Lee 

et al. Nanoscale Research Letters (2022) 17:84

https://doi.org/10.1186/s11671-022-03722-3

RESEARCH Open Access

Multi‑level Cells and Quantized Conductance

Characteristics of ­Al2O3‑Based RRAM Device
for Neuromorphic System
Yunseok Lee1, Jongmin Park1, Daewon Chung1, Kisong Lee2 and Sungjun Kim1* 

Abstract 
Recently, various resistance-based memory devices are being studied to replace charge-based memory devices to
satisfy high-performance memory requirements. Resistance random access memory (RRAM) shows superior per-
formances such as fast switching speed, structural scalability, and long retention. This work presented the different
filament control by the DC voltages and verified its characteristics as a synaptic device by pulse measurement. Firstly,
two current–voltage (I–V) curves are characterized by controlling a range of DC voltages. The retention and endur-
ance for each different I–V curve were measured to prove the reliability of the RRAM device. The detailed voltage
manipulation confirmed the characteristics of multi-level cell (MLC) and conductance quantization. Lastly, synaptic
functions such as potentiation and depression, paired-pulse depression, excitatory post-synaptic current, and spike-
timing-dependent plasticity were verified. Collectively, we concluded that Pt/Al2O3/TaN is appropriate for the neuro-
morphic device.
Keywords:  Neuromorphic system, Memristor, Al2O3, Quantized conductance, MLC

Introduction information proceed through parallel chemical interac-

In an environment where data demand is rapidly increas- tions. Information processing and memory capabilities
ing, a breakthrough is needed in computing performance vary depending on various factors such as the size, hold-
limitations due to serial processing of CPU and memory ing time, and a repetition time of external signals and
[1]. It is necessary to change the computing structure and stimuli [5–7].
improve the materials of the memory device to solve the Among various memories, the RRAM exhibits a fast
memory wall. Neuromorphic computing architecture is switching speed and a low operating voltage [8–15]. In
emerging as a structural solution to the bottleneck. The addition, RRAM could be implemented in a simple struc-
neuromorphic computing system mimics the neuron and ture such as a metal-oxide-metal (MIM) with various
synapses of the human brain [2–4]. This system is suita- structural expandability [16–21] such as the connection
ble for the process of complex and unstructured informa- of transistor with each memory cell, an array structure,
tion. First of all, to implement neuromorphic computing, and a 3D vertical structure.
it is necessary to understand how the human brain pro- The switching of RRAM occurs by the formation and
cesses information. The human brain includes numer- rupture of filament in an insulator between the met-
ous synapses and neurons, and learning and memory of als [22–26]. The resistance of RRAM is varied through
a conductive filament composed of oxygen vacancy in
the insulator existing between the top electrode (TE)
*Correspondence: sungjun@dongguk.edu
and bottom electrode (BE) and has two basic switching
1
Division of Electronics and Electrical Engineering, Dongguk University, states (high and low) to process the data storage pro-
Seoul 04620, Republic of Korea
Full list of author information is available at the end of the article
cess [20, 27–30]. In the case of the high-resistance state

© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.
Lee et al. Nanoscale Research Letters (2022) 17:84 Page 2 of 10

