ABSTRACT:

Embedded actives are to bury thinned active chips into package substrates, as opposed to surface mounted devices (SMDs), which can achieve smaller form factor, better electrical performance and higher functionality than the SMD technology. While many embedded actives have been explored so far, they are based on chip-first and -middle approaches, in which the active chips are embedded before and during the build-up processes of package substrates, respectively. The most concern with those two current approaches is the loss accumulation associated with the build-up layer processes carried out right on top of the embedded chips, which is highly likely to lose the embedded chips during their packaging process. The reworkability to replace the faulty chips embedded with good ones and thermal management of the embedded chips are also issues since the embedded chips are totally surrounded by hard-cured polymers. In this paper, chip-last embedded active has been proposed to address some of the issues that are reported in current chip-first and -middle approaches, in which chips are embedded after all the package substrate processes including the build-up layers are completed, just like conventional SMD packaging. In the chip-last approach, a cavity is introduced within the build-up layers of package substrate and a chip is directly embedded into the cavity. A first proto-type of the chip-last embedded active will be demonstrated by developing various cavity formation processes within the build-up layers and then embedding 100 $mu{rm m}$ thick chips into the defined cavities.