There is a Packaging Problem to Solve for Silicon Carbide Devices

Published by Olivier Mathieu, Market Development Manager
Advanced Electronics Solutions

There is currently a lot of interest for silicon carbide (SiC) as a semiconductor material because its properties make it more promising than silicon for power electronics applications. At wafer and device level, significant progress has been achieved. Today diodes and transistors with voltage ratings up to 1.2kV are available in mass production from multiple sources of supply and some further technology improvements are already in progress. High voltage SiC devices are not yet available in large volumes but are expected to come soon. While everyone is excited about the new opportunities that will appear with SiC devices, those devices also come with some new challenges and particularly regarding the assembly, interconnection and packaging technologies. Innovative packaging solutions are required!

What are the advantages of SiC devices?

SiC is a wide band gap (WBG) semiconductor material. The band gap generally refers to the energy difference in electron volts (eV) between the valence band and the conduction band. It is the energy required to promote a valence electron bound to an atom to become a conduction electron, which is free to move within the crystal lattice and conducts electric current as a charge carrier. Insulators have a very large band gap, typically higher than 4eV.

Both are semiconductor materials but SiC provides a number of advantages over silicon (Si) as shown in table 1.

Table 1: properties of SiC vs. Si

Compared to Si, SiC has a band gap about 3 times wider and an electric breakdown field about 10 times larger. This means that in SiC devices, the distance of drift region can be reduced to about one-tenth of that of Si devices having the same blocking voltage. Moreover, in SiC devices, the dopant concentration in drift region can be about 100 times higher compared to Si devices. The majority of the on-resistance of a power device having a high blocking voltage is drift region resistance. Therefore, the on-resistance of SiC devices (RDSon) can be about a thousandth that of Si power devices having the same blocking voltage.

The electron drift velocity of SiC is about twice as fast as that of Si. In addition, the drift distance of SiC devices is shorter than that of Si ones with the same blocking voltage. These properties indicate that SiC devices can operate at higher switching frequency than Si devices.

Finally, the thermal conductivity of SiC is about three times that of Si. In addition, SiC has a very high intrinsic semiconductor temperature well above 1000°C. As a consequence, SiC devices can be stable at higher temperatures compared to Si devices.

Impact on market and applications

SiC devices are superior to Si devices in terms of low-loss operation, fast switching time and high temperature operation stability. These characteristics are very attractive for next-generation power modules as they result in multiple system benefits.

High temperature stability means that SiC can either be operated at higher temperatures or tolerate some temperature peaks that may occur depending on the mission profile. Besides, operating at higher switching frequency leads to a system size and weight reduction as the bulky magnetic components are replaced by components with a smaller form factor. Last but not least, system efficiency increases as the switching and conduction losses are reduced thanks to the fast switching time, respectively, the low on-state resistance.

Even though SiC is (much) more expensive at component level, system costs usually can be reduced. This however has to be thoroughly investigated case by case as SiC does not make sense for all common power electronics applications. Associations like Power America and European Center for Power Electronics (ECPE) have published WBG roadmaps indicating primary markets and applications for SiC-based power modules. It appears that photovoltaic inverters, uninterruptible power supplies (UPS) and inverters for electric vehicles could benefit from SiC in the very short term while high voltage applications may come later.

Packaging is a limiting factor

Significant efficiency gains as well as volume and weight reductions can be achieved at system level thanks to SiC devices. The prerequisite is that the integration of the devices into the system does not eliminate all those advantages. Packaging requires special care as it is the first step towards system integration.

As indicated before, switching and conduction losses per chip will be significantly reduced but the chip area will shrink even more. Ultimately, the loss power density will increase and the packaging has to be carefully selected to deal with this higher waste heat flux. In addition, the devices may be operated at higher junction temperatures and the temperature rise may increase. Requirements regarding temperature stability, cooling and reliability will be much tougher and the packaging materials used for die encapsulation, die attach, die interconnections, as die carriers and for waste heat dissipation will have to be selected accordingly. New materials for die encapsulation will likely be required to tackle the higher operating temperature. New technologies for interconnecting dies will replace the conventional heavy aluminum wires. Silver sintering of dies on silicon nitride active metal brazed (Si3N4 AMB) substrates appears to be the better choice today to cope with these cooling and reliability challenges. We can also expect some innovative solutions with thicker copper layers, fewer thermal barriers and integrated heat sinks to optimize the heat capacity, heat spreading and overall heat dissipation for the junction-to-coolant fluid distance.

Next to the thermal and reliability challenges, some electrical challenges will appear with fast switching SiC devices. A drop in current within a very short switching time at turn-off from conductive-to-blocking mode will generate a steep current slope and may cause significant overvoltage and oscillations. Slowing down the devices is not a smart solution. Such problems can be avoided thanks to a low inductive current flow within and in the vicinity of the power module. Short current loops, current flowing in opposite planes and multiple, symmetrical current paths are some basic guidelines to consider for the design of the power module, the DC busbars including capacitors and their connections to each other. Another problem is related to the capacitive coupling between the AC current and the ground. At very high switching speeds, this coupling becomes critical as it generates significant electromagnetic interferences. Here again a smart design of the power module can help to reduce this coupling effect as much as possible.

Finally, cost is probably the biggest challenge for a wide adoption of SiC devices. Though these devices are expensive, they can lead to significant system cost reductions. However, as the devices are driving the system costs, one should use as many devices as required but as few as possible. Therefore, thermal management solutions are required to reduce the chip area while maximizing the output power in a compact and lightweight package.

SiC devices will change the game as their packaging requires new materials, new joining techniques and new components. Innovation is the key to overcome the hurdles. As a manufacturer of metallized ceramic substrates and cooling solutions, Rogers PES team is available to help you as your development partner. Do you have any questions or require some support with the design and selection of a suitable substrate or cooler for your application? Please contact us for assistance.

Related Products:
curamik Ceramic Substrates

Tags:
Olivier's Twist Blog, Automotive & EV/HEV, General Industrial, Rail, Wind & Solar, Wired Infrastructure

Published on Mar 27, 2019

Return to Blog