- Email: [email protected]
Direct laser writing of aluminum and copper on glass surfaces from metal powder
Applied Surface Science 174 (2001) 118±124
Direct laser writing of aluminum and copper on glass surfaces from metal powder Hirofumi Hidai*, Hitoshi Tokura Department of Mechanical Sciences and Engineering, Tokyo Institute of Technology, 2-12-1 O-okayama, Meguro, Tokyo 152-8552, Japan Received 11 August 2000; accepted 26 December 2000
Abstract In this paper, a new, simple, high-speed method of selective metal deposition on glass substrates is proposed. The method is as follows: metal powder is placed on a glass substrate, then an argon ion laser is irradiated through the glass from the other side, consequently the powders are deposited on the glass substrates. Soda glass, Pyrex glass and silica glass were used as substrates, because they are popular materials and their thermal properties were varied. Aluminum and copper powders, with grain sizes of 7.0 and 4.6 mm, respectively, were chosen. Glass substrates and metal powder were placed in a chamber to enable control of the atmosphere, the chamber was ®xed on an electronically controlled X±Y±Z stage. Aluminum and copper can be deposited on all three types of glass. Aluminum deposited on the soda glass were 80±800 mm in width and 10±120 mm in height. The deposited aluminum and copper had high conductivity and resistances of 0.017±0.64 and 0.0014±0.2 O/mm (1 mm long), respectively. The adhesion between deposited copper and soda glass was stronger than 3 MPa. The interface between the glass substrate and deposited metals have a complicated shape, but the border is distinct and aluminum was not diffused, as determined by observation of the cross section and etching the deposited metal. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Argon ion laser; Metal deposition; Aluminum; Copper; Glass
1. Introduction There is a growing need for the deposition of metal ®lms on insulators. In particular, the use of focused lasers for direct writing or maskless patterned depositions have been extensively described in the literatures. The laser-induced chemical vapor deposition (LCVD) [1,2] has been extremely studied, because it is a relatively fast and generally applicable process for * Corresponding author. Tel.: 81-3-5734-2160; fax: 81-3-5734-2893. E-mail addresses: [email protected] (H. Hidai), [email protected] (H. Tokura).
depositing a wide range metals and substrates, but well-developed gas exhaust and vacuum systems are required. Because of these expensive equipments and the need to use toxic gas, there is strong pressure to develop newer, safer and cheaper metal deposition technologies such as laser-assisted forward transfer (LIFT) [3,4], laser-assisted deposition from organometallic solutions [5±7], laser-enhanced electro- [8] or electroless [9,10] plating, and photothermal decomposition of metal-doped polymer ®lms [11,12]. Typically, the adherence of metal depositions to a smooth surface, such as glass or fused quartz, is low [13]. In this paper, a new, simple, high-speed method of selective metal deposition on glass substrates is
0169-4332/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 4 3 3 2 ( 0 1 ) 0 0 0 6 5 - 4
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
119
proposed, in which metal powders placed on a glass substrate are irradiated by a laser through the glass from the other side. The thickness and width of the deposited metal were observed as functions of scanning speed and laser power. The conductivity and adherence of deposited metal were estimated. The surface morphology of the glass substrates and the cross sectional views of the deposited metal were observed by a scanning electronic microscope. 2. Experimental Soda glass, Pyrex glass and silica glass were used as substrates, because they are popular materials and their thermal properties were varied as shown in Table 1. Aluminum and copper powders, with grain sizes of 7.0 and 4.6 mm, respectively, were chosen. In the experiments, the beam of an argon ion laser (Coherent, DBW20) was used at 488 nm wavelength, because the glasses have high transparency of visible light. The laser beam was focused by a convex lens, with a focal length of 170 mm, down to a spot size of about 120 mm. Glass substrates and metal powder were placed in a chamber to enable control of the atmosphere; the chamber was ®xed on an electronically controlled X±Y±Z stage, as shown in Fig. 1. Table 2 shows experimental conditions. The thickness of the metal powder layer was approximately 2 mm and the powder was compressed with a roller. After irradiation, excessive or loosely adhered powder was brushed off, then the substrate was cleaned by ultra sonication.
Soda glass
735
Pyrex glass Silica glass
820 1600
a
At 5±3008C.
