PBGA package warpage and impact on traditional MSL classification for Pb-free assembly: the authors call for new classification criteria in light of warpage results.

Traditional moisture sensitivity classification per IPC/JEDEC J-STD-020C outlines how a device manufacturer must classify a given package type for moisture/reflow performance. (1) Device classification is based on electrical test results and C-SAM inspection for cracking/delamination at a given moisture sensitivity level (MSL) and peak reflow temperature. One potential shortfall of J-STD-020C is a missing requirement for the evaluation of high-temperature package body warpage response as a function of moisture content and peak reflow temperature. High-temperature component warpage is a concern known to affect the robust assembly performance of area array package types. (2-4) J-STD-020C does not address excessive high-temperature component warpage as a failure criterion for MSL classification. The specification does, however, include as a failure criterion room temperature coplanarity and its associated affect on package stand-off dimensions. Unfortunately, this measurement alone is not sufficient to address high-temperature component assembly robustness, especially for Pb-free board level assembly.

The results obtained from our earlier work (2) suggested that PBGA package warpage may lead to board mount issues for Pb-free SMT assembly, which requires higher peak temperatures to ensure proper reflow of the solder. The potential for board assembly-related issues is further compounded by the fact that component users generally require the same MSL classification ratings for Pb-free SMT assembly so that manufacturing exposure time (i.e., "floor life") is preserved.

The intent of the present study is twofold:

1. To determine the interdependence of body size and die size on the high-temperature warpage response for PBGA packages and the impact on SMT assembly robustness.

2. To determine the impact of high-temperature warpage on SMT assembly for Pb-free capability and how this will ultimately affect traditional MSL classification criteria.

Nine test packages were evaluated in this study. All test devices were overmolded plastic ball grid array (PBGA) packages having 1.0 mm BGA ball pitch and ranging in body size from 23 X 23 mm to 37.5 X 37.5 mm. The die size ranged from 5 X 5 mm to 11 X 11 mm and in one limiting case, the package was manufactured with no silicon die. Package substrates consisted of four-metal layer Bismaleimide Triazine (BT) laminates ranging in thickness from 0.56 to 0.61 mm, depending on the particular package body size.

Packages were baked for 72 hours at 125[degrees]C to remove all residual moisture content. They were then subjected to MSL-3 (192 hours at 30[degrees]C/60%RH) soak conditions per IPC/JEDEC J-STD-020C. Next, the packages were assembled to an eight-layer, 2.36 mm thick, high Tg FR-4 PCB containing 0.5 mm diameter non-soldermask defined (NSMD) copper pads.

PCBs with both immersion gold and immersion silver finishes were used in the assembly experiments. A type 3 no-clean solder paste was screen-printed onto the boards using a 0.13 mm thick stainless steel stencil with 0.432 mm square apertures and 0.076 mm radius rounded corners. Both eutectic SnPb and Pb-free pastes were used in the board assembly experiments. After component placement, the PCBs were sent through a five-zone convection reflow oven. Reflow profiles were developed by monitoring the top body surface temperature of the package using thermocouples attached to the mold compound surface using a thermally conductive epoxy. Peak reflow temperatures ranged between 240[degrees] and 260[degrees]C.

Post-assembly inspection consisted of x-ray analysis to detect solder ball shorts and scanning acoustic microscopy (C-SAM) to check for internal package delamination effects.

For each package TV included in the board assembly study, thermal shadow moire was used to measure PBGA package warpage over the reflow profile temperature range. Shadow moire is a noncontact technique that permits measurement of the topography of solid surfaces as a function of induced temperature changes.

Shadow moire measurements were performed on a subset of packages not used for board assembly. Both dry and moisture-exposed devices were measured. All packages used for moire analysis were manufactured without solder balls to facilitate measurements on the substrate side of the device. Even though solder balls were not attached, the moire test packages were subjected to the BGA ball attach reflow profile to expose the packages to the same thermal treatment as the samples used in the assembly builds.

Results

PBGA package body warpage increases with ingressed moisture content and peak reflow temperature, as required for Pb-free SMT assembly applications. PBGA package body warpage increased with package body size and decreasing silicon die size. Larger PBGA packages (35 X 35 mm and greater with 1.0 mm solder ball pitch) passed traditional moisture/reflow classification testing, but failed board assembly due to solder ball shorts as a result of excessive package body warpage. PBGA packages with body sizes 27 X 27 mm and less (with a 1.0 mm solder ball pitch) successfully passed board assembly due to smaller total package body warpage. Our tests reveal that IPC/JEDEC moisture/reflow classification criteria should be modified to include the effects of package body warpage to ensure robust Pb-free SMT assembly capability for PBGA packages.

Ed.: For the full article, please see circuitsassembly.com/cms/content/view/1344/

Ed.: This paper was presented at SMTA International in September 2004 and is used with permission.

Brian T. Vaccaro, Richard L. Shook, E. Thomas, J.J. Gilbert, C. Horvath, A. Dairo and G.J. Libricz are with Agere Systems (agere.com); [e-mail].