Panel Level Packaging – Where are the Technology Limits?

Abstract

Fan-out Wafer and Panel Level Packaging are two of the dominating trends in microelectronics packaging. Both approaches with different flavors as RDL last face-up or face- down have reached maturity and are introduced in high volume manufacturing. For Fan-out Wafer Level Packaging (FOWLP) clear application trends and technology roadmaps do exist. These range from low density core technology for e.g. RF or PMIC (power management IC) packaging over high density application processor packaging to ultra-high-density applications for networking servers etc.. For panel level packaging it is still not fully clear if the same performance can be achieved as on wafer level as larger process / panel sizes may have higher challenges in process control, accuracy and consistency, material and equipment or handling.Main driver for moving to panel level packaging is of course lowering the packaging cost. More packages can be processed in parallel and panel formats have a much better area utilization (ratio between panel/wafer size and package size) than round wafer shapes. Also, environmentally PLP is advantageous by e.g. lower waste and smaller carbon footprint. However, for both aspects processes with sufficiently high yield are required. This is especially true for FOWLP/PLP RDL last processes as a failure in the RDL will also lead to a loss of packaged die(s).This paper describes current technology developments to access the limits of the panel level packaging technology. Warpage, die shift and fine line capabilities are the main topics here. To better understand the compression molding process as the technological basis of the reconfigured panel and its influence on warpage and die shift a dedicated sensor mold tool has been developed. By integration of temperature, pressure, dielectric and fiber Bragg grating sensors the flowing and curing behavior of epoxy molding compound can be studied in-situ. Results will support process simulations for warpage prediction and more accurate die shift compensation.For large panel processing an adaptive patterning approach might be needed anyhow to achieve a high yield. Here the combination of an intelligent assembly strategy for high speed and sufficient accuracy, capabilities to measure each die position and a maskless lithography process adapting the redistribution layer (RDL) to each die position may lead to a cost-effective high yield process.In addition, a clear trend towards finer lines and spaces as well as smaller via diameters is also demanded for large panel RDL processes. Process developments towards 2 μm lines and spaces and via shrinking on 610x457 mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> (24"x18") panels are shown including material and process options.In summary this paper will show current PLP technology developments for future high-end applications and will cover at the same time economic and environmental aspects.