On this page
Finish selection for wire bonding is a decision about a named wire and bond mechanism. The relevant surface must support the intended first or second bond without transferring unacceptable damage into the fired pad or ceramic. Comparing attractive pull-force numbers before matching wire, loop geometry and test configuration can lead to the wrong finish choice.
Key design decisions
- Treat wire material, diameter and bond mechanism as eligibility requirements for the finish.
- Separate the ceramic-side bond from the other end of the wire when interpreting test results.
- Compare finish candidates using a defined wire geometry and recorded failure locations.
Start with the wire specification and the actual bond
Identify whether the design uses round wire or ribbon, its material and dimensions, and the intended bonding mechanism. Gold, aluminum and copper systems do not present identical surface and process questions. A material described as wire bondable may have been evaluated using only one wire and one technique. That description is a reason to examine its documentation, not permission to exchange the wire freely. Preserve the particular first-bond and second-bond requirements when reviewing the candidate finish.
The die pad can constrain the wire choice before the ceramic finish is selected. Review that endpoint with the device supplier or assembly authority, then trace the same wire to the ceramic landing. A ceramic pad with excellent local behavior does not solve an incompatible die-side joint. If two wire options remain open, describe them as two complete bond systems and compare their package consequences separately rather than assuming that a finish ranking survives the wire substitution.
Specify the surface encountered by the bonding tool
A directly bonded fired conductor and a conductor carrying an additional deposited finish have different interface inventories. The fired route depends on its processed metal and bonding constituents; the added finish also introduces adhesion between deposited layers and the underlying pad. Name every relevant layer and the condition in which it reaches bonding. A color description cannot identify whether the tool touches a fired surface, plating, a thin film or contamination over any of those constructions.
The eligible route must fit the actual ceramic and the full layer sequence. A supplier result on one alumina grade supplies useful screening information but does not settle another substrate or an altered series of refires. Compare the final pad after component attachment and cleaning with the documented material state. When the evidence does not cover that history, reserve the candidate for a targeted trial instead of describing it as approved because the original metal name still appears on the drawing.
Avoid attributing a wire failure to the wrong finish
A pull test loads both ends of a wire, and the result can be limited by the die bond, the ceramic bond or a damaged section of wire. Record the complete bond map and identify which endpoint uses the finish under selection. A break elsewhere can show that the ceramic bond survived that particular loading event, but it does not measure its ultimate strength. That distinction becomes especially important when the finish candidates produce different limiting locations.
Keep the observation before and after pulling. Footprint shape, heel condition and pad attachment can explain why one specimen fails differently from another. If a candidate produces more wire breaks and fewer interface separations, report that change alongside force rather than treating all failures as interchangeable data. A process that simply weakens the wire can make the ceramic bond appear stronger by moving the failure away from it; the finish decision must preserve the intended integrity of the complete interconnect.
Compare the wire size behind a breaking-load record
A finish comparison must retain the cross-section of the wire used to produce its supporting data. For a straight, round wire carrying known axial tension, average tensile stress is force divided by the circular cross-sectional area. This is a material-test reporting quantity. It is not the local stress under a bonding tool, and a hook force applied to a completed loop cannot be inserted as the axial tension without resolving that test geometry separately.
Consider a hypothetical axial load of 0.120 N. A round wire of 25 micrometers diameter has an area of about 491 square micrometers and an average tensile stress of approximately 244 MPa. At 32 micrometers diameter, the area is about 804 square micrometers and the corresponding stress falls to roughly 149 MPa. The same force therefore represents different demands on the two wires. Neither value is an allowable wire or pad stress. The example explains why a finish supported by a small-wire breaking-load record needs new evidence when another wire size, deformation response or bond footprint is proposed.
σwire,avg = T/Awire = 4T/(πd²)
- T is known axial wire tension in N, not an unresolved wire-loop hook reading.
- d is the diameter of an assumed solid circular wire in mm.
- Awire is cross-sectional area in mm²; σwire,avg is in N/mm², equivalent to MPa.
Uniform axial loading of an intact round cross-section. Local bond deformation, necking, residual stress, wire texture and hook mechanics are excluded; this is not a bond-interface strength or material acceptance limit.
