Dual-Mechanism-Controlled Ternary Memory ... - ACS Publications

Nov 6, 2012 - Moreover, when combined with the 3D-stacking architecture, the data-storage capacity of the device may significantly be increased. There...
53 downloads 11 Views 3MB Size
Article pubs.acs.org/JPCC

Dual-Mechanism-Controlled Ternary Memory Devices Fabricated by Random Copolymers with Pendent Carbazole and Nitro-Azobenzene Hao Zhuang,† Xiaoping Xu,† Yuanhua Liu,† Qianhao Zhou,† Xufeng Xu,† Hua Li,*,† Qingfeng Xu,† Najun Li,† Jianmei Lu,*,†,‡ and Lihua Wang† †

Key Laboratory of Organic Synthesis of Jiangsu Province, College of Chemistry, Chemical Engineering and Materials Science, Soochow University, 199 Ren’ai Road, Suzhou 215123, People's Republic of China ‡ Institute of Chemical Power Sources, Soochow University, Suzhou, Jiangsu 215006, People's Republic of China S Supporting Information *

ABSTRACT: In the designed copolymers of poly(9-(4-vinyl benzyl)-9H-carbazole) (PVCz)-random-poly(1-(4-nitro-azophenyl)-pyrrole-2, 5-dione) (PMIDO3), carbazole donor and nitro-azobenzene acceptor are introduced into the lateral chains to induce charge transfer under an electric field, and the resulting copolymers have a progressive increase in the nitrogen content to induce filamentary conduction. The fabricated devices with a simple sandwich configuration distinctively exhibit three conductivity states when a negative bias is applied which can be encoded as “0”, “1” and “2” for future ternary data storage. The fabricated devices could endure 108 read cycles and showed a long retention time of 104 s. It is worth noting that the switch-OFF voltages and the OFF-currents are significantly affected by the MIDO3 content within the polymers. The experimental results indicate that charge transfer between the donor and acceptor is responsible for the first switching and the second switching can be attributed to the filamentary rupture. This approach of achieving devices with multistable states through combination of different switching mechanisms in one device may provide a framework for the design and selection of multilevel electrical memory materials in future research.

1. INTRODUCTION As an effective solution to the ever-increasing requirements for high density data storage, electrically tristable organic and polymeric memory materials have recently attracted much attention.1,2 Compared with conventional inorganic siliconbased counterparts, organic and polymeric materials have advantages in their ease of miniaturization and tailored properties through molecular design.3,4 Moreover, polymeric materials are more favorable due to their easy processability, high mechanical strength, and 3D-stacking capability to realize low-cost high-capacity memory devices for practical use.5−8 Significant effort has been made to study polymer-based ternary memory devices since the first report of small-molecule based nonvolatile ternary memory devices1 for permanent data storage. Until, various binary polymeric memory materials have been reported. Their memory behaviors generally cover the following types: write-once read-many-times (WORM) memory,9−14 flash memory,6,11,15−18 dynamic random access memory (DRAM),19,20 and static random access memory (SRAM).20 In these devices, several mechanisms have been proposed to understand the switching phenomena, such as carrier trapping and detrapping,21 charge transfer between the donor and acceptor moieties,12,22 redox,23 field-induced conformational change,17,24 and the filamentary conduction theory.16,25 To the best of our knowledge, nearly all of the reported devices are determined by one single mechanism during the conductivity transition process. A memory device © 2012 American Chemical Society

