Flexible Nonvolatile Cu/CuxO/Ag ReRAM Fabricated Using Ink-Jet Printing Technology 
 
by 
 
Simin Zou 
 
 
 
 
A thesis submitted to the Graduate Faculty of 
Auburn University 
in partial fulfillment of the 
requirements for the Degree of 
Master of Science 
 
Auburn, Alabama 
May 4, 2014 
 
 
 
 
Keywords: ReRAM, Nonvolatile Memory, Memristor, Flexible Memory, Ink-jet Printing 
 
Copyright 2014 by Simin Zou 
 
 
Approved by 
 
Michael C. Hamilton, Chair, Assistant Professor of Electrical and Computer Engineering 
 Guofu Niu, Alumni Professor of Electrical and Computer Engineering 
Robert Dean, Associate Professor of Electrical and Computer Engineering
ii 
 
 
 
 
 
 
 
Abstract 
 
 
 The design and fabrication of flexible Cu/CuxO/Ag ReRAM devices is presented in this 
work. We have investigated Cu/CuxO/Ag capacitor-like structure which exhibits bipolar resistive 
switching behavior under low-range direct current sweep in the temperature range from 255K to 
355K. Ink-jet printing technology is used to fabricate the device. The device displays a resistive 
switching ratio of more than 30 between high resistance state and low resistance state at room 
temperature. The device displays a resistive switching ratio of more than 20 between HRS and 
LRS over 100 cycles.  Memory states are reproducible and remained over 500 times. The device 
has the ability to operate well after over 1000 flexes. The good ductility of printed silver and 
electroplated copper electrodes and simple cross-point structure of the memory cell result in 
good flexibility and mechanical robustness, which indicates great potential for future flexible 
memory applications. 
 The physical mechanism of Cu/CuxO/Ag ReRAM is also presented. The temperature-
dependent I-V measurement and cell-size dependent test reveal that the bipolar resistive 
switching behavior is attributed to the formation and rupture of the conducting filament paths in 
the insulating material. Furthermore, a resistive switching physical model of the device is 
presented.  
 
 
 
 
 
iii 
 
 
 
 
 
 
 
Acknowledgments 
 
 
 First and foremost, the author would like to express her great gratitude and thanks to her 
academic advisor Dr. Michael C. Hamilton for providing significant guidance, assistance, and 
financial support throughout this work and her graduate studies. Dr. Niu and Dr. Dean are also 
thanked for serving in her committee. Their comments and advices are very valuable and helpful.  
  The author would also like to thank Dr. Michael Bozack for his help with the X-ray 
photoelectron spectroscopy (XPS) analysis for the CuxO thin film and thank Mr. Charles Ellis, 
Pingye Xu and George A. Hernandez for their assistance in many aspects of this work. 
 A personal note of appreciation goes to Xiaohui Zou, Yao Ma and Di Zhang for their 
support, care and love throughout the author?s academic endeavors. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
iv 
 
 
 
 
 
 
 
Table of Contents 
 
 
Abstract ......................................................................................................................................... ii 
Acknowledgments ....................................................................................................................... iii 
List of Tables ............................................................................................................................... vi 
List of Figures ............................................................................................................................. vii 
List of Abbreviations ................................................................................................................... ix 
1      Introduction ........................................................................................................................... 1 
2      Background   ......................................................................................................................... 5 
    2.1    Classification of Semiconductor Memory ....................................................................... 5 
    2.2    Resistive Random-Access Memory ................................................................................. 8 
             2.2.1    Classification of Resistive Switching Behaviors ................................................. 8 
             2.2.2    Mechanism for Resistive Switching .................................................................... 9 
             2.2.3    Flexible Resistive Switching Memory Devices ................................................. 10 
3      Fabrication of Cu/CuxO/Ag ReRAM Devices .................................................................... 13 
4      Device Characterization and Testing .................................................................................. 17 
    4.1    X-ray Photoelectron Spectroscopy ................................................................................ 17 
    4.2    Current-Voltage Characteristics of Cu/CuxO/Ag ReRAM Devices .............................. 19 
             4.2.1    Electroforming Process ...................................................................................... 19 
             4.2.2    I-V Characteristics after Electroforming ........................................................... 21 
             4.2.3    Temperature-Dependent I-V Characteristics ..................................................... 23 
v 
 
4.3    Device HRS and LRS Yields ......................................................................................... 24 
    4.4    Switching Endurance and Data Retention Performance ................................................ 25 
    4.5    Mechanical Bending Tests and Flex Tests..................................................................... 27 
             4.5.1    Mechanical Bending Tests ................................................................................. 27 
             4.5.2    Flex Tests ........................................................................................................... 29 
    4.6    Switching Time Tests .................................................................................................... 31 
5      Resistive Switching Mechanism ......................................................................................... 34 
    5.1    Non-Cu-Oxidation Control Group ................................................................................. 34 
    5.2    Physical Mechanism of Cu/CuxO/Ag ReRAM Devices ................................................ 37 
    5.3    Resistive Switching Mechanism Model......................................................................... 43 
6      Conclusion and Future Work .............................................................................................. 45 
References   ................................................................................................................................. 47 
Appendix A    Processing Equipment and Materials .................................................................. 54 
 
 
 
 
 
 
 
 
 
 
vi 
 
 
 
 
 
 
 
List of Tables 
 
 
Table 4.1    XPS surface elemental composition (at%) versus sputter etching time  ................. 19 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
vii 
 
 
 
 
 
 
 
List of Figures 
 
 
Figure 1.1     Cu/CuxO/Ag ReRAM devices fabricated on flexible Kapton substrate .................. 4 
Figure 1.2     Fabricated Cu/CuxO/Ag ReRAM array ................................................................... 4 
Figure 2.1     Classification of standard semiconductor memories ............................................... 7 
Figure 2.2     I-V curves of resistive switching in ReRAM devices ............................................. 9 
Figure 2.3     The conductivity mapping results of TiO2 at (a) LRS and (b) HRS ..................... 10 
Figure 2.4     Photograph of different flexible ReRAM devices ................................................. 12 
Figure 3.1     Schematic fabrication flow for the Cu/CuxO/Ag ReRAM devices ....................... 14 
Figure 3.2     Photographic process flow for Cu/CuxO/Ag ReRAM array ................................. 15 
Figure 3.3     The SEM image of the cross section of the Cu/CuxO/Ag ReRAM device ........... 16 
Figure 4.1     The XPS spectra of CuxO thin film before sputter etching ................................... 18 
Figure 4.2     The XPS spectra of CuxO thin film after 2 min or 4 min sputter etching ............. 19 
Figure 4.3     Electrical forming process and I-V characteristics of the device .......................... 21 
Figure 4.4     R-V characteristics of the Cu/CuxO/Ag ReRAM .................................................. 22 
Figure 4.5     Multi-cycle I-V characteristics for the 1st, 250 th and 500 th cycles .................... 23 
Figure 4.6     Temperature dependence of I-V curves in semi-log scale .................................... 24 
Figure 4.7     Distribution of HRS and LRS of fabricated devices ............................................. 25 
Figure 4.8     Switching endurance test of the device during 100 cycles .................................... 26 
Figure 4.9     Data retention performance of the device ............................................................. 26 
Figure 4.10     Photograph of the device bent at radius = 11.5mm ............................................. 28 
viii 
 
