Thermal Shock Test Chamber For Semiconductors Per JESD22 A106

What thermal shock is

Thermal shock is the abrupt cousin of thermal cycling. Where a cycling test ramps the temperature up and down at a measured rate, taking many minutes to cross between its extremes, JESD22-A106 makes the crossing almost instant, moving the parts from a cold zone to a hot one in seconds and holding each end only long enough for them to reach temperature. The whole purpose is the speed of the change.

A part that has settled at a deep cold, plunged at once into heat, cannot warm evenly; its outer skin heats while its core is still cold, and the mismatch tears at everything inside. The test repeats that violent crossing for hundreds or thousands of transitions, hunting out the cracks and separations that only a steep, sudden gradient will open.

Why the speed is the stress

Speed is the weapon.

The harm in thermal shock is the gradient the speed creates. When a part changes temperature slowly the whole of it moves together, and the internal stress stays modest; when it changes in seconds the surface races ahead of the interior, and for a brief moment the package holds a steep temperature difference between its outside and its middle. That transient gradient is the stress, since the materials at different temperatures want to be different sizes at the same instant, and the shear at their boundaries spikes far higher than a gentle ramp ever produces.

The faster the transition, the steeper the gradient and the harder the shock, so the transition time is the lever the whole method turns on. Where a cycling ramp might move the air ten or fifteen degrees a minute, a shock crosses two hundred degrees in seconds, an effective rate hundreds of times steeper at the surface. A part survives the slow working of cycling and still cracks under the sudden slam of shock, which is exactly what the test is built to expose.

The surface arrives before the core, and the part tears in between.

Shock against cycling

The two methods share extremes and differ in the journey between them. Thermal cycling, the JESD22-A104 method, controls the rate of change, ramping at perhaps ten or fifteen degrees a minute so the part follows the air smoothly, and it reproduces the gradual swings of real service. Thermal shock throws away the controlled ramp and makes the transition as fast as the equipment allows, a harsher stress that packs the same temperature span into seconds rather than an hour.

Shock finds weaknesses faster and screens harder, catching marginal die-attach and package cracks a gentle cycle would take far longer to reveal, while cycling stays closer to the field and is read as the more representative life test. The choice between them is a choice of severity against realism, and a thorough programme uses each for what it does best.

The two-zone air chamber

The common machine for A106 keeps two air chambers always at temperature, one held deep in the cold and one held high in the heat, and moves the parts between them rather than changing the air around a single space. A basket of devices rides on a lift or a shuttle that drops them from the hot zone into the cold and back, and because each zone never has to heat or cool itself, the parts meet the new extreme almost the instant they arrive, which is the whole purpose, since the shock lives in how fast the surface of a part changes temperature while its core lags behind. To deliver that, each zone carries far more thermal mass and reserve than the load, so when a cold basket lands in the hot chamber the air barely sags and recovers in seconds, and the part is driven to the set point fast rather than dragged there slowly. The transfer itself is timed and kept short, since seconds spent in the open air between zones soften the blow and let the standard slip. Some machines add a third, room-temperature zone so a part can be parked without a shock between phases, and the better chambers seal the moving basket and steer the airflow so every device sees the same plunge at the same moment, because a shock that reaches one corner late is no longer the shock the test specifies.

Liquid to liquid, the fierce original

The harshest form of the test drops the parts not through air but through fluid. Two baths of an inert, electrically non-conductive fluorocarbon liquid sit side by side, one chilled and one heated, and a basket lifts the load out of one and lowers it into the other. Liquid carries heat into and out of a part far faster than air, its heat-transfer coefficient higher by a wide margin, so a liquid-to-liquid shock is the steepest gradient the method can deliver and the fiercest version of the test.

That ferocity comes at a price the trade has grown wary of: the fluorocarbon fluids are costly, they drag out of the bath clinging to every part and basket and have to be recovered, and many belong to the family of persistent compounds now under environmental scrutiny. The liquid bath was the original A106 chamber and remains the severest, while the cleaner air-to-air design has taken over much of the everyday work.

