On this page
Gold on the drawing is an incomplete bonding specification. A fired paste surface, a plated finish and the wire attached to them are distinct materials with distinct histories. The useful selection asks whether a named conductor can form the intended bond on the actual supported stack and retain that connection through the remaining assembly and service conditions.
Key design decisions
- Name the wire or ribbon and bond method before interpreting bondability language.
- Evaluate the processed gold surface at the stage when bonding actually occurs.
- Compare bond results using matched geometry and recorded failure locations.
Describe both sides of the intended bond
Write an interface description containing the conductor product, wire or ribbon material, dimensions, bonding method and the order in which the two ends are made. A gold wire landing on a fired gold pad is not the same material pair as an aluminum wire landing on that pad. The other end may also use another metallization, so the complete interconnection can contain a different aging-sensitive interface even when the ceramic end is well matched. Include both endpoints in the review.
Separate thermosonic, ultrasonic and thermocompression processes in the candidate evidence. Their combinations of deformation, vibration and heat differ, and a favorable result from one method does not define the operating window of another. Identify which bond belongs on the ceramic, which tool approaches it and whether a stitch, wedge or ball footprint is expected. This description gives a material supplier enough information to discuss an actual interface instead of responding to an undefined request for bondable gold.
Read gold-paste compatibility as a conditional statement
A conductor data sheet can support a candidate without qualifying the proposed bond. A mixed-bonded gold formulation may identify particular companion conductor and dielectric combinations, while another product description explicitly addresses gold-wire bonding. Such statements narrow the material question; they do not establish suitability for every wire alloy, bonding tool or substrate beneath the pad. Obtain clarification when the intended wire or stack is absent from the supplier's scope, and preserve that distinction in the trial definition.
Preserve the specific document and material revision used for screening. A generic gold percentage does not describe the fired surface's glass distribution, roughness or adhesion to the layer below. Nor can bulk gold properties substitute for measurements on a porous or compositionally complex printed film. Place unverified items beside the candidate, particularly the actual substrate and post-fire treatment. A technically plausible candidate should advance to a representative trial with those questions identified, rather than inherit a blanket approval because another gold circuit bonded successfully.
Select against the last surface operation before bonding
Trace the surface from conductor firing to the moment the tool touches it. Further glass firing, cleaning, probing, package attachment and storage can each change the relevant condition. A freshly fired witness can be useful for understanding the material, but it does not represent a pad that later contacted a tray or received assembly residue. Identify the last operation that can reach the landing area and prepare selection specimens through that operation.
Do not introduce a cleaning treatment merely to maximize the initial test force. The treatment must fit the gold formulation, underlying layers and later assembly. If a cleaning trial is necessary, retain an untreated comparison and examine whether the treatment changes the film or only removes contamination. Probing should also respect the intended bond region: a scratch or transferred material at the landing location can make the selection test a test of probe damage. Document permissible contact zones and keep diagnostic measurements outside the landing area where practical.
Reproduce the support that receives bonding force
The same gold material can experience different deformation when the ceramic is supported on a flat carrier, spans a cavity or rests on a compliant adhesive. Draw the support present during bonding, including the location of stops, vacuum contact and already installed components. The final enclosure support may be irrelevant if it is added after wire bonding. A candidate must be tested in the condition that actually reacts the tool load.
Inspect local film continuity and substrate condition around the intended landing region. Pad thickness and topography can change tool engagement, while a nearby overglaze boundary can interfere before the tool reaches the nominal center. These are reasons to use the real constrained pad in the selection coupon. Broad open pads may simplify initial process development, but they cannot establish tool access or ceramic integrity at a narrow production feature. Keep support changes distinguishable from material changes so an improvement caused by the fixture is not attributed to the gold paste.
Normalize the comparison geometry before ranking pull forces
A wire-pull reading is a force applied to a loop. Its relationship to tension at either bond depends on loop shape, hook position and pull direction. For a simple symmetric model with equal-height endpoints, equal straight legs and a centered vertical pull, equilibrium gives the hook force as twice the leg tension times the sine of the leg angle above the pad plane. Real wire stiffness, deformation and asymmetric loops can require another treatment.
