [36]. This band could come from adsorbed
water since the samples remained exposed to the
environment or could be associated with Al–OH
bonds. Furthermore, another IR band is observed at
2050 cm
-1
associated with the –CO bond, which
probably comes from the remains of OA in the AlOx-
UVO-OA thin film [37]. Finally, when analyzing the
AlOx-OA layer (green curve in Fig. 5) in the IR region
between 4000 and 500 cm
-1
, we again observe the
presence of the band at *940 cm
-1
, which is the
band associated with Al–O bonds, and a band at
3620 cm
-1
corresponding to the –OH groups. All the
IR bands found in the different active layers are
shown in Table 1. Although the FTIR technique
allows us to determine the vibrations corresponding
to the different types of bonds in the samples, it does
not allow us to determine if the vibrations come from
Fig. 5 FTIR spectra for the different active layers; aIn the 4000
to 2000 cm
-1
region and bin the 2000 to 500 cm
-1
region
Table 1 Assignment of the observed bands in FTIR spectra for
glass/Al bottom electrode/Al-Ox, glass/Al bottom electrode/Al-
UVO, Al bottom electrode/Al-UVO-OA, and glass/Al bottom
electrode/Al-OA thin films
IR Bands [cm
-1
] Assignment References
3620 0(–OH) [36]
2100 0(–CO) [37]
940 0(Al–O) [35,36]
1150 Page 10 of 15 J Mater Sci: Mater Electron (2023) 34:1150
the atomic bonds on the sample surface or in the
sample bulk.
XPS is a surface technique allowing the determi-
nation of the oxidation states of atoms between 0 and
10 nm from the surface of the sample. Therefore, this
technique allows for determining the physicochemi-
cal properties of the AlOx layers. From the electrical
characterizations, it is observed that the Al bottom
electrode continued to work as a metallic electrode.
Furthermore, the X-ray diffractograms do not show
the crystalline planes corresponding to the AlOx film.
Therefore, differences in electrical behavior should be
related to differences in the physicochemical prop-
erties of the active layers. Thus, the physicochemical
properties of the surfaces of the different active layers
were measured using XPS. In Fig. 6, we present the
O1s XPS spectra of the various active layers of the
memory devices. From Fig. 6a, we observe that the
native AlOx layer presents Al in a hydroxide/
oxyhydroxide (Al–OH) state (532.4 eV) as well as an
Al oxide (Al
2
O
3
) state (531.3 eV) [38]. In the case of
the AlOx-UVO active layer (Fig. 6b), we could ade-
quately fit the XPS measurement using one band
corresponding to the Al–O bonds. The presence of
the Al–O band shows that one effect of UVO treat-
ment is the almost complete removal of Al hydroxide
on the sample surface. For the active layer corre-
sponding to Al-UVO-OA, we obtain the presence of
Al in the hydroxide/oxyhydroxide state and Al
oxide. However, the ratio of the Al
2
O
3
state is higher
than that of the Al–OH state. Something similar
occurs for the Al-OA active layer, but in this case, the
ratio of the Al hydroxide state to the Al oxide state is
almost identical. Therefore, one of the effects of using
the oxalic acid solution is to increase Al hydroxides
on the Al surface. From the XPS characterization, we
observe that the active layer of the AlOx-UVO devi-
ces is primarily alumina due to the removal of V
O
sby
Fig. 6 O1s XPS spectra and their deconvoluted peaks from the different active layers: aglass/Al bottom electrode/AlOx, bglass/Al
bottom electrode/AlOx-UVO, cglass/Al bottom electrode/Al-UVO-OA and dglass/Al bottom electrode/Al-OA
J Mater Sci: Mater Electron (2023) 34:1150 Page 11 of 15 1150
UVO irradiation. Therefore, the MIM devices based
on AlOx-UVO active layer present a voltage thresh-
old in the I–Vcurves (Fig. 1b
2
). In contrast, those
modified using oxalic acid, AlOx-UVO-OA, and Al-
OA, are a mix between alumina and Al hydroxide.
The MIM device based on Al-OA active layer does
not present a voltage threshold in the I–Vcurves due
to many V
O
s, facilitating the electroforming process
(Fig. 1b
4
).
