Full Paper
ꢀ
1
PMNN, which had a significant influence on the memory be-
havior of the corresponding devices. Thus, we wanted to
design an “ideal” polymer that could improve upon the indi-
vidual performances of PMNB and PMNN. In addition, our pre-
viously designed polymers were all obtained through random
copolymerization; thus, we also wanted to consider the effect
of other polymerization methods, such as block polymeri-
zation, on memory performance.
PMCz -b-PMBNa , was 12551 gmol , with a PDI of 1.33. Thus,
8 2
the relative ratio of MCz/MBNa in block polymer PMCz -b-
8
PMBNa was 8:2. Both polymers exhibited good thermal stabili-
2
ty, with a 5% decomposition temperature (T ) of about 2808C
and a glass-transition temperature (T ) of about 1008C, thus
implying good heat endurance for both polymers PMCz -co-
d
g
8
PMBNa and PMCz -b-PMBNa in their corresponding memory
2
8
2
devices.
Carbazole and 1,8-naphthalimide groups both possess a
[
43–45]
high degree of p-conjugated planarity,
which can induce a
Current–Voltage (I–V) Characteristics of PMCz -co-PMBNa
8
2
conformational change between the Cz and Na planes under
an external electrical field. Furthermore, the Cz group, which
contains a O=CꢀOꢀCꢀC spacer group in its side chain, has
been shown to exhibit binary “write once read many” (WORM)
memory behavior through a conformational-change mecha-
and PMCz -b-PMBNa2
8
Figure 2a shows the structure of the memory devices, in which
a thin film (thickness: ca. 80 nm) was sandwiched between an
aluminum (Al) electrode and an ITO electrode (for a cross-sec-
tional SEM image, see Figure 2b). Typical I–V characteristics of
[
33]
nism.
donors and Na chromophores are well-known electron accept-
In addition, Cz groups are also attractive electron
ꢀ
1
the memory devices were determined at a scan rate 0.1 Vs .
[
24]
ors, thereby readily allowing charge-transfer or charge-trap-
ping mechanisms in the side chains. Finally, the same flexible
Figure 2c shows the I–V characteristics of an ITO/PMCz -co-
8
PMBNa /Al memory device, which was initially in its low-con-
2
3
O=CꢀOꢀCꢀC chains are connected to nitrogen atoms with sp
ductivity state (OFF or “0” state). When a negative voltage
from 0 to ꢀ6 V was applied to the cell, two abrupt increases in
current were observed at threshold voltages of ꢀ1.7 and
ꢀ3.3 V, thus indicating the occurrence of two successive elec-
trical transitions from the OFF state to an intermediate conduc-
tance state (ON1 or “1” state) and from the ON1 state to a
high-conductivity state (ON2 or “2” state). The corresponding
OFF–ON1 and ON1–ON2 transitions could both serve as the
“writing” process in the memory device. The memory device
remained in the ON2 state under a subsequent voltage sweep
from 0 to ꢀ6.0 V. The long-term stability and performance of
the device was also evaluated from the retention time and a
stimulus effect test under the same conditions. As shown in
the Supporting Information, Figure S1, the PMCz -co-PMBNa -
hybridization at pendant Na moieties, which is theoretically fa-
vorable for the arrangement of Cz and Na moieties under an
external stimulus.
Herein, we report the synthesis of two polymers, PMCz -co-
8
PMBNa and PMCz -b-PMBNa , through random copolymeriza-
2
8
2
tion and block polymerization, respectively. I–V measurements
showed that both polymers exhibited ternary memory per-
formance, and the corresponding threshold voltages were
ꢀ
1.7 V/ꢀ3.3 V and ꢀ2.7 V/ꢀ3.8 V, respectively. However, ITO/
PMCz -b-PMBNa /LiF/Al (ITO=indium tin oxide) only exhibited
8
2
binary memory performance, whereas ITO/PMCz -co-PMBNa /
8
2
LiF/Al still exhibited the typical ternary memory performance.
Based on AFM, XRD, UV/Vis, and photoluminescence (PL) spec-
troscopy experiments, we attributed the ternary memory prop-
erties of PMCz -co-PMBNa and PMCz -b-PMBNa to the opera-
8
2
8
based device endured over 1ꢁ10 continuous read pluses of
ꢀ1.0 V, and no significant degradation in the current for any
state was observed after 100 min.
8
2
8
2
tion of different conduction mechanisms. We expect that this
work will guide the design of new polymers with improved
memory performance.
Subsequently, to exclude the possibility of aluminum parti-
cles penetrating the active layer during the electrical measure-
ments, a LiF layer (thickness: ca. 5 nm) was vacuum-deposited
onto the PMCz -co-PMBNa2 film as a buffer layer to prevent
8
Results and Discussion
direct contact with the Al electrode. Then, the electrical behav-
ior of the ITO/PMCz -co-PMBNa /LiF/Al device was measured
8
2
Synthesis and Characterization of PMCz -co-PMBNa and
8
2
under the same conditions as before. As shown in Figure 3a,
this device also exhibited ternary memory behavior, thus indi-
cating that this behavior originated from the intrinsic proper-
ties of the active layer itself. Thus, a PMCz -co-PMBNa -based
PMCz -b-PMBNa2
8
Scheme 1 shows our synthetic route to polymers PMCz -co-
8
PMBNa and PMCz -b-PMBNa ; for detailed procedures, see the
2
8
2
8
2
Supporting Information. The chemical structures of these poly-
memory device exhibited nonvolatile ternary WORM-type
memory behavior.
1
mers were confirmed by using H NMR spectroscopy (Figure 1).
Both polymers could be spin-cast into uniform thin films from
Similarly, the ITO/PMCz -b-PMBNa /Al device also showed
8
2
ꢀ
1
solutions in cyclohexanone (12 mgmL ). The relative molecu-
typical ternary WORM-type data-storage performance, as
shown in Figure 2d, and the threshold voltage for the OFF–
ON1 and ON1–ON2 transitions were ꢀ2.7 and ꢀ3.8 V, respec-
tively, much larger than the corresponding values for the
PMCz -co-PMBNa -based memory device. Then, the LiF layer
lar weight (M ) of random copolymer PMCz -co-PMBNa was
n
8
2
ꢀ
1
2
0205 gmol , with a polydispersity index (PDI) of 1.52, and
the ratio of MCz/MBNa segments was measured by using the
UV/Vis absorption of a solution of MBNa in DMF (see ref [41]).
8
2
Subsequently, the M of polymer R-PMCz [R=Reversible Addi-
was also applied to consider the possibility of metal-filament
formation in the electrical measurements. Unexpectedly, the
ITO/PMCz -b-PMBNa /LiF/Al device only showed binary
n
tion-Fragmentation Chain Transfer Polymerization (RAFT)] was
9
447, with a PDI of 1.26, whilst that of the final block polymer,
8
2
&
&
Chem. Asian J. 2018, 00, 0 – 0
2
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!