Novel Electrical Properties of Nanoscale Thin ... - Semantic Scholar

applying a forward or a reverse voltage pulse.1-3 This charac- teristic has mostly been studied for devices in bulk ... To whom correspondence should ...
0 downloads 9 Views 306KB Size
Download PDF
9024

Langmuir 2007, 23, 9024-9030

Novel Electrical Properties of Nanoscale Thin Films of a Semiconducting Polymer: Quantitative Current-Sensing AFM Analysis Jiyoun Kim,†,‡ Shinhyo Cho,†,‡ Seungchel Choi,†,‡ Sungsik Baek,‡ Dongjin Lee,§ Ohyun Kim,*,§ Su-Moon Park,*,‡ and Moonhor Ree*,‡ Department of Chemistry, Department of Electronic & Electrical Engineering, National Research Lab for Polymer Synthesis & Physics, Center for Integrated Molecular Systems, and BK21 Program, Pohang UniVersity of Science and Technology (Postech), Pohang 790-784, Republic of Korea ReceiVed March 16, 2007. In Final Form: June 4, 2007 Thin films (20-150 nm thickness) of poly(o-anthranilic acid) with various doping levels were prepared on silicon substrates with deposited indium tin oxide, and their topography and current-voltage (I-V) characteristics were quantitatively investigated using current-sensing atomic force microscopy with a platinum-coated tip. The films were found to have a surface morphology like that of orange peel, with a periodic modulation and a surface roughness. The films exhibited nonuniform current flows and I-V characteristics that depended on the doping level as well as on the film thickness. Films with a high doping level were found to exhibit Zener diode switching behavior, whereas the films with a very low doping level (or that were dedoped) exhibited no current flow at all, and so are insulators. Interestingly, self-doped films (which are at an intermediate doping level) were found to have a novel electrical bistability, i.e., a switching characteristic like that of Schottky diodes, and increasingly insulating characteristics as the film thickness was increased. The films with thickness e62 nm, which exhibited this novel and stable electrical bistability, can potentially be used in the fabrication of high-density, stable, high-performance digital nonvolatile memory devices based only on transistors. The measured current images and I-V characteristics indicate that the electrical switching and bistability of the films are governed by local filament formation and charge traps.

Introduction Voltage-controlled electrical switching in organic and polymer molecules has attracted great attention due to its potential applications in electronic devices.1-3 The bistable phenomena of active organic and polymer molecules are of renewed interest because of their applications in switching and memory devices, in which reversible switching between two distinct conducting states, the “on-state” that is suitable for electron transfer and the “off-state” in which electrons are trapped, can be carried out by applying a forward or a reverse voltage pulse.1-3 This charac* To whom correspondence should be addressed: Phone: +82-54-2792120. Fax: +82-54-279-3399. E-mail: [email protected] (M.R.), smpark@ postech.edu (S.M.P.), [email protected] (O.K.). † J.K., Sh.C., and S.C. contributed equally to this study. ‡ Department of Chemistry, National Research Lab for Polymer Synthesis & Physics, Center for Integrated Molecular Systems, and BK21 Program. § Department of Electronic & Electrical Engineering and BK21 Program. (1) (a) Seiz, R.; Walter, A.; Engl, R.; Maltenberger, A.; Schumann, J.; Kund, M.; Dehm, C. IEDM Technol. Dig. 2003, 10.2.1-10.2.4. (b) Pinnow, C.-U.; Mikolajick, T. J. Electrochem. Soc. 2004, 151, K13-K19. (c) Yang, Y.; Ouyang, J.; Ma, L.; Tseng, R. J.-H.; Chu, C.-W. AdV. Funct. Mater. 2006, 16, 1001-1014. (2) (a) Donhauser, Z. J.; Mantooth, B. A.; Kelly, K. F.; Bumm, L. A.; Monnell, J. D.; Stapleton, J. J.; Price, D. W., Jr.; Rawlett, A. M.; Allara, D. L.; Tour, J. M.; Weiss, P. S. Science 2001, 292, 2303-2307. (b) Tsujioka, T.; Kondo, H. Appl. Phys. Lett. 2003, 83, 937-939. (c) Bandyopadhyay, A.; Pal, A. J. Appl. Phys. Lett. 2003, 82, 1215-1217. (d) Bandyopadhyay, A.; Pal, A. J. AdV. Mater. 2003, 15, 1949-1952. (e) Bandyopadhyay, A.; Pal, A. J. J. Phys. Chem. B.2003, 107, 2531-2536. (f) Ma, L. P.; Liu, J.; Yang, Y. Appl. Phys. Lett. 2002, 80, 2997-2999. (g) Sakai, K.; Matsuda, H.; Kawada, H.; Eguchi, K.; Nakagiri, T. Appl. Phys. Lett. 1988, 53, 1274-1276. (h) Bozano, L. D.; Kean, B. W.; Deline, V. R.; Salem, J. R.; Scotta, J. C. Appl. Phys. Lett. 2004, 84, 607-609. (3) (a) Mo¨ller, S.; Perlov, C.; Jackson, W.; Taussig, C.; Forrest, S. R. Nature 2003, 426, 166-169. (b) Ling, Q.; Song, Y.; Ding, S. J.; Zhu, C.; Chan, D. S. H.; Kwong, D. L.; Kang, E.-T.; Neoh, K.-G. AdV. Mater. 2005, 17, 455-459. (c) Majumdar, H. S.; Bolognesi, A.; Pal, A. J. Synth. Met. 2004, 140, 203-206. (d) Ma, D.; Aguiar, M.; Freire, J. A.; Huemmelgen, I. A. AdV. Mater. 2000, 12, 1063-1066. (e) Lai, Y.-S.; Tu, C.-H.; Kwong, D.-L.; Chen, J. S. Appl. Phys. Lett. 2005, 87, 122101-122103. (f) Beinhoff, M.; Bozano, L. D.; Scott, J. C.; Carter, K. R. Macromolecules 2005, 38, 4147-4156. (g) Krieger, Ju. H.; Trubin, S. V.; Vaschenko, S. B.; Yudanov, N. F. Synth. Met. 2001, 122, 199-202. (h) Vorotyntsev, M. A.; Skompska, M.; Pousson, E.; Goux, J.; Moise, C. J. Electrochem. Soc. 2003, 552, 307-317. (i) Henisch, H. K. Thin Solid Films 1978, 51, 265-274.

