- Email: [email protected]
Microstructure and electrical properties of high power laser thermal annealing on inkjet-printed Ag films
Microelectronic Engineering 87 (2010) 2230–2233
Contents lists available at ScienceDirect
Microelectronic Engineering journal homepage: www.elsevier.com/locate/mee
Microstructure and electrical properties of high power laser thermal annealing on inkjet-printed Ag films Yo Han Yoon a,b, Seol-Min Yi a, Jung-Ryoul Yim a, Ji-Hoon Lee a, George Rozgonyi b, Young-Chang Joo a,* a b
Department of Materials Science and Engineering, Seoul National University, Seoul 151-744, Republic of Korea Department of Materials Science and Engineering, North Carolina State University, Raleigh, NC 27695-7907, USA
a r t i c l e
i n f o
Article history: Received 23 April 2009 Received in revised form 2 February 2010 Accepted 15 February 2010 Available online 19 February 2010 Keywords: CW Nd:YAG laser Furnace annealing Inkjet printing Thermal diffusion Resistivity
a b s t r a c t In this work, a high power continuous-wave (CW) Nd:YAG laser was used for thermal treatment of inkjetprinted Ag films – resulting in the elimination of organic additives (dispersant, binder, and organic solvent) in the Ag ink and annealing of Ag nano-particles. By optimizing laser parameters such as laser power and defocusing value, the laser energy can be totally converted into heat energy, which is used for thermal treatment of inkjet-printed Ag films. This results in the microstructure and the resistivity of the films to be controlled. We investigated the thermal diffusion mechanisms during laser annealing and the resulting microstructures. The impact of high power laser annealing on microstructures and electrical characteristics of inkjet-printed Ag films was compared to those of the films annealed by a conventional furnace annealing. Focused ion beam (FIB) channeling images show that the laser annealed Ag films have large columnar grains and a dense void-free structure, while furnace annealed films have much smaller grains and exhibit void formation. As a result, the laser annealed films have better electrical properties (low resistivity) compared to furnace annealed samples. Ó 2010 Published by Elsevier B.V.
1. Introduction Traditional thin film technology, comprised of physical vapor deposition and lithographic processes for patterning, has achieved feature resolutions in the nanometer (nm) range. However, following the introduction of advanced device materials, flexible large-area electronic products additionally demand simple fabrication processes that reduce cost by decreasing the number of fabrication steps [1]. A promising approach to reduce processing steps is to directly print the device materials, thereby eliminating many steps such as lithography and etching [2]. Inkjet printing, microcontact printing, and screen printing have attracted considerable attention as highly efficient methods for the fabrication of electronic products. Among these printings, pattern-on-demand technology in the form of inkjet printing has demonstrated a rapid growth during the last decade [3,4]. The ink used in inkjet printing is essentially composed of two parts: organic additives (dispersant, binder, and organic solvent) and nano-particles. Organic additives are stabilizers for nanoparticles based ink, but act as inhibitors for grain growth of nano-particles. Thus, drying processes for these organic additives and post annealing procedures for nano-particle grain growth are needed to optimize the electrical performance of inkjet* Corresponding author. E-mail addresses: [email protected] (Y.H. Yoon), [email protected] (Y.-C. Joo). 0167-9317/$ - see front matter Ó 2010 Published by Elsevier B.V. doi:10.1016/j.mee.2010.02.008
printed films, as recent studies demonstrated [5,6] for thermally sintered silver conductors whose conductivities approached half of that for bulk materials. Although traditional furnace annealing is generally used in the post annealing process, it is time-consuming and does not allow for locally annealing of a specific area of a device, such as an inkjetted pattern. Furthermore, the furnace annealed films typically have a microstructure consisted of tiny grains and pores. The microstructure of films is obviously an important determinant of the film resistivity and mechanical properties and to achieve high performance. Laser annealing is a possible solution to this drawback since a high power laser can deliver heat to the inkjet-printed films up to the melting point, thus providing its desired microstructure (large grain size, void-free dense structure). In this paper, a Nd:YAG continuous-wave (CW) laser was used for annealing inkjet-printed films. Excimer lasers are widely used in semiconductor and thin film applications, but their energy density (several milli-joule) is generally too low to evaporate organic additives and to anneal silver nano-particles. We investigated the microstructure of 1.5 lm thick inkjet-printed Ag films annealed by a Nd:YAG CW laser. This wavelength (1064 nm) of Nd:YAG is suitable for the reaction with silver nano-particles. In particular, we examined the use of different laser energy densities to improve the film quality and correlated the microstructure of the Ag films with the electrical resistivity.