From these results, we see that depending on the
procedure, the properties of the active layers are
entirely different. The differences in the oxidation
states observed between the native and UVO-treated
AlOx layers are related to the interaction of the Al
electrode with the environment and the presence of
contaminants on the Al electrode after their deposi-
tion since UVO cleaners are used to remove organic
contaminants. In addition, due to the use of ozone for
removing organic contaminants, this process increa-
ses the thickness of the AlOx layer and modifies its
physicochemical properties. On the other hand, it has
been reported that oxalic acid interacts differently
depending on the presence of different surface pha-
ses exposed of metal oxides and metals [39,40]. Thus,
the interaction of oxalic acid with the surface of the
Al electrode will depend on the initial conditions
(native AlOx or AlOx-UVO) of the surface of the
active layers. From the results of the electrical char-
acterization, we observe that oxalic acid interacting
directly on Al bottom electrode, without any previ-
ous UVO treatment (native AlOx), generated excel-
lent results. Note that the memory devices based on
the AlOx-OA layer did not show a threshold voltage
and did not reach the compliance current
(I=110
-1
A), which could be related to its layer
thickness, its low roughness, and the increased
presence of Al hydroxide on its surface layer com-
pared to devices based on the other active layers.
4 Conclusions
The methods presented in this work proved to be an
easy and fast technique for modifying the native Al
oxide (AlOx) of the bottom electrode in the different
types of devices (thickness, homogeneity, and oxi-
dation states). The first method consisted of UV-
ozone irradiation for 5 min of the native AlOx layer
on the bottom electrode; in the second method, a 1 M
oxalic acid solution was deposited using spin coating
for 20 s, and immediately the bottom electrode with
the modified AlOx layer was dried at 60 °C for
10 min; the last method was a combination of the
UVO treatment with the deposition of the oxalic acid
solution on the native AlOx layer on the bottom
electrode. Depending on the treatment performed on
the AlOx layer, different physical, chemical, and
electrical properties were obtained on the active layer
of the WORM devices. The devices that showed a
better ON/OFF ratio were the Al/AlOx-UVO-OA/Al
devices with an ON/OFF ratio of *1910
4
.In
contrast, the Al/AlOx-UVO/Al and Al/AlOx-OA/
Al showed similar ON/OFF ratios (*1910
2
),
although their electrical, physical, and chemical
properties differ. The differences in the active layer
properties are related to the type of interaction
between the AlOx layer with UVO and oxalic acid.
The UVO treatment removes the Al hydroxide layer
on the Al surface, while the oxalic acid treatment
increases this layer.
Additionally, the devices did not show physical
deformations on the top electrode, implying that the
mechanism responsible for electrical switching is not
related to the anti-fuse mechanism, which our group
reported in a previous work where the native AlOx
layer received a UVO treatment for 30 min. There-
fore, the mechanism responsible for electrical
switching from the OFF to ON state is related to the
increase in oxygen vacancies (AlOx electro-reduction
process) during the first voltage sweep. Finally, a
suitable method for obtaining an AlOx active layer is
directly using a 1 M solution of oxalic acid on the Al
electrode for a few seconds since exposing the elec-
trodes to a previous UVO treatment was
unnecessary.
Acknowledgements
The authors acknowledge M. Sc. Beatriz A. Rivera, M.
Sc. Ana I. Pen
˜a, Dr. Gladis J. Labrada, Dr. Ignacio G.
Becerril, and Dr. Hector G. Silva-Pereyra for technical
assistance as well as LINAN-IPICYT for providing
access to its facilities. We also acknowledge M. Sc.
Victoria Gonza
´lez-Rodrı
´guez of Polymer Laboratory-
IPICYT and M. Sc. Eva. M. Barrera-Rendo
´n of Labo-
ratory of New Materials and Heterogeneous Envi-
ronmental Catalysis-IPICYT for facilitating the use of
their facilities.
1150 Page 12 of 15 J Mater Sci: Mater Electron (2023) 34:1150
Author contributions
All authors contributed to the study conception and
design. Material preparation, data collection, and
analysis were performed by BPR and JDSV. The first
draft of the manuscript was written by BPR, MRR,
and RLS, and all authors commented on previous
versions. Finally, all authors read and approved the
final manuscript.
Funding
CONACYT supported this work through Research
Scholarships Nos. 722355 (B.P.R) and 707791 (J.D.S.V.).
Data availability
All data generated or analyzed during this study are
included in this published article (and its supple-
mentary information files).
Declarations
Conflict of interest The authors declare that they
have no conflict of interest.
Supplementary Information: The online version
contains supplementary material available at http
s://doi.org/10.1007/s10854-023-10597-2.
