Home
Search
Collections
Journals
About
Contact us
My IOPscience
Improving mechanical fatigue resistance by optimizing the nanoporous structure of inkjetprinted Ag electrodes for flexible devices
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 125706 (http://iopscience.iop.org/0957-4484/25/12/125706) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 169.230.243.252 This content was downloaded on 27/02/2015 at 16:12
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 25 (2014) 125706 (6pp)
doi:10.1088/0957-4484/25/12/125706
Improving mechanical fatigue resistance by optimizing the nanoporous structure of inkjet-printed Ag electrodes for flexible devices Byoung-Joon Kim1,4 , Thomas Haas1 , Andreas Friederich1 , Ji-Hoon Lee2 , Dae-Hyun Nam2 , Joachim R Binder1 , Werner Bauer1 , In-Suk Choi3 , Young-Chang Joo2 , Patric A Gruber1 and Oliver Kraft1 1
Institute for Applied Materials, Karlsruhe Institute of Technology, Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany 2 Department of Materials Science and Engineering, Seoul National University, Seoul 151-744, Republic of Korea 3 High Temperature Energy Materials Research Center, Korea Institute of Science and Technology, Seoul 136-791, Republic of Korea E-mail: [email protected] and [email protected] Received 24 October 2013, revised 23 December 2013 Accepted for publication 9 January 2014 Published 27 February 2014
Abstract
The development of highly conductive metallic electrodes with long-term reliability is in great demand for real industrialization of flexible electronics, which undergo repeated mechanical deformation during service. In the case of vacuum-deposited metallic electrodes, adequate conductivity is provided, but it degrades gradually during cyclic mechanical deformation. Here, we demonstrate a long-term reliable Ag electrode by inkjet printing. The electrical conductivity and the mechanical reliability during cyclic bending are investigated with respect to the nanoporous microstructure caused by post heat treatment, and are compared to those of evaporated Ag films of the same thickness. It is shown that there is an optimized nanoporous microstructure for inkjet-printed Ag films, which provides a high conductivity and improved reliability. It is argued that the nanoporous microstructure ensures connectivity within the particle network and at the same time reduces plastic deformation and the formation of fatigue damage. This concept provides a new guideline to develop an efficient method for highly conductive and reliable metallic electrodes for flexible electronics. Keywords: inkjet printing, fatigue lifetime, flexible electrode, nanoporous (Some figures may appear in colour only in the online journal)
1. Introduction
opment of metallic interconnects/electrodes with high conductivity, productivity, and reliability is of great importance for the efficient and stable electrical connection inside and between electronic devices. Inkjet printing using suspensions of metal nanoparticles is one of the most prominent techniques to replace the conventional expensive vacuum processes due to its low cost, high throughput, and high substrate compatibility [6–9].
Flexible electronics fabricated on compliant substrates are under active development for a wide range of applications such as flexible displays [1], flexible batteries [2–4], or electronic skin [5]. To realize flexible electronics, the devel4 Present address: Surface Technology Division, Korea Institute of Materials
Science (KIMS), Changwon, 641-831, Republic of Korea. 0957-4484/14/125706+06$33.00
1
c 2014 IOP Publishing Ltd
Printed in the UK
Nanotechnology 25 (2014) 125706
B-J Kim et al
The typical process chain to form conductive metallic electrodes by inkjet printing is that a suspension of metal nanoparticles coated by an organic shell is printed on a flexible substrate and a sequential post heat treatment is conducted to achieve high conductivity electrodes by decomposing the organic shell and forming a dense film-like structure [10–13]. The relationship between electrical conductivity and microstructural evolution such as particle growth, neck formation, and porosity during post heat treatment has been widely studied [13, 14]. The electrical conductivity can be further improved by forming a dense microstructure and reducing the porosity by controlling annealing temperature [10–13], annealing time [13, 14], heat treatment method [15], and ambient atmosphere [16, 17]. The mechanical properties of the inkjet-printed film are also very sensitive to its microstructure. In particular, the mechanical reliability of the nanoparticulate film under various mechanical deformation modes such as monotonic tensile or cyclic deformation (fatigue) is a critical issue for flexible device application. For monotonic tensile testing, it has been generally reported that the inkjet-printed film has a lower fracture strain compared to an evaporated film because nanopores may act as stress concentrators and accelerate crack nucleation [16, 18, 19]. On the other hand, for fatigue tests, it has been reported that inkjet-printed films may exhibit a lower electrical resistance increase during fatigue compared to evaporated films [20], or at least a comparable fatigue lifetime may be achieved [19]. Based on this it is still unclear how the nanoporous microstructure affects the fatigue behavior of inkjet-printed films. Therefore, a systematic study of its effect on fatigue behavior is required to develop long-term reliable metallic electrodes by inkjet printing. In this study, we show a fatigue-resistant and highly conductive Ag electrode on polymer substrate using inkjet printing. The nanoporous structure has been optimized efficiently by post heat treatment with respect to electrical as well as mechanical performance. The optimized nanoporous electrode exhibited a considerably low initial resistivity of 4.28 µ cm, which is only 1.3 times higher than that of an evaporated film having same thickness. Furthermore, it showed a significantly improved reliability during cyclic bending deformation. It is shown that the printed nanoporous film maintained its initial resistivity for 5 × 105 cycles, while that of the conventionally evaporated film increased by a factor of two for the same number of cycles. This is attributed to the specifics of the microstructure of the printed film with nanoscale particles covered with remaining organic shell and nanopores. Our study suggests a highly efficient fabrication route for flexible metallic electrodes with high conductivity and long-term reliability.
Figure 1. Schematic illustration of the bending fatigue test system
with in situ electrical resistivity measurement.