3. Results and discussion 3.1. Laser irradiation of the glasses The laser beam was focused on glass substrates, hence, the glass substrates could be damaged, even when metal powder was not placed on the glass substrates. The laser beam (laser power: 7 W maximum, exposure time: 60 s, no scanning) was irradiated on each glass without metal powder. As a result, we found that there were no changes, except for the soda glass which became slightly warmer. Therefore, absorption of the glasses is considered to be negligible in this paper.
Table 2 Experimental conditions
Table 1 Properties of glass substratesa Softening point (8C)
Fig. 1. Illustration of the experimental set-up.
Thermal expansion coefficienta (10ÿ7 Kÿ1)
Transmission rate (%)
99
Approximately 90 90±95 90±95
32 5.5
Laser Wavelength Power Beam mode Scanning speed of moving stage Substrate Thickness of substrate Metal powder (grain size) Powder layer thickness Atmosphere
Ar ion laser 488 nm 0±7.0 W TEM00 0.1±5.0 mm/s Soda, Pyrex, silica glass 2 mm Al (7 mm), Cu (4.6 mm) 2 mm Air, vacuum (3±5 Pa)
120
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
Fig. 2. SEM micrographs of the deposited metals on glasses: laser power 7 W, scanning speed 1.0 mm/s: (a) aluminum on soda glass; (b) aluminum on Pyrex glass; (c) aluminum on silica glass; (d) copper on soda glass.
3.2. Irradiation of the metal powder on the glass Glass substrates, metal powders, laser power and scanning speed were varied in deposition experiments. Both aluminum and copper were deposited on all the glasses under certain irradiation conditions. Fig. 2 shows scanning electron micrographs of the deposited metal lines at a laser power of 7 W and scanning speed of 1 mm/s. These micrographs show that metal powders are deposited well on any glass. Fig. 2a shows aluminum deposited on soda glass. Individual powder grains remain and cracks are observed on the glass due to rapid heating by laser irradiation and cooling. Aluminum deposited on Pyrex glass (Fig. 2b) does
not differ from that on soda glass. On the other hand, aluminum powder on silica glass (Fig. 2c) is different from the others. In some parts of the deposited aluminum, grain shape cannot be recognized. There are no cracks on the glass because silica glass has good thermal endurance. Copper deposited on soda glass is shown in Fig. 2d. Detachment is recognized at the interface of the glass substrate and the deposited copper. Fig. 3 shows the deposited state with changing laser power, scanning speed, metal powder and glass substrate. Metal powders are deposited better in each combination of metal and glass, in the case when laser power is higher and scanning speed is slower.
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
121
Fig. 3. In¯uence of on the states of the deposited metal powers: (}) good conductor (<10 MO (1 mm long)); (*) well-deposited but poor conductor (>10 MO (1 mm long)); (~) deposited in some parts and not deposited in other parts; () not deposited at all. (a) Aluminum on soda glass; (b) aluminum on Pyrex glass; (c) aluminum on silica glass; (d) copper on soda glass; (e) copper on Pyrex glass; (f) copper on silica glass.
Aluminum was conductive on any glass. Copper was also deposited on all three types of glass, but adhesion strength on silica glass was poor, so the deposited copper was easily detached during brushing or wash-
ing in an ultrasonic bath. Copper deposited on soda glass and Pyrex glass was conductive, but on silica glass it was not conductive under any condition. Both metals adhered strongest to the soda glass and poorest
122
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
Fig. 5. Resistance of 1 mm long aluminum and copper lines on soda glass as a function of scanning speed and laser power.
increasing laser power and decreasing scanning speed. That of aluminum was 0.017±0.64 O (1 mm long), and that of copper was 0.0014±0.2 O (1 mm long). 3.3. Adhesion strength
Fig. 4. Scanning speed dependence of the thickness and width of the deposited line.
