Almost nobody chooses a die attach method. It arrives with the quote, because the line that priced the build already runs one, and the thermal design quietly forms around that decision. That works until the part gets hot, or the cavity has to seal, or a qualification temperature turns out to sit above what the joint can take. Eutectic and epoxy are not interchangeable, and the difference between them surfaces late — usually in thermal data, and occasionally in a reliability failure nobody budgeted for.
What each one actually is
Epoxy die attach is an adhesive. A filled polymer is dispensed or stamped, the die is placed into it, and the assembly is cured — well below the temperatures a metal joint needs. It comes conductive or non-conductive, it tolerates placement variation, and it is fast. For a large share of real builds it is entirely correct, and the cheaper answer besides.
Eutectic die attach is metallurgy. A eutectic alloy — gold-tin and gold-silicon are the usual systems — melts at one fixed temperature rather than over a mushy range, wets the die and the substrate metallisation, and freezes into a metal joint. There is no polymer in the stack. You get a short thermal path, no organic outgassing, and a bond that behaves like metal because it is metal. You pay for it in process temperature, surface preparation and cycle time.
The three things that actually decide it
1. The thermal path
This is the usual forcing function. A eutectic joint moves heat roughly an order of magnitude better than a filled epoxy — as an industry generality, tens of watts per metre-kelvin against low single digits. The precise numbers belong to the specific alloy and adhesive you are actually buying, not to a constant worth memorising.
What matters is the consequence: for a die dissipating real power, the attach layer is a term in the thermal budget. Teams routinely specify a heatsink and a substrate carefully and then leave the 50 microns directly under the die to whatever the assembly house happens to run.
2. Hermeticity and outgassing
Seal a polymer inside a cavity and it keeps releasing volatiles for the life of the part. In a hermetic build, or an optical one where a condensate film on a facet is a failure, that is disqualifying on its own — regardless of how the thermal numbers come out. This is why eutectic and hermetic packaging tend to appear in the same sentence: the requirement is chemical, not thermal.
3. What temperature your die and substrate can survive
The constraint that runs the other way. Eutectic attach means taking the whole assembly up to the alloy's melting point. If there is a temperature-sensitive die, a previously attached component, or a substrate with its own limit, the process window may simply not exist — and then the honest answer is epoxy with the thermal design adjusted to suit, not eutectic forced through on a reflow profile that damages the part.
Where silver sinter fits
Sinter is the third answer, and it is the right one more often than its reputation suggests. Silver sinter forms its bond below the melting point of silver, so the finished joint tolerates operating temperatures that would soften a solder joint, and it avoids the intermetallic ageing that eventually degrades one. For power devices and anything running hot continuously, that combination is hard to get any other way.
It is not a drop-in substitute. Sinter wants clean, properly metallised surfaces, and pressure-assisted sinter adds tooling and cycle time. Reach for it when operating temperature is the thing limiting the part, rather than when peak thermal conductivity looks attractive on a datasheet.
The honest rule
If the die does not dissipate much, the cavity is not hermetic, and process temperature is tight or cost matters, epoxy is the right answer and the cheaper one. Go eutectic when the thermal path is load-bearing, when the package has to seal, or when nothing organic is allowed inside it. Go silver sinter when the part has to keep working at a temperature that would soften solder.
What should never decide it is which method the assembly line already owns.
How the joint gets verified
A die attach decision is only as good as the evidence that the joint came out the way the process intended. Die-attach void content is verified by X-ray when void measurement is required by the part or the specification, alongside automated optical inspection and metrology; cross-section and failure analysis are available where a programme needs destructive confirmation.
Placement is captured per unit, which is the part teams tend to appreciate later — accuracy is auditable after the fact instead of asserted up front. Ours is ±2.5µm @3σ, and it is a measured number rather than a brochure one.
What this looks like in practice
We run epoxy, eutectic, silver sinter and fluxless attach in-house, which is the only reason we can give a straight answer here instead of steering you toward the single process we happen to own. Configurations run die-to-wafer, die-to-die, die-to-substrate and die-to-PCB, on wafers from 2 in through 12 in including thin and fragile die, on silicon, sapphire, aluminium nitride and ceramic carriers. Full process detail, design rules and build data are available under NDA.
Where to start
Three numbers settle this faster than a spec sheet will, and all three are usually known long before anyone quotes the assembly:
- How much power the die actually dissipates, and what junction temperature you have to hold.
- Whether the package has to be hermetic — or contains anything an outgassing polymer would spoil.
- The maximum temperature the die, substrate and any already-attached parts can take during assembly.
More on all four attach families, the accuracy data and the configurations we support on our die bonding services page. If the build has to seal, the hermetic packaging side of the decision is worth reading with it. If you are weighing this for a specific stack, we are happy to compare notes.
Die attach, answered.
What is eutectic die attach?
Eutectic die attach bonds a die to its substrate with a metal alloy that melts at a single fixed temperature rather than over a range. Gold-tin (AuSn) and gold-silicon are the common systems. Because the joint is metal, the result is a metallurgical bond with a short thermal path and no organic content — which is what makes it the usual choice for hermetic and high-power builds.
When should you use epoxy die attach instead of eutectic?
When the thermal path is not the limiting factor, and process temperature or cost is. Epoxy cures well below eutectic reflow temperatures, tolerates more placement variation, and comes in conductive and non-conductive forms. For sensors, general assembly and most prototypes it is the faster, lower-risk answer, and choosing eutectic anyway buys thermal headroom the device never uses.
What is the thermal difference between eutectic and epoxy die attach?
Roughly an order of magnitude, and often more. As an industry generality a gold-tin eutectic joint carries heat in the tens of watts per metre-kelvin, while a filled epoxy typically sits in the low single digits — the exact figures belong to the specific alloy and adhesive on your bill of materials, not to a universal constant. The practical read: if your die dissipates real power, the attach layer is part of your thermal design, not an afterthought.
Is silver sinter a better alternative to solder or eutectic die attach?
For high-temperature and power work it often is. Silver sinter forms its bond below the melting point of the metal, so the finished joint survives temperatures that would soften solder, and it does not have solder's intermetallic ageing behaviour. The trade is process control: sinter wants clean, well-metallised surfaces, and pressure-assisted sinter adds tooling. It is not a drop-in substitute — it is the right answer when operating temperature is what limits the part. We run pressure and pressureless silver sinter alongside eutectic and epoxy.
How do you verify die-attach quality and voids?
Die-attach void content is verified by X-ray when void measurement is required by the part or the spec, alongside automated optical inspection and metrology. Placement is captured per unit, so accuracy is auditable after the fact rather than asserted. Where a programme calls for destructive confirmation, cross-section and failure analysis are available.
Where can I get eutectic and epoxy die attach in the US?
We run both in-house in Halethorpe, Maryland, on US soil, along with silver sinter and fluxless attach — four attach families rather than one house default. Wafer sizes run 2 in through 12 in, including thin and fragile die, and configurations cover die-to-wafer, die-to-die, die-to-substrate and die-to-PCB.