By Dirk. M Baars, Rogers Corp.
The expectation that embedded passives technology will be required to meet size, performance and cost requirements of devices has been a driver of considerable R&D activity in many materials suppliers’ laboratories. In particular, the proliferation of multiple-function handheld devices with multiple wireless functions should require that discrete components be integrated into the package to save precious space and minimize electrical distances for optimum performance. Our scientists have been feverishly developing new materials that expand the use of low-loss dielectric technology down to the sub-50-µm thicknesses desired in packaging substrate layers. But, so far, there seems to still be very little actual market adoption of these materials. Each new generation handheld or wireless card continues to come out with just more of the same when it comes to incorporation of needed passive functions.
Open up the iPhone and you’ll find that discrete passive components are in fact multiplying inside these latest handhelds. Granted, these discretes have become so small, the can only be handled by the daintiest of fingers or robots. Continuing shrinkage in size and cost of discrete devices certainly is a major reason that embedding passives has not become the norm. But reading SMT magazine tells me, however, that issues with mounting these little specks onto the circuit board are requiring more precise and expensive equipment and putting downward pressure on yields.
Embedded inductors, capacitors and resistors were all first deployed around 1999, which means that the technology is not new. What else may be holding up wider adoption of embedded passives into package substrates? Experience suggests that the key bottleneck may be the lack of design experience and necessary design libraries that hinder progress. Since the passive element is no longer a discrete device with specified performance, but rather a printed element on a circuit layer, there are no specifications available to characterize the element’s performance. Furthermore, when various elements are placed in close proximity, their fields may interact with each other and need further model refinement to fully predict performance. The learning curve may lead to lower yields in initial implementation attempts may discourage broader implementation.
Embedded passives have been the norm when you look at what is going on defense and aerospace. Here, the performance gains achieved are key, and ceramics materials are the standard materials of choice typically processed in panels barely larger than the size of today’s semiconductor wafers. Design libraries have been developed by individual players in these segments but are not broadly available and in some cases, maybe even classified. Even where the information is publicly available, the solutions are often too expensive for application in the commercial sector.
Commercial millimeter wave applications such as automotive radar, high-frequency imaging, and emerging potential of 60GHz wireless applications are driving the embedding functions for the same performance reasons, but are increasingly looking to organic materials as a way of increasing production scale and reducing total cost. It may indeed be these emerging high-end commercial applications that will finally cause the chasm to be crossed between niche early adoption and broad commercial acceptance.
If you have any thoughts about what is needed to proliferate the use of embedded passives into semiconductor package substrates, drop me a line. In the meantime, we’ll continue to get ready to supply the materials to help make it happen.
DIRK M. BAARS, director, advanced materials group, may be contacted at Rogers Corporation, 1 Technology Dr., Rogers CT, 06263; 860/779/4772; E-mail: [email protected].