to the silica glass. This result was due to a difference in the softening point. Width and height of the deposited aluminum on the soda glass were measured using an optical microscope, and results are shown in Fig. 4. As scanning speed became slower and laser power became larger, both width and height became larger. However, width was not <80 mm because the diameter of the laser spot was approximately 120 mm. The scanning speed dependence of height for 7 W case is different from other case. The difference may be due to sintered layer in 7 W case was thicker than others. Resistance values of the deposited metal were measured by the double bridge method. When the deposited metals were poor conductors (>1 MO (1 mm long)), as shown in Fig. 3, there were gaps in the deposited metals. For the deposited aluminum and copper having good conductivity, resistance values are shown in Fig. 5. Resistance values decreased with
Adhesion strength between the deposited metal and glass substrates was estimated by the following method: copper was deposited on soda glass by scanning the laser beam for 1 mm in length. Then wire was soldered on the copper and it was pulled up. Adhesion strength was determined, by measuring the applied force when detachment occurred. Irradiation conditions were as follows: laser power 3.0, 5.0 and 7.0 W; scanning speed 0.3, 0.6 and 1.0 mm/s. Fig. 6 shows the results of adhesion strength measurement. Adhesion strength was stronger for metal deposited with higher laser power and slower scanning speed, however, the glass broke in the case when laser power was higher and scanning speed was slower. The SEM observation of the glass substrates revealed that deposited copper was detached not at the interface of the glass substrate and the sintered copper, but at the crack in the glass substrate. Adhesion strength between deposited metal and glass substrate was stronger than 3 MPa. 3.4. Deposition in vacuum In powder metallurgy, metal is usually sintered in an inert or reductive atmosphere, because oxygen in air inhibits metal sintering. To reveal the in¯uence of
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
123
Fig. 6. Adhesive strength between deposited copper and soda glass substrate as a function of scanning speed and laser power: (*) adhesive strength >3 MPa; () adhesive strength <3 Mpa.
oxygen in air, deposition was performed under the same conditions as listed in Table 2 in vacuum (3±5 Pa). As a result, metal powders were sintered well, but the adhesion strength at the interface of the glass substrates and the sintered metals was weaker and detachment occurred when the glass substrate was cleaned with a brush. These results reveal that oxygen is necessary in order for the metal powders to be securely adhered to the glass substrates. 3.5. Observation of cross section and interface of the glass substrate The mechanism of deposition was also investigated by cross sectional observation. Samples for observation were prepared as follows: metals were deposited on the glass substrates at a laser power of 7 W and scanning speed of 1 mm/s, then the deposited metal and glass substrates were buried in resin and cut and polished. Fig. 7 shows cross sectional SEM images of aluminum deposited on the soda glass. A semicircular crack with a radius of 800 mm is observed in Fig. 7a, Fig. 7b is the magni®ed image of Fig. 7a at the boundary surface of the glass substrate and the deposited aluminum. Two layers can be distinguished in the deposited aluminum: one is a layer of melted aluminum just above the glass substrate in which grains cannot be recognized; the other is sintered aluminum in which grains remain. The sintered layer completely covers the melted layer. The interface between the glass substrate and melted metals has an intricate
Fig. 7. Cross sectional images of the deposited aluminum on soda glass. Laser power 7 W; scanning speed: 1 mm/s. (a) A semicircular crack with a radius of 800 mm, (b) the magni®ed image of (a) at the boundary surface of the glass substrate and the deposited aluminum.
shape, but the boundary is distinct and aluminum is not diffused into glass. Cross sections of the other metal and glass combinations were also observed. The deposited metals consisted of two layers in all cases. However, in the case of aluminum deposited on silica glass, the sintered layer was thinner than the others. Consequently, the melted part could be observed at the surface, and surface morphology was different from the others, as shown in Fig. 2c. The boundary surface of the glass substrate was observed by etching the deposited metal. Fig. 8 shows a SEM micrograph of the soda glass surface after the deposited aluminum was etched by
124
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
0.017±0.64 and 0.0014±0.2 O (1 mm long), respectively. The adhesion between deposited copper and soda glass was stronger than 3 MPa. The interface between the glass substrate and deposited metals have a complicated shape, but the border is distinct and aluminum was not diffused, as determined by observation of the cross section and etching the deposited metal. Acknowledgements
Fig. 8. SEM micrograph of the soda glass surface after deposited aluminum was etched by hot concentrated hydrochloric acid. Laser power: 7 W; scanning speed: 1 mm/s.
