0.3 mm CSPs: closer than you think; Revealed: the best-suited material set for next-generation chips.

Not only are 0.3 mm CSPs literally closer (as in pitch), they are also figuratively closer (as in entry into mainstream manufacturing) than most would like to think. The truth is that in 12 to 18 months, 0.3 mm CSPs will be incorporated into high-volume SMT production for certain applications--like it or not! Already, major cellphone manufacturers are evaluating these devices and the corresponding manufacturing requirements necessary to push them into the mix for high UPH (units per hour) production.

If you think moving from 0.5 mm to 0.4 mm CSPs was challenging, just wait until your BoM includes 0.3 mm chips. To help customers prepare for the inevitable, we undertook a study to investigate the printing process as it relates to 0.3 mm CSPs. The results were interesting, indeed; some you'd expect, some not so predictable.

To obtain a complete picture of manufacturability and optimum screen-printing parameters, it was important to analyze the entire process--print, place, reflow and build quality--not just look at printing and deposit volumes. Remember, a process like this would have been in the low-volume semiconductor realm not so long ago, so understanding all implications for high-volume SMT assembly was imperative. For solder, we used type 3 and 4 pastes from the same supplier. As type 3 paste is most common in mainstream SMT, we thought it prudent to analyze its behavior, even though we anticipated it would not be a high-volume choice. And, obviously, the type 4 paste's finer particle size was expected to be more conducive to ultra-fine-pitch printing. It also bears noting that we intentionally excluded type 5 paste because, at this point, the cost of the material is likely to outweigh the benefits. Stencil selection included three laser-cut stainless steel stencils (two of superior quality at thicknesses of 50 [micro]m and 75 [micro]m, and one of lesser quality at 75 [micro]m) and a 66 [micro]m electroformed stencil of very high quality. Because stencil production method and laser precision are critical to the overall deposit result--for shape, volume and location--we added the inferior stainless steel stencil as a measure of comparison against the higher quality stencils. Finally, aperture dimensions ranging from 125 [micro]m to 210 [micro]m were used to determine the most efficient ratios.

As stated, many of the results were expected and confirmed experimentally. First, as projected, type 4 solder paste was the clear winner from a materials perspective. The experiment proved that type 3 paste can be used in certain situations (a 50 [micro]m mask thickness with 180 [micro]m aperture dimensions), but it is not a high-volume alternative. The 50 [micro]m stencil would most certainly cause insufficient material volumes with other devices on the board and would not comply with heterogeneous assembly requirements. Second, it became apparent early on that the stainless steel stencil produced using an inferior laser delivered poor results with numerous I/C failures--even with type 4 solder paste--indicating that stencil quality and manufacturing method must be well-defined for a process this precise. We expected the electroformed stencil to give superior results and, in fact, it did deliver outstanding, beautifully spherical and repeatable deposits. But, surprisingly, the well-manufactured stainless steel stencil produced very acceptable results, with a price/performance ratio more in line with low-cost, high-volume production. Finally, the study confirmed that the 0.66 aperture ratio rule can be broken, but only when suitable materials and inputs are used in the print process.

What's the magic recipe? While further analysis is necessary to evaluate the results of varying squeegee blade material types and delve deeper into some inconsistencies we found with nitrogen versus air reflow atmospheres, we have proven that 0.3 mm CSPs can be produced in a repeatable, high-volume environment. The material set that is best suited to 0.3 mm CSP production is as follows:

* Type 4 solder paste.

* 75 [micro]m thick stencil.

* Extremely high-quality stencil--either stainless steel or electroformed.

* Aperture dimensions ranging from 170-190 [micro]m.

* Stable, high-accuracy printer platform.

Although we are confident that our "magic recipe" provides the optimum ingredients, putting them together for a five-star result will take tremendous expertise and superior process control. Depositing 170-180 [micro]m dots of material in a high-volume SMT environment is no easy task, especially when you consider all the materials have to be compatible with the other components on the board. Oh yeah, let's not forget that this has to happen at sub-7 sec. cycle times! Pretty incredible when you think about how far we've come in the past few years. Keep looking over your shoulder: 0.3mm CSPs will be on your back in no time and you'd better be ready.

Clive Ashmore is global applied process engineering manager at DEK (dek.com); [e-mail]. His column appears bimonthly.

[ILLUSTRATION OMITTED]