printing frequency of 500 Hz. The sample stage was heated to 110 ◦ C during printing. The dimensions of the electrodes were 4 mm × 62 mm (50 drops × 775 drops). After drying on the sample stage, the films were post-annealed in a convection oven (150, 200, 250 ◦ C for 30 min) or a furnace (350 ◦ C for 60 min). All heat treatments were conducted in air. The thickness of the inkjet-printed films was around 800 nm. In order to identify the decomposition behavior of the organic shell, thermogravimetric analysis (TGA) was conducted in air using different target temperatures between 150 and 600 ◦ C and a heating rate of 10 K min−1 . For comparison, a conventional 800 nm thick Ag thin film was also deposited on a PI substrate by thermal evaporation (base pressure 5 × 10−6 Torr, deposition rate 25 nm s−1 ). The remaining organics and Ag surface fraction after the different post heat treatments were analyzed by x-ray photoelectron spectroscopy (XPS). The mechanical reliability of inkjet-printed and evaporated films was investigated by a custom-built bending fatigue test system (CK Trading Company, Seoul, Korea). A schematic diagram of the test setup is shown in figure 1. The Ag films on compliant PI substrates were gripped at both ends by metal grips, which allowed for in situ conductivity measurements. While the top board was fixed, the bottom board was moving backward and forward over a range of 10 mm, as indicated by the arrow in figure 1. The repeating linear motion leads to cyclic straining of the sample and the evolution of fatigue damage in the metallic films as discussed in detail in our previous work [21, 22]. In this testing scheme, the stressed volume is relatively large compared to a three-point bending test system since the film is cyclically deformed at a total length of about 20 mm. The gap between the two boards was 11.4 mm, which results in a maximum tensile strain of 1.1% along the sample [22]. The testing frequency was 5 Hz. The change in the resistivity of the Ag films during fatigue testing was measured using a custom-built two-point measurement (Agilent 34410A, Keithley 7007). The microstructure of the samples was investigated after fatigue testing using field emission scanning electron microscopy (FESEM) and focused ion beam (FIB) microscopy. Particles and pores were marked on SEM images and digitized by the image analysis software Scion Image for Windows. From the measured area of the particles, the average diameter of the particle was calculated assuming a circular shape. The surface porosity was defined as
2. Experiment
A solvent based Ag nanoparticle ink with 20 wt% silver content (EMD5603, Sun Chemical) was used in this study. The Ag ink was printed on 125 µm thick polyimide (PI) substrate R (DuPont, Kapton HN) using a single nozzle drop-on-demand inkjet printer (Autodrop Professional, Microdrop), a printhead with 70 µm orifice diameter, 80 µm drop distance, and a 2
Nanotechnology 25 (2014) 125706
B-J Kim et al
the ratio of pore area to total image area. The Ag film thickness was measured by cutting cross-sections by FIB and analyzing the corresponding SEM images. 3. Results and discussion
For a suspension of metallic nanoparticles surrounded by an organic shell, the decomposition temperature (Td ) is one of the most critical factors for defining the microstructure and the resulting electrical/mechanical properties of the final metallic film. In order to optimize the heat treatment conditions for controlling the microstructure of the film, we first determined Td of our Ag ink by TGA as shown in figure 2(a). The mass of the sample decreased steadily up to 250 ◦ C and saturated at 20% during further heating up to 600 ◦ C, which is about the same as the initial Ag content of nanoparticle ink. This is attributed to the evaporation of the solvent and decomposition of the organic shells, which should occur simultaneously during continuous heating up to 250 ◦ C. The decomposition temperature, Td , of our ink was thereby determined to be 250 ◦ C. By considering Td , we applied four different heat treatment conditions: 150 ◦ C, 200 ◦ C, and 250 ◦ C for 30 min, and 350 ◦ C for 60 min to represent the conditions below, around, and above Td , respectively. Figure 2(b) shows the mass change as a function of annealing time when the maximum temperatures were set according to the heat treatments defined above. For maximum temperatures of 250 ◦ C and 350 ◦ C, the mass decreased rapidly to around 20% after 45 min and 35 min, respectively. The mass of the sample annealed at 200 ◦ C decreased at a lower rate. Here, it took about 150 min to reach 20% of the initial mass. The relative mass for the annealing at 150 ◦ C never reached 20%, but saturated at 40% even after 300 min. This already suggests that the annealing temperature needs to be above 200 ◦ C to ensure complete evaporation of the solvent and decomposition of the organic shells. Therefore, some remaining organics are expected for the samples with heat treatment at 150 and 200 ◦ C and possibly even at 250 ◦ C. The chemical composition at the surface of samples with different heat treatments was analyzed by XPS. The results are shown in figure 2(c). About 20% of carbon was detected in samples annealed at 150 and 200 ◦ C, and less than 20% was obtained in the sample annealed at 250 ◦ C. In contrast, no carbon was detected at the surface of the sample annealed at 350 ◦ C. This again indicates that some organics still remained due to the lower annealing temperature. Remaining organics generally deteriorate the electrical conductivity [13]. However, they could improve mechanical properties such as interfacial strength due to the carbon bridging effect [23]. Figures 3(a)–(d) show plane view and cross-sectional SEM images of the inkjet-printed Ag films for the four different annealing conditions. The microstructural evolution was quantified by SEM image analysis. The particle size was found to be distributed lognormally, and the median particle size, surface porosity, and film thickness are listed in table 1. The surface porosity was calculated by the ratio of pore area to total area except for the sample annealed at 150 ◦ C, because for this sample the pore size was too small to be detected.
Figure 2. (a) Relative mass change of the Ag nanoparticle-based ink as a function of temperature using a heating rate of 10 K min−1 .
(b) Mass change of the Ag ink as a function of annealing time when the temperature was held at 150, 200, 250, and 350 ◦ C. (c) XPS analysis of the chemical composition at the surface of the inkjet-printed Ag films.
For increasing annealing temperature, the particle size increased, and the porosity decreased, respectively. The microstructural evolution of inkjet-printed Ag films during annealing can be characterized as a sequential process of neck formation, particle coalescence, particle coarsening, and secondary grain growth [13]. The inkjet-printed Ag film annealed at 150 ◦ C showed the smallest median particle size (31.0 ± 12.2 nm), which was still larger than that of as-dried ink (about 10 nm). It is argued that in this sample the nanoparticles remained as isolated particles because the temperature was too low to decompose the organic shell and to initiate neck formation. In contrast, the particle size for the sample annealed at 200 ◦ C has increased (65.8 ± 15.4 nm) and neck formation between the nanoparticles was observed. Therefore, this sample is in the stage of neck formation and coalescence. For the sample annealed at 250 ◦ C, a further increase of particle size (145 ± 41.1 nm) and pore size was found, which 3
Nanotechnology 25 (2014) 125706
B-J Kim et al
Figure 3. Plane view and cross-sectional SEM images of the inkjet-printed Ag films after annealing at (a) 150 ◦ C, (b) 200 ◦ C, and (c) 250 ◦ C for 30 min, and (d) 350 ◦ C for 60 min. Table 1. Median grain size, surface porosity, and film thickness of inkjet-printed Ag films after different heat treatments.