teristic has mostly been studied for devices in bulk systems with metal-insulator-metal structures.2,3 Moreover, these devices have generally been fabricated with top and bottom electrodes and dimensions of several hundred micrometers or millimeters. However, the metal conductor electrodes used in electrical devices in the markets of interest are in fact much smaller than 100 µm; the possibility of controlled scaling of electrically active organic and polymer materials means that they can be used in the fabrication of advanced switching and memory devices. The nanometer-scaling and charging of such organic and polymer materials mean that they are in high demand, in particular for the development of ultrahigh density data storage devices.4-12 In the present study, we aimed to quantitatively analyze the electrical characteristics of thin nanoscale films of a π-conjugate polymer, poly(o-anthranilic acid) (PARA), using current-sensing atomic force microscopy (CS-AFM), which is a powerful tool to analyze nanoscopic topographies and electrical properties of a variety of samples under precisely controlled load force between the tip and the sample.13 A series of thin PARA films with thicknesses in the range 20-150 nm and various doping levels (4) Quyang, J. Y.; Chu, C. W.; Tseng, R. J. H.; Prakash, A.; Yang, Y. Proc. IEEE 2005, 93, 1287-1296. (5) Tseng, J.; Huang, J.; Quyang, J.; Kaner, R.; Yang, Y. Nano Lett. 2005, 5, 1077-1080. (6) Wu, H. M.; Song, Y. L.; Du, S. X.; Liu, H. W.; Gao, H. J.; Jiang, L.; Zhu, D. B. AdV. Mater. 2003, 15, 1925-1929. (7) Kazimierski, P.; Tyczkowski, J. J. Surf. Coat. Technol. 2003, 174, 770773. (8) Lee, D.-W.; Ono, T.; Abe, T.; Esashi, M. Microelectromech. Syst. 2002, 11, 215-221. (9) Lee, H. W.; Kim, Y. M.; Jeon, D. J.; Kim, E.; Kim, J.; Park, K. Opt. Mater. 2002, 21, 289-293. (10) Shin, H.; Hong, S.; Moon, J.; Jeon, J. U. Ultramicroscopy 2002, 91, 103-110. (11) Vettiger, P.; Brugger, J.; Despont, M.; Drechsler, U.; Diirig, U.; Hgberle, W.; Lutwyche, M.; Rothuizen, H.; Stutz, R.; Widmer, R.; Binnig, G. Microelectron. Eng. 1999, 46, 11-17. (12) Ma, L. P.; Pyo, S.; Ouyang, J.; Xu, Q. F.; Yang, Y. Appl. Phys. Lett. 2003, 82, 1419-1421.