Keep material eligibility separate from process evidence
The selection grid should distinguish a documented material match from a demonstrated assembly match. A finish may be chemically plausible for the wire but lack evidence at the required diameter. Another may have a relevant material history but require an unsupported package arrangement during bonding. Do not collapse those differences into a single score whose weights can hide a mandatory requirement. Remove an incompatible route or identify the missing evidence before spending effort on fine ranking.
Use the narrowest valid description for each result. A successful ball shear observation supports a different loading configuration from a double-wire pull. A coupon with a flat, fully supported ceramic landing does not reproduce a pad spanning a poorly supported package region. The comparison remains useful when these boundaries are explicit: the trial can then be designed to close the actual gap instead of reproducing a convenient demonstration that leaves the decision unchanged.
| Gate | Evidence to match | Reason a candidate remains conditional |
|---|---|---|
| Wire identity | Material, alloy designation, diameter or ribbon section | Documentation concerns another wire |
| Bond mechanism | Ball or wedge and the ceramic-side bond role | Only the opposite bond configuration was evaluated |
| Exposed finish | Layer stack and final processed state | Additional finish or cleaning changes the contacted layer |
| Support during bonding | Ceramic and package constraint at the landing | Coupon support differs from the assembly |
| Mechanical assessment | Method, loop geometry and failure location | Published force cannot be compared directly |
| Downstream compatibility | Encapsulation, storage and required environmental sequence | Only immediate bond formation was checked |
Recognize damage hidden by stronger bonding settings
Increasing force or ultrasonic input may change the footprint and the failure location while producing a larger measured number. That is not sufficient reason to select a finish. Examine whether the resulting pad remains attached and whether the wire heel or ceramic beneath it has been damaged. Failure signatures include a crater, a lifted metal edge and excessive wire deformation; record each separately even when initial continuity remains satisfactory. The selected process must satisfy all relevant criteria simultaneously.
Avoid an uncontrolled search in which each candidate receives a different surface cleaning, support fixture and wire geometry. Such a comparison selects a bundle of changes, not an isolated finish improvement. Sometimes a distinct process is necessary for a candidate; record it as part of that candidate rather than pretending the comparison held it constant. Preserve unsuccessful attempts and the reason each adjustment was made so the final choice reflects a reproducible route instead of the single bond that happened to survive.
Validate the ceramic endpoint through the complete package sequence
Prepare specimens that include the intended ceramic landing, neighboring topography and support used at bonding. Identify the wire lot, tool, process program and surface history without inventing settings from a generic guide. Keep an appropriate stable comparison in the run to detect drift in the bonding setup. The evaluation should include observations before destructive work, because the original footprint and electrical condition cannot be reconstructed after a bond has been pulled away.
Continue the comparison through the package operations that can change the bond system. Encapsulation, cure, lid installation and cleaning may alter the local environment or constrain the loop. Select environmental and electrical loading from the actual product requirements, and specify whether measurements occur during exposure or after return to a defined state. Retain specimen-level force, failure location and electrical events together. A batch average cannot show whether one candidate develops localized pad separation while another remains limited by the wire.
Release one wire, finish and bonding configuration
Write the approved wire specification and ceramic-side bond role beside the finish definition. Link the pad drawing to the final surface preparation, permitted handling, bonding support and applicable evaluation method. The purchase description should make clear which substitutions reopen the decision. Changing diameter, wire supplier formulation, added finish or package support can invalidate evidence even when the drawing continues to call the connection a gold wire bond.
The RFQ should include the die-side constraint as well as the ceramic pad. Provide the intended current and pulse duty, loop envelope, neighboring components and downstream materials so an eligible finish can be evaluated in context. Existing results are most useful when they retain the test geometry and endpoint that failed. With those details, the technical review can distinguish missing material compatibility from missing process evidence and define the smallest representative trial needed to resolve the selection.
Specify the complete wire and finish system
Provide both bond endpoints and the evidence needed to compare the ceramic-side finish.
- Wire material and exact diameter or ribbon dimensions, die-pad metallurgy and first/second bond assignment.
- Ceramic pad stack, final finish, full processing history and allowed material alternatives.
- Bond map, loop shape, tool approach, neighboring topography and support during bonding.
- Pull or shear records with hook geometry, failure location, electrical observations and required post-package validation.
The drawing-upload form loads as you reach this section.