(HRS), a low current flows in HRS, and in the case of the of the oxygen vacancy near the TaON/Al2O3 interface
low-resistance state (LRS), it means a state has low resist- [38].
ance and good conductivity. Accordingly, the on/off state In order to confirm the TaON layer, the X-ray photo-
could be monitored through the read voltage. The repeti- electron spectroscopy (XPS) spectra fittings were con-
tion of set and reset processes cause the device to move ducted. Figure 2a shows the Al 2p XPS spectra in which
back and forth between the HRS and LRS, which can be peak intensity is located at 75 eV for Al–O bonding [39].
described as a memory that stores 0 and 1 from a digital Figure  2b and c shows Ta 4f and N 1s XPS peak for the
perspective. TaON layer. In Fig. 2b, small peaks exist at higher binding
In this paper, the gradual resistive switching is con- energy than general Ta 4f XPS peaks. This indicates that
ducted on Pt/Al2O3/TaN device, including A ­ l2O3 high-k the binding Ta–O or Ta–Al energy also affected the Ta
dielectric [31–35], which was deposited by atomic layer 4f XPS peaks with binding Ta–N energy [40, 41]. From
deposition (ALD) equipment. The characteristics using Fig. 2c, through combination with oxygen, N 1s XPS peak
basic DC current sweep and on/off endurance character- shows more biased to higher binding energy than the
istics were measured, and the suitability of neuromorphic normal N 1s peak [42]. As a result, a thin layer of TaON
devices was also measured through synaptic measure- exists between the ­Al2O3 insulator and the TaN BE.
ment, including potentiation, depression, PPD, EPSC, Next, we investigate two types of bipolar resistive
and STDP. switching by DC sweep. All of the above I–V charac-
teristics were measured at a step voltage of 0.01 V. Rep-
Experiments resentative feature of this device shows forming-free
Pt/Al2O3/TaN device was fabricated as follows. Firstly, characteristics in Fig.  3a [43]. The set process has simi-
TaN as BE was deposited by the sputtering system on lar I–V curves as the forming process, and the set pro-
­SiO2/Si wafer. A 5-nm-thick A ­ l2O3 film was deposited cess occurs at − 2  V or higher, and the reset process is
by the ALD process. In the ALD process, TMA precur- induced by applying a 2.75 V. This is referred to as a deep
sors and ­O3 were used at stage temperature 450 °C. Then reset curve. At this time, the on/off ratio is about 45,000
a 100-nm-thick Pt as TE was deposited by evaporator in based on the read voltage of 0.5 V, which is a character-
which the top pattern was formed in a circular pattern by istic due to a large band gap of ­Al2O3. Set shows abrupt
using a shadow mask with a diameter of 100 µm. For the behavior, and in the reset process, it shows a curve that
measurement environment, all measurements were per- returns to the HRS state with a stepwise drop from 1 V
formed at room temperature and ambient atomic pres- or higher to 2.7 V. In the case of Fig. 3b, unlike Fig. 3a, it
sure. Electrical data were measured using the Keithley can be implemented by adjusting reset voltage less than
4200-SCS semiconductor parameter ultrafast module 2.75 V. This is referred to as a partial reset curve, and the
and in pulse mode using a 4225-PMU ultrafast module. on/off ratio at this time is about 13 at the read voltage of
0.5  V. Compared to the I–V curves with fully reset, the
Results and Discussion I–V curves with partial reset process shows more gradual
Figure  1a shows the schematic illustration of the fabri- characteristics in the set and reset processes. Both I–V
cated Pt/Al2O3/TaN device. In Fig.  1b, the cross section characteristics have self-compliance characteristics [44].
of the Pt/Al2O3/TaN RRAM device is inspected by a The method of connecting the two differences in Fig. 3a
transmission electron microscope (TEM). The thickness and b can be confirmed by a continuous DC sweep in
of the ­Al2O3 insulator layer deposited by the ALD system Fig. 3c. The deep reset occurs when the larger voltage is
is about 5  nm. In Fig.  1c, energy dispersion X-ray spec- applied, indicating that the strength of the reset can be
troscopy (EDS) mapping of each element was performed controlled by the voltage adjustment. The current flows
to investigate possible chemical interactions. EDS map- in the HRS induced by the partial reset and an additional
ping shows the spatial distribution of elements in Pt/ reset occur, which lowers the current level due to addi-
Al2O3/TaN. EDS maps of Pt, Al, O, Ta, and N elements tional filament decomposition. Figure 3d exhibits a pos-
were collected in the area shown in the electronic image. sible switching mechanism of partial reset (left) and deep
A region where O and Ta overlap is observed, indicat- reset (right) curves. As confirmed in Fig. 2, Al–O bond-
ing The TaON interface layer between the ­Al2O3 insula- ing has higher binding energy than that Ta–O bonding.
tor and TaN BE is formed by a chemical redox reaction This suggests that switching depends on the TaON layer
between the TaN BE and the lower ­Al2O3 layer due to the when the small electric field is applied and on the A ­ l2O3
strong oxygen binding of TaN [36–38]. Because of the layer when it is a large electric field. Thus, oxygen ions
formation of the TaON interface layer by extracting oxy- formed between TaN and TaON affect the conduction
gen from the A ­ l2O3 layer by TaN, better switching char- mechanism of the device and are estimated to result in
acteristics could be exhibited according to the formation MLC characteristics [3, 45, 46]. Gradual partial reset with
 Lee et al. Nanoscale Research Letters (2022) 17:84 Page 3 of 10

(a) (b)
Pt

Al2O3 5nm

TaN TaN Al2O3 Pt

5 nm

(c)

Fig. 1  a Schematics image of Pt/Al2O3/TaN device, b cross-sectional TEM image, c EDS mapping images of Pt, Al, O, Ta and N elements collected
from the area indicated in the TEM image of the Pt/Al2O3/TaN device

(a) Al 2p (b) Ta 4f (c) N 1s

Raw Raw Raw
Intensity
25.0k Intensity Intensity
Background Background Background
4.0k
Intensity (a.u.)

Al-O 2p 20.0k Ta-N 4f 7/2 Nitrides

Intensity (a.u.)

Intensity (a.u.)

Ta-N 4f 5/2 Oxynitrides

15.0k
Ta-O-N 4f 7/2 5.0k Ta 4p 3/2
Ta-O-N 4f 5/2
Al-Ta-O 4f 7/2
2.0k 10.0k Al-Ta-O 4f 5/2
Ta-O 4f 7/2
Ta-O 4f 5/2
5.0k
0.0 0.0 0.0
78 76 74 72 32 30 28 26 24 22 20 410 405 400 395
Binding energy (eV) Binding energy (eV) Binding energy (eV)
Fig. 2  XPS spectra of a Al 2p, b Ta 4f and c N 1s of the device
Lee et al. Nanoscale Research Letters (2022) 17:84 Page 4 of 10

(a) (b) (c)

-2 -2
10 10
10-3 10-3 10-3
10-4 10-4 10-4
10-5 10-5
Current (A)

Current (A)

Current (A)
10-5
10-6 10-6
10-6
10-7 10-7
10-8 10-8 10-7
1st
10-9 10-9 10-8
2nd
10-10 10-10 10-9
-3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3 0.0 0.5 1.0 1.5 2.0 2.5
Voltage (V) Voltage (V) Voltage (V)

(d) (e) (f)

Pt Pt 10-3 -1.5

Set voltage (V)

Current (A)