Annealing condition
150 ◦ C, 30 min
200 ◦ C, 30 min
250 ◦ C, 30 min
350 ◦ C, 60 min
Median particle size (nm) Surface porosity (%) Film thickness (nm)
31.0 ± 12.2 — 800 ± 35.5
65.8 ± 15.4 9.6 755 ± 47.0
145 ± 41.1 6.6 780 ± 22.6
239 ± 83.1 1.1 780 ± 33.7
of pores—about 5 times smaller than the sample annealed at 150 ◦ C and only 17% larger than that of the sample annealed at high temperature (350 ◦ C). By applying a higher annealing temperature, the particle size further increases, and the porosity decreases. However, the effect on the resistivity is not as significant as the effect of neck formation and organic decomposition occurring at Td [13, 14, 17]. This implies that a sufficiently high conductivity can be achieved by optimizing the nanoporous microstructure. Figure 4(a) shows the electrical resistance change of the inkjet-printed and evaporated Ag as a function of the number of bending cycles using a bending strain of 1.1%. The evaporated Ag film showed poor fatigue properties; the electrical resistance increased to 110% after 5 × 105 cycles. The cross-sectional SEM image of this film shows typical fatigue damage of evaporated metallic thin films. Extrusions and intrusions were formed by irreversible dislocation motion during cyclic deformation, which leads to the crack formation shown for Cu [24, 25] and Ag thin films [26]. In contrast, inkjet-printed Ag films showed superior fatigue resistance compared to that of the evaporated film. The electrical resistivity of the samples annealed at 150, 200, and 250 ◦ C did not increase significantly from their initial values (1R/R0 < 5%) up to 5 × 105 cycles. Only the sample annealed at 350 ◦ C showed a resistivity increase of more than 80%. Figures 4(b)–(e) show SEM images after 5 × 105 bending cycles of all inkjet-printed films. All images were captured at 52◦ tilt, and the loading direction was horizontal. No damage was observed in the samples annealed at 150, 200, and 250 ◦ C. However, in the sample annealed at 350 ◦ C long cracks and extrusions were observed, which explains the strong increase in electrical resistivity. It is interesting that the more porous films (samples annealed at 150, 200, and 250 ◦ C) have a superior fatigue resistance compared to dense films, although the pores in metal films have been commonly regarded as a mechanically weak site due to the reduced cross-sectional area and stress concentration [16, 18, 19].
may be defined by the stage of coarsening. After annealing at 350 ◦ C for 60 min the sample showed an almost dense microstructure consisting of fairly large grains, which may be related to secondary grain growth. Concomitantly to particle coarsening and secondary grain growth, the porosity, as seen at the surface, decreased from 9.6% to 1.1% with increasing annealing temperature. In the cross-sectional images, similar trends of particle growth, neck formation and evolution of pore and overall porosity were observed. The film thickness of inkjet-printed films was reported to be either decreasing during high temperature annealing due to the densification of the films [15, 16], or not to change significantly [13, 23]. For our samples, the thicknesses varied between 755 ± 47.0 nm and 800 ± 35.5 nm, but no significant trend was observed. The electrical resistivity of inkjet-printed film was found to vary for the different annealing conditions corresponding to the described microstructural features. The electrical resistivities of the samples annealed at 150 ◦ C, 200 ◦ C, 250 ◦ C, and 350 ◦ C were 23 ± 0.9, 6.9 ± 0.6, 4.3 ± 0.2, and 3.7 ± 0.2 µ cm, respectively, while that of the evaporated Ag film with 800 nm film thickness was measured to be 3.3 ± 0.1 µ cm and that of bulk Ag is 1.6 µ cm. These values are comparable to those from other reports for the resistivity of solution based Ag films in the range of 2.64–12.5 µ cm [15, 18]. The high resistivity of the sample annealed at 150 ◦ C is attributed to the lack of a conducting path due to the isolated particles and remaining organic shell. The resistivity decreased for the sample annealed at 200 ◦ C mainly due to neck formation between the nanoparticles. With further increase of annealing temperature to 250 and 350 ◦ C, the resistivity decreased continuously to a level similar to that of the evaporated film by particle coarsening and secondary grain growth. Finally, it was only 10–30% higher than that of the evaporated Ag film. It is noteworthy that the sample annealed at 250 ◦ C shows an excellent low electrical resistivity despite the large portion 4
Nanotechnology 25 (2014) 125706
B-J Kim et al
Figure 4. (a) Electrical resistivity change of the inkjet-printed Ag films and the evaporated Ag film during a bending fatigue test. The bending strain was 1.1%. Errors imply the standard deviation for at least three samples. The top-right image is the cross-section of an evaporated Ag film after 5 × 105 cycles. The thickness of the evaporated Ag film was 800 nm, which is similar to those of the inkjet-printed films. ((b)–(e)) Surface morphologies of the inkjet-printed Ag films after 5 × 105 cycles after annealing at (b) 150 ◦ C for 30 min, (c) 200 ◦ C for 30 min, (d) 250 ◦ C for 30 min, and (e) 350 ◦ C for 60 min. All SEM images were captured at 52◦ tilt. (f) Schematic illustrations of fatigue behavior of inkjet-printed films with different microstructures.
