Hermetic Seam Sealing for Hybrid Modules

1. Introduction

Hybrid microelectronic modules house bare silicon dies, passive components, and interconnect structures inside a protective package. For most commercial applications — smartphones, consumer electronics, automotive infotainment — plastic overmolded packaging is entirely adequate. But in radar systems, satellite communications, implantable medical devices, and military electronics, a different standard applies: the package must seal its interior from the external environment indefinitely.

Hermetic sealing is the engineering solution to this problem. A hermetic seal is defined as one that prevents gaseous and liquid ingress into the package interior. The term comes from the Greek hermētikos ("closed"), and in microelectronics it carries a specific quantitative meaning: the leak rate must be below 1 × 10⁻⁸ atm·cm³/s of air equivalent.

The alternative — non-hermetic, plastic packaging — relies on moisture barrier coatings and encapsulants. These work by slowing moisture diffusion, but they cannot stop it entirely. Over time, moisture accumulates inside plastic packages, creating failure mechanisms that hermetic packaging eliminates by design.

Moisture Ingress Failure Modes

When moisture penetrates a hybrid module, several failure mechanisms activate:

• Galvanic corrosion: Dissimilar metals within the module (e.g., aluminum wire bonds adjacent to copper traces) form galvanic couples when electrolyte is present. Even trace amounts of moisture enable ionic conduction, accelerating metal dissolution at the anode.
• ionic contamination: Residual flux activators, handling salts, or process chemicals dissolve in accumulated moisture, creating conductive paths between adjacent conductors. A 1 MΩ leakage path on a high-impedance input can shift bias points by volts.
• Metal migration: Applied electric fields drive metal ion migration (silver, copper, tin) across insulating surfaces, creating conductive dendrites that eventually short adjacent traces.
• Delamination and popcorn cracking: During solder reflow, absorbed moisture vaporizes and expands. In plastic packages, this pressure can delaminate the die-attach interface or crack the molding compound — the "popcorn" failure mode well documented in plastic IC packaging.

Hermetic packaging eliminates all of these failure modes by keeping moisture out entirely. The interior atmosphere can also be controlled — typically dry nitrogen (N₂), helium (He), or forming gas (N₂/H₂) — which further suppresses corrosion reactions even if a microscopic leak were to develop.

2. Why Hermeticity Matters

The performance requirement for hermeticity is not arbitrary. It derives from the physics of gas permeation and the reliability targets of the end application.

Corrosion Failure Mechanisms

Two primary corrosion mechanisms threaten hybrid modules in the field:

• Galvanic corrosion without moisture: In a dry, inert atmosphere (N₂-filled package), galvanic corrosion rates are negligible. The absence of an electrolyte essentially halts the electrochemical reaction, regardless of the nobility difference between metal pairs.
• Galvanic corrosion with moisture: Once water enters the package — even at parts-per-million levels — it adsorbs onto hydrophilic surfaces and becomes the electrolyte for electrochemical reactions. The corrosion rate becomes a function of humidity, temperature, and the ionic content of the water film. In marine environments, saltwater aerosol can enter through leaks and create highly conductive electrolytes.

Application-Specific Requirements

Different industries mandate hermetic packaging for distinct reasons:

• Military/Aerospace (MIL-PRF-38534, MIL-STD-883): The U.S. Department of Defense requires hermetic packaging for Class H and Class K hybrid circuits. The justification is twofold: field repair is often impossible (satellites, missiles), and service life must exceed 20 years in uncontrolled environments.
• Medical implants: Active implantable devices such as pacemakers, cochlear implants, and neurostimulators require hermetic seals because body fluids are highly corrosive and the device must function for the patient's lifetime without replacement.
• Space electronics: The vacuum of space causes outgassing from non-hermetic materials. Additionally, thermal cycling between sun-facing and shadow sides of a spacecraft subjects packages to extreme temperature swings that can drive moisture condensation on internal surfaces.
• Oil and gas sensing: Downhole tools operate at temperatures above 200°C and in hydrogen sulfide (H₂S) environments. Only hermetic metal packages survive these conditions.

Internal Gas Composition

The atmosphere sealed inside a hybrid module is engineered for specific purposes:

• Nitrogen (N₂): The most common fill gas. Inert, dry, and suppresses combustion/corrosion reactions. Typical purity: 99.999% (5N) or better.
• Helium (He): Used as a tracer gas for leak detection (see Section 6). Helium is chosen because its small atomic radius (0.28 Å) makes it the most likely gas to leak through a real defect — if He can't get in, nothing can. Some military modules are sealed directly with He fill so that leak checking can be performed throughout the module's service life.
• Forming gas (N₂/H₂, typically 95/5): Used in some power semiconductor modules where a reducing atmosphere prevents surface oxidation of bond pads and wire bonds during high-temperature operation.

3. Seam Welding Parameters

Resistance seam welding is the dominant method for closing hermetic hybrid module lids. It is a solid-state welding process — the lid and header are not melted; instead, the faying surfaces are heated by electrical resistance and forged together under force, creating a metallurgical bond.