-1.0
10-4
Al2O3 Al2O3
TaON TaON -0.5
10-5

0.0
TaN TaN 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
Partial reset Deep reset Cell (#) Cell (#)
Fig. 3  Bipolar resistive switching of Pt/Al2O3/TaN device: a deep reset, b partial reset, c process of inducing from partial reset to deep reset, and d
deep and partial reset mechanism schematic diagram. Uniformity of e HRS and LRS, f set voltage

MLC occurs in the TaON layer within the − 2.2 V region. on the reset voltage at the boundary between the par-
However, the more electric field induces the filament tial reset and deep reset, the reset process was repeat-
decomposition inside the A ­ l2O3 and causes the abrupt edly measured while increasing 0.025  V from 1.8 to
current decrease during the reset process. In Fig. 3e, HRS 2.35 V. It could be verified that the current level gradually
and LRS were confirmed in the read operation of 0.5 V to decreases, and this could prove the existence of various
demonstrate state uniformity. Since the filament decom- multi-level states.
position depends on the magnitude of reset voltage, HRS The property of conductance quantization [49–52]
varies more severely than LRS. Also, more decomposi- was confirmed. This is thought to be due to the quanti-
tion demands more set voltage to re-form the filaments. zation effect of conductive filament during the reset pro-
Variation of set voltage is shown in Fig.  3f and it varied cess. When the conductive filament is well controlled, it
from − 1.25 to − 0.75 V in accordance with the previous is possible to implement more state and higher density
reset cycle process. memory through this phenomenon. As shown in Fig. 5a,
In Fig.  4a, the endurance characteristics were also this phenomenon can be observed when the conductive
measured for partial I–V conditions using pulse for ­105 filament is modified in atomic units. The step voltage
cycles. It shows that HRS and LRS can be switched even of 0.002  V and delay time of 0.3  s every step is used to
at ­105 or more times. In Fig.  4b, it is the result of per- observe quantization in multiple cycles, and only elemen-
forming the retention test for each I–V characteristic tal disruption of the filament was measured during the
including partial and deep resets. HRS and LRS were reset process. The conductance quantum, represented
measured at the read voltage of 0.15  V, and both states by the symbol G0, is the quantized unit of electrical con-
were maintained for ­104  s. These results show the Pt/ ductance. It is defined by the elementary charge e and
Al2O3/TaN device has good non-volatile memory prop- Planck constant h as G0 = 2e2/h = 7.74809 × ­10–5 S. The
erties. Multi-level cell characteristics are very beneficial device takes an integer multiple of G0 or an intermedi-
for practical applications such as high-density memory ate value between integers. In the end, LRS is changed to
and neuromorphic device [43, 47, 48]. Figure  4c shows HRS. The statistical analysis is essential through multi-
a reset process by increasing the reset voltage by 0.2  V ple cycles [53–57]. Figure  5b shows the histogram plot-
for each cycle. Through this process, as the reset volt- ting, and it can be seen that even in various conductance
age increases, multiple HRS is achieved. In Fig. 4d, based steps, there is a high tendency near a multiple of G0 or
 Lee et al. Nanoscale Research Letters (2022) 17:84 Page 5 of 10

(a) (b)
10-3

10-4

Conductance (S)
10-5 Full HRS
Full LRS
Partial HRS
10-6 Partial LRS

10-7

10-8
102 103 104

(c) (d) Time (s)

10-3
Current (A)

10-4

Reset MLC
10-5 1.8V to 2.35V
@+0.025V

0.0 0.5 1.0 1.5 2.0

Voltage (V)
Fig. 4  a Endurance characteristic of the device for ­105 cycles and b each deep and partial HRS and LRS retention characteristics for 1­ 04 s. MLC
characteristics of c different reset voltages, d consecutive incremental voltage repetition

a half multiple [58–61]. It is noted the values between A neuromorphic computing system can be imple-
0.5G0 and 3G0 are distinctly distinguishable. It may be mented using multi-level cells in Pt/Al2O3/TaN devices.
necessary to make the conducting filament smaller by As shown in Fig.  6a, the conductive filament connect-
means of a method such as making the device smaller in ing the TE and BE of RRAM can be expressed very
order to distinguish the quantized values. Pulse measure- similarly to the human’s biological system [43, 65–67].
ments were performed in Fig.  5c and d to describe the In order to confirm the suitability of neuromorphic
quantized conductance [62–64]. Conductance calculated computing, pulse measurements were conducted. In
with the voltage of 0.5  V was induced by adding write Fig. 6b, conductance control is continuously performed
pulses at 0.5 s intervals. In Fig. 5c, an incremental write through 5 cycles of potentiation and depression by
pulse increased by − 25 mV from − 0.7 to − 1.775 V was applying the pulses. Potentiation and depression were
used and the abrupt set operation occurred at a voltage set to − 1.15  V and 1.3  V, respectively, and both pulse
of − 1  V or higher. The conductance in HRS increases widths were set to 10  sµ. From the I–V characteristic
more than 10G0 at a time due to the abrupt characteristic of the set process, relatively abruptness in the poten-
in the set region. This characteristic was also confirmed tiation can be confirmed. It could be verified that the
in the I–V curve in the inset image of Fig. 5c, which MLC depression part has a more gradual characteristic.
implemented by limiting compliance current. In contrast, Moreover, we demonstrate more gradual and sym-
conductance quantization with the erase pulses com- metric resistance-change characteristics by control-
posed by 25 mV from 1.5 to 2.175 V were ranged of about ling the voltage amplitude of pulses in Fig. 6c [68, 69].
G0. From those two different conductance ranges show Each 6 potentiation and depression segments are used
that it is more ease to implement MLC during reset pro- to increase and decrease the conductance. The voltage
cess due to the clear state distinction. varied from − 0.9 to − 1.4 V for potentiation and from
Lee et al. Nanoscale Research Letters (2022) 17:84 Page 6 of 10