Figure 4.11     Resistance of HRS and LRS as a function of the bending radius ....................... 28 
Figure 4.12     VSET and VRESET as a function of the bending radius .......................................... 29 
Figure 4.13     Resistance of HRS and LRS during 1~1000 flexes ............................................. 30 
Figure 4.14   The Cu/CuxO/Ag ReRAM device was bent into a convex shape ......................... 30 
Figure 4.15   Switching time measurements of (a) RET and (b) RESET operation ................... 32 
Figure 5.1     Hysteretic I-V behavior of the Cu/CuxO/Ag ReRAM device ............................... 35 
Figure 5.2     I-V curve of a non-oxidation device ...................................................................... 35 
Figure 5.3    Resistance dependence of HRS and LRS on number of switching cycles ............. 36 
Figure 5.4    Temperature dependence of HRS, IRS and LRS ................................................... 38 
Figure 5.5    Temperature-dependent I-V characteristics at HRS in log-log scale ..................... 39 
Figure 5.6    The LnI versus 1/KT curve for HRS ...................................................................... 40 
Figure 5.7    Activation energy at various voltages .................................................................... 40 
Figure 5.8     Logarithmic plot of I-V characteristic of positive voltage region ......................... 42 
Figure 5.9     Photograph of the fabricated different cell-size ReRAM array ............................. 42 
Figure 5.10   Resistance of HRS and LRS versus cell size of Cu/CuxO/Ag ReRAM devices ... 43 
Figure 5.11   Schematic diagram of the resistive switching mechanism model ......................... 44 
Figure A.1     FUJIFILM Dimatix Material Printer .................................................................... 54 
Figure A.2     Kapton polymide substrate ................................................................................... 54 
Figure A.3     Silver nanoparticle ink .......................................................................................... 55 
Figure A.4     Hot plate  .............................................................................................................. 55 
Figure A.5     YES (Yield Engineering Systems) vacuum cure oven ......................................... 56 
Figure A.6     Copper electroplating equipment .......................................................................... 56 
Figure A.7     Keithley 4200 (Semiconductor Characterization Systems) parameter analyzer .. 57 
ix 
 
Figure A.8     Micro-manipulated cryogenic probe system ........................................................ 57 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
x 
 
 
 
 
 
 
 
List of Abbreviations 
 
 
BRS Bipolar Resistive Switching 
HRS  High Resistance State 
IRS            Initial Resistance State 
LRS Low Resistance State 
MOM Metal-Oxide-Metal  
RAM         Random-Access Memory 
ReRAM Resistive Random-Access Memory 
ROM Read-Only Memory 
SEM Scanning Electron Microscope  
URS Unipolar Resistive Switching 
 
 
 
1 
 
 
 
 
 
 
 
Chapter 1 
Introduction 
 
In the modern world, semiconductor nonvolatile memory devices, such as programmable 
read-only memory (PROM) and Flash memory, have been successfully scaled down to achieve 
high-capacity memories by perfecting the lithography technology. However, how to satisfy the 
increasing demand for the scaling of conventional memory devices is the most serious problem 
encountered due to the physical limitation. Compared to conventional memories, ReRAMs have 
attracted wide attention due to their various characteristics, such as high switching speed, high 
data density, low power consumption, excellent scalability and simple structure. ReRAMs have 
been regarded as one of the promising candidates of the next generation nonvolatile memory. 
Moreover, ReRAMs have neuromorphic and biological circuit applications, signal processing 
and programmable logic applications. 
A ReRAM device consists of a two-terminal metal-oxide-metal (MOM) sandwich cross 
structure. Scientists such as Leon Chua has pointed that all two-terminal nonvolatile memory 
devices including ReRAM should be considered memristors, which are the missing non-linear 
passive electrical components relating electric charge (Q) and magnetic flux linkage (?). There 
are two kinds of resistive switching behaviors reported in ReRAM, which are unipolar resistive 
switching (URS) and bipolar resistive switching (BRS). The switching direction of URS depends 
on the applied voltage amplitude whereas BRS depends on the polarity of the applied voltage. 
The capacitor-like memory cell is characterized by four resistive states: initial resistance state 
2 
 
(IRS), electroformed resistance state (ERS), high resistance state (HRS) and low resistance state 
(LRS). Reversible resistive switching between HRS and LRS can be achieved by applying either 
a current or a voltage bias to the memory cell. The MOM cell shows hysteretic current-voltage 
(I-V) characteristics and this hysteretic behavior keeps the resistive switching in HRS or LRS 
unless the voltage bias reaches or exceeds the threshold voltages, which are called RESET 
voltage and SET voltage, respectively. Since most of resistive switching materials are insulators, 
IRS is the most resistive state of other states. For most ReRAM devices, a pretreatment process 
named electroformed process is essential, leading to a great increase in the conductance by 
applying a high current or voltage to the memory cell at IRS. Generally, ERS has a low 
resistance, which is comparable to LRS and IRS has higher resistance than HRS. 
Resistive switching is a significant physical effect in the ReRAM operation. The resistive 
switching effect has been studied for more than 45 years. Research in addressing ReRAM has 
been investigated by an increasing number of research groups [1-8]. For example, Hewlett-
Packard laboratories have studied Pt/Organic monolayer/Ti memory devices with a tunable 
resistor more than 102-105 ? range under current or voltage control [2]. Gergel-Hackett, et al. 
have investigated flexible Al/TiO2/Al solution-processed memory devices with an ON/OFF ratio 
of greater than 10000:1 and long data retention time of over 1.2?106 s [3]. Jin-Woo Han, et al. 
fabricated resistive switching function embedded e-texile with Cu/CuxO/Pt sandwich structure 
[4].  
Up to now, various resistive switching models have been suggested, including filament-type 
model [5], space charge limited current (SCLC) model [6], Schottky barrier model [7], oxygen 
migration model [8] and so on. However, these physical mechanisms cannot completely explain 
the universal resistive switching characteristics.  
3 
 
In this research, we have investigated the characterization of the BRS in Cu/CuxO/Ag 
(1<x<2) memory devices for flexible ReRAM applications. Our measurements consist of two-
point probe DC measurement. The material we used to fabricate memory devices are copper and 
silver, which are relatively abundant in the natural world. We use ink-jet printing technology 
during the fabrication process which has extensive potential for roll-to-roll processing in 
industrial production. In addition, the entire fabrication processing we used is low temperature 
processing, which are lower than 200 ?C. Kapton polyimide film is used as the substrate, 
allowing a good flexibility of the memory devices. Fig 1.1 shows a photograph of fabricated 
Cu/CuxO/Ag memory cells at the bending condition. Fig 1.2 shows the Cu/CuxO/Ag ReRAM 
array. 
The fabricated ReRAM devices have the characteristics of high resistive switching 
repeatability of more than 500 times, low-voltage operation (less than 1 V), reliable switching 
endurance of over 100 cycles, data retention for nearly 2 weeks, low-temperature fabrication, 
HRS/LRS ratio of greater than 20 under room temperature and operated well after being flexed 
over 1000 times. Furthermore, to explore the physical mechanism of the BRS characteristics in 
the device, temperature dependence of resistive switching and cell-size dependent measurement 
have been investigated, which reveals the conductive filament path has the most probability to 
lead to the BRS behavior of Cu/CuxO/Ag ReRAM devices. 
 
 
 
 
 
4 
 
 
Figure 1.1: Cu/CuxO/Ag ReRAM devices fabricated on flexible Kapton substrate. 
 
 
 
Figure 1.2: Fabricated Cu/CuxO/Ag ReRAM array 
 
 
 
 
 
 
5 
 
 
 
 
 
 
 
Chapter 2 
Background 
 
2.1 Classification of Semiconductor Memory 
As an important electronic data storage device, semiconductor memories have sustained 
increasing advances in the integrated circuit technologies. In addition, portable electronics are 
heavily reliant on the development of the fast-speed, high-density, low-power consumption low-
cost and simple-manufacture-technique memories. Basically, semiconductor memories can be 
divided into random access memories (RAMs) and read-only memories (ROMs). Compared to 
RAMs, ROMs are designed to hold permanent data and instead of being written to, ROMs are 
normally read-only.  
RAMs can be further classified as volatile RAMs and nonvolatile RAMs. The volatile 
RAMs include dynamic RAMs (DRAMs) and static RAMs (SRAMs) and the volatile RAMs 
lose stored data once the power to the memory chip is down. DRAMs are usually used as the 
main memory in the computer for their feature of high density and low cost. SRAMs are faster 
and less memory refresh required, which are mostly used for cache memories in the computer. 
Depending on different materials or operating principles, nonvolatile RAMs are classified as 
ferroelectric RAMs (FeRAMs) [9], magnetoresistive RAMs (MRAMs) [10], phase-change 
RAMs (PRAMs) [11], flash memory and resistive RAMs (ReRAMs) [1-8]. FeRAM is a kind of 
RAM which has similar construction to DRAM. However, FeRAM has a thin film of 
ferroelectric material (lead zirconate titanate) different from the dielectric layer in a DRAM.  
6 
 
