Role of oxadiazole moiety in different D-A polyazothines and related resistive switching properties

Article abstract of DOI:10.1039/c3tc30826j

Two donor-acceptor (D-A) polyazothines (PAs), incorporating the oxadiazole entity either acting as an electron acceptor (A) to form D-A structured PA-1 with the triphenylamine donor (D), or acting as a donor to form D-A structured PA-2 with the 3,3′-dinitro-diphenylsulfone acceptor, have been successfully synthesized via a polycondensation reaction. The variation in the role of the oxadiazole moiety in the D-A polymers, together with the use of different top electrode metals, leads to interesting electronic transport properties and various resistive switching behaviors of the present polyazothines. Pt-electrode devices based on a PA-1 active layer show a rewritable memory effect with poor endurance (less than 20 cycles), whereas the PA-2 based Pt devices exhibit write-once read-many-times (WORM) memory behavior. For the Al-electrode devices, both PAs demonstrate a much improved resistive switching effect, and the endurance of the PA-2 devices is better than that of the PA-1 devices. The difference in the electronic transport and memory properties of the four devices may originate from the different charge injection/extraction and electron transfer processes of the sandwich systems, and will provide guidelines for selecting both the proper D and A moieties in D-A polymers and electrode metals for high-performance resistance random access memories (RRAMs). The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c3tc30826j

Journal of
Materials Chemistry C
View Article Online
View Journal | View Issue
PAPER
Role of oxadiazole moiety in different D–A
polyazothines and related resistive switching
Cite this: J. Mater. Chem. C, 2013, 1,
4556
properties†
Liang Pan,‡abc Benlin Hu,‡ab Xiaojian Zhu,ab Xinxin Chen,ab Jie Shang,ab
c
ab
ab
Hongwei Tan,ab Wuhong Xue,ab Yuejin Zhu, Gang Liu
*
*
and Run-Wei Li
Two donor–acceptor (D–A) polyazothines (PAs), incorporating the oxadiazole entity either acting as an
electron acceptor (A) to form D–A structured PA-1 with the triphenylamine donor (D), or acting as a
donor to form D–A structured PA-2 with the 3,30-dinitro-diphenylsulfone acceptor, have been
successfully synthesized via a polycondensation reaction. The variation in the role of the oxadiazole
moiety in the D–A polymers, together with the use of different top electrode metals, leads to interesting
electronic transport properties and various resistive switching behaviors of the present polyazothines.
Pt-electrode devices based on a PA-1 active layer show a rewritable memory effect with poor endurance
(less than 20 cycles), whereas the PA-2 based Pt devices exhibit write-once read-many-times (WORM)
memory behavior. For the Al-electrode devices, both PAs demonstrate a much improved resistive
switching effect, and the endurance of the PA-2 devices is better than that of the PA-1 devices. The
difference in the electronic transport and memory properties of the four devices may originate from the
different charge injection/extraction and electron transfer processes of the sandwich systems, and will
provide guidelines for selecting both the proper D and A moieties in D–A polymers and electrode
metals for high-performance resistance random access memories (RRAMs).
Received 2nd May 2013
Accepted 27th May 2013
DOI: 10.1039/c3tc30826j
www.rsc.org/MaterialsC
Since 1992, when the donor–acceptor (D–A) approach was
introduced for the rst time to the design of conjugated poly-
Introduction
mers,30 D–A polymers have attracted much attention for prom-
ising electronic applications in light-emitting diodes, eld
effect transistors, solar cells and memory devices.5,31,32 In
particular, electronic memories based on D–A polymers with a
large on/off ratio, good endurance, long retention time, low
power consumption and fast switching speed have been inves-
tigated extensively over the past few years.33–36 Nevertheless,
there are still challenges in D–A polymer memories, where the
lack of a comprehensive and clear understanding of the elec-
tronic origin that causes the resistance transition hinders the
development of high-performance materials for practical
memory applications.37–39
The resistive switching effect, which refers to the electrical
phenomenon that the resistance of a metal/insulator/metal
sandwich structure can be reversibly regulated by an external
DC or pulse voltage stress,1–6 has demonstrated potential
applications in resistance random access memories (RRAM)7–20
and articial neural synapses21–23 to mimic the functionalities of
data storage and logic calculus devices.24–28 In comparison to
their inorganic counterparts, polymeric materials are equally
attractive for resistive switching applications due to their
alternative advantages of rich chemical diversity, easy solution
processability, high mechanical exibility, low fabrication cost,
and three-dimensional stacking capability.29
Poly(Schiff base) materials are a family of conjugated poly-
mers containing imine groups (C]N) in the backbone and
exhibiting excellent thermal stability, good mechanical prop-
erties, metal-chelating ability and molecular doping-controlled
electrical properties.20,40–42 On the other hand, oxadiazole-based
polymers are well known as electron-transporting materials in
photoelectric devices, due to their high electron affinity, as well
as thermal, hydrolytic, and optical stability.43,44 Recently, poly-
mers containing either Schiff base or oxadiazole segments have
been explored for memory applications, and it is reported that
the oxadiazole moieties may turn from an electron acceptor into
aKey Laboratory of Magnetic Materials and Devices, Ningbo Institute of Material
Technology and Engineering, Chinese Academy of Sciences, Ningbo 315201, China.