Recovery is the hard part

The engineering challenge hides in what happens after the load lands. A basket of cold parts dropped into the hot zone dumps a slug of cold into that air and chills it, and the zone has to drive its temperature back to set point quickly so the parts still see the full hot soak the test demands; the same happens in reverse when a hot load enters the cold zone.

A shock chamber answers this with heavy thermal mass in each zone and generous heating and refrigeration to recover fast, since a zone that sags for minutes after each transfer robs the parts of the extreme they were sent to meet. The recovery time after the load enters is itself a specified figure, the measure of whether the chamber can keep its promise under the repeated insult of cold baskets and hot baskets arriving every few minutes. A box that recovers slowly quietly softens the test it claims to run, the parts seeing perhaps a hundred and thirty degrees where the recipe called for a hundred and fifty.

Transition time, the defining number

One figure separates a real shock from a brisk cycle. JESD22-A106 sets a maximum transition time, the seconds allowed for the parts to cross from one extreme to the other, and a transfer slower than that limit fails to deliver the steep gradient the method depends on. In an air-to-air chamber that limit is often ten seconds or less, and a liquid transfer beats it easily, the load crossing between baths in a second or two.

The standard counts the crossing from the moment the load leaves one zone to the moment it is sealed in the next, and the chambers are built and timed to beat that figure with the heaviest load the basket will carry. It is the one specification that defines whether a machine is shocking the parts or merely cycling them quickly, and it is checked rather than assumed when a chamber is qualified for the method.

What the shock breaks

The failures are the familiar thermomechanical ones, hit harder and found sooner. The steep transient gradient cracks the die where the silicon meets its attach, lifts wire bonds at their heels, fractures the moulding compound at sharp internal corners, and separates the layers of a laminate where their expansion disagrees. The same mismatch of materials that powers thermal cycling drives shock too, only the suddenness concentrates the strain in the moments just after each transfer, when the gradient is steepest.

Parts with internal interfaces under stress, large dies, mismatched materials, voids in the attach, surrender to shock long before they would to a gentle cycle, and the test reads their early failure as a clear warning that the design or the process needs attention.

Extremes, dwell, and the count

The recipe names the two temperatures, the dwell, and the number of crossings. A common pairing runs from around minus sixty-five degrees to a hundred and fifty, the span chosen to bracket the part's rated range with margin, and the dwell at each end is set long enough for the whole load to reach the extreme before the next transfer rather than by the clock alone. The count of cycles, run into the hundreds or thousands, is the result the qualification records, with the crossing to a defined failure standing as the figure of merit. A dwell cut too short leaves the core of a dense part short of the extreme, softening the shock, so the dwell is sized to the load rather than trimmed to save time.

Choosing shock or a gentle cycle

The decision rests on what the programme needs from the test. Thermal shock earns its place as a fast, fierce screen, flushing out package and interconnect weaknesses in fewer cycles and less calendar time than a gentle ramp would take. Thermal cycling earns its place as the more lifelike measure, its controlled rate closer to the swings a part meets switching on and off and warming with the seasons. Many qualification plans run both, the shock to find the gross weaknesses quickly and the cycle to estimate the working life, each answering a different question about the same part.

Pulling it together

The A106 chamber is built around speed: two zones, one cold and one hot, held ready at once, and a carrier that throws the parts between them faster than either can change gently. The sudden crossing forces a steep transient gradient through each package, and that gradient cracks the dies, lifts the bonds, and separates the layers that a slow cycle would only ever work loose over far longer. Whether the transfer is through clean air or fierce fluorocarbon liquid, the transition time defines the severity, the recovery of each zone defends it against the cold and hot loads landing in turn, and the dwell makes sure every part reaches the extreme before it is flung to the other. Run that way, thermal shock is the test that finds in an afternoon what a gentle cycle might take weeks to admit.