controlled by two or more mechanisms to achieve more than two states in each memory cell has rarely been studied. In this work, we designed and synthesized a series of sidechain conjugated random copolymers, poly(9-(4-vinyl benzyl)9H-carbazole)-random-poly(1-(4-nitro-azo-phenyl)-pyrrole-2, 5-dione), denoted P(VCz)x(MIDO3)y (Figure 1a). Electrondonating (i.e., carbazole) and -withdrawing (i.e., nitroazobenzene) entities with high nitrogen contents in the lateral chains are to induce charge transfer and filamentary conduction. The as-fabricated devices underwent an abrupt current increase from OFF- to ON-state during the negative voltage scan, and could only be partly switched off to an intermediate-state under a higher bias, consequently exhibiting electrically tristable states. Measurements and theoretical analyses were carried out to investigate the switching mechanisms. All of the results indicate that the first transition from OFF- to ON-state originated from the field-induced charge transfer and the second transition could be attributed to the filamentary rupture. It is noteworthy that the switch-OFF voltages varied with the composition of the copolymers while the switch-ON voltage was independent of the copolymer’s composition. Received: July 19, 2012 Revised: October 30, 2012 Published: November 6, 2012 25546

dx.doi.org/10.1021/jp307156c | J. Phys. Chem. C 2012, 116, 25546−25551

The Journal of Physical Chemistry C

Article

Figure 1. (a) Synthesis of the random copolymers; (b) Scheme of the prototype device fabricated with the P(VCz)x(MIDO3)y and aluminum (Al) top and indium−tin oxide (ITO) glass bottom electrode.

Table 1. Molecular Weight (Mn), Polydispersity (PDI), Composition, and Thermal Properties of the Random Copolymers GPC result polymer

[VCz]0/[MIDO3]0

P(VCz)45(MIDO3)5 P(VCz)41(MIDO3)9 P(VCz)18(MIDO3)7 P(VCz)33(MIDO3)16 P(VCz)29(MIDO3)21

20/1 10/1 5/1 3/1 1/1

a

Mn

PDI

MIDO3b %

Td (°C)c

6900 7900 8800 7100 7400

1.72 1.79 1.66 1.70 1.65

10.4 17.6 28.0 33.2 42.0

220 263 265 319 318

a

The initial feeding ratio of the two monomers. bThe weight percentage of MIDO3 moieties in the copolymers were calculated according to the UV−vis absorption standard curve. cThermal decomposition temperature (5% weight-loss).

2. EXPERIMENTAL SECTION 2.1. Materials. Reagent-grade Disperse Orange 3 (4-Nitro4′-aminoazobenzene) was purchased from Acros. 4-chloromethyl styrene was ordered from Fluka. Carbazole and maleic anhydride were obtained from Alfa. AIBN was purchased from Sigma-Aldrich and used after recrystallization with methanol. Cyclohexanone was distilled before use. All other materials were used as received. 2.2. Instrumentation and Characterizations. 1H NMR spectra were measured on a NMRststem 300 MHz and INOVA 400 MHz spectrometer. Elemental analysis was performed on a Carlo Erba-MOD1106 elemental analyzer. UV−visible absorption spectra were recorded on a Perkin-Elmer λ-17 UV−vis spectrometer. Weight-average (Mw) and number-average molecular weight (Mn) were determined by gel permeation chromatography (GPC) on Waters1515 Gel Chromatograph with tetrahydrofuran (THF) as the eluent. Monodispersed polystyrene samples were used as the standards. Cyclic voltammetry (CV) measurements were carried out in 0.1 M acetonitrile solution of tetra-n-butyl-ammonium hexafluorophosphate (TBAPF6) with a CHI-660C electrochemical workstation with a platinum gauze auxiliary electrode and an Ag/AgCl reference electrode. The devices were characterized at

room temperature under ambient conditions, using a HewlettPackard 4145B semiconductor parameter analyzer with Hewlett-Packard 8110A pulse generator. 2.3. Synthesis of Monomers. The synthesis and characterization of the two monomers are shown in Scheme S1 and Figure S1 in the Supporting Information. 2.4. General Polymerization Procedure of Random Copolymers. Synthetic routes for the copolymers are illustrated in Figure 1a. The polymerization was carried out at 80 °C in cyclohexanone for 48 h with AIBN as the initiator. The product was precipitated in methanol and the resulting solid was further washed with methanol repeatedly and finally dried in a vacuum oven. The characterizations, including details for the calculation of the random copolymers’ compositions (relative ratios of D/A segments) are provided in the Supporting Information (Figures S2−S6). 2.5. Fabrication and Measurements of the ITO/ P(VCz)x(MIDO3)y/Al Electrical Memory Devices. The indium−tin oxide (ITO) glass was precleaned by sonication for 15 min each with deionized water, acetone and ethanol, in that order. Then 1,2-dichloroethane solution of P(VCz)x(MIDO3)y (10 mg mL−1) was spin-coated onto the ITO-glass substrate at a rotational speed of 2000 rpm for 40 s. The resulting film was annealed at 80 °C in a vacuum oven for 25547