Y.H. Yoon et al. / Microelectronic Engineering 87 (2010) 2230–2233
2. Experimental procedure Ag nano-particle suspension with an average Ag particle size between 10 and 15 nm was used for inkjet printing. The inkjet printing was conducted on 4-inch Si(1 0 0) wafers, 500-lm-thick. Inkjet Ag films were fabricated on a Dimatix materials printer (DMP-2800). Substrate and ink temperatures were 30 °C and 50 °C, respectively. All the inkjet-printed metal structures we used had a pad shape with dimension 5 5 mm2. The films as jetted were dried in a conventional convection oven at 100 °C for 60 min under ambient air prior to irradiation with CW Nd:YAG laser (k = 1064 nm, maximum power: 500 W). The maximum laser power density was 1.5 106 W/cm2, which was adjusted using various defocusing values for a 1 s irradiation time. Surface morphology/microstructure was obtained by using a field emission scanning electron microscope (FE-SEM) (Hitachi S48000), while cross-sectional images examined by a focused ion beam (FIB) (SMI3050SE). An image analyzer with lognormal distribution fitting was used to measure the planar and cross-sectional grain size, while the sheet resistance was obtained with a fourpoint-probe. 3. Results and discussion The concept of laser annealing for inkjet-printed films relies on two parts: free from surface damage and sufficient laser power density for melting inkjet-printed films. The laser parameters should be precisely controlled to anneal only inkjet-printed Ag films, not to evaporate or delaminate them. If it is optimized, in molten films organic, additives and pores can migrate to the surface of films and removed. Therefore, appraisal of the laser parameters and those of their interactions are critical to understand the dynamics of the organic additives and pores and to optimize the resulting laser annealed Ag microstructure. 3.1. Laser parameter optimization Among the various laser parameters, we first demonstrated that the surface damage of the laser annealed film was entirely dependent on the degree of laser defocusing and laser power which determines the laser power density. For constant laser output energy, the crater diameters of a defocused laser are larger than a focused laser [7]. Fig. 1 shows the relationship between the average grain size of laser annealed Ag nano-particles and the degree of defocusing (in mm) from the best focal plane (0 mm) and laser power. At defocusing range within 35 mm, little change of the
Fig. 1. Average grain size of silver nano-particle with laser power (350, 400, and 500 W) and defocusing value.
