Ceramic circuit board and method for producing same

Abstract

[Problem] To obtain a ceramic circuit board having superior crack-resistance with respect to ultrasonic bonding. [Solution] The abovementioned problem is solved by a ceramic circuit board characterized in that a metal circuit board is bonded to one surface of a ceramic substrate and a metal heat radiation plate is bonded to the other surface of the ceramic substrate, wherein the crystal grain size in the metal circuit board is at least 20 m and at most 70 m. This ceramic circuit board can be manufactured by arranging the metal circuit board on one surface of the ceramic substrate and arranging the metal heat radiation plate on the other surface of the ceramic substrate, and bonding in a vacuum of at most 110.sup.3 Pa, at a bonding temperature of at least 780 C. and at most 850 C., for a retention time of at least 10 minutes and at most 60 minutes.

Claims

1. A ceramic circuit board comprising a metal circuit board bonded to one surface of a ceramic substrate and a metal heat radiation plate bonded to the other surface of the ceramic substrate, wherein a crystal grain size in the metal circuit board is at least 20 m and at most 70 m.

2. The ceramic circuit board according to claim 1, wherein the ceramic substrate comprises aluminum nitride.

3. A manufacturing method for the ceramic circuit board according to claim 1, comprising arranging a metal circuit board on one surface of a ceramic substrate and arranging a metal heat radiation plate on the other surface of the ceramic substrate, and bonding the metal circuit board and the metal heat radiation plate to the ceramic substrate in a vacuum of at most 1103 Pa, at a bonding temperature of at least 780 C. and at most 850 C., for a retention time of at least 10 minutes and at most 60 minutes.

4. A manufacturing method for the ceramic circuit board according to claim 2, comprising arranging a metal circuit board on one surface of a ceramic substrate and arranging a metal heat radiation plate on the other surface of the ceramic substrate, and bonding the metal circuit board and the metal heat radiation plate to the ceramic substrate in a vacuum of at most 1103 Pa, at a bonding temperature of at least 780 C. and at most 850 C., for a retention time of at least 10 minutes and at most 60 minutes.

Description

EXAMPLES

Example 1

(1) An active metal brazing filler metal containing 3.5 parts by mass of titanium (TSH-350 manufactured by Osaka Titanium Technologies) with respect to 90 parts by mass of silver powder (AgC-BO manufactured by Fukuda Metal Foil & Powder Co., Ltd.) and 10 parts by mass of copper powder (SRC-Cu-20 manufactured by Fukuda Metal Foil & Powder Co., Ltd.) coming to a total of 100 parts by mass, was applied to both surfaces of a 0.635 mm thick aluminum nitride substrate. Following this, a bonded body of copper plates and an aluminum nitride substrate was manufactured by stacking a copper plate for forming circuits and a copper plate for forming a heat radiation plate (both of which were composed of electrolytic copper foil), respectively, on the front surface and the rear surface of the aluminum nitride substrate and holding them at a temperature of 800 C. for 20 minutes in a vacuum atmosphere (6.510.sup.4 Pa).

(2) Circuits were formed by etching the bonded circuit board with an etching solution comprising copper chloride. Further, the brazing filler metal layers were etched with an ammonium fluoride/hydrogen peroxide etching solution and an aluminum nitride circuit board was manufactured.

(3) Measurements of the crystal grain size in the metal circuit board of the obtained aluminum nitride circuit board, ultrasonic bonding test, and heat-cycle test were performed as follows.

(4) <Crystal Grain Size>

(5) The front surface of the metal circuit board of the ceramic circuit board was observed using an optical microscope (50). The longest crystal distance in 10 random 1 mm1 mm regions was confirmed and the average was calculated to arrive at the crystal grain size. The results are shown in Table 1.

(6) <Ultrasonic Bonding Test>

(7) Bonding was performed with a 1.5 mm thick copper electrode material (UP-Lite 3000 manufactured by Adwelds), with a load of 1200 N, a frequency of 20 kHz, an amplitude of 50 m, and a bonding time of 0.4 seconds. After bonding, the copper electrode and copper circuit board were removed by etching and the front surface of the ceramic substrate was observed with an optical microscope. Ultrasonic bonding was performed at 16 locations for each condition and the locations where cracks occurred were ranked in the following three categories.

(8) A: 0 locations, B: 1-8 locations, and C: 9-16 locations

(9) <Heat Cyde Test>

(10) The manufactured ceramic circuit board was exposed to thermal shock testing and after 500 thermal shock testing cycles in which one cycle is 30 minutes at 40 C. and 30 minutes at 125 C., the copper plates were removed by etching and the condition regarding the occurrence of cracks in the front surface of the ceramic substrate was observed with an optical microscope (50). Heat cycle testing was performed on 10 plates and these were ranked in the following three categories by the largest crack length value of the measured crack lengths.

(11) A: No cracks were observed.

(12) B: Observed cracks were less than 100 m in length.

(13) C: Observed cracks were 100 m or longer in length.

(14) Considering each evaluation, a comprehensive ranking by the following three ranks was performed.

(15) Good: Plates with an A rank in both the ultrasonic bonding test and heat cycle test.

(16) Fair: Plates with a B rank in either the ultrasonic bonding test or the heat cycle test.

(17) Poor: Plates with a C rank in either the ultrasonic bonding test or the heat cycle test.

Examples 2-5 and Comparative Examples 1-5

(18) In cases other than the changes in conditions indicated in Table 1, testing and evaluation was performed as in Example 1.

(19) TABLE-US-00001 TABLE 1 Crystallization Grain Size in the Bonding Metal Circuit Evaluation Bonding Retention Board after (Crack Occurrence) Temperature Time Bonding Ultrasonic Heat Cycle Metal Plate C. Minutes m Bonding Resistance Judgment Example 1 Electrolytic 810 20 31 A A Good Copper Foil Example 2 Electrolytic 790 20 22 A A Good Copper Foil Example 3 Electrolytic 840 20 55 A A Good Copper Foil Example 4 Electrolytic 810 10 20 A A Good Copper Foil Example 5 Electrolytic 810 50 62 A A Good Copper Foil Comparative Electrolytic 770 20 18 No Bond No Bond Example 1 Copper Foil Comparative Electrolytic 865 20 81 A B Fair Example 2 Copper Foil Comparative Electrolytic 810 5 15 No Bond No Bond Example 3 Copper Foil Comparative Electrolytic 810 90 91 A B Fair Example 4 Copper Foil Comparative Electrolytic 950 20 172 B C Poor Example 5 Copper Foil

(20) As shown above, due to the present invention, when bonding copper electrodes to a ceramic circuit board by ultrasonic bonding, bonding can be performed without cracks occurring in the ceramic substrate, ceramic circuit boards that are capable of improving module reliability are provided.