10.1021/la700785h CCC: $37.00 © 2007 American Chemical Society Published on Web 07/13/2007

Nanoscale Thin Films of a Semiconducting Polymer

Langmuir, Vol. 23, No. 17, 2007 9025

was fabricated on indium tin oxide (ITO) deposited silicon substrates with thermal oxide layer. Our CS-AFM observations demonstrated that the nanometer-scale PARA thin films exhibit novel on-off switching and nonvolatile memory characteristics depending on the doping level and the film thickness, so they are suitable for use in advanced switching and nonvolatile memory devices. The current flows of the polymer films in on-states with switching and memory characteristics were found to be inhomogeneous rather than homogeneous throughout the PARA film layer. Moreover, the CS-AFM observations indicate that the electrical on-off switching and electrical bistability of the PARA thin film based devices are governed by local filament formation and charge traps. Experimental Section PARA was prepared via the chemical oxidative polymerization of o-anthranilic acid with an oxidant, ammonium persulfate, in an aqueous acidic medium (1.0 M HCl), i.e., with the same method as used for polyaniline derivatives.14 In the polymerization an oxidant/ monomer ratio of 0.7 was used. o-Anthranilic acid (50 g, 365 mmol) was dissolved in 200 mL of 1.0 M HCl aqueous solution and cooled to below 5.0 °C with an ice-water bath. A precooled 1.0 M HCl aqueous solution containing ammonium persulfate (83.2 g, 365 mmol) was added dropwise to the o-anthranilic acid monomer solution with stirring over a period of 1 h. The reaction mixture was further stirred for 48 h in the ice-water bath and then filtered through a Bu¨chner funnel. The resulting brown precipitate was washed with 0.1 M HCl aqueous solution until the filtrate became colorless. The washed precipitate (HCl-doped PARA) was collected and dried under vacuum at room temperature for 48 h. The HCl-doped PARA was converted to its base form free from HCl dopant and then further doped at three different levels as follows. The dried HCl-doped PARA was dissolved in 0.1 M NH4OH aqueous solution by stirring for 1 day and then divided into three parts. The first part of the dissolved PARA was again precipitated by neutralizing the solution to pH 7.0 with 0.01 M HCl solution. The precipitate was filtered, washed several times with deionized (DI) water, and dried under vacuum at room temperature for 2 days. The intrinsic viscosity of the polymer product (i.e., self-doped PARA polymer) was determined to be 0.11 in N-methyl-2-pyrrolidinone at 25.0 °C by using an Ubbelohde capillary viscometer. The second part of the PARA solution was also precipitated by adjusting the solution to pH 3.0 with 0.01 M HCl solution. The precipitate (HCl-doped PARA polymer) was filtered, washed several times with DI water, and dried under vacuum at room temperature for 2 days. The final part of the PARA solution was adjusted to pH 9.0 with 0.1 M NH4OH solution, concentrated, and dried. The dried part was again dissolved in acetonitrile and then filtered through a Bu¨chner funnel. The resulting solution was filtered through PTFE-membrane microfilters with a pore size of 0.45 µm. The filtered solution was again concentrated and finally dried under vacuum at room temperature for 2 days, giving dedoped PARA polymer. Homogeneous solutions (0.05-0.10 wt % solid content) of the synthesized individual PARA samples (Figure 1) were prepared by dissolving them in acetonitrile. The resulting solutions were filtered through PTFE-membrane microfilters with a pore size of 0.45 µm. Thin films with 20-150 nm thickness were prepared from the filtered solutions by spin-coating onto precleaned ITO-deposited silicon (13) (a) Kelley, T. W.; Granstrom, E. L.; Frisbie, C. D. AdV. Mater. 1999, 11, 261-264. (b) Beebe, J. M.; Engelkes, V. B.; Miller, L. L.; Frisbie, C. D. J. Am. Chem. Soc. 2002, 124, 11268-11269. (c) Nakamura, T.; Yasuda, S.; Miyamae, T.; Nozoye, H.; Kobayashi, N.; Kondoh, H.; Nakai, I.; Ohta, T.; Yoshimura, D.; Matsumoto, M. J. Am. Chem. Soc. 2002, 124, 12642-12643. (d) Gardner, C. E.; Macpherson, J. V. Anal. Chem. 2002, 74, 576A-584A. (e) Han, D. H.; Park, S.-M. J. Phys. Chem. B.2004, 108, 13921-13927. (f) Lee, H. J.; Park, S.-M. J. Phys. Chem. B.2004, 108, 1590-1595. (g) Park, S.-M.; Lee, H. J. Bull. Korean Chem. Soc. 2005, 26, 697-706. (h) Cho, S. H.; Park, S.-M. J. Phys. Chem. B.2006, 110, 25656-25664. (14) Baek, S.; Ree, J. J.; Ree, M. J. Polym. Sci.: Polym. Chem. 2002, 40, 983-994.