2231
average grain size is observed regardless of laser output power. In this case, the irradiated film area is locally evaporated or delaminated from the silicon substrate, due to the excessive laser power density. If the laser power density is over the critical value, the excess laser energy will evaporate Ag nano-particles and organic additives instead of thermal annealing the Ag nano-particles. Note that the surface damage is extremely reduced, and grain growth of Ag nano-particles occurs when the defocusing value is within the range of 38 mm. In this situation, all of the laser energy is converted into heat energy. However, when the defocusing value is larger than the optimum point (38–40 mm), the average grain size decreases with increasing defocusing since the laser power density is then insufficient to evaporate organic additives and to anneal the Ag nano-particles. The correlation between the laser power density and average grain size is discussed below. 3.2. Microstructure of inkjet-printed Ag film Fig. 2 shows SEM and FIB images of as-dried, see Fig. 2(a), and annealed samples, see Fig. 2(b–d). Fig. 2 compares SEM and FIB channeling images of conventional furnace annealed (250 °C for 60 min under air) inkjet-printed Ag films, see Fig. 2(b), with laser annealed (the laser power densities of (c and d) are 5.67 105 W/cm2 and 7.09 105 W/cm2, respectively. Lasing condition is 38 mm (defocusing), 1 s (duration time)) samples, see Fig. 2(c and d). Surface grains of conventional furnace annealed inkjet-printed Ag films are sufficiently annealed like the laser annealed samples. Note that FIB image of Fig. 2(b) shows many tiny nano-particles, which not completely sintered and with pores scattered throughout the films. This is attributed to the two step densification process that occurs with furnace annealing; namely organic additives are removed with oxygen and then grain growth occurs. At the very beginning during furnace annealing process, oxygen in the air reacts initially with organic additives existed at the film surface. Therefore, organics decomposition and grain growth of Ag nano-particles starts at the film surface, thus the dense film surface blocks the path that brings the reaction between organic additives in the film and the oxygen in the air. Fig. 2(c and d) correspond to regions directly annealed by the incident laser beam at power densities of 5.67 105 W/cm2 (the laser power is 400 W) and 7.09 105 W/cm2 (the laser power is 500 W), respectively, Depending on the lasing conditions, microstructural characteristics result to be entirely different, as evident in the FIB cross-sectional view images. Though the SEM determined surface average grain sizes are similar, an extremely sharp increase in the grain size through the thickness occurs with a small increase in the laser power density. Such remarkable differences of microstructure evolution reveal that high laser energy is used to evaporate organic additives, which are capped by Ag nano-particles, and anneal the Ag nano-particles. Due to these large columnar grains and their dense structures, it is believed that the direct laser annealed zone undergoes melting. In this condition, the dense microstructures are free of organic additives formed in the films though the large pores are partially scattered. Deposited material, pore and grain size play a critical role in the densification process during sintering. In the lower level of laser power density, organic additives cannot be entirely eliminated from the film due to the lack of melting. Those organic additives which still remain in grains and at grain boundaries restrict Ag film grain growth; thus, tiny equi-axial grains are formed throughout the film thickness [Fig. 2(c)]. Another significant observation is the mechanism of thermal diffusion from laser irradiation which leads to a heat affected zone (HAZ) near the laser beam irradiated area. HAZ is the area, surrounding the melted region whose microstructure is thermally affected by the heat input from the lasing process. Note that the
2232
Y.H. Yoon et al. / Microelectronic Engineering 87 (2010) 2230–2233
Fig. 2. SEM images (in-plane view) and FIB images (cross-section view) of as-dried (a), furnace annealed (b) and laser annealed samples (c and d). Furnace annealing condition of sample (b) is 250 °C for 60 min under air. The laser power densities of (c) and (d) are 5.67 105 W/cm2 and 7.09 105 W/cm2, respectively. Lasing condition is 38 mm (defocusing), 1 s (duration time).
Fig. 3. SEM image (a) and FIB channeling image (b) for heat affected zone (HAZ) of laser annealed sample. The laser power density is 7.09 105 W/cm2. Drying condition of all specimens is 100 °C for 60 min and lasing condition is 38 mm (defocusing), 1 s (duration time).
duration time of the laser pulse is ultrashort (femtosecond), and the penetration length of thermal diffusion in the material is very limited, leading to a very small HAZ [8]. In this study, however, laser duration time is relatively long (1 s) and the laser power density is sufficient to melt the film, thus HAZ formation is inevitable. Fig. 3 shows the HAZ in the laser annealed sample (the laser power density is 7.09 105 W/cm2 and lasing condition is 38 mm (defocusing), 1 s (duration time)). In inkjet-printed Ag films, organic additives capped by silver nano-particles and the point contact between silver nano-particles act as diffusion barriers, enabling thermal diffusion to occur through the silicon substrate. Fig. 3 shows the microstructure of HAZ formed by this thermal diffusion mechanism. As the heat transfer direction is from silicon substrate to films, the microstructure is denser at the bottom of the films. This is because organic additives migrate to the film surface and only partially decompose since the amount of heat in the HAZ is insufficient.
