On this page
A dielectric is selected for the interfaces and process history it will actually encounter. Substrate identity, both conductor levels and every subsequent firing belong in the material decision before an electrical thickness is assigned.
Key design decisions
- Name the substrate grade and the compositions touching each face of the dielectric.
- Confirm the complete print, dry and fire sequence rather than matching peak temperatures alone.
- Use representative crossings and openings to qualify the selected stack.
Start with the substrate the dielectric must follow
The substrate is part of the dielectric specification, even when the electrical drawing shows it only as a supporting rectangle. Record the ceramic composition and grade, surface condition, thickness and any earlier metallization. A dielectric described for one alumina system does not acquire evidence for aluminum nitride, glass or insulated metal merely because it can be printed on them. The candidate must have a credible route to adhesion and dimensional stability on the actual surface. Surface preparation should therefore be included in the comparison, especially when an approved material is moving from bare ceramic to a previously fired layer.
Treat a substrate substitution as a stack change. It may alter contraction during cooling, support of a large coated area and the chemistry presented to the glass phase. A purchasing description that says only ceramic leaves these decisions unresolved. Request substrate-specific compatibility information early, then use the intended substrate in qualification coupons instead of allowing a convenient laboratory tile to become the unspoken reference.
Approve both sides of the insulating layer
Draw the stack in process order and identify every direct contact. The lower conductor must survive dielectric deposition and firing; the upper conductor must form its intended connection on the fired dielectric. Those are different compatibility questions. A satisfactory adhesion result for the lower metal on ceramic cannot stand in for an upper bond pad on glass. Likewise, a dielectric paired with one conductor series should not be accepted with another series solely because both contain silver. Glass additives, metal loading and the thermal route are part of the pairing.
Include the function of the upper feature in this interface map. A passive crossover, a soldered terminal and a wire bond pad impose different evidence needs after firing. Where the upper feature will be attached, ask for data covering that attachment route or schedule a representative attachment trial. Keep the material names tied to layer numbers so a conductor substitution cannot disappear inside a drawing revision that appears to change only a terminal.
Compare complete firing histories
Peak furnace temperature is only one coordinate of compatibility. The sequence also includes drying, whether dielectric prints are fired separately, the time spent at elevated temperature, atmosphere and later conductor or protective-layer firings. Two proposed stacks can share a peak temperature while giving the first dielectric layer different total exposure. Write a layer-by-layer route before comparing materials. Mark which films are already fired at each step and which films are still developing their structure together.
The published preparation route for a candidate is a starting condition, not a license to combine steps. If the material instruction calls for separately fired dielectric layers, a proposed combined firing needs its own justification. Record the reason for any deviation and the property it might affect, such as coverage, conductor interaction or substrate bow. The useful purchasing question is whether the entire intended route is supported. It is not whether each paste has an individually familiar firing temperature on its container label.
Use expansion mismatch as a screening variable
Thermal expansion helps identify a pairing that deserves closer mechanical review. In a simple free-expansion comparison, two materials with different coefficients would change length by different amounts over the same temperature interval. Bonding constrains that freedom, but the difference alone does not calculate stress, crack probability or acceptable film thickness. Elastic properties, geometry, glass relaxation and the actual strain-free state would be needed for a more complete model. Do not automatically make the furnace peak the strain-free reference temperature.
For an illustrative comparison, assume constant coefficients of eight and seven parts per million per kelvin for the dielectric and substrate over a four hundred kelvin interval. Their free-strain difference is four hundred microstrain. Replacing the hypothetical dielectric coefficient with eleven gives sixteen hundred microstrain, four times the first result. None of these values describes a selected paste or a production firing window. The result ranks an assumed mismatch for investigation; observed bow and interface integrity still decide whether the real stack is acceptable.
Δε_free = ∫[α_d,t(T) − α_s,t(T)] dT ≈ (α_d − α_s) ΔT
- Δε_free: difference in unconstrained linear thermal strain over the stated interval
- α_d,t and α_s,t: tangent linear expansion coefficients of dielectric and substrate
- α_d and α_s: constant coefficients used only in the illustrative approximation
- ΔT: temperature interval in kelvin
Common starting temperature, small linear strains and a single direction. The integral uses tangent coefficients; the numerical example assumes constant coefficients. This is not a residual-stress or fracture calculation.