Figure 1. Chemical structure of PARA polymer synthesized in this study and its films at different levels. substrates with a thermal oxide layer, followed by drying on a hot plate at 110 °C in a vacuum for 1 h. The resulting films were determined to have a thickness of 20-150 nm by using a spectroscopic ellipsometer (model VASE, Woollam) and a surface profiler (Veeco Instrument Dektak3). CS-AFM measurements were conducted in ambient air using a contact mode current-sensing atomic force microscope (model PicoSPM, Molecular Imaging) equipped with a Nanosensor platinum-coated cantilever (with a 0.12 N/m spring constant). The microscope was set to measure currents up to 1.0 µA by the manufacturer. The load force was maintained below 10 nN to avoid damage to the cantilever tip and the polymer film. In these measurements, the bias voltage between the substrate (ITO electrode) and the conducting platinum (Pt) cantilever (which was grounded) was scanned in a forward or backward mode to obtain the CS-AFM images. In addition, I-V curves were recorded by performing forward and reverse voltage scans between -3.0 V and +3.0 V at a scan rate of 100 mV/s. Topographical AFM images were obtained concurrently at a bias voltage between 1.0 and 3.0 V. Further, the electrical dc conductivities of thick films (70-75 µm) of the polymer samples were measured using a Signatone four-point probe station with tungsten carbide electrodes connected to a Keithley constant-current source voltammeter; each conductivity was determined by measurement of the voltage drop across the two inner probes while an appropriate constant current of 1-50 µA was applied to the two outer probes. In addition, cyclic voltametry (CV) was carried out in aqueous 0.1 M HCl solution using an electrochemical workstation (IM6ex impedance analyzer) with a platinum gauze counter electrode and an Ag/AgCl (3.8 M KCl) reference electrode. A scan rate of 100 mV/s was used. Ultraviolet-visible (UV-vis) spectra were measured at 0.8 cm-1 resolution using a Sinco UV-vis spectrophotometer.

Results and Discussion Using the CS-AFM technique, we examined in detail the PARA films at various doping levels (Figure 1). The topographic and current images we obtained are presented in Figure 2. Figure 2A shows the surface topography of a 63 nm thick PARA film, which was prepared at pH 3.0. The film surface has morphology like that of an orange peel with a periodicity of around 260 nm. The root-mean-square (rms) roughness is 5.2 nm over an area of 4 × 4 µm2. The surface morphology and roughness of the film are probably mainly due to the aggregation and molecular ordering of the polymer chains that occur during the spin-coating and subsequent drying processes. Further, the surface morphology might partly be due to the roughness (1.2 nm rms roughness) of the surface of the ITO-deposited substrate we employed. Similar orange-peel surface morphologies were found in the AFM images

9026 Langmuir, Vol. 23, No. 17, 2007

Figure 2. Topographic (A, B, and C) and current (a, b, and c) AFM images measured simultaneously for PARA films prepared from solutions at pH 3.0, 7.0, and 9.0: (A and a) 63 nm thick film prepared at pH 3.0, (B and b) 70 nm thick film prepared at pH 7.0, and (C and c) 70 nm thick film prepared at pH 9.0. In these measurements, platinum-coated cantilevers were used. A bias voltage of 1.0 V was used for the film prepared at pH 3.0 and a bias voltage of 3.0 V was used for the films prepared at pH 7.0 and 9.0.

of the other PARA films with different doping levels (Figure 2B,C). The periodicity of the orange-peel morphology appears around every 160 nm for a 70 nm thick film prepared at pH 7.0 and around 200 nm for a 70 nm thick film prepared at pH 9.0. The rms roughness over a surface area of 4 × 4 µm2 is 5.4 nm for the film prepared at pH 7.0 and 2.2 nm for the film prepared at pH 9.0. Figure 2a displays the current image of a PARA film prepared at pH 3.0. The current image clearly shows that the application of a voltage (1.0 V) induces current flow nonuniformly at a number of localized spots across the polymer film, which is between the platinum-coated cantilever tip and the ITO electrode layer. The observation of such locally conducting spots in the film was reproducible. The observation of inhomogeneous current flow is a clue to the formation of filaments at spots localized throughout the film at applied voltages above a critical or threshold voltage of the film. Similar current heterogeneities have been reported for electrochemically synthesized polyaniline, polypyrrole, and poly(3,4-ethylenedioxythiophene) films.15-17 The heterogeneities in the response current image might result from heterogeneities in the film morphology and the doping level. Localized spots (i.e., morphological heterogeneities) in the polymer film in which the aggregation and arrangement of the (15) (a) Hong, S. Y.; Park, S.-M. J. Phys. Chem. B.2005, 109, 9305-93120. (b) Wu, C. G.; Chang, S.-S. J. Phys. Chem. B. 2005, 109, 825-832. (16) Han, D.-H.; Kim, J.-W.; Park, S.-M. J. Phys. Chem. B. 2006, 110, 1487414880. (17) Han, D.-H.; Lee, H. J.; Park, S.-M. Electrochim. Acta 2005, 50, 30853092.

