On this page
A neat die-attach fillet does not guarantee a clean wire-bond region. A thin resin-rich halo can extend beyond the visible filled adhesive and approach nearby pads without producing an electrical short. The engineering question is whether that migrating surface film reaches a region needed for the next assembly operation, and which combination of adhesive state, surface condition and elapsed time controls its movement. Evaluate that question separately from die coverage, bondline thickness and the external fillet size.
Key design decisions
- Map the filled adhesive boundary and the resin-rich halo as separate features, referenced to the actual bond landing region.
- Compare equivalent material and surfaces while changing one defined variable; adhesive lot, surface preparation and waiting time must remain traceable.
- Verify the downstream bond interface after any corrective treatment rather than accepting a visually smaller halo as the only outcome.
Separate squeeze-out from resin-rich migration
Bulk adhesive displacement occurs when the applied material moves under placement force and forms a fillet. Resin bleed involves a mobile resin-rich portion extending beyond the filled body. A clear or amber perimeter mark can be consistent with that process, but appearance alone does not identify its chemistry. Preserve the original surface and compare it with the known adhesive and previous assembly states before assigning the cause.
For ceramic hybrids, distinguish bare ceramic, fired conductor, plated finish and protective glass along the path. The halo may not advance equally over these surfaces. A measurement taken only along the widest bare-ceramic direction can miss a shorter but functionally important path toward a bond pad. Do not treat a fired gold pad as equivalent to a gold-plated finish merely because both appear gold in a photograph.
Measure to the usable landing region
Define where the wire bond actually needs clean metal, including its footprint and placement allowance. The nearest pad edge may not be the correct functional boundary when only part of the pad is used for bonding. Conversely, reserving only the nominal wire contact can ignore the tool's working region or a subsequent second bond. Use the approved assembly geometry to define what must remain untouched.
Measure distance from the observed halo front to that protected region at the same orientation and magnification. Record the edge-selection method for a faint halo. A change in illumination can move the apparent boundary even when the material is stationary. Retain both the raw image and the measured overlay so later reviewers can distinguish a real advance from a different image threshold.
Build a directional clearance budget
Consider a hypothetical nominal distance of 0.40 mm from the intended filled-adhesive boundary to the protected bond region. If placement and bulk spread can consume 0.08 mm in that direction, observed resin extension is 0.12 mm, and an additional 0.03 mm is reserved for measurement uncertainty, the remaining conservative geometric margin is 0.17 mm. Each number is an explicit example input, not a recommended assembly clearance.
If a later qualified observation finds 0.25 mm of resin extension under a different waiting condition, the same arithmetic leaves only 0.04 mm. This result identifies sensitivity to the waiting condition; it does not prove the remaining metal is clean enough to bond. Do not add placement variation twice if the extension measurement already starts from the actual displaced adhesive boundary. State the origin of every distance before summing allowances.
m = g - a - b - u
- g: nominal directional separation between the intended adhesive boundary and protected bond region, in mm.
- a: permitted placement and bulk-spread movement toward that region, in mm.
- b: resin extension beyond the corresponding filled-adhesive boundary, in mm.
- u: additional conservative measurement allowance, in mm; m: remaining geometric margin.
A worst-direction geometric screen with explicitly non-overlapping allowances. A positive margin is not a chemical-cleanliness or bond-strength acceptance result.
Follow the same deposit through the relevant delay
Choose observation times from the actual assembly queue: immediately after placement, before the planned cure begins, after cure and before wire bonding. Use the same specimen orientation and image settings. This reveals whether the critical change occurs during an uncured wait, during thermal exposure or during a later handling step. A final image alone cannot establish when the surface changed.
For a batch operation, include the first and last placed devices because they can have different pre-cure residence times. Record the actual substrate temperature and adhesive preparation history, not only room temperature and a programmed oven setting. Avoid extending an uncured wait beyond the adhesive's approved processing conditions merely to make a dramatic comparison. A diagnostic experiment still needs a bounded, compatible material process.