(a) (b)

10-3
20
Conductance (S)

Counts (#)
5G0
4G0
3G0
2.5G0
2G0
10
1.5G0
10-4
G0

0.5G0
0
1.0 1.5 1 2 3 4 5 6 7
Voltage (V) G/G0
(c) (d)
22
20 4
18
16 3
14
12
G0

G0
-3
10
Current (A)

10 2
10-4
8
6 10-5
1
4
2 -1.0
Voltage (V)
-0.5 0.0

0 0
10 15 20 0 5 10
Time (s) Time (s)
Fig. 5  a Quantized conductance characteristic using DC measurement on reset part, and b histogram of quantized conductance levels. Quantized
conductance characteristic using pulse measurement c in set region, d in reset region

(a) Top
Pre-synapse Electrode

Conducting
Synaptic Filament
Information

Bottom
Post-synapse Electrode

(b) (c) (d)

1.5 Set -1.15 V 1.2
-0.9 V
80
Conductance (mS)

Conductance (mS)

-1 V
Reset 1.3 V -1.1 V
70
Accuracy (%)

Pulse width 10 s -1.2 V

1.0 -1.3 V
-1.4 V 60
1.0 1.35 V
0.8 1.65 V 50
1.95 V
2.25 V
2.55 V
40
0.6 2.85 V
Set -0.9 V to -1.4 V 30
0.5 gradual
Reset 1.35 V to 2.85 V
0.4 20 abrupt
Pulse width 10 μs
10
0 100 200 300 400 500 0 20 40 60 80 100 120 0 3 6 9 12 15 18 21
Pulse Number (#) Pulse Number (#) Epochs (#)
Fig. 6  a Schematic illustration of similarity between human synaptic neural structure and RRAM device structure. Potentiation and depression
characteristics: b consecutive same voltage pulse, c single incremental voltage pulse. d MNIST pattern recognition simulation results of (b) and (c)
 Lee et al. Nanoscale Research Letters (2022) 17:84 Page 7 of 10

1.35 to 2.85 V for depression. Figure 6d shows MNIST increases, indicating that the device is suitable for imple-
pattern recognition simulation results by using the con- menting STP. Figure  7b illustrates conductance changes
ductance results of Fig. 6b and c [70, 71]. The result of before and after giving five identical write pulses and
using Fig. 6c shows higher accuracy for each epoch. In summarizes them with pulse amplitudes. As the voltage
other words, pulse improvement measurement pro- amplitude increases, both potentiation and depression
vides a better learning process. have a larger synaptic weight change. Continuous stim-
Synaptic functions, such as PPD, EPSC, and STDP ulation raises EPSC; the degree of weight strengthening
measurements, were performed to determine suit- can be adjusted according to the amplitude. The strength
ability for the neuromorphic application [72–74]. Fig- of connections between neurons in biological synapses
ure  7a shows the device’s PPD measurement data, the can be controlled by STDP. Therefore, if we can elucidate
ratio change between two pulses was confirmed when the detailed mechanisms of biological synaptic action and
the seven different intervals were used. Synaptic weight imitate the action behavior, it will be possible to mimic
changed with the time interval ranging from 20  μs to the energy-efficient processing of the human brain. Fig-
5  ms between two consecutive depression pulses. The ure  7c explains the configuration of the STDP protocol.
amount of synaptic weight change was expressed as When the pre-spike signal and post-spike signal, which
ΔW = (A2 − A1)/A1 × 100 (%). As a result, the current vary with the interval, were applied to the biological
responded by the second pulse decreases as the interval synapses, the weight was changed and implemented

Fig. 7  a PPD measurement, b EPSC data according to amplitude. STDP characteristics: c Schematic for measurement imitation between synaptic
neural structure and RRAM, d pulse authorization for STDP measurement at Δt = 60 μs and e result of STDP measurement
Lee et al. Nanoscale Research Letters (2022) 17:84 Page 8 of 10