dx.doi.org/10.1021/jp307156c | J. Phys. Chem. C 2012, 116, 25546−25551

The Journal of Physical Chemistry C

Article

12 h. The thickness of the film was typically about 70−80 nm. Finally, Al was thermally evaporated onto the film surface at 5 × 10−4 Pa through a shadow mask to yield top electrodes with thickness of about 100 nm and area of 0.0314 mm2. The fabricated memory device is shown in Figure 1 b.

threshold, the device could then be switched off to an intermediate-state (i.e., a conductivity between the ON-state and the OFF-state). The observed behavior could be reproducibly obtained in other cells of the device and in three more devices, excluding the possibility that the tristable states may be arising from the instability of the device. The distinctive low-, intermediate-, and high-conductivity states (i.e., different responses to external electric field) can be programmed to correspond to “0”, “1”, and “2” signals, respectively, suggesting the device’s potential application for ternary data storage. Moreover, when combined with the 3Dstacking architecture, the data-storage capacity of the device may significantly be increased. Therefore, this ternary ITO/ P(VCz)29(MIDO3)21/Al memory device could be utilized as a WORM-type memory for permanent large-capacity data storage.1 To investigate whether this particular switching behavior is common in all these VCz-MIDO3-based random copolymers, the memory characteristics of copolymers with different compositions (i.e., different content of donor/acceptor segments) were systematically studied. The memory performance of random copolymers with different compositions was shown in Figure S8 of the Supporting Information. During the sweep from 0 to −6.0 V, the devices were initially in a relatively low conductivity state (current level ∼10−6 A), and when the bias reached the threshold of about −1.8 V (the switch-ON voltage), the devices were all switched from the pristine OFFstate to the ON-state. However, when applying a higher bias, the switch-OFF voltages (i.e., threshold voltages for the transition from ON-state to intermediate-state) varied considerably. The switch-OFF voltage for P(VCz)29(MIDO3)21 was the highest, and interestingly the switch-OFF voltages decreased with the decrease in the MIDO3 content within the copolymers. It is worth noting that the second transition from the ON-state to intermediate-state become less obvious with decreasing MIDO3 content, eventually leading to the disappearance of the second switching for P(VCz)45(MIDO3)5. Therefore, the memory behavior of the devices could be controlled by the content of MIDO3 entities in the copolymers from single switching (bistable state) to double switching (tristable state) with various switch-OFF voltages. The stability performance of the three conductive states was evaluated for the ITO/P(VCz)29(MIDO3)21/Al memory device at room temperature under ambient air conditions. Figure 3a shows representative results of the retention time test under a constant stress of −1 V for the ON-, intermediate- and OFF-states of the device. As can be seen, during the entire test period of 104 s, no significant degradation in the current occurred for all the three distinctive states. Moreover, the effect of continuous read pulses of −1 V on the ON-, intermediateand OFF-states was also investigated (Figure 3b). A pulse with width and duration of 1 and 2 μs, respectively, which are typical values in practical devices, was adopted. No obvious degradation in the current was observed for the ON-, intermediate- and OFF-states after more than one hundred million (108) continuous read cycles, indicating that the three states were all insensitive to the read cycles. Figure 4 shows the UV−vis optical absorption spectrum and the cyclic voltammetry (CV) results for the copolymer P(VCz)29(MIDO3)21 thin films. The optical energy barrier for P(VCz)29(MIDO3)21 is about 2.27 eV according to the onset absorption wavelength of the copolymer. The energy levels of the highest occupied molecular orbital (HOMO) and

