Helium leak test: a practical guide to hermetic welds
The helium leak test (also known as a helium leak inspection) is the established validation method for hermetic welds on laboratory capsules, sensor housings, medical implants, and high-reliability electronic enclosures. Typical acceptance limits range from 10⁻⁵ mbar·l/s for general industrial specifications to less than 10⁻⁹ mbar·l/s for high vacuum and aerospace applications; the relevant standards are MIL-STD-883 (Test Method 1014) and DIN EN ISO 20485. Three factors determine whether a component passes the test: weld preparation, welding parameters, and test method. This guide explains what is tested, how it is tested, and how the welding process prepares the component for the test.
Why is the helium leak test so important?
A hermetic weld is only as good as the test that verifies it. This applies to laboratory capsules containing radioactive seeds, biological samples, or air-sensitive substances; to medical implants in body tissue; and to sensor housings in the deep sea or in the vacuum of space.
The helium leak test is the industry standard because:
- Helium has the smallest kinetic diameter of all practically usable tracer gases. It penetrates weld defects that other test gases do not detect.
- the background concentration of helium in the atmosphere is low (about 5 ppm), resulting in a high signal-to-noise ratio.
- Helium-calibrated mass spectrometers are commercially available and can achieve sensitivities below 10⁻¹² mbar·l/s in the laboratory.
To clarify the terminology: “Leak test” is the standard generic term for the pass/fail verification, while “leak detection” refers to the process of locating the leak. “Helium leak test” and “helium leak inspection,” on the other hand, are used interchangeably in the market.
In most hermetic applications, the question is not, “Can the weld pass any leak test?” but rather, “Can it meet the leak rate required by the application?” Therefore, the target leak rate is determined first, and then the weld and test method are designed to meet that target.
Acceptable leak rates by application
The leak rate is measured in mbar·l/s (millibars × liters per second), which is the volume of gas that escapes per unit of time under standard conditions. Applications tolerate a wide range of values:
| Application | Typical acceptance leak rate | Test Method |
|---|---|---|
| General-purpose industrial hermetic (connectors, sensor housings) | 10⁻⁴ to 10⁻⁵ mbar·l/s | Sniffer probe, pressure drop |
| Brachytherapy seeds (radioisotope capsules) | 10⁻⁶ to 10⁻⁸ mbar·l/s, application-specific | Vacuum chamber + bombing |
| High-reliability electronics (aviation, medical implants, military) | Volume-dependent: 5×10⁻⁸ to 1×10⁻⁶ mbar·l/s (air equivalent), space/Class K up to 10⁻⁹ | Vacuum chamber in accordance with MIL-STD-883, Test Method 1014 |
| Glass-to-metal feedthroughs (hermetic enclosures) | 10⁻⁸ mbar·l/s (typical) | Vacuum chamber bombing |
| Vacuum-tight components (UHV, particle physics) | 10⁻⁹ to 10⁻¹² mbar·l/s | Vacuum chamber, calibrated mass spectrometer |
| Implantable cardiac and neurological devices | 10⁻⁸ to 10⁻⁹ mbar·l/s | FDA/ISO-specific protocols |
| Laboratory capsules for high-pressure research (diamond-piston cells, platinum) | 10⁻⁶ to 10⁻⁸ mbar·l/s | Vacuum chamber bombing |
| Lithium-ion cell sealing (series) | ≤ 10⁻⁶ mbar·l/s, trend toward stricter limits | Sniffing + production scaling |
Note on MIL limits: MIL-STD-883, Test Method 1014, grades the rejection limits based on the internal volume of the enclosure: up to 0.05 cm³, the limit is 5×10⁻⁸; from 0.05 to 0.4 cm³, it is 1×10⁻⁷; and above 0.4 cm³, it is 1×10⁻⁶ mbar·l/s, in each case expressed as an air equivalent. For space hybrids (Class K), the limits are one decade stricter in each case. General values such as “10⁻⁸ to 10⁻⁹” describe the strictest end of this scale.
Overview of testing methods
Vacuum chamber (highest accuracy)
The component is placed in a vacuum chamber connected to a helium mass spectrometer. If the component has been previously loaded with helium (through helium bombing or filling during sealing), helium diffuses outward through any leaks and is detected. Highest sensitivity: up to 10⁻¹² mbar·l/s in an optimized laboratory setup; typical detection limit in practice at service providers is around 10⁻⁹ mbar·l/s. Considered the standard for medical applications, aviation, and brachytherapy seeds. However, the process is slower per component due to the required chamber pumping time.