increase. Nevertheless, extrusions and voids at the interface to the substrate are observed, from which cracks are nucleated. With decreasing film thickness, fewer and smaller extrusions are observed, and it can be argued that defects are annihilated at free surfaces or grain boundaries [24]. In our test, the uniformly distributed nanoscale pores may act as an annihilation site for single dislocations and thus prevent damage formation [20, 21]. Consequently, the formation of extrusions and subsequent cracking are reduced due to the nanoporous microstructure of the films. It should be noted that the stiffness of the printed thin film decreases with increasing porosity [28]. In the elastic regime, at a given strain, the low stiffness of the nanoporous films leads to lower film stresses, and the driving force for fatigue damage formation is reduced. However, the total applied strain of 1.1% is fairly large, and thus plastic deformation is still expected to occur. In contrast, many cracks and extrusions were observed in the samples annealed at 350 ◦ C. Initially, this film was dense and had the lowest electrical resistivity, but finally it showed the worst mechanical reliability due to extensive fatigue damage formation. Due to the higher annealing temperature, particle growth was strongly enhanced and a dense and coarse microstructure was obtained. Dislocation plasticity is further facilitated in the large grains, and dislocations are accumulated and form extrusions and intrusions as shown in figure 4(e). This finally leads to crack formation, and a similar fatigue behavior as for the evaporated Ag film is observed. Figure 5 shows the electrical conductivity and fatigue lifetime of the evaporated film and the inkjet-printed films with
The improvement of fatigue resistance in the porous films is mainly attributed to the small particle size and uniform distribution of the nanopores. The schematic diagram in figure 4(f) shows the fatigue behaviors of inkjet-printed films with different microstructures. After annealing at low temperatures such as 150 ◦ C, the microstructure of the films still consists of small isolated metallic particles surrounded by organic shell. Dislocation plasticity is restricted within the volume of the small particles, as dislocations annihilated at the surface of the particles, leading to a dislocation-starved condition [27]. Furthermore, the remaining organic shell may prevent the formation of extrusions and intrusions. Thereby the nanoparticulate film is more resistant to fatigue damage, but it has the inherent drawback of high initial electrical resistivity due to the absence of neck formation between the particles and the remaining organic shell. In the samples annealed at 200 and 250 ◦ C, the neck formation and particle coarsening has occurred, while some organics may still have remained. Due to the larger particle size, the activation stress for dislocation plasticity is reduced compared to the sample annealed at 150 ◦ C. Nevertheless, the particle size is still so small that the formation of long-range ordered dislocation structures such as cells, veins, or even persistent slip bands is not expected [25]. It is well known that in fatigue bulk metals, fatigue damage such as extrusions and intrusions is formed where persistent slip bands intersect with the surface. In contrast, in metallic thin films with thickness below 1 µm, such dislocation structures are no longer observed [24–26], and the fatigue resistance tends to 5
Nanotechnology 25 (2014) 125706
B-J Kim et al
Acknowledgments
This work was supported by the Global Leading Technology Program of the Office of Strategic R&D Planning (OSP) funded by the Korean government (Ministry of Trade, Industry and Energy) (No. 10042537). Oliver Kraft is grateful to the Robert Bosch Foundation for supporting his Endowed Chair for Functional Nanomaterials, and Byoung-Joon Kim gratefully acknowledges the KIT guest program. References [1] Forrest S R 2004 Nature 428 911 [2] Koo M, Park K I, Lee S H, Jeon D Y, Choi J W, Kang K and Lee K J 2012 Nano Lett. 12 4810 [3] Nam K T, Kim D W, Yoo P J, Meethong N, Hammond P T, Chiang Y M and Belcher A M 2006 Science 312 885 [4] Li Y, Lee D K, Kim J Y, Kim B, Park N G, Kim K, Shin J H, Choi I S and Ko M J 2012 Energy Environ. Sci. 5 8950 [5] Yeo W H et al 2013 Adv. Mater. 25 2773 [6] Ko S H, Pan H, Grigoropoulos C P, Luscombe C K, Fr´echet J M J and Poulikakos D 2007 Nanotechnology 18 345202 [7] Van Osch T H J, Perelaer J, De Laat A W M and Schubert U S 2008 Adv. Mater. 20 343 [8] Lee H M, Choi S Y, Kim K T, Yun J Y, Jung D S, Park S B and Park J 2011 Adv. Mater. 23 5524 [9] Lee H M, Choi S Y, Jung A and Ko S H 2013 Angew. Chem. Int. Edn. 52 7718 [10] Lee H H, Chou K S and Huang K C 2005 Nanotechnology 16 2436 [11] Jeong S, Kim D, Lee S, Park B K and Moon J 2006 Mol. Cryst. Liq. Cryst. 459 35 [12] Park J W and Baek S G 2006 Scr. Mater. 55 1139 [13] Jung J K, Choi S H, Kim I, Jung H C, Joung J and Joo Y C 2008 Phil. Mag. 88 339 [14] Greer J R and Street R A 2007 Acta Mater. 55 6345 [15] Kim N R, Lee J H, Yi S M and Joo Y C 2011 J. Electrochem. Soc. 158 K165 [16] Lee J H, Kim N R, Kim B J and Joo Y C 2012 Carbon 50 98 [17] Yi S M, Lee J H, Kim N R, Oh S, Jang S, Kim D, Joung J and Joo Y C 2010 J. Electrochem. Soc. 157 K254 [18] Kim S, Won S, Sim G D, Park I and Lee S B 2013 Nanotechnology 24 085701 [19] Sim G D, Won S and Lee S B 2012 Appl. Phys. Lett. 101 191907 [20] Lee H Y, Yi S M, Lee J H, Lee H S, Hyun S and Joo Y C 2010 Met. Mater. Int. 6 947 [21] Kim B J, Cho Y, Jung M S, Shin H A S, Moon M W, Han H N, Nam K T, Joo Y C and Choi I S 2012 Small 8 3300 [22] Kim B J, Shin H A S, Jung S Y, Cho Y, Kraft O, Choi I S and Joo Y C 2013 Acta Mater. 61 3473 [23] Lee I, Kim S, Yun J, Park I and Kim T S 2012 Nanotechnology 23 485704 [24] Schwaiger R, Dehm G and Kraft O 2003 Phil. Mag. 83 693 [25] Zhang G P, Volkert C A, Schwaiger R, Wellner P, Arzt E and Kraft O 2006 Acta Mater. 54 3127 [26] Schwaiger R and Kraft O 2003 Acta Mater. 51 195 [27] Julia R G and William D N 2006 Phys. Rev. B 73 245410 [28] Sanders P G, Eastman J A and Weertman J R 1997 Acta Mater. 45 4019
Figure 5. Electrical conductivities and fatigue lifetimes of the inkjet-printed Ag films and the evaporated Ag films. The fatigue lifetime was defined as the cycle number at which a 20% increase in resistivity was observed.
