Ink-Jet Printing of Cu−Ag-Based Highly Conductive Tracks on a

Langmuir 2009, 25, 429-433

429

Ink-Jet Printing of Cu-Ag-Based Highly Conductive Tracks on a Transparent Substrate Kyoohee Woo,† Dongjo Kim,† Jang Sub Kim,‡ Soonkwon Lim,‡ and Jooho Moon*,† Department of Materials Science and Engineering, Yonsei UniVersity, Seoul 120-749, Korea and LCD R&D Center, Samsung Electronics Co., Ltd., Gyeonggi-Do 449-711, Korea ReceiVed July 9, 2008. ReVised Manuscript ReceiVed October 22, 2008 We have developed a Cu-Ag-based mixed metal conductive ink from which highly conductive tracks form on a flexible substrate after annealing at low temperature. Addition of small Ag particles significantly improves the particle packing density by filling the interstices formed between the larger Cu particles, which in turn facilitates better conductivity compared to pure Cu metal film. The particle size and volume ratio of the Ag particles added should be carefully controlled to achieve maximum packing density in the bimodal particle system, which is consistent with the theoretical considerations of the Furnas model. In addition, we demonstrate direct writing of complex patterns that exhibit high conductivity upon annealing at sufficiently low temperature (175-210 °C) to not damage the transparent plastic substrate such as polyethersulphone (PES).

Introduction Recent years have seen growing attention on printed electronics by means of ink-jet printing technology. This is because ink-jet printing has significant advantages over conventional photolithographic technique, i.e., reduction in time, cost, and waste produced in manufacture.1-7 In contrast to vacuum-based methods, it is environmentally friendly, simpler, less expensive, and highly productive when it comes to the production of printed electronics. As printable conducting inks, several materials have been studied such as doped conjugated polymers or metal-organic complexes,8,9 but these materials are characterized by low electrical conductivity and poor electrical and thermal stabilities.10-12 By contrast, metals that possess high conductivity and operational stability can be applied in the form of nanoparticles.13,14 Most of relevant recent studies have focused on novel metals such as gold and silver nanoparticles,6,7,15 but the high cost of these metals has discouraged their use. Copper is a good candidate material because it is highly conductive but significantly cheaper than Au and Ag. Copper * To whom correspondence should be addressed. E-mail: jmoon@ yonsei.ac.kr. † Yonsei University. ‡ Samsung Electronics Co., Ltd.

(1) Lee, H.; Chou, K.; Huang, K. Nanotechnology 2005, 16, 2436. (2) Gans, B.-J.; Duineveld, P. C.; Schubert, U. S. AdV. Mater. 2004, 16, 203. (3) Kim, D.; Jeong, S.; Park, B.; Moon, J. Appl. Phys. Lett. 2006, 89, 264101. (4) Huang, D.; Liao, F.; Molesa, S.; Redinger, D.; Subramanian, D. J. Electrochem. Soc. 2003, 150, G412. (5) Park, B.; Kim, D.; Jeong, S.; Moon, J. Thin Solid Films 2007, 515, 7706. (6) Perelaer, J.; Gans, B.-J.; Schubert, U. S. AdV. Mater. 2006, 18, 2101. (7) Osch, Thijs, H. J.; Perelaer, J.; Laat, Antonius, W. M.; Schubert, U. S AdV. Mater. 2008, 20, 343. (8) Liu, Z.; Su, Y.; Varahramyan, K. Thin Solid Films 2005, 478, 275. (9) Sele, C. W.; Werne, T.; Friend, R. H.; Sirringhaus, H. AdV. Mater. 2005, 17, 997. (10) Gelinck, G. H.; Genus, T. C. T.; de Leeuw, D. M Appl. Phys. Lett. 2000, 77, 1487. (11) Drury, C. J.; Mutsaers, C. M. J.; Hart, C. M.; Mater, M.; de Leeuw, D. M Appl. Phys. Lett. 1998, 73, 108. (12) Sirringhaus, H.; Kawasem, T.; Friend, R. H.; Shimoda, T.; Inbasekaran, M.; Wu, W.; Woo, E. P. Science 2000, 290, 2123. (13) Fuller, S. B.; Wilhelm, E. J.; Jacobson, J. M. J. Microelectromech. Syst. 2002, 11, 54. (14) Bieri, N. R.; Chung, J.; Poulikakos, D.; Grigoropoulos, C. P. Superlattices Microstruct. 2004, 35, 437. (15) Noguchi, Y.; Sekitani, T.; Yokota, T.; Someya, T. Appl. Phys. Lett. 2008, 93, 043303.