3. RESULTS AND DISCUSSION The obtained random copolymers have good thermal stability with the decomposition temperature of P(VCz)33(MIDO3)16 and P(VCz)29(MIDO3)21 higher than 310 °C (see Figure S7 in the Supporting Information). Furthermore, the synthesized copolymers exhibit good solubility in common organic solvents such as tetrahydrofuran (THF), dichloromethane, and odichlorobenzene. Therefore, a simple spin-coating method could be employed to prepare the nanoscale thin film. The number-average molecular weight (Mn) of the five copolymers are 6900, 7900, 8800, 7100, and 7400 with the corresponding PDI of 1.72, 1.79, 1.66, 1.70, and 1.65, respectively, as shown in Table 1. Figure 2 shows the typical I−V characteristics of the memory device fabricated with P(VCz)29(MIDO3)21 as the active layer.

Figure 2. Typical I−V curves of the ITO/P(VCz)29(MIDO3)21/Al memory device with current compliance at 0.05A.

As can be seen from the figure, during the voltage sweep, the as-fabricated device exhibits a conductance switching behavior different from those previously reported. Initially, the device was in a low-conductivity (OFF) state (current level ∼10−6 A). However, as the negative bias increased, a sharp transition from the low-conductivity (OFF) state to a high-conductivity (ON) state was observed at −1.7 V (the switch-ON voltage), as was indicated by the abrupt increase in the current (i.e., from 10−6 A to 10−2 A) (sweep 1). This transition from OFF- to ON-state could serve as the “writing” process for practical data storage application. The obtained ON-state could be well retained, when the power was shut off or even during the reverse voltage sweep to +3 V (sweep 2 and 3). However, when the negative bias was even higher (i.e., the fourth sweep from 0 to −6.0 V), the current underwent an abrupt decrease to 10−4 A (an intermediate state) around −5.3 V (the switch-OFF voltage), and similar behavior could be observed during the high positive sweep (see Figure S11 of the Supporting Information). The device maintained well in the intermediate-state during the subsequent fifth scan (from 0 to −6.0 V) or the reverse voltage sweep (the sixth sweep), fulfilling the characteristic of a WORM device. Overall, the memory device ITO/P(VCz)29(MIDO3)21/Al could be switched from a lowconductivity (OFF) state to a high-conductivity (ON) state under a relatively low bias, but once the bias reached a higher 25548