Compared to conventional RAMs, MRAMs store data by magnetic storage elements not current 
flow or electric charge. The magnetic storage element is formed by two ferroelectric plates, 
holding two magnetic fields and sandwiching an insulating layer. PRAM is also known as 
PCRAM and C-RAM, exploiting its unique characteristics of chalcogenide glass. ReRAMs have 
cross-point architecture and a fairly large number of organic and inorganic material systems 
show resistive switching effects. The physical mechanism of ReRAMs have not been explained 
clearly now. Mechanism model such as filament-like model [5], oxygen migration model [6], 
Schottky barrier model [7], space charge limited current (SCLC) model [8] have been published. 
ROMs can be classified as mask ROMs (MROMs), programmable ROMs (PROMs), 
erasable programmable ROMs (EPROMs), electrically erasable programmable ROMs 
(EEPROMs) and ultraviolet erasable EPROMs (UVEPROMs). Fig 2.1 shows the categories of 
semiconductor memories.  
7 
 
 
 
 
 
Figure 2.1: Classification of standard semiconductor memories. 
 
8 
 
2.2 Resistive Random-Access Memory 
ReRAM is a kind of emerging nonvolatile memory. It is based on various new materials, 
such as organic compounds [2], metal oxide [3, 4], showing a resistive switching behavior. The 
ReRAM memory cells have a capacitor-like architecture composed of semiconducting or 
insulating material sandwiched between top and bottom electrodes.  
Hysteretic current-voltage (I-V) characteristic in metal-oxide-metal (MOM) structure of 
Al/Al2O3/Al was first published by T. W. Hickmott in the year 1962 [12]. His research indicated 
that resistive switching was induced by the applied electric fields. Since then, a variety of metal 
oxides have been subsequently reported showing resistive switching, such as TiO2 [3], NiO [13, 
14] and so on.  
 
2.2.1 Classification of Resistive Switching Behaviors 
The resistive switching behavior of the ReRAM can be categorized into two types: BRS 
and URS. In BRS, the resistive switching depends on the polarity of the applied voltage bias. 
This type of resistive switching behavior occurs with various semiconducting oxides, such as: 
HfO2 [15], CuxO [16], SrTiO3 [17], Ta2O5 [18], Pr0.7Ca0.3MnO3 [19], Si3N4 [20], Nb2O5 [21], 
CeO2 [22]. In URS, the resistive switching direction relies on the amplitude of the applied 
voltage bias. In other words, the different resistance states of the system can be switched by 
application of continuous electric stress of the same or the opposite polarities. The URS has been 
reported in materials such as: CoO [23], SnO2 [24], MgO [25], Lu2O3 [26] and so on. 
Interestingly, materials such as: TiO2 [27], ZnO [1, 28], BST [29], NiO [13, 14], ZrO2 [30], 
YMnxO3 [31] have been observed exhibiting the coexistence of BRS and URS. Fig 2.2 shows 
9 
 
the I-V characteristics of URS in a Pt/NiO/Pt memory cell [32] and BRS in a Al/TiO2/Al 
memory cell [3].  
 
Figure 2.2: I-V curves of resistive switching in ReRAM devices. (a) URS in a Pt/NiO/Pt cell [32]. 
(b) BRS in a Al/TiO2/Al cell [3]. 
 
2.2.2 Mechanism of Resistive Switching 
The physical origin of the resistance variation in the resistive switching materials is not 
fully understood yet, despite the fact that this variation is observed in every ReRAM reported. 
Generally, there are two basic driving mechanism, one is filament-type resistive switching, the 
other is interface-type resistive switching.  
 In filament-type resistive switching mechanism, the resistive switching is attributed to the 
formation and rupture of the conductive filament paths in the insulating material. The filament?
type resistive switching model was first published in 1967 by J. G. Simmons and R .R. Verderber 
[33]. In some previous works, high spatial resolution provided the evidence of the formation and 
rupture of the conductive paths [34, 35]. Fig 2.3 shows the conductivity mapping results of TiO2 
films at LRS and HRS by using HVAFM (high vacuum atomic force microscopy). Conductive 
10 
 
spots can be observed clearly in the electroforming process SET process. In the Cu/ZrO2:Cu./Pt 
memory cell, by conducting temperature-dependent measurements, Guan, et al. investigated Cu 
conductive bridge as the conductive filament path is formed in the SET process and the 
dissolving of the existing Cu conductive bridge is caused by the Joule heating assisted oxidation 
[36]. 
 
 
Figure 2.3: The conductivity mapping results of TiO2 at (a) LRS and (b) HRS [35]. 
 
 In interface?type resistive switching mechanism, the resistive switching is caused by the 
change of charge carriers or mobile ions which contribute to current transport. Various models 
have been proposed for physical mechanism involving an interface-type conductive path, 
including: (electrochemical) migration of oxygen vacancies [6, 37, 38], trapping and detrapping 
of charge carriers (electrons or holes) [39] and Schottky barrier model [7, 40]. 
 
 
 
11 
 
2.2.3 Flexible Resistive Switching Memory Devices 
 Flexible electronics systems such as wearable electronics and built-in elements have 
become an emerging field. Based on this trend, the demand for a flexible memory will also 
increase to support flexible electronics. Flexible ReRAM has been demonstrated to be a 
promising technology by several research groups. 
 For instance, Sungho et al., implemented Al/TiOx/Al ReRAM fabricated on transparent 
and flexible substrate of polyethersulfone (PES) [41]. The I-V correlation of this ReRAM 
follows the SCLC theory. Meanwhile, Yongsung et al., implemented a 8?8 crossbar array of 
organic ReRAM devices on polyethylene terephthalate (PET) flexible substrate [42]. This group 
studied HRS/LRS ratio as a function of the bending radius as well as the number of bendings. 
The authors of this work noted there were no significant difference in the HRS/LRS ratio 
regardless the bending radius and the cycles. Furthermore, Lu-Hao et al., fabricated graphene 
oxide resistive switching memories on PET with high RESET speed of 100 ns and a SET speed 
of approximately 100 ?s [43]. The above flexible ReRAM devices are shown in Fig 2.4. 
 
 
  
12 
 
 
Figure 2.4: Photograph of different flexible ReRAM devices. (a) The Al/TiOx/Al ReRAM [41]. 
(b) Ti/Au/Al/PI:PCBM/Al organic memory devices [42]. (c) Graphene oxide based ReRAMs 
[43]. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
13 
 
 
 
 
 
 
 
Chapter 3  
 
Fabrication of Cu/CuxO/Ag ReRAM Devices 
 
The Cu/CuxO/Ag ReRAM Device was fabricated on Dupont Kapton polyimide substrate. A 
silver thin film seed layer was printed on Kapton substrate. Ink-jet printing was performed using 
a FUJIFILM Dimatix Material Printer 2831 with a piezoelectric print head cartridge. The ink 
used in this work was U5603 from Sun Chemical (20 wt% suspension of silver nanoparticles 
dispersed in ethanol and ethanediol, particles size < 150 nm).  
After printing, the sample was thermally cured at 200 ?C for 2 h in a nitrogen environment 
to remove solvent and other impurities from the printed silver ink. Copper was electroplated on 
the silver lines at a current density of approximately 6 mA/cm2 for 5 min. We used a copper 
plating solution consisting of sulfuric acid (100 g/L), copper sulfate (60 g/L), Enthone SC MD (8 
ml/L), hydrochloric acid (2 ml/L), and Enthone SC Lo (2 ml/L). The copper coated sample was 
placed on a hot plate in air with heating temperature of 200 ?C for 3 h to create a thin oxide 
(CuxO) layer on the Cu. Note that we also fabricated a subset of control samples that did not 
undergo the Cu oxidation step. Next, a second nanoparticle silver film pattern was printed on top 
of the CuxO layer to create the cross structure of the device as well as Ag top electrodes (TE). 
This layer was thermally cured at 200 ?C for 2 h in a nitrogen environment. Last, a cotton swab 
with 1 % HCl solution was used to remove the copper oxide thin layer in a region away from the 
Ag overlay to allow better connection to the Cu bottom electrode (BE) [43]. Note that more 
information of the fabrication equipment and materials can be found in Appendix A. The 
14 
 
fabrication process flow is schematically shown in Fig 3.1. Fig 3.2 shows the photograph of the 
fabrication process flow. The SEM image of the cross-sectional view of the Cu/CuxO/Ag MOM 
device fabricated on a Kapton substrate is shown in Fig 3.3. 
The memory cell has a cell size of 20 ?m?20 ?m, which is the smallest deposit feature limit 
of the cartridge. In the cell-size dependent test, different cell-size memory devices have also been 
fabricated in the same process. The ink-jet printer we used is the FUJIFILM DMP (Dimatix 
Material Printer) 2831 with a piezoelectric print head cartridge. 
 