hot concentrated hydrochloric acid. The surface of the glass substrate became rough over the same width as that of the deposited aluminum. The etched glass was not clear and color of aluminum was observed. This result revealed that aluminum oxide or perfectly penetrated aluminum remained on glass. 4. Conclusions A novel method of metal deposition on glass substrates is proposed. The method is as follows: metal powder is placed on a glass substrate, then an argon ion laser is irradiated through the glass from the other side. Aluminum and copper can be deposited on soda, Pyrex and silica glass. In this research, cracks were created on soda glass and Pyrex glass. However, crack free deposition is expected by improving the irradiation conditions, because crack free deposition was achieved on silica glass. The deposited aluminum and copper had high conductivity and resistances of
This work was ®nancially supported (partly) by the Sasagawa Scienti®c Research Grant from the Japan Science Society. References [1] G.E. Blonder, G.S. Higashi, C.G. Fleming, Appl. Phys. Lett. 50 (12) (1987) 766. [2] J.Y. Tsao, D.J. Ehrlich, Appl. Phys. Lett. 45 (6) (1984) 617. [3] E. Pogarassy, C. Fuchs, F. Kerherve, G. Hauchecorne, J. Perriere, J Mater. Res. 4 (5) (1989) 1082. [4] H. Esrom, J. Zhng, U. Kogelschatz, A.J. Pedraza, Appl. Surf. Sci. 86 (1995) 202. [5] A. Gupta, R. Jagannathan, Appl. Phys. Lett., 51 (26) (1987) 2254. [6] K. Bali, T. Szorenyi, M.R. Brook, G.A. Shafeev, Appl. Surf. Sci. 69 (1993) 75. [7] T. Lin, H.Y. Lee, M.A. Souto, J. Mater. Res. 6 (4) (1991) 760. [8] R.J. von Gutfeld, E.E. Tynan, R.L. Melcher, S.E. Blum, Appl. Phys. Lett. 35 (9) (1079) 651. [9] K.G. Mishra, R.K. Paramguru, J. Electrochem. Soc. 143 (2) (1996) 510. [10] G. Schrott, B. Braren, D.J.M. O'Sullivan, R.F. Saraf, P. Bailey, J. Roldan, J. Electrochem. Soc. 142 (3) (1995) 944. [11] M.E. Gross, G.J. Fisanik, P.K. Gallagher, K.J. Schnoes, M.D. Fennell, Appl. Phys. Lett. 47 (9) (1985) 923. [12] H.G. Muller, Appl. Phys. Lett. 56 (10) (1990) 904. [13] G.A. Shafeev, L. Bellard, J.-M. Themlin, W. Marine, A. Cros, Appl. Surf. Sci. 86 (1995) 387.
Direct laser writing of aluminum and copper on glass surfaces from metal powder Hirofumi Hidai*, Hitoshi Tokura Department of Mechanical Sciences and Engineering, Tokyo Institute of Technology, 2-12-1 O-okayama, Meguro, Tokyo 152-8552, Japan Received 11 August 2000; accepted 26 December 2000
Abstract In this paper, a new, simple, high-speed method of selective metal deposition on glass substrates is proposed. The method is as follows: metal powder is placed on a glass substrate, then an argon ion laser is irradiated through the glass from the other side, consequently the powders are deposited on the glass substrates. Soda glass, Pyrex glass and silica glass were used as substrates, because they are popular materials and their thermal properties were varied. Aluminum and copper powders, with grain sizes of 7.0 and 4.6 mm, respectively, were chosen. Glass substrates and metal powder were placed in a chamber to enable control of the atmosphere, the chamber was ®xed on an electronically controlled X±Y±Z stage. Aluminum and copper can be deposited on all three types of glass. Aluminum deposited on the soda glass were 80±800 mm in width and 10±120 mm in height. The deposited aluminum and copper had high conductivity and resistances of 0.017±0.64 and 0.0014±0.2 O/mm (1 mm long), respectively. The adhesion between deposited copper and soda glass was stronger than 3 MPa. The interface between the glass substrate and deposited metals have a complicated shape, but the border is distinct and aluminum was not diffused, as determined by observation of the cross section and etching the deposited metal. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Argon ion laser; Metal deposition; Aluminum; Copper; Glass
1. Introduction There is a growing need for the deposition of metal ®lms on insulators. In particular, the use of focused lasers for direct writing or maskless patterned depositions have been extensively described in the literatures. The laser-induced chemical vapor deposition (LCVD) [1,2] has been extremely studied, because it is a relatively fast and generally applicable process for * Corresponding author. Tel.: 81-3-5734-2160; fax: 81-3-5734-2893. E-mail addresses: [email protected] (H. Hidai), [email protected] (H. Tokura).