dx.doi.org/10.1021/jp307156c | J. Phys. Chem. C 2012, 116, 25546−25551

The Journal of Physical Chemistry C

Article

respectively. The CV results of the five random copolymers with different composition were presented in Table 2 (curves shown in Figure S9 of the Supporting Information). As can be seen from the obtained data, the random copolymers show both p-doping and n-doping behaviors during the positive and negative voltage scans. Furthermore, with the increase in the MIDO3 content (or the decrease in the VCz content) within the polymer, the half-peak potentials for both oxidation and reduction become higher and the calculated HOMO and LUMO energy levels simultaneously decrease. To gain insights into the switching mechanism for the memory devices, an energy level diagram for the ITO/ P(VCz)x(MIDO3)y/Al devices is summarized in Figure 5. In the case of copolymer P(VCz)29(MIDO3)21, the energy barrier for electron injection from Al cathode to the LUMO of the polymer is only 0.77 eV, indicating that electron injection is a favored process under the negative bias. The lateral chains of the copolymers comprise the carbazole and the 4-nitroazobenzene entities, which are known as electron donors (i.e., hole transport media)26 and acceptors (i.e., electron transport media),12 respectively. In such donor−acceptor−based polymers, both inter- and intrachain charge transfer processes can easily occur when applying an external field.22b This charge transfer process can readily generate an excited state, which could be responsible for the transition from the OFF- to ONstate. Here, the energy barrier for donor−acceptor charge transfer is 2.32 eV (i.e., the energy difference between the HOMO and LUMO energy levels of P(VCz)29(MIDO3)21), which is much higher than the charge injection barrier (i.e., 0.77 eV), and therefore the switching from OFF- to ON-state is mainly dominated by the charge transfer process. Table 2 shows the HOMO and LUMO energy levels, Eg and the switchON voltages of the random copolymers. Although both the HOMO and LUMO energy levels decrease with the increase in the MIDO3 content, their energy differences still maintain at a relatively stable level (i.e., 2.32−2.38 eV). Hence the switchON voltages of the copolymers maintain in the range of −1.70 to −1.90 V, and no clear trend was observed with increase in the MIDO3 content (Figure 6). It has been reported that aluminum electrodes tend to penetrate into the polymer films during the vacuum deposition process, interacting with strongly coordinating heteroatom, nitrogen or sulfur, in the conjugated polymers to form conductive metal filaments.25,27 As for our designed copolymers P(VCz)x(MIDO3)y, large quantities of nitrogen atoms exist in MIDO3 entities of the polymer with conjugated lateral chains. Therefore, it is very likely to form aluminum filaments when large numbers of electrons are injected into the film at the highconductivity (ON) state. During the voltage sweep, the filament current steadily increased with increase in the bias. However, the maximum current a filament able to conduct is limited. Excessive current generated too much Joule heat, which would rupture the originally formed metal filaments, exhibiting an abrupt decrease in the current level (i.e., the second switching from ON- to intermediate-state). The switch-OFF voltages for the ON-intermediate state transition decrease with the decrease in the MIDO3 content within the polymers, causing no second switching phenomenon for P(VCz)45(MIDO3)5 with the lowest nitrogen content (Figure 6), which supports our filamentary rupture mechanism for the second transition. Nevertheless, the carbazole moieties in the polymer still promised a relatively high conductivity due to the local regioregular arrangement formed under the electric field

Figure 3. Stability tests of the fabricated ITO/P(VCz)29(MIDO3)21/ Al device: (a) time response of the retention ability for the ON-, intermediate- and OFF-states with a constant reading voltage of −1 V; (b) Effect of read pulse of −1 V on the ON-, intermediate- and OFFstates. The inset shows the pulse shape employed.

Figure 4. (a) UV−vis optical absorption and (b) CV response of the P(VCz)29(MIDO3)21 film on an ITO-coated glass substrate. The scan rate was 1 mVs−1.

lowest unoccupied molecular orbital (LUMO) can be calculated from the CV data to be −5.84 and −3.52 eV, 25549

dx.doi.org/10.1021/jp307156c | J. Phys. Chem. C 2012, 116, 25546−25551

The Journal of Physical Chemistry C

Article

Table 2. Electrochemical Properties and the ON-Switching Voltages of the Studied Random Copolymers E/V (vs Ag/AgCl) polymer

EOx (onset)

ERed (onset)

HOMO (eV)a

LUMO (eV)a

Eg (eV)b

ON-Switching Voltage (V)

P(VCz)45(MIDO3)5 P(VCz)41(MIDO3)9 P(VCz)18(MIDO3)7 P(VCz)33(MIDO3)16 P(VCz)29(MIDO3)21

0.91 1.18 1.22 1.40 1.47

−1.46 −1.18 −1.16 −0.96 −0.85

−5.28 −5.55 −5.59 −5.77 −5.84

−2.91 −3.19 −3.21 −3.41 −3.52

2.37 2.36 2.38 2.36 2.32

−1.82 −1.70 −1.82 −1.74 −1.89

a The HOMO and LUMO energy levels were calculated from cyclic voltammetry with reference to ferrocene (0.43 eV). bThe data were calculated by the energy difference between HOMO and LUMO.

Figure 7. Typical I−V characteristics of the ITO/P(VCz)29(MIDO3)21/Al device under different compliance currents: the first sweep was performed to switch the device on; the device was switched off when compliance current achieved 0.05 A.