Kim et al.

polymer chains are more dense might exhibit better current responses to the applied voltages. Localized spots with a higher doping level might also produce better current responses. We suggest that these localized spots act as filaments that conduct when a voltage is applied. The average interdistance of the nonuniform current-flow spots in the film is much shorter than the periodicity (hill-to-hill or valley-to-valley distance) of the orange-peel morphology. This suggests that the morphological heterogeneities found in the response current image measurements are much smaller than those found in the surface topographic images. The PARA film prepared at pH 7.0 also exhibits current flow but only at one local spot in the 4 × 4 µm2 area (Figure 2b), which is quite different from the behavior of the film prepared at pH 3.0. The observation of such the locally conducting spot in the film was also found to be reproducible. Here it is noted that the observed current flow was obtained at an applied voltage of 3.0 V, which is higher than that (1.0 V) applied for the film prepared at pH 3.0; at 1.0 V, the film revealed no current flow characteristics. On the other hand, the PARA film prepared at pH 9.0 exhibits no current flow characteristics in any part of the film layer over an applied voltage up to 3.0 V (Figure 2c). As discussed above, the PARA films have nearly identical thicknesses (63-70 nm) and exhibit similar surface morphologies, although there are some scale differences in their surface modulations, roughnesses, and aggregations. In order to further investigate the conductivities of the films, we attempted to determine the conductivities of the local spots that exhibit current flow in the current images (Figure 2a,b). The conductivity σ⊥ (i.e., the conductivity across the film thickness) of the spot exhibiting the highest current flow in each current image was determined from the resistance R measured for the spot in the current measurement obtained with the Pt-coated cantilever tip from the relationship, σ⊥ ) (1/R)(l/A), where l and A are the film thickness and the area of the spot contacted by the probing cantilever tip, respectively.17-19 σ⊥ was found to be 2.86 S/cm for the PARA film prepared at pH 3.0 and 0.83 S/cm for the film prepared at pH 7.0. The conductivity measurements were extended to PARA films with 70-75 µm thickness by using a four-probe technique, and we obtained the bulk conductivity σbulk. The measured σbulk values of polymer films prepared at pH 3.0, 7.0, and 9.0 were 1.8 × 10-3, 8.8 × 10-8, and 9.5 × 10-9 S/cm, respectively. These conductivity results indicate that the PARA films prepared at pH 3.0 and 7.0 are semiconductors, and that the PARA film prepared at pH 9.0 is essentially an insulator. In general, the conductivity of a π-conjugated polymer film is known to depend on the doping level of the dopant of the film.14-16 In the case of a polyaniline film as a representative π-conjugated polymer, electrochemical analysis found that the electrochemical activity and conductivity become deteriorated as the pH value in the doping increases (i.e., the degree of protonation in the film decreases).15,16 Taking this fact and the conductivity data into account, we conclude that the differences in the observed current flow behaviors might be due to the differences in their doping levels, rather than the morphological differences. Furthermore, even for a given PARA film, the doping level is inhomogeneous throughout the film layer. The current image and conductivity data suggest that local spots with a higher doping level in a given PARA film produce higher current flows. We carried out CS-AFM measurements for PARA films of various thicknesses in the range 20-150 nm. Figure 3 presents (18) Irraelachvili, J. Intermolecular and Surface forces; Academic: London, 1992. (19) Lee, H. J.; Park, S.-M. J. Phys. Chem. B. 2005, 109, 13247-13254.

Nanoscale Thin Films of a Semiconducting Polymer

Figure 3. Surface topographic (A, B, and C) and current (a, b, and c) AFM images measured simultaneously for PARA films with various thicknesses prepared from a solution at pH 7.0: (A and a) 150 nm thick film, (B and b) 62 nm thick film, and (C and c) 20 nm thick film. In these measurements, platinum-coated cantilevers and a bias voltage of 3.0 V were used.

representative topographic and current images, which were obtained from films prepared at pH 7.0. As can be seen in Figure 3A-C, the films have similar surface topographies, regardless of the film thickness. These surface morphologies are comparable to that of the 70 nm thick film. However, the current images were found to be sensitive to the film thickness. No current flow was observed for films with thicknesses in the range 80-150 nm; a representative CS-AFM image is shown in Figure 3a. As discussed earlier, the 70 nm thick film exhibits only one current flow spot in an area of 4 × 4 µm2 (Figure 2b). When the film thickness was further reduced to 62 nm, current flow spots appear in large numbers even over an area of 1 × 1 µm2 (Figure 3b). The number of spots producing current flow is even higher for the 20 nm thick film (Figure 3c). These current image data indicate that the inhomogeneous current flow of the PARA film prepared at pH 7.0 is strongly dependent on the film thickness. Films thicker than 70 nm do not exhibit current flow, and so are probably insulators. The films that are 70 nm or thinner produce inhomogeneous current flows: the thinner the film, the stronger and denser the current flow. The PARA films of 20-150 nm thickness prepared at pH 3.0 all exhibit surface morphologies similar to that of the 63 nm thick film shown in Figure 2A. The current image measurements of these films indicate that they exhibit a high density of inhomogeneous current flows (data not shown), as was also observed for the 63 nm thick film (Figure 2a). Thinner films produce stronger and denser current flows. In contrast, there were no current flows evident in the current image measurements for the films of 20-150 nm thickness prepared at pH 9.0, regardless of the film thickness for the thickness range considered (data not shown). These results confirm that the PARA films prepared at pH 9.0 are insulators.