 
Figure 3.1: Schematic fabrication flow for the Cu/CuxO/Ag ReRAM devices. (a) Silver seed 
layer printed. (b) Electroplated copper. (c) CuxO layer created. (d) Silver top electrode printed. 
 
15 
 
 
 
 
                
                                               (a)                                           (b) 
        
                                                    (c)                                           (d) 
                                 
                                                   (e) 
Figure 3.2: Photographic process flow for Cu/CuxO/Ag ReRAM array. (a) Ink-jet printed silver 
seed layer. (b) Electroplated copper lines. (c) Copper oxidation. Create a thin CuxO layer on the 
copper. (d) Ink-jet printed silver top electrodes (TE). (e) Copper bottom electrodes (BE) created 
by removing copper oxide. 
 
 
16 
 
 
 
 
 
 
 
 
 
 
Figure 3.3: The SEM image of the cross section of the Cu/CuxO/Ag ReRAM device. 
 
 
 
 
 
17 
 
 
 
 
 
 
 
Chapter 4 
Device Characterization and Testing 
4.1  X-ray Photoelectron Spectroscopy 
   X-ray photoelectron spectroscopy is one of the most common tools for thin film 
characterization. To analyze the oxide material induced BRS behavior, we did X-ray 
photoelectron spectroscopy (XPS) on the original CuxO surface and the same sample after 
sputter etching for 2 min and 4 min.  
   Fig 4.1 shows the X-ray photoelectron spectroscopy (XPS) of the original oxide film 
before sputter etching, using a sputter rate of 25 ?/min. The calculated XPS surface element 
composition (at %) versus sputter time (depth) is shown in Table 4.1. For the initial CuxO film, 
high resolution scans over the Cu2p peak in the XPS spectra indicate a native Cu oxide state, 
which is a CuO film. This is corroborated by the approximately 50/50 at% Cu/O elemental ratio 
of the ?as received? surface. 
 However, as shown in Fig 4.2, over the sputtered depth of ~100?, there is a transition 
from native CuO to Cu2O/Cu, shown by the shape change in the Cu2p peak and accompanying 
Cu2p3/2 binding energy changes observed in the spectra. Additionally, the layer composition 
shows an increase in Cu with depth. This indicates that the copper oxide has transitioned from 
CuO to Cu2O/Cu over a film thickness of ~100?. We say ?Cu2O/Cu? due to the well-known 
limitation of XPS to distinguish Cu(0) = elemental and Cu(I) = Cu2O, since the Cu2p3/2 Cu(0) 
and Cu(I) peak shapes and binding energies are nearly identical [44]. Since the CuxO thickness, 
18 
 
indicated by Fig 3.3, is on the order of 0.1 micron = 1000?, we presume we have not yet reached 
the Cu substrate and that the Cu2p feature at a depth of ~100? is largely Cu2O. 
 
 
 
 
Figure 4.1: The XPS spectra of CuxO thin film before sputter etching. 
 
19 
 
 
Figure 4.2: The XPS spectra of CuxO thin film after 2 min or 4 min sputter etching. 
(Corresponding to removal of ~50/100? of the initial CuxO surface). 
 
 
 
 
 
 
Table 4.1: XPS surface elemental composition (at%) versus sputter etching time. 
 
4.2 Current-Voltage Characteristics of Cu/CuxO/Ag ReRAM Devices 
4.2.1 Electroforming Process 
Electroforming process is a pretreatment procedure for resistive switching in most binary-
transition-oxide ReRAM such as TiO2 [27], ZnO [1] and NiO [45]. These ReRAMs show 
Time (min) Cu (%) O (%) 
0 47.9 52.1 
2  63.7 36.3 
4 69.6 30.4 
20 
 
electrically insulating characteristics initially, with relatively high resistance. For different 
ReRAM devices, by applying high electrical stress (a voltage or a current) to one of the two 
electrodes (TE or BE), the resistance of the device can be decreased by several orders of 
magnitude (corresponding to a dramatic increase in current) suddenly. The specific 
voltage/current which causes this sudden change is called the electroforming voltage/current or 
threshold voltage/current. The electroforming procedure plays a significant role in resistive 
switching. In some literatures, electroforming is believed to activate resistive switching in the 
switching material layer and lead to the formation or growth of some conductive objects such as 
filaments [46]. 
 In this work, a Keithley 4200 semiconductor characterization system was used to measure 
the current-voltage (I-V) characteristics of the fabricated devices. The electroforming process is 
labeled as (1) in the Fig 4.3 [47]. By applying a voltage sweep from 0 V to +2 V as the arrow 
shows, a large increase (corresponding to a dramatic reduction in resistance) happens in the 
memory cell?s current at approximately +1.8 V (the electroforming voltage), indicating that the 
strong electrical stress leads to the formation of conducting filaments. 
 
 
 
 
 
 
 
 
21 
 
 
 
Figure 4.3: Electrical forming process and I-V characteristics of the memory. 
 
4.2.2      I-V Characteristics after Electroforming 
 Voltage sweeps were performed in lower range. In the Fig 4.3, the arrows indicate the 
voltage sweeping direction (-1 V? 0  V? +1  V? 0  V? -1 V), the I-V hysteresis curve is 
obtained. During the voltage scan from -1 V to -0.8 V (SET voltage), a sharp current decrease is 
observed before step (2), which is called the SET process, resulting in a significant change of 
resistance in the ReRAM memory cell from the low resistance state (LRS) to the high resistance 
state (HRS). Next, when the sweeping voltage increases to +0.8 V (RESET voltage), a rapid 
increase in current is observed at the end of step (3), which is called the RESET process. During 
the RESET process, the resistance of the ReRAM device is switched from HRS to LRS. The 
magnitudes of both the SET voltage and the RESET voltage are less than 1 V, which shows the 
low-voltage operation feature of the Cu/CuxO/Ag memory devices.  
22 
 
 Fig 4.4 shows the corresponding resistance-voltage (R-V) curve during the same voltage 
sweep. Two resistance states (HRS and LRS) are demonstrated. When the voltage reaches the 
VSET, the resistance state of the device is switched to HRS. Alternatively, LRS is achieved after 
the voltage bias reaches VRESET. The resistance at HRS and LRS is approximately 150 ? and 10 
? with a ratio of 15. 
 
 
 
Figure 4.4: R-V characteristics of the Cu/CuxO/Ag ReRAM. 
 
 The device shows excellent reproducible behavior, as shown in Fig 4.5. The 
measurement was conducted in a lab environment at room temperature. The arrows indicated the 
sweeping direction, which starts from -1 V to +1 V then back to -1 V. During all of the voltage 
sweeps, the I-V curves show reproducible and stable behavior, and the ratio of HRS/LRS is 
almost constant. 
23 
 
 
 
 
Figure 4.5: Multi-cycle I-V characteristics for the 1st, 250 th and 500 th cycles. 
 
