A hermetic seal is a joint that blocks not just water and dust, but gases and vapor at the molecular level. The gold standard for hermeticity is a helium leak rate at or below 1×10⁻⁷ atm·cc/sec, a threshold so tight that virtually no air, moisture, or contaminant can pass through over the life of the sealed component. Making one requires choosing the right method and materials for your application, then verifying the result with proper testing.
Hermetic vs. Airtight: Why the Distinction Matters
Many people confuse hermetic sealing with waterproofing or high IP ratings. An IP68-rated enclosure is fully dust-tight and survives continuous water immersion, but it only addresses liquid water and solid particles. It does not prevent gases, chemical vapors, or humidity from slowly migrating through over time. A true hermetic seal stops everything, including individual gas molecules, which is why it’s required in electronics, medical implants, aerospace hardware, and anywhere long-term reliability in harsh environments is non-negotiable.
Glass-to-Metal Seals
The most established method for creating a hermetic seal is fusing glass directly to metal. This technique is used in electrical feedthroughs, sensor housings, and military-grade connectors. The process works because molten glass chemically bonds to an oxidized metal surface, forming a joint with no organic material that could degrade or outgas.
The critical challenge is thermal expansion. Glass and metal expand at different rates as temperature changes, and if those rates don’t match closely enough, the seal will crack when it cools or when it experiences temperature swings in service. Engineers solve this in two ways:
- Matched seals pair a metal and glass with nearly identical expansion rates. The classic combination is Kovar (an iron-nickel-cobalt alloy) with borosilicate glass. Kovar’s thermal expansion coefficient sits around 5.1 to 5.5 × 10⁻⁶ m/m·K between 30°C and 300°C, which closely tracks borosilicate glass through the critical cooling range.
- Compression seals use a metal housing that contracts slightly more than the glass as both cool, squeezing the glass plug and keeping it under constant compression. This approach tolerates a wider range of material pairings but requires careful geometry.
During manufacturing, the glass is heated to between 500°C and 600°C to flow and bond to the metal, then cooled at a controlled rate. Research at Sandia National Laboratories shows that cooling speed dramatically affects residual stress. Slow cooling at 0.5°C per minute produces a very different stress profile than a rapid quench at 100°C per minute. Controlled, slow cooling (annealing) minimizes internal stress and reduces the risk of cracking later.
Why Glass-to-Metal Seals Fail
The most common failure mode is meniscus cracking. When molten glass wets the metal pin or housing wall, it climbs slightly up the surface, forming a thin curved edge called the meniscus. This zone is fragile. If too much glass wets the metal and then the housing applies high compression during cooling, the meniscus can fracture. Research at Clemson University traced seal failures during shock loading directly to breakdown at this meniscus zone along the glass-metal interface.
Other failure causes include mismatched thermal expansion (CTE mismatch) that accumulates compressive stress at the sealing interface, and improper oxidation of the metal surface. Some oxidation is actually necessary, since the chemical bond between glass and metal depends on metal ions bridging both materials. But too much oxidation weakens the bond and signals a production problem. Too little, and the glass won’t wet the metal at all.
Laser Welding for Metal Enclosures
When both halves of an enclosure are metal, laser welding is the preferred method for achieving hermeticity. This is the standard approach for implantable medical devices like pacemakers and cochlear implants, where a titanium case must remain sealed inside the human body for years.
Pulsed laser welding delivers energy in precisely controlled bursts. For titanium housings, typical settings run between 1 and 4 joules per pulse with pulse durations of 1 to 8 milliseconds. A representative setup might use 3 joules per pulse at a 3-millisecond width and 10 pulses per second, delivering 30 watts of average power. The process is done under a shield of inert gas (argon or helium) to prevent the molten weld pool from reacting with oxygen or nitrogen in the air. Without shielding gas, the weld becomes brittle and porous, destroying hermeticity.
Laser welding produces a narrow, clean weld with minimal heat input to surrounding components, which is why it’s favored for sealing enclosures that contain sensitive electronics.
