Long-term (contact mode, 8 mm) and short-term (noncontact mode) surface-roughness values of the silver films, prepared using the MOD silver paste as well as the high-temperature silver paste screen-printed on polished and nonpolished alumina, are shown in Table 6-1. The short-term surface roughness values of polished and nonpolished bare alumina substrates were found to be 1.93 and 389.9 nm, respectively. The Al2O3 substrates used in this study were fabricated using tape casting technology. After sintering, pores or gain boundaries usually existed on the substrate surface, which determines the surface roughness. For the silver films studied, the short-term surface roughness values, shown in Table 6-1, are slightly larger than the long-term surface roughness values. This is due to the fact that the short-term measurement employs area scanning, which traces the local structure of the surface, while long-term measurement operates on a line scanning, which may not map out all
- 113 -
the defects or bumpy areas. Comparison of both types of silver film shows that the surface roughness values of the fired films printed on the nonpolished Al2O3 substrate are slightly less than those on the polished substrate. The fluidity and wettability of the pastes on the substrate surface may correspond to the differences in the surface roughness of the films. The surface free energies of the polished and nonpolished Al2O3 substrates, measured using a goiniometric system, were found to be 36.69 and 29.81 mN/m, respectively. The surface energy of nonpolished alumina is slightly lower than that of polished alumina, because the former has a higher surface roughness. An increase in the surface roughness and number of pores generally causes a decrease in the wettability of hydrophilic oxide surfaces[10]. It is reported that macroscopic pores are generally filled with low-surface-energy contaminants or absorb water. Silver pastes screen-printed on the nonpolished Al2O3 with lower surface free energy leads to a better wetting property and fluidity, which allows an improved paste leveling on the substrate and results in a reduced surface roughness of the film.
For films prepared using the low-curing-temperature silver paste heat-treated at 250°C, the surface roughness slightly changes with soaking time.
The surface roughness of the films slightly increases as the soaking time of the heat treatment is increased from 5 to 10 min, which is due to the shrinkage associated with the burnout of the organics. For the films prepared using the high-temperature silver paste, the high-temperature firing process causes the sintering of silver particles. Also, the softening of the glass binder at high temperature levels the silver film and enhances the adhesion with the Al2O3 substrate. Sintering and leveling reduces the thickness and surface roughness of the silver film.
Table 6-2 shows the results of DC resistivity of the silver films prepared
- 114 -
using the MOD silver paste as well as the high-temperature silver paste screen-printed on polished and nonpolished alumina, measured using the four-point probe method. The films prepared using the low-curing-temperature MOD silver paste have electrical conductivities ranging from 8.1 x 106 to 10.5 x 106 S/m after the thermal treatment, which is relatively close to the bulk resistivity of Ag. Although the differences are insignificant, it appears that the electrical conductivity slightly increases with soaking time, owing to the organic burnout and coalescence of the silver particles. The films prepared using the high-temperature silver paste have higher electrical conductivity ranging from 4.08 x 107 to 4.13 x 107 S/m, because the high-temperature firing process leads to a better connectivity of the silver particles. Apparently, the slight difference in the surface-roughness of substrate has no effect on the DC conductivity of the screen-printed silver films.
Table 6-3 shows the results of the surface resistance and effective conductivity of the silver films prepared using the MOD silver paste as well as the high-temperature silver paste screen-printed on polished and nonpolished alumina, measured at the microwave frequency range. The surface resistance (Rs) was determined at a frequency of ≈ 4.3 GHz using the TE011 mode of the resonator cavities method. The skin depths of the electric field at microwave frequency for the silver films were calculated, as shown in Table 6-3. The results confirm that the thicknesses of the silver films, shown in Table 6-2, are at least 5 times larger than the skin depth values. The films prepared using the high-temperature silver pastes have surface resistance values smaller than those prepared using the low-curing-temperature MOD silver paste. Comparison of the films prepared using the low-curing-temperature MOD paste after heat treatment at 250°C for various times shows that the surface roughness increases
- 115 -
slightly and the surface resistance slightly decreases with soaking time. This implies that the decrease in the surface resistance of the films is mainly due to the improved connectivity between the silver particles. The surface roughness of the substrate and the film’s free surface has no effect on the high-frequency surface resistance of the screen-printed silver films. This may be due to the fact that the films are sufficient to eliminate the surface scattering (size effect) [11, 12], or the differences in the magnitude of the surface roughness are too small to cause a discrepancy in the high-frequency surface resistance. The effective conductivities of the films were determined using Eq. 6-3 and the results are shown in Table 6-3. The calculated effective conductivities at 4.3 GHz are slightly lower than the DC conductivity of the films. Moreover, similar to the DC conductivity, there is no significant difference in the high-frequency effective conductivity among the silver films prepared using the low-curing-temperature MOD silver paste.
