Characterization and Kinetics of the Interfacial Reactions in Solder Joints of Tin-Based Solder Alloys on Copper Substrates

Transcription

1 Characterization and Kinetics of the Interfacial Reactions in Solder Joints of Tin-Based Solder Alloys on Copper Substrates J. C. Madeni*, S. Liu* and T. A. Siewert** *Center for Welding, Joining and Coatings Research George S. Ansell Department of Metallurgical & Materials Engineering Colorado School of Mines Golden, Colorado 841, U.S.A. **Materials Reliability Division National Institute of Standards and Technology Boulder, Colorado 83, U.S.A. 1. INTRODUCTION Health and environmental concerns in the electronic packaging industry regarding the use of lead-based alloys led to the search for lead-free substitute alloys. Previous studies have shown that Sn-based alloys such as Sn-3.Ag, Sn-.7Cu, and Sn-3.2Ag-.8Cu exhibit promising mechanical properties 1 and can be considered as serious candidates to replace the Sn-Pb alloy. Nevertheless, more information is needed before joint life and reliability predictions can be made using these materials since brittle intermetallics that form are detrimental to the mechanical integrity of the solder joints. 2,3 This study provides information about the intermetallic layer formation and growth, between the Sn-3.Ag, Sn-.7Cu, Sn-3.2Ag-.8Cu and Sn-9Zn alloys, and copper-plated substrates. 2. EXPERIMENTAL PROCEDURE The solders alloys were Sn-3.Ag, Sn-.7Cu, Sn-3.2Ag-.8Cu and Sn- 9Zn, and the substrates were Cu-plated printed circuit boards (PCB). The substrates were cut in pieces of 2.4 x 2.4 mm. (1x1 ). Prior to producing the joints, the s were cleaned with alkaline and acidic solutions in the case of Sn-3.Ag, Sn-.7Cu and Sn-3.2Ag-.8Cu; and a commercial sulfuric acid based paste in the case of the Sn-9Zn alloy. This procedure removed the oxide layer and promoted wetting. The joints were cooled in air to room temperature. Then, they were subjected to thermal aging treatment for 2,, 2, hours

2 at 7,, and o C. Each sample was analyzed using light microscopy, SEM and EDS. 3. RESULTS AND DISCUSSION Micrographs of the Cu/Solder interface at o C for hours in Figure 1 show the intermetallic compounds (IMCs),, and Ag 3 Sn. The first two were observed in three of the materials systems, Cu/Sn-3.2Ag-.8Cu, Cu/Sn-3.Ag, and Cu/Sn-.7Cu. In the case of Cu/Sn-3.Ag few small particles of Ag 3 Sn can be observed at the /Sn-3.Ag interface 4. The /Solder interface is mostly irregular and in some areas developing into elongated needlelike and elongated scalloped structures growing far into the solder matrix. The total intermetallic layer thickness at o C, for hours in the four material systems were found to be the following: 13.3 µm for Cu/Sn-3.Ag, 13.6 µm for Cu/Sn-3.2Ag-.8Cu, 14.1 µm for Cu/Sn-.7 Cu, and 19.1 µm for the Cu/Sn-9Zn alloy. The growth kinetics for the IMCs as a function of aging time and temperature was analyzed by measuring the thickness of the IMC layers, and a parabolic relationship was observed, Figure 2. The linear behavior of the data reminds diffusion-controlled growth mechanism. Using the intermetallics formed in the Cu/Sn-3.Ag system as example, the activation energies Q can be calculated for each of the two intermetallic compounds, Figure 3. The results for all other solder alloy systems are shown in table 1. The data are in agreement with results reported by other researchers. 3,4,

3 Ag 3 Sn particle Sn-3.Ag Sn-3.2Ag-.8Cu Needle-like Ag 3 Sn µm µm (a) (b) Sn-.7Cu Sn-9Zn Zn Cu Zn 8 µm µm (c) (d) Figure 1. Micrographs of the Cu/solder interfaces at o C for hours, for the four material systems, indicating their respective intermetallic layers, (a) Cu/Sn- 3.Ag, (b) Cu/Sn-3.8Ag-.8Cu, (c) Cu/Sn-.7Cu, and (d) Cu/Sn-9Zn. 2 Time. (hous). 2 Time. (hours). 2 Time. (hour). Sn-3.2Ag-.8Cu Sn-3.Ag Sn-.7Cu Sn-9Zn (a) (b) (c) Figure 2. Total IMC growth ( + ) at the Cu/solder interface at three temperatures, (a) 7 o C, (b) o C, and (c) o C.

4 2 - - IMC thickness (µm) Time. (hours). Log D (1/T)* 3 Log D (1/T)* 3 (a) (b) (c) Figure 3. Experimental data for the Cu/Sn-3.Ag system, (a) growth of and IMCs, (b) diffusion data for the individual and IMCs plotted to determine the activation energy Q, and (c) diffusion data for the total IMC ( + ) plotted to determine the activation energy Q, for the total IMC growth. Table 1. Activation energies, Q, determined experimentally for the total and individual IMCs in the Cu/Sn-3.8Ag-.8Cu, Cu/Sn-3.Ag, Cu/Sn-.7Cu, and Cu/Sn-9Zn joint systems. ALLOY IMC Q (kj/mol) Q (ev/atom) Sn-3.2Ag-.8Cu Total Sn-3.Ag Total Sn-.7Cu Total Sn-9Zn Cu Zn Conclusions Two layers of intermetallic compounds and, form at the Cu/Sn-3.Ag, Cu/Sn-3.2Ag-.8Cu, and Cu/Sn-.7 Cu interfaces. In the case of the Cu/Sn-9Zn system, the IMC is Cu Zn 8. The three intermetallic layers grow by thermal activation in a parabolic manner. If IMC formation were the sole criterion for determination of candidates for replacing the Sn-Pb alloys, the ranking would

5 be: Sn-3.Ag, Sn-3.2Ag-.8Cu, Sn-.7Cu and Sn-9Zn, being Sn-3.Ag the system that exhibited least IMC growth. It also seems to be more advantageous since the activation energy to promote IMC formation is the highest,.8 ev/atom. References 1. J. Madeni, S, Liu, T. Siewert, Casting of lead-free solder bulk specimens with various solidification rates, ASM International - TMS Annual Conference, Indianapolis, Westbrook, J. H. (Ed.), Intermetallic compounds, John Wiley & Sons Ltd. 1967, New York, NY and Chichester. 3. P. G. Harris and K. S. Chaggar, The role of intermetallic compounds in leadfree soldering, Soldering and Surface Mount Technology /3, 1998, pp Dr. Flanders, E.G. Jacobs and R.F. Pinizzotto, Activation energies of intermetallic growth of Sn-Ag eutectic solder on copper substrates, Journal of Electronic Materials, Col. 26, No7, P. T. Vianco, A. C. Kilgo, R. Grant, Intermetallic compound layer growth kinetics in non lead bearing solders, Sandia National Laboratories, Albuquerque, NM, 199.