different heat treatments. The fatigue lifetime was defined as the cycle number at which a 20% increase in the resistivity of the individual sample was observed. In the case of no fatigue damage up to 5 × 105 cycles, fatigue lifetime was plotted at 5 × 105 cycles and open symbols were used. The electrical conductivity of the inkjet-printed films increases with increasing annealing temperature and is highest for the highest annealing temperature. However, the fatigue lifetime of this sample is deteriorated due to fatigue damage formation as discussed above. This indicates that an optimum compromise of high conductivity and long-term reliability of flexible metallic electrodes can be achieved by establishing a nanoporous microstructure of the material using an inkjet printing process with nanoparticulate suspensions and appropriate annealing conditions. 4. Conclusions
The fatigue behavior and electrical performance of inkjetprinted Ag films with different nanoporous microstructures were investigated using a bending fatigue test system. Samples annealed at temperatures below the decomposition temperature Td of the organic shells of the nanoparticulate Ag ink (Td = 250 ◦ C) showed superior fatigue resistance, but had a high initial resistivity due to the absence of neck formation between the particles. Samples annealed at high temperature above Td showed a low initial resistivity due to the dense and coarse microstructure, but the fatigue lifetime is deteriorated due to the early formation of extrusions and cracks. By annealing around Td , a nanoporous microstructure is achieved, having not only high electrical conductivity but also beneficial fatigue resistance. The electrical conductivity is obtained due to neck formation within the particle network, whereas the enhanced fatigue resistance results from hindered dislocation plasticity due to the small particle size and dislocation annihilation at free surfaces. This systematic study of the effect of nanopores on electrical and mechanical properties can provide a helpful solution to develop highly conductive and reliable flexible metallic electrodes using inkjet printing. 6
Search
Collections
Journals
About
Contact us
My IOPscience
Improving mechanical fatigue resistance by optimizing the nanoporous structure of inkjetprinted Ag electrodes for flexible devices
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 125706 (http://iopscience.iop.org/0957-4484/25/12/125706) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 169.230.243.252 This content was downloaded on 27/02/2015 at 16:12
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 25 (2014) 125706 (6pp)
doi:10.1088/0957-4484/25/12/125706
Improving mechanical fatigue resistance by optimizing the nanoporous structure of inkjet-printed Ag electrodes for flexible devices Byoung-Joon Kim1,4 , Thomas Haas1 , Andreas Friederich1 , Ji-Hoon Lee2 , Dae-Hyun Nam2 , Joachim R Binder1 , Werner Bauer1 , In-Suk Choi3 , Young-Chang Joo2 , Patric A Gruber1 and Oliver Kraft1 1
Institute for Applied Materials, Karlsruhe Institute of Technology, Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany 2 Department of Materials Science and Engineering, Seoul National University, Seoul 151-744, Republic of Korea 3 High Temperature Energy Materials Research Center, Korea Institute of Science and Technology, Seoul 136-791, Republic of Korea E-mail: [email protected] and [email protected] Received 24 October 2013, revised 23 December 2013 Accepted for publication 9 January 2014 Published 27 February 2014
Abstract
The development of highly conductive metallic electrodes with long-term reliability is in great demand for real industrialization of flexible electronics, which undergo repeated mechanical deformation during service. In the case of vacuum-deposited metallic electrodes, adequate conductivity is provided, but it degrades gradually during cyclic mechanical deformation. Here, we demonstrate a long-term reliable Ag electrode by inkjet printing. The electrical conductivity and the mechanical reliability during cyclic bending are investigated with respect to the nanoporous microstructure caused by post heat treatment, and are compared to those of evaporated Ag films of the same thickness. It is shown that there is an optimized nanoporous microstructure for inkjet-printed Ag films, which provides a high conductivity and improved reliability. It is argued that the nanoporous microstructure ensures connectivity within the particle network and at the same time reduces plastic deformation and the formation of fatigue damage. This concept provides a new guideline to develop an efficient method for highly conductive and reliable metallic electrodes for flexible electronics. Keywords: inkjet printing, fatigue lifetime, flexible electrode, nanoporous (Some figures may appear in colour only in the online journal)
1. Introduction
opment of metallic interconnects/electrodes with high conductivity, productivity, and reliability is of great importance for the efficient and stable electrical connection inside and between electronic devices. Inkjet printing using suspensions of metal nanoparticles is one of the most prominent techniques to replace the conventional expensive vacuum processes due to its low cost, high throughput, and high substrate compatibility [6–9].
Flexible electronics fabricated on compliant substrates are under active development for a wide range of applications such as flexible displays [1], flexible batteries [2–4], or electronic skin [5]. To realize flexible electronics, the devel4 Present address: Surface Technology Division, Korea Institute of Materials
Science (KIMS), Changwon, 641-831, Republic of Korea. 0957-4484/14/125706+06$33.00
1
c 2014 IOP Publishing Ltd
Printed in the UK
Nanotechnology 25 (2014) 125706
B-J Kim et al
The typical process chain to form conductive metallic electrodes by inkjet printing is that a suspension of metal nanoparticles coated by an organic shell is printed on a flexible substrate and a sequential post heat treatment is conducted to achieve high conductivity electrodes by decomposing the organic shell and forming a dense film-like structure [10–13]. The relationship between electrical conductivity and microstructural evolution such as particle growth, neck formation, and porosity during post heat treatment has been widely studied [13, 14]. The electrical conductivity can be further improved by forming a dense microstructure and reducing the porosity by controlling annealing temperature [10–13], annealing time [13, 14], heat treatment method [15], and ambient atmosphere [16, 17]. The mechanical properties of the inkjet-printed film are also very sensitive to its microstructure. In particular, the mechanical reliability of the nanoparticulate film under various mechanical deformation modes such as monotonic tensile or cyclic deformation (fatigue) is a critical issue for flexible device application. For monotonic tensile testing, it has been generally reported that the inkjet-printed film has a lower fracture strain compared to an evaporated film because nanopores may act as stress concentrators and accelerate crack nucleation [16, 18, 19]. On the other hand, for fatigue tests, it has been reported that inkjet-printed films may exhibit a lower electrical resistance increase during fatigue compared to evaporated films [20], or at least a comparable fatigue lifetime may be achieved [19]. Based on this it is still unclear how the nanoporous microstructure affects the fatigue behavior of inkjet-printed films. Therefore, a systematic study of its effect on fatigue behavior is required to develop long-term reliable metallic electrodes by inkjet printing. In this study, we show a fatigue-resistant and highly conductive Ag electrode on polymer substrate using inkjet printing. The nanoporous structure has been optimized efficiently by post heat treatment with respect to electrical as well as mechanical performance. The optimized nanoporous electrode exhibited a considerably low initial resistivity of 4.28 µ cm, which is only 1.3 times higher than that of an evaporated film having same thickness. Furthermore, it showed a significantly improved reliability during cyclic bending deformation. It is shown that the printed nanoporous film maintained its initial resistivity for 5 × 105 cycles, while that of the conventionally evaporated film increased by a factor of two for the same number of cycles. This is attributed to the specifics of the microstructure of the printed film with nanoscale particles covered with remaining organic shell and nanopores. Our study suggests a highly efficient fabrication route for flexible metallic electrodes with high conductivity and long-term reliability.
Figure 1. Schematic illustration of the bending fatigue test system
with in situ electrical resistivity measurement.