power density. All of these data were obtained at optimum defocusing (38 mm). As shown in Fig. 4, the film resistivity is reduced, while the average planar grain size increases with laser power density. Note that the average planar grain size is saturated for the laser power densities higher than a critical value of 5.67 105 W/
3.3. Electrical properties of inkjet-printed Ag films The microstructures mentioned in Section 3.2 affect the resistivity of laser annealed films. In case of as-dried sample [Fig. 2(a)], its resistivity is several hundreds lX cm because of the remaining organic additives and nano-size Ag particles. However, following a second lasing process makes the resistivity decrease to 3 lX cm. Fig. 4 shows the average planar grain sizes and the resistivity of laser annealed films as a function of laser
Fig. 4. Average grain size and resistivity as a function of laser power density. Drying condition of all specimens is 100 °C for 60 min and lasing condition is 38 mm (defocusing), 1 s (duration time).
Y.H. Yoon et al. / Microelectronic Engineering 87 (2010) 2230–2233
cm2 and also a sharp decrease in resistivity occurs at this point. This is directly related to the occurrence of a non-melted or melted condition. In case of non-melted condition, see Fig. 2(b), the equiaxial tiny grains are distributed through the thickness of the film, producing a planar grain size (688 nm) essentially the same as the overall grain size (683 nm). However, following melting, see Fig. 2(d), the planar grains remain unaltered, while large columnar grains are present throughout the film thickness, producing a sharp decrease in resistivity. Also furnace annealed films with large pore density and small grains produce unsatisfactory values of resistivity (5.32 lX cm) compared to the values obtained in laser annealed films.
4. Conclusions In this work, inkjet-printed Ag films were annealed by a continuous-wave Nd:YAG laser. The surface morphology of the laser annealed films was strongly dependent on the laser parameters – defocusing value and laser power. For an optimum defocusing value the laser power is totally converted to heat energy, the surface damage of laser annealed films is minimized, and the average grain size of Ag nano-particles increases with increasing laser power. An important characteristic of the laser power density applied to the inkjet-printed films is the occurrence of entirely different crosssectional microstructures due to slight differences of laser power density. With a small increase in the laser power density, an extre-
2233
mely sharp increase in the grain size through the thickness occurs though the surface average grain size of films is similar. The major results of this work on the impact of laser annealing versus conventional furnace annealing on the microstructure and electrical properties of inkjet-printed films are that laser annealed films are void-free, have significantly large columnar grains and lower resistivity than furnace annealed films. Acknowledgment This research was partially supported by HANA Engineering, which is providing the technical support on the basis of the optimized laser system, in Korea. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (0417-20090028). References [1] C.M. Hong, S. Wagner, IEEE Electron Device Lett. 21 (2000) 384–386. [2] H. Gleskova, S. Wagner, D.S. Shen, J. Non-Cryst. Solids 227–230 (1998) 1217– 1220. [3] B.-J. de Gans, P.C. Duineveld, U.S. Schubert, Adv. Mater. 16 (2004) 203–213. [4] O.A. Basaran, AIChE J. 48 (2002) 1842–1848. [5] J.K. Jung, S.H. Choi, I. Kim, H.C. Jung, J. Joung, Y.C. Joo, Philos. Mag. 88 (3) (2008) 339–359. [6] D. Kim, J. Moon, Electrochem. Solid-State Lett. 8 (2005) J30–J33. [7] Masaki Ohata, Yoshihiro Iwasaki, Naoki Furuta, Isaac B. Brenner, Spectrochim. Acta, Part B 57 (2002) 1713–1725. [8] R. Le Harzic, N. Huot, E. Audouard, C. Jonin, P. Laporte, Appl. Phys. Lett. 80 (2002) 3886–3888.