Langmuir, Vol. 23, No. 17, 2007 9027

Figure 4. Current-voltage curves measured at the conducting spots of PARA films: (a) 63 nm thick film prepared at pH 3.0, (b) 70 nm thick film prepared at pH 7.0, (c) 62 nm thick film prepared at pH 7.0. In these measurements, platinum-coated cantilevers were used.

We investigated the current-voltage (I-V) characteristics of the PARA films by applying voltages in the forward and backward scanning modes with a scanning rate of 100 mV/s. Figure 4a displays the I-V curve of a 63 nm thick PARA film prepared at pH 3.0, which was measured at a conducting spot in the film. In the forward scanning direction (here, “forward” is defined as a positive voltage on the ITO), a local spot of the film sandwiched between the ITO electrode and the Pt cantilever tip was found to exhibit a very low conductivity state, similar to that of an insulator, up to 1.29 V, and then an electrical transition was found at 1.30 V, with a current increase. The current of the conducting state of this device increases abruptly with further increases in the applied voltage. It was found that applying a positive forward bias to the device causes rectification; the device acts under forward bias as a diode. This behavior might be due to the charge injection that results from tunneling through Schottky-like barriers at the interfaces between the film and the electrodes.20 However, in the backward scanning direction (“backward” is defined as a negative voltage on the ITO), there is an abrupt increase in the current in the device at -0.92 V and below, which is similar to the electrical breakdowns found in Zener diodes. This behavior can be interpreted in terms of the tunneling of electrons into the conduction band and of holes into the valence band of the polymer.20 Such Zener diode behavior has been previously reported for devices based on poly(3octylthiophene)20 and poly(N-methylaniline) thin films21 as well as for devices based on poly(3-methylthiophene) and polypyrrole (20) Garten, F.; Schlatmann, A. R.; Gill, R. E.; Vrijmoeth, J.; Klapwijk, T. M.; Hadziioannou, G. Appl. Phys. Lett. 1995, 66, 2540-2542.

9028 Langmuir, Vol. 23, No. 17, 2007

Figure 5. (a) UV-vis absorption spectrum and (b) CV response of a PARA film prepared at pH 7.0. The CV measurement was carried out in aqueous 0.1 M HCl solution using a platinum gauze counter electrode and an Ag/AgCl (3.8 M KCl) reference electrode. A scan rate of 100 mV/s was used.

thin films with and without cadmium selenide nanoparticles.22 The threshold voltages of the spot were determined from the I-V curve to be -0.92 and 1.30 V. Note that the threshold voltages were found to vary from local spot to local spot, which had various conductivity levels. Here the difference in the absolute magnitudes of the determined threshold voltages is attributed to the dissimilarity of the used metal electrodes (i.e., ITO and Pt). From these threshold voltages, the band gap energy of the spot is estimated to be 2.22 eV, according to a method reported previously.16 The estimated band gap is close to that (2.71 eV) determined from the film by the UV-vis spectroscopy measurement (UV-vis spectrum not shown but very similar to that in Figure 5a). Further, the highest occupied molecular orbital (HOMO) level of the film was estimated to be 4.45 eV from the first oxidation potential (0.05 V) measured by CV analysis (CV data not shown but very similar to that in Figure 5b). I-V curve measurements for the PARA film were further carried out by scanning the applied voltage from -3.0 to +3.0 V, as well as by reverse scanning the applied voltage from +3.0 to -3.0 V; a scanning rate of 100 mV/s was employed. The applied voltage scans resulted in identical I-V curves, which are the same as that shown in Figure 4a. Similar I-V curves were also observed for films with other thicknesses (20-150 nm) (data not shown). I-V curve measurements were also carried out for the PARA films prepared at pH 7.0, and current flows were found. These films were determined to have a band gap of 2.96 eV and a HOMO level of 4.45 eV from UV-vis spectroscopy and CV analysis (see a representative UV-vis spectrum and CV data in Figure 5). Figure 4b shows the I-V curve of a 70 nm thick PARA film prepared at pH 7.0, which was measured at a conducting spot in the film. This I-V curve is different from that observed for the film prepared at pH 3.0. In the forward scanning direction, the device consisting of the local spot with ITO and (21) Abthagir, P. S.; Saraswathi, R. J. Mater. Sci.: Mater. Electron. 2004, 15, 81-86. (22) Cassagneau, T.; Mallouk, T. E.; Fendler, J. H. J. Am. Chem. Soc. 1998, 120, 7848-7859.

Kim et al.