according to the learning behavior. This process was Competing interests

The authors declare they have no competing interests.
mimicked on the memristor in the same way. The pulse
protocol in Fig. 7d was used for the measurements. The Author details
same pre and post-signal were composed, but the differ- 1
 Division of Electronics and Electrical Engineering, Dongguk University,
Seoul 04620, Republic of Korea. 2 Department of Information and Communica-
ent shape of pulses was finally configured and applied tion Engineering, Dongguk University, Seoul 04620, Republic of Korea.
according to the interval. Since the final pulse configura-
tion was different, the synaptic weights over time had dif- Received: 26 June 2022 Accepted: 24 August 2022
ferent weight changes, as shown in Fig. 7e [2, 75, 76]. In
general, the shorter the absolute time of the spike time
difference, the greater the change in conductance change
like a biological synapse. References
1. Wang R, Yang J-Q, Mao J-Y et al (2020) Recent advances of volatile mem-
ristors: devices, mechanisms, and applications. Adv Intell Syst. https://​doi.​
Conclusions org/​10.​1002/​aisy.​20200​0055
2. Kim S, Kim H, Hwang S et al (2017) Analog synaptic behavior of a silicon
As a result, the MLC characteristics and quantized con- nitride memristor. ACS Appl Mater Interfaces 9:40420–40427. https://​doi.​
ductance were confirmed through the A ­l2O3-based org/​10.​1021/​acsami.​7b111​91
RRAM device deposited with ALD, and excellent biologi- 3. Ismail M, Abbas H, Sokolov A et al (2021) Emulating synaptic plasticity
and resistive switching characteristics through amorphous ­Ta2O5 embed-
cal characteristics were investigated through pulse meas- ded layer for neuromorphic computing. Ceram Int 47:30764–30776.
urement. DC I–V bipolar switching characteristics were https://​doi.​org/​10.​1016/j.​ceram​int.​2021.​07.​257
verified through DC measurement, and it was verified 4. Li Q, Tao Q, Chen Y et al (2021) Low voltage and robust InSe memristor
using van der Waals electrodes integration. Int J Extreme Manuf. https://​
that switching characteristics of two different characteris- doi.​org/​10.​1088/​2631-​7990/​ac2296
tics could be easily controlled only by adjusting a voltage. 5. Ryu H, Kim S (2020) Self-rectifying resistive switching and short-term
Multi-levels in various cases were confirmed by varying memory characteristics in Pt/HFO2/TaOx/TiN artificial synaptic device.
Nanomaterials 10:2159. https://​doi.​org/​10.​3390/​nano1​011215
the amount of voltage that adjusts different characteris- 6. Park M, Kang M, Kim S (2021) Pulse frequency dependent synaptic
tics, and the conductance quantization phenomenon was characteristics in Ta/SiN/Si memristor device for neuromorphic system. J
also confirmed within the reset section and pulse meas- Alloys Compd 882:160760. https://​doi.​org/​10.​1016/j.​jallc​om.​2021.​160760
7. Ryu JH, Mahata C, Kim S (2021) Long-term and short-term plasticity of
urements. This MLC phenomenon was connected with ­Ta2O5/HfO2 memristor for hardware neuromorphic application. J Alloys
pulse measurement to measure potentiation and depres- Compd. https://​doi.​org/​10.​1016/j.​jallc​om.​2020.​156675
sion, and it was possible to maximize MLC characteris- 8. Kim D, Kim S, Kim S (2021) Logic-in-memory application of CMOS com-
patible silicon nitride memristor. Chaos Solitons Fract 153:11154. https://​
tics through voltage control of each segment. Including doi.​org/​10.​1016/j.​chaos.​2021.​111540
PPD and EPSC, through the measurement of STDP, the 9. Lin KL, Hou TH, Shieh J et al (2011) Electrode dependence of filament
change in the conductance weight of the device was con- formation in ­HfO2 resistive-switching memory. J Appl Phys 109:084104.
https://​doi.​org/​10.​1063/1.​35679​15
firmed by imitating the synapse. In conclusion, the MLC 10. Rodriguez-Fernandez A, Aldana S, Campabadal F et al (2017) Resistive
characteristics of the device and the suitability of neuro- switching with self-rectifying tunability and influence of the oxide layer
morphic computing were successfully completed. thickness in Ni/HfO2/n+-Si RRAM devices. IEEE Trans Electron Devices
64:3159–3166. https://​doi.​org/​10.​1109/​TED.​2017.​27174​97
Acknowledgements 11. Hu G, An H, Xi J et al (2021) A ZnO micro/nanowire-based photonic
Not applicable. synapse with piezo-phototronic modulation. Nano Energy 89:106282.
https://​doi.​org/​10.​1016/j.​nanoen.​2021.​106282
Author Contributions 12. Khan SA, Lee GH, Mahata C et al (2021) Bipolar and complementary resis-
All authors read and approved the final manuscript. tive switching characteristics and neuromorphic system simulation in a
Pt/ZnO/TiN synaptic device. Nanomaterials 11:315. https://​doi.​org/​10.​
Funding 3390/​nano1​10203​15
This research was supported in part by National R&D Program through the 13. Sun J, Tan JB, Chen T (2020) Investigation of electrical noise signal trig-
National Research Foundation of Korea (NRF) funded by Ministry of Science gered resistive switching and its implications. IEEE Trans Electron Devices
and ICT (2021K1A3A1A49098073) and in part by Korea Institute of Energy 67:4178–4184. https://​doi.​org/​10.​1109/​TED.​2020.​30148​41
Technology Evaluation and Planning (KETEP) grant funded by the Korea Gov- 14. Shin J, Kang M, Kim S (2021) Gradual conductance modulation of Ti/WOx/
ernment (MOTIE) under Grant 20224000000020. Pt memristor with self-rectification for a neuromorphic system. Appl Phys
Lett 119:012102. https://​doi.​org/​10.​1063/5.​00534​78
Availability of Data and Materials 15. Chen ZX, Fang Z, Wang Y et al (2014) Impact of Ni concentration on the
All data generated or analysed during this study are included in this article. performance of Ni silicide/HfO2/TiN resistive RAM (RRAM) cells. J Electron
Mater 43:4193–4198. https://​doi.​org/​10.​1007/​s11664-​014-​3309-9
16. Luo Q, Xu X, Gong T et al (2018) 8-Layers 3D vertical RRAM with excellent
Declarations scalability towards storage class memory applications. In: Technical
digest—international electron devices meeting, IEDM
Ethics Approval and Consent to Participate 17. Yu S, Chen HY, Gao B et al (2013) ­HfOx-based vertical resistive switching
Not applicable. random access memory suitable for bit-cost-effective three-dimensional
cross-point architecture. ACS Nano 7:2320–2325. https://​doi.​org/​10.​1021/​
Consent for Publication nn305​510u
Not applicable. 18. Al-Haddad A, Wang C, Qi H et al (2016) Highly-ordered 3D vertical resis-
tive switching memory arrays with ultralow power consumption and
 Lee et al. Nanoscale Research Letters (2022) 17:84 Page 9 of 10