Contents lists available at ScienceDirect
Microelectronic Engineering journal homepage: www.elsevier.com/locate/mee
Microstructure and electrical properties of high power laser thermal annealing on inkjet-printed Ag films Yo Han Yoon a,b, Seol-Min Yi a, Jung-Ryoul Yim a, Ji-Hoon Lee a, George Rozgonyi b, Young-Chang Joo a,* a b
Department of Materials Science and Engineering, Seoul National University, Seoul 151-744, Republic of Korea Department of Materials Science and Engineering, North Carolina State University, Raleigh, NC 27695-7907, USA
a r t i c l e
i n f o
Article history: Received 23 April 2009 Received in revised form 2 February 2010 Accepted 15 February 2010 Available online 19 February 2010 Keywords: CW Nd:YAG laser Furnace annealing Inkjet printing Thermal diffusion Resistivity
a b s t r a c t In this work, a high power continuous-wave (CW) Nd:YAG laser was used for thermal treatment of inkjetprinted Ag films – resulting in the elimination of organic additives (dispersant, binder, and organic solvent) in the Ag ink and annealing of Ag nano-particles. By optimizing laser parameters such as laser power and defocusing value, the laser energy can be totally converted into heat energy, which is used for thermal treatment of inkjet-printed Ag films. This results in the microstructure and the resistivity of the films to be controlled. We investigated the thermal diffusion mechanisms during laser annealing and the resulting microstructures. The impact of high power laser annealing on microstructures and electrical characteristics of inkjet-printed Ag films was compared to those of the films annealed by a conventional furnace annealing. Focused ion beam (FIB) channeling images show that the laser annealed Ag films have large columnar grains and a dense void-free structure, while furnace annealed films have much smaller grains and exhibit void formation. As a result, the laser annealed films have better electrical properties (low resistivity) compared to furnace annealed samples. Ó 2010 Published by Elsevier B.V.
1. Introduction Traditional thin film technology, comprised of physical vapor deposition and lithographic processes for patterning, has achieved feature resolutions in the nanometer (nm) range. However, following the introduction of advanced device materials, flexible large-area electronic products additionally demand simple fabrication processes that reduce cost by decreasing the number of fabrication steps [1]. A promising approach to reduce processing steps is to directly print the device materials, thereby eliminating many steps such as lithography and etching [2]. Inkjet printing, microcontact printing, and screen printing have attracted considerable attention as highly efficient methods for the fabrication of electronic products. Among these printings, pattern-on-demand technology in the form of inkjet printing has demonstrated a rapid growth during the last decade [3,4]. The ink used in inkjet printing is essentially composed of two parts: organic additives (dispersant, binder, and organic solvent) and nano-particles. Organic additives are stabilizers for nanoparticles based ink, but act as inhibitors for grain growth of nano-particles. Thus, drying processes for these organic additives and post annealing procedures for nano-particle grain growth are needed to optimize the electrical performance of inkjet* Corresponding author. E-mail addresses: [email protected] (Y.H. Yoon), [email protected] (Y.-C. Joo). 0167-9317/$ - see front matter Ó 2010 Published by Elsevier B.V. doi:10.1016/j.mee.2010.02.008
printed films, as recent studies demonstrated [5,6] for thermally sintered silver conductors whose conductivities approached half of that for bulk materials. Although traditional furnace annealing is generally used in the post annealing process, it is time-consuming and does not allow for locally annealing of a specific area of a device, such as an inkjetted pattern. Furthermore, the furnace annealed films typically have a microstructure consisted of tiny grains and pores. The microstructure of films is obviously an important determinant of the film resistivity and mechanical properties and to achieve high performance. Laser annealing is a possible solution to this drawback since a high power laser can deliver heat to the inkjet-printed films up to the melting point, thus providing its desired microstructure (large grain size, void-free dense structure). In this paper, a Nd:YAG continuous-wave (CW) laser was used for annealing inkjet-printed films. Excimer lasers are widely used in semiconductor and thin film applications, but their energy density (several milli-joule) is generally too low to evaporate organic additives and to anneal silver nano-particles. We investigated the microstructure of 1.5 lm thick inkjet-printed Ag films annealed by a Nd:YAG CW laser. This wavelength (1064 nm) of Nd:YAG is suitable for the reaction with silver nano-particles. In particular, we examined the use of different laser energy densities to improve the film quality and correlated the microstructure of the Ag films with the electrical resistivity.
