Underfill in PCB Assembly: When Buyers Should Specify It
Learn what underfill does in PCB assembly, where it improves reliability, how it affects rework and inspection, and what OEM buyers should define before release.
Underfill is used to reduce solder-joint fatigue when package and board expansion rates do not match.
Many boards do not need it. The business case depends on package geometry, environment, and field-failure cost.
Once cured, underfill can turn a routine component replacement into a pad-damage or scrap decision.
The package list, cure window, inspection plan, and rework policy should be frozen before first article.
Underfill is one of those process decisions that buyers usually hear about only after a reliability concern appears. A field drop failure, a thermal-cycling target, or a fragile wafer-level package suddenly forces the team to ask whether the solder joints alone are enough. That is where underfill enters the discussion. It is not a cosmetic step and it is not a generic upgrade. It is a deliberate material and process choice that trades extra assembly complexity for improved mechanical robustness at the package level.
For neutral background, review underfill, ball grid array packaging, and coefficient of thermal expansion. In a real PCB assembly program, those topics matter because the package body and the laminate do not expand at the same rate. Repeated temperature change transfers stress into the solder joints, especially when the package stand-off is small.
Buyers should treat underfill as a reliability control with a cost. The cost is not only material. It includes dispensing time, cure time, inspection effort, cleaner process discipline around keep-out zones, and harder rework. If you specify underfill without understanding that trade, you can improve one failure mode while creating three new production arguments.
"On a 0.4 mm or 0.5 mm pitch package, the solder joint can be electrically perfect and still be mechanically exposed. Underfill becomes rational when the field environment will cycle that joint thousands of times, not when the team simply wants extra insurance."
— Hommer Zhao, Technical Director
What underfill actually does
After the component has been soldered and cleaned as required, a low-viscosity epoxy is dispensed around one or more package edges. By capillary action, the material flows under the package, surrounding the solder interconnects and filling the gap between the package body and the board surface. The assembly is then cured according to the material supplier's profile. Once cured, the material shares mechanical load with the solder joints and reduces the local strain each joint sees during thermal cycling, drop, or vibration.
That matters most for area-array and chip-scale packages where the joints are hidden and short. The lower the stand-off, the less joint compliance you usually have. If the package has a very different expansion behavior from the PCB, fatigue can accumulate quickly. The same logic is one reason buyers already care about moisture-sensitive handling and hidden-joint X-ray control. Package-level reliability is rarely controlled by one variable.
Underfill does not fix a defective assembly. If solder volume is wrong, warpage is uncontrolled, or wetting is poor, encapsulating the site after reflow can hide a process problem rather than solve it. Buyers should expect the package to pass normal workmanship and inspection gates before underfill is even considered beneficial.
When underfill is worth considering
The best reasons to specify underfill are specific, not emotional. Common triggers include WLCSP and fine-pitch CSP devices on handheld products, BGA modules on assemblies that see repeated thermal cycling, industrial electronics exposed to vibration, and automotive or transportation hardware where mechanical shock is part of the life profile. In those cases, the field-failure cost can be far greater than the additional assembly process cost.
Underfill can also make sense when a board has very limited package stand-off, a large package-to-board CTE mismatch, or a drop-test requirement that has already exposed corner-joint weakness. It is less compelling on benign indoor products with moderate temperature swings, roomy package geometry, and low replacement cost. Many buyers overuse the phrase "for reliability" without defining the actual failure mechanism. That is too vague to support a real manufacturing change.
| Scenario | Underfill value | Main benefit | Main penalty |
|---|---|---|---|
| WLCSP in portable device | High | Improves drop and thermal-cycle robustness | Rework becomes difficult and slower |
| Fine-pitch BGA in industrial control | Medium to high | Reduces fatigue from repeated temperature change | Needs clear X-ray and repair policy |
| Large BGA on stable indoor equipment | Case dependent | May improve margin if CTE mismatch is severe | Can add cost with little field return reduction |
| Prototype board with frequent engineering rework | Usually low | Little value during rapid iteration | Component changes become expensive |
| Automotive or transport module | Often high | Improves shock and vibration durability | Requires strict material and cure control |
That table is why underfill should be tied to the qualification plan. Ask what test is being improved: thermal cycling, powered cycling, drop, vibration, or field transportation abuse. If the supplier cannot connect the process to a measurable failure mode, the request may be coming from habit instead of engineering evidence.
"A buyer should never approve underfill as a default note across every BGA. On many programs, only 10% to 20% of the package list is truly fatigue-sensitive enough to justify the extra process and rework burden."
— Hommer Zhao, Technical Director
Process controls buyers should ask about
If you decide underfill is justified, the next question is whether the supplier can run it repeatably. Start with the package list. The drawing or PO should name the exact reference designators or package families that receive underfill. Then define the material family and cure profile. A vague note such as "apply epoxy for reinforcement" is not enough because dispense behavior, cure temperature, glass transition, modulus, and rework behavior vary substantially by chemistry.