References
1. C. Rossel, G.I. Meijer, D. Bre´maud, D. Widmer, Electrical
current distribution across a metal-insulator-metal structure
during bistable switching. J. Appl. Phys. 90, 2892–2898
(2001). https://doi.org/10.1063/1.1389522
2. J.J. Yang, M.D. Pickett, X. Li, D.A.A. Ohlberg, D.R. Stewart,
R.S. Williams, Memristive switching mechanism for metal/
oxide/metal nanodevices. Nat. Nanotechnol. 3, 429–433
(2008). https://doi.org/10.1038/nnano.2008.160
3. L. Qingjiang, A. Khiat, I. Salaoru, C. Papavassiliou, X. Hui,
T. Prodromakis, Memory impedance in TiO
2
based metal-
insulator-metal devices. Sci. Rep. 4, 1–6 (2014). https://doi.
org/10.1038/srep04522
4. S. Slesazeck, T. Mikolajick, Nanoscale resistive switching
memory devices: a review. Nanotechnology. 30, 352003
(2019). https://doi.org/10.1088/1361-6528/ab2084
5. P. Liu, T.P. Chen, X.D. Li, Z. Liu, J.I. Wong, Y. Liu, K.C.
Leong, Realization of write-once-read-many-times memory
device with O
2
plasma-treated indium gallium zinc oxide thin
film. Appl. Phys. Lett. 104, 033505 (2014). https://doi.org/10.
1063/1.4862972
6. S. Wu, X. Chen, L. Ren, W. Hu, F. Yu, K. Yang, M. Yang, Y.
Wang, M. Meng, W. Zhou, D. Bao, S. Li, Write-once-read-
many-times characteristics of Pt/Al
2
O
3
/ITO memory devices.
J. Appl. Phys. 116, 074515 (2014). https://doi.org/10.1063/1.
4893660
7. W. Zhu, J. Li, L. Zhang, X.C. Hu, A reversible bipolar
WORM device based on AlOxNy thin film with Al nano
phase embedded. Solid State Electron. 129, 134–137 (2017).
https://doi.org/10.1016/j.sse.2016.11.014
8. J.A. A
´vila-Nin˜o, M. Reyes-Reyes, O. Nu´n˜ ez-Olvera, R.
Lo´pez-Sandoval, A simple method for fabrication of antifuse
WORM memories. Appl. Surf. Sci. 454, 256–261 (2018). h
ttps://doi.org/10.1016/j.apsusc.2018.05.126
9. M. Das, A. Kumar, B. Mandal, M.T. Htay, S. Mukherjee,
Impact of Schottky junctions in the transformation of
switching modes in amorphous Y
2
O
3
-based memristive sys-
tem. J. Phys. D Appl. Phys. 51, 315102 (2018). https://doi.
org/10.1088/1361-6463/aacf14
10. S. Chandrasekaran, F.M. Simanjuntak, R. Aluguri, T.Y.
Tseng, The impact of TiW barrier layer thickness dependent
transition from electro-chemical metallization memory to
valence change memory in ZrO
2
-based resistive switching
random access memory devices. Thin Solid Films. 660,
777–781 (2018). https://doi.org/10.1016/j.tsf.2018.03.065
11. M. Ismail, S.A. Khan, M.K. Rahmani, J. Choi, Z. Batool,
A.M. Rana, S. Kim, Oxygen annealing effect on resistive
switching characteristics of multilayer CeO
2
/Al/CeO
2
resis-
tive random-access memory. Mater. Res. Express. 7, 016307
(2019). https://doi.org/10.1088/2053-1591/ab61b1
12. A. Kumari, S.M. Shanbogh, I. Udachyan, S. Kandaiah, A.
Roy, V. Varade, A. Ponnam, Interface-driven multifunction-
ality in two-dimensional TiO
2
nanosheet/poly(dimercaptoth-
iadiazole-triazine) hybrid resistive random access memory
device. ACS Appl. Mater. Interfaces. 12, 56568–56578
(2020). https://doi.org/10.1021/acsami.0c16451
13. J.H. Hur, D. Lee, Universal memory characteristics and
degradation features of ZrO
2
-based bipolar resistive memory.