Figure 1. SEM image of (a) copper nanoparticles with a mean size of 65 ( 3 nm and (b) silver nanoparticles with a mean size of 21 ( 2 nm. (c) SEM images showing the cross-section of the printed films as a function of the Cu/Ag mixing ratio. The films were annealed at 200 °C for 90 min under vacuum (10-3 Torr). The annealed films were coated with tungsten of thickness ∼10 nm to protect the conductive films from the damage by FIB-assisted cross-sectioning. (d) The binary image showing the metal and pore structures as a function of the Cu/Ag mixing ratio. The average porosities calculated from the image analysis are shown.

nanoparticles synthesized in ambient atmospheric temperature and pressure inevitably have surface oxide layers because the Cu oxide phases are thermodynamically more stable than pure Cu. We have recently demonstrated that the use of Cu nanoparticles with minimal surface oxidation improves the electrical properties of ink-jet printed Cu conductive tracks.16 The presence of surface oxide increases the annealing temperature due to its higher melting (16) Jeong, S.; Woo, K.; Kim, D.; Lim, S.; Kim, J.; Shin, H.; Xia, Y.; Moon, J. AdV. Funct. Mater. 2008, 18, 679.

10.1021/la802182y CCC: $40.75  2009 American Chemical Society Published on Web 11/26/2008

430 Langmuir, Vol. 25, No. 1, 2009

Figure 2. The comparison between the calculated packing density based on the Furnas model and the experimentally measured annealed density for the films prepared with various mixing ratios of Cu/Ag.

point (Tm) and reduces the electrical conductivity. For example, bulk CuO has a higher melting point (Tm ) 1330 °C) and much higher resistivity (5.1 × 107 µΩ · cm) than copper (Tm ) 1083 °C, resistivity ) 1.72 µΩ · cm). Cu nanoparticles with minimal surface oxidation require annealing at 250 °C to produce conductive tracks with reasonably low resistivity (∼50 µΩ · cm). Only polyimide (PI), among the various flexible substrates, withstands this annealing temperature, limiting the use of less thermally stable plastic substrates with better transparency and lower costs. In this paper, we demonstrate conductive patterns with a reasonably low resistivity (38.6 µΩ · cm), which were annealed at 175 °C on a transparent flexible substrate. The film conductivity improves as a result of enhanced particle packing density when small silver nanoparticles are mixed in that effectively fill the interparticle pores between Cu nanoparticles. It is well-known that the packing density of micrometer-sized metal granular films can be increased by adding smaller particles, which improve the film’s electrical conductivity.17-20 To the best of our knowledge, however, there are no reports on conductivity enhancement of nanometer-sized mixed metal granular films. It should be noted here that silver nanoparticles were selected because they are easily synthesized with good size controllability. Any metal nanoparticles that meet physical dimension requirements as well as have high conductivity and low melting point would be usable. The microstructure and conductivity evolution of granular films as a function of the mixing ratio between Cu and Ag were investigated. We fabricated complex conductive tracks on a transparent substrate by ink-jet printing.

Experimental Section 1. Metal Nanoparticle Synthesis. Copper nanoparticles of mean size 65 ( 3 nm were synthesized in ambient atmosphere by the polyol method. Details of the synthesis procedure have previously been described.21 Poly(N-vinylpyrrolidone) (Mw ) 40 000 g/mol, Sigma-Aldrich), acting as a capping molecule to hinder the formation of an oxide layer and coagulation of particles, was dissolved in diethyleneglycol (DEG, 99%, Sigma-Aldrich). Sodium phosphinate monohydrate (NaH2 · PO2 · H2O, Junsei), used as a reducing agent, was added to the DEG solution, and the solution was heated to a reaction temperature of 140 °C. An aqueous solution of copper (II) sulfate pentahydrate (98%, Sigma-Aldrich) was then injected into (17) Sumirat, I.; Ando, Y.; Shimamura, S. J. Porous Mater. 2006, 13, 439. (18) Solonin, S. M.; Chernyshev, L. I. Powder Metall. Met. Ceram. 2006, 45, 226. (19) Grandjean, S.; Absi, J.; Smith, D. S. J. Eur. Ceram. Soc. 2006, 26, 2669. (20) Montes, J. M.; Rodriguez, J. A.; Herrera, E. J. Powder Metall. 2003, 46, 251. (21) Park, B.; Jeong, S.; Kim, D.; Lee, S.; Moon, J. J. Colloid Interface Sci. 2007, 311, 417.

Woo et al.