Y.H. Yoon et al. / Microelectronic Engineering 87 (2010) 2230–2233
2. Experimental procedure Ag nano-particle suspension with an average Ag particle size between 10 and 15 nm was used for inkjet printing. The inkjet printing was conducted on 4-inch Si(1 0 0) wafers, 500-lm-thick. Inkjet Ag films were fabricated on a Dimatix materials printer (DMP-2800). Substrate and ink temperatures were 30 °C and 50 °C, respectively. All the inkjet-printed metal structures we used had a pad shape with dimension 5 5 mm2. The films as jetted were dried in a conventional convection oven at 100 °C for 60 min under ambient air prior to irradiation with CW Nd:YAG laser (k = 1064 nm, maximum power: 500 W). The maximum laser power density was 1.5 106 W/cm2, which was adjusted using various defocusing values for a 1 s irradiation time. Surface morphology/microstructure was obtained by using a field emission scanning electron microscope (FE-SEM) (Hitachi S48000), while cross-sectional images examined by a focused ion beam (FIB) (SMI3050SE). An image analyzer with lognormal distribution fitting was used to measure the planar and cross-sectional grain size, while the sheet resistance was obtained with a fourpoint-probe. 3. Results and discussion The concept of laser annealing for inkjet-printed films relies on two parts: free from surface damage and sufficient laser power density for melting inkjet-printed films. The laser parameters should be precisely controlled to anneal only inkjet-printed Ag films, not to evaporate or delaminate them. If it is optimized, in molten films organic, additives and pores can migrate to the surface of films and removed. Therefore, appraisal of the laser parameters and those of their interactions are critical to understand the dynamics of the organic additives and pores and to optimize the resulting laser annealed Ag microstructure. 3.1. Laser parameter optimization Among the various laser parameters, we first demonstrated that the surface damage of the laser annealed film was entirely dependent on the degree of laser defocusing and laser power which determines the laser power density. For constant laser output energy, the crater diameters of a defocused laser are larger than a focused laser [7]. Fig. 1 shows the relationship between the average grain size of laser annealed Ag nano-particles and the degree of defocusing (in mm) from the best focal plane (0 mm) and laser power. At defocusing range within 35 mm, little change of the
Fig. 1. Average grain size of silver nano-particle with laser power (350, 400, and 500 W) and defocusing value.