E-mail: liug@nimte.ac.cn; runweili@nimte.ac.cn; Fax: +86-574-8668-5163; Tel: +86-
574-86685030
bZhejiang Province Key Laboratory of Magnetic Materials and Application Technology,
Ningbo Institute of Material Technology and Engineering, Chinese Academy of
Sciences, Ningbo 315201, China
cDepartment of Physics, Ningbo University, Ningbo 315211, China
† Electronic supplementary information (ESI) available: See DOI:
10.1039/c3tc30826j
‡ Liang Pan and Benlin Hu contributed equally to this work.
4556 | J. Mater. Chem. C, 2013, 1, 4556–4564
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Journal of Materials Chemistry C
an electron donor when bonded to a functional group with was recrystallized from tetrahydrofuran–water (2 : 1, v/v), and a
stronger electron withdrawing ability.7,45–48
colorless needle crystal with a yield of 81.2% (8.91 g) was
In this work, polyazothines based on triphenylamine-oxadi- obtained. FT-IR (cmꢁ1, KBr): 2762 (aldehyde C–H stretching),
azole (TPA-OXD) and oxadiazole-3,30-dinitro-diphenylsulfone 1698 (aldehyde C]O stretching), 1562 and 1321 (–NO2
(OXD-NPS) D–A structures have been synthesized via poly- stretching), 1338 and 1150 (–SO2– stretching). Elemental
condensation reactions. With the different roles of the oxadi- microanalysis, anal. calcd for C26H16N2O8S3: C, 53.78; H, 2.78;
azole moiety in the D–A polymers, PA-1 and PA-2 exhibit N, 4.82; S, 16.57. Found: C, 53.52; H, 2.83; N, 4.62; S, 16.38%. 1H
rewritable and write-once read-many-times (WORM) resistive NMR (400 MHz, DMSO-d6, d, ppm): 9.98 (2H, s), 8.77 (2H, d, J ¼
switching behaviors in the Pt/PA/Pt sandwich structures, 2.4 Hz), 8.31 (2H, dd, J ¼ 8.8 Hz, 2.4 Hz), 7.99 (2H, d, J ¼ 8.8 Hz),
respectively. When Al is used as the top electrode instead, both 7.43–7.36 (6H, m).
polymers exhibit rewritable switching behaviors due to the
Synthesis of PA-1. 4,40-diformyl-triphenylamine (0.30 g, 1.0
enhanced carrier extraction under reversed electric elds. These mmol) and M1 (0.47 g, 1.0 mmol) were dissolved in dry N-
results could provide new insight into the design strategy of methyl-2-pyrrolidone (NMP) (6 mL) with p-toluenesulfonic acid
high-performance D–A polymers and their resistance random (0.040 g) as the catalyst. Aer stirring at 120 ꢀC for 20 h, the
access memory (RRAM) devices.
mixture was cooled to room temperature and poured slowly into
ethanol (100 mL). The precipitate was collected by ltration,
washed with hot methanol (on a Soxhlet apparatus) for 24 h,
and dried at 80 ꢀC overnight under reduced pressure to give 0.52
g (70.8% yield) of the yellowish PA-1 powder. FT-IR (cmꢁ1, KBr)
Experimental section
Materials and synthesis
Hydrazine sulfate, p-aminothiophenol, p-uorobenzoic acid, (Fig S1†): 1621 (C]N stretching in oxadiazole unit), 1599 (C]N
triphenylamine, 4,40-dichlorodiphenyl sulfone and p-hydrox- stretching in imine unit). Elemental microanalysis, anal. calcd
ybenzaldehyde were purchased from Aladdin Reagent Co. and for C46H31N5OS2 (repeating unit): C, 75.28; H, 4.26; N, 9.54; S,
1
used as received. 2,5-bis(4-uorophenyl)-1,3,4-oxadiazole, 1,3- 8.74. Found: C, 75.31; H, 4.24; N, 9.58; S, 8.71%. H NMR (400
bis(4-chloro-3-nitrophenyl)sulfone49 and 4,40-diformyl-triphe- MHz, DMF-d7, d, ppm): 10.14 (2H, s), 8.08–6.97 (29H, m).
nylamine50 were synthesized and puried according to the
Synthesis of PA-2. This poly(Schiff base) was prepared by a
literature. All the other reagents and solvents were of analytical similar procedure to the synthesis of PA-1. 0.85 g (86.6% yield)
grade and used without further purication.
of pale PA-2 powder was obtained. FT-IR (cmꢁ1, KBr) (Fig. S1†):
Synthesis of 2,5-bis(p-aminophenylthiophenyl)-1,3,4-oxadi- 1627 (C]N stretching in oxadiazole unit), 1595 (C]N stretch-
azole (M1 in Scheme 1). A mixture of 2,5-bis(4-uorophenyl)- ing in imine unit), 1498 and 1384 (–NO2 stretching in NPS unit),
1,3,4-oxadiazole (5.16 g, 0.020 mol), p-aminothiophenol (5.10 g, 1321 and 1158 (–SO2– stretching in NPS unit). Elemental
0.041 mol), potassium carbonate (5.67 g, 0.041 mol) and N,N- microanalysis, anal. calcd for C45H31N5O6S3 (repeating unit): C,
dimethylformamide (DMF, 40 mL) was placed in a two-necked 64.81; H, 3.75; N, 8.40; S, 11.53. Found: C, 64.91; H, 3.71; N, 8.51;
1
round-bottom ask and heated to 100 ꢀC for 7 h under vigorous S, 11.52%. H NMR (400 MHz, DMF-d7, d, ppm): 10.10 (2H, s),
stirring in an argon atmosphere. The mixture was then cooled 8.93–7.32 (30H, m).
