Power Device Corrosion Testing: Standards, Methods, and Reliability Assessment

Power device corrosion testing evaluates the resistance of power semiconductors such as MOSFETs, IGBTs, thyristors, and power modules against moisture, salt fog, and corrosive gas environments that accelerate electrochemical degradation. By applying internationally recognized standards such as IEC 60068-2, IEC 60749, JEDEC JESD22, and AEC-Q101, manufacturers can detect corrosion-induced failure mechanisms early and ensure long-term reliability in automotive, industrial, and renewable energy applications. This article outlines the sample types, test items, methodologies, and acceptance criteria involved in a comprehensive corrosion testing program for power devices.

Technical Overview

Power semiconductor devices, including silicon MOSFETs, IGBTs, thyristors, and wide-bandgap devices based on SiC and GaN, operate under high voltage, high current, and elevated temperature conditions. Their long-term reliability is strongly influenced by the ability of the die, metallization, interconnects, and encapsulation system to withstand corrosive environments. Corrosion testing for power devices is therefore a core element of reliability qualification and failure analysis programs.

Corrosion in power devices is an electrochemical process that requires three simultaneous conditions: an electrolyte (typically condensed moisture or ionic contamination), a corrosive medium (chloride, sulfide, or acid gas species), and an electrochemical driving force (applied bias or galvanic potential between dissimilar metals). Typical corrosion-related failure mechanisms include:

• Aluminum metallization corrosion and bond pad attack caused by chloride or phosphate ions combined with moisture ingress.
• Galvanic corrosion at the interface between copper leadframes, aluminum bond wires, and solder or silver-plated layers.
• Electrochemical migration and dendrite growth between adjacent metallization traces under humidity and bias.
• Corrosion of copper and aluminum baseplates, pins, and terminals exposed to salt fog or industrial atmospheres.
• Mold compound degradation, delamination, and pinhole formation that create pathways for moisture and contaminants to reach active die surfaces.

Accelerated corrosion tests compress years of field exposure into days or weeks by elevating temperature, humidity, contaminant concentration, and electrical stress. The results are used to qualify package designs, screen molding compounds and plating finishes, validate process changes, and support automotive and industrial reliability certifications.

Testing Samples

Discrete Power Devices

Packaged discrete devices such as power diodes, Schottky diodes, MOSFETs, IGBTs, bipolar transistors, thyristors, and TRIACs in through-hole or surface-mount packages are the most common test articles. These samples are tested in their final packaged form, often with leads trimmed or mounted on test boards to reflect real-use conditions.

Power Modules and Assemblies

IGBT power modules, SiC MOSFET modules, intelligent power modules (IPM), and half-bridge or six-pack assemblies are tested to evaluate the corrosion resistance of direct bonded copper (DBC) and active metal brazed (AMB) substrates, baseplates, terminal solder joints, silicone gels, and epoxy encapsulants.

Materials and Test Vehicles

Bare dies, leadframes, bonding wires, substrate coupons, plating samples, and purpose-built interconnect test vehicles are used for material-level screening. These samples allow rapid comparison of plating finishes, die-attach materials, and encapsulation systems before full device qualification.

Board-Level Assemblies

Assembled power boards, including inverter stages, DC-DC converters, and gate driver boards, are tested to assess solder joint corrosion, conformal coating performance, and creepage-path degradation in humid and contaminated environments.

Testing Items

A comprehensive corrosion evaluation for power devices typically combines climatic, contaminant-exposure, and biased-stress tests, followed by electrical and physical characterization:

• Neutral salt spray (NSS) exposure for terminals, leadframes, baseplates, and external metallization.
• Cyclic salt spray and salt mist with controlled humidity transitions for marine and automotive underbody environments.
• Steady-state temperature-humidity bias testing (THB, 85°C / 85% RH with applied bias).
• Highly Accelerated Stress Test (HAST) with and without bias at elevated pressure and temperature.
• Unbiased autoclave or pressure cooker testing (PCT, 121°C / 100% RH) for package moisture resistance.
• Flowing mixed gas (FMG) corrosion testing with controlled concentrations of H2S, SO2, NO2, and Cl2.
• Single-gas exposure tests for sulfur dioxide and hydrogen sulfide atmospheres.
• Moisture sensitivity level (MSL) preconditioning per moisture/reflow sensitivity classification standards.
• Electrochemical migration and insulation resistance testing under humid conditions.
• Post-test electrical characterization, visual inspection, cross-sectioning, and SEM/EDS failure analysis.

Testing Methods

Salt Fog Testing

Neutral salt spray testing is performed to IEC 60068-2-11 or ASTM B117 using a 5% sodium chloride solution at 35°C with a collection rate of 1 to 2 mL per 80 cm² per hour and a solution pH between 6.5 and 7.2. Cyclic salt mist testing to IEC 60068-2-52 alternates salt fog exposure with storage at high humidity, reproducing wet-dry cycling seen in coastal and road-salt environments. Military-grade devices may be assessed using the salt atmosphere method of MIL-STD-883. Typical exposure durations range from 24 to 1000 hours depending on the application severity level.

Damp Heat and Biased Humidity Testing

The steady-state humidity bias life test per JEDEC JESD22-A101 and IEC 60749 subjects devices to 85°C and 85% relative humidity while a DC bias is applied to active terminals. The bias reverses the local electrochemical potential and drives electrolytic corrosion at vulnerable nodes, making this test highly effective at exposing aluminum corrosion and bond pad attack. Automotive discrete device qualification under AEC-Q101 requires extended 85°C / 85% RH biased exposure, commonly 1000 hours, with intermediate electrical readouts.

