Flip Chip

Flip Chip Package-in-Package (fcPiP)

STATS ChipPAC’s Flip Chip Package-in-Package (fcPiP) is an innovative family of 3D packages that stack minimally packaged die and bare die into a single molded package. A pre-tested Internal Stacking Module (ISM) Land Grid Array (LGA) and one or more bare die are stacked with at least one bare die connected to the substrate using flip chip interconnection. The flip chip die is underfilled using a special fine fillet underfilling process. The ISM is inverted and attached to the top surface of the Flip Chip device leaving exposed bond fingers for subsequent wirebond interconnection to the base substrate. The wirebondable bare die is attached onto the top surface of the ISM-LGA. Finally both the ISM-LGA and bare die are wirebonded to a common base substrate. This is followed by a single step overmolding process.

A typical fcPiP integrates an ASIC or DSP logic device (generally with flip chip interconnection) with memory die(s) (in the ISM configuration) and analog die(s) (in bare die configuration), the latter two interconnected to the common substrate using wire bond interconnection. A typical package could have a footprint of 15x15 mm, 1.4mm maximum thickness and incorporate a 0.5 or 0.4mm ball pitch.


3D packaging is driven by wireless and consumer products that require package level functional integration in the smallest footprint, lowest profile and lowest cost. Stacked die packages for Flash, SRAM and DRAM memories in an FBGA package footprint using wirebond interconnection are widely available today. However, with the onset of more complex device designs, two new challenges have arisen: (a) higher I/O density and performance requirements on logic chips dictate the use of flip chip interconnection, and (b) the complexity of memory chips makes it difficult to obtain them in “known good die (KGD)” form.

fcPiP addresses these concerns: the use of flip chip interconnection for the logic die addresses the need for high I/O density and low parasitics, while the procurement of memory in the form of pre-tested ISM packages ensures that the memory devices are “known good.”

An additional benefit is the high level of reliability made possible by the overmolded structure which is typically superior to flip chip–only packages. In addition, the fcPiP configuration obviates certain problems associated with the competing configuration known as fcPoP, such as warpage management and SMT process difficulties associated with mounting the PoP top and bottom packages.


  • Combines flip chip and wirebond interconnection for ultimate in performance and functional integration
  • Package-in-Package configuration facilitates use of pre-tested memory and high module yields
  • Mechanical form, fit and function identical to JEDEC standard FBGA packages
  • Package/die stack: 1 flip chip die + 1 ISM + 1 or more bare wirebondable die
  • Package/die and passive component integration: 2 flip chip die side by side + ISM + passive component
  • 10 x 10mm to 23 x 23mm body sizes
  • Package body size can equal ISM LGA size + 2mm
  • Overall package height including solder balls at 1.2mm and 1.4mm max
  • 0.4mm to 0.8mm ball pitch
  • Flip chip with 28 / 45nm Si node ELK and lead-free bump
  • Logic, Memory and Analog functions can be combined
  • JEDEC standard package outlines available for ISM & FBGA
  • Thin die capability down to 100mm Si thickness
  • Thin mold cap down to 200µm for ISM LGA
  • Higher memory capacity ISM from 256MB to 512MB/1GB
  • Low loop wirebonding, <50µm>
  • BOL, nSOP, MUF for lower package cost structure

End Applications

  • Portable electronics (Cellular phones, Gaming, PDAs, Digital Cameras, Camcorders, Wireless products)
  • Integration of digital base band, memory and analog functions in a high end cellular handset architecture
  • Mixed technology integration at package level (SIP) as a trade off for SOC designs where die yields are compromised by Si-level integration of disparate functional blocks
  • Any 3D application requiring the use of flip chip interconnection for performance reasons


Print page