to room temperature and poured slowly into a water–ethanol
Since neither polymer is soluble in tetrahydrofuran (THF) for
mixture (10 : 1, v/v). The white precipitate was ltered, washed the determination of molecular weight by standard gel perme-
with deionized water and dried under reduced pressure at ation chromatography (GPC) analysis, the inherent viscosity of
120 ꢀC overnight. The crude product was recrystallized from the PA–NMP solutions were used to evaluate the formation of
tetrahydrofuran–ethanol (13 : 10, v/v), and a colorless needle macromolecules. The as-synthesized PA-1 and PA-2 show
crystal with a yield of 76.8% (7.20 g) was obtained. FT-IR (cmꢁ1
,
inherent viscosities of 0.45 dL gꢁ1 and 0.51 dL gꢁ1 in NMP,
KBr): 2754 (aldehyde C–H stretching), 3450 and 3371 (NH2 respectively, which allow the fabrication of thin polymer lms
stretching), 1613 (C]N stretching). Elemental microanalysis, by spin-coating processes.
anal. calcd for C26H20N4OS2: C, 66.64; H, 4.30; N, 11.96; S, 13.69.
Found: C, 66.49; H, 4.36; N, 11.98; S, 13.65%. 1H NMR (400
Memory device fabrication
MHz, DMSO-d6, d, ppm): 7.93 (4H, d, J ¼ 8.0 Hz), 7.23 (4H, d, J ¼
8.0 Hz), 7.16 (4H, d, J ¼ 8.0 Hz), 6.66 (4H, d, J ¼ 8.0 Hz) and 5.65 The Pt/Ti/SiO2/Si substrates were pre-cleaned in an ultrasonic
(4H, s).
Synthesis
bath with ethanol, acetone and isopropanol for 20 min each in
of
bis-(3-nitro-4-p-formylphenyloxy-phenyl) that order. A 0.1 mL cyclohexanone solution of the PAs was spin-
sulfone. A mixture of bis(4-chloro-3-nitrophenyl)sulfone (7.54 g, coated onto the substrates at a spinning speed of 400 rpm for
0.020 mol), p-hydroxybenzaldehyde (4.98 g, 0.041 mol), potas- 12 s and then 2000 rpm for 40 s, followed by a vacuum-drying
sium carbonate (5.80 g, 0.042 mol) and N,N-dimethylacetamide treatment at 80 ꢀC for 8 h. Before spin-coating, the solution was
(DMAc, 100 mL) was placed in a three-necked round-bottom ltered through polytetrauoroethylene (PTFE) membrane
ask and heated to 100 ꢀC for 7 h under vigorous stirring in an micro-lters with a pore size of 0.45 mm to remove undissolved
argon atmosphere. The mixture was then cooled to room particles. The thickness of the polymer lms was about 90 nm,
temperature and poured slowly into water (300 mL). The orange as measured with
a spectroscopic ellipsometer (model
precipitate was ltered, washed with deionized water and dried M2000DI, Woollam). To construct the metal/polymer/Pt sand-
ꢀ
under reduced pressure at 80 C overnight. The crude product wich structures, metal top electrodes (Pt or Al) with a diameter
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. C, 2013, 1, 4556–4564 | 4557

View Article Online
Journal of Materials Chemistry C
Paper
Scheme 1 Synthesis routes for the monomers and poly(Schiff base)s.
of 100 mm and thickness of 150 nm were deposited at room Differential scanning calorimetry (DSC) measurements were
temperature by E-beam evaporation through a metal shadow carried out wꢁit1h a Mettler Toledo DSC system at a heating rate
of 10 C min under a nitrogen ow of 50 mL minꢁ1, where
ꢀ
mask.
the glass transition temperature was recorded as the temper-
ature at the middle of the thermal transition from the second
heating scan. Gel permeation chromatography traces of the
samples were recorded on a Viscotek T60A/LR40 GPC system,
where chloroform was used as the eluent, and linear poly-
styrene was used as the standard. Cyclic voltammetry (CV)
measurements were performed on a CHI 650D electrochemical
workstation (Chenhua, Shanghai, China) using a three-elec-
trode cell under argon atmosphere. The polymer lm coated
on a Pt/Ti/SiO2/Si substrate was scanned in an electrolyte
Measurements
The current–voltage (I–V) characteristics of the metal/polymer/
Pt devices were measured in an ambient atmosphere using a
Keithley 4200 semiconductor characterization system in the
voltage sweeping mode. The sweeping step was 0.01 V. During
the measurement, a bias voltage was applied to the top elec-
trodes while the Pt coated substrates were always grounded.
Elemental microanalysis (for C, H and N) was performed on an
Elementar Vario EL III analyzer. 1H nuclear magnetic resonance
(1H NMR) spectra were measured at 400 MHz on a Bruker
400 AVANCE III spectrometer, using dimethyl sulfoxide-d6
(DMSO-d6) or N,N-dimethylformamide-d7 (DMF-d7) as the
solvent. Fourier transform infrared (FT-IR) spectra were recor-
ded on a Thermo Nicolet 6700 FT-IR spectrometer by dispersing
the samples in KBr pellets. The UV-visible absorption spectra
were recorded on a Perkin-Elmer lambda 950 spectrophotom-
eter at room temperature.
solution
of
tetrabutylammonium
hexauorophosphate
(n-Bu4NPF6) in acetonitrile (0.1 M), with a platinum gauze and
Ag/AgCl (3.8 M KCl) as the counter and reference electrode,
respectively. A scan rate of 100 mV sꢁ1 was used during the CV
measurements.
