A static strike of a few hundred volts can puncture a semiconductor junction and leave no visible mark. The diode still lights. It still meets forward voltage at the test current used on the line. Then reverse leakage climbs over weeks and the part fails in the fixture. That delay is why ESD control is a process discipline rather than an inspection step.
Three Models, Three Ratings
ESD ratings are quoted against a model rather than a single voltage, because the shape of the discharge pulse decides how much energy reaches the die.
| Model | Charge source | Governing standard | Typical LED rating |
|---|---|---|---|
| HBM | A charged person touching the part | ANSI/ESDA/JEDEC JS-001 | 1 kV to 8 kV, class 1A to 3B |
| CDM | The part itself, charged then discharged | JS-002 | Often lower than the HBM figure |
| MM | A charged machine or metal tool | Largely withdrawn | Legacy sheets only |
An HBM class 1A part survives 250V to 499V. Many LED chips are quoted at HBM 2 kV, which sounds generous until the CDM number appears beside it and turns out to be far lower. A datasheet that lists one figure without naming the model is telling you nothing usable.
What a Strike Actually Does
The failure path in an InGaN or AlInGaP chip usually runs through the mesa sidewall or a bond pad edge, where the dielectric is thinnest. Three outcomes follow. Small damage raises reverse leakage and slightly shifts forward voltage. Larger damage creates a local hot spot that degrades light output over time. The worst case shorts the junction and the die goes dark immediately. Only the third case is visible at assembly.
Test Methods That Find Latent Damage
- Reverse leakage at rated reverse voltage, measured before and after any handling change.
- Forward voltage at a low current such as 1mA, compared with the population median.
- Curve tracer sweep of the reverse branch, looking for a soft knee rather than a sharp breakdown.
- Thermal imaging of a powered strip, where a damaged die runs cooler or hotter than its neighbors.
A single forward voltage check at rated current hides small damage, because the junction still conducts normally in that region.
Handling Rules That Change the Outcome
1. Wrist straps at 1 megohm to a common ground, tested at the start of each shift.
2. Dissipative work surfaces and containers, never plain plastic bags or foam.
3. Ionizers wherever insulators cannot be grounded, because tape, labels and bare housings hold charge.
4. Relative humidity held between 40% and 60%, since dry air lets charge build and stay.
5. Grounded soldering irons and reflow carriers, checked on the same schedule as the wrist straps.
6. No peeling of carrier tape by hand over an open bin, because the separation charges the tape.
7. Incoming and outgoing transit in closed conductive tubes or reels.
Where It Goes Wrong in Practice
The most common escape is a rework bench. High-power LEDs get pulled, re-potted and re-soldered by hand, often outside the ESD area set up for the SMT line. The second is an ungrounded vacuum pickup. Both produce parts that pass the immediate test.
For incoming checks, define the sample and the parameter before the shipment arrives. A practical screen is reverse leakage measured at the rated reverse voltage on 30 pieces per lot, plus forward voltage at 1mA on the same pieces, with a reject limit set from the supplier’s own distribution. Record the ambient humidity during the test, because the number moves with it.
Related reading
