Kovar Alloy and Glass Matching Evaluation: A Technical Guide to Glass-to-Metal Seal Compatibility As

Kovar alloy and glass matching evaluation is the systematic verification of thermal expansion compatibility between iron-nickel-cobalt sealing alloys and borosilicate glasses, the decisive factor in producing leak-tight, stress-controlled glass-to-metal seals. This article explains the metallurgical and glass science fundamentals behind the match, the sample types, test items, and evaluation methods used by accredited third-party laboratories, along with acceptance criteria and typical application scenarios for high-reliability hermetic products.

Technical Overview

Kovar is a controlled-expansion iron-nickel-cobalt sealing alloy, standardized in ASTM F15 (UNS K94610, Chinese grade 4J29), with a nominal composition of 28.5–29.5% nickel, 16.8–17.8% cobalt, and balance iron. Its mean coefficient of thermal expansion (CTE) of approximately 5.0–5.5 ×10⁻⁶/°C between room temperature and about 400°C closely matches that of borosilicate "iron-sealing" glasses, which is the basis of the classic matched glass-to-metal seal (GTMS).

The matching logic is governed by the sealing thermal cycle. During seal formation the glass wets the metal above its dilatometric softening region, then stiffens at the setting point (typically 430–480°C for matched borosilicates). As the assembly cools to room temperature, any cumulative expansion mismatch between the setting point and ambient temperature is converted into residual stress in the glass and at the interface. Excessive mismatch produces tensile stress in the glass, which can cause radial or circumferential cracking, interfacial delamination, or delayed seal failure under thermal cycling.

A distinctive feature of Kovar is its magnetic Curie transition near 430°C, above which the CTE rises steeply toward values of 8–10 ×10⁻⁶/°C. Successful matched seals are designed so that glass setting occurs near or below this inflection, keeping the effective integrated expansion of metal and glass nearly equal. As a practical acceptance guideline, the mean CTE difference between alloy and glass from room temperature to the glass setting point is normally held within approximately ±10% (roughly ±0.5 ×10⁻⁶/°C), ideally with the metal marginally higher so that the finished glass sits in slight compression.

A complete matching evaluation therefore covers four pillars: expansion behavior of both materials, surface chemistry and oxidation control of the alloy, seal fabrication quality on representative coupons, and verification of hermeticity, residual stress, and endurance on the final assembly.

Testing Samples

Raw Alloy Specimens

Dilatometer bars (typically 25–50 mm long, 3–6 mm diameter or square) machined from the same heat, lot, and mill product form (rod, wire, strip, ring, or machined preform) used in production. The heat-treatment condition (annealed, stress-relieved, pre-oxidized) must be documented, since prior thermal history shifts the measured CTE curve.

Raw Glass Specimens

Dilatometer bars cut or pressed and sintered from the candidate sealing glass, commonly matched borosilicate grades such as glass codes 7052, 8250/8252, and DM-308. Glass from the same melting lot as production performs should be used whenever possible to capture batch-to-batch expansion variation.

Seal Test Coupons

Standardized seal geometries fabricated for qualification: bead seals (glass ring fused around a metal rod), ribbon or tubular seals, disc or button seals, and pin-in-plate seals. Coupons are produced with the production firing profile and pre-oxidation treatment so that interface quality is representative.

Product-Level Assemblies

Finished hermetic components such as feedthroughs, connector inserts, transistor-outline headers, and package lids, evaluated for lot acceptance and periodic qualification including hermeticity, thermal endurance, and destructive cross-sectioning.

Testing Items

Depending on the stage of evaluation, the following items are typically assessed:

• Chemical composition of the alloy: Ni, Co, Fe balance, C, Mn, Si, Cr content against ASTM F15 or 4J29 grade limits
• Mean CTE of the alloy over intervals such as 20–100°C, 20–200°C, 20–300°C, 20–400°C, and 20–500°C, including heating/cooling hysteresis across the Curie transition
• Mean CTE of the glass over 20–300°C and 20–400°C, glass transition temperature (Tg), dilatometric softening point, and effective setting point
• Curie temperature of the alloy to locate the expansion inflection relative to the glass setting range
• Pre-oxidation behavior: oxide weight gain per unit area, oxide thickness, uniformity, and adhesion after the wet-hydrogen firing treatment
• Metallographic examination: grain size, non-metallic inclusions, oxide layer thickness at the seal interface, interfacial bubble and void content, and wetting continuity
• Residual stress magnitude and sign (compressive or tensile) in the sealed glass, measured photoelastically
• Hermeticity: helium fine-leak rate and gross-leak screening
• Thermal endurance: thermal shock and thermal cycling resistance with post-test leak and visual verification
• Mechanical integrity: push-out strength of pins, seal shear or tensile strength, and torque resistance where applicable

Testing Methods

Dilatometric Expansion Measurement

Mean CTE of the alloy is measured with a horizontal push-rod dilatometer per ASTM E228 or GB/T 4339, and the glass per ISO 7991. Specimens are heated at 3–5°C/min in an inert or dry atmosphere against a calibrated alumina or fused-silica reference, typically to 600–700°C for the alloy and 20–50°C beyond the glass softening range. The test reports mean CTE intervals, the tangent-derived Tg and softening point of the glass, the alloy Curie inflection, and the integrated expansion difference between room temperature and the glass setting point, which is the primary matching criterion.