4.2.3      Temperature-Dependent I-V Characteristics 
 Temperature dependence of BRS characteristics after electroforming is investigated, which 
is shown in Fig 4.6. The temperature-dependent measurement was conducted in a micro-
manipulated cryogenic probe station. The workable temperature range of the cryogenic probe 
station is from 11 K to 370 K. However, an excessive high or cold temperature can affect proper 
functioning of the ReRAM device. When the temperature is above 365 K or below 250 K, the 
ReRAM device does show BRS behavior. In Fig 4.6, the temperature range is from 255 K to 315 
K with a step size of 15 K. No strong temperature dependence was observed for the switching 
voltages VSET and VRESET, which are approximately -0.6 V and +0.6 V respectively. 
 
24 
 
 
Figure 4.6: Temperature dependence of I-V curves in semi-log scale. 
 
4.3      Device HRS and LRS Yields 
We have fabricated and electrically tested more than 220 devices. From 121 functional 
memory cells, the HRS yield and LRS yield are shown in the Fig 4.7. The average value of 
resistance at HRS and LRS is 95.3 ? and 13.7 ?, respectively. There is no overlap between HRS 
and LRS and the ratio of HRS and LRS is approximately 6.9. The variation of the resistance at 
HRS and LRS might be caused by: (a) the thickness of CuxO film was not perfectly uniform and 
(2) the defect in the fabrication processing. 
25 
 
 
Figure 4.7: Distribution of HRS and LRS of fabricated devices. 
 
4.4      Switching Endurance and Data Retention Performance  
 The switching endurance and data retention characteristics were studied to evaluate the 
reliability of the Cu/CuxO/Ag ReRAM devices. Fig 4.8 shows the data endurance at 295 K. For 
the switching cycle measurements, -0.8 V and +0.8 V were applied to the device as the SET 
voltages and RESET voltages, respectively, switching the device to HRS and LRS. Since once 
the device is switched to one state (i.e. HRS/LRS), the resistance of the device will not be 
changed until the voltage bias reaches to RESET voltage/SET voltage to switch the device to 
LRS/ HRS. The resistance of the device can be read at a low voltage range. In this study, the 
resistances of HRS and LRS were recorded at low voltage range from -0.01 V to +0.01 V. 
During 100 switching cycles, resistance value of HRS was approximately 120 ? and LRS was 
around 6 ? without switching failure. The resistance value of HRS and LRS remained stable 
with a ratio of 20. Fig 4.9 shows the retention characteristics of the device at room temperature. 
26 
 
After switching the device to one resistive state (HRS or LRS), no electrical power was needed 
to maintain the resistance in this state. As shown in Fig 4.9, no significant change of the 
resistance at both HRS and LRS remained for nearly 2 weeks, which further proved the 
nonvolatile nature of the Cu/CuxO/Ag ReRAM device. 
 
Figure 4.8: Switching endurance test of the device during 100 cycles. 
 
Figure 4.9: Data retention performance of the device. 
27 
 
4.5      Mechanical Bending Tests and Flex Tests 
  Flexibility is very important for future electronic applications such as wearable 
electronic devices. The mechanical robustness test and flex tests of the Cu/CuxO/Ag ReRAM 
devices were performed in order to examine the mechanical stability of the devices. Thus, 
excellent electrical characteristic under bending condition is crucial for practical flexible 
electronic devices. This required property of our Cu/CuxO/Ag memory cells was investigated in 
detail. 
 
4.5.1      Mechanical Bending Tests 
  In the mechanical bending tests, I-V measurements were performed under various degree 
of bending. We investigated the HRS/LRS ratios as a function of the bending radius of the 
Cu/CuxO/Ag memory cells. The device was bent to form a convex shape with different radius of 
55 mm, 35 mm, 26 mm, 17.5 mm, 11.5 mm and 5.5 mm and the device was measured while 
being flexed, which is shown in the Fig 4.10. The devices exhibit stable I-V hysteretic behavior 
under bending condition. With the increase of curvature, the resistances at HRS and LRS 
remained stable with a ratio of greater than 7, which is shown in Fig 4.11. In particular, we 
investigated the distribution of SET voltages and RESET voltages according to the bending 
radius, which is shown in Fig 4.12. The VSET was at a range from -0.58V to -0.80 V and the 
VRESET was from +0.67 V to +0.79 V. There was no overlap between SET voltages and RESET 
voltages observed during the bending process, indicating the device can be controlled precisely 
and operated well at low voltage, even under significant flexing conditions. 
 
28 
 
 
 
Figure 4.10: Photograph of the device bent at radius = 11.5 mm. 
 
 
Figure 4.11: Resistance of HRS and LRS as a function of the bending radius. 
 
 
29 
 
 
Figure 4.12: VSET and VRESET as a function of the bending radius. 
 
4.5.2      Flex Tests 
  The statistical data of mechanical robustness test for the Cu/CuxO/Ag ReRAM memory 
cell is shown in Fig 4.13. The device was physically flexed for over 1000 cycles. For each flex 
cycle, the 25-mm-long device was bent into a convex shape and then into a concave shape with a 
radius of 8mm (convex shape shown in Fig 4.14). During 1000 flex cycles, the ratio of HRS and 
LRS was approximately 20 and remained relatively stable. 
 
 
30 
 
 
 
Figure 4.13: Resistance of HRS and LRS during 1~1000 flexes. 
 
 
 
Figure 4.14: The Cu/CuxO/Ag ReRAM device was bent into a convex shape. 
 
31 
 
4.6      Switching Time Tests 
 Fig 4.15 illustrates the switching time (or writing time) of the SET/RESET operation 
measured at room temperature. In Fig 4.15(a), a SET voltage of -0.8 V was applied to the top 
electrode of the memory device. The current that flowed through the cell was recorded as the 
time changed. It can be observed that a drastic reduction of current happens at a time of 
approximately 2 s, which means that it takes 2 s to switch the resistance from LRS to HRS. 
Correspondingly, it takes roughly 1 s to switch from HRS to LRS by applying a RESET voltage 
of +0.8 V and the result is shown in Fig 4.15(b). 
 
 
 
 
 
 
 
  
32 
 
 
 
 
Figure 4.15: Switching time measurements of (a) RET and (b) RESET operation. 
 
 
 
33 
 
To explore the temperature effect and the SET/RESET voltage effect on the switching 
speed, we conducted the same experiments under different temperatures (from 255 K to 355 K), 
and applied different SET/RESET voltage biases (from ?0.6 V to ?1 V) to switch to the memory 
cell. However, there was no stable trend for the temperature and voltage bias effects observed. 
Therefore, no corresponding results are shown in this thesis. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
34 
 
 
 
 
 
 
 
Chapter 5 
Resistive Switching Mechanism 
5.1    Non-Cu-Oxidation Control Group 
          Fig 5.1 shows the I-V curve of a Cu/CuxO/Ag ReRAM device. Compared to these fully 
fabricated devices, we fabricated a subset of devices without the Cu oxidation procedure and 
tested the I-V characteristics, which is shown in Fig 5.2. We note that the resistance of devices 
without Cu oxidation does not show BRS. Furthermore, -0.8 V and +0.8 V were applied to the 
device as the SET voltages and RESET voltages, switching both the fully fabricated device and 
the non-oxidation device to HRS and LRS, respectively. The resistances of HRS and LRS were 
recorded at a low voltage range from -0.01 V to +0.01 V. The results for the fully fabricated 
device and the no-oxidation device are shown in Fig 5.3(a) and Fig 5.3(b). We noted that the 
fully fabricated Cu/CuxO/Ag ReRAM device shows HRS and LRS with a ratio of 20. However, 
during the 50 switching cycles, no resistance difference can be confirmed when writing the no-
oxidation device to the SET voltage and the RESET voltage.  
In conclusion, the Cu oxidation process is necessary for realization of our devices. This 
also provides some insight into the physical mechanism occurring in these devices that leads to 
BRS. 
 
 
 
35 
 
 
 
Figure 5.1: Hysteretic I-V behavior of the Cu/CuxO/Ag ReRAM device. 
 
 
  
Figure 5.2: I-V curve of a non-oxidation device. 
 
36 
 
 
 
 
 
 
Figure 5.3: Resistance dependence of HRS and LRS on number of switching cycles. (a) A fully 
fabricated Cu/CuxO/Ag ReRAM device. (b) A Cu/Ag device without Cu oxidation step. 
 