printing frequency of 500 Hz. The sample stage was heated to 110 ◦ C during printing. The dimensions of the electrodes were 4 mm × 62 mm (50 drops × 775 drops). After drying on the sample stage, the films were post-annealed in a convection oven (150, 200, 250 ◦ C for 30 min) or a furnace (350 ◦ C for 60 min). All heat treatments were conducted in air. The thickness of the inkjet-printed films was around 800 nm. In order to identify the decomposition behavior of the organic shell, thermogravimetric analysis (TGA) was conducted in air using different target temperatures between 150 and 600 ◦ C and a heating rate of 10 K min−1 . For comparison, a conventional 800 nm thick Ag thin film was also deposited on a PI substrate by thermal evaporation (base pressure 5 × 10−6 Torr, deposition rate 25 nm s−1 ). The remaining organics and Ag surface fraction after the different post heat treatments were analyzed by x-ray photoelectron spectroscopy (XPS). The mechanical reliability of inkjet-printed and evaporated films was investigated by a custom-built bending fatigue test system (CK Trading Company, Seoul, Korea). A schematic diagram of the test setup is shown in figure 1. The Ag films on compliant PI substrates were gripped at both ends by metal grips, which allowed for in situ conductivity measurements. While the top board was fixed, the bottom board was moving backward and forward over a range of 10 mm, as indicated by the arrow in figure 1. The repeating linear motion leads to cyclic straining of the sample and the evolution of fatigue damage in the metallic films as discussed in detail in our previous work [21, 22]. In this testing scheme, the stressed volume is relatively large compared to a three-point bending test system since the film is cyclically deformed at a total length of about 20 mm. The gap between the two boards was 11.4 mm, which results in a maximum tensile strain of 1.1% along the sample [22]. The testing frequency was 5 Hz. The change in the resistivity of the Ag films during fatigue testing was measured using a custom-built two-point measurement (Agilent 34410A, Keithley 7007). The microstructure of the samples was investigated after fatigue testing using field emission scanning electron microscopy (FESEM) and focused ion beam (FIB) microscopy. Particles and pores were marked on SEM images and digitized by the image analysis software Scion Image for Windows. From the measured area of the particles, the average diameter of the particle was calculated assuming a circular shape. The surface porosity was defined as
2. Experiment
A solvent based Ag nanoparticle ink with 20 wt% silver content (EMD5603, Sun Chemical) was used in this study. The Ag ink was printed on 125 µm thick polyimide (PI) substrate R (DuPont, Kapton HN) using a single nozzle drop-on-demand inkjet printer (Autodrop Professional, Microdrop), a printhead with 70 µm orifice diameter, 80 µm drop distance, and a 2
Nanotechnology 25 (2014) 125706
B-J Kim et al
the ratio of pore area to total image area. The Ag film thickness was measured by cutting cross-sections by FIB and analyzing the corresponding SEM images. 3. Results and discussion
For a suspension of metallic nanoparticles surrounded by an organic shell, the decomposition temperature (Td ) is one of the most critical factors for defining the microstructure and the resulting electrical/mechanical properties of the final metallic film. In order to optimize the heat treatment conditions for controlling the microstructure of the film, we first determined Td of our Ag ink by TGA as shown in figure 2(a). The mass of the sample decreased steadily up to 250 ◦ C and saturated at 20% during further heating up to 600 ◦ C, which is about the same as the initial Ag content of nanoparticle ink. This is attributed to the evaporation of the solvent and decomposition of the organic shells, which should occur simultaneously during continuous heating up to 250 ◦ C. The decomposition temperature, Td , of our ink was thereby determined to be 250 ◦ C. By considering Td , we applied four different heat treatment conditions: 150 ◦ C, 200 ◦ C, and 250 ◦ C for 30 min, and 350 ◦ C for 60 min to represent the conditions below, around, and above Td , respectively. Figure 2(b) shows the mass change as a function of annealing time when the maximum temperatures were set according to the heat treatments defined above. For maximum temperatures of 250 ◦ C and 350 ◦ C, the mass decreased rapidly to around 20% after 45 min and 35 min, respectively. The mass of the sample annealed at 200 ◦ C decreased at a lower rate. Here, it took about 150 min to reach 20% of the initial mass. The relative mass for the annealing at 150 ◦ C never reached 20%, but saturated at 40% even after 300 min. This already suggests that the annealing temperature needs to be above 200 ◦ C to ensure complete evaporation of the solvent and decomposition of the organic shells. Therefore, some remaining organics are expected for the samples with heat treatment at 150 and 200 ◦ C and possibly even at 250 ◦ C. The chemical composition at the surface of samples with different heat treatments was analyzed by XPS. The results are shown in figure 2(c). About 20% of carbon was detected in samples annealed at 150 and 200 ◦ C, and less than 20% was obtained in the sample annealed at 250 ◦ C. In contrast, no carbon was detected at the surface of the sample annealed at 350 ◦ C. This again indicates that some organics still remained due to the lower annealing temperature. Remaining organics generally deteriorate the electrical conductivity [13]. However, they could improve mechanical properties such as interfacial strength due to the carbon bridging effect [23]. Figures 3(a)–(d) show plane view and cross-sectional SEM images of the inkjet-printed Ag films for the four different annealing conditions. The microstructural evolution was quantified by SEM image analysis. The particle size was found to be distributed lognormally, and the median particle size, surface porosity, and film thickness are listed in table 1. The surface porosity was calculated by the ratio of pore area to total area except for the sample annealed at 150 ◦ C, because for this sample the pore size was too small to be detected.
Figure 2. (a) Relative mass change of the Ag nanoparticle-based ink as a function of temperature using a heating rate of 10 K min−1 .
(b) Mass change of the Ag ink as a function of annealing time when the temperature was held at 150, 200, 250, and 350 ◦ C. (c) XPS analysis of the chemical composition at the surface of the inkjet-printed Ag films.
For increasing annealing temperature, the particle size increased, and the porosity decreased, respectively. The microstructural evolution of inkjet-printed Ag films during annealing can be characterized as a sequential process of neck formation, particle coalescence, particle coarsening, and secondary grain growth [13]. The inkjet-printed Ag film annealed at 150 ◦ C showed the smallest median particle size (31.0 ± 12.2 nm), which was still larger than that of as-dried ink (about 10 nm). It is argued that in this sample the nanoparticles remained as isolated particles because the temperature was too low to decompose the organic shell and to initiate neck formation. In contrast, the particle size for the sample annealed at 200 ◦ C has increased (65.8 ± 15.4 nm) and neck formation between the nanoparticles was observed. Therefore, this sample is in the stage of neck formation and coalescence. For the sample annealed at 250 ◦ C, a further increase of particle size (145 ± 41.1 nm) and pore size was found, which 3
Nanotechnology 25 (2014) 125706
B-J Kim et al
Figure 3. Plane view and cross-sectional SEM images of the inkjet-printed Ag films after annealing at (a) 150 ◦ C, (b) 200 ◦ C, and (c) 250 ◦ C for 30 min, and (d) 350 ◦ C for 60 min. Table 1. Median grain size, surface porosity, and film thickness of inkjet-printed Ag films after different heat treatments.