Molecular simulation
Calculations of the optimized geometry and electronic struc-
The inherent viscosities were measured with an Ubbelohde tures, including the dipole moment, HOMO, LUMO and elec-
ꢀ
viscometer at 30 ꢂ 0.1 C in NMP at a concentration of 0.5 g trostatic potential (ESP) surfaces of the donor–acceptor
dLꢁ1. Thermogravimetric analysis (TGA) was performed on a moieties and the basic unit of the PA copolymers were carried
Perkin-Elmer Diamond TG/DTA instrument at a heating rate of out using the Gaussian 09 program package and density func-
10 ꢀC minꢁ1 under a nitrogen ow of 50 mL minꢁ1
4558 | J. Mater. Chem. C, 2013, 1, 4556–4564
.
tional theory (DFT) methods at the B3LYP/6-31 G(d) level.51
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Journal of Materials Chemistry C
Fig. 2 Current–voltage (I–V) characteristics of the (a) Pt/PA-1/Pt and (b) Pt/PA-
2/Pt devices.
Fig. 1 (a) UV-visible absorption spectra of the free-standing PA films. (b) Cyclic
voltammogram profiles of the PA films coated on Pt/Ti/SiO2/Si substrates.
However, the presence of a stronger electron acceptor gives
rise to a stronger reduction peak of PA-2 as compared to that of
PA-1 during the cathodic scan. The reduction peak of PA-2 is
observed with an onset potential of ꢁ0.85 V and a peak
maximum at ꢁ1.16 eV, and the corresponding values for PA-1
are ꢁ0.80 and ꢁ1.03 V, respectively. Therefore, the HOMO
(highest occupied molecular orbital)/LUMO (lowest unoccupied
molecular orbital) energy levels of PA-1 and PA-2 are ꢁ5.05 eV/
ꢁ2.17 eV, and ꢁ6.70 eV/ꢁ3.57 eV, respectively, as deduced
according to eqn (S1)–(S4).†
Results and discussion
Optical and electrochemical properties
Fig. 1(a) shows the UV-visible absorption spectra of the PA lms
which were detached from the glass substrate upon casting and
dried from the NMP solutions. Both poly(Schiff base)s exhibit
two absorption peaks at around 275 nm and 338 nm, respec-
tively, which are attributed to the coupling of the p–p* and
n–p* transitions of the imine-containing polymer backbones
and the benzene rings of the oxadiazole moieties,52 and the
spin-allowed p–p* transitions involving the azomethine-
Memory characteristics of the polymers
phenyl-oxadiazole segment.53 The optical band gaps of PA-1 and The resistive switching behaviors of the present poly(Schiff
PA-2, which are estimated to be 2.88 and 3.12 eV, respectively, base) lms are demonstrated by the current–voltage (I–V)
are also deduced from the onset optical absorbance wavelength characteristics of Fig. 2. At rst, platinum was used as the top
of the polymers.
electrode to fabricate the sandwich devices. As estimated from
Fig. 1(b) shows the cyclic voltammogram of the poly(Schiff Fig. 3(a), the initial resistance of the Pt/PA-1/Pt device is ꢃ3.2 ꢄ
base) lms coated on platinum working electrodes. During the 107 U, and this state can be dened as the high resistance state
anodic scan, a broad yet moderate oxidation peak of the tri- (HRS), or off state, of a memory device. As the voltage sweeps
phenylamine moieties is observed in the PA-1 prole with an from 0 V to ꢁ3.3 V, the current of the device increases slowly in
onset potential of 0.63 V and a peak maximum at 1.17 V, while the range of 2.0 ꢄ 10ꢁ10 to 3.4 ꢄ 10ꢁ7 A. As the sweeping voltage
the oxidation peak in PA-2 is relatively weak since the oxadiazole further increases, an abrupt increase of the device current to the
unit does not exhibit a strong electron donating ability.
magnitude of the compliance current (1 mA) is observed,
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. C, 2013, 1, 4556–4564 | 4559

View Article Online
Journal of Materials Chemistry C
Paper
Fig. 3 Endurance performance of the Pt/PA-1/Pt device.
Fig. 4 Current–voltage (I–V) curves of the (a) Al/PA-1/Pt and (b) Al/PA-2/Pt
indicating that the device has been switched from a HRS to a
low resistance state (LRS, or on state) (sweep 1 in Fig. 2(a)).