Pt electrodes exhibits a very low conductivity state, like that of an insulator, up to 1.69 V, and then an electrical transition was found at 1.70 V, with a current increase. The current of the conducting state increases abruptly with further increases in the applied voltage. This rectifying behavior might also be due to charge injection that results from tunneling through the interfaces between the film and the electrodes over a Schottky-like barrier. In contrast, in the reverse scanning direction the spot remains in a very low conductivity insulator-like state and does not exhibit any Zener breakdown behavior, as was observed for the film prepared at pH 3.0. Similar I-V curves with rectifying properties have previously been reported for electrochemically prepared poly(3-methylthiophene) with the appropriate doping levels.23 Similar I-V curves were also observed for spots in the films with a thickness less than 70 nm. In addition, these PARA thin films were found to exhibit an electrical bistability, which could be used in the fabrication of digital nonvolatile memory devices. Such electrically bistable behavior was not observed for any of the PARA films prepared at pH 3.0. As can be seen in Figure 4c, a 62 nm thick PARA film prepared at pH 7.0 produces an I-V curve that is almost identical to those obtained in the forward and reverse applied voltage scanning directions (see Figure 4b) between -3.0 and +3.0 V. During the subsequent reverse scan from +3.0 to -3.0 V, the film was found to exhibit an I-V curve with a shape that is similar to that obtained by scanning from -3.0 to +3.0 V, but which shows a hysteresis. As can be seen in Figure 4c, the film-based device (ITO/PARA film/Pt tip) undergoes an electrical transition at 1.70 V with an abrupt current increase (i.e., the switch-on voltage is 1.70 V) during scanning of the applied voltage from -3.0 to +3.0 V. The high conductivity state (i.e., the on-state) is retained when the applied voltage is reversibly scanned to +3.0 V from -3.0 V. However, during this reverse scanning, the high conductivity state was found to suddenly return to the low conductivity state (i.e., the off-state) with an electrical transition between 0.56 and 0.08 V, and then the low conductivity state was retained below +0.08 V and for negative voltages (Figure 4c). After the device has returned to the low conductivity state, it can be switched back to the high conductivity state by again scanning the applied voltage to +3.0 V from -3.0 V. Such on-off switching behavior was also observed for films thinner than 62 nm. These unipolar switching behaviors demonstrate that PARA films of e62 nm thickness prepared at pH 7.0 can be used in nonvolatile memory devices. In particular, the unipolar switching characteristics of PARA films discovered in this study can enable the fabrication of simple structured nonvolatile memory devices based only on transistors, rather than on transistor and resistor pairs, as required in the manufacture of memory devices based on bipolar switching characteristics.1 Thus, high performance, high-density memory devices based on a multilayer structure, which are in high demand in the world market, can easily be fabricated with these PARA films. For the devices exhibiting excellent bistability, we performed further retention tests in ambient air. Figure 6 shows representative results of these retention tests. As can be seen in this figure, once the device (i.e., the local spot) is switched to the on-state by applying a positive voltage of 2.60 V, which is higher than the threshold voltage (i.e., the critical switch-on voltage, 1.70 V), the on-state is retained for 680 s or more. When the on-state is switched back to the off-state by applying a negative voltage, the off-state is also retained for 680 s or more. For the films prepared at pH 9.0, a band gap of 3.00 eV and a HOMO level of 4.44 eV were measured from UV-vis (23) Lee, H. J.; Park, S.-M. J. Phys. Chem. B. 2004, 108, 16365-16371.

Nanoscale Thin Films of a Semiconducting Polymer

Figure 6. Time response (i.e., retention time) of the on-state at the conducting spot of a 62 nm thick PARA film prepared at pH 7.0. A bias voltage of 2.6 V was used.