Figure 6-3 shows the SEM surface images of the films prepared using the low-curing-temperature MOD silver paste screen-printed on polished as well as nonpolished substrates after heat treatment at 250°C for 10 and 30 min. Silver grains with neck growth and interconnected porosities are found in the microstructures, which confirm the low resistivities of the silver films. It seems that all the samples have very similar microstructures, although the silver grains are slightly bigger in size for the films being heat treated for 30 min, compared with those films heat treated for 10 min. It is apparent that the nature of the substrate has no effect on the microstructural evolution of the films. Figure 6-4 shows the SEM images of the cross-sectional view of the films shown in Figure 6-3. It can be seen that the microstructures of the films are porous and the coalescent grains and porosities become larger as the soaking time increases.
- 116 -
The surface morphologies of the polished and nonpolished substrates do not change after firing because there is no chemical reaction or interdiffusion at the interface at the firing temperature of 250°C. Figures 6-5 and 6-6 show the SEM images of the surfaces and the crosssections of the films, respectively, prepared using the high-temperature silver paste screen-printed on polished and nonpolished substrates. The sizes of the silver grains ranged between 5 to 10 μm.
Both the films printed on the polished and nonpolished substrates have an irregular interface. In the high-temperature firing process, the glass added to the film not only densifies the silver film but also wets and reacts with the substrate.
This results in a perfect adherence as well as an irregular interface between the films and substrate.
Devices of T-type resonator circuits were designed according to the effective conductivities of the silver films at frequencies of 4.24~4.30 GHz. The simulated and measured Q-values and resonance frequency values of T-type resonators prepared from the films using both low-curing-temperature MOD silver paste and high-temperature silver paste are shown in Table 6-4. The resonator circuit was designed to have a resonance frequency at 4.3~4.5 GHz.
The measured Q-values of T patterns were used to calculate the ∆Q value [∆Q = (QHF-QM)/QHF; QHF: simulated value; QM: measured value]. ∆Q indicates the difference between the simulated and measured values of Q. The results indicate that films screen-printed on the polished substrate have a higher Q and a lower
∆Q value than those of the films screen-printed on the nonpolished substrate. It is known that the device loss at high frequency is mainly due to the conductor loss, dielectric loss, and radiation loss. In this study, the higher roughness of the substrate surface results in a higher ∆Q value (Figure 6-4), while the free-surface roughness of the silver film does not have a significant effect on the
- 117 -
∆Q value. For the silver films prepared using the high-temperature silver paste, the Q values are higher than those of the films prepared using the low-curing-temperature MOD paste, because the former has a dense microstructure (Figure 6-5), which leads to a higher electrical conductivity and lower conductor loss. However, the difference between the simulated and measured Q values for the films printed on the polished substrate is similar to that printed on the nonpolished substrate, which is the largest among the films studied. This is due to the fact that the glass in the film reacts with the substrate during the high-temperature firing process, which results in a perfect adherence as well as an irregular interface between the film and substrate, as shown in Figure 6-4, and a higher dielectric loss.
6-5. Summary
The electrical properties of silver films, prepared using a low-curing-temperature MOD paste and a high-temperature silver paste screen-printed on polished and nonpolished alumina substrates, at microwave frequency were characterized in this study. On the basis of the results obtained, several conclusions could be made as follows:
a. The surface roughnesses of the fired films printed on nonpolished Al2O3 substrate are less than those printed on polished substrate owing to the lower surface energy of the former.
b. The calculated effective conductivities at 4.3 GHz are slightly less than the DC conductivities of the films.
- 118 -
c. The films prepared using the high-temperature silver paste have a higher effective conductivity ranging from 4.08 x 107 to 4.13 x 107 S/m, because the high-temperature firing leads to a better connectivity of the silver particles.
d. The results for the device with T-type resonator circuits indicate that the films screen-printed on the polished substrate have a higher Q and a lower ∆Q value than those of the films screen-printed on the nonpolished substrate.
For the silver films prepared using the high-temperature silver paste, the Q values are higher than those of the films prepared using the low-curing-temperature MOD paste, because the former has a dense microstructure that leads to a higher electrical conductivity and a lower conductor loss. However, the ∆Q values of the films prepared using the high-temperature silver paste are the largest among the films studied, owing to the interfacial reaction between the glass contained in the film and the substrate during the high-temperature firing process, which results in an irregular interface between the film and substrate and leads to a higher dielectric loss.