Resistance Seam Weld Process

In resistance seam welding, two rotating copper electrodes (the weld wheels) apply force to the lid-to-header interface while a pulsed or continuous AC/DC current passes through the joint. The heat is generated by the electrical resistance at the interface — primarily at the point where the lid flange and header rim meet.

Unlike spot welding, which creates discrete weld nuggets, seam welding produces a continuous gas-tight weld along the entire perimeter of the lid. The weld appears as a overlapping series of weld spots, each fused into the next.

Weld Schedule Parameters

Four primary parameters govern seam weld quality. These must be balanced for each lid/header material combination and geometry.

Lid Thickness vs. Weld Current

Lid thickness is the primary driver for weld current selection. A thicker lid has greater thermal mass, requiring more current to achieve the forging temperature at the interface. As a rule of thumb:

• Kovar lid, 0.25 mm (10 mil): ~1,000–1,500 A
• Kovar lid, 0.38 mm (15 mil): ~1,500–2,200 A
• Kovar lid, 0.50 mm (20 mil): ~2,200–3,000 A

Gold plating adds complexity: plating thickness must be accounted for because the Au layer (which melts at 1,064°C) partially melts during welding. Excessive Au thickness can cause the weld to be " Au-rich" and prone to cracking during solidification.

Weld Head Designs: Transverse vs. Longitudinal

Transverse weld heads (wheel axis perpendicular to weld direction) are the most common. The weld wheel spins continuously as it traverses the lid perimeter, producing a smooth, consistent weld. Suitable for rectangular and square lids.

Longitudinal weld heads (wheel axis parallel to weld direction) use flat-faced electrodes that advance along the seam. They offer superior force control and are preferred for very long straight seams or when the lid geometry doesn't accommodate wheel rotation at corners.

Heat-Affected Zone Concerns

The heat-affected zone (HAZ) adjacent to the weld seam is a region where the base metal properties are altered by thermal exposure without melting. In gold-plated Kovar, the HAZ can extend 0.5–2 mm from the weld line. Key concerns:

• Gold diffusion: At elevated temperatures, Au diffuses into the Kovar substrate, potentially forming Au-Kovar intermetallics that are brittle.
• Residual stress: Uneven thermal expansion during welding creates tensile stresses in the lid that can cause delayed cracking (stress relief cracking) hours or days after welding.
• Substrate thermal shock: If the header is attached to an alumina or AlN substrate, excessive heat input can crack the ceramic at the seal ring attachment point.

4. Gold-Plated Kovar Lids

Why Kovar?

Kovar (Fe-29Ni-17Co, ASTM F15) is a controlled-expansion alloy engineered to match the thermal expansion coefficient (CTE) of alumina (Al₂O₃, CTE ≈ 6.5–7.0 ppm/°C) and aluminum nitride (AlN, CTE ≈ 4.5 ppm/°C) substrates over the relevant temperature range (−55°C to +125°C or beyond).

Without CTE matching, differential thermal expansion between the lid and the substrate creates shear stresses at the seal interface. Over temperature cycling, these stresses accumulate and eventually cause cracking — either in the lid, the solder seal, or the glass-to-metal feedthroughs.

Why Gold Plate?

Gold plating on Kovar lids serves multiple functions:

• Corrosion resistance: Bulk Kovar corrodes in humid environments. The Au layer acts as a barrier coating, protecting the underlying alloy from oxidation and moisture attack.
• Solderability: For solder-sealed lids (alternative to seam welding), the Au surface solders readily to gold-tin (AuSn) or tin-lead (SnPb) preforms.
• Seam weld compatibility: The Au layer at the interface melts during resistance welding and forms a transient Au-rich liquid phase that promotes wetting and fusion of the Kovar surfaces. The Au layer must be thin enough to avoid creating a continuous brittle Au-rich weld seam.

Plating Thickness and Kirkendall Voiding

Standard plating thickness for hybrid lid applications: 50–150 μin (1.3–3.8 μm) of electrodeposited soft gold over 50–150 μin of electrodeposited nickel. The nickel layer serves as a diffusion barrier between the Kovar and the Au — without it, Au would diffuse rapidly into Kovar at elevated temperatures.

Kirkendall voiding is a classic failure mode in Au/Ni/Kovar systems. When heated above ~300°C (such as during AuSn die attach or rework), Au and Ni interdiffuse at different rates. The difference in diffusion rates causes vacancy supersaturation at the Au/Ni interface, which collapses into microscopic voids. These voids reduce the effective contact area and can lead to electrical opens or mechanical delamination. Mitigation strategies include specifying minimum Ni thickness and avoiding prolonged high-temperature exposures.

5. Header Designs

The header (also called a "base," "foot," or "seal ring") is the structural foundation of the hermetic package. It provides the sealing surface for the lid, mechanical attachment points for the substrate, and electrical feedthroughs for signals and power.