spectroscopy and CV analysis (UV-vis spectrum and CV data not shown but very similar to those in Figure 5). The band gap and HOMO level are very close to those of the films prepared at pH 3.0 and 7.0. However, unlike the PARA films prepared at pH 3.0 and 7.0, the film prepared at pH 9.0 exhibits no current flow when the applied voltage is scanned from -4.0 to 4.0 V, indicating that this PARA film is probably an insulator. As shown above, the PARA films showed that the band gap ranges 2.71-3.00 eV and the HOMO level varies in the range of 4.44-4.45 eV, depending on the doping levels. The measured CV response data suggest that the PARA films may undergo oxidation and reduction reaction during applying forward and backward bias voltages. Further, the oxidation and reduction reaction is accompanied with changes in the π-conjugation of the PARA polymer backbone, consequently causing conformational changes in the polymer molecule. Therefore, we may consider a redox mechanism as the switching mechanism for the current flows observed for the PARA films. However, all the PARA films prepared at pH 9.0 exhibit no electrical switching behavior, although their band gap and HOMO level are very close to those of the films prepared at pH 3.0 and 7.0. Further, in the case of the films (which were prepared at pH 3.0 and 7.0) revealing electrical current flows, their electrical on-off switching behaviors are strongly dependent on the film thickness. The electrical bistability observed for only the films prepared at pH 7.0 is also strongly dependent on the film thickness. These results suggest that the electrical switching behaviors of the PARA films in our study are far from a redox mechanism. If a redox mechanism is involved in the films, its contribution to the device switching is small. Taking into account the nonuniform current flows observed in the current images, the observed I-V characteristics suggest that the electrical on-off switching and bistability of the PARA thin films are governed by local filament formation and charge traps. As can be seen in Figure 1, the PARA polymer chain has one carboxylic acid unit and one amino linkage per chemical repeat unit. In the polymer chain, the carboxylic acid units can play as electrophilic sites, and the amino linkages can play as nucleophilic sites; in contrast, the proton-doped amino linkages can play as electrophilic sites, depending on the doping level. These carboxylic acid units and amino linkages may play as charge trapping sites, depending on their associations and doping levels. When bias voltage is applied to the film, these trap sites become filled with the injected electrons, and at the critical voltage (i.e., switch-on voltage) the trapped charges may flow through the trapped sites by a hopping process (i.e., through the filament formation), consequently causing current flows through the film between the ITO electrode and the Pt tip. When reverse bias voltage is applied to the film, the hole injection is enhanced by the trapped electrons, and thus a high current state is retained.

Langmuir, Vol. 23, No. 17, 2007 9029

With a further increase of the reverse bias voltage, the trapped electrons may be detrapped at the critical voltage (i.e., switchoff voltage), and consequently, the current flow will keep the low conductance state. The PARA films have relatively low conductivity and thus have a resistive barrier for the applied voltage. Namely, thicker film has larger resistive barrier. Thicker PARA film may need higher bias voltage to inject electrons into the trap sites. Further, the distribution of such trap sites across the film may depend on the film thickness. These factors may work together cooperatively with the thickness-dependent electrical on-off switching and bistability of the PARA thin films.

Conclusions The polymer poly(o-anthranilic acid) (PARA) was synthesized, and PARA films with various thicknesses were prepared with various doping levels by using aqueous hydrochloric acid and ammonium hydroxide on ITO-deposited silicon substrates. The surface topographies and I-V characteristics of the PARA films deposited on the ITO electrodes were investigated in detail using CS-AFM analysis. In addition, dc conductivity measurements were performed using the four-probe technique. All the PARA films were found to have surface morphologies similar to orange peel with a periodicity of 160-260 nm. The rms surface roughnesses were found to be in the range 2.2-5.4 nm. The surface periodicity and roughness were found to vary with the doping level. However, no correlations between the doping level and the periodicity and roughness of the film were found. The PARA films were found to have conductivities in the range from 1.8 × 10-3 to 9.5 × 10-9 S/cm, depending on the doping level; higher doping levels in the films were found to result in higher conductivities. These PARA films have interesting current flow and I-V characteristics that vary with the doping level and the film thickness as follows. First, it was demonstrated that inhomogeneous current flow could be induced in the PARA films at a number of localized spots across a polymer film sandwiched between the ITO electrode and Pt cantilever tip, depending on the doping level as well as the film thickness. The films with high and intermediate doping levels prepared at pH 3.0 and 7.0 were found to exhibit current flows, and both the density and the intensity of the current flows were found to increase with increases in the doping level and with decreases in the film thickness. In contrast, the films with a very low doping level (or that were dedoped) that were prepared at pH 9.0 were found to exhibit no current flows at all and so are like insulators. Second, the films with a high doping level prepared at pH 3.0 were found to exhibit a remarkable on-off switching behavior, like that of a Zener diode, for the thickness range 20-150 nm, which means that they are suitable for the fabrication of diodes with Zener breakdown or bipolar characteristics. Third, the PARA films with an intermediate doping level prepared at pH 7.0 were found to have electrical properties that are sensitive to film thickness. The films with g80 nm thickness were found to exhibit no current flows and so are probably insulators. In contrast, a 70 nm thick film was found to exhibit a remarkable on-off switching behavior, like that of a Schottky diode, and is thus suitable for the fabrication of diodes with a Schottky barrier. Moreover, the films of e62 nm thickness were found to exhibit very interesting nonvolatile memory characteristics based on a unipolar on-off switching behavior. This electrically bistable characteristic means that PARA films of e62 nm thickness can potentially be used in the fabrication of

9030 Langmuir, Vol. 23, No. 17, 2007

high-density, very stable, high-performance digital nonvolatile memory devices based only on transistors. Finally, the measured current images and I-V characteristics suggest that the electrical on-off switching and electrical bistability of the PARA thin films are governed by local filament formation and charge traps.

Kim et al.

Acknowledgment. This study was supported by the Korea Science & Engineering Foundation (National Research Lab Program and SRC Program), the Samsung Electronics Co., and the Ministry of Education (BK21 Program). LA700785H