37 
 
5.2     Physical Mechanism of Cu/CuxO/Ag ReRAM Devices 
  Temperature-dependent resistance measurements on Cu/CuxO/Ag ReRAM devices were 
performed to explore the physics of the resistive switching conduction mechanism.  
The resistance of devices at initial resistance state (IRS, resistance state before 
electroforming), HRS and LRS as a function of temperature is shown in Fig 5.4. The temperature 
range is from 255 K to 355 K with a step size of 15 K. Reproducible BRS can be observed in this 
temperature range. The resistance of all states is read by sweeping the voltage bias in a small 
range from -0.01 V to +0.01 V to prevent the device from switching to a different resistive state. 
The IRS and HRS can be considered to exhibit semiconductor behavior because when the 
temperature increases, the resistance of the device decreases. On the contrary, a weak metallic 
behavior can be observed at LRS. There is potential possibility that the filamentary conductive 
paths were generated in the RESET process, which results in very low resistance. However, the 
rupture of conductive filament paths happened at SET process, thus the temperature-dependent 
HRS is dominated by the semiconductor behavior of the CuxO. Furthermore, the ratio of 
HRS/LRS drops from approximately 82 to 13 over the temperature range from 255 K to 355 K. 
 
 
 
 
 
 
 
 
38 
 
 
 
 
 
 
Figure 5.4: Temperature dependence of HRS, IRS and LRS. 
 
 
 
 
 
 
 
 
39 
 
To exclude the effect of the SET and RESET operations on the original physical state, the 
temperature-dependent I-V measurement was conducted after the device was switched to HRS 
by applying a SET voltage (-0.8 V). Fig 5.5 shows the I-V curves in logarithm scale measured at 
various temperatures from 255 K to 315 K, which is in a positive voltage region (0 V to +0.5 V) 
low enough to prevent the RESET process. It was shown that with the increase of temperature, 
the current increases gradually. The ln(I) versus 1/KT for HRS was plotted in Fig 5.6. The plot 
shows five generally straight lines at different voltages. Then, the activation energy (Ea) at each 
voltage bias was calculated from the slopes of fitting lines. Fig 5.7 shows the calculated Ea as a 
function of applied voltages.  
 
 
 
Figure 5.5: Temperature-dependent I-V characteristics at HRS in log-log scale. 
 
 
 
40 
 
 
Figure 5.6: The LnI versus 1/KT curve for HRS. 
 
 
Figure 5.7: Activation energy at various voltages. 
 
 
 
41 
 
To further clarify the resistive switching mechanism of Cu/CuxO/Ag memory devices, the 
double logarithmic plot of I-V curves for the positive region was replotted, which is shown in 
Fig 5.8. As indicated in Fig 5.8, in the positive voltage bias region, starting from 0 V, the I-V 
behavior of Cu/CuxO/Ag memory cell in HRS is very close to linear and the linear dependence 
for applied voltage is smaller than +0.60 V. In this low-voltage region, the slope of I-V curve is 
very close to 1, in agreement with Ohm?s law. Next, a dramatic increase of current is observed at 
approximately +0.8 V (VRESET) due to the formation of conductive filaments through the device 
[4, 16], which does not follows space charge limited current (SCLC) theory [48]. Finally, the 
Ohmic conduction behavior was observed again in the decreasing voltage bias scan in the LRS. 
Different cell-size ReRAM devices were fabricated. The cell sizes were from 20?20 ?m2 
to 800?800 ?m2, which is shown in Fig 5.9. Fig 5.10 shows the cell-size dependence of the HRS 
and LRS of Cu/CuxO/Ag memory cells. Different cell-size Cu/CuxO/Ag memory devices were 
fabricated in the same condition. All devices show BRS during DC voltage scan. The best linear 
fitting lines are also shown in the figure. With the increase of cell size, the ratio of HRS/LRS 
deceases roughly from 30 to 10. The resistance of HRS decreases from approximately 264 ? to 
46 ?, whereas the resistance of LRS only shows slight dependence of cell sizes, decreasing from 
8.32 ? to 4.46 ?. The result reveals that the resistive switching behavior is a local phenomenon 
in the Cu/CuxO/Ag memory cell, dominated by the local conductive filament path, further 
indicating the resistive switching mechanism of the Cu/CuxO/Ag MOM cross-point structure is 
attributed to the formation and rupture of the conductive filament paths [16, 49, 50]. 
 
 
 
42 
 
 
 
Figure 5.8: Logarithmic plot of I-V characteristic of positive voltage region. 
 
 
 
 
 
Figure 5.9: Photograph of the fabricated different cell-size ReRAM array. 
 
43 
 
 
Figure 5.10: Resistance of HRS and LRS versus cell size of Cu/CuxO/Ag ReRAM devices. 
 
5.3     Resistive Switching Mechanism Model 
     Based on the temperature-dependent measurement and cell-size dependent measurement 
analysis, we are able to conceive an intuitive diagram of the BRS in the device. Fig 5.11 shows 
the schematic diagram of the resistive switching mechanism model of the Cu/CuxO/Ag ReRAM 
devices. Note that the geometry and distribution of this model are ideally simplified and the 
actual conduction is considered to be more complicated. Fig 5.11(a) shows the fresh device at 
IRS. After the electroforming process (Fig 5.11(b)), a strong electrical field leading to the 
formation of a conductive filament is taking in the insulating matrix between the Ag top 
electrode and the Cu bottom electrode. Next, at HRS, the filaments are ruptured under a large 
negative voltage. Consider the SET process, a huge SET current will generate high Joule heating 
at local filament paths, leading to the rupture of conductive paths, which is shown in Fig 5.11(c). 
In this case, the carriers can only transport through the semiconductor material CuxO thin film, 
exhibiting a semiconductor behavior. On the other hand, the conductive filament paths are rebuilt 
44 
 
under a certain positive voltage at LRS (Fig 5.11(d)). In this case, the current flowing through 
the device is mainly confined in conductive filament paths. Thus a metallic behavior is shown at 
LRS. Note that the SET process and the RESET process are reversible processes, leading to the 
switching of HRS and LRS in a working device. 
 
 
 
Figure 5.11: Schematic diagram of the resistive switching mechanism model. (a) IRS. (b) ERS. 
(c) HRS. (d) LRS. 
 
 
 
 
 
 
45 
 
 
 
 
 
 
 
Chapter 6 
 
Conclusion and Future Work 
  In this work, we discussed the design and fabrication of the Cu/CuxO/Ag ReRAM device. 
We described the process for fabricating ReRAM devices. We used ink-jet printing technology 
during the fabrication process which has great potential for roll-to-roll processing in the 
industrial production. The temperature of entire fabrication process was under 200 ?C. The 
memory devices were fabricated based on Cu and Ag, which are abundant and easy to process. 
The electroformed devices exhibited reproducible and reliable BRS characteristics under a 
temperature range from 255 K to 355 K. The fabricated device showed good switching 
endurance (more than 100 cycles), excellent data retention performance (nearly 2 weeks). In 
addition, as a flexible device, the device had great mechanical robustness properties.  
  The temperature-dependent and cell-size dependent measurement results reveal that the 
physical mechanism of the memory device is attributed to the formation and rupture of the local 
conductive filament paths in the insulating matrix between two electrodes. The device shows 
semiconductor-like behavior at HRS and IRS, and exhibits metallic behavior at LRS. Moreover, 
a physical mechanism model of the device has been presented. The observed trend of increased 
ratio of HRS to LRS as the cell area decreases can be regarded as a benefit of device scaling. The 
simple, flexible inexpensive low-temperature-fabricated low-voltage operation ink-jet printed 
Cu/CuxO/Ag memory cell is expected to provide opportunities for flexible memory electronics 
and neuromorphic electronic devices. 
46 
 
  In terms of future work, we are studying various methods to optimize the fabrication 
process to increase the ratio of HRS/LRS and decrease the writing time. Possible solutions 
include: oxidizing high-quality CuxO and decreasing the memory cell size. Our current approach 
using an oxidation temperature of 200 ?C gets a layer of CuxO with a thinkness of approximately 
23 ?. In addition, since the smallest deposit feature limit of the cartridge is 20 ?m?20 ?m, a 
smaller cell size needs a bran-new design utilizing photolithography technology to be achieved. 
Moreover, we are in the process of measuring the low frequency noise (flicker noise or 1/f 
noise), which will be a useful tool and offer new ideas in understanding the operation mechanism 
of the ReRAM devices.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
47 
 
