News

Flip chip is a method of chip attachment to a substrate with essentially no conventional package. Here are the methods used by the EMPF* to assemble and validate flip-chip technology.

1. Decide if flip chip is warranted for the design. Flip-chip attachment is used for high I/O devices where there is insufficient peripheral length to accommodate bond pads for all of the I/Os needed for the function of the die. This often occurs in DSPs or microprocessors with over 1,000 I/O connections per chip. However, flip chip is also used in some low I/O count chips that require very good signal integrity and a very low impedance to each I/O. This is known as low parasitics. Flip chip is also used when size is at an extreme premium (airborne applications, or medical products such as “in the ear canal” hearing aids). Figure 1 illustrates several types of flip-chip assembly.

2. Select the IC die (chip) to be flip-chip attached. Three key considerations in the selection of a chip (also known as a die) to be used in the flip-chip attachment process are:

Bump type. Is the bump solder or metal? Solder bumps are more convenient if the flip chip is to be attached to the assembly at the same time as other surface-mount components. Only one solder reflow bake process is needed to simultaneously attach the flip chips, conventional packages and discrete parts. Metal-bumped chips could, by contrast, require special solders or adhesives to perform the flip-chip process, which would then require two separate attachment sequences to build the assembly. We prefer solder-bumped die.

Availability. Sometimes the different dies to be attached are not available in solder- or metal-bumped versions. Dies that are made for the normal wire bonding of the I/O pads can be bumped individually using what is called a “stud bump.” This bump is applied to the die pad with a ball-bonding wire bonder. The bond is made to the die pad and then the wire is broken off. A die can be converted in this manner from a face-up, wire-bondable die to a stud-bumped flip-chip die. When the die is flipped, either solder or an adhesive (usually electrically conductive) can be used to make the bond at the stud bumps.

Die “shrinks.” “Shrinking” the die by getting more die per wafer processed and lowering its cost should be considered. If frequent shrinks are anticipated, be advised that the substrate pattern will have to change to accommodate the smaller die pattern. If planned at frequent intervals, however, the costs for new substrates could become an issue.

3. Select the substrate material to be used in the flip-chip assembly. The critical choice here is whether to use ceramic or organic substrate material. Ceramic has a CTE much closer to most IC dies than do common organic materials. An exception to this rule is that some organics, such as liquid crystal polymers, can be engineered to have similar CTE properties to some IC materials. If common organics such as FR-4 epoxy-glass composite must be used, an underfill epoxy will be needed to secure the die to the substrate to limit the CTE mismatch reliability consequences. Many ceramic materials have CTEs low enough to be used for flip-chip substrates without underfill.

4. Select the underfill to be used. This choice can range from none to various underfill materials. Underfill materials are usually filled epoxy systems that have intermediate CTE between the die and substrate. Underfills increase the flip chip’s ability to withstand thermal cycling but add dispense and cure steps to the manufacturing process.

5. Determine the assembly sequence. Without underfill, flip chip (with solder bumps) requires no change in the typical surface-mount process. If underfilling is required, provisions for the dispense and cure of the underfill material are necessary.

6. Rework and repair. If rework and repair of assemblies is contemplated, the flip-chip process with underfill presents a considerable issue. Reworkable epoxy underfill materials are not well proven, and must be selected on a case-by-case basis. Flip chip with no underfill is less of an issue, but either can be accommodated.

Without underfill, solder flip chip presents only the challenge of fine-pitch surface mount. If underfill is required, as for reliability on organic substrates, a more complex assembly process, including dispense and cure, is needed. Rework of underfilled flip chip then presents additional issues, such as removal of cured underfill before die replacement.

*The EMPF (empf.org) aids industry in improving electronics manufacturing processes required to build military systems.

The American Competitiveness Institute (aciusa.org) is a scientific research corporation dedicated to the advancement of electronics manufacturing processes and materials for the Department of Defense and industry. This column appears monthly.