Make the coupon contain the difficult features
A flat dielectric patch leaves important parts of the selection untested. Include a lower conductor step, the intended crossover, an opening where one is required and the edge of a broad covered area. These features expose different interactions between paste flow, film formation and the supporting surface. Keep their dimensions tied to the actual design. A large uniform coupon can be useful for material characterization while still missing the location most likely to govern the product.
Compare the candidates on the same representative topography and process route. Examine coverage at conductor shoulders, continuity around openings and the condition of upper metal after its own firing. Local observations should retain their location on the coupon rather than becoming one undifferentiated defect count. The table identifies separate compatibility decisions; it does not prescribe universal dimensions or acceptance limits. If a feature cannot be reproduced reliably in the trial, resolve that limitation before treating a favorable flat-area measurement as a complete stack qualification.
| Coupon feature | Question it exposes | Record needed |
|---|---|---|
| Lower conductor crossing | Does the dielectric cover the metal step after the specified route? | Crossing identity and local section or coverage observation |
| Upper attachment pad | Does the upper metal retain its intended attachment interface? | Metal composition, attachment process and failure location |
| Required opening | Does the opening remain usable without unwanted conductive contact? | Opening geometry and continuity or isolation result |
| Broad dielectric area | Does the assembly distort or separate during the route? | Substrate condition and bow or interface observation |
Assign electrical requirements after identifying the stack
Compatibility narrows the material candidates but does not establish their electrical suitability. State which conductor pair the dielectric separates, the applied waveform, the operating environment and the consequence of leakage or coupling. Insulation resistance, dielectric withstand and capacitance describe different responses. A material can satisfy an isolation test while still producing unwanted coupling in a sensitive circuit. Conversely, a low capacitance estimate does not demonstrate freedom from a local conductive path.
Keep the electrical design calculation attached to the selected composition and its actual fired geometry. Do not borrow a dielectric constant or voltage result from a different grade simply to complete a calculation sheet. The supplier's measurement conditions must be retained when interpreting a published property. Set application acceptance conditions with the circuit owner, including the state before and after relevant exposures. This guide selects the material stack; the resulting electrical dimensions and test levels still require their own justified design decisions using the named materials and intended operating conditions.
Validate the route and preserve failure locations
Validation should follow the coupon through the route instead of inspecting only the finished top surface. Record the condition after dielectric firing, after upper metallization and after any subsequent thermal or assembly step. This sequence can distinguish a film-formation problem from an interaction introduced later. Use the same coupon identity for electrical readings, dimensional observations and microscopy so a local anomaly can be connected to its processing history. Include uncoated substrate or simpler stack controls where they help interpret distortion or attachment changes.
Relevant failure signatures include cracking concentrated near a metal step, separation along an interface, new bow after an added firing, upper-pad detachment and leakage localized to a crossing. These observations suggest where to investigate; none proves a cause by itself. Risks include contamination, incomplete coverage and an unsupported conductor substitution. Preserve failed locations for examination rather than repeatedly retesting until an acceptable reading appears. Acceptance requires the agreed electrical response and physical integrity on the representative stack after the complete approved sequence.
Release a material combination with explicit change triggers
The output of selection should identify a combination, not an interchangeable category called dielectric paste. Name the substrate, lower conductor, dielectric, upper conductor and protective materials that touch or thermally affect the stack. Attach the process sequence and the representative feature set to that record. If multiple combinations are allowed, give each its own evidence trail rather than assuming a successful result for one permutation covers every possible mixture of approved materials.
Define which changes require review before purchase or production. A different substrate grade, a revised conductor composition, a combined firing step or a changed attachment process can all move the design beyond the evidence used for selection. The review need not repeat every test automatically; it should identify the interfaces and functions affected by the change and justify the validation scope. A clear stack record lets purchasing request the correct materials and lets engineering determine whether a substitution preserves the demonstrated conditions or creates a new material decision.
Inputs for a dielectric stack review
Provide the material interfaces and complete processing route so the selection can be tied to a defined stack.
- Substrate composition, grade, thickness and supplied surface condition.
- Named lower and upper conductor compositions, layer order and upper-pad attachment method.
- Print, dry and fire sequence including every subsequent thermal operation and proposed deviation.
- Representative crossings and openings, electrical requirements, exposure conditions and agreed acceptance owner.
The drawing-upload form loads as you reach this section.