Resistor Electromigration Testing Methods — What Actually Works in the Lab

Electromigration kills resistors slowly. You will not see it on day one. You will not see it after a thousand hours either. But somewhere between those two points, the resistive element starts losing material atom by atom, the resistance value creeps upward, and eventually the part fails open or shorts out. That is why electromigration testing is not something you skip — it is something you do before the field does it for you.

Google trends show searches for "resistor electromigration test" and "thin film resistor reliability" have been climbing steadily. More engineers are realizing that DC bias life testing alone does not tell the full story. You need dedicated electromigration validation, and here is how to do it right.

Why Electromigration Happens Inside a Resistor

Most people think electromigration is a PCB trace problem. It is, but resistors suffer from it too — especially thin film and thick film types where the resistive element is a narrow metal or metal-oxide line.

When current flows through that narrow path, electrons collide with metal atoms and push them in the direction of electron flow. Over time, atoms accumulate at one end and vacancies form at the other. The cross-section of the resistive element thins out. Resistance goes up. Eventually the line breaks.

Temperature makes this worse. Every 10°C increase roughly doubles the electromigration rate. That is why a resistor running hot under load will fail faster than one sitting at room temperature with the same current through it.

The failure is silent. No smoke, no pop, no warning. Just a slow resistance drift that pushes your circuit out of spec.

The Two Main Test Approaches

Accelerated DC Bias Life Testing

This is the workhorse method. You apply a DC voltage or current well above normal operating levels, crank up the temperature, and watch how long the resistor lasts before its resistance drifts beyond a defined limit — usually 1% or 5% depending on the spec.

The acceleration comes from two levers: temperature and current density. You push both up to compress years of field life into weeks of lab time. Typical test temperatures range from 125°C to 175°C. Current density can go 10x or more above rated value.

The key is to run enough samples. Five parts is not enough. You need at least 20 to 30 for statistical validity. Record resistance at regular intervals — every 100 hours is a good cadence. Plot the drift curve. If the curve bends upward sharply, you are seeing the onset of electromigration damage.

One thing people get wrong: they run the test at one temperature and one current, then declare the part "passed." That tells you almost nothing. You need at least two temperature-current combinations to extract an activation energy. Without that, you cannot extrapolate to real-world conditions with any confidence.

Temperature-Humidity Bias Testing

Electromigration does not happen in a vacuum. Moisture gets in, especially around the terminations and the resistive element edges. Water molecules accelerate metal ion migration and can cause corrosion-assisted electromigration.

So you run the same DC bias test but inside a humidity chamber. Typical conditions are 85°C / 85% RH. Some standards push it to 1000 hours minimum.

This test catches failures that dry bias testing misses. A resistor might pass 1000 hours at 150°C with no humidity and then fail at 600 hours when moisture is introduced. That difference matters a lot in automotive or outdoor applications.

In-Situ Monitoring — The Method That Changes Everything

Traditional electromigration testing measures resistance before and after. That gives you a pass/fail but no insight into what happened in between.

In-situ monitoring changes that. You measure resistance continuously during the test — every second, every minute, whatever your data logger can handle. The result is a real-time drift curve instead of a before-and-after snapshot.

What you are looking for is the shape of that curve. A smooth, linear drift usually means uniform degradation. A sudden step change means a localized failure — a void opened up or a crack propagated. That information tells you exactly where the weak point is in the resistor construction.

Some labs go further and use focused ion beam cross-sectioning on failed samples. You cut the resistor in half and look at the resistive element under SEM. You will see the voids, the hillocks, the grain boundary separation. It is ugly but it is the most honest data you can get.

Thermal imaging during bias testing is another trick worth using. A hot spot on the resistor body means current is crowding in one area. That is where electromigration will start first. Catch it early and you understand the failure mechanism without destroying the part.

Test Parameters That Actually Matter

Current density is the single most important variable. Not total current — current density, which is current divided by the cross-sectional area of the resistive element. Two resistors with the same rated current can have very different current densities depending on their construction. A thin film resistor with a narrow resistive track will electromigrate much faster than a thick film part with a wide track, even at the same nominal current.

Temperature must be controlled precisely. A ±2°C variation at 150°C changes the electromigration rate by roughly 15%. Use a calibrated oven, not a hot plate. Hot plates have hot spots and temperature gradients that ruin your data.

Dwell time between measurements matters too. If you check resistance every 1000 hours, you will miss the inflection point where the drift rate accelerates. Check every 50 to 100 hours during the first half of the test, then you can space it out as the curve stabilizes.

What the Standards Actually Say

IEC 60115 covers general testing for fixed resistors and includes sections on load life and surge, but electromigration specifically falls under the endurance testing clauses. MIL-STD-202 has method 108 for electromigration of thin film resistors — it is old but still referenced in many military specs.

AEC-Q200 for automotive applications requires electromigration-aware life testing at elevated temperature with bias. The exact parameters depend on the resistance range and termination style, but the principle is the same: push the part hard, measure drift, and fail it if it exceeds the allowed tolerance.

JEDEC standards also have relevant guidance, especially for thin film networks used in precision applications. If you are working in those spaces, check JESD22-A114 for electromigration test procedures.

Common Mistakes That Waste Your Time

Running the test at rated current and calling it an electromigration test. That is just a life test. Electromigration requires current density well above normal operating levels. If you are not pushing at least 3x to 5x the rated current density, you are not testing electromigration — you are testing something else.

Not measuring post-test resistance drift under no-bias conditions. After you remove the bias, the resistor cools down. Some of the drift you measured at high temperature recovers. You need to let the part sit at room temperature for 24 hours, then measure again. That room-temperature reading is what matters for spec compliance.

Ignoring termination failure. Electromigration damage often starts at the termination-to-element interface, not in the middle of the resistive track. If you only measure end-to-end resistance, you might miss a partial failure where one end has lifted off. Use four-wire Kelvin measurement to catch this.

The Bottom Line on Test Planning

Electromigration testing is not about running a test and getting a pass/fail stamp. It is about understanding how your resistor degrades over time under stress. The data you collect becomes the foundation for your derating curves, your field life predictions, and your design margins.

If you skip this test, you are guessing. And in power electronics, automotive, and aerospace, guessing is not an option.