Figure 5. Energy level diagram for the ITO/Polymer/Metal devices: wherein, V45M5 stands for P(VCz)45(MIDO3)5. The rest may be deduced by analogy. The work functions of ITO and Al electrodes are 4.8 and 4.29 eV, respectively.

when it reached 0.01 A. When the compliance current was set at 0.05 A, however, an abrupt decrease in the current occurred at about −5.3 V, indicating the transition from the ON-state to the intermediate-state. The dependence of resistance on temperature in the ON state was also tested after the device was switched on (see Figure S10 of the Supporting Information). These results demonstrated that the formed filaments would not rupture unless the current exceeded a certain value and the ON−state resistance increased almost linearly with temperature, which is consistent with the features of filamentary conductivity.28 Moreover, the electric performance measurement was also carried out with Hg top electrode for P(VCz)29(MIDO3)21 copolymer (Figure 8) because Hg droplet owns a high surface tension, which makes it difficult to permeate into the polymer film to form conductive filaments. Accordingly, during the voltage sweep (0 to −6.0 V), only a switch-ON phenomenon was observed around −2.0 V, but no

Figure 6. Typical switch-ON and switch-OFF voltages for P(VCz)x(MIDO3)y copolymers with different MIDO3 content: no switch-OFF phenomenon was found for P(VCz)45(MIDO3)5.

and the excellent hole-transport ability of the carbazole moieties.24 Thus, these devices could only be reprogrammed to an intermediate-state instead of the pristine OFF-state. With the decrease in the weight percentage of MIDO3 (i.e., reduction of nitrogen content), fewer filaments were formed, hence exhibiting a lower switch-OFF voltage. To further confirm our assumption, the same voltage sweep was performed with various compliance currents on different cells in the device. Figure 7 shows the typical I−V characteristics of the tested cells in the ITO/P(VCz)29(MIDO3)21/Al device under different compliance currents. During the first sweep from 0 to −6.0 V (compliance current: 0.001 A), the device was switched on at a threshold voltage around −1.8 V. Then the current remained at 0.001 A, but no abrupt decrease in the current occurred under an enhanced electric field. Further the compliance current was raised, and no switch-OFF phenomenon was observed even

Figure 8. I−V characteristics of the ITO/P(VCz)29(MIDO3)21/Hg device: the device was found able to be switched on around −2 V but could not be switched off. 25550