This off-to-on transition can be viewed as the “set” or “write”
process in a memory device. The device stays at the on state
devices.
when the power is turned off or during the following voltage are more than ve orders of magnitude during the 20 switching
sweep (sweep 2), indicating the non-volatile characteristic of the cycles. The resistance of the LRS uctuates slightly more severely
resistive switching of Pt/PA-1/Pt device. A reverse sweeping than that of HRS. Fig. 3(b) shows the distribution of the threshold
voltage of 0.9 V can recover the initial off state (sweep 3), and voltages in the cyclic operations. Corresponding to the lighter
this positively-biased sweeping serves as the “reset” or “erase” uctuation of the LRS resistance, the distribution range of the
process of a rewritable device. It is found that a high on/off ratio reset voltages is narrower than that of set process. The average set
(read at 0.1 V) of 107 can be achieved in the present device, and reset voltages for the 20 consecutive cycles are ꢁ2.6 V and
which leads to a low misreading rate for potential memory 0.7 V, respectively. The retention performance of Pt/PA/Pt devices,
applications. However, the transport properties of the Pt/PA-2/ which maintain a stable on/off ratio of more than 105 for at least
Pt devices are completely different (Fig. 2(b)). Although the set 24 h, is shown in Fig. S3.†
process is similar to that of the Pt/PA-1/Pt device, the threshold
Beyond the chemical structure of the conjugated copolymers
voltages (the average switching voltage is ꢁ4.6 V for more than that affects the electronic transport properties of the poly-
20 tested samples) are found to be larger than that of the Pt/PA- azothine derivatives, the choice of electrode metals also plays a
1/Pt devices. Moreover, the device could not be reprogrammed key role on the resistive switching behavior of the metal/poly-
to its initial off state with a positive-voltage sweeping operation, mer/metal sandwich structures. With the chemically more
and behaves as a typical write-once read-many-times (WORM) active aluminum as top electrodes, the Al/PA-1/Pt shows the
memory with an on/off ratio of more than 107.
rewritable resistive switching properties with a set voltage of
Cyclic switching operations have been conducted to ꢁ2.5 V and a lower on/off ratio of ꢃ105, as shown in Fig. 4(a).
further investigate the endurance performance of the rewritable
Interestingly, the memory behavior of the PA-2 based devices
Pt/PA-1/Pt memory devices. Generally, the devices can be changes from that of a WORM type to a rewritable type when the
repeatedly switched between off and on states for less than thirty Pt top electrode is replaced by Al (Fig. 4(b)). The on/off ratio for
cycles. Fig. 3(a) shows the evolution of the resistance of the two the Al/PA-2/Pt is about 3 ꢄ 104. Moreover, the current–voltage
well-resolved states for about 20 cycles. The resistance values responses of both the PA-1 and PA-2 devices with Al top elec-
were read out at 0.1 V in each voltage sweep. The resistance ratios trodes are smoother and more stable than those of the
4560 | J. Mater. Chem. C, 2013, 1, 4556–4564
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Journal of Materials Chemistry C
Fig. 5 (a and c) Endurance and (b and d) threshold voltage distribution of the Al/PA/Pt devices.
Pt-electrode devices. The Al/PA-1/Pt devices exhibit an on/off
operation window of 105, and the average set and reset voltages
are ꢁ2.2 V and 0. 6 V, respectively (Fig. 5(a) and (b)). For the case
of the Al/PA-2/Pt devices, the on/off window of more than 103
can be well maintained for over 100 cycles (Fig. 5(c)). The
average set and reset voltages are found to be ꢁ2.5 V and 0.8 V,
respectively (Fig. 5(d)). Additionally, the retention times of the
Al/PA/Pt devices are also longer than 105 s, during which no
signicant degradation of the resistance states was observed
(Fig. S4†).
Proposed switching mechanisms
To better understand the switching mechanisms of the metal/
PA/metal sandwich structures, molecular simulation, and
spectroscopic as well as electrochemical characterization were
jointly employed to describe the role of the oxadiazole moiety in
the two D–A polyazothines and the electronic scenario of the
memory devices. The molecular orbital and electrostatic
potential surfaces, calculated and experimental HOMO and
LUMO energy levels, and dipole moments of the donor,
acceptor and the repeating units of the PAs, together with the
Fig. 6 HOMO, LUMO, electrostatic potential (ESP) surfaces, and dipole moments of TPA, OXD, NPS and PAs. The color bar indicates the revolution of electrostatic
potential. a Estimated from the UV-visible absorption spectra and CV profiles.
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. C, 2013, 1, 4556–4564 | 4561

View Article Online
Journal of Materials Chemistry C
Paper
providing the electron donating functionality in the PA-2 D–A
structure. Due to the different roles of the TPA, OXD and NPS
groups in the D–A structures of PA-1 and PA-2, it is can be
concluded that the HOMO of PA-1 comes from the TPA donor
with its similar energy level, while the LUMO arises from that of
the OXD acceptor. Similarly, the HOMO of PA-2 is from the OXD
donor while the LUMO is from the NPS acceptor. In the ground
state, the stronger electron donating ability of the TPA moieties
leads to more intense charge transfer interactions with the OXD
acceptors. Therefore, the pristine PA-1 thin lm demonstrates a
lower resistance (read at ꢁ1 V) as compared to the pristine PA-2
thin lm. Under a negative electric eld with low magnitude,
electrons are injected from the electrode into the polymer lm
and trapped by the OXD acceptors to form a space charge layer,
which in turn hinders the further injection of electron charge
carriers. At the threshold voltage, charge transfer and separa-
tion occur between the TPA and OXD D–A pairs of PA-1, as well
as between the OXD and NPS moieties of PA-2, and the straight
arrows in Fig. 7(b) represent the occurrence of electron transi-
tion from the HOMO into the LUMO of the polymers. The
charge separated state of the D–A pairs increases the carrier
concentration inside the solid state polymer thin lm signi-
cantly, and switches the device from the high resistance state to
the low resistance state. The curved arrows in Fig. 7(b) indicate
the transport of separated electrons to the metal electrode,
which is done in a closed circuit to conduct DC current. Due to
the stronger electron push–pull interaction in the D–A struc-
ture, the CT effect occurs earlier and more easily in the thin lm
of PA-1, resulting in a lower switching threshold voltage than
that of the PA-2 device. Meanwhile, the moderate electron
withdrawing ability of the OXD acceptor makes the CT inter-
action between the TPA-OXD pairs reversible, and a sufficient
reverse electric eld can initiate the back transfer of the sepa-
rated charge carriers. As a result, PA-1 recovers its initial neutral
state and the device returns to its pristine high resistance state
with rewritable memory behavior. However, the strong electron
withdrawing ability of the NPS moieties causes an irreversible
CT interaction between the OXD-NPS pairs, and the PA-2 based
devices behave as a WORM memory. As platinum is chemically
inert, the electrode metal in the Pt/PA/Pt devices will not be
ionized into cations nor driven into the polymer lms under
electric elds. The formation of a conductive lament by
diffusion of Pt nanoparticles into the polymer thin lm during
E-beam evaporation process is unlikely to occur either, as the
as-fabricated device should be in its initial low resistance state
instead of the observed high resistance state. Thus, the resistive
switching behaviors in the present Pt-electrode devices may not
be accounted for by the formation/rupture of metal laments.