Separate adhesive state from surface history
Use a small comparison matrix that preserves at least one unchanged condition. For example, compare the same adhesive preparation on two traceable surface histories, then compare two permitted waiting intervals on the unchanged surface. Replicate the relevant combinations rather than making one deposit per condition. A lone favorable dot cannot distinguish an interaction from local surface variation.
Keep applied quantity and placement conditions stable while examining resin migration. Reducing material volume can shrink the visible affected area while producing inadequate die coverage, which is a different failure. Retain the bondline and die-attach acceptance checks alongside the halo observations. The correction must preserve the intended joint rather than exchanging a visible surface problem for a hidden attachment defect.
| Variable under review | Keep controlled | Record before choosing a correction |
|---|---|---|
| Permitted waiting interval | Adhesive preparation, surface and dispense quantity | Halo position versus elapsed time on each relevant surface |
| Surface preparation | Adhesive state, placement and cure sequence | Migration direction and the downstream bond response |
| Adhesive preparation or lot | Substrate lot and storage history | Filled-body shape, halo extension and joint coverage |
| Compatible cure route | Initial deposit and waiting interval | Actual temperature history, final halo and cured joint condition |
| Pad proximity | Same material route and observation method | Clearance to the protected landing region, not only total stained area |
Keep corrective treatments inside material limits
Surface condition, adhesive rheology and cure history can influence resin bleed. That does not justify applying the hottest available cure or an aggressive cleaning treatment to every hybrid. The die, metallization, resistor, overglaze and any installed components impose their own limits. Select a proposed change from the compatible process window and verify its effect on both attachment and subsequent bonding.
A cleaning or plasma step can also alter the surface that controls adhesive wetting. Do not assume that a visually cleaner surface must reduce migration. Compare the actual construction after the treatment, including any newly exposed interfaces. If a changed adhesive is considered, reassess its electrical, thermal and mechanical role in the joint; lower visible bleed alone is not sufficient material-equivalence evidence.
Test the bond region after the full sequence
Inspect and evaluate bonds made after the complete intended die-attach sequence. Compare a region near the adhesive with an appropriate unchanged region where the construction permits. Hold wire, tool, support and bonding setup consistent so a process adjustment at the bonder does not hide the influence of the surface condition being investigated.
Record bond location and failure interface, not only a maximum force. A passing bond located well away from the halo does not qualify a closer landing region. An initial continuity reading also cannot establish mechanical integrity at a contaminated interface. If the trial changes from interfacial separation to wire fracture, retain that distinction when interpreting the numerical result and the remaining uncertainty about the affected region.
Define the decision for affected assemblies
When a halo reaches a protected region, preserve affected specimens before attempting cleanup. Mark the assembly identities, nearest bond regions, adhesive preparation and time history. Keep uncertain parts separate from those with demonstrated acceptable interfaces. Whether rework is permissible depends on the component, adhesive and bonding requirements; removing a visible stain is not by itself a rework qualification.
The final process record should connect a controlled adhesive condition, surface route, placement envelope and allowed queue history to the demonstrated bond outcome. Reopen the comparison when a finish, cleaning method, adhesive preparation or cure route changes. This turns a recurring visual complaint into a specific interface-control problem, with measurable spatial boundaries and a downstream functional check.
Provide the resin-bleed and bond-region evidence
Send time-linked images and the adjacent bonding requirement so the relevant migration path can be evaluated without confusing it with ordinary adhesive squeeze-out.
- Die, filled-adhesive footprint, wire-bond landing regions and permitted assembly clearances on one coordinate drawing.
- Adhesive identity, lot, preparation, applied quantity, placement conditions and actual waiting and cure histories.
- Substrate and finish identities, overglaze boundaries, surface preparation and storage conditions.
- Images before and after the critical process steps, halo measurements, joint-coverage observations and bond results with failure locations.
The drawing-upload form loads as you reach this section.