Annealing condition
150 ◦ C, 30 min
200 ◦ C, 30 min
250 ◦ C, 30 min
350 ◦ C, 60 min
Median particle size (nm) Surface porosity (%) Film thickness (nm)
31.0 ± 12.2 — 800 ± 35.5
65.8 ± 15.4 9.6 755 ± 47.0
145 ± 41.1 6.6 780 ± 22.6
239 ± 83.1 1.1 780 ± 33.7
of pores—about 5 times smaller than the sample annealed at 150 ◦ C and only 17% larger than that of the sample annealed at high temperature (350 ◦ C). By applying a higher annealing temperature, the particle size further increases, and the porosity decreases. However, the effect on the resistivity is not as significant as the effect of neck formation and organic decomposition occurring at Td [13, 14, 17]. This implies that a sufficiently high conductivity can be achieved by optimizing the nanoporous microstructure. Figure 4(a) shows the electrical resistance change of the inkjet-printed and evaporated Ag as a function of the number of bending cycles using a bending strain of 1.1%. The evaporated Ag film showed poor fatigue properties; the electrical resistance increased to 110% after 5 × 105 cycles. The cross-sectional SEM image of this film shows typical fatigue damage of evaporated metallic thin films. Extrusions and intrusions were formed by irreversible dislocation motion during cyclic deformation, which leads to the crack formation shown for Cu [24, 25] and Ag thin films [26]. In contrast, inkjet-printed Ag films showed superior fatigue resistance compared to that of the evaporated film. The electrical resistivity of the samples annealed at 150, 200, and 250 ◦ C did not increase significantly from their initial values (1R/R0 < 5%) up to 5 × 105 cycles. Only the sample annealed at 350 ◦ C showed a resistivity increase of more than 80%. Figures 4(b)–(e) show SEM images after 5 × 105 bending cycles of all inkjet-printed films. All images were captured at 52◦ tilt, and the loading direction was horizontal. No damage was observed in the samples annealed at 150, 200, and 250 ◦ C. However, in the sample annealed at 350 ◦ C long cracks and extrusions were observed, which explains the strong increase in electrical resistivity. It is interesting that the more porous films (samples annealed at 150, 200, and 250 ◦ C) have a superior fatigue resistance compared to dense films, although the pores in metal films have been commonly regarded as a mechanically weak site due to the reduced cross-sectional area and stress concentration [16, 18, 19].
may be defined by the stage of coarsening. After annealing at 350 ◦ C for 60 min the sample showed an almost dense microstructure consisting of fairly large grains, which may be related to secondary grain growth. Concomitantly to particle coarsening and secondary grain growth, the porosity, as seen at the surface, decreased from 9.6% to 1.1% with increasing annealing temperature. In the cross-sectional images, similar trends of particle growth, neck formation and evolution of pore and overall porosity were observed. The film thickness of inkjet-printed films was reported to be either decreasing during high temperature annealing due to the densification of the films [15, 16], or not to change significantly [13, 23]. For our samples, the thicknesses varied between 755 ± 47.0 nm and 800 ± 35.5 nm, but no significant trend was observed. The electrical resistivity of inkjet-printed film was found to vary for the different annealing conditions corresponding to the described microstructural features. The electrical resistivities of the samples annealed at 150 ◦ C, 200 ◦ C, 250 ◦ C, and 350 ◦ C were 23 ± 0.9, 6.9 ± 0.6, 4.3 ± 0.2, and 3.7 ± 0.2 µ cm, respectively, while that of the evaporated Ag film with 800 nm film thickness was measured to be 3.3 ± 0.1 µ cm and that of bulk Ag is 1.6 µ cm. These values are comparable to those from other reports for the resistivity of solution based Ag films in the range of 2.64–12.5 µ cm [15, 18]. The high resistivity of the sample annealed at 150 ◦ C is attributed to the lack of a conducting path due to the isolated particles and remaining organic shell. The resistivity decreased for the sample annealed at 200 ◦ C mainly due to neck formation between the nanoparticles. With further increase of annealing temperature to 250 and 350 ◦ C, the resistivity decreased continuously to a level similar to that of the evaporated film by particle coarsening and secondary grain growth. Finally, it was only 10–30% higher than that of the evaporated Ag film. It is noteworthy that the sample annealed at 250 ◦ C shows an excellent low electrical resistivity despite the large portion 4
Nanotechnology 25 (2014) 125706
B-J Kim et al
Figure 4. (a) Electrical resistivity change of the inkjet-printed Ag films and the evaporated Ag film during a bending fatigue test. The bending strain was 1.1%. Errors imply the standard deviation for at least three samples. The top-right image is the cross-section of an evaporated Ag film after 5 × 105 cycles. The thickness of the evaporated Ag film was 800 nm, which is similar to those of the inkjet-printed films. ((b)–(e)) Surface morphologies of the inkjet-printed Ag films after 5 × 105 cycles after annealing at (b) 150 ◦ C for 30 min, (c) 200 ◦ C for 30 min, (d) 250 ◦ C for 30 min, and (e) 350 ◦ C for 60 min. All SEM images were captured at 52◦ tilt. (f) Schematic illustrations of fatigue behavior of inkjet-printed films with different microstructures.
