A sensor package is asked to do two contradictory things. It has to seal a cavity against the outside world, and it has to let electrical connections cross that same wall. Every hermetic sensor design is some resolution of that contradiction, and the components that resolve it — glass-to-metal feedthroughs, matched-expansion lids — are worth understanding before the mechanical design is frozen around a guess.
The problem is thermal expansion, not sealing
Sealing two pieces of metal together is straightforward. The difficulty arrives the moment the assembly changes temperature, and it changes temperature during sealing, during qualification, and throughout service.
Different materials expand at different rates. Bond a metal lid to a ceramic body with a mismatched expansion coefficient and every thermal excursion works the joint — the seal that passed a leak test at room temperature becomes a seal that leaks after temperature cycling. The failure is not in the seal process; it is in the material pairing.
This is why Kovar appears everywhere in hermetic packaging and looks otherwise unremarkable. It is a nickel-cobalt-iron alloy whose expansion behaviour is close to that of certain hard glasses and ceramics over a useful temperature range. Its whole reason for existing is compatibility, not strength.
How a connection crosses a sealed wall
A glass-to-metal seal is the standard answer. A conductor passes through a glass insulator, which is itself sealed to a metal body — usually Kovar, for the expansion reason above. Done properly, the glass bonds chemically to the oxide on the metal, and the result is a genuine hermetic feedthrough that carries a signal.
Two consequences follow that catch teams out. First, the number and spacing of feedthroughs is a mechanical constraint as much as an electrical one — every penetration is a potential leak path, and they cannot be placed arbitrarily close. Second, glass is glass: the feedthrough region has real mechanical fragility, and a design that treats it as ordinary metal will discover this during handling.
Choosing how to close the lid
The seal method has to suit what is already inside the cavity, and that is usually the binding constraint rather than the seal's own performance.
Seam welding puts a localised weld around the lid perimeter, keeping bulk heat away from the contents — which matters when the cavity contains a die that has already been attached and wire bonded. Eutectic seam sealing, in Au-Si, Au-In or Sn-Ag systems, gives a metallurgical seal with no polymer anywhere in the cavity. Solder and epoxy seals are lower-temperature and less demanding on surface preparation; epoxy carries the outgassing consideration that follows any polymer sealed inside a package for the life of the part.
The sequence problem
Sealing is the last operation, which means every earlier decision has already constrained it. The die attach method has fixed how much heat the assembly can now take. The wire bond loops have fixed how much internal clearance the lid needs. The feedthrough layout has fixed where the seal perimeter can run.
Designing the seal first and working backwards is the version of this that goes smoothly. Doing it in build order — and discovering at the lid step that the only seal process compatible with the cavity contents is one the geometry does not allow — is the version that costs a revision.
Where to start
Four facts scope a hermetic sensor package:
- How many electrical connections have to cross the wall, and at what spacing.
- What is inside the cavity, and the maximum temperature it can now tolerate.
- The body and lid materials already fixed by the mechanical design — this is where expansion mismatch is usually decided.
- The environmental profile, since a seal's leak rate after thermal cycling is the number that matters.
We run Kovar lids, glass-to-metal seals, seam welding, eutectic seam sealing and solder or epoxy seals in Halethorpe, Maryland, with hermetic testing on request. More on hermetic packaging and on sensor packaging, and the leak-rate side of the decision is covered in what Method 1014 actually proves.
Answered.
What is a glass-to-metal seal?
A hermetic electrical feedthrough: a conductor passes through a glass insulator that is itself sealed to a metal body, usually Kovar. Done properly the glass bonds chemically to the oxide on the metal, giving a genuinely hermetic path for a signal to cross a sealed cavity wall.
Why is Kovar used in hermetic packages?
For thermal expansion compatibility, not strength. Kovar is a nickel-cobalt-iron alloy whose expansion behaviour closely matches certain hard glasses and ceramics over a useful temperature range. Pair mismatched materials and every thermal excursion works the joint until a seal that passed leak testing starts to leak.
What is the difference between seam welding and eutectic seam sealing?
Seam welding forms a localised weld around the lid perimeter, keeping bulk heat away from cavity contents — useful when a die is already attached and wire bonded inside. Eutectic seam sealing uses an alloy system such as Au-Si, Au-In or Sn-Ag to form a metallurgical seal with no polymer anywhere in the cavity.
How many feedthroughs can a hermetic sensor package have?
It is a mechanical question as much as an electrical one. Every penetration is a potential leak path and they cannot be spaced arbitrarily closely, so feedthrough count and spacing constrain where the seal perimeter can run. It is worth settling early, because sealing is the last operation and inherits every earlier decision.
Do you offer glass-to-metal sealed sensor packaging in the US?
Yes — Kovar lids, glass-to-metal seals, seam welding, eutectic seam sealing and solder or epoxy seals, in Halethorpe, Maryland, on US soil, with hermetic testing available on request.