The dipole moment of the basic unit of the polymers is an
important issue that inuences the volatility of the resistive
Fig. 7 (a) Energy diagram of M/PA/M sandwich systems, and the different RS
mechanisms for the (b) Pt/PA/Pt and (c) Al/PA/Pt devices.
work functions of the metal electrodes, are summarized in
Fig. 6 and Fig. 7.
From the molecular orbital energy level diagram of the TPA, switching behaviors, and larger dipole moments favor the reten-
OXD and NPS moieties, it can be clearly seen that both the tion of separated charge carriers and is benecial for a non-volatile
HOMO and LUMO levels of the OXD groups are lower than that memory effect. Beyond the dipole moment, there do exist other
of the TPA moieties. Therefore, TPA is the donor component in strategies that will affect the volatility of the polymer memories.
the PA-1 D–A structure and OXD is the acceptor component. In For instance, one reported work54 claimed that by varying the
comparison with those of the NPS moiety, the molecular orbital thickness of a thin lm constructed from a D–A polymer PI (3,7-
energy levels of the OXD group are relatively higher, thus APDBT-6FDA) with the small dipole moment of 2.20 debye,
4562 | J. Mater. Chem. C, 2013, 1, 4556–4564
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Journal of Materials Chemistry C
various electrical responses, namely high-conductance ohmic 2009CB930803, 2012CB933004), the Chinese Academy of
current ow, non-volatile negative differential resistance, dynamic Sciences (CAS), the National Natural Science Foundation of
random access memory and insulation are demonstrated. With China, the Ningbo Science and Technology Innovation Team
the azothine-based copolymers carrying larger dipole moments of (2009B21005, 2011B82004), and the Zhejiang and Ningbo
2.5009 debye and 2.5839 debye, respectively, which are more Natural Science Foundations.
favorable for holding separated charge carriers, as well as the
much greater thickness (100 nm) of the lm which may provide
additional trapping centers via chain end, chain folds, etc., it is
Notes and references
reasonable that the present devices based on PA-1 and PA-2
demonstrate non-volatile resistive switching behaviors.
1 Z. J. Donhauser, B. A. Mantooth, K. F. Kelly, L. A. Bumm,
J. D. Monnell, J. J. Stapleton, D. W. Price Jr, A. M. Rawlett,
D. L. Allara, J. M. Tour and P. S. Weiss, Science, 2001, 292,
2303.
2 H. Li, Z. Jin, N. Li, Q. Xu, H. Gu, J. M. Lu, X. Xia and L. Wang,
J. Mater. Chem., 2011, 21, 5860.
3 B. Cho, S. Song, Y. Ji, T. W. Kim and T. J. Lee, Adv. Funct.
Mater., 2011, 21, 2806–2829.
4 N. G. Kang, B. Cho, B. G. Kang, S. Song, T. J. Lee and J. S. Lee,
Adv. Mater., 2012, 24, 385–390.
5 S. M. Miao, Y. X. Zhu, H. Zhuang, X. P. Xu, H. Li, R. Sun,
N. J. Li, S. J. Ji and J. M. Lu, J. Mater. Chem. C, 2013, 1, 2320.
6 Y. G. Ko, W. Kwon, H. J. Yen, C. W. Chang, D. M. Kim,
K. Kim, S. G. Hahm, T. J. Lee, G. S. Liou and M. Ree,
Macromolecules, 2012, 45, 3749.
7 S. J. Liu, W. P. Lin, M. D. Yi, W. J. Xu, C. Tang, Q. Zhao,
S. H. Ye, X. M. Liu and W. Huang, J. Mater. Chem., 2012,
22, 22964.
When the Pt electrodes are replaced by the more chemically
active metal Al, the Schottky barrier at the metal–polymer inter-
faces becomes smaller (Fig. 7(c)). Consequently, it is easier for
electron injection from Al into the LUMO of PA-1 to occur (as
shown by the curved arrows of Fig. 7(c)), and the Al-electrode
devices exhibit lower switching threshold voltages compared to
those of the Pt-electrode devices. Similarly, the back transfer also
occurs more easily to give a lower reset voltage. It is also note-
worthy that the interfacial effect arising from the presence of an
ultrathin layer of Al2O3 cannot be excluded at the moment, and the
resistive switching behaviors of the PAs, PA-2 in particular, will be
affected. For instance, the Al-electrode devices based on PA-2
demonstrate a rewritable characteristic, rather than the WORM
behavior of the Pt-electrode device. To date, the inuence of the
interfacial effect on the memory behavior of polymer devices is
still unclear and is beyond the scope of the present study.