dx.doi.org/10.1021/jp307156c | J. Phys. Chem. C 2012, 116, 25546−25551

The Journal of Physical Chemistry C

Article

(2) Fang, Y.-K.; Liu, C.-L.; Li, C.-X.; Lin, C.-J.; Mezzenga, R.; Chen, W.-C. Adv. Funct. Mater. 2010, 20, 3012−3024. (3) Raymo, F. M. Adv. Mater. 2002, 14, 401−414. (4) Forrest, S. R. Nature 2004, 428, 911−918. (5) Stikeman, A. Technol. Rev. 2002, 105, 31. (6) (a) Ouyang, J. Y.; Chu, C. W.; Szmanda, C. R.; Ma, L. P.; Yang, Y. Nat. Mater. 2004, 3, 918−922. (b) Collier, C. P.; Mattersteig, G.; Wong, E. W.; Luo, Y.; Beverly, K.; Sampaio, J.; Raymo, F. M.; Stoddart, J. F.; Heath, J. R. Science 2000, 289, 1172−1175. (7) Tang, W.; Shi, H. Z.; Xu, G.; Ong, B. S.; Popovic, Z. D.; Deng, J. C.; Zhao, J.; Rao, G. H. Adv. Mater. 2005, 17, 2307−2311. (8) Scott, J. C.; Bozano, L. D. Adv. Mater. 2007, 19, 1452−1463. (9) Choi, S.; Hong, S. H.; Cho, S. H.; Park, S.; Park, S. M.; Kim, O.; Ree, M. Adv. Mater. 2008, 20, 1766−1771. (10) Moller, S.; Perlov, C.; Jackson, W.; Taussig, C.; Forrest, S. R. Nature 2003, 426, 166−169. (11) (a) Liu, G.; Ling, Q. D.; Teo, E. Y. H.; Zhu, C. X.; Chan, D. S. H.; Neoh, K. G.; Kang, E. T. ACS Nano 2009, 3, 1929−1937. (b) Li, H.; Li, N. J.; Gu, H. W.; Xu, Q. F.; Yan, F.; Lu, J. M.; Xia, X. W.; Ge, J. F.; Wang, L. H. J. Phys. Chem. C 2010, 114, 6117−6122. (12) Zhuang, X. D.; Chen, Y.; Liu, G.; Zhang, B.; Neoh, K. G.; Kang, E. T.; Zhu, C. X.; Li, Y. X.; Niu, L. J. Adv. Funct. Mater. 2010, 20, 2916−2922. (13) Liu, J. Q.; Lin, Z. Q.; Liu, T. J.; Yin, Z. Y.; Zhou, X. Z.; Chen, S. F.; Xie, L. H.; Boey, F.; Zhang, H.; Huang, W. Small 2010, 6, 1536− 1542. (14) (a) Liu, J. Q.; Yin, Z. Y.; Cao, X. H.; Zhao, F.; Lin, A. P.; Xie, L. H.; Fan, Q. L.; Boey, F.; Zhang, H.; Huang, W. ACS Nano 2010, 4, 3987−3992. (b) Zhuang, X. D.; Chen, Y.; Li, B. X.; Ma, D. G.; Zhang, B.; Li, Y. Chem. Mater. 2010, 22, 4455−4461. (15) Chu, C. W.; Quyang, J.; Tseng, J. H.; Yang, Y. Adv. Mater. 2005, 17, 1440−1443. (16) Baek, S.; Lee, D.; Kim, J.; Hong, S. H.; Kim, O.; Ree, M. Adv. Funct. Mater. 2007, 17, 2637−2644. (17) Xie, L. H.; Ling, Q. D.; Hou, X. Y.; Huang, W. J. Am. Chem. Soc. 2008, 130, 2120−2121. (18) (a) Zhuang, X. D.; Chen, Y.; Liu, G.; Li, P. P.; Zhu, C. X.; Kang, E. T.; Neoh, K. G.; Zhang, B.; Zhu, J. H.; Li, Y. X. Adv. Mater. 2010, 22, 1731−1735. (b) Liu, J. Q.; Zeng, Z. Y.; Cao, X. H.; Lu, G.; Wang, L.-H.; Fan, Q. L.; Huang, W.; Zhang, H. Small 2012, DOI: 10.1002/ smll.201200999. (19) (a) Ling, Q. D.; Chang, F. C.; Sang, Y.; Zhu, C. X.; Liaw, D. J.; Chan, D. S. H.; Kang, E. T.; Neoh, K. G. J. Am. Chem. Soc. 2006, 128, 8732−8733. (b) Li, H.; Li, N. J.; Sun, R.; Gu, H. W.; Ge, J. F.; Lu, J. M.; Xu, Q. F.; Xia, X. W.; Wang, L. H. J. Phys. Chem. C 2011, 115, 8288−8294. (20) Kuorosawa, T.; Chueh, C. C.; Liu, C. L.; Higashihara, T.; Ueda, M.; Chen, W. C. Macromolecules 2010, 43, 1236−1244. (21) Ma, L.; Pyo, S.; Ouyang, J.; Xu, Q.; Yang, Y. Appl. Phys. Lett. 2003, 82, 1419−1421. (22) (a) Quyang, J.; Chu, C. W.; Szmanda, C. R.; Ma, L.; Yang, Y. NanoLett 2005, 5, 1077−1080. (b) Fang, Y.-K.; Liu, C.-L.; Yang, G.-Y.; Chen, P.-C.; Chen, W.-C. Macromolecules 2011, 44, 2604−2612. (23) Choi, T.-L.; Lee, K.-H.; Joo, W.-J.; Lee, S.; Lee, T.-W.; Chae, M. Y. J. Am. Chem. Soc. 2007, 129, 9842−9843. (24) Lim, S. L.; Ling, Q. D.; Teo, E. Y. H.; Zhu, C. X.; Chan, D. S. H.; Kang, E. T.; Neoh, K. G. Chem. Mater. 2007, 19, 5148−5157. (25) Lei, B.; Kwan, W. L.; Shao, Y.; Yang, Y. Org. Electron. 2009, 10, 1048−1053. (26) Fulghum, T. M.; Taranekar, P.; Advincula, R. C. Macromolecules 2008, 41, 5681−5687. (27) Joo, W.-J.; Choi, T.-L.; Lee, Jaeho; Lee, S. K.; Jung, M.-S.; Kim, N.; Kim, J. M. J. Phys. Chem. B 2006, 110, 23812−23816. (28) (a) Hahm, S. G.; Choi, S.; Hong, S.-H.; Lee, T. J.; Park, S.; Kim, D. M.; Kwon, W.-S.; Kim, K.; Kim, O.; Ree, M. Adv. Funct. Mater. 2008, 18, 3276−3282. (b) Hu, W.; Qin, N.; Wu, G. H.; Lin, Y. T.; Li, S. W.; Bao, D. H. J. Am. Chem. Soc. 2012, 134, 14658−14661.