- 119 -
References:
1. C. A. Lu, P. Lin, H. C. Lin, and S. F. Wang: Jpn. J. Appl. Phys. 45 (2006) 6987.
2. C. A. Lu, P. Lin, H. C. Lin, and S. F. Wang: Jpn. J. Appl. Phys. 46 (2007) 251.
3. J. Krupka, M. Klinger, M. Kuhn, A. Baranyak, M. Stiller, and J. Hinken:
IEEE Trans. Appl. Supercond. 3 (1993) 3043.
4. C. Wilker, Z. Y. Shen, V. X. Nguyen, and M. S. Brenner: IEEE Trans. Appl.
Supercond. 3 (1993) 1457.
5. J. Krupka, IEEE MTT-S Int. Symp. Dig., 2007, p. 515.
6. K. Fuchs: Proc. Cambridge Philos. Soc. 34 (1938) 100.
7. R. J. Good: J. Adhes. Sci. Technol. 12 (1992) 1269.
8. J. Krupka, R. G. Geyer, J. B. Jarvis, and J. Ceremuga: DMMA’ 96 Conf., Bath, U.K. 23-26 Sept. 1996, p. 21.
9. J. Barker-Jarvis, M. D. Janezic, B. Riddle, C. Holloway, N. G. Paulter, and J.
E. Blendell: NIST Tech. Note 1520 (2001).
10. M. Harju, E. Levanen, and T. Mantyla: Appl. Surf. Sci. 252 (2006) 8514.
11. U. Jacob, J. Vancea, and H. Hoffmann: Phys. Rev. B 41 (1990) 11852.
12. E. Z. Luo, S. Heun, M. Kennedy, J. Wollschlager, and M. Henzler: Phys. Rev.
B 49 (1994) 4858.
- 120 -
Figure 6-1. Schematic of dielectric resonator
Metal Film
BZT Dielectric Resonator Al2O3 Substrate
Cu Cavity
Coaxial Line
- 121 - (a)
(b)
Figure 6-2. Test patterns of “T-type” microstrip resonator for silver film printed on (a) polished alumina substrate and (b) nonpolished alumina substrate,
resonated at 4.32 GHz.
0.48mm
50mm
6.85mm
0.61mm
50mm
6.85mm
- 122 -
Figure 6-3. SEM surface images of films prepared using
low-curing-temperature MOD silver paste screen-printed on polished substrate and fired at 250°C for (a) 10 and (c) 30 min, as well as on nonpolished
substrates and fired at 250°C for (b) 10 and (d) 30 min.
(a) (b)
(c) (d)
- 123 -
Figure 6-4. SEM images of crosssections of the films shown in Fig. 3.
(c) (b) (a)
(d)
- 124 -
(a) High‐temperature silver paste on polished substrate (top view, 500 x).
(b) High‐temperature silver paste on nonpolished substrate (top view, 500 x).
Figure 6-5. SEM surface images of films prepared using high-temperature silver paste screen-printed on (a) polished and (b) nonpolished substrate and fired at
800°C for 10 min (a)
(b)
- 125 -
Figure 6-6. SEM images of the crosssections of the films shown in Fig. 5.
(a)
- 126 -
Table 6-1 Long-term and short-term surface roughness values of silver films prepared using MOD silver paste as well as high-temperature silver paste screen-printed on polished and nonpolished alumina
Paste Substrate Temperature (°C) Soaking time (min) Ra (long-terma), 8 mm), μm Ra (short-termb), 640 μm), μm
b) Noncontact mode optical profile surface morphological measurement.
- 127 -
Table 6-2 DC resistivities of silver films prepared using MOD silver paste as well as high-temperature silver paste screen-printed on polished and nonpolished alumina, measured using the four-point probe method.
Paste Substrate Temperature (°C)
b) MOD: low-curing-temperature silver paste containing silver 2-ethylhexanoate [6].
c) HT: commercial LTCC internal electrode application (DuPont, U. S. A.)
- 128 -
Table 6-3 Surface resistance and effective conductivity of silver films prepared using MOD silver paste as well as high-temperature silver paste screen-printed on polished and nonpolished alumina, measured at microwave frequency range.
Paste Substrate Temperature (°C)
- 129 -
Table 6-4 Simulated and measured Q-values and resonance frequency values of T-type resonators prepared from the films using both low-curing-temperature MOD silver paste and high-temperature silver paste.
Paste Substrate
- 130 -