Hermetic Package Welding: Leak, Cross-Section and Functional Validation

Validate hermetic package laser welding with three separate checks: inspect the weld structure, test the complete package for leakage, and verify the enclosed device still functions. A good-looking seam or a sound cross-section cannot establish whole-package hermeticity; a leak-test pass cannot establish electrical or optical performance. Define all three acceptance routes before sample welding, using the actual package construction and the customer's governing specification.

For laser source and joint selection, use the semiconductor package laser welding guide. This guide focuses on the evidence needed to accept the welded package.

Define separate acceptance evidence

The following matrix is a starting point for the customer-approved validation plan. Set numerical limits, sample quantities and disposition rules before testing.

AMADA WELD TECH separately lists metallographic sectioning, seal testing and weld-strength testing among its application-development tools. These methods answer different questions, rather than replacing one another. AMADA application support.

Choose a leak test for the complete package

Include the lid seam, package body, windows, feedthroughs and other seals in the assessment. A leak may not originate in the laser weld. Where required by the acquisition specification, apply MIL-STD-883 Method 1014 with its specified conditions. Do not copy a leak limit from an unrelated package or treat the standard's name alone as a complete test procedure.

Record the variables behind the result

Agree the internal free cavity volume, package pressure capability, tracer gas, exposure pressure and duration, delay before measurement, temperature and method compatibility. Fine and gross leak coverage must follow the selected route: a fine-leak reading alone is not a universal substitute for gross-leak detection. Confirm any fluid exposure or tracer treatment is compatible with the device and subsequent tests.

Keep the measured tracer-gas result distinct from any equivalent standard leak rate used for acceptance. State units, reference conditions and the prescribed conversion. Retain actual timestamps rather than only a pass label. The applicable method determines test sequence and retest permission.

Control background and calibration

NASA's hermeticity correlation work found that helium desorption from package materials could affect measurements and examined differences between test systems. Review background signals, representative package behavior and instrument capability when setting the procedure, particularly for small cavities. NASA NEPP hermeticity study.

Request the reference-leak certificate, uncertainty information and system verification record appropriate to the measurement. NIST describes helium-flow and leak calibration using primary and comparative flow standards; this supports traceable measurement, not certification of a welding machine or a package. NIST flow and leak calibration.

Use cross-sections to examine weld structure

Agree a destructive sampling plan covering representative straight seams, corners, start and stop regions, and geometry transitions where applicable. Map every section to the package and welding program. Document preparation, magnification and measurement references so another reviewer can interpret the images.

Measure the fusion geometry required by the joint design, and assess cracks, pores, incomplete fusion, thinning and damage near sensitive features against agreed criteria. Do not apply one penetration percentage to every lid thickness and joint type. A single section samples one location; it cannot prove continuity around the entire perimeter or the integrity of feedthrough seals.

Process signals can help identify variation, but they need correlation with accepted part results. AMADA's laser-welding material describes both destructive checks and process monitoring. A stable monitoring trace should not automatically replace the package acceptance tests. AMADA laser welding monitoring and troubleshooting.

Verify function and risk based stability

Test the device before sealing and after welding under comparable conditions. Select relevant measurements such as leakage current, insulation resistance, output, offset, optical alignment or RF response from the device specification. Define bias, temperature, fixtures and allowable changes. Separate welding-related damage from assembly or measurement variation.

Choose environmental or operating stress from the product's qualification plan and failure risks. Where the plan requires it, repeat leak and functional checks after stress to assess change. Record stress conditions, sequence and sample allocation; destructive section samples cannot also serve as intact packages for later leak testing. Additional internal-gas or contamination checks may be needed when the internal atmosphere is itself an acceptance requirement.

Send a usable validation RFQ

• Package drawing, lid and housing materials, coatings, seal geometry and tolerance stack.
• Internal free volume, feedthrough and window construction, and pressure or fluid restrictions.
• Device thermal budget, protected regions, pre-seal cleaning and atmosphere requirements.
• Governing test method and revision, leak-rate definition, units, limits and permitted retest.
• Section locations, defect criteria, functional limits, stress sequence and sample quantities.
• Required traceability, raw-data format, independent laboratory responsibilities and production throughput.

Provide these inputs with the equipment RFQ checklist when reviewing the semiconductor packaging precision laser welding system.

Carry accepted results into equipment handover

Link each sample to its material lot, fixture, program revision, weld settings and test records. Investigate a failure against all three evidence routes before changing settings. Freeze the accepted process window and define which material, geometry or equipment changes trigger revalidation. Use the FAT and SAT checklist to agree repeatability trials and delivery records; factory acceptance alone does not replace product qualification.

Common questions

Does a good cross-section prove hermeticity

No. It describes the sectioned location. Whole-package leak testing is needed to assess leakage through the complete enclosure under the selected procedure.

Is one helium leak limit suitable for every package

No. Use the governing specification and applicable conditions for the cavity, package construction and required device class. Keep measured helium and air-equivalent values clearly identified.

Can a leak-test pass replace functional testing

No. Seal integrity and device performance are different acceptance questions. Retain both results for the same identified sample where the test sequence permits.

Must every device receive the same stress test

No. Select stresses, severity, sequence and post-stress checks through the customer's qualification plan and the intended operating risks.

Explore the semiconductor manufacturing and laser process map, or submit the package drawing and validation requirements.