depositing a wide range metals and substrates, but well-developed gas exhaust and vacuum systems are required. Because of these expensive equipments and the need to use toxic gas, there is strong pressure to develop newer, safer and cheaper metal deposition technologies such as laser-assisted forward transfer (LIFT) [3,4], laser-assisted deposition from organometallic solutions [5±7], laser-enhanced electro- [8] or electroless [9,10] plating, and photothermal decomposition of metal-doped polymer ®lms [11,12]. Typically, the adherence of metal depositions to a smooth surface, such as glass or fused quartz, is low [13]. In this paper, a new, simple, high-speed method of selective metal deposition on glass substrates is
0169-4332/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 4 3 3 2 ( 0 1 ) 0 0 0 6 5 - 4
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
119
proposed, in which metal powders placed on a glass substrate are irradiated by a laser through the glass from the other side. The thickness and width of the deposited metal were observed as functions of scanning speed and laser power. The conductivity and adherence of deposited metal were estimated. The surface morphology of the glass substrates and the cross sectional views of the deposited metal were observed by a scanning electronic microscope. 2. Experimental Soda glass, Pyrex glass and silica glass were used as substrates, because they are popular materials and their thermal properties were varied as shown in Table 1. Aluminum and copper powders, with grain sizes of 7.0 and 4.6 mm, respectively, were chosen. In the experiments, the beam of an argon ion laser (Coherent, DBW20) was used at 488 nm wavelength, because the glasses have high transparency of visible light. The laser beam was focused by a convex lens, with a focal length of 170 mm, down to a spot size of about 120 mm. Glass substrates and metal powder were placed in a chamber to enable control of the atmosphere; the chamber was ®xed on an electronically controlled X±Y±Z stage, as shown in Fig. 1. Table 2 shows experimental conditions. The thickness of the metal powder layer was approximately 2 mm and the powder was compressed with a roller. After irradiation, excessive or loosely adhered powder was brushed off, then the substrate was cleaned by ultra sonication.
Soda glass
735
Pyrex glass Silica glass
820 1600
a
At 5±3008C.
3. Results and discussion 3.1. Laser irradiation of the glasses The laser beam was focused on glass substrates, hence, the glass substrates could be damaged, even when metal powder was not placed on the glass substrates. The laser beam (laser power: 7 W maximum, exposure time: 60 s, no scanning) was irradiated on each glass without metal powder. As a result, we found that there were no changes, except for the soda glass which became slightly warmer. Therefore, absorption of the glasses is considered to be negligible in this paper.
Table 2 Experimental conditions
Table 1 Properties of glass substratesa Softening point (8C)
Fig. 1. Illustration of the experimental set-up.
Thermal expansion coefficienta (10ÿ7 Kÿ1)
Transmission rate (%)
99
Approximately 90 90±95 90±95
32 5.5
Laser Wavelength Power Beam mode Scanning speed of moving stage Substrate Thickness of substrate Metal powder (grain size) Powder layer thickness Atmosphere
Ar ion laser 488 nm 0±7.0 W TEM00 0.1±5.0 mm/s Soda, Pyrex, silica glass 2 mm Al (7 mm), Cu (4.6 mm) 2 mm Air, vacuum (3±5 Pa)
120
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
Fig. 2. SEM micrographs of the deposited metals on glasses: laser power 7 W, scanning speed 1.0 mm/s: (a) aluminum on soda glass; (b) aluminum on Pyrex glass; (c) aluminum on silica glass; (d) copper on soda glass.
3.2. Irradiation of the metal powder on the glass Glass substrates, metal powders, laser power and scanning speed were varied in deposition experiments. Both aluminum and copper were deposited on all the glasses under certain irradiation conditions. Fig. 2 shows scanning electron micrographs of the deposited metal lines at a laser power of 7 W and scanning speed of 1 mm/s. These micrographs show that metal powders are deposited well on any glass. Fig. 2a shows aluminum deposited on soda glass. Individual powder grains remain and cracks are observed on the glass due to rapid heating by laser irradiation and cooling. Aluminum deposited on Pyrex glass (Fig. 2b) does
not differ from that on soda glass. On the other hand, aluminum powder on silica glass (Fig. 2c) is different from the others. In some parts of the deposited aluminum, grain shape cannot be recognized. There are no cracks on the glass because silica glass has good thermal endurance. Copper deposited on soda glass is shown in Fig. 2d. Detachment is recognized at the interface of the glass substrate and the deposited copper. Fig. 3 shows the deposited state with changing laser power, scanning speed, metal powder and glass substrate. Metal powders are deposited better in each combination of metal and glass, in the case when laser power is higher and scanning speed is slower.