Figure 3. The resistivity variation as a function of annealing temperature for the mixed metal films with various Cu/Ag mixing ratios. The inset shows the resistivity of the Cu/Ag ) 3:1 film prepared with larger silver nanoparticles (50 nm).

the hot reaction medium via a syringe pump at a rate of ∼2 mL/min. After 1 h of reaction, the solution was cooled to room temperature, and the particles were separated by centrifugation and then washed with methanol. Silver nanoparticles with two different mean sizes of 21 ( 2 nm and 50 ( 3 nm were also synthesized in a similar manner.22 The mean particle size and the standard deviation of the synthesized particles were determined from image analysis of scanning electron microscopy (SEM) micrographs of each particle. 2. Preparation of Inks and Metal Particulate Films. The synthesized copper and silver nanoparticles were mixed at varying volume ratios from Cu/Ag ) 2:1 to 4:1. The metal particles were dispersed in a mixed solvent of methanol, 2-methoxy ethanol, and ethylene glycol. The solid loading of all inks was 20 wt %. The formulated inks were milled by a planetary milling machine for 60 min, followed by filtration through a 5 µm nylon mesh. The prepared conductive inks were printed with an ink-jet printer on a polyethersulphone (PES; thickness ∼ 0.125 mm) substrate. The printer setup consisted of a drop-on-demand piezoelectric ink-jet nozzle manufactured by Microfab Technologies, Inc. (Plano, TX) with a nozzle diameter of 30 µm. Uniform ejection of the droplets was achieved by applying a 35 V pulse lasting 10 µs at a frequency of 1000 Hz. The diameter and velocity of the ejected droplets were about 36 µm and 3 m/s, respectively. The ink-jet printed films were annealed at various temperatures ranging from 150 to 325 °C for 90 min under vacuum conditions (10-3 Torr) to form interparticular connections for developing electrical conductivity. 3. Characterization. The printed films were cross-sectioned by a focused ion beam (FIB, SMI-2050, SII Nanotechnology, Inc.) and their microstructure was observed by high-resolution SEM (JSM6700F, JEOL). Image analysis (Leica Qwin) was performed to generate metal/pore binary images from which the average porosity was calculated from at least five two-dimensional images obtained as the FIB progressively exposed new cross-sectioned surfaces. The sheet resistivities of the conductive films as a function of annealing temperatures were measured in air by a 4-point probe (CMT-SR200N, Chang Min Co., Ltd.), and the film thickness was measured using a surface profiler (AS500, KLA-Tencor Co.). The three-dimensional morphology of the ink-jet printed conductive patterns was observed by confocal laser scanning microscopy (LEXT OLS3000, Olympus). The thermal profile of the Cu-Ag mixed particles was measured under argon atmosphere by differential scanning calorimetry (DSC, STA409C, Netzsch). Exothermic peaks associated with the pyrolysis of the capping molecules were observed at above 250 °C accompanying a weight loss.

Results and Discussion The synthesized copper and silver nanoparticles were separately synthesized in ambient atmosphere by the polyol process. The mean particle size of Cu nanoparticles is 65 ( 3 nm (as shown in Figure 1a). The Cu particle is surrounded by a thin layer of (22) Kim, D.; Jeong, S.; Moon, J. Nanotechnology 2006, 17, 4019.

Ink-Jet Printing of Cu-Ag-Based ConductiVe Tracks

Langmuir, Vol. 25, No. 1, 2009 431

Figure 4. (a) The variation in resistivity of pure Cu film and Cu-Ag mixed film as a function of annealing temperature. (b) DSC analysis of the mixed Cu-Ag (3:1) nanoparticles. (c) SEM images showing the microstructure of pure Cu (the upper) and Cu/Ag ) 3:1 mixed film (the lower) annealed at 175, 250, and 325 °C for 90 min. Arrows indicate the presence of Ag particles between Cu particles. All scale bars ) 100 nm.

CuO with a thickness of ∼1.6 nm.16 When these Cu particles are consolidated into the particulate film, they form tetrahedral interstices whose ideal diameter is calculated to be 27 nm. The synthesized silver nanoparticles have a mean particle size of 21 ( 2 nm (Figure 1b). The added silver nanoparticles are small enough to fill the interstices between the Cu particles, improving the packing density of the particulate film. To verify the effect of silver particle addition on film density, three different conductive inks with varying ratios of Cu/Ag ) 2:1 by volume (denoted 2Cu1Ag), Cu/Ag ) 3:1 (denoted 3Cu1Ag), and Cu/Ag ) 4:1 (denoted 4Cu1Ag) were prepared. After annealing at 200 °C, the printed granular films were cross-sectioned by FIB for better microstructural observation, as shown in Figure 1c. The cross-sectional SEM images reveal that the annealed microstructures vary depending upon the mixing ratio. An image analysis was performed to generate metal/pore binary images from which the average porosity was calculated (Figure 1d). We expected that, as silver nanoparticles were added, the film density would increase because the presence of small silver nanoparticles would facilitate sintering. Contrary to our expectation, the 3Cu1Ag film is more densely packed than the 2Cu1Ag and 4Cu1Ag films. (23) Furnas, C. C. Ind. Eng. Chem 1931, 23, 1052. (24) German, R. M. Metall. Trans., A 1992, 23A, 1455.