For an illustrative hook force of 10 mN, a 30-degree leg angle gives 10 mN tension in each idealized leg. At a 15-degree angle the same hook force gives about 19.3 mN leg tension. The test has changed its mechanical demand without changing the displayed force. These hypothetical values set no acceptable bond strength. They show why two candidate gold surfaces should not be ranked from pull numbers collected with different loops, and why a geometry correction alone cannot identify the failed interface.
F_hook = 2 T_leg sin θ; T_leg = F_hook/(2 sin θ)
- F_hook: centered vertical pull force in newtons.
- T_leg: axial tension in either symmetric wire leg in newtons.
- θ: the angle of each straight leg above the common bond plane.
Quasi-static symmetric two-leg equilibrium with negligible bending stiffness, equal-height endpoints and no hook friction or asymmetric deformation. It is a comparison model, not a bond acceptance criterion.
Find a usable formation window instead of a single strong bond
A candidate that works only at one carefully tuned setting may provide little room for the expected variation in surface and assembly. Evaluate a planned region of the bonding variables that the responsible process owner can control. Keep wire, tool condition and support identified while examining the response. The purpose is to find a reproducible combination that forms the bond without unacceptable wire deformation, film damage or ceramic damage, not to maximize one destructive-test reading.
Observe the bond footprint alongside electrical and mechanical results. A larger footprint may indicate more deformation, but it can also approach the usable pad boundary or produce a damaged heel. Record changes in failure location across the candidate settings. Where all high-force results end in substrate damage, the available evidence does not establish a favorable gold interface window. Refine the construction or process with the appropriate owner and repeat the relevant comparison. Machine settings and permissible ranges remain project-specific; no universal bonding recipe follows from the material name.
| Selection question | Keep comparable | Record separately |
|---|---|---|
| Does the interface form consistently? | Wire, tool, surface history and support | Footprint and missed-bond observations |
| Does a pull comparison favor one candidate? | Loop geometry, hook location and pull direction | Force distribution and failure location |
| Does stronger formation damage the stack? | Local pad and ceramic geometry | Film lift, heel deformation and substrate damage |
| Does the connection remain suitable later? | Assembly and environmental sequence | Electrical change and aged failure modes |
Validate the selected interface after the remaining assembly history
Reserve specimens for confirmation after the candidate window has been established. Carry them through the subsequent heat, encapsulation or other operations relevant to the product, then apply the agreed environmental sequence. Initial formation and retained connection behavior are different observations. Keep bonds from the same circuit and process session associated in the data so numerous tests on one favorable substrate do not appear to demonstrate independent material-lot repeatability.
Classify wire break, heel break, interface lift, separation within the printed film and loss of film from the underlying layer individually. A shift toward pad lift after conditioning challenges a different part of the stack from wire weakening. Intermittent resistance under a controlled mechanical disturbance may reveal a connection problem that a static continuity reading misses. Preserve the original orientation and examine relevant specimens before destructive testing removes the surfaces. Confirmation should support the specific failure consequences and service conditions of the circuit, not a generic reliability description.
Specify the gold choice with its supported bonding boundary
The material decision should identify the accepted gold product, processed surface, substrate stack, wire specification, bonding method and representative process window. Link the conclusion to the actual coupon and product geometry used to demonstrate it. If only open-pad trials exist, record the constrained production pad as an outstanding item. If a requested aluminum-wire route has only gold-wire evidence, preserve that gap rather than treating the conductor's gold content as proof of equivalence.
Provide purchasing and assembly with the changes that reopen the decision: conductor revision, wire metallurgy, surface preparation, pad support, tool geometry and later thermal exposure. A replacement material with similar resistance can still present another bonding surface. Conversely, a changed wire loop can alter the mechanical test without changing the gold. Keeping those changes visible makes future comparison possible. The resulting quotation can specify a focused evaluation or a documented construction; it should not imply an unrestricted gold bonding capability, allowable current or lifetime from the screening work.
Provide the complete gold bonding interface
A useful selection package describes the wire, both endpoints and the supported surface at bond formation.
- Gold paste candidate and substrate/layer stack with fired pad geometry and all cleaning, probing, storage and assembly steps before bonding.
- Wire or ribbon specification, both endpoint metallurgies, bond method, bond order, tool approach and intended footprints.
- Ceramic and package support during bonding, proposed process-variable study and restrictions on film or substrate deformation.
- Loop geometry, mechanical-test method, electrical observations, failure-location records and required post-assembly environmental validation.
The drawing-upload form loads as you reach this section.