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
121
Fig. 3. In¯uence of on the states of the deposited metal powers: (}) good conductor (<10 MO (1 mm long)); (*) well-deposited but poor conductor (>10 MO (1 mm long)); (~) deposited in some parts and not deposited in other parts; () not deposited at all. (a) Aluminum on soda glass; (b) aluminum on Pyrex glass; (c) aluminum on silica glass; (d) copper on soda glass; (e) copper on Pyrex glass; (f) copper on silica glass.
Aluminum was conductive on any glass. Copper was also deposited on all three types of glass, but adhesion strength on silica glass was poor, so the deposited copper was easily detached during brushing or wash-
ing in an ultrasonic bath. Copper deposited on soda glass and Pyrex glass was conductive, but on silica glass it was not conductive under any condition. Both metals adhered strongest to the soda glass and poorest
122
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
Fig. 5. Resistance of 1 mm long aluminum and copper lines on soda glass as a function of scanning speed and laser power.
increasing laser power and decreasing scanning speed. That of aluminum was 0.017±0.64 O (1 mm long), and that of copper was 0.0014±0.2 O (1 mm long). 3.3. Adhesion strength
Fig. 4. Scanning speed dependence of the thickness and width of the deposited line.
to the silica glass. This result was due to a difference in the softening point. Width and height of the deposited aluminum on the soda glass were measured using an optical microscope, and results are shown in Fig. 4. As scanning speed became slower and laser power became larger, both width and height became larger. However, width was not <80 mm because the diameter of the laser spot was approximately 120 mm. The scanning speed dependence of height for 7 W case is different from other case. The difference may be due to sintered layer in 7 W case was thicker than others. Resistance values of the deposited metal were measured by the double bridge method. When the deposited metals were poor conductors (>1 MO (1 mm long)), as shown in Fig. 3, there were gaps in the deposited metals. For the deposited aluminum and copper having good conductivity, resistance values are shown in Fig. 5. Resistance values decreased with
Adhesion strength between the deposited metal and glass substrates was estimated by the following method: copper was deposited on soda glass by scanning the laser beam for 1 mm in length. Then wire was soldered on the copper and it was pulled up. Adhesion strength was determined, by measuring the applied force when detachment occurred. Irradiation conditions were as follows: laser power 3.0, 5.0 and 7.0 W; scanning speed 0.3, 0.6 and 1.0 mm/s. Fig. 6 shows the results of adhesion strength measurement. Adhesion strength was stronger for metal deposited with higher laser power and slower scanning speed, however, the glass broke in the case when laser power was higher and scanning speed was slower. The SEM observation of the glass substrates revealed that deposited copper was detached not at the interface of the glass substrate and the sintered copper, but at the crack in the glass substrate. Adhesion strength between deposited metal and glass substrate was stronger than 3 MPa. 3.4. Deposition in vacuum In powder metallurgy, metal is usually sintered in an inert or reductive atmosphere, because oxygen in air inhibits metal sintering. To reveal the in¯uence of
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
123
Fig. 6. Adhesive strength between deposited copper and soda glass substrate as a function of scanning speed and laser power: (*) adhesive strength >3 MPa; () adhesive strength <3 Mpa.
oxygen in air, deposition was performed under the same conditions as listed in Table 2 in vacuum (3±5 Pa). As a result, metal powders were sintered well, but the adhesion strength at the interface of the glass substrates and the sintered metals was weaker and detachment occurred when the glass substrate was cleaned with a brush. These results reveal that oxygen is necessary in order for the metal powders to be securely adhered to the glass substrates. 3.5. Observation of cross section and interface of the glass substrate The mechanism of deposition was also investigated by cross sectional observation. Samples for observation were prepared as follows: metals were deposited on the glass substrates at a laser power of 7 W and scanning speed of 1 mm/s, then the deposited metal and glass substrates were buried in resin and cut and polished. Fig. 7 shows cross sectional SEM images of aluminum deposited on the soda glass. A semicircular crack with a radius of 800 mm is observed in Fig. 7a, Fig. 7b is the magni®ed image of Fig. 7a at the boundary surface of the glass substrate and the deposited aluminum. Two layers can be distinguished in the deposited aluminum: one is a layer of melted aluminum just above the glass substrate in which grains cannot be recognized; the other is sintered aluminum in which grains remain. The sintered layer completely covers the melted layer. The interface between the glass substrate and melted metals has an intricate
Fig. 7. Cross sectional images of the deposited aluminum on soda glass. Laser power 7 W; scanning speed: 1 mm/s. (a) A semicircular crack with a radius of 800 mm, (b) the magni®ed image of (a) at the boundary surface of the glass substrate and the deposited aluminum.