Adv. Electron. Mater. 6, 1–7 (2020). https://doi.org/10.1002/
aelm.202000368
14. S.M. Kim, H.G. Moon, H.S. Lee, Resistive switching char-
acteristics of directly patterned Y-doped CeO
2
by photo-
chemical organic-metal deposition. Ceram. Int. 46,
22831–22836 (2020). https://doi.org/10.1016/j.ceramint.202
0.06.051
J Mater Sci: Mater Electron (2023) 34:1150 Page 13 of 15 1150
15. D.W. Tao, Z.J. Jiang, J.B. Chen, B.J. Qi, K. Zhang, C.W.
Wang, Stable resistive switching characteristics from highly
ordered Cu/TiO
2
/Ti nanopore array membrane memristors.
Appl. Surf. Sci. 539, 148161 (2021). https://doi.org/10.1016/
j.apsusc.2020.148161
16. C.C. Hsu, W.C. Jhang, Resistive switching behavior of tita-
nium oxynitride fabricated using a thermal nitridation pro-
cess. IEEE Electron. Device Lett. 42, 990–993 (2021). h
ttps://doi.org/10.1109/LED.2021.3080328
17. D.W. Kim, H.J. Kim, W.Y. Lee, K. Kim, S.H. Lee, J.H. Bae,
I.M. Kang, K. Kim, J. Jang, Enhanced switching reliability of
sol–gel-processed Y
2
O
3
RRAM devices based on Y2O3
surface roughness-induced local electric field. Materials 15,
1943 (2022). https://doi.org/10.3390/ma15051943
18. M.A. Islam, D.W. Morton, B.B. Johnson, B.K. Pramanik, B.
Mainali, M.J. Angove, Metal ion and contaminant sorption
onto aluminium oxide-based materials: a review and future
research. J. Environ. Chem. Eng. 6, 6853–6869 (2018). h
ttps://doi.org/10.1016/j.jece.2018.10.045
19. H.K. Hami, R.F. Abbas, E.M. Eltayef, N.I. Mahdi, Applica-
tions of aluminum oxide and nano aluminum oxide as
adsorbents: review. Samarra J. Pure Appl. Sci. 2, 19–32
(2021). https://doi.org/10.54153/sjpas.2020.v2i2.109
20. M. Shukla, S.K. Dhakad, P. Agarwal, M.K. Pradhan, Char-
acteristic behaviour of aluminium metal matrix composites: A
review. Mater. Today Proc. 5, 5830–5836 (2018). https://doi.
org/10.1016/j.matpr.2017.12.180
21. K.K. Alaneme, M.H. Adegun, A.G. Archibong, E.A. Okotete,
Mechanical and wear behaviour of aluminium hybrid com-
posites reinforced with varied aggregates of alumina and
quarry dust. J. Chem. Technol. Metall. 54, 1361–1370 (2019)
22. I.A. Rosales-Gallegos, J.A. Avila-Nin˜ o, M. Reyes-Reyes, O.
Nu´n˜ez-Olvera, R. Lo´pez-Sandoval, Effect of the oxidation of
aluminum bottom electrode in a functionalized-carbon nan-
otube based organic rewritable memory device. Thin Solid
Films. 619, 10–16 (2016). https://doi.org/10.1016/j.tsf.2016.
10.046
23. A. Kuznetsova, J.T. Yates, G. Zhou, J.C. Yang, X. Chen,
Making a superior oxide corrosion passivation layer on alu-
minum using ozone. Langmuir. 17, 2146–2152 (2001). http
s://doi.org/10.1021/la001300x
24. K. Park, J.-S. Lee, Reliable resistive switching memory based
on oxygen-vacancy-controlled bilayer structures. RSC Adv.
6, 21736–21741 (2016). https://doi.org/10.1039/
C6RA00798H
25. S. Nigo, M. Kubota, Y. Harada, T. Hirayama, S. Kato, H.
Kitazawa, G. Kido, Conduction band caused by oxygen
vacancies in aluminum oxide for resistance random access
memory. J. Appl. Phys. 112, 033711 (2012). https://doi.org/
10.1063/1.4745048
26. M.D. Groner, J.W. Elam, F.H. Fabreguette, S.M. George,
Electrical characterization of thin Al
2
O
3
films grown by
atomic layer deposition on silicon and various metal substrate.
Thin Solid Films. 413, 186–197 (2002). https://doi.org/10.