increase. Nevertheless, extrusions and voids at the interface to the substrate are observed, from which cracks are nucleated. With decreasing film thickness, fewer and smaller extrusions are observed, and it can be argued that defects are annihilated at free surfaces or grain boundaries [24]. In our test, the uniformly distributed nanoscale pores may act as an annihilation site for single dislocations and thus prevent damage formation [20, 21]. Consequently, the formation of extrusions and subsequent cracking are reduced due to the nanoporous microstructure of the films. It should be noted that the stiffness of the printed thin film decreases with increasing porosity [28]. In the elastic regime, at a given strain, the low stiffness of the nanoporous films leads to lower film stresses, and the driving force for fatigue damage formation is reduced. However, the total applied strain of 1.1% is fairly large, and thus plastic deformation is still expected to occur. In contrast, many cracks and extrusions were observed in the samples annealed at 350 ◦ C. Initially, this film was dense and had the lowest electrical resistivity, but finally it showed the worst mechanical reliability due to extensive fatigue damage formation. Due to the higher annealing temperature, particle growth was strongly enhanced and a dense and coarse microstructure was obtained. Dislocation plasticity is further facilitated in the large grains, and dislocations are accumulated and form extrusions and intrusions as shown in figure 4(e). This finally leads to crack formation, and a similar fatigue behavior as for the evaporated Ag film is observed. Figure 5 shows the electrical conductivity and fatigue lifetime of the evaporated film and the inkjet-printed films with
The improvement of fatigue resistance in the porous films is mainly attributed to the small particle size and uniform distribution of the nanopores. The schematic diagram in figure 4(f) shows the fatigue behaviors of inkjet-printed films with different microstructures. After annealing at low temperatures such as 150 ◦ C, the microstructure of the films still consists of small isolated metallic particles surrounded by organic shell. Dislocation plasticity is restricted within the volume of the small particles, as dislocations annihilated at the surface of the particles, leading to a dislocation-starved condition [27]. Furthermore, the remaining organic shell may prevent the formation of extrusions and intrusions. Thereby the nanoparticulate film is more resistant to fatigue damage, but it has the inherent drawback of high initial electrical resistivity due to the absence of neck formation between the particles and the remaining organic shell. In the samples annealed at 200 and 250 ◦ C, the neck formation and particle coarsening has occurred, while some organics may still have remained. Due to the larger particle size, the activation stress for dislocation plasticity is reduced compared to the sample annealed at 150 ◦ C. Nevertheless, the particle size is still so small that the formation of long-range ordered dislocation structures such as cells, veins, or even persistent slip bands is not expected [25]. It is well known that in fatigue bulk metals, fatigue damage such as extrusions and intrusions is formed where persistent slip bands intersect with the surface. In contrast, in metallic thin films with thickness below 1 µm, such dislocation structures are no longer observed [24–26], and the fatigue resistance tends to 5
Nanotechnology 25 (2014) 125706
B-J Kim et al
Acknowledgments
This work was supported by the Global Leading Technology Program of the Office of Strategic R&D Planning (OSP) funded by the Korean government (Ministry of Trade, Industry and Energy) (No. 10042537). Oliver Kraft is grateful to the Robert Bosch Foundation for supporting his Endowed Chair for Functional Nanomaterials, and Byoung-Joon Kim gratefully acknowledges the KIT guest program. References [1] Forrest S R 2004 Nature 428 911 [2] Koo M, Park K I, Lee S H, Jeon D Y, Choi J W, Kang K and Lee K J 2012 Nano Lett. 12 4810 [3] Nam K T, Kim D W, Yoo P J, Meethong N, Hammond P T, Chiang Y M and Belcher A M 2006 Science 312 885 [4] Li Y, Lee D K, Kim J Y, Kim B, Park N G, Kim K, Shin J H, Choi I S and Ko M J 2012 Energy Environ. Sci. 5 8950 [5] Yeo W H et al 2013 Adv. Mater. 25 2773 [6] Ko S H, Pan H, Grigoropoulos C P, Luscombe C K, Fr´echet J M J and Poulikakos D 2007 Nanotechnology 18 345202 [7] Van Osch T H J, Perelaer J, De Laat A W M and Schubert U S 2008 Adv. Mater. 20 343 [8] Lee H M, Choi S Y, Kim K T, Yun J Y, Jung D S, Park S B and Park J 2011 Adv. Mater. 23 5524 [9] Lee H M, Choi S Y, Jung A and Ko S H 2013 Angew. Chem. Int. Edn. 52 7718 [10] Lee H H, Chou K S and Huang K C 2005 Nanotechnology 16 2436 [11] Jeong S, Kim D, Lee S, Park B K and Moon J 2006 Mol. Cryst. Liq. Cryst. 459 35 [12] Park J W and Baek S G 2006 Scr. Mater. 55 1139 [13] Jung J K, Choi S H, Kim I, Jung H C, Joung J and Joo Y C 2008 Phil. Mag. 88 339 [14] Greer J R and Street R A 2007 Acta Mater. 55 6345 [15] Kim N R, Lee J H, Yi S M and Joo Y C 2011 J. Electrochem. Soc. 158 K165 [16] Lee J H, Kim N R, Kim B J and Joo Y C 2012 Carbon 50 98 [17] Yi S M, Lee J H, Kim N R, Oh S, Jang S, Kim D, Joung J and Joo Y C 2010 J. Electrochem. Soc. 157 K254 [18] Kim S, Won S, Sim G D, Park I and Lee S B 2013 Nanotechnology 24 085701 [19] Sim G D, Won S and Lee S B 2012 Appl. Phys. Lett. 101 191907 [20] Lee H Y, Yi S M, Lee J H, Lee H S, Hyun S and Joo Y C 2010 Met. Mater. Int. 6 947 [21] Kim B J, Cho Y, Jung M S, Shin H A S, Moon M W, Han H N, Nam K T, Joo Y C and Choi I S 2012 Small 8 3300 [22] Kim B J, Shin H A S, Jung S Y, Cho Y, Kraft O, Choi I S and Joo Y C 2013 Acta Mater. 61 3473 [23] Lee I, Kim S, Yun J, Park I and Kim T S 2012 Nanotechnology 23 485704 [24] Schwaiger R, Dehm G and Kraft O 2003 Phil. Mag. 83 693 [25] Zhang G P, Volkert C A, Schwaiger R, Wellner P, Arzt E and Kraft O 2006 Acta Mater. 54 3127 [26] Schwaiger R and Kraft O 2003 Acta Mater. 51 195 [27] Julia R G and William D N 2006 Phys. Rev. B 73 245410 [28] Sanders P G, Eastman J A and Weertman J R 1997 Acta Mater. 45 4019
Figure 5. Electrical conductivities and fatigue lifetimes of the inkjet-printed Ag films and the evaporated Ag films. The fatigue lifetime was defined as the cycle number at which a 20% increase in resistivity was observed.
different heat treatments. The fatigue lifetime was defined as the cycle number at which a 20% increase in the resistivity of the individual sample was observed. In the case of no fatigue damage up to 5 × 105 cycles, fatigue lifetime was plotted at 5 × 105 cycles and open symbols were used. The electrical conductivity of the inkjet-printed films increases with increasing annealing temperature and is highest for the highest annealing temperature. However, the fatigue lifetime of this sample is deteriorated due to fatigue damage formation as discussed above. This indicates that an optimum compromise of high conductivity and long-term reliability of flexible metallic electrodes can be achieved by establishing a nanoporous microstructure of the material using an inkjet printing process with nanoparticulate suspensions and appropriate annealing conditions. 4. Conclusions
The fatigue behavior and electrical performance of inkjetprinted Ag films with different nanoporous microstructures were investigated using a bending fatigue test system. Samples annealed at temperatures below the decomposition temperature Td of the organic shells of the nanoparticulate Ag ink (Td = 250 ◦ C) showed superior fatigue resistance, but had a high initial resistivity due to the absence of neck formation between the particles. Samples annealed at high temperature above Td showed a low initial resistivity due to the dense and coarse microstructure, but the fatigue lifetime is deteriorated due to the early formation of extrusions and cracks. By annealing around Td , a nanoporous microstructure is achieved, having not only high electrical conductivity but also beneficial fatigue resistance. The electrical conductivity is obtained due to neck formation within the particle network, whereas the enhanced fatigue resistance results from hindered dislocation plasticity due to the small particle size and dislocation annihilation at free surfaces. This systematic study of the effect of nanopores on electrical and mechanical properties can provide a helpful solution to develop highly conductive and reliable flexible metallic electrodes using inkjet printing. 6