We utilized the random packing model proposed by Furnas to understand the observed result.23-25 For an ideal twocomponent (binary) mixture consisting of particles with infinitely different sizes, the maximum fractional packing density (PFMAX) occurs when the larger particles (e.g., copper) constitute a packing at a density of PFL and the remaining interstitial volume (1 PFL) between the larger particles is filled by the smaller particles (e.g., silver) at a density of PFS, giving

PFMAX ) PFL + (1 - PFL)PFS The fractional packing density of the monosized spherical particle system proves to be 0.65, regardless of the particle size, if the ratio of the diameter of the container divided by the particle diameter is over 10.26 Therefore, the theoretical maximum fractional packing density of a bimodal mixture system is 0.8775. On the basis of this model, the bimodal mixture system approaches the maximum packing density when the volume ratio of the larger particles to the smaller particles is 2.86:1. In addition, the diameter ratio of the larger to the smaller particles should be at least 2.4, which is calculated based on the assumption that the smaller particles should fit into the tetragonal interstices between the larger particles. (25) German, R. M. Int. J. Powder Met. 1992, 28, 301.

432 Langmuir, Vol. 25, No. 1, 2009

Woo et al.

Figure 5. (a) Ink-jet printed complex conductive patterns on flexible and transparent PES substrates. The ink contains a mixture of Cu and Ag nanoparticles at a ratio of Cu/Ag ) 3:1. (b) The cross-sectional profiles and the confocal laser scanning images of a single ink droplet after drying and a single printed line.

This random packing model only predicts the initial packing density (green body packing density) in the bimodal system, instead of the annealed density. However, for instances of anneals performed at relatively low temperatures, the green body density should correlate well with the annealed density, so that it can be reasonably estimated from the measured density of the films annealed at 200 °C. Figure 2 compares the calculated packing density based on the Furnas model and the experimentally measured annealed density as a function of mixing ratio. In the Furnas model, the packing density increases as smaller particles are added to the larger particles. At Cu/Ag ) 4:1, the packing density is predicted to be 81.3%. The packing density reaches a maximum (87.0%) at Cu/Ag ) 3:1 and any further increase in the portion of small particles reduces the packing density (85.0% at Cu/Ag ) 2:1). The predicted packing density is in good agreement with the experimentally measured annealed density. As determined by image analysis, when the volume ratio of the Cu particles to the smaller Ag particles varies from 4 to 3 to 2, the annealed density of the corresponding film is 79% ( 1.5, 86% ( 1.7, and 82% ( 2.1, respectively. Close agreement between the predicted and experimental values indicates that the random packing model originally developed for a micrometersized particle system can be applied to nanometer-sized metal films. Addition of small Ag particles with specific particle size and mixing ratio with respect to the larger Cu particles effectively (26) Mcgeary, R. K. J. Am. Ceram. Soc. 1961, 44, 513.

maximizes the packing density by filling the interstices. The enhancement in the packing density means that there are more interparticle contact areas, which enables the granular film to develop a denser nanoparticulate film structure when annealed, resulting in better conductivity. To verify the effect of the density improvement, we measured the film resistivity as a function of the mixing ratio as shown in Figure 3. The film conductivity for all samples increases with increasing annealing temperature. The film composition of 3Cu1Ag exhibited the lowest electrical resistivity (23.6 ( 2.5 µΩ · cm when annealed at 200 °C) compared to 2Cu1Ag (48.3 ( 0.9 µΩ · cm) and 4Cu1Ag (50.8 ( 1.2 µΩ · cm). The observed resistivity also matches well with the annealed density variation of the corresponding films. It was expected that the 2Cu1Ag film would exhibit higher conductivity than other films because of its higher silver content than copper and no tendency to form resistive surface oxides. On the contrary, the 3Cu1Ag film, with denser annealed density, became the better conductive track. This finding indicates that the packing density of the printed metal particulate film is the predominant factor that determines its conductivity. A densely packed structure involves interparticle pores with a smaller curvature radius, which provides a higher driving force for sintering. The metal film with higher packing density undergoes densification at lower temperature, resulting in better conductivity after annealing. The particle size of the Ag nanoparticles plays an important role in improving the film conductivity (see the inset of Figure 3). We prepared the metal thin film from the mixed conductive ink containing the larger Ag particles (50 ( 3 nm) at the specific