Header Types

• Pin Grid Array (PGA) headers: Feature a matrix of through-hole feedthroughs. Used where high lead counts are needed (microprocessors, memory hybrids). Feedthrough spacing typically 2.54 mm (0.1") or 1.27 mm (0.05").
• Flat pack headers: Feature leads extending from two or four sides. Common in RF modules where coaxial connectors or microstrip launches are integrated into the header. Lead counts range from 8 to 100+.
• Cavity-up vs. cavity-down: In cavity-up designs, the substrate is attached to the floor of a recessed cavity in the header, and the lid closes over the cavity. In cavity-down designs, the substrate is mounted on the exterior face and the lid seals against the header's outer rim. Cavity-up provides better mechanical protection for dies; cavity-down offers lower thermal resistance to the heatsink (if header is direct-mounted).

Glass-to-Metal Seals (GTMS)

The electrical feedthrough pins must penetrate the header wall while maintaining hermeticity. Glass-to-metal sealing is the dominant technology. A glass preform (typically a borosilicate glass such as Corning 7070 or equivalent) is melted around the pin and the header hole at high temperature. On cooling, the glass contracts and compresses the pin, creating a permanent hermetic seal.

Header Plating: Soft Gold Over Nickel

Header sealing surfaces (the flange or seal ring) are plated with soft gold over electrodeposited nickel. The Ni provides corrosion resistance and a diffusion barrier; the Au provides the solderable/weldable surface. Typical specification: 50–100 μin Au over 100–200 μin Ni, per MIL-G-45204 and MIL-N-45207.

6. Leak Testing: MIL-STD-883 Method 1014

MIL-STD-883, Method 1014 (Hermeticity) is the canonical test method for verifying hermetic seal quality in microelectronics packages. It defines two categories of leak testing: gross leak and fine leak, each with multiple test conditions.

Gross Leak vs. Fine Leak

• Gross leak detects large, obvious leaks that would allow liquid penetrant to enter the package. Typically corresponds to leak rates >10⁻⁵ atm·cm³/s.
• Fine leak detects smaller leaks that allow tracer gas (helium or fluorocarbon) to pass, but would not admit liquid. Range: 10⁻⁵ to 10⁻¹¹ atm·cm³/s.

Helium Mass Spectrometer Leak Detection

The fine leak test (Test Condition B or C) uses helium as the tracer gas. The sequence:

1. Package is placed in a pressurized chamber filled with He at 45–75 psig (absolute) for 2–4 hours (bombing).
2. Pressure is released, external He is blown off, and the package is transferred to the mass spectrometer test chamber within 1 hour (shorter for very small packages).
3. The spectrometer measures He evolution from the package. A leak rate is computed from the He signal.

Bubble Test (Fluorocarbon Method)

The gross leak test (Test Condition A) submerges the package in a bath of fluorocarbon liquid (Galden or equivalent) at 125°C. The liquid is preheated to above its boiling point. Any gross leak allows internal gas to expand and bubble out of the package. Bubble detection is visual and operator-dependent, making it less sensitive and repeatable than He testing.

MIL-STD-883 Method 1014 Acceptance Criteria

Note: R₁ is the measured leak rate corrected for package internal volume per Method 1014 Equation 1. Smaller packages have tighter R₁ limits due to their smaller internal gas volume.

7. Leak Rate vs. Bubble Test

Choosing between helium leak detection and bubble testing depends on the reliability requirement and the package geometry.

Bubble test limitations: The bubble test detection threshold is approximately 10⁻⁵ atm·cm³/s — meaning a leak must be large enough to allow visible bubble formation. Smaller leaks (10⁻⁶ to 10⁻¹¹ atm·cm³/s) will not produce detectable bubbles, yet are still catastrophically leaky from a reliability standpoint. The bubble test is also destructive (packages are immersed in hot liquid) and operator-dependent.

Helium leak detection advantages: He leak testing is non-destructive (the package survives the test), quantitative, and orders of magnitude more sensitive. Standard He bomb-and-detect can reliably find leaks down to 10⁻⁹ atm·cm³/s. With accumulation methods (trapping He emitted from the package into a small volume over time), detection limits can reach 10⁻¹¹ atm·cm³/s — the level required for military space-grade hardware.

8. Common Failure Modes

9. Conclusion

Hermetic seam sealing is a mature but demanding technology. Getting it right requires controlling metallurgy (Kovar/Au plating quality), process parameters (weld current, force, speed), feedthrough quality (glass-to-metal seals), and test coverage (MIL-STD-883 Method 1014 fine and gross leak testing).

The penalty for getting it wrong is severe: field failures in defense systems, medical implants, or space hardware cannot be recalled. The investment in proper hermetic sealing — qualified personnel, controlled processes, statistical process control, and thorough screening — is justified by the elimination of moisture-driven failure modes over a 20+ year service life.

For design engineers specifying hybrid modules for harsh environments, the key decisions are: hermetic vs. non-hermetic packaging choice, header and lid material selection, seam weld parameter development with the assembly partner, and a leak test plan that matches the application's reliability requirement.