Sintered Copper (Cu): Chemistry, Process, and Reliability
This chapter describes Cu-sintered joints for power semiconductor devices. Cu-based die bonding is the least expensive and yields better thermal and electrical properties than other sintered or solder joints. Their thermal characteristics and and reliability, i.e. power cycle and thermal cycle tests, are investigated in this chapter.
The Cu nanoparticles with fatty acids and amines were fabricated and examined in ir thermal and reliability tests. The Cu joint samples were fabricated under a pressure of 1 MPa in a H2 atmosphere at a temperature of 250–350 °C. The specimens consisted of simple structures which eliminated other elements, from top to down; they consisted of an Al2O3 heater chip, a joint consisting of Cu nanoparticles and a baseplate.
This test structure was fixed to a water-circulated cooler using a thermal grease. The temperature was measured using a thin thermocouple, and then the thermal resistances were calculated. The thermal resistances varied with joining temperatures and the type of Cu nanoparticles. The thermal conductivities were also estimated using a finite element analysis. The specimens with the lowest thermal resistance were estimated to be at least 125 W/mK for their thermal conductivity.
The joint with significant higher thermal conductivity might behave thermal spreading effect. The thermal resistance reached a minimum for a certain thickness of the joint, corresponding to the thermal conductivities of the joint and the baseplate are 400 and 207 W/mK, respectively.
The power cycle tests were carried out under constant applied voltage; the voltage applied 10 s on and 50 s off. Temperature change was not found in the 65/200 °C and 65/250 °C tests for the Cu nanoparticle joints, whereas deterioration was found in 65/200 °C test for the Sn-0.7Cu solder joint. However, some vertical cracks of Cu nanoparticle joints were observed after the power cycle tests. In addition, varied CTE (coefficient of thermal expansion) samples for the chip and the baseplate were examined. Significant changes were not seen for Cu-65Mo, Cu-40Mo, and Al-SiC baseplates; however, the specimens using Cu baseplate showed an increase in resistance.
In addition, conventional thermal cycle tests were also conducted, and cross-sectional observations using SEM (scanning electron microscope) were evaluated. The tests were first subject to 1000 cycles of −40/150 °C test and then 1000 cycles of −40/200 °C test. No deterioration was found after −40/150 °C test; however, delamination from the baseplate was found after −40/200 °C test for a specimen jointed at 300 °C. On the contrary, no deterioration was observed after −40/200 °C test for a specimen jointed at 350 °C.
KeywordsPower semiconductor device Copper (Cu) nanoparticle Thermal resistance Thermal simulation Power cycle test Thermal cycle test Reliability CTE (coefficient of thermal expansion) Young’s modulus
This study was carried out at Toyota Central R&D Labs, Inc. and Daido University. The former partner prepared the nanoparticles and joint specimens, and the latter partner carried out the thermal and reliability evaluations.
- 2.Y. Yamada, Y. Takaku, Y. Yagi, I. Nakagawa, T. Atsumi, M. Shirai, I. Ohnuma, K. Ishida, Pb-free high temperature solder joints for power semiconductor devices. Trans. JIEP 2(1), 79–84 (2009)Google Scholar