Characterizing the TIM as a reference
We placed the TIM between the external cooling surface of a power semiconductor device and a heat sink. The semiconductor was fixed to the heat sink's surface by constant force, maintained by a special fixture (see figure 1). We first characterized the highest temperature elevation and the total junction-to-ambient thermal resistance at the device's operating parameters to obtain reference values.
Figure 1. The measurement system for TIM reliability testing.
This type of test is done to evaluate the durability of TIMs for temperature cycling, induced by power cycles. The surface temperature of the cold-plate can be set and maintained at any temperature between 5C and 100C. In our tests, we used the Mentor Graphics T3Ster transient tester and its high-power accessory that provides the temperature rise in the device. After reaching the hot thermal steady state, the high power was switched off and the cooling transient was measured by the transient tester. These test conditions resembled the device's normal operation. If the quality of the TIM gets worse, the same powering results in higher junction temperatures for each cycle.
Testing TIMs in the package
We used a PNP-type power transistor as the heater element in a TO-220 type package to compare the effects on various TIMs under various temperatures and power applications. We attached the package to a cold-plate. Then we generated the power stepping required for the measurements using a voltage-stepped method: a constant 2A current was forced through the open transistor, while the base-collector voltage was switched in 1-μs steps between -1V and -10V.
After power stepping, we used the base-emitter voltage as a temperature-sensitive parameter to indicate the temperature response. The power step was 18.2W, and the sensitivity of the transistor resulted in -1.1 mV/įC. All measurements lasted for 20 seconds, and it took 10 seconds for the transistor junction to heat up completely, and another 10 seconds to cool down.
We set the surface temperature of the heat sink to 40C during the measurements, which allowed the junction temperature to increase to approximately 120C in each cycle. Of course, this value may change slightly depending on the actual TIM used.
The most straightforward way to see the difference between two measurements is to compare the transient responses. The transients should run together until the point in time when the main trajectory of the heat leaves the package boundary and enters the grease layer. If no other structural element changes the thermal properties in the TIM's heat-flow path, the difference between the total elevations will illustrate the change in the thermal resistance, which can be easily verified using structure functions.
Verifying results using structure functions
The structure functions are generated from the transient results using a mathematical procedure called the NID method. They represent the heat-flow path from the place where the power step takes place (the driving point) toward the outside environment. Both the thermal resistance of each structure in the heat-flow path and the thermal capacitance of the corresponding layer in which the heat spreads can be identified this way.
Because of the extremely high repeatability of the thermal transient method2, as long as the heat flows in the same structure, the structure functions are exactly the same. If the heat-flow path's structure changes, the difference immediately appears on these functions and shows the exact location (thermal resistance value) and magnitude (thermal capacitance value) of the change. Using this method allowed us to determine whether it was actually the TIM that changes thermal resistance in the assembly.
For each TIM, we conducted 2,500 cycles and captured the cooling transient for each cycle using the transient response and structure functions only. This many cycles makes it difficult to handle all of them together and to show clear trends. So we plotted the temperature difference between an early point of the transient and the highest elevation after each measurement.