1016/S0040-6090(02)00438-8
27. A. Rose, Space-charge-limited currents in solids. Phys. Rev.
97, 1538 (1955). https://doi.org/10.1103/PhysRev.97.1538
28. M.A. Lampert, Simplified theory of space-charge-limited
currents in an insulator with traps. Phys. Rev. 103, 1648
(1956). https://doi.org/10.1103/PhysRev.103.1648
29. F.-C. Chiu, A review on conduction mechanisms in dielectric
films. Adv. Mater. Sci. Eng. 2014, 1–18 (2014). https://doi.
org/10.1155/2014/578168
30. J. Joshua Yang, F. Miao, M.D. Pickett, D.A.A. Ohlberg, D.R.
Stewart, C.N. Lau, R.S. Williams, The mechanism of elec-
troforming of metal oxide memristive switches. Nanotech-
nology. 20, 215201 (2009). https://doi.org/10.1088/0957-
4484/20/21/215201
31. J.G. Simmons, Poole-Frenkel effect and Schottky effect in
metal-insulator-metal systems. Phys. Rev. 155, 657 (1967). h
ttps://doi.org/10.1103/PhysRev.155.657
32. O. Sanz, F.J. Echave, J.A. Odriozola, M. Montes, Aluminum
anodization in oxalic acid: controlling the texture of Al
2
O
3
/Al
monoliths for catalytic applications. Ind. Eng. Chem. Res. 50,
2117–2125 (2011). https://doi.org/10.1021/ie102122x
33. T.W. Hickmott, Voltage-dependent dielectric breakdown and
voltage-controlled negative resistance in anodized Al–Al
2
O
3
–
Au diodes. J. Appl. Phys. 88, 2805 (2000). https://doi.org/10.
1063/1.1287116
34. M. Zhang, V. Kamavaram, R.G. Reddy, New electrolytes for
aluminum production: ionic liquids. Jom. 55, 54–57 (2003). h
ttps://doi.org/10.1007/s11837-003-0211-y
35. A. Va´ zquez, T. Lo´pez, R. Go´ mez, A. Bokhimi, O. Morales,
Novaro, X-ray diffraction, FTIR, and NMR characterization
of sol–gel alumina doped with lanthanum and cerium. J. Solid
State Chem. 128, 161–168 (1997). https://doi.org/10.1006/
jssc.1996.7135
36. R. Zolfaghari, B. Rezai, Z. Bahri, M. Mahmoudian, Influ-
ences of new synthesized active seeds and industrial seed on
the aluminum hydroxide precipitation from sodium aluminate
solution. J. Sustain. Metall. 6, 643–658 (2020). https://doi.
org/10.1007/s40831-020-00302-6
37. M.E. Pandelia, H. Ogata, L.J. Currell, M. Flores, W. Lubitz,
Inhibition of the [NiFe] hydrogenase from desulfovibrio
vulgaris Miyazaki F by carbon monoxide: an FTIR and EPR
spectroscopic study. Biochim. Biophys. Acta-Bioenerg. 1797,
304–313 (2010). https://doi.org/10.1016/j.bbabio.2009.11.
002
38. N. Kumar, K. Biswas, Cryomilling: an environment friendly
approach of preparing large quantity ultra-refined pure
1150 Page 14 of 15 J Mater Sci: Mater Electron (2023) 34:1150
aluminum nanoparticles. J. Mater. Res. Technol. 8, 63–74
(2019). https://doi.org/10.1016/j.jmrt.2017.05.017
39. A. Berna´ , A. Rodes, J.M. Feliu, Oxalic acid adsorption and
oxidation at platinum single crystal electrodes. J. Electroanal.
Chem. 563, 49–62 (2004). https://doi.org/10.1016/j.jelechem.
2003.07.043
40. C.B. Mendive, T. Bredow, M.A. Blesa, D.W. Bahnemann,
ATR-FTIR measurements and quantum chemical calculations
concerning the adsorption and photoreaction of oxalic acid on
TiO
2
. Phys. Chem. Chem. Phys. 8, 3232–3247 (2006). http
s://doi.org/10.1039/b518007b
Publisher’s Note Springer Nature remains neutral with
regard to jurisdictional claims in published maps and
institutional affiliations.
Springer Nature or its licensor (e.g. a society or other
partner) holds exclusive rights to this article under a
publishing agreement with the author(s) or other rightsh-
older(s); author self-archiving of the accepted manuscript
version of this article is solely governed by the terms of
such publishing agreement and applicable law.
J Mater Sci: Mater Electron (2023) 34:1150 Page 15 of 15 1150
A preview of this full-text is provided by Springer Nature.
Content available from Journal of Materials Science: Materials in Electronics
This content is subject to copyright. Terms and conditions apply.