Coverage rules matter just as much. Buyers should ask whether the process is corner-bond only, edge-bond, full capillary underfill, or a no-flow material integrated with the soldering sequence. Each choice changes throughput, inspection points, and rework difficulty. On mixed technology builds, also verify that connectors, test pads, optical fiducials, and keep-out areas are protected from stray dispense.
Inspection planning should be explicit. Because underfill can obscure later visual access, many teams align it with prior verification steps such as solder paste inspection, placement review, and X-ray inspection. If a package needs X-ray evidence, decide whether that happens before underfill, after underfill, or at both checkpoints.
| Control item | What should be defined | Why it matters |
|---|---|---|
| Package scope | Reference designators, package family, or BOM line | Prevents supplier assumptions and quote gaps |
| Material | Approved underfill chemistry and revision | Controls modulus, cure behavior, and rework response |
| Cure profile | Temperature, duration, and any ramp limits | Incomplete cure changes long-term reliability |
| Coverage rule | Corner bond, edge bond, or full underfill | Changes mechanical benefit and process time |
| Inspection gate | X-ray, visual boundary check, and retained records | Prevents hidden defects from being locked in |
Rework, scrap, and the hidden cost of getting it wrong
The hardest commercial issue with underfill is not dispensing it. It is deciding what happens after a defect is found. On a non-underfilled site, replacing a package may be routine. On an underfilled site, the same event may require local heating, controlled material removal, higher pad-lift risk, and longer engineering time. For some dense boards, the safest disposition becomes board scrap rather than device replacement.
Buyers should force that conversation before the quote is approved. Ask which underfilled packages are considered reworkable, which are limited-rework only, and which become scrap on failure. That answer affects spare quantity planning, warranty cost, and how much first article evidence you should retain. It also ties directly to whether the supplier has a practical process for removing cured material without damaging adjacent parts or solder mask.
Underfill also changes how engineering teams iterate prototypes. If the program is still tuning land pattern, stencil aperture, or reflow window, early underfill can lock a volatile design into a high-cost debug path. That is why many teams delay underfill until the package and process baseline are already stable.
"If the supplier cannot tell you in advance whether an underfilled device is reworkable, limited to one repair cycle, or automatic scrap, the quote is missing a real cost driver. That answer belongs in first-article planning, not in the failure review meeting."
— Hommer Zhao, Technical Director
Practical recommendation for OEM buyers
Specify underfill only when the package physics and field environment justify it. Then define the package list, material, cure profile, coverage rule, inspection checkpoint, and rework disposition before release. That is the minimum viable control plan. Anything less leaves the supplier guessing, and guesswork is expensive once hidden joints and cured epoxy are involved.
The strongest programs treat underfill as part of the total assembly strategy alongside package MSL control, hidden-joint inspection, workmanship criteria, and repair policy. If you need the reliability gain, buy the whole control system, not just the material.
Teams often add an underfill note after a reliability complaint but leave package scope, cure profile, and repair disposition undefined. That usually shifts uncertainty from field returns to production delays instead of solving the root control gap.
Frequently asked questions about underfill in PCB assembly
What is underfill in PCB assembly?
Underfill is an epoxy material dispensed under area-array or bottom-terminated packages after reflow. It fills the gap between the package and the PCB, then cures to reduce solder-joint strain from thermal cycling, shock, and vibration.
Which packages usually need underfill?
The most common candidates are CSP, WLCSP, fine-pitch BGA, and some QFN or module assemblies where stand-off is low and board-to-package CTE mismatch is high. Portable, automotive, and industrial products with repeated temperature swings are the usual drivers.
Does underfill replace good soldering process control?
No. A weak stencil design, poor paste release, or unstable reflow profile will still create defects. Underfill is a reliability reinforcement after a sound solder joint has already been formed and verified.
Why do buyers worry about underfill and rework?
Because cured underfill usually makes component removal slower, raises pad-damage risk, and often changes whether a site can be reworked economically. On some dense assemblies, the right decision is to scrap the board instead of reworking one underfilled device.
Should every BGA be underfilled?
No. Many BGA programs ship reliably without underfill. Buyers should specify it only when the package geometry, thermal-cycling requirement, drop requirement, or field environment justify the extra process, inspection, and rework burden.
What should be defined on the PO or assembly drawing?
Define the exact package list, underfill material family, dispense coverage rule, cure profile, keep-out or no-flow areas, inspection expectation, and rework disposition. Leaving those items vague is how first-article delays start.
Need help deciding whether underfill belongs in your build?
We can review the package list, reliability target, X-ray plan, and rework implications before underfill becomes an avoidable cost or a missing reliability control.