A comparative fatigue study of solder/electroless-nickel and solder/copper interfaces
- PDF / 1,624,952 Bytes
- 9 Pages / 612 x 792 pts (letter) Page_size
- 32 Downloads / 181 Views
The fatigue resistance of the interface between electroless–nickel and the eutectic tin–lead solder alloy was examined in the as-reflowed and aged conditions and compared to fatigue behavior of the copper/solder interface under the same conditions. In the as-reflowed state, the fatigue resistance of the solder/electroless-nickel interface was slightly superior to that of the solder/copper interface. However, after long-term aging, the fatigue resistance of the solder/electroless-nickel interface became far worse in the high crack growth rate regime. Examinations of interfacial microstructures and crack growth mechanisms indicated that the differences in fatigue resistance between the two interfaces were not directly related to the thickness of the intermetallic phase at the interface, as commonly believed, but were due to differences in crack growth mechanisms. I. INTRODUCTION
Nickel is widely used as an interconnect metal in microelectronics because of its high electrical conductivity and chemical stability. Its rather limited solubility in tin and slow reaction rate with tin-based solder alloys have made it an ideal choice for the diffusion barrier in controlling intermetallic growth between copper and tin, as exemplified in the under-bump-metallurgy (UBM) of ball-grid array devices.1 In this capacity, a thin nickel film of a few micrometers in thickness is either electroplated or electrolessly plated onto the underlying copper metallization. Electroless nickel is chosen for the simplicity of the plating process, the uniformity of the coating, and its desirable physical and chemical properties. Electroless nickel plating is an autocatalytic process where nickel is reduced from an ionic nickel solution because of the difference in chemical potentials of the nickel in and out of the solution. It requires no electrical current and therefore can be applied to both conducting and nonconducting substrates. The most common process is based on hypophosphite reducing agent, which produces deposits of nickel–phosphorus alloys with P concentration up to 12 wt%.2 Depending on the P concentration and plating condition, a wide range of microstructures and properties may be obtained in Ni–P deposits.3–12 On contact with liquid solder alloys, the electroless nickel develops good wetting and is highly solderable.12 When copper is covered with a few micrometer thick Ni film, the reaction between copper and Sn is essentially stopped. Instead, Sn reacts with Ni to form Ni–Sn intermetallic compounds. The most common reaction product is the Ni3Sn4 intermetallic phase.13–20 When alloyed with J. Mater. Res., Vol. 15, No. 11, Nov 2000
http://journals.cambridge.org
Downloaded: 30 Mar 2016
P, the electroless Ni also develops a P-rich region at the solder interface.15,16,19,20 In both liquid and solid states, the kinetics of the Sn–Ni reaction is much slower than that of the Sn–Cu reaction, and the thickness of the Ni3Sn4 intermetallic phase is only a fraction of the thickness of the Sn–Cu intermetallic formed under the identical reflow
Data Loading...
