Research on Thermal Testing Method of Double-sided Heat Dissipation Automotive IGBT Module

Research on Thermal Testing Method of Double-sided Heat Dissipation Automotive IGBT Module

Compared with traditional single-sided cooling (SSC) power modules, double-sided cooling (DSC) modules have more robust heat dissipation capabilities and lower parasitic parameters. To further improve the efficiency, power density, and reliability of automotive motor controllers, the application of double-sided cooling power modules in electric vehicles has received more and more attention. With the successful mass application of double-sided cooling automotive IGBT devices in manufacturers such as Toyota (Denso), GM (Delphi), and Tesla (ST), the market demand for double-sided cooling IGBT modules has increased sharply.

The double-sided heat dissipation power module adopts an advanced three-dimensional packaging structure with multiple heat transfer channels. The existing thermal resistance test method still uses the thermal resistance test method of single-channel heat transfer. Unlike traditional single-sided cooling IGBT modules, double-sided cooling automotive IGBT modules simultaneously conduct heat to both the front and back sides. Its heat dissipation method is similar to that of a press-fit IGBT module. Due to the different package structures, the internal heat dissipation path and thermal resistance will be quite different, and the evaluation method needs to be reconsidered. Only a few major manufacturers, such as Infineon, have launched serialized products for double-sided heat dissipation.

Thermal Test Protocol

IGBT junction temperature testing methods mainly include the thermal parameter, finite element simulation, sensor calibration, infrared scanning, etc. The structural characteristics of the double-sided cooling IGBT module determine that it requires very high contact thermal resistance. However, the camber problem introduced by the unique process of the X module will lead to a poor direct crimping effect between the heat dissipation surface and the heat sink. The premise of the double-interface method test is to ensure that the chip junction-case heat dissipation paths are consistent under the two interface conditions. Direct press-fitting will make it inconsistent with the course of the silicone grease interface, and the front part of the structure-function curve will not coincide, resulting in the inability to test the thermal resistance accurately. Therefore, the traditional double-interface method is unsuitable for thermal testing automotive IGBT modules with double-sided heat dissipation. It is necessary to develop new interface materials instead of direct crimping to ensure the consistency of the heat dissipation paths of the two interfaces.

To solve the above problems, this paper innovatively proposes a thermal testing method for a dual-interface heat dissipation structure. To optimize the traditional double interface method, two different thermal interface materials A and B, separate the structure-function curves. The test method of double-sided coupling thermal resistance of double-interface materials is shown in Figure 1, and the steps are as follows.

The module's primary and secondary cooling surfaces are covered with thermal interface materials A and B, which are press-fitted on the heat sink. The radiator continues to pass water, the thermal resistance test starts, and the structure-function curves 1-1 1-2 are obtained.

The overlapping part of the two structure function curves is the double-sided junction-case thermal resistance of the module.

The metal surface structure of double-sided cooling automotive IGBT products only conducts heat and not electricity. Its heat dissipation path can be understood as two devices with different power and thermal resistance parameters driving heat to two surfaces simultaneously back to back. As long as we try to achieve double-sided heat dissipation and single-sided heat dissipation for automotive products, the structure-function curves of the primary and secondary heat dissipation surfaces of the IGBT module can be measured in theory. The way to eliminate the coupling effect of double-sided heat conduction is to realize the single-sided conduction of heat. The test method of single-sided thermal resistance of double-interface materials is shown in Figure 2, and the specific steps are as follows.

1) The main heat dissipation surface of the module is covered with heat-insulating material, and the secondary heat dissipation surface is press-fitted on the radiator through interface materials A and B. The radiator continues to pass water to dissipate heat, and the structure-function curves of the secondary heat dissipation surface are measured 2-1, 2-2.

2) The secondary heat dissipation surface of the module is covered with a heat-insulating material, and the primary heat dissipation surface is press-fitted on the radiator through interface materials A and B. The radiator continues to pass water to dissipate heat, and the structure-function curves of the primary heat dissipation surface are measured 3-1, 3-2.

The junction-to-case thermal conduction paths are identical for the two measurements on one side, and only the case-to-sink thermal resistance is different. Therefore, the two structure function curves are separated at the cooling surface of the module, and the overlapping part is the corresponding junction-case thermal resistance. The related single-sided junction-to-case thermal resistance can be obtained through the above method.

Starting from the structural characteristics and heating characteristics of the X module, through structural design, simulation analysis, and optimization, the radiator design with high heat exchange efficiency is realized. Under the condition of the maximum heating power of the device, the temperature difference between the upper and lower cooling surfaces is within 1°C, and the temperature difference between the inlet and outlet water is within 2°C. The overall design of the X-module thermal test fixture is shown in Figure 3.