shape, but the boundary is distinct and aluminum is not diffused into glass. Cross sections of the other metal and glass combinations were also observed. The deposited metals consisted of two layers in all cases. However, in the case of aluminum deposited on silica glass, the sintered layer was thinner than the others. Consequently, the melted part could be observed at the surface, and surface morphology was different from the others, as shown in Fig. 2c. The boundary surface of the glass substrate was observed by etching the deposited metal. Fig. 8 shows a SEM micrograph of the soda glass surface after the deposited aluminum was etched by
124
H. Hidai, H. Tokura / Applied Surface Science 174 (2001) 118±124
0.017±0.64 and 0.0014±0.2 O (1 mm long), respectively. The adhesion between deposited copper and soda glass was stronger than 3 MPa. The interface between the glass substrate and deposited metals have a complicated shape, but the border is distinct and aluminum was not diffused, as determined by observation of the cross section and etching the deposited metal. Acknowledgements
Fig. 8. SEM micrograph of the soda glass surface after deposited aluminum was etched by hot concentrated hydrochloric acid. Laser power: 7 W; scanning speed: 1 mm/s.
hot concentrated hydrochloric acid. The surface of the glass substrate became rough over the same width as that of the deposited aluminum. The etched glass was not clear and color of aluminum was observed. This result revealed that aluminum oxide or perfectly penetrated aluminum remained on glass. 4. Conclusions A novel method of metal deposition on glass substrates is proposed. The method is as follows: metal powder is placed on a glass substrate, then an argon ion laser is irradiated through the glass from the other side. Aluminum and copper can be deposited on soda, Pyrex and silica glass. In this research, cracks were created on soda glass and Pyrex glass. However, crack free deposition is expected by improving the irradiation conditions, because crack free deposition was achieved on silica glass. The deposited aluminum and copper had high conductivity and resistances of
This work was ®nancially supported (partly) by the Sasagawa Scienti®c Research Grant from the Japan Science Society. References [1] G.E. Blonder, G.S. Higashi, C.G. Fleming, Appl. Phys. Lett. 50 (12) (1987) 766. [2] J.Y. Tsao, D.J. Ehrlich, Appl. Phys. Lett. 45 (6) (1984) 617. [3] E. Pogarassy, C. Fuchs, F. Kerherve, G. Hauchecorne, J. Perriere, J Mater. Res. 4 (5) (1989) 1082. [4] H. Esrom, J. Zhng, U. Kogelschatz, A.J. Pedraza, Appl. Surf. Sci. 86 (1995) 202. [5] A. Gupta, R. Jagannathan, Appl. Phys. Lett., 51 (26) (1987) 2254. [6] K. Bali, T. Szorenyi, M.R. Brook, G.A. Shafeev, Appl. Surf. Sci. 69 (1993) 75. [7] T. Lin, H.Y. Lee, M.A. Souto, J. Mater. Res. 6 (4) (1991) 760. [8] R.J. von Gutfeld, E.E. Tynan, R.L. Melcher, S.E. Blum, Appl. Phys. Lett. 35 (9) (1079) 651. [9] K.G. Mishra, R.K. Paramguru, J. Electrochem. Soc. 143 (2) (1996) 510. [10] G. Schrott, B. Braren, D.J.M. O'Sullivan, R.F. Saraf, P. Bailey, J. Roldan, J. Electrochem. Soc. 142 (3) (1995) 944. [11] M.E. Gross, G.J. Fisanik, P.K. Gallagher, K.J. Schnoes, M.D. Fennell, Appl. Phys. Lett. 47 (9) (1985) 923. [12] H.G. Muller, Appl. Phys. Lett. 56 (10) (1990) 904. [13] G.A. Shafeev, L. Bellard, J.-M. Themlin, W. Marine, A. Cros, Appl. Surf. Sci. 86 (1995) 387.