Ceramic circuit board

Abstract

It is an object of the present invention to obtain a ceramic circuit substrate having high bonding strength, excellent heat cycle resistance, enhanced reliability of operation as an electronic device, and excellent heat dissipation properties. The present invention provides a ceramic circuit substrate in which metal plates, particularly copper plates, and both main surfaces of a ceramic substrate are bonded vial silver-copper brazing material layers. The silver-copper brazing material layers are formed from a silver-copper brazing material including i) 0.3-7.5 parts by mass of carbon fibers, and ii) 1.0-9.0 parts by mass of at least one active metal selected from titanium, zirconium, hafnium, niobium, tantalum, vanadium, and tin; with respect to iii) a total of 100 parts by mass of a) 75-98 parts by mass of silver powder and b) 2-25 parts by mass of copper powder. The carbon fibers having an average length of 15-400 m, an average diameter of 5-25 m and an average aspect ratio of 3-28.

Claims

1. A ceramic circuit substrate comprising two main surfaces bonded to a copper metal plate via a silver-copper brazing material layer, wherein the silver-copper brazing material layer is formed from a silver-copper brazing material comprising carbon fibers in a range of 0.3-7.5 parts by mass and at least one active metal selected from the group consisting of titanium, zirconium, hafnium, niobium, tantalum, vanadium, and tin in a range of 1.0-9.0 parts by mass with respect to a total of 100 parts by mass of a silver powder in a range of 75-98 parts by mass and a copper powder in a range of 2-25 parts by mass, the carbon fibers have an average length in a range of 15-400 m, an average diameter in a range of 5-25 m, and an average aspect ratio in a range of 3-28, and the ceramic substrate has a thickness in a range of 0.1 to 1.0 mm and the metal plate has a thickness in a range of 1.0 to 1.5 mm.

2. The ceramic circuit substrate of claim 1, wherein the carbon fibers have an average length in a range of 15-160 m and an average aspect ratio in a range of 3-8.

3. The ceramic circuit substrate of claim 1, wherein the carbon fibers have an average length in a range of 22-160 m, an average diameter in a range of 7.5-25 m, and an average aspect ratio in a range of 3-8.

4. The ceramic circuit substrate of claim 1, wherein the ceramic substrate comprises at least one selected from the group consisting of silicon nitride, zirconium oxide, silicon carbide and lanthanum boride.

5. The ceramic circuit substrate of claim 1, wherein the carbon fibers have an average aspect ratio in a range of 3-24.

6. The ceramic circuit substrate of claim 1, wherein the carbon fibers have an average aspect ratio in a range of 8-28.

7. The ceramic circuit substrate of claim 1, wherein the carbon fibers have an average aspect ratio in a range of 15-28.

8. The ceramic circuit substrate of claim 1, wherein the carbon fibers have an average length in a range of 120-360 m, an average diameter in a range of 8-20 m, and an average aspect ratio in a range of 3-24.

9. A ceramic circuit substrate comprising two main surfaces bonded to a copper metal plate via a silver-copper brazing material layer, wherein the silver-copper brazing material layer is formed from a silver-copper brazing material comprising carbon fibers in a range of 0.3-7.5 parts by mass and titanium in a range of 1.0-9.0 parts by mass with respect to a total of 100 parts by mass of a silver powder in a range of 75-98 parts by mass and a copper powder in a range of 2-25 parts by mass, the carbon fibers have an average length in a range of 15-400 m, an average diameter in a range of 5-25 m, and an average aspect ratio in a range of 3-28, a ceramic used in the ceramic circuit substrate is selected from the group consisting of silicon nitride, aluminum nitride, aluminum oxide, zirconium oxide, silicon carbide, and lanthanum boride.

Description

EXAMPLES

Example 1

(1) An active metal brazing material comprising carbon fibers (made by Nippon Graphite Fiber Corporation: XN-100-15M) having an average length of 120 m, an average diameter of 15 m, and an average aspect ratio of 8 at 1.5 parts by mass and titanium (made by Osaka Titanium Technologies Co., Ltd: TSH-350) at 3.5 parts by mass with respect to a total of 100 parts by mass of a silver powder (made by Fukuda Metal Foil & Powder Co., LTD: AgC-BO) at 90 parts by mass and a copper powder (made by Fukuda Metal Foil & Powder Co., LTD: SRC-Cu-20) at 10 parts by mass was applied to both main surfaces of a silicon nitride substrate having a thickness of 0.25 mm, and oxygen-free copper plates having a thickness of 1.0 mm and 1.0 mm were respectively bonded to the circuit surface and the rear surface in a vacuum at 830 degrees Celsius for 20 minutes.

(2) The bonded circuit substrate was etched with an etching solution including copper chloride and circuits were formed. Further, the brazing material layer was etched with an etching solution of ammonium fluoride/hydrogen peroxide, and the silicon nitride circuit substrate was manufactured.

