23812
J. Phys. Chem. B 2006, 110, 23812-23816
Metal Filament Growth in Electrically Conductive Polymers for Nonvolatile Memory Application Won-Jae Joo,*,† Tae-Lim Choi,‡ Jaeho Lee,† Sang Kyun Lee,† Myung-Sup Jung,† Nakjoong Kim,§ and Jong Min Kim† Materials Center, Samsung AdVanced Institute of Technology, P. O. Box 111, Suwon 440-600, Korea, Electronic Chemical Materials DiVision, Cheil Industries Inc., 332-2 Gocheon-dong, Uiwang-si, Gyeonggi-do, Korea, and Department of Chemistry, Hanyang UniVersity, Seoul 133-791, Korea ReceiVed: August 3, 2006; In Final Form: September 20, 2006
Solution processable polymers that can reproducibly form metal filament by applying voltage are investigated for nonvolatile memory application. Up to present, the understanding of materials enabling to make the metal filament has not been well-documented and the vacuum deposition methods were dominantly used in device fabrication. After screening various polymers, we found that only the polymers having two functionalities, the presence of strongly coordinating heteroatom (S or N) with metal ions and the electrical conductivity, showed the reproducible filament formation behavior. Among the polymers screened, the regiorandom poly(3-hexylthiophene) showed the best switching endurance over 30 000 write-read-erase-read cycles without any switching failure.
1. Introduction As a new application of organic electronic devices, nonvolatile memory is attracting lots of attention due to the low fabrication cost and the high potential for a high-density memory via 3-D stacking structure.1-5 Extensive studies toward new material and device structure have been carried out to show good memory performance such as a large on-off ratio over 103, low operation voltage under 5 V, relatively long retention time of several months, and high endurance of 104, which are basic requirements for memory application.6-8 As examples of new materials and device systems, Rose Bengal,9 organic bilayer system containing metal interlayer,10 metal nanoparticle containing system,11 and Cu-TCNQ charge-transfer complex system12 were reported in the literature. However, the effort for finding new materials and systems is still desirable since there have been no good examples of the organic memory devices that satisfy all basic requirements. Especially achieving long retention time, thermal stability, and high endurance seems to be serious challenges in organic memory.3-12 We have focused on the development of organic memory systems based on the metal filament formation because they have unique potential in aspects of the retention time and the thermal stability.7 According to the previous papers, the metal filament formed within organic layer by applying voltage is very stable for several months at room temperature and for several hours at a high temperature of 110 °C, contrast to the organic memory systems using the charge trap or charge-transfer behavior which suffers from the poor thermal stability.7 Polystyrene (PS) film fabricated by a glow-discharge deposition technique had been most frequently used in the filament based organic memory device.13-16 However, the phenomenon * Corresponding author. Telephone: +82-31-280-6731. Fax: +82-31280-9349. E-mail:
[email protected]. † Samsung Advanced Institute of Technology. ‡ Cheil Industries Inc. § Hanyang University.
of metal filament formation is not intrinsic characteristics of PS, because spin-coated PS does not show memory behavior. Recently, Yang et al. reported that vacuum deposited 2-amino4,5-imidazoledicarbonitrile (AIDCN), tris-8-(hydroxyquinoline) aluminum (Alq3), and zinc 2,9,16,23-tetra-tert-butyl-29H,31Hphthalocyanine (ZnPc) showed metal filament based memory properties.7 But, up to the present, the most of the organic materials are fabricated by vacuum deposition techniques, such as thermal evaporation, and glow-discharge deposition technique and the understanding of organic materials that enables the metal filament formation is not well-studied. In this paper, we investigated the solution processable organic materials which can reproducibly form metal filament by applying voltage. To emphasize the advantages of low cost and large-area fabrication, wet processable materials are favored in organic electronic devices. Several polymers having different structure and physical properties would be screened and studied to find the mechanism of metal filament formation. Then, the switching endurance of the best polymer among the ones screened would be measured in order to evaluate the memory performance of the metal filament based organic memory. 2. Experimental Methods All the organic materials used in this work except poly(siloxane carbarzole) were commercially obtained from Aldrich and were used to prepare the test device without any purification. Poly(siloxane carbarzole) was synthesized according to the known procedure.17 Generally the test devices are fabricated with the metalpolymer-metal structure with aluminum and copper as bottom and top electrodes, respectively. Si wafer deposited with 200 nm thick silicon oxide layer was used as substrate after sonication cleaning in the acetone solution for 30 min. Aluminum bottom electrode (80 nm) was thermally evaporated onto the substrate followed by spin coating of the active organic layer with approximately 50 nm thickness. Regiorandom and regio-
10.1021/jp0649899 CCC: $33.50 © 2006 American Chemical Society Published on Web 11/09/2006
Metal Filament Based Polymer Memory
Figure 1. Typical current-voltage curve of regioregular P3HT device in which the sudden increase in current is attributed to the metal filament formation. Parts a-f indicate the sequences of the voltage sweep.