HAST and Pressure Cooker Testing

The Highly Accelerated Stress Test per JEDEC JESD22-A110 combines approximately 130°C, 85% relative humidity, and elevated pressure (around 2.3 atmospheres), reducing test duration to 96 or 264 hours while achieving acceleration factors far beyond conventional damp heat. Unbiased autoclave testing at 121°C and 100% relative humidity per JEDEC JESD22-A102 evaluates bulk package moisture resistance and delamination without electrical stress.

Corrosive Gas Testing

Industrial and outdoor atmospheres contain sulfur-bearing and chlorine-bearing gases that attack silver, copper, and nickel surfaces. Single-gas tests to IEC 60068-2-42 (sulfur dioxide) and IEC 60068-2-43 (hydrogen sulfide) expose devices at controlled concentrations and humidity. Flowing mixed gas testing to IEC 60068-2-60, Method 1 through Method 4, combines multiple gas species at parts-per-billion to parts-per-million levels to simulate petrochemical, urban, and marine industrial environments. Exposure durations typically range from 4 to 21 days, with severity selected according to the intended service environment.

Post-Test Evaluation and Failure Analysis

After exposure, devices undergo electrical parameter measurement including threshold voltage, on-state resistance, forward voltage drop, leakage current, and insulation resistance, with drift limits defined by the applicable product specification or qualification plan. Physical evaluation includes external visual examination for corrosion products, discoloration, and lead damage, followed where necessary by acoustic microscopy for delamination, cross-sectioning, and scanning electron microscopy with energy-dispersive X-ray spectroscopy to identify corrosive species and failure sites. Wire pull and die shear testing quantify the residual mechanical strength of interconnects affected by corrosion.

Applications

Corrosion testing of power devices supports qualification and reliability assurance across every major power electronics market:

• Automotive: electric vehicle traction inverters, on-board chargers, DC-DC converters, and 48V mild hybrid systems qualified under AEC-Q101 and OEM-specific environmental schedules.
• Renewable energy: solar string and central inverters, wind turbine converters, and energy storage systems exposed to humidity, salt fog, and pollution gases for decades in the field.
• Industrial: motor drives, soft starters, welding equipment, and Uninterruptible Power Supply (UPS) systems operating in chemically aggressive factories.
• Rail transit and marine: traction converters and auxiliary power units facing high salinity and condensation cycles.
• Power transmission and smart grid: high-voltage direct current (HVDC) valve assemblies, solid-state transformers, and grid-tied equipment installed in coastal or polluted atmospheres.
• Consumer and data center: fast chargers, server power supplies, and graphics processing unit (GPU) power stages where miniaturization increases sensitivity to ionic contamination.

Test results feed directly into design robustness reviews, package material selection, conformal coating decisions, and warranty lifetime models based on humidity and contamination acceleration factors.

FAQ

What standards are most commonly used for power device corrosion testing?

The core references are IEC 60068-2-11 and IEC 60068-2-52 for salt mist, IEC 60068-2-42, IEC 60068-2-43, and IEC 60068-2-60 for corrosive gases, IEC 60749 for semiconductor-specific climatic tests, JEDEC JESD22-A101, JESD22-A102, and JESD22-A110 for humidity and HAST, ASTM B117 for neutral salt spray, MIL-STD-883 for defense-grade devices, GB/T 4937 for the Chinese semiconductor test method series, and AEC-Q101 for automotive discrete device qualification.

How long does a typical corrosion test take?

Salt spray tests commonly run 24 to 1000 hours, damp heat and THB tests typically run 500 to 1000 hours, HAST compresses equivalent stress into 96 to 264 hours, and flowing mixed gas tests generally run 4 to 21 days. The exact duration depends on the qualification standard, the severity class of the target environment, and the acceptance criteria agreed with the customer.

Why is electrical bias applied during humidity testing?

Bias creates local electrolytic cells that dramatically accelerate corrosion of aluminum metallization and promote electrochemical migration between biased terminals. Unbiased humidity tests mainly evaluate bulk package sealing and moisture ingress, while biased tests reveal the field-realistic failure modes that occur when devices operate in humid, contaminated conditions.

What acceptance criteria are applied after corrosion exposure?

Acceptance is typically based on electrical parameter drift remaining within specification, absence of visible corrosion on leads and terminals beyond defined limits, no delamination or cracking detected by acoustic microscopy, retention of bond strength above minimum values, and confirmation through failure analysis that no corrosion has reached active die areas.

Can accelerated corrosion results be extrapolated to field lifetime?

Yes, when combined with humidity, temperature, and contaminant acceleration models such as Peck's law and Eyring-based models. Accelerated test data, together with knowledge of the actual service environment, allows reliability engineers to estimate field lifetime and define maintenance or derating strategies, although extrapolation should always be validated against real-world failure data.

Should corrosion testing be performed by an internal laboratory or an accredited third-party laboratory?

Both approaches are valid. Internal laboratories provide rapid feedback during development, while accredited third-party laboratories offer standardized chambers, calibrated gas delivery systems, traceable measurement, and internationally recognized test reports required for automotive, industrial, and export certifications.