Ceramic-to-Metal Brazing
Some applications need the electrical insulation or high-temperature resistance of ceramics combined with the structural strength of metal. Active brazing makes this possible in a single step by using a brazing alloy that contains a reactive element, typically titanium, mixed into a silver-copper base. The titanium reacts with the ceramic surface and creates a true metallurgical bond, something conventional brazing alloys cannot do on ceramics.
One widely used active brazing alloy is Incusil-ABA (a silver-copper-indium-titanium alloy), applied at 730°C for 10 to 20 minutes inside a vacuum furnace. A study funded by the Department of Defense used this alloy to bond aluminum oxide ceramic tiles to titanium alloy and Kovar backing structures, with thin copper sheets placed between layers to absorb thermal stress. The vacuum environment is essential because oxygen would react with the titanium in the alloy before it could bond to the ceramic.
Scaling up from small lab samples to large panels requires extra time in the furnace to account for the added thermal mass, since the entire assembly must reach brazing temperature uniformly.
Epoxy Seals for Lower-Demand Applications
Not every application needs a glass or metal seal. Hermetic-grade epoxies can provide a near-hermetic barrier for connectors and enclosures where the environment is less extreme. These materials are tested to NASA’s ASTM E-595 outgassing standard, which measures how much material evaporates in vacuum at 125°C. A qualifying epoxy might show less than 0.22% weight loss and less than 0.002% volatile condensable material, with no visible deposits on a test surface.
The practical limit of epoxy is moisture. Even high-performance epoxies allow some water vapor to pass through. A typical hermetic-grade epoxy has a moisture vapor transmission rate around 0.7 grams per square meter per day through a thin film. For many commercial and industrial applications this is perfectly adequate, but for components that must survive decades in space or inside the body, glass-to-metal or welded seals remain the standard.
Induction Cap Sealing
If you’re sealing bottles or jars for food, pharmaceuticals, or supplements, induction sealing is the most common method for creating a hermetic closure at production scale. A thin foil liner sits inside the cap, and an electromagnetic coil generates an alternating field (typically in the 47 to 220 kHz range) that heats the foil. The foil’s polymer coating melts and bonds to the container rim, creating a tamper-evident, airtight seal.
Systems as small as 1.5 kW can power the coil head, and the resulting bonds can withstand air pressures up to 590 kPa (roughly 85 psi). For dusty production environments like coffee or nutraceutical packaging, the induction system itself is often housed in a sealed, zero-ventilation enclosure to keep particulates out of the electronics.
How Hermetic Seals Are Tested
A seal is only hermetic if you can prove it. The military standard MIL-STD-883, Method 1014 defines the two main categories of leak testing:
Fine leak testing uses a tracer gas, almost always helium, because its small atomic size lets it find paths that larger molecules cannot. The sealed component is placed in a pressurized helium chamber, then moved to a mass spectrometer that detects any helium escaping. This method reliably measures leak rates from 10⁻⁴ down to 10⁻¹⁰ atm·cc/sec. A radioactive isotope of krypton (Kr-85) is sometimes used as an alternative tracer, particularly when components are too small for helium bombing.
Gross leak testing catches larger defects that helium testing actually misses, since a wide-open leak lets helium flow through and escape before the detector measurement begins. Gross leak methods include submerging the sealed part in a fluorocarbon liquid bath and watching for bubbles, checking for weight gain after pressurized exposure to a heavy fluid, or using penetrant dye that seeps through visible cracks. Some newer cumulative helium leak detection (CHLD) systems can screen for both fine and gross leaks in a single test cycle.
There is no single universally agreed-upon maximum allowable leak rate. As the FDA notes, the specific threshold depends on the application, the internal volume of the package, and how long the seal must last. The 1×10⁻⁷ atm·cc/sec helium leak rate is a widely cited benchmark, but critical applications like satellite components or implantable devices often demand rates several orders of magnitude tighter.