regular poly(3-hexylthiophene) (P3HT) was dissolved in chlorobenzene to make 1.3 wt % solution, leucoemeladine polyaniline (PA) in N-methylpyrrolidone to make 2.5 wt % solution, organic acids doped polypyrrole in water to make 1.5 wt % solution, poly(siloxane carbarzole) (PSX-Cz), poly(flourene) (PF), poly(styrene) (PS), poly(methyl methacrylate) (PMMA), poly(2-vinyl pyridine) (P2VP), poly(4-vinyl pyridine) (P4VP), poly(2-methoxy-5-(2′-ehtylhexyloxy)-1,4-phenylenevinylene) (MEH-PPV), and poly(phenylene vinylene)-disperse red 1 (PPVDR1) in chloroform to make 0.5 wt % solution. After spin casting, remaining solvent was removed by heating the device on a hot plate at the temperature a little below the boiling point of each solvent. Last, the copper was thermally evaporated with a deposition rate of 0.8 Å/s as a top electrode (80 nm) and the final test cell size was 0.25 mm2. The electric properties were measured in air with the programmable power supply of Yokogawa instrument (model 7631) and digital oscilloscope (model TDS3502B). Measurement in nitrogen did not give any significant difference in memory phenomenon. Arbitrary waveform generator (model Agilent 33220A) and voltage amplifier (model BOP 100) were used to make write-read-erase-read pulses. For measurement of device resistance at low temperature, cryostat (model 125, Austin Scientific) and temperature controller (model 331, Lakeshore) were used to control the temperature of sample. Resistance was measured with a source meter (model 2400, Keithley). The copper top electrode was connected to signal line, and the bottom electrode was used as ground. 3. Results and Discussion From a metal-polymer-metal device using P3HT, we observed the sudden change in conductivity under positive and negative voltage sweep (Figure 1). Aluminum and copper were thermally deposited as bottom and top electrodes, respectively, and regioregular P3HT was spin-coated with a thickness of 50 nm. During the positive voltage sweep, the current increased with voltage due to the intrinsic conductivity of P3HT. On the other hand, a sharp increase in the current to about 30 mA was observed at -2 V, which corresponds to the resistance of 20 Ω, exclusive of the wire resistance. The higher conductive state was returned to the lower state at higher voltage. This phenomenon occurred when highly diffusive copper was used
J. Phys. Chem. B, Vol. 110, No. 47, 2006 23813
Figure 2. Resistance of higher conductive (set) state as a function of temperature. Open circles are data measured, and the solid line shows the fitting result from the equation.
as top electrode. Since such a high conductivity cannot be coming from semiconducting P3HT, we consider this sudden increase in conductivity of the device is attributed to the formation of copper filament. To prove out the metallic conductance, the temperature dependence of resistance was measured in the higher conductive state switched at -2 V using a cryostat system. Figure 2 shows the resistance as a function of temperature from 160 to 300 K. The resistance increases linearly with the temperature, which is well-known typical behavior of metal conductor having free electrons. In the temperature region between 100 and 1000 K, the resistance of copper can be represented by R ) Ro[1 + R(T - To)], where R is the temperature coefficient, Ro is the resistance measured at reference temperature To, usually 293 K.18 We obtained the temperature coefficient of 0.0028 by fitting result, which differs a little from the value of 0.0039 known in the literature18 which is due to the thermal loss occurred at the contact between the sample and the cooling rod of cryostat. If we make assumptions that the metal filament is a straight rod connecting two electrodes, and that only a single filament is formed each time, the resistance of the filament can be estimated from the equation; R ) FL/A where F, L, and A are the resistivity, the length, and the cross-section area of an object, respectively. In our case, considering the resistivity of copper (1.72 × 10-8 Ω m) and the length (50 nm), the diameter of the filament is calculated to be 7.4 nm from the cell resistance of 20 Ω. Although the validity of our assumption may be in question, one can imagine nanometer-sized metal bridge connects between the two electrodes. As regioregular P3HT is the first good example of solution processable material for metal filament based organic memory, it is worthwhile to investigate in detail what the crucial factors are that lead to reproducible metal filament formation. Immediately two distinguishable structural points are noticed in regioregular P3HT, conjugated polymer backbone allowing some current to flow and a presence of strongly coordinating heteroatom, sulfur, at every repeat unit of the polymer backbone. From these two factors, we decided to screen various polymers to figure out what functionalities are required for the metal filament formation. The chosen polymers are listed in Figure 3.
23814 J. Phys. Chem. B, Vol. 110, No. 47, 2006
Joo et al.
Figure 3. Chemical structures of various polymers screened. Polymers in the box showed reproducible memory behavior.
TABLE 1: Switching Properties of Various Polymers Screened mobilitya (cm2/(V s)) P3HT regioregular regiorandom polyaniline PPV-DR1 PSX-Cz polypyrrole MEH-PPV polyfluorene P4VP P2VP polyvinylpyrrolidon PS PMMA
10-3
10-6 10-5 10-5 10-5 10-1 10-5 10-3