switch-OFF phenomenon was seen. This result also evidently indicates that the first transition to the ON-state (i.e., the switch-ON process) arises from the field-induced charge transfer between the donor/acceptor entities and is not dominated by the penetrated nanoparticles during the thermal evaporation process, since the penetration of metal electrode can be ruled out by a cold-deposited Hg droplet.

4. CONCLUSIONS Novel donor−acceptor random copolymers of P(VCz)x(MIDO3)y have been successfully synthesized. The electrical memory performance of the random copolymers with different compositions was systematically investigated on the devices with the sandwich configuration of ITO/Polymer/ Metal. The memory behavior of the devices is controllable from single switching (bistable states) to double switching (tristable states) by the composition of the copolymers. The first switchON threshold voltages for the OFF-ON transition are almost constant (i.e., around −1.80 V), but the switch-OFF voltages decrease with the MIDO3 content. The experimental results suggest that the transition from the OFF- to ON-state is mainly attributed to the intramolecular and intermolecular charge transfer between the donor/acceptor moieties and that from the ON- to intermediate-state under a higher bias is due to the rupture of the conductive filaments. The fabricated memory devices exhibit a long retention time of 104 s and are insensitive to one hundred million (108) read cycles, which may render good stability for practical applications. Our approach of achieving devices with multistable states through the combination of different switching mechanisms in one device may provide a strategy for the design of nonbinary electrical memory materials for future ternary permanent data storage.



ASSOCIATED CONTENT

S Supporting Information *

Detailed preparation procedure and the characterization of the monomers and copolymers, including optical absorption, thermogravimetric analysis results, and cyclic voltammograms. This information is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was financially supported by the Chinese Natural Science Foundation (21076134, 21176164 and 21206102), NSF of Jiangsu Province (BK2010208), a project of Jiangsu Education Department (12KJB430011), Suzhou Nanoproject (ZXG2012023), and Project supported by the Specialized Research Fund for the Doctoral Program of Higher Education of China (Grant No. 20113201130003). A Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).



REFERENCES

(1) Li, H.; Xu, Q. F.; Li, N. J.; Sun, R.; Ge, J. F.; Lu, J. M.; Gu, H. W.; Yan, F. J. Am. Chem. Soc. 2010, 132, 5542−5543. 25551

dx.doi.org/10.1021/jp307156c | J. Phys. Chem. C 2012, 116, 25546−25551