8 C. J. Chen, Y. C. Hu and G. S. Liou, Chem. Commun., 2013, 49,
2536.
Conclusions
9 H. Li, Q. Xu, N. Li, R. Sun, J. Ge, J. M. Lu, H. Gu and F. Yan,
J. Am. Chem. Soc., 2010, 132, 5542.
Two D–A type polyazothine derivatives based on triphenyl-
amine, oxadiazole and 3,30-dinitro-diphenylsulfone moieties 10 X. D. Zhuang, Y. Chen, G. Liu, P. P. Li, C. X. Zhu, E. T. Kang,
have been successfully synthesized for memory device applica-
tions. By varying the chemical structure of the electroactive
K. G. Noeh, B. Zhang, J. H. Zhu and Y. X. Li, Adv. Mater.,
2010, 22, 1731.
polymers, as well as the top electrode metal, the four kinds of 11 Y. Ma, X. Cao, G. Li, Y. Wen, Y. Yang, J. Wang, S. X. Du,
memory devices, Pt/PA-1/Pt, Pt/PA-2/Pt, Al/PA-1/Pt and Al/PA-2/
Pt, are found to exhibit different resistive switching properties.
L. M. Yang, H. Gao and Y. L. Song, Adv. Funct. Mater.,
2010, 20, 803.
In the case of the Pt/PA/Pt devices, PA-1 shows rewriteable 12 F. Zhao, J. Liu, X. Huang, X. Zou, G. Lu, P. Sun, S. Wu, W. Ai,
memory behavior with poor endurance of less than 20 cycles,
while PA-2 exhibits write-once read-many-times (WORM)
M. Yi, X. Qi, L. Xie, J. Wang, H. Zhang and W. Huang, ACS
Nano, 2012, 6, 3027.
memory characteristics. The different memory properties of the 13 K. Kim, H. J. Yen, Y. G. Ko, C. W. Chang, W. Kwon, G. S. Liou
Pt-electrode devices originate from the different HOMO–LUMO and M. Ree, Polymer, 2012, 53, 4135.
band gaps of the PAs, which inuence the degree of intra- and 14 G. Liu, Q.-D. Ling, E.-T. Kang, K.-G. Neoh, D. J. Liaw,
intermolecular charge transfer interaction between the electron
donor–acceptor pairs. By using Al to replace the Pt top elec-
F. C. Chang, C.-X. Zhu and D. S. H. Chan, J. Appl. Phys.,
2007, 102, 024502.
trodes, the Al/PA/Pt devices exhibit a repeatable resistive 15 B. Hu, R. Quhe, C. Chen, F. Zhuge, X. Zhu, S. Peng, X. Chen,
switching effect, although the endurance of the PA-1 devices is
worse than that of the PA-2 devices. Besides the reduced
L. Pan, Y. Wu, W. Zheng, Q. Yan, J. Lu and R.-W. Li, J. Mater.
Chem., 2012, 22, 16422.
Schottky barrier at the metal–polymer interfaces that promotes 16 B. Y. Ji, B. Cho, S. Song, T. W. Kim, M. Choe, Y. H. Kahng and
charge carrier injection from the electrode into the LUMO of the T. Lee, Adv. Mater., 2010, 22, 3071.
polymer material, the presence of an Al2O3 layer may also play 17 G. Liu, B. Zhang, Y. Chen, C.-X. Zhu, L. Zeng, D. Siu-Hung
an important role in regulating the electrical properties of the
metal/polymer/metal sandwich structures.
Chan, K.-G. Neoh, J. Chen and E.-T. Kang, J. Mater. Chem.,
2011, 21, 6027.
18 B. Hu, X. Zhu, X. Chen, L. Pan, S. Peng, Y. Wu, J. Shang,
G. Liu, Q. Yan and R.-W. Li, J. Am. Chem. Soc., 2012, 134,
17408.
Acknowledgements
The authors acknowledge nancial support from the State Key 19 J. J. Yang, M. D. Pickett, X. Li, D. A. A. Ohlberg, D. R. Stewart
Project of Fundamental Research of China (973 Program,
and R. S. Williams, Nat. Nanotechnol., 2008, 3, 429.
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. C, 2013, 1, 4556–4564 | 4563

View Article Online
Journal of Materials Chemistry C
Paper
20 A. Iwan and D. Sek, Prog. Polym. Sci., 2008, 33, 289.
37 T. Kurosawa, Y. C. Lai, T. Higashihara, M. Ueda, C. L. Liu
21 A. Nayak, T. Ohno, T. Tsuruoka, K. Terabe, T. Hasegawa,
and W. C. Chen, Macromolecules, 2012, 45, 4556.
J. K. Gimzewski and M. Aono, Adv. Funct. Mater., 2012, 22, 38 T. Kurosawa, Y. C. Lai, A. D. Yu, H. C. Wu, T. Higashihara,
3606.
M. Ueda and W. C. Chen, J. Polym. Sci., Part A: Polym.
Chem., 2013, 51, 1348.