ultrahigh density. ACS Appl Mater Interfaces 8:23348–23355. https://​doi.​ 41. Misha SH, Tamanna N, Woo J et al (2015) Effect of nitrogen doping on
org/​10.​1021/​acsami.​6b054​24 variability of ­TaOx-RRAM for low-power 3-bit MLC applications. ECS Solid
19. Yu M, Cai Y, Wang Z et al (2016) Novel vertical 3D structure of ­TaOx-based State Lett 4:P25–P28. https://​doi.​org/​10.​1149/2.​00115​04ssl
RRAM with self-localized switching region by sidewall electrode oxida- 42. Cristea D, Cunha L, Gabor C et al (2019) Tantalum oxynitride thin films:
tion. Sci Rep 6:1–10. https://​doi.​org/​10.​1038/​srep2​1020 assessment of the photocatalytic efficiency and antimicrobial capacity.
20. Jeong DS, Thomas R, Katiyar RS et al (2012) Emerging memories: resistive Nanomaterials 9:476. https://​doi.​org/​10.​3390/​nano9​030476
switching mechanisms and current status. Rep Prog Phys 75:076502 43. Ismail M, Mahata C, Kim S (2022) Forming-free Pt/Al2O3/HfO2/HfAlOx/
21. Lanza M, Wong HSP, Pop E et al (2019) Recommended methods to study TiN memristor with controllable multilevel resistive switching and neuro-
resistive switching devices. Adv Electron Mater 5:1800143 morphic characteristics for artificial synapse. J Alloys Compd 892:162141.
22. Waser R, Aono M (2007) Nanoionics-based resistive switching memories. https://​doi.​org/​10.​1016/j.​jallc​om.​2021.​162141
Nat Mater 6:833–840. https://​doi.​org/​10.​1038/​nmat2​023 44. Ryu H, Kim S (2021) Gradually modified conductance in the self-com-
23. Sawa A (2008) Resistive switching in transition metal oxides. Mater Today pliance region of an atomic-layer-deposited Pt/TiO2/HfAlOx/TiN rram
11:28–36 device. Metals 11:1199. https://​doi.​org/​10.​3390/​met11​081199
24. Yu S (2014) Overview of resistive switching memory (RRAM) switching 45. Chen MC, Chang TC, Chiu YC et al (2013) The resistive switching char-
mechanism and device modeling. In: Proceedings—IEEE international acteristics in TaON films for nonvolatile memory applications. Thin Solid
symposium on circuits and systems Films 528:224–228. https://​doi.​org/​10.​1016/j.​tsf.​2012.​09.​081
25. Li YT, Long SB, Liu Q et al (2011) An overview of resistive random access 46. Ismail M, Abbas H, Mahata C et al (2022) Optimizing the thickness of
memory devices. Chin Sci Bull 56:3072–3078 ­Ta2O5 interfacial barrier layer to limit the oxidization of Ta ohmic interface
26. Choi J, Kim S (2020) Improved stability and controllability in ZrN-based and ­ZrO2 switching layer for multilevel data storage. J Mater Sci Technol
resistive memory device by inserting ­TiO2 layer. Micromachines 11:905. 106:98–107. https://​doi.​org/​10.​1016/j.​jmst.​2021.​08.​012
https://​doi.​org/​10.​3390/​mi111​00905 47. Lin J, Wang S, Liu H (2021) Multi-level switching of al-doped ­HfO2 RRAM
27. Kim S, Chen J, Chen YC et al (2019) Neuronal dynamics in H ­ fOx/AlOy- with a single voltage amplitude set pulse. Electronics (Switzerland)
based homeothermic synaptic memristors with low-power and homo- 10:731. https://​doi.​org/​10.​3390/​elect​ronic​s1006​0731
geneous resistive switching. Nanoscale 11:237–245. https://​doi.​org/​10.​ 48. Wu J, Ye C, Zhang J et al (2016) Multilevel characteristics for bipolar resis-
1039/​c8nr0​6694a tive random access memory based on hafnium doped ­SiO2 switching
28. Zhang K, Sun K, Wang F et al (2015) Ultra-low power Ni/HfO2/TiOx/TiN layer. Mater Sci Semicond Process 43:144–148. https://​doi.​org/​10.​1016/j.​
resistive random access memory with sub-30-nA reset current. IEEE Elec- mssp.​2015.​12.​012
tron Device Lett 36:1018–1020. https://​doi.​org/​10.​1109/​LED.​2015.​24642​ 49. Li Y, Long S, Liu Y et al (2015) Conductance quantization in resistive
39 random access memory. Nanoscale Res Lett 10:1–30
29. Sun QQ, Gu JJ, Chen L et al (2011) Controllable filament with electric field 50. Celano U, Goux L, Belmonte A et al (2015) Understanding the dual nature
engineering for resistive switching uniformity. IEEE Electron Device Lett of the filament dissolution in conductive bridging devices. J Phys Chem
32:1167–1169. https://​doi.​org/​10.​1109/​LED.​2011.​21597​70 Lett 6:1919–1924. https://​doi.​org/​10.​1021/​acs.​jpcle​tt.​5b006​33
30. Kwon O, Kim Y, Kang M, Kim S (2021) Comparison of synaptic properties 51. Sharath SU, Vogel S, Molina-Luna L et al (2017) Control of switching
considering dopant concentration and device operation polarity in Cu/ modes and conductance quantization in oxygen engineered ­HfOx based