 
 
 
 
 
 
References 
 
 
[1] Lee, Seunghyup, Heejin Kim, Jinjoo Park, and Kijung Yong, "Coexistence of unipolar 
and bipolar resistive switching characteristics in ZnO thin films." Journal of Applied 
Physics, vol. 108, no. 7, pp. 076101, 2010.  
[2] Stewart, D. R., D. A. A. Ohlberg, P. A. Beck, Y. Chen, R. Stanley Williams, Jan Oskar 
Jeppesen, K. A. Nielsen, and J. Fraser Stoddart, "Molecule-independent electrical 
switching in Pt/organic monolayer/Ti devices." Nano Letters, vol. 4, no. 1, pp. 133-136, 
2004. 
[3] Gergel-Hackett, Nadine, Behrang Hamadani, Barbara Dunlap, John Suehle, Curt Richter, 
Christina Hacker, and David Gundlach, "A flexible solution-processed 
memristor." Electron Device Letters, IEEE, vol. 30, no. 7, pp. 706-708, 2009. 
[4] Han, Jin-Woo, and M. Meyyappan, "Copper oxide resistive switching memory for e-
textile." AIP Advances, vol. 1, no. 3, pp. 032162, 2011. 
[5] Guan, Weihua, Ming Liu, Shibing Long, Qi Liu, and Wei Wang, "On the resistive 
switching mechanisms of Cu/ZrO2: Cu/Pt." Applied Physics Letters, vol. 93, no. 22, pp. 
223506, 2008. 
[6] Liu, Qi, Weihua Guan, Shibing Long, Rui Jia, Ming Liu, and Junning Chen, "Resistive 
switching memory effect of ZrO2 films with Zr+ implanted." Applied physics 
letters, vol. 92, no. 1, pp. 012117-012117, 2008. 
48 
 
[7] Fujii, T., M. Kawasaki, A. Sawa, H. Akoh, Y. Kawazoe, and Y. Tokura, "Hysteretic 
current?voltage characteristics and resistance switching at an epitaxial oxide Schottky 
junction SrRuO3/SrTi0.99Nb0.01O3." Applied Physics Letters, vol. 86, no. 1, pp.  012107-
012107, 2005. 
[8] Williams, R, "How we found the missing memristor." Spectrum, IEEE, vol. 45, no. 12, pp. 
28-35, 2008. 
[9] Mikolajick, T., C. Dehm, W. Hartner, I. Kasko, M. J. Kastner, N. Nagel, M. Moert, and C. 
Mazure, "FeRAM technology for high density applications."Microelectronics 
Reliability, vol. 41, no. 7, pp. 947-950, 2001. 
[10] Tehrani, S., E. Chen, M. Durlam, T. Zhu, and H. Goronkin, "High density nonvolatile 
magnetoresistive RAM." In Proceeding International Electron Devices Meeting, pp. 8-11. 
1996. 
[11] Wong, H-SP, Simone Raoux, SangBum Kim, Jiale Liang, John P. Reifenberg, Bipin 
Rajendran, Mehdi Asheghi, and Kenneth E. Goodson, "Phase change 
memory." Proceedings of the IEEE, vol. 98, no. 12, pp. 2201-2227, 2010. 
[12] Hickmott, T. W, "Low?Frequency Negative Resistance in Thin Anodic Oxide 
Films." Journal of Applied Physics, vol. 33, no. 9, pp. 2669-2682, 2004. 
[13] Oka, Keisuke, Takeshi Yanagida, Kazuki Nagashima, Hidekazu Tanaka, and Tomoji 
Kawai, "Nonvolatile bipolar resistive memory switching in single crystalline NiO 
heterostructured nanowires." Journal of the American Chemical Society, vol. 131, no. 10, 
pp. 3434-3435, 2009. 
49 
 
[14] Panda, D., A. Dhar, and S. K. Ray, "Nonvolatile and unipolar resistive switching 
characteristics of pulsed laser ablated NiO films." Journal of Applied Physics, vol.108, no. 
10, pp. 104513, 2010. 
[15] Chen, Y. S., H. Y. Lee, P. S. Chen, P. Y. Gu, C. W. Chen, W. P. Lin, W. H. Liu et al, 
"Highly scalable hafnium oxide memory with improvements of resistive distribution and 
read disturb immunity." In Electron Devices Meeting (IEDM), 2009 IEEE International, 
pp. 1-4. IEEE, 2009. 
[16] Wang, Sheng-Yu, Chin-Wen Huang, Dai-Ying Lee, Tseung-Yuen Tseng, and Ting-Chang 
Chang, "Multilevel resistive switching in Ti/CuxO/Pt memory devices." Journal of 
Applied Physics, vol. 108, no. 11, pp. 114110, 2010. 
[17] Yan, X. B., Y. D. Xia, H. N. Xu, X. Gao, H. T. Li, R. Li, J. Yin, and Z. G. Liu, "Effects of 
the electroforming polarity on bipolar resistive switching characteristics of SrTiO3? ? 
films." Applied Physics Letters, vol. 97, no. 11, pp. 112101, 2010. 
[18] Lee, Chang Bum, Dong Soo Lee, Anass Benayad, Seung Ryul Lee, Man Chang, 
Myoung-Jae Lee, Jihyun Hur, Chang Jung Kim, and U-In Chung, "Highly uniform 
switching of tantalum embedded amorphous oxide using self-compliance bipolar resistive 
switching." Electron Device Letters, IEEE, vol. 32, no. 3, pp. 399-401, 2011. 
[19] Fujimoto, Masayuki, Hiroshi Koyama, Yuji Nishi, and Toshimasa Suzuki, "Resistive 
switching properties of high crystallinity and low-resistance Pr0.7Ca0.3MnO3 thin film 
with point-contacted Ag electrodes." Applied Physics Letters, vol. 91, no. 22, pp. 223504-
223504, 2007. 
[20] Kim, Hee-Dong, Ho-Myoung An, Kyoung Chan Kim, Yujeong Seo, Ki-Hyun Nam, 
Hong-Bay Chung, Eui Bok Lee, and Tae Geun Kim, "Large resistive-switching 
50 
 
phenomena observed in Ag/Si3N4/Al memory cells." Semiconductor Science and 
Technology, vol. 25, no. 6, pp. 065002, 2010. 
[21] Chen, Lin, Qing-Qing Sun, Jing-Jing Gu, Yan Xu, Shi-Jin Ding, and David Wei Zhang, 
"Bipolar resistive switching characteristics of atomic layer deposited Nb2O5 thin films for 
nonvolatile memory application." Current Applied Physics, vol. 11, no. 3, pp. 849-852, 
2011. 
[22] Younis, Adnan, Dewei Chu, Ionsecu Mihail, and Sean Li, "Interface-Engineered Resistive 
Switching: CeO2 Nanocubes as High-Performance Memory Cells."ACS applied materials 
& interfaces, vol. 5, no. 19, pp. 9429-9434, 2013. 
[23] Nagashima, Kazuki, Takeshi Yanagida, Keisuke Oka, Masateru Taniguchi, Tomoji Kawai, 
Jin-Soo Kim, and Bae Ho Park, "Resistive switching multistate nonvolatile memory 
effects in a single cobalt oxide nanowire." Nano letters, vol. 10, no. 4, pp. 1359-1363, 
2010. 
[24] Nagashima, Kazuki, Takeshi Yanagida, Keisuke Oka, and Tomoji Kawai, "Unipolar 
resistive switching characteristics of room temperature grown SnO2 thin films." Applied 
Physics Letters, vol. 94, no. 24, pp. 242902, 2009. 
[25] Yoshida, Chikako, Masaki Kurasawa, Young Min Lee, Masaki Aoki, and Yoshihiro 
Sugiyama, "Unipolar resistive switching in CoFeB?MgO?CoFeB magnetic tunnel 
junction." Applied Physics Letters, vol. 92, pp. 113508, 2008. 
[26] Gao, Xu, Yidong Xia, Bo Xu, Jizhou Kong, Hongxuan Guo, Kui Li, Haitao Li et al, 
"Unipolar resistive switching behaviors in amorphous lutetium oxide films."Journal of 
Applied Physics, vol. 108, no. 7, pp. 074506, 2010. 
51 
 