(3) The bonding characteristics of the copper plates and the silicon nitride substrate and the heat cycle resistance of the circuit substrate were assessed with the following methods.

(4) <Bonding Characteristics of the Copper Plates and the Silicon Nitride Substrate>

(5) The bonding characteristics of the copper plates and the silicon nitride substrate were assessed with a peel strength measurement. The measurement method was as follows an edge of a pattern having a 5 mm width, this being a portion of the copper circuit pattern bonded to the silicon nitride substrate, was torn off with pliers, this bonded substrate was secured on a pedestal of a tension testing apparatus, and the edge of the pattern was attached to a chuck of a pull testing apparatus. At this time, these were arranged such that the angle of the surface of the silicon nitride substrate and the torn copper circuit pattern was 90 degrees (vertical direction). Subsequently, the tension testing apparatus was operated, the torn pattern was pulled and moved upwards via the chuck, and the maximum pealing load at this time was measured. The maximum peeling load was divided by a width (0.5 cm) and the bonding strength was measured. The results are displayed in Table 3.

(6) <Assessment of the Heat Cycle Resistance>

(7) The manufactured silicon nitride substrate was subjected to a heat cycle resistance test wherein one cycle was set at 40 degrees Celsius for 30 minutes, 25 degrees Celsius for 10 minutes, 150 degrees Celsius for 30 minutes, and 25 degrees Celsius for 10 minutes for 2001) repeated cycles. Subsequently, the copper plates and the brazing material layer were separated from the silicon nitride circuit substrate by using a copper chloride solution and an etching solution of ammonium fluoride/hydrogen peroxide, a horizontal crack area of the surface of the silicon nitride substrate was binarized and calculated with an image analysis software GIMP 2 (threshold value of 140), following which, a crack rate (%) was calculated from the horizontal crack area/area of circuit pattern (that is, a ratio of the horizontal crack area to the circuit pattern area). The :results are displayed in Table 3.

(8) As displayed in Table 3, when bonding the copper plates to the silicon nitride substrate, by using a blend comprising carbon fibers having an average length of 120 m, an average diameter of at most 15 m, and an average aspect ratio of 8 at 1.5 parts by mass and titanium at 3.5 parts by mass with respect to a total of 100 parts by mass of a silver powder at 90 parts by mass and a copper powder at 10 parts by mass, and bonding at 780-875 degrees Celsius for 10-60 minutes like in Example 1, a circuit substrate having a crack rate of 0.01% assessed with the heat cycle resistance, without reducing the bonding characteristics, was obtained, thus establishing that circuit substrates having a crack rate of at most 1% can be obtained.

Examples 2-24 and Comparative Examples 1-13

(9) Aside from changing the conditions displayed in Tables 1 and 2, these were conducted in the same way as Example 1. Further, in Comparative Example 13, in place of carbon fibers, non-fibrous carbon particles were used. The assessments of the bonding characteristics of the copper plates and the silicon nitride substrate and the heat cycle resistance were performed in the same way as Example 1. The results are displayed in Tables 3 and 4.

(10) TABLE-US-00001 TABLE 1 Brazing Material Composition Carbon Fiber Active Metal Structural Blending Blending Member Thickness Ag Cu amount amount Silicon (Parts (Parts (Parts Length Diameter Aspect (Parts Nitride Copper No. by mass) by mass) by mass) (m) (m) Ratio Type by mass) mm mm Example 1 90 10 1.5 120 15 8 Ti 3.5 0.25 1 Example 2 95 5 1.5 120 15 8 Ti 3.5 0.25 1 Example 3 77 25 1.5 120 15 8 Ti 3.5 0.25 1 Example 4 90 10 1.5 120 15 8 Ti 1.5 0.25 1 Example 5 90 10 1.5 120 15 8 Ti 5.5 0.25 1 Example 6 90 10 1.5 120 15 8 Ti 8 0.25 1 Example 7 90 10 0.5 120 15 8 Ti 3.5 0.25 1 Example 8 90 10 3.5 120 15 8 Ti 3.5 0.25 1 Example 9 90 10 6.5 120 15 8 Ti 3.5 0.25 1 Example 10 90 10 1.5 360 15 24 Ti 3.5 0.25 1 Example 11 90 10 1.5 45 15 3 Ti 3.5 0.25 1 Example 12 90 10 1.5 120 8 15 Ti 3.5 0.25 1 Example 13 90 10 1.5 120 20 6 Ti 3.5 0.25 1 Example 14 90 10 1.5 24 8 3 Ti 3.5 0.25 1 Example 15 90 10 1.5 160 8 20 Ti 3.5 0.25 1 Example 19 90 10 1.5 120 15 8 Ti 3.5 0.25 1 Example 20 90 10 1.5 120 15 8 Ti 3.5 0.25 1 Example 21 90 10 1.5 120 15 8 Ti 3.5 0.25 1 Example 22 90 10 1.5 120 15 8 Ti 3.5 0.25 1 Example 23 90 10 1.5 120 15 8 Ti 3.5 0.4 1 Example 24 90 10 1.5 120 15 8 Ti 3.5 0.25 0.5