SiN/SiO2/p-Si devices for neuromorphic system. Appl Surf Sci 563:150101. memristive devices. Adv Funct Mater 27:1700432. https://​doi.​org/​10.​
https://​doi.​org/​10.​1016/j.​apsusc.​2021.​150101 1002/​adfm.​20170​0432
31. Vishwanath SK, Woo H, Jeon S (2018) Enhancement of resistive switching 52. Park J, Lee S, Lee K, Kim S (2021) Conductance quantization behavior in
properties in ­Al2O3 bilayer-based atomic switches: Multilevel resistive pt/sin/tan rram device for multilevel cell. Metals 11:1918. https://​doi.​org/​
switching. Nanotechnology 29:235202. https://​doi.​org/​10.​1088/​1361-​ 10.​3390/​met11​121918
6528/​aab6a3 53. Sun Y, Wen D (2018) Conductance quantization in nonvolatile resistive
32. Chen L, Gou HY, Sun QQ et al (2011) Enhancement of resistive switch- switching memory based on the polymer composite of zinc oxide nano-
ing characteristics in ­Al2O3-based RRAM with embedded ruthenium particles. J Phys Chem C 122:10582–10591. https://​doi.​org/​10.​1021/​acs.​
nanocrystals. IEEE Electron Device Lett 32:794–796. https://​doi.​org/​10.​ jpcc.​8b011​20
1109/​LED.​2011.​21257​74 54. Gao S, Zeng F, Chen C et al (2013) Conductance quantization in a Ag
33. Chen L, Xu Y, Sun QQ et al (2010) Highly uniform bipolar resistive switch- filament-based polymer resistive memory. Nanotechnology 24:335201.
ing with ­Al2O3 buffer layer in robust NbAlO-based RRAM. IEEE Electron https://​doi.​org/​10.​1088/​0957-​4484/​24/​33/​335201
Device Lett 31:356–358. https://​doi.​org/​10.​1109/​LED.​2010.​20411​83 55. Zhu X, Su W, Liu Y et al (2012) Observation of conductance quantization
34. Kim S, Park BG (2016) Nonlinear and multilevel resistive switching in oxide-based resistive switching memory. Adv Mater 24:3941–3946.
memory in Ni/Si3N4/Al2O3/TiN structures. Appl Phys Lett 108:212103. https://​doi.​org/​10.​1002/​adma.​20120​1506
https://​doi.​org/​10.​1063/1.​49527​19 56. Mehonic A, Vrajitoarea A, Cueff S et al (2013) Quantum conductance in
35. Chen C, Pan F, Wang ZS et al (2012) Bipolar resistive switching with self- silicon oxide resistive memory devices. Sci Rep 3:1–8. https://​doi.​org/​10.​
rectifying effects in Al/ZnO/Si structure. J Appl Phys 111:013702. https://​ 1038/​srep0​2708
doi.​org/​10.​1063/1.​36728​11 57. Gao S, Chen C, Zhai Z et al (2014) Resistive switching and conductance
36. Wang Z, Yaegashi O, Sakaue H et al (2003) Suppression of native oxide quantization in Ag/SiO2/indium tin oxide resistive memories. Appl Phys
growth in sputtered TaN films and its application to Cu electroless plat- Lett 105:063504. https://​doi.​org/​10.​1063/1.​48932​77
ing. J Appl Phys 94:4697–4701. https://​doi.​org/​10.​1063/1.​16096​44 58. Long S, Lian X, Cagli C et al (2013) Quantum-size effects in hafnium-oxide
37. Zhou Q, Zhai J (2014) Study of the bipolar resistive-switching behaviors in resistive switching. Appl Phys Lett 102:183505. https://​doi.​org/​10.​1063/1.​
Pt/GdOx/TaNx structure for RRAM application. Physica Status Solidi A Appl 48022​65
Mater Sci 211:173–179. https://​doi.​org/​10.​1002/​pssa.​20133​0098 59. Tsuruoka T, Hasegawa T, Terabe K, Aono M (2012) Conductance quantiza-
38. Zhou P, Yin M, Wan HJ et al (2009) Role of TaON interface for ­CuxO resis- tion and synaptic behavior in a ­Ta2O5-based atomic switch. Nanotechnol-
tive switching memory based on a combined model. Appl Phys Lett ogy 23:435705. https://​doi.​org/​10.​1088/​0957-​4484/​23/​43/​435705
94:053510. https://​doi.​org/​10.​1063/1.​30788​24 60. Chen C, Gao S, Zeng F et al (2013) Conductance quantization in oxygen-
39. Fang R-C, Sun Q-Q, Zhou P et al (2013) High-performance bilayer flexible anion-migration-based resistive switching memory devices. Appl Phys
resistive random access memory based on low-temperature thermal Lett 103:043510. https://​doi.​org/​10.​1063/1.​48167​47
atomic layer deposition. Nanoscale Res Lett 8:1–7. https://​doi.​org/​10.​ 61. Lv H, Xu X, Sun P et al (2015) Atomic view of filament growth in electro-
1186/​1556-​276x-8-​92 chemical memristive elements. Sci Rep 5:1–8. https://​doi.​org/​10.​1038/​
40. Zhou J, dan Nie D, Jin XB, Xiao W (2020) Controllable nitridation of ­Ta2O5 srep1​3311
in molten salts for enhanced photocatalysis. Int J Miner Metall Mater 62. Milano G, Aono M, Boarino L et al (2022) Quantum conductance in
27:1703–1710. https://​doi.​org/​10.​1007/​s12613-​020-​2050-z memristive devices: fundamentals, developments, and applications. Adv
Mater. https://​doi.​org/​10.​1002/​adma.​20220​1248
Lee et al. Nanoscale Research Letters (2022) 17:84 Page 10 of 10

