-
- News
- Books
Featured Books
- smt007 Magazine
Latest Issues
Current IssueBox Build
One trend is to add box build and final assembly to your product offering. In this issue, we explore the opportunities and risks of adding system assembly to your service portfolio.
IPC APEX EXPO 2024 Pre-show
This month’s issue devotes its pages to a comprehensive preview of the IPC APEX EXPO 2024 event. Whether your role is technical or business, if you're new-to-the-industry or seasoned veteran, you'll find value throughout this program.
Boost Your Sales
Every part of your business can be evaluated as a process, including your sales funnel. Optimizing your selling process requires a coordinated effort between marketing and sales. In this issue, industry experts in marketing and sales offer their best advice on how to boost your sales efforts.
- Articles
- Columns
Search Console
- Links
- Events
||| MENU - smt007 Magazine
Minimizing Voids in Lead-free Array Packages
December 31, 1969 |Estimated reading time: 5 minutes
Minimizing Voids in Lead-free Array Packages
Voids in lead-free alloys and their effect on reliability, causes, and classification have been debated extensively. Only process voids can be laid firmly at the feet of the solder paste, though alloy can influence via-in-pad voids. Process voids typically are the most common and often the easiest to manage.
BY Brian Toleno, Ph.D., and Neil Poole, Ph.D., Henkel
Lead-free solder alloys, such as tin/silver/copper (SAC), all have higher surface tension than tin/lead alloys by nature. This impedes the escape of voids from the solder joint once formed. Because of this, the most effective route to minimize voids is to prevent their formation. Optimizing the process and materials can effectively deter void formation. The following analysis will show that best results can be achieved by judicious choice of profile and pad finish, for any given solder paste. To isolate the effects, only two solder pastes, N and W, were chosen. A screening design of experiment (DoE) helped determine optimum profile type and pad finish for each material.
Experiment
For the first part of the investigation (solder paste W), a standard 0.062"-thick, 6" × 4" test board was selected as representative of the final application. A 1.2-mm-pitch, 256-I/O BGA was the array component. The part was a dummy component complete with dummy die bumped with a SAC alloy. In some cases, the actual board and component can be used for the evaluation, providing costs do not make the approach impractical. When choosing test systems, it is desirable to mimic the same component sizes and pitches as much as possible. A 150-μm-thick stainless steel stencil with round apertures, using a 1:1 aperture-to-pad ratio, was used to print the paste. Four pad finishes - immersion silver (ImmAg), immersion tin (ImmSn), copper (Cu), and electroless nickel immersion gold (ENIG) - were available. Boards were reflowed using one of three reflow profiles (Figure 1). The completed boards were then evaluated with X-ray.
The second phase of the investigation (solder paste N) was an abbreviated version of the first, using a smaller 132-I/O chip-scale package (CSP), with SAC bumps, mounted on a test board. Only one soak profile was used: the cool soak and linear profile. The completed boards also were evaluated using an X-ray system.
Results
When analyzing the voiding performance of any paste, two factors should be considered: void-size distribution and total voiding. It has been suggested that multiple small voids are less of a reliability issue than a single large void. The recorded data was analyzed for void-size distribution as well as total void instances for the different pad finishes and profiles. Simple statistical methods were used to determine if apparent differences were statistically significant. Figures 2a through 2d illustrate the void-size distribution for the different profiles on each of the different pad finishes used with paste W. Visual inspection of the data reveals fundamental differences in void-size distribution between the three profiles across the various substrates. With the exception of the copper pads, the linear profile created smaller voids and fewer large voids. When using the linear profile with the copper pads, fewer voids of all sizes were generated. The two soak profiles have qualitatively similar distribution profiles with long tails of single voids up to 10% of the ball volume. Typically, the hot soak profile produced more voids of all sizes than the cool soak. The size distribution, however, does not present an evaluation of total void levels in joints as a comparison from one profile to another, another factor.
Figure 1. Reflow profiles used for the voiding investigation.
While the distribution of void sizes is very similar, the compilation of the individual void data into total voiding-per-solder-interconnect reveals differences between the various profiles for pastes N and W. Significance testing showed that each profile was statistically different from the others for paste W and illustrated statistically significant linear profiles and cool soak profiles for paste N. The data for paste W suggests an elevated void level associated with increasing heat in the pre-liquidus phase of the profile, while paste N displayed the reverse. This reinforces the need for each solder paste to be evaluated on its own merits. The increase in voiding associated with temperature elevation in the profiles for paste W suggest that this is a result of increased oxidation levels on the powder surface during preheat. A higher amount of oxides on the particle surfaces uses up valuable activator; this activator is then no longer available at liquidus to lower the metal’s surface tension and allow effective coalescence without trapped vapor. Paste N behaved very differently, however, which is most likely the result of additional volatile material being driven off by the extra heat and/or reacting away materials that are thermally unstable at or above solder-liquidus temperatures.
Figure 2A. Voids on ENIG pads with different profiles. B. Voids on ImmAg pads with different profiles. C. Voids on ImmSn pads with different profiles. D. Voids on Cu pads with different profiles.
Analysis of the compiled total voiding data with respect to substrate pad finish also points to substrate-dependent differences - smaller, but statistically significant. The pad finishes used with paste W fall into two groups: immersion tin and ENIG are indistinguishable from each other and generate less overall voiding than immersion silver and copper, which are also indistinguishable from each other. The same ENIG/copper effect is observed for paste N, suggesting that it may be induced more by the chemistry of the pad finishes than their interaction with the flux systems. Wettability of the varying pad finishes also may factor into the results.
Conclusion
Reflow profile has a major influence on the size and number of process voids formed, with the optimal profile dependent on the flux system used. Understanding how various flux systems react differently to profile changes, it is evident that further work is required to improve flux chemistry to deliver more robust materials that can withstand profile changes and deliver lower overall voiding levels. In this study, though the paste chemistries differed greatly, similar voiding levels were achieved. Pad finish also affects voiding, albeit comparatively less than the profile. For assemblers to optimize processes, they must understand the materials’ effect on reflow and the paste/process interactions that suppress voiding and improve yield.
REFERENCES
For a list of references, contact the authors.
Brian Toleno, Ph.D., TS&D manager; and Neil Poole, Ph.D., senior application specialist, the electronics group of Henkel, may be contacted at 15350 Barranca Parkway, Irvine, CA 92608.