Sniff probe
The component is filled with helium, and a handheld probe scans the seam to detect helium escaping into the ambient air. This method is ideal for production line testing, as it can be performed quickly and without the need for a chamber. Sensitivity typically ranges from 10⁻⁵ to 10⁻⁷ mbar·l/s. The method is portable and can pinpoint leak locations, which is an advantage for repair workflows.
Pressure bombing (helium precharge)
A sealed component is placed in a pressure vessel filled with helium for a specified duration, and helium diffuses through any leaks into the interior. The helium content is then tested in the vacuum chamber. This enables post-seal testing of components that were not filled with helium during welding. The measurement must be performed within one hour after the bombing process, but may be extended to up to four hours using a qualified procedure. The fixed conditions cover internal volumes of up to only 20 cm³, which in practice means that larger housings are filled directly with helium during sealing (e.g., in a glove box) rather than being bombed afterward.
Pressure drop and mass flow methods
For larger volumes where helium tracer tests are impractical, the pressure drop test (sealed component under pressure, with pressure drop measured over time) provides a coarser leak measurement. This method is less sensitive than helium tracer methods, but it is acceptable for general industrial hermetic specifications and is independent of the tracer gas.
Fine leak and gross leak: the test sequence matters
Paradoxically, a large leak may appear “tight” in a fine-leak test because the tracer gas has already escaped through the large leak before the measurement is taken, causing the mass spectrometer to detect no helium particles. MIL-STD-883, Test Method 1014, therefore combines the fine-leak and gross-leak tests in a defined sequence with tight time windows following the bombing. The simple bubble test in a water bath is suitable here as a quick preliminary check in the workshop, but it is unsuitable for detecting small leaks.
Interpreting false readings: not every signal on the leak detector indicates a weld defect. Permeation is the passage of gas through the material itself, which is particularly relevant for polymer seals and glass. In this case, the solution lies in the choice of material, not in weld correction. Virtual leaks are trapped volumes of gas within the component (cavities, gaps behind the weld, blind holes); during a vacuum test, they outgas like a real leak, but no leak point can be detected from the outside. Therefore, gap geometries behind the weld should be avoided in the design.
Units and conversion
Three units dominate this field. Rule of thumb: 1 mbar·l/s = 0.1 Pa·m³/s.
| Unit | Conversion | Typical context |
|---|---|---|
| mbar·l/s | Unit of reference in this guide | European practice, leak detector displays |
| Pa·m³/s | 1 Pa·m³/s = 10 mbar·l/s | SI unit; ISO standards, ASTM scope specifications |
| atm·cm³/s | 1 atm·cm³/s ≈ 1.013 mbar·l/s | MIL-STD-883, U.S. specifications |
Two pitfalls in U.S. specifications: MIL-STD-883 specifies leak rates in atm·cm³/s; these values are practically identical to mbar·l/s. Furthermore, the rejection limits are based on air equivalents; measured helium rates must be converted, using a factor of 0.37 for fixed conditions and the howl-mann equation for flexible conditions.
Standards and regulations
| Rules | Content and relevance |
|---|---|
| MIL-STD-883, Test Method 1014 “Seal” | U.S. benchmark for microelectronics packages: fine/gross leak test conditions, bombing parameters, volume-dependent reject limits. Base document currently at Revision L with change 1 dated June 27, 2025; most recent documented method revision 1014.17. |
| DIN EN ISO 20485:2018 (ISO 20485:2017) | Key civil standard for leak testing using tracer gas. |
| DIN EN 1779:1999-10 | Criteria for the selection of test methods and test procedures; draft standard 2024-12 in progress. |
| ASTM E493/E493M-11 (2022) | Inside-out test mode for pre-sealed components (semiconductors, hermetic relays); fits directly onto welded capsules and housings. |
| ASTM E498/E498M-11 (2022) | Tracer probe mode: test specimen evacuated, tracer gas external; scope starting at 1×10⁻⁸ Pa·m³/s. |
| ASTM E499/E499M-11 (2017) | Detector-probe mode (sniffing); scope starting at 1×10⁻⁷ Pa·m³/s. |
| ASTM F2391-22 | Helium leak test for implant and medical device packaging; classifies leaks from large (10⁻² to 10⁻⁵ Pa·m³/s) to ultra-fine (10⁻⁹ to 10⁻¹¹ Pa·m³/s). |
The terminology standard DIN EN ISO 20484 (which replaces DIN EN 1330-8) defines the relevant terminology.