2231
average grain size is observed regardless of laser output power. In this case, the irradiated film area is locally evaporated or delaminated from the silicon substrate, due to the excessive laser power density. If the laser power density is over the critical value, the excess laser energy will evaporate Ag nano-particles and organic additives instead of thermal annealing the Ag nano-particles. Note that the surface damage is extremely reduced, and grain growth of Ag nano-particles occurs when the defocusing value is within the range of 38 mm. In this situation, all of the laser energy is converted into heat energy. However, when the defocusing value is larger than the optimum point (38–40 mm), the average grain size decreases with increasing defocusing since the laser power density is then insufficient to evaporate organic additives and to anneal the Ag nano-particles. The correlation between the laser power density and average grain size is discussed below. 3.2. Microstructure of inkjet-printed Ag film Fig. 2 shows SEM and FIB images of as-dried, see Fig. 2(a), and annealed samples, see Fig. 2(b–d). Fig. 2 compares SEM and FIB channeling images of conventional furnace annealed (250 °C for 60 min under air) inkjet-printed Ag films, see Fig. 2(b), with laser annealed (the laser power densities of (c and d) are 5.67 105 W/cm2 and 7.09 105 W/cm2, respectively. Lasing condition is 38 mm (defocusing), 1 s (duration time)) samples, see Fig. 2(c and d). Surface grains of conventional furnace annealed inkjet-printed Ag films are sufficiently annealed like the laser annealed samples. Note that FIB image of Fig. 2(b) shows many tiny nano-particles, which not completely sintered and with pores scattered throughout the films. This is attributed to the two step densification process that occurs with furnace annealing; namely organic additives are removed with oxygen and then grain growth occurs. At the very beginning during furnace annealing process, oxygen in the air reacts initially with organic additives existed at the film surface. Therefore, organics decomposition and grain growth of Ag nano-particles starts at the film surface, thus the dense film surface blocks the path that brings the reaction between organic additives in the film and the oxygen in the air. Fig. 2(c and d) correspond to regions directly annealed by the incident laser beam at power densities of 5.67 105 W/cm2 (the laser power is 400 W) and 7.09 105 W/cm2 (the laser power is 500 W), respectively, Depending on the lasing conditions, microstructural characteristics result to be entirely different, as evident in the FIB cross-sectional view images. Though the SEM determined surface average grain sizes are similar, an extremely sharp increase in the grain size through the thickness occurs with a small increase in the laser power density. Such remarkable differences of microstructure evolution reveal that high laser energy is used to evaporate organic additives, which are capped by Ag nano-particles, and anneal the Ag nano-particles. Due to these large columnar grains and their dense structures, it is believed that the direct laser annealed zone undergoes melting. In this condition, the dense microstructures are free of organic additives formed in the films though the large pores are partially scattered. Deposited material, pore and grain size play a critical role in the densification process during sintering. In the lower level of laser power density, organic additives cannot be entirely eliminated from the film due to the lack of melting. Those organic additives which still remain in grains and at grain boundaries restrict Ag film grain growth; thus, tiny equi-axial grains are formed throughout the film thickness [Fig. 2(c)]. Another significant observation is the mechanism of thermal diffusion from laser irradiation which leads to a heat affected zone (HAZ) near the laser beam irradiated area. HAZ is the area, surrounding the melted region whose microstructure is thermally affected by the heat input from the lasing process. Note that the
2232
Y.H. Yoon et al. / Microelectronic Engineering 87 (2010) 2230–2233
Fig. 2. SEM images (in-plane view) and FIB images (cross-section view) of as-dried (a), furnace annealed (b) and laser annealed samples (c and d). Furnace annealing condition of sample (b) is 250 °C for 60 min under air. The laser power densities of (c) and (d) are 5.67 105 W/cm2 and 7.09 105 W/cm2, respectively. Lasing condition is 38 mm (defocusing), 1 s (duration time).
Fig. 3. SEM image (a) and FIB channeling image (b) for heat affected zone (HAZ) of laser annealed sample. The laser power density is 7.09 105 W/cm2. Drying condition of all specimens is 100 °C for 60 min and lasing condition is 38 mm (defocusing), 1 s (duration time).
duration time of the laser pulse is ultrashort (femtosecond), and the penetration length of thermal diffusion in the material is very limited, leading to a very small HAZ [8]. In this study, however, laser duration time is relatively long (1 s) and the laser power density is sufficient to melt the film, thus HAZ formation is inevitable. Fig. 3 shows the HAZ in the laser annealed sample (the laser power density is 7.09 105 W/cm2 and lasing condition is 38 mm (defocusing), 1 s (duration time)). In inkjet-printed Ag films, organic additives capped by silver nano-particles and the point contact between silver nano-particles act as diffusion barriers, enabling thermal diffusion to occur through the silicon substrate. Fig. 3 shows the microstructure of HAZ formed by this thermal diffusion mechanism. As the heat transfer direction is from silicon substrate to films, the microstructure is denser at the bottom of the films. This is because organic additives migrate to the film surface and only partially decompose since the amount of heat in the HAZ is insufficient.