39 J. C. Chen, Y. C. Liu, J. J. Ju, C. J. Chiang and Y. T. Chern,
Polymer, 2011, 52, 954.
40 A. Bolduc, S. Dufresne and W. G. Skene, J. Mater. Chem.,
2010, 20, 4820.
22 S. Yu, B. Gao, Z. Fang, H. Yu, J. Kang and H. S. P. Wong, Adv.
Mater., 2013, 25, 1774.
23 S. H. Jo, T. Chang, I. Ebong, B. B. Bhadviya, P. Mazumder
and W. Lu, Nano Lett., 2010, 10, 1297.
24 M. Ziegler, R. Soni, T. Patelczyk, M. Ignatov, T. Bartsch,
P. Meuffels and H. Kohlstedt, Adv. Funct. Mater., 2012, 22, 41 S. Barik and W. G. Skene, Macromolecules, 2010, 43, 10435.
2744.
42 H. H. Chang, C. E. Tsai, Y. Y. Lai, D. Y. Chiou, S. L. Hsu,
25 W. Lu, Nat. Mater., 2013, 12, 93.
C. S. Hsu and Y. J. Cheng, Macromolecules, 2012, 45, 9282.
26 M. D. Pickett, G. Medeiros-Ribeiro and R. S. Williams, Nat. 43 G. Hughes and M. R. Bryce, J. Mater. Chem., 2005, 15, 94.
Mater., 2013, 12, 114.
27 Q. Xia, W. Robinett, M. W. Cumbie, N. Banerjee,
T. J. Cardinali, J. J. Yang, W. Wu, X. Li, W. M. Tong,
44 J. A. Mikroyannidis, I. K. Spiliopoulos, T. S. Kasimis,
A. P. Kulkarni and S. A. Jenekhe, Macromolecules, 2003, 36,
9295.
D. B. Strukov, G. S. Snider, G. Medeiros-Ribeiro and 45 Y. L. Liu, K. L. Wang, G. S. Huang, C. X. Zhu, E. S. Tok,
R. S. Williams, Nano Lett., 2009, 9, 3640. K. G. Neoh and E. T. Kang, Chem. Mater., 2009, 21, 3391.
28 J. Borghetti, G. S. Snider, P. J. Kuekes, J. J. Yang, D. R. Stewart 46 Y. K. Fang, C. L. Liu and W. C. Chen, J. Mater. Chem., 2011,
and R. S. Williams, Nature, 2010, 464, 873. 21, 4778.
29 Q. D. Ling, F. C. Chang, Y. Song, C. X. Zhu, D. J. Liaw, 47 W. Kwon, B. Ahn, D. M. Kim, Y. G. Ko, S. G. Hahm, Y. Kim,
D. S. H. Chan, E. T. Kang and K. G. Neoh, J. Am. Chem.
Soc., 2006, 128, 8732.
30 E. E. Havinga, W. Tenhoeve and H. Wynberg, Polym. Bull.,
1992, 29, 119.
31 V. Mullekom, J. Vekemans, E. E. Havinga and E. W. Meijer,
Mater. Sci. Eng., R, 2001, 32, 1.
32 P. M. Beaujuge, C. M. Amb and J. R. Reynolds, Acc. Chem.
Res., 2010, 43, 1396.
H. Kim and M. Ree, J. Phys. Chem. C, 2011, 115, 19355.
48 L. Zeng, G. Liu, B. Zhang, J. Chen, Y. Chen and E. T. Kang,
Polym. J., 2012, 44, 257.
49 Y. Imai, M. Ueda and K. Otaira, J. Polym. Sci., Polym. Chem.
Ed., 1977, 15, 1457.
50 D. Sek, A. Iwan, B. Jarzabek, B. Kaczmarczyk, J. Kasperczyk,
Z. Mazurak, M. Domanski, K. Karon and M. Lapkowski,
Macromolecules, 2008, 41, 6653.
33 S. G. Hahm, S. Choi, S. H. Hong, T. J. Lee, S. Park, D. M. Kim, 51 M. J. Frish, G. W. Trucks, H. B. Schiegel, G. E. Scuseria,
W. S. Kwon, K. Kim, O. Kim and M. Ree, Adv. Funct. Mater.,
2008, 18, 3276.
M. A. Robb and J. R. Cheeseman, Gaussian 09 (Revision
A.02), Gaussian Inc., Wallingford CT, 2009.
34 J. Hu, Y. Li, Z. Ji, G. Jiang, L. Yang, W. Hu, H. Gao, L. Jiang, 52 E. Wolarz, E. Chrzumnicka, T. Fischer and J. Stumpe, Dyes
Y. Wen, Y. L. Song and D. Zhu, J. Mater. Chem., 2007, 17, 3530. Pigm., 2007, 75, 753.
35 Q. Liu, K. Jiang, L. Wang, Y. Wen, J. Wang, Y. Ma and 53 L. Marin, M. D. Damaceanu and D. Timpu, So Mater., 2009,
Y. Song, Appl. Phys. Lett., 2010, 96, 213305. 7, 1.
36 J. J. Yang, D. B. Strukov and D. R. Stewart, Nat. Nanotechnol., 54 C. L. Liu, T. Kurosawa, A. D. Yu, T. Higashihara, M. Ueda and
2013, 8, 13.
W. C. Chen, J. Phys. Chem. C, 2011, 115, 5930.
4564 | J. Mater. Chem. C, 2013, 1, 4556–4564
This journal is ª The Royal Society of Chemistry 2013