[27] Jeong, Doo Seok, Herbert Schroeder, and Rainer Waser, "Coexistence of Bipolar and 
Unipolar Resistive Switching Behaviors in a Pt? TiO2? Pt Stack."Electrochemical and 
solid-state letters, vol. 10, no. 8, G51-G53, 2007. 
[28] Zhang, Jian, Hui Yang, Qi-long Zhang, Shurong Dong, and J. K. Luo, "Bipolar resistive 
switching characteristics of low temperature grown ZnO thin films by plasma-enhanced 
atomic layer deposition." Applied Physics Letters, vol. 102, no. 1, pp. 012113, 2013. 
[29] Shen, Wan, Regina Dittmann, and Rainer Waser, "Reversible alternation between bipolar 
and unipolar resistive switching in polycrystalline barium strontium titanate thin 
films." Journal of Applied Physics, vol. 107, no. 9, pp. 094506, 2010. 
[30] Lin, Chih-Yang, Chung-Yi Wu, Chen-Yu Wu, Tseung-Yuen Tseng, and Chenming Hu, 
"Modified resistive switching behavior of ZrO2 memory films based on the interface 
layer formed by using Ti top electrode." Journal of Applied Physics, vol. 102, no. 9, pp. 
094101, 2007. 
[31] Yan, Z. B., S. Z. Li, K. F. Wang, and J-M. Liu, "Unipolar resistive switching effect in 
YMn1-?O3 thin films." Applied Physics Letters, vol. 96, no. 1 (2010): pp. 012103-012103. 
[32] Sawa, Akihito, "Resistive switching in transition metal oxides." Materials today11, no. 6, 
pp. 28-36, 2008. 
[33] Simmons, J. G., and R. R. Verderber, "New conduction and reversible memory 
phenomena in thin insulating films." Proceedings of the Royal Society of London. Series 
A. Mathematical and Physical Sciences, vol. 301, no. 1464, pp. 77-102, 1967. 
[34] Lee, Myoung-Jae, Seungwu Han, Sang Ho Jeon, Bae Ho Park, Bo Soo Kang, Seung-Eon 
Ahn, Ki Hwan Kim et al, "Electrical manipulation of nanofilaments in transition-metal 
oxides for resistance-based memory." Nano letters, vol. 9, no. 4, pp. 1476-1481, 2009. 
52 
 
[35] Choi, B. J., D. S. Jeong, S. K. Kim, C. Rohde, S. Choi, J. H. Oh, H. J. Kim et al, 
"Resistive switching mechanism of TiO2 thin films grown by atomic-layer 
deposition." Journal of Applied Physics, vol. 98, no. 3, pp. 033715, 2005. 
[36] Guan, Weihua, Ming Liu, Shibing Long, Qi Liu, and Wei Wang, "On the resistive 
switching mechanisms of Cu/ZrO2:Cu/Pt." Applied Physics Letters, vol. 93, no. 22, pp. 
223506, 2008. 
[37] Baikalov, A., Y. Q. Wang, B. Shen, B. Lorenz, S. Tsui, Y. Y. Sun, Y. Y. Xue, and CWc 
Chu, "Field-driven hysteretic and reversible resistive switch at the Ag?Pr0.7Ca0.3MnO3 
interface." Applied Physics Letters, vol. 83, no. 5, pp. 957-959, 2003. 
[38] Sawa, A., T. Fujii, M. Kawasaki, and Y. Tokura, "Interface resistance switching at a few 
nanometer thick perovskite manganite active layers." Applied physics letters, vol. 88, no. 
23, pp. 232112-232112, 2006. 
[39] Hickmott, T. W, "Impurity conduction and negative resistance in thin oxide 
films." Journal of Applied Physics, vol. 35, no. 7, pp. 2118-2122, 2004. 
[40] Cowley, A. M., and S. M. Sze, "Surface States and Barrier Height of 
Metal?Semiconductor Systems." Journal of Applied Physics, vol. 36, no. 10, pp. 3212-
3220, 2004. 
[41] Kim, Sungho, Oktay Yarimaga, Sung-Jin Choi, and Yang-Kyu Choi, "Highly durable and 
flexible memory based on resistance switching." Solid-State Electronics, vol. 54, no. 4, 
pp. 392-396, 2010. 
[42] Ji, Yongsung, Byungjin Cho, Sunghoon Song, Tae Wook Kim, Minhyeok Choe, Yung Ho 
Kahng, and Takhee Lee, "Stable switching characteristics of organic nonvolatile memory 
on a bent flexible substrate." Advanced Materials, vol. 22, no. 28, pp. 3071-3075, 2010. 
53 
 
[43] Zou, S., P. Xu, and M. C. Hamilton, "Resistive switching characteristics in printed 
Cu/CuO/(AgO)/Ag memristors." Electronics Letters, vol. 49, no. 13, pp. 829-830, 2013. 
[44] Biesinger, Mark C., Leo WM Lau, Andrea R. Gerson, and Roger St C. Smart, "Resolving 
surface chemical states in XPS analysis of first row transition metals, oxides and 
hydroxides: Sc, Ti, V, Cu and Zn." Applied Surface Science, vol. 257, no. 3, pp. 887-898, 
2010. 
[45] Gibbons, J. F., and W. E. Beadle, "Switching properties of thin NiO films."Solid-State 
Electronics, vol. 7, no. 11, pp. 785-790, 1964. 
[46] Jeong, Doo Seok, Reji Thomas, R. S. Katiyar, J. F. Scott, H. Kohlstedt, A. Petraru, and 
Cheol Seong Hwang, "Emerging memories: resistive switching mechanisms and current 
status." Reports on Progress in Physics, vol. 75, no. 7, pp. 076502, 2012. 
[47] Zou, S. and M. C. Hamilton, ?Flexible non-volatile Cu/CuxO/Ag ReRAM memory 
devices fabricated using ink-jet printing technology,? presented at IEEE ECTC, 2014, 
Orlando, FL, May 27-30, 2014. 
[48] Dong, R., D. S. Lee, W. F. Xiang, S. J. Oh, D-J. Seong, S. H. Heo, H. J. Choi et al, 
"Reproducible hysteresis and resistive switching in metal-CuxO-metal 
heterostructures." Applied physics letters, vol. 90, no. 4, pp. 042107-042107, 2007. 
[49] Kim, Sungho, Hanul Moon, Dipti Gupta, Seunghyup Yoo, and Yang-Kyu Choi, 
"Resistive switching characteristics of sol?gel zinc oxide films for flexible memory 
applications." Electron Devices, IEEE Transactions on, vol. 56, no. 4, pp. 696-699, 2009. 
[50] Zou, S. and M. C. Hamilton, ?Ink-jet Printed Cu/CuxO/Ag ReRAM Memory Devices 
Fabricated on Flexible Substrates? presented at the MRS Spring Meeting 2014, San 
Francisco, CA, April 21-25, 2014. 
54 
 
 
 
 
 
 
 
Appendix A 
Processing Equipment and Materials 
 
 
Figure A.1: FUJIFILM Dimatix Material Printer. 
 
 
Figure A.2:  Kapton polymide substrate. 
55 
 
 
 
Figure A.3:  Silver nanoparticle ink. 
 
 
 
 
 
 
 
Figure A.4:  Hot plate. 
56 
 
 
Figure A.5:  YES (Yield Engineering Systems) vacuum cure oven. 
 
 Figure A.6:  Copper electroplating equipment. 
57 
 
 
 
Figure A.7:  Keithley 4200 (Semiconductor Characterization Systems) parameter analyzer. 
 
 
 
 
 
Figure A.8:  Micro-manipulated cryogenic probe system.