The use of sinter pastes in modern power semiconductor devices

This article will tell you about the use and assembly of modern power modules IGBT and MOSFET modules, problems encountered in the operation and assembly of these modules and how to solve these problems.

Modern power semiconductors

Power semiconductors have become very widespread in the modern world. They can be found in household appliances, electric trains and electric vehicles, welding machines and so on. Most often, power semiconductor devices are used in voltage converters and motor drivers.
')
Today, MOSFET and IGBT power modules are most common. Moreover, MOSFET is more used in a small power range, and IGBT is more often used in power electronics. Together, IGBT and MOSFET power modules occupy 80% of the power semiconductor market. The traditional devices from which power electronics began: thyristors (SCR), including lockable (GTO), bipolar transistors (BPT) - in recent years, more and more replaced by devices with field control.

IGBT (Insulated - gate bipolar transistor) is an insulated gate bipolar transistor (IGBT) - a three-electrode power electronic device used mainly as a powerful electronic key in switching power supplies, inverters, and electrical drive control systems. Symbols IGBT presented in the figure below. By its internal structure, the IGBT is a cascade connection of two electronic keys: the input key on a field-effect transistor controls a powerful terminal key on a bipolar transistor. The control electrode is called a gate, as in a field-effect transistor, the other two electrodes — an emitter and a collector, as in a bipolar one. Such a composite inclusion of field and bipolar transistors allows combining the advantages of both types of semiconductor devices in one device. Currently, IGBT power modules are available for currents from 10 to 2400 A and switching voltage up to 3.3 kV. Often you can find intelligent power modules (IPM - Intelligent power module). Intelligent power modules (IPM) modules are characterized by the presence of a control card, which contains sensors, protection driver circuits, diagnostics.



MOSFET (metal-oxide-semiconductor field effect transistor) is a MOS structure (metal oxide-semiconductor) - the most widely used type of field-effect transistors. The structure consists of a metal and a semiconductor, separated by a layer of silicon dioxide (SiO2). In general, the structure is called MDP (metal - dielectric - semiconductor). MOSFET transistors are called field-effect or MOS transistors. Symbols, shown in the figure.



An example of IGBT and MOSFET power modules is shown in the figure below.



These power modules are created using DCB technology (Direct Copper Bonding), which literally translates as direct landing on copper. The illustration shows the DCB section of the power module.



Lead-free solder is used as solder in modern power devices, which imposes a number of restrictions on the characteristics of devices. The low melting point of lead-free solder (≈ 250 ° C) imposes restrictions on the maximum operating temperature of 100 - 120 ° C, which makes it impossible to assemble a DCB structure using a gallium arsenide chip or silicon carpide, which have an operating temperature above 200 ° C . Also solder badly withstands thermal cycling - heating to the maximum temperature and subsequent cooling. After several thousand cycles, pores appear cracked on the solder.



Due to the different coefficients of thermal expansion (silicon CTE = 4 µm / mK, solder CTD = 25-30 µm / mK), different parts of the transistor expand in different ways, which leads to the bending of the structure. Due to bending, contact with the radiator is deteriorating, and in the worst case scenario, the silicon chip may crack. This phenomenon is called the bimetallic effect. Due to the bimetallic effect, it is impossible to assemble structures with a silicon chip with a diameter greater than 100 mm. This problem can be solved if the material with a smaller KTR is used as solder than in lead-free solder. Such a material is a paste paste (CTE = 19–20 µm / mK).

Sinter paste

Sinter paste is a silver powder with solvents. Since the technology of paste paste is quite new, there is no established term in the Russian language that could be used to call them. The literal translation sounds like sinter paste. After sinning, the sinter paste is pure silver, all solvents evaporate. The structure is shown below. Grain size from 0.1 to 1 micron. On the left is a photograph of a real sample, on the right is a picture from the documentation for mAgic sinter paste sinter paste from Herause. The table below compares the sinter paste and lead-free solder. It can be seen that Sinter Paste has better thermal and electrical conductivities, as well as lower CTE values.





Sinter paste baked in 2 stages. Before fusing, paste paste is applied by screen printing on the surfaces to be joined. Thick layer 150 microns. The first stage is pre-drying at 80 ° C for 15 minutes. After that, it is sintered at a pressure of 30 MPa and a temperature of 270 ° C for 5 minutes. No-pressure fusion is allowed for chips with a diameter greater than 150 mm.



To study the properties of the paste paste, DCB structure models were used. Copper plates 2 mm thick were used as DCB ceramics. When applied, a stencil with a thickness of 150 µm was used, in which a square hole of 10 by 10 mm was made. To study the homogeneity of the paste paste, the thermo-emf method was used. For this, a thermo-EMF measurement installation was assembled. A heater from a Lukey 702 soldering iron was used as a heater. The same soldering station was the temperature setter. A tungsten needle was attached to the heater. Thermo-EMF values ​​were measured with a UT71D multimeter. All this was mounted on a frame from a PMT-3 microscope, which has a micrometer table and a spring-loaded suspension, on which a heater is fixed.




Samples sintered at various pressures (10, 20, 30, 40 and 50 MPa) were measured on this unit. Measurements were taken in 1 mm increments. Heater temperature 320 ° C. After the measurements, a matrix of 10 by 10 values ​​was obtained. The scattering parameter r was chosen as the homogeneity criterion. It is derived from the thermo-emf value as follows:



The results are tabulated and the homogeneity dependence on the sintering time is plotted.



From the results, it can be concluded that homogeneity increases with increasing pressure, but with increasing pressure it becomes more difficult to observe the plane-parallelism of the connected parts, which can lead to damage to the silicon chip.

To study the dependence of the porosity of the sinter paste on pressure, 5 samples were made, sintered at pressures from 10 to 50 MPa. On specimens sintered at a pressure of 30 MPa and higher, areas of specular gloss are visible, which indicates that in these regions the density approaches that of silver. Since after contacting the contact is pure silver, by calculating the density of the contact and comparing it with the density of silver, one can find out the porosity of the paste. The contacts were separated from the copper base, measured with a micrometer and weighed on electronic scales. The measurement results and calculations are summarized in the table according to which the graph of the dependence of porosity on the sintering pressure is plotted.




To study the dependence of the adhesion of the paste to the coating, 4 samples were made, matched at a pressure of 30 MPa. Coatings of gold, silver and nickel were used. The last sample was uncoated. As can be seen from the table, the best way was gold, the worst - copper.

Conclusion

Sinter paste is shown to be suitable for use in power semiconductor modules as a replacement for lead-free solder. The best indicators of heat and electrical conductivity can increase the power of the devices while maintaining the same dimensions. A higher melting point allows the use of silicon carbide and gallium arsenide chips, which can operate at temperatures of 200 ° C and above. Also, the sinter paste does not impair its properties as a result of thermal cycling, which increases the reliability of the devices.