63. Zhao X, Xu J, Xie D et al (2021) Natural acidic polysaccharide-based

memristors for transient electronics: highly controllable quantized con-
ductance for integrated memory and nonvolatile logic applications. Adv
Mater 33:1–10. https://​doi.​org/​10.​1002/​adma.​20210​4023
64. Banerjee W, Hwang H (2019) Quantized conduction device with 6-bit
storage based on electrically controllable break junctions. Adv Electron
Mater 5:1–10. https://​doi.​org/​10.​1002/​aelm.​20190​0744
65. Strukov DB, Kohlstedt H (2012) Resistive switching phenomena in thin
films: Materials, devices, and applications. MRS Bull 37:108–114
66. Zhang SR, Zhou L, Mao JY et al (2019) Artificial synapse emulated by
charge trapping-based resistive switching device. Adv Mater Technol
4:1800342. https://​doi.​org/​10.​1002/​admt.​20180​0342
67. Rahmani MK, Ismail M, Mahata C, Kim S (2020) Effect of interlayer on resis-
tive switching properties of ­SnO2-based memristor for synaptic applica-
tion. Results Phys 18:103325. https://​doi.​org/​10.​1016/j.​rinp.​2020.​103325
68. Park J, Ryu H, Kim S (2021) Nonideal resistive and synaptic characteristics
in Ag/ZnO/TiN device for neuromorphic system. Sci Rep 11:1–7. https://​
doi.​org/​10.​1038/​s41598-​021-​96197-8
69. Cho H, Kim S (2020) Enhancing short-term plasticity by inserting a thin
­TiO2 layer in ­WOx-based resistive switching memory. Coatings 10:908.
https://​doi.​org/​10.​3390/​COATI​NGS10​090908
70. Ismail M, Mahata C, Kwon O, Kim S (2022) Neuromorphic synapses with
high switching uniformity and multilevel memory storage enabled
through a Hf-Al-O alloy for artificial intelligence. ACS Appl Electron Mater.
https://​doi.​org/​10.​1021/​acsae​lm.​2c000​23
71. Lee Y, Mahata C, Kang M, Kim S (2021) Short-term and long-term synaptic
plasticity in Ag/HfO2/SiO2/Si stack by controlling conducting filament
strength. Appl Surf Sci 565:150563. https://​doi.​org/​10.​1016/j.​apsusc.​2021.​
150563
72. Feng G, Jiang J, Zhao Y et al (2020) A sub-10 nm vertical organic/inor-
ganic hybrid transistor for pain-perceptual and sensitization-regulated
nociceptor emulation. Adv Mater 32:1–11. https://​doi.​org/​10.​1002/​adma.​
20190​6171
73. Li Y, Yin K, Diao Y et al (2022) A biopolymer-gated ionotronic junctionless
oxide transistor array for spatiotemporal pain-perception emulation in
nociceptor network. Nanoscale 14:2316–2326. https://​doi.​org/​10.​1039/​
d1nr0​7896h
74. Jiang J, Hu W, Xie D et al (2019) 2D electric-double-layer phototransistor
for photoelectronic and spatiotemporal hybrid neuromorphic integra-
tion. Nanoscale 11:1360–1369. https://​doi.​org/​10.​1039/​c8nr0​7133k
75. Shen Z, Zhao C, Qi Y et al (2020) Advances of RRAM devices: Resistive
switching mechanisms, materials and bionic synaptic application. Nano-
materials 10:1437
76. Lu K, Li Y, He WF et al (2018) Diverse spike-timing-dependent plastic-
ity based on multilevel HfOx memristor for neuromorphic comput-
ing. Appl Phys A Mater Sci Process 124:1–9. https://​doi.​org/​10.​1007/​
s00339-​018-​1847-3

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in pub-
lished maps and institutional affiliations.