How weld preparation affects test results
A leak test cannot save a poorly prepared weld. The factors that determine whether the test is passed occur before the welding step:
- Cleanliness of the joint surfaces. Contaminants in pores, recesses, or crevices burn off during welding and cause porosity in the weld. Ultrasonic cleaning is the standard practice in the laboratory; for deep recesses or threads, a stiff brush and abrasive cleaner should also be used.
- Joint geometry. A weld can only seal surfaces that are geometrically in contact with each other. A 0.1-mm gap can typically be bridged, whereas a 0.5-mm gap typically no longer meets the hermetic specifications.
- Material compatibility. Proven for hermetic micro-TIG welding: 316L stainless steel (DIN 1.4404), titanium grades 2 and 5, platinum and platinum-rhodium alloys, Kovar, nickel-based superalloys, gold, and gold-palladium alloys. Tantalum, niobium, and zirconium typically require preliminary test welds.
- Pulse parameters matching the wall thickness. For thin capsule walls (0.1–0.3 mm), low energy and short pulse durations (15–25% energy, 1.0–3.5 ms) prevent burn-through and porosity. For thicker specimens, longer pulses with deeper penetration ensure complete fusion of the seam.
- Inert gas integrity. Argon ≥ 99.9% (argon 4.6) at approximately 2 l/min with automatic pre- and post-flow protects the weld pool from atmospheric contamination. Insufficient inert gas causes oxide inclusions, which act as leak paths.
- First-article validation. For new geometries or materials, cut a sample weld cross-sectionally before starting series production. Full penetration depth and oxide-free weld metal in the cross-section are the most reliable indicators that the helium leak test will be successful.
Common errors and how to correct them
| Symptoms | Common cause | Correction |
|---|---|---|
| Local porosity | Surface contamination caused by burning | Ultrasonic cleaning, brushing, repeat preparation |
| Insufficient penetration depth | Pulse duration too short for wall thickness | Increase pulse duration in 0.5-ms increments; perform a cross-section cut for verification |
| Oxide inclusions | Insufficient argon pre-flow | Check that the argon flow is ≥ 2 l/min and that the pre-flow is active |
| Leak at a seam corner / edge | Gap geometry or fixture offset | Improve the fixture; place tack welds before the through-weld |
| Weld cracks | Excessive heat input with crack-prone alloys | Reduce energy, shorten the pulse, or change the material program if necessary |
| Fluctuating leak rate from batch to batch | Operator parameter drift or electrode wear | Use the program memory and electrode counter (the MAW provides both via Modbus) |
How the Lampert Micro Arc Welder prepares parts for leak tests
The Lampert Micro Arc Welder (MAW) is designed to produce reproducible, leak-test-compatible hermetic welds:
- 5–1,200 a peak current and 0.1–34 ms pulse duration. The energy sector ranges from thin laboratory capsules to rugged industrial sensor housings.
- 12 preinstalled material programs for stainless steel, titanium, platinum, copper, aluminum, and more. No more guesswork with parameters during initial sample qualification.
- Patented process monitoring. Real-time defect detection (contact, pressure, electrode wear, temperature, “electrode sticking”). Defective welds are flagged before they reach the leak tester.
- Modbus TCP/IP with 21 documented registers for production integration, parameter logging, electrode wear tracking, and weld spot counting. Parameter logging supports traceability requirements for typical medical and aerospace applications.
- Automatic argon pre- and post-flow. Consistent inert gas integrity at approximately 2 l/min of argon ≥ 99.9%.
- Glove-box compatible. For air- or moisture-sensitive encapsulation and for enclosures whose internal volume exceeds the practical bombing limit: if you fill the enclosure directly with helium during sealing, you won’t need to perform bombing afterward.
The article “Hermetic Sealing with micro-TIG welding” provides an overview of applications, including materials and recommended parameters; the Lampert welding principle explains how the welding pulse is generated.