Material Selection

According to the dual-interface material thermal test plan, it is necessary to select a suitable thermal conductivity and thermal insulation material as the interface to realize the double-sided and single-sided thermal impedance test of the IGBT module. Considering the physical and chemical properties of various materials, it was decided to choose heat-dissipating graphite film and thermally conductive silicone grease as the thermal interface materials for the thermal resistance test of the X module.

Flexible airtel and polyurethane PU glue were screened out as candidate thermal insulation interface materials for the X module thermal test. To verify the actual thermal insulation performance of the two materials, two materials are used to insulate the secondary heat dissipation surface respectively. The thermal resistance comparison test of the main heat dissipation surface uses the heat-conducting material homogeneous graphite film.

The comparison results show that the junction-ring thermal resistance and the highest junction temperature measured under the adiabatic conditions of polyurethane PU adhesive are lower than those of aerogel. It shows that under the same pressing torque, the thermal insulation ability of Airtel is better than that of polyurethane PU glue. Therefore, airtel was chosen as the thermal insulation material for the thermal test.

Thermal Testing and Results Analysis

The press-fit force comparison test conditions were designed to study the influence of the press-fit torque on the junction-case thermal resistance of the X module. As the press-fit torque increases, the device junction-ring thermal resistance of the X-module IGBT and FRD decreases. However, the junction-case thermal resistance test showed no significant change. It shows that for double-sided heat dissipation IGBT modules, different press-fit torques only affect the contact thermal resistance between the device and the heat sink and have no effect on the test of its junction-case thermal resistance.

The thermal simulation model of the X module is shown in Figure 4.

According to the double-interface material method, the single-sided thermal resistance test of the primary and secondary heat dissipation surfaces of the X module is carried out. The test results show that the single-sided thermal resistance of the upper and lower tubes measured by this method has a good consistency.

The deviation between the measured and simulated values of the double-sided thermal resistance of IGBT and FRD is within ±5%. The divergence between the estimated value and the fake value of the thermal resistance of the primary heat dissipation surface is within ±10%. The measured value of the thermal resistance of the secondary cooling surface deviates from the simulated value by about 70%.

Since there is no absolute thermal insulation material, the thermal coupling effect of the double-sided cooling module cannot be eliminated. For the thermal insulation condition of the secondary heat dissipation surface, most of the heat flows through the primary heat dissipation surface without heat insulation material to form an ideal single-sided heat dissipation, and the deviation of the test value is slight. For the adiabatic condition of the primary heat dissipation surface, some heat will flow through the primary heat dissipation surface with heat-insulating material, and the coupling effect is relatively significant. This leads to a large deviation between the test and simulated values.

For the thermal coupling effect in the thermal test of the secondary heat dissipation surface, the correction method is to eliminate the coupling effect by inverting the thermal resistance of the secondary heat dissipation surface based on the measured thermal resistance of the primary heat dissipation surface and both sides. The error of the corrected result is reduced to within 25%.

Repeated press-fit and thermal resistance tests were performed on the X module to verify the repeatability of the thermal test method of the double-sided heat dissipation automotive IGBT module. The results show that the deviation of the five test results is within ±2%, indicating that the thermal test method has good repeatability and generalizability.

Conclusion

Based on the comparative study of thermal test tooling design, interface material selection, and press-fitting methods, a thermal test method for dual heat dissipation interface materials suitable for double-sided heat dissipation automotive IGBT modules is proposed. This method can realize double-sided and single-sided junction-case thermal resistance testing. For the double-sided heat dissipation automotive IGBT module, different press-fit torques within a specific range do not affect the test of its junction-case thermal resistance. The deviation between the measured and simulated values of the double-sided thermal resistance of IGBT and FRD obtained by the double-sided thermal resistance test method is within ±5%. The deviation between the measured and simulated value of single-sided thermal resistance is within ±10%. This can accurately realize the thermal resistance test of the double-sided heat dissipation automotive IGBT, and the results are of the reference value. The thermal coupling effect of the double-sided heat dissipation module cannot be eliminated, resulting in the small matter of the measured and simulated thermal resistance of one side. The coupling effect can be reduced, and the test results can be corrected by inverting the thermal resistance of the secondary heat dissipation surface based on the measured thermal resistance of the primary heat dissipation surface and both sides. The thermal test method has good repeatability and generalizability.