Composition and Microstructure Analysis

Alloy composition is confirmed by ICP-OES optical emission spectrometry with combustion analysis for carbon. Metallography follows ASTM E3 preparation, ASTM E407 etching, and ASTM E112 grain-size rating. Seal cross-sections are examined optically and by SEM with EDS mapping to verify a continuous 1–3 µm interfacial oxide partially dissolved into the glass, the chemical bonding layer, and to quantify bubble area percentage and interfacial voids.

Pre-Oxidation Control

Coupon oxidation is performed in wet hydrogen at approximately 800–1000°C, then verified gravimetrically. A controlled weight gain on the order of 0.2–0.7 mg/cm², yielding a uniform gray-green adherent oxide, is targeted. Insufficient oxidation causes dewetting and lack of chemical bond; excessive oxidation produces gas bubbles, oxide flaking, and leak paths.

Photoelastic Stress Measurement

Residual stress in the sealed glass is measured with a polariscope or polarimeter using Senarmont compensation, applying the stress-optical coefficient of the glass. The evaluation confirms that the seal periphery remains in net compression and that tensile stress peaks stay below the design limit for the glass thickness.

Hermeticity Testing

Fine leak testing uses helium mass spectrometry in the hard-vacuum mode, with helium bombing pressures and dwell times per MIL-STD-883 Method 1014 or equivalent internal criteria. Common rejection limits are 5 ×10⁻⁸ atm·cm³/s He for microelectronic packages and tighter limits of 1 ×10⁻⁹ Pa·m³/s class for vacuum and optical devices. Gross leaks are screened by fluorocarbon bubble or weight-gain methods.

Thermal Endurance Testing

Seal assemblies are subjected to thermal shock and thermal cycling per IEC 60068-2-14 and MIL-STD-883 Method 1010, commonly −65°C to +150°C for 10 to 100 cycles with transfer times of a few minutes. After cycling, specimens are re-inspected visually for cracks, re-tested for leak rate, and destructively cross-sectioned to confirm interface integrity.

Mechanical Strength Testing

Pin and terminal retention is verified by push-out or pull testing at controlled loading rates, with acceptance based on the minimum retention force or failure occurring in the glass body rather than at the interface, the latter indicating a properly bonded matched seal.

Applications

Verified Kovar-glass matching is essential wherever hermeticity, long-term vacuum integrity, or thermal reliability is required:

• Hermetic microelectronic and MEMS packages, crystal oscillator headers, and transistor-outline packages
• Aerospace, defense, and industrial hermetic connectors and feedthroughs subjected to wide temperature swings
• Pressure and oxygen sensor transducers, inertial sensor housings, and down-hole electronics
• Hermetic terminals for lithium primary and secondary battery cells requiring long-life leak tightness
• Vacuum electron devices, X-ray tube feedthroughs, and vacuum envelope insulator seals
• Medical implantable device feedthroughs and explosion-proof equipment terminals

FAQ

What CTE mismatch between Kovar and glass is acceptable?

As a rule of thumb, the mean CTE difference from room temperature to the glass setting point should be within about ±10%, or roughly ±0.5 ×10⁻⁶/°C. Slightly higher metal expansion is preferred so the glass ends in mild compression, which improves crack resistance.

Which glasses are matched to Kovar?

Borosilicate iron-sealing glasses with mean CTE of roughly 4.8–5.5 ×10⁻⁶/°C between 20°C and 300–400°C, such as glass codes 7052, 8250, and DM-308. Soda-lime and high-expansion glasses are not suitable for matched seals with Kovar.

Why does the Kovar CTE curve change near 430°C and why does it matter?

Above its Curie temperature, the alloy loses its low-expansion magnetic contribution and expands more rapidly. Sealing schedules are designed so the glass sets near or below this inflection, and the evaluation must use the measured full CTE curve rather than datasheet room-temperature values alone.

Why is pre-oxidation of Kovar necessary before sealing?

A thin, adherent iron-nickel-cobalt oxide partially dissolves into the glass, forming the chemical bond required for a strong hermetic seal. The oxidation treatment must be optimized: too little oxide causes dewetting, while too much creates bubbles and weak, leaky interfaces.

What leak rate defines a hermetic seal?

For most microelectronic-grade packages, a helium fine-leak rate of 5 ×10⁻⁸ atm·cm³/s or better per MIL-STD-883 Method 1014 is standard; high-reliability vacuum and optical applications often specify rates in the 10⁻⁹ Pa·m³/s range.

What is the difference between a matched seal and a compression seal?

A matched seal uses materials with nearly equal expansion so stresses remain low in both components, suitable for strong, complex geometries. A compression seal deliberately uses a housing with higher expansion than the glass so the metal squeezes the glass on cooling, common with steel or stainless bodies and higher-expansion glasses.

How is residual stress in the sealed glass evaluated?

Photoelastic polarimetry on the finished seal quantifies the magnitude and sign of stress, and destructive cross-sectioning with metallography confirms the interface condition. Compressive peripheries with tensile peaks below the allowable design stress indicate a well-matched seal.

Which standards govern the evaluation?

Key references include ASTM F15 for the alloy, ASTM E228 and GB/T 4339 for metallic expansion, ISO 7991 for glass expansion and transition temperature, ASTM E3/E407/E112 for metallography, MIL-STD-883 Methods 1010 and 1014 for cycling and hermeticity, and IEC 60068-2-14 for thermal shock. Accredited third-party laboratories can conduct the full evaluation chain against these standards.