Device recommendation: the Lampert Micro Arc Welder
| Specification | Value |
|---|---|
| Peak current (TIG) | 5 to 1,200 a |
| Pulse duration | 0.1 to 34 ms |
| Minimum workpiece thickness | 0.1 mm |
| Weld spot diameter | 0.2 to 4.0 mm; for depths over 1 mm, use a 1.3 mm electrode |
| Material sets (default) | 12 (universal, gold, silver, platinum, palladium, bronze, stainless steel, titanium, aluminum, tin, brass, copper) |
| Aluminum mode | Specially optimized (HF-superimposed welding curve) |
| Industry 4.0 interface | Modbus TCP/IP via LAN (21 documented registers) |
| Patented welding process control | Yes (real-time error detection) |
| inert gas | Argon ≥ 99.9%, approx. 2 L/min |
| Weight | 10.9 kg |
| Certification | EN 60974-6, EN 61000-6-2/-6-4, RoHS 2011/65/EU; UKCA-compliant |
| Warranty | 1 year; developed and manufactured in Germany |
| Investment (full starter package) | Starting at approximately 7,000 EUR net |
Frequently asked questions about the helium leak test
There is no single, universally accepted definition. “Hermetic” is application-specific: medical implant standards typically require 10⁻⁸ to 10⁻⁹ mbar·l/s; general industrial hermetic standards accept 10⁻⁴ to 10⁻⁵; and high-vacuum and aerospace applications require values below 10⁻⁹. For context: “bacteria-tight” corresponds to approximately 10⁻⁴ mbar·l/s, while “virus-tight” ranges from 10⁻⁶ to 10⁻⁸ mbar·l/s, depending on the source. Define the target value before designing the weld.
Yes. The sniffer probe achieves a sensitivity of 10⁻⁵ to 10⁻⁷ mbar·l/s and is suitable for production line and field testing. For higher sensitivity (10⁻⁸ and better), the vacuum chamber is the industry standard.
No. Pressure bombing after welding fills the sealed component with helium and allows for leak testing even on components that were sealed in argon or air. Bombing pressure (30 to 75 psia) and dwell time are defined in MIL-STD-883, Test Method 1014, based on volume; the measurement must be performed within 1 hour after bombing, or up to 4 hours using a qualified procedure.
Only to a limited extent. The fixed conditions of Test Method 1014 cover internal volumes up to 20 cm³; above that, either the bombing times increase significantly (up to 190 hours are listed for Class K hybrids) or the detection sensitivity decreases. Practical recommendation: when sealing larger enclosures, fill them directly with helium, for example, when welding in a glove box.
Yes, with proper preparation, the right range of materials, validated pulse parameters, and inert gas integrity. Micro-TIG welds reliably meet typical medical and electronic hermetic specifications using validated parameters. For high-vacuum or research specifications (below 10⁻⁹), validate sample welds in advance.
316L stainless steel (DIN 1.4404), Grade 2 and 5 titanium, platinum and platinum-rhodium alloys, Kovar, and nickel-based superalloys are the established standards. Tantalum, niobium, and zirconium are suitable under certain conditions; request a test weld.
Yes. With the proper preparation, the Lampert MAW can operate in inert gas environments. This is important for encapsulations that are sensitive to air or moisture, and for components whose internal volume makes subsequent gas flushing impractical.
Yes. The standard carrier gas is a mixture of 5% H₂ in 95% N₂ (non-flammable), which is less expensive than helium. Helium costs can also be reduced through recovery or by using lower concentrations. For maximum sensitivity, helium remains the standard.
For wall thicknesses of 0.1 to 0.3 mm: 15 to 25% energy, 0.5 to 1.5 ms pulse duration, tapered tungsten electrode, argon at approximately 2 l/min. Validate the initial sample using a cross-section.
Contact Lampert application engineering at [email protected]. Upon request, we provide free sample welds with a written welding report, particularly useful when the target leak rate, geometry, or material differ from the standard.
Conclusion: first the target leakage rate, then the weld seam
The helium leak test is the gold standard for hermetic welds, ranging from general industrial specifications at 10⁻⁵ mbar·l/s to aerospace and medical implant standards below 10⁻⁹. Manufacturers of hermetic components first define the target leak rate, then select the test method (vacuum chamber for maximum sensitivity, sniffer probe for production lines, bombing for post-seal testing), and consistently tailor the welding preparation to these requirements. Clean joint surfaces, controlled gap dimensions, appropriate pulse parameters, and intact inert gas coverage determine the test result long before the component even enters the leak tester.
The Lampert Micro Arc Welder provides the tools for this: 12 material programs, patented process monitoring, automatic argon management, and Modbus logging for traceability. With proper preparation, MAW welds reliably meet typical industrial hermetic specifications.
Free sample welds with a written welding report included: for questions regarding specific applications, please contact Lampert application engineering at [email protected].