A methodology to establish baseline metrics for assessing the isothermally aging of Sn-Pb solder interconnects P.T. Vianco Sandia National Laboratories, Albuquerque, NM J.A. Rejent Sandia National Laboratories, Albuquerque, NM Introduction The long-term functionality of solder interconnects is paramount to the reliability of electronic assemblies in the field. It was deemed advantageous to have a methodology by which to quantitatively assess the age of solder joints as a function of their history, which includes the manufacturing cycle, storage and transportation conditions, as well as the subsequent service environment(s). The first step in the development of such a methodology was to identify those metrics of the solder joint microstructure that are sensitive to time and temperature. The second step was to establish the baseline or initial conditions from which to measure the extent of long-term, solid-state aging. The third step was to determine the rate kinetics by which those metrics change as a function of the time and temperature conditions which support the solid-state aging processes. Two properties were identified as suitable aging metrics; they were (1) the thickness of the intermetallic compound layer that forms at the interface between the tin-lead (Sn-Pb) solder and the copper (Cu) substrate and (2) the size distribution of the lead (Pb)-rich phase particles in the solder microstructure. First, the intermetallic compound layer thickness metric will be discussed. An optical micrograph of the intermetallic compound layer formed in an as-fabricated solder joint is shown in Figure 1a(Vianco et al., 1994). The layer initially forms between the molten solder and the Cu substrate during assembly. Once the solder has solidified, the layer can continue to thicken, albeit more slowly, supported by solid-state diffusion processes. The growth of the intermetallic compound layer is supported by solid-state diffusion processes and hence, is thermally activated. The accelerated development of the layer under elevated temperatures is exemplified by the micrograph in Figure 1b. The aging conditions were 1708C for 402 days. The layer in Figure 1b was comprised of two sub-layer compositions. The darker layer adjacent to the Cu substrate had the composition Cu 3 Sn while that neighboring the solder field has the Cu 6 Sn 5 composition. The rate of growth exhibited by the intermetallic compound layer is not significantly affected by residual stresses generated in the joint (Vianco and Rejent, 1997). The growth. The second aging metric for Sn-Pb solder joints is the size distribution of the Pb-rich phase particles in the solder. The Sn-Pb solder microstructure has two phases which are identified in Figure 2. The matrix phase is essentially pure Sn, as there is a negligible solid solubility for Pb in Sn. The particles are the Pb-rich phase, comprised primarily of Pb, but with a finite solubility for Sn on the order of 0–5 wt.%; the exact Sn concentration depends upon the solidification and cooling conditions experienced by the solder. The consequences of aging are that the Pb-rich phase particles coarsen and become fewer in number. The coarsening is supported by solid-state diffusion processes, and like intermetallic compound layer growth, is thermally activated. However, the coarsening rate is also sensitive to mechanical deformation resulting from residual or externally applied stresses. It is this deformation-enhanced coarsening that is the signature artifact of thermal mechanical fatigue degradation of Sn-Pb solder joints (Frear, 1990). Thus, the Pb-rich phase particle size serves as an aging metric describing the cumulative effects of both temperature and deformation. In this study, the Pb-rich phase size distribution referred to the area of each particle projected through a metallographic cross section. By directly measuring the areas of the particles through the availability of image analysis software, a more accurate account could be made of the phase particle size distribution as opposed to the use of linear intercept methods and their associated correction methodologies. Growth of the intermetallic compound layer and coarsening of the Pb-rich particles were thermally activated processes. Therefore, the respective rate kinetics/were presumed to have the following form: x ¼ x 0 þ At n expð2DH=RT Þ ð1Þ where x is the parameter of interest, either the mean layer thickness or the mean Pb-rich phase particle size (projected area); x 0 is the baseline or as-fabricated parameter for each of the respective metrics, A is the pre-exponential constant, t is the time parameter, n is the time exponent, DH is the apparent activation energy, R is the universal gas constant (8.314 J/mol-K), and T is the temperature. This methodology has been used in previous studies (e.g., Vianco et al., 1994). The logarithm of equation (1) was taken, resulting in equation (2) below: lnðx 2 x 0 Þ¼ ln A þ n ln ðtÞ 2 DH=RT ð2Þ Experimental data were fitted to equation (2) by performing a multivariable, linear regression analysis. The independent variables were ln(t ) and 1/T; the dependent variable was ln(x2x 0 ). The output data were the intercept, ln(A ), and the slopes of the two independent variables: the time exponent, n, and the parameter, DH/R. The latter slope value was multiplied by R to arrive at the apparent activation energy, DH. The current issue and full text archive of this journal is available at http:/www.emeraldinsight.com/0954-0911.htm Keywords Solder joints, Printed circuit boards Abstract A methodology was developed to establish baseline metrics for assessing the isothermal aging of 63Sn-37Pb (or 60Sn-40Pb) solder joints in circuit board assemblies. Those metrics were the intermetallic compound layer thickness at the Sn-Pb solder/Cu interface and the Pb-rich phase particle size distribution in the solder. The baseline, or as- fabricated, values for these metrics were 0.71 ^ 0.27 mm and 3.2 ^ 6.5 £ 10 26 mm 2 , respectively. The rate kinetics were determined for growth of the intermetallic compound layer and coarsening of the Pb-rich phase particles by isothermal aging experiments. The/were: (1) intermetallic compound layer thickness (mm)¼0.714 þ 3.265 £ 10 3 t 0.58 exp(252200/ RT); and (2) Pb-rich phase particle size (mm 2 )¼3.2 £ 10 26 þ 1.47 £ 10 23 t 0.32 exp(231000/RT). Received: October 2001 Accepted: February 2002 (Received 29 October 2001; In final form 20 February 2002) Soldering and Surface Mount Technology 14/2 [2002] 26–34 q MCB UP Limited [ISSN 0954-0911] DOI: 10.1108/09540910210427790 [26]