(11) TABLE-US-00002 TABLE 2 Brazing Material Composition Carbon Fiber Active Metal Structural Blending Blending Member Thickness Ag Cu amount amount Silicon (Parts (Parts (Parts Length Diameter Aspect (Parts Nitride Copper No. by mass) by mass) by mass) (m) (m) Ratio Type by mass) mm mm Comparative 99 1 1.5 120 15 8 Ti 3.5 0.25 1 Example 1 Comparative 99 1 1.5 120 15 8 Ti 3.5 0.25 1 Example 2 Comparative 70 30 1.5 120 15 8 Ti 3.5 0.25 1 Example 3 Comparative 70 30 1.5 120 15 8 Ti 3.5 0.25 1 Example 4 Comparative 90 10 1.5 120 15 8 Ti 0.5 0.25 1 Example 5 Comparative 90 10 1.5 120 15 8 Ti 12 0.25 1 Example 6 Comparative 90 10 0.1 120 15 8 Ti 3.5 0.25 1 Example 7 Comparative 90 10 8 120 15 8 Ti 3.5 0.25 1 Example 8 Comparative 90 10 1.5 450 15 35 Ti 3.5 0.25 1 Example 9 Comparative 90 10 1.5 35 15 2 Ti 3.5 0.25 1 Example 10 Comparative 90 10 1.5 120 3 40 Ti 3.5 0.25 1 Example 11 Comparative 90 10 1.5 120 30 4 Ti 3.5 0.25 1 Example 12 Comparative 90 10 1.5 (carbon) 15 Ti 3.5 0.25 1 Example 13

(12) TABLE-US-00003 TABLE 3 Bonding Bonding Thermal Temper- Bonding Charac- Crack Resis- ature Time teristics rate tance No. C. Minutes N/cm % C./W Example 1 830 20 180 0.01 0.19 Example 2 835 20 160 0.39 0.19 Example 3 840 20 165 0.35 0.19 Example 4 830 20 141 0.56 0.19 Example 5 830 20 193 0.35 0.19 Example 6 830 20 208 0.68 0.19 Example 7 830 20 192 0.46 0.19 Example 8 830 20 176 0.05 0.19 Example 9 830 20 162 0.53 0.19 Example 10 830 20 155 0.40 0.19 Example 11 830 20 185 0.01 0.19 Example 12 830 20 150 0.01 0.19 Example 13 830 20 173 0.22 0.19 Example 14 830 20 185 0.01 0.2 Example 15 830 20 151 0.41 0.21 Example 19 865 20 196 0.22 0.19 Example 20 790 20 166 0.18 0.19 Example 21 830 10 170 0.07 0.19 Example 22 830 55 208 0.09 0.19 Example 23 830 20 188 0.01 0.28 Example 24 830 20 193 <0.01 0.33

(13) TABLE-US-00004 TABLE 4 Bonding Bonding Thermal Temper- Bonding Charac- Crack Resis- ature Time teristics rate tance No. C. Minutes N/cm % C./W Comparative 830 20 Bonding Bonding 0.19 Example 1 Not Not Possible Possible Comparative 880 20 145 2.94 0.19 Example 2 Comparative 830 20 Bonding Bonding 0.19 Example 3 Not Not Possible Possible Comparative 880 20 149 2.32 0.19 Example 4 Comparative 830 20 75 2.86 0.19 Example 5 Comparative 830 20 235 2.45 0.19 Example 6 Comparative 830 20 192 1.4 0.19 Example 7 Comparative 830 20 70 2.99 0.19 Example 8 Comparative 830 20 110 2.50 0.19 Example 9 Comparative 830 20 102 2.40 0.19 Example 10 Comparative 830 20 98 2.40 0.19 Example 11 Compatative 830 20 84 2.40 0.19 Example 12 Comparative 830 20 199 1.01 0.23 Example 13

(14) From Tables 3 and 4, when bonding the copper plates to the silicon nitride substrate, by using a blend comparing carbon fibers having an average length of 15-400 m, an average diameter of 5-25 m, and an average aspect ratio of 3-28 at 0.3-7.5 parts by mass and at least one active metal selected from titanium, zirconium, hafnium, niobium, tantalum, vanadium, and tin at 1.0-9.0 parts by mass with respect to a total of 100 parts by mass of a silver powder at 75-98 parts by mass and a copper powder at 2-25 parts by mass, and bonding at 780-875 degrees Celsius for 10-60 minutes, a circuit substrate having a crack rate of at most 1% assessed with the heat cycle resistance, without reducing the bonding characteristics, was obtained.