Underfill and Corner Bonding for BGA Board Level Reliability
A ball grid array joint has to survive thousands of thermal cycles while the package and the board expand at different rates. Underfill is the material that spreads that strain across the whole array instead of concentrating it in the corner balls. It is not a universal remedy, and applied badly it creates as many problems as it solves, but where the fatigue risk is real it is one of the few effective countermeasures.
Why Underfill Exists
When a package heats, the laminate expands more than the silicon, and the difference is absorbed by the solder joints. The outermost balls experience the greatest displacement and fail first, usually as a crack that begins at the package side of the joint and grows with every cycle. Underfill reduces that displacement by mechanically coupling the package to the board.
The material fills the gap between the package and the substrate and bonds to both surfaces. Because the cured resin carries much of the load, the solder joints see far less strain. The result is a dramatic improvement in thermal fatigue life, which is why underfill is standard for large dies, fine-pitch area arrays, and products that must survive automotive or industrial temperature ranges.
Capillary Flow and Material Selection
Most underfill is applied along one or two edges of the package and drawn into the gap by capillary action. The flow depends on the filler particle size, the viscosity, the gap height, and the surface energy of the solder mask and the package. A material that flows quickly may settle unevenly, while one that flows slowly may fail to reach the far corner before curing begins.
Filler content controls the coefficient of thermal expansion of the cured resin. A high filler loading lowers expansion and improves stiffness, which benefits fatigue life, but raises viscosity and makes flow more difficult in a tight gap. Selection is therefore a balance between mechanical performance and the ability to fill the specific gap on the board.
Corner Bond Versus Full Underfill
Corner bonding applies a small amount of adhesive at the corners of the package rather than filling the whole gap. It is faster, easier to inspect, and reworkable, and it improves mechanical shock resistance noticeably. What it does not do is reduce the thermal fatigue of the joints in the middle of the array, because the material never reaches them.
Full underfill delivers the fatigue benefit but costs more, takes longer, and makes rework difficult. The choice follows the failure mode that matters: a product exposed to drop and vibration may be adequately protected by corner bonds, while one that cycles between temperature extremes needs the full treatment. Many specifications allow corner bond on smaller packages and require full underfill above a size threshold.

Dispensing Patterns and Equipment
The dispensing pattern determines whether the material fills the gap uniformly. A simple L-shaped or dot pattern along two adjacent edges is common, with the material drawn across the package by capillary action. The volume must be enough to fill the gap and form a small fillet without flowing onto neighbouring components or into vias.
Equipment ranges from a simple time-pressure dispenser to a volumetric pump with a heated needle and a pre-heating stage. Substrate preheat lowers viscosity and speeds flow, but too much heat cures the material before it has travelled. Programming the pattern, the volume, and the dwell requires iteration against actual flow results rather than a datasheet figure.
Cure Schedules and Inspection
Underfill cures either by heat or by a combination of heat and ultraviolet light. A typical schedule holds the assembly at an elevated temperature for a defined period, and the cure must complete under the package where no light reaches unless the material is dual-cure. Incomplete cure leaves the resin soft, and it can continue to flow or to release volatiles during later thermal excursions.
Inspection is difficult because the material sits beneath the package. Visual checks confirm the fillet and the absence of flow onto adjacent parts, while scanning acoustic microscopy detects voids, delamination, and incomplete fill. Weight gain measurements on witness samples verify cure, and cross-sections confirm that the material reached the far edge of the die shadow.

CTE Mismatch and Mechanical Benefit
The benefit of underfill comes directly from the mismatch between silicon and laminate. If both materials expanded identically, there would be no shear on the joints and nothing for the underfill to do. Because the mismatch is unavoidable, the reinforced joint is the practical answer, and its effectiveness improves as the die grows larger.
Package type matters as well. A large body with a high standoff sees less benefit than a thin package with a low gap, where the joints are stiffer and the stress more concentrated. Ceramic packages, with an expansion coefficient closer to silicon, need less reinforcement than plastic ball grid arrays on a standard laminate.
Rework and Repair Limits
Underfill makes rework much harder. The cured material must be removed around the package before the part can be heated and lifted, and the removal process risks damaging the solder mask, the laminate, and neighbouring components. The site then has to be cleaned and refilled after the new package is placed.
Some materials are designed to be reworkable, softening at a defined temperature so the package can be removed with less mechanical effort. Even then, the process is slow and the yield lower than for an unfilled assembly. Where a design is likely to need rework, that consideration belongs in the material selection, not in the production plan.
Process Control and Defect Modes
Common defects include incomplete fill, voids, voids at the die shadow edge, material on neighbouring pads, and contamination that prevents adhesion. Incomplete fill usually points to insufficient volume, low substrate temperature, or a gap that is too small for the filler particles. Voids come from trapped air or from volatiles released during cure.
Control begins with the material. Store it according to the supplier’s instructions, allow frozen syringes to reach room temperature before use, and respect the pot life once opened. Record the lot, the dispense volume, the substrate temperature, and the cure schedule for each build so that a defect can be traced to a specific material or parameter change.
When Underfill Is Worth the Cost
Underfill adds material, equipment, cycle time, and rework difficulty. It is worth the cost when the package is large, the thermal environment is demanding, the product must pass a drop test, or a field failure would be expensive. For small packages in benign conditions, the added process complexity rarely pays back.
The decision should be based on data rather than habit. Thermal cycling and drop tests on a representative assembly show whether the unfilled joints have adequate margin. Where the margin is thin, underfill or a corner bond buys reliability cheaply; where the margin is comfortable, the process step is simply an extra way to lose yield.
FAQ
How do I know if the underfill has filled completely? Visual inspection only shows the fillet at the package edge. Scanning acoustic microscopy reveals voids and unfilled regions beneath the package, and cross-sections confirm the extent of flow at the die shadow. Weighing a witness sample before and after cure verifies that the material cured rather than merely gelled.
Can an underfilled BGA be reworked? It can, but the process is slow and risky. The material must be removed locally without damaging the mask or laminate, and the site must be cleaned and refilled after the replacement part is soldered. Where rework is likely, choose a reworkable formulation and plan the removal step in advance.
Is corner bonding a substitute for full underfill? Only for some failure modes. Corner bonds improve shock and vibration resistance and reduce corner joint stress, but they do not reinforce the joints under the middle of the die. If thermal fatigue of the central balls is the concern, full underfill is the appropriate measure.