power density. All of these data were obtained at optimum defocusing (38 mm). As shown in Fig. 4, the film resistivity is reduced, while the average planar grain size increases with laser power density. Note that the average planar grain size is saturated for the laser power densities higher than a critical value of 5.67 105 W/
3.3. Electrical properties of inkjet-printed Ag films The microstructures mentioned in Section 3.2 affect the resistivity of laser annealed films. In case of as-dried sample [Fig. 2(a)], its resistivity is several hundreds lX cm because of the remaining organic additives and nano-size Ag particles. However, following a second lasing process makes the resistivity decrease to 3 lX cm. Fig. 4 shows the average planar grain sizes and the resistivity of laser annealed films as a function of laser
Fig. 4. Average grain size and resistivity as a function of laser power density. Drying condition of all specimens is 100 °C for 60 min and lasing condition is 38 mm (defocusing), 1 s (duration time).
Y.H. Yoon et al. / Microelectronic Engineering 87 (2010) 2230–2233
cm2 and also a sharp decrease in resistivity occurs at this point. This is directly related to the occurrence of a non-melted or melted condition. In case of non-melted condition, see Fig. 2(b), the equiaxial tiny grains are distributed through the thickness of the film, producing a planar grain size (688 nm) essentially the same as the overall grain size (683 nm). However, following melting, see Fig. 2(d), the planar grains remain unaltered, while large columnar grains are present throughout the film thickness, producing a sharp decrease in resistivity. Also furnace annealed films with large pore density and small grains produce unsatisfactory values of resistivity (5.32 lX cm) compared to the values obtained in laser annealed films.
4. Conclusions In this work, inkjet-printed Ag films were annealed by a continuous-wave Nd:YAG laser. The surface morphology of the laser annealed films was strongly dependent on the laser parameters – defocusing value and laser power. For an optimum defocusing value the laser power is totally converted to heat energy, the surface damage of laser annealed films is minimized, and the average grain size of Ag nano-particles increases with increasing laser power. An important characteristic of the laser power density applied to the inkjet-printed films is the occurrence of entirely different crosssectional microstructures due to slight differences of laser power density. With a small increase in the laser power density, an extre-
2233
mely sharp increase in the grain size through the thickness occurs though the surface average grain size of films is similar. The major results of this work on the impact of laser annealing versus conventional furnace annealing on the microstructure and electrical properties of inkjet-printed films are that laser annealed films are void-free, have significantly large columnar grains and lower resistivity than furnace annealed films. Acknowledgment This research was partially supported by HANA Engineering, which is providing the technical support on the basis of the optimized laser system, in Korea. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (0417-20090028). References [1] C.M. Hong, S. Wagner, IEEE Electron Device Lett. 21 (2000) 384–386. [2] H. Gleskova, S. Wagner, D.S. Shen, J. Non-Cryst. Solids 227–230 (1998) 1217– 1220. [3] B.-J. de Gans, P.C. Duineveld, U.S. Schubert, Adv. Mater. 16 (2004) 203–213. [4] O.A. Basaran, AIChE J. 48 (2002) 1842–1848. [5] J.K. Jung, S.H. Choi, I. Kim, H.C. Jung, J. Joung, Y.C. Joo, Philos. Mag. 88 (3) (2008) 339–359. [6] D. Kim, J. Moon, Electrochem. Solid-State Lett. 8 (2005) J30–J33. [7] Masaki Ohata, Yoshihiro Iwasaki, Naoki Furuta, Isaac B. Brenner, Spectrochim. Acta, Part B 57 (2002) 1713–1725. [8] R. Le Harzic, N. Huot, E. Audouard, C. Jonin, P. Laporte, Appl. Phys. Lett. 80 (2002) 3886–3888.