Ink-Jet Printing of Cu-Ag-Based ConductiVe Tracks

mixing ratio where maximum packing density can be achieved, i.e., Cu/Ag ) 3:1. The diameter ratio of the Cu to the Ag particles was 1.3:1.0, lower than the minimum value calculated according to geometric considerations. The resistivity of the metal film based on the Cu-Ag(50 nm) was approximately 2-3 times higher than that based on the Cu-Ag(20 nm). This result confirms that the 50 nm Ag nanoparticles are unable to improve the packing density because of their larger size with respect to the pore size between the Cu nanoparticles. The size ratio of the constituent particles should be controlled to be at least over 2.4 to improve the packing density in a randomly packed bimodal system. The effect of particle mixing becomes more evident when comparing the resistivity of the metal film composed only of Cu nanoparticles (Figure 4a).16 There was significant improvement in the film conductivity (23.60 µΩ · cm when annealed at 200 °C) when annealed between 150 and 250 °C as compared to that of the pure copper film (1.294 × 107 µΩ · cm). Formation of better conductive tracks at lower annealing temperatures can be attributed to the presence of smaller Ag nanoparticles rather than to the larger Cu particles. The small size of Ag particles can further depress the melting point as well as improve packing density, which allows faster densification compared to when only large Cu nanoparticles are present in the film. This result is supported by the observation of exothermic peaks in DSC analysis (Figure 4b). The formation of necks between particles involves surface diffusion of unstable surface atoms, which is characterized by the exothermic peak.27 The sintering process of smaller silver particles begins at around 135 °C, whereas larger copper particles sinter at around 165 °C. The DSC result clearly reveals that the addition of silver nanoparticles expedites the sintering process at lower temperatures. SEM analysis reveals that the 3Cu1Ag mixed metal film has a much denser microstructure compared to the pure Cu film when annealed at the same temperature, because the former forms more interparticle junctions (see Figure 4c and Supporting Information, Figure S1). To demonstrate the applicability of the mixed metal-based conductive ink (Cu/Ag ) 3:1) from which high conductive tracks form by annealing at low temperature, we direct-wrote complex patterns on transparent plastic substrates. Figure 5 shows conductive patterns printed on flexible PES substrate. Solvent (27) Moon, K.-S.; Dong, H.; Maric, R.; Pothukuchi, S.; Li, Y.; Wong, C. P. J. Electron. Mater. 2005, 34, 168.

Langmuir, Vol. 25, No. 1, 2009 433

evaporation from the printed single ink droplet produces a spherical dot pattern of diameter ∼80 µm. The line pattern is generated by reducing the dot-to-dot distance. The separated dots begin to merge together at a distance of 100 µm, and printing at an interdot distance of 90 µm results in a continuous line with a ∼85 µm line-width and relatively smooth edge definition. The conductive features exhibited relatively low electrical resistivity of 38.6 µΩ · cm after annealing at 175 °C for 90 min and 23.6 µΩ · cm (10 times higher than theoretical resistivity) at 200 °C.

Conclusions We have developed ink-jet printable ink containing Cu and Ag nanoparticles of different sizes, which can be used to form highly conductive tracks on a flexible substrate after annealing at low temperature. Addition of small Ag particles significantly improves the particle packing density by filling the pores between the larger Cu particles, which in turn improve conductivity at lower temperatures. The particle size and volume ratio of the Ag particles with respect to the larger Cu particles should be carefully controlled to achieve maximum packing density in the bimodal particle system, which is in good agreement with theoretical considerations based on the Furnas random packing model. The maximum annealed density of the film was determined to be 86% when Ag with a mean size of 20 nm was added to Cu with a mean size of 65 nm at the ratio of Cu/Ag ) 3:1. The film resistivity reached 23.6 µΩ · cm when annealed at 200 °C. Importantly, this mixed metal-based conductive ink allows us to directly write conductive features on transparent plastic substrates, suggesting the potential of creating plastic electronics for applications such as radio frequency identification (RFID) tags and thin-film transistor (TFT) circuits in a simple and economic manner. Acknowledgment. This work was supported by the Korea Science and Engineering Foundation (KOSEF) through the National Research Laboratory. This program is funded by the Ministry of Education, Science and Technology (No. R0A-2005000-10011-0). Supporting Information Available: The back scattered images and EDS analysis of Cu/Ag ) 3:1 mixed film. This material is available free of charge via the Internet at http://pubs.acs.org. LA802182Y