High-performance energy-saving chips
Switching to an electrical future with 3300 volts
The energy transition is a central lever for achieving climate goals, in other words, to bring net CO2 emissions to zero. In order to do this, as an energy supplier, it is necessary to comprehensively replace fossil fuels in as many sectors as possible.
A piece of the puzzle for achieving this goal is modern semiconductor technology. Caspar Leendertz and Konrad Schrami from Infineon Technologies and Thomas Basler from the University of Technology Chemnitz have made a name for themselves in this field. The team succeeded in developing the first silicon carbide transistor in the world (to be more precise, a SiC-MOSFET) with copper contacting in the 3300 V voltage class [1]. The new power semiconductor modules are “smaller, more lightweight, more reliable, more high-performance and more efficient than their silicon predecessors,” according to a press release from the University of Technology Chemnitz [2]. They have the potential to save energy and to electrify new areas of infrastructure that were previously reliant on gasoline.
In recognition of their research, the team was nominated for The Federal President’s Award for Technology and Innovation in 2024. We will take a look at the groundbreaking work and its background.
Conductivity between a conductor and an insulator
Semiconductors are the basis of our modern electrical engineering. They act as controllable switches, can precisely control electrical current, convert alternating current into direct current and change voltages. A semiconductor’s electric conductivity is between that of a conductor and that of an insulator. Whereas conductors (e.g. metals or water) conduct electricity well and insulators (e.g. plastic or wood) conduct electricity poorly, a semiconductor’s conductivity depends on the prevailing conditions (e.g. temperature).
A well-known semiconductor material is monocrystalline silicon which is primarily used in solar cells and converts sunlight into electrical current with an efficiency of over 20 per cent in this application. In electric vehicles, silicon transistors control the battery and ensure safe and fast charging.
The components that contain the semiconductors are called converters. They work because of the semiconductor material’s electronic properties. A semiconductor is distinguished by its band gap, an energy barrier between the valence band and the conduction band. By influencing this band gap in a targeted manner (usually by doping the material with foreign atoms), the voltage required can be precisely adjusted.
ROTH explains for non-physicists
Current always flows in a conductor, but it never flows in a non-conductor. In a semiconductor, current sometimes flows under certain conditions. A semiconductor therefore works in a similar way to a switch that switches the current on and off.
You can switch on this switch in a targeted manner, in other words, you can make current flow in the semiconductor by putting in a certain amount of energy. For example, this can take the form of heat, light or current. This is how simple photovoltaic modules work – if the sun is shining on them, the pump will run.
The amount of energy required to switch on the semiconductor is specific to each individual semiconductor and is called the “band gap”. In this case, the “bands” are the two energy levels of the semiconductor, at which current flows (“on”) or no current flows (“off”). The band gap can be practically adjusted to a required amount of energy by mixing the semiconductor material (e.g. silicon) with a certain amount of a conductor (e.g. phophorus or copper) – “doping” the semiconductor.
In electric cars, converters control braking and accelerating processes by varying the current, voltage and frequency. As they switch thousands of times per second to do this and each switching operation creates losses in the form of waste heat, the module quickly heats up and must be laboriously cooled in order to remain operational [3]. This is where traditional silicon semiconductor chips reach their physical limits.
For further basics about semiconductors:
https://www.leistungselektronik.de/halbleiter-a-ae0f529fa95e12e1d345fd6913593c02
How much energy could the new chips save?
Semiconductors made of silicon carbide (SiC), a silicon carbon compound with improved thermal stability and a larger band gap, are much more robust. While the band gap for silicon is 1.1 electron-volts, it is 3.3 electron-volts for silicon carbide. This enables the SiC semiconductor to be used at higher operating voltages and temperatures. Leendertz, Schraml and Basler have used this to develop a semiconductor module that opens up new application areas for electrification. The researchers reduced the switching losses in their SiC module by 90 per cent compared with traditional silicon transistors [3]. This is sufficient for simple air cooling, rather than the water cooling otherwise required. This also means that the technology is practical for large industrial drives, such as in rail vehicles or heavy construction machinery that were previously dependent on fossil fuels. The newly developed SiC chips are ten times more reliable than traditional silicon modules [3].
With the newly developed power saving chip, a single electric locomotive can save approx. 300 MWh per year compared with the previous silicon-based solution. This is approximately equivalent to the annual energy demand of 100 single-family homes, the researchers calculate [2].
Table 1: Comparison of silicon and silicon carbide as semiconductor materials [4]
| Si | SiC | |
| Melting point | 1414 °C | approx. 2700 °C |
| Thermal conductivity | 1,5 to 1,7 W/(mK) | 3 to 4,9 W/(mK) |
| Band gap | 1,1 eV | 2,2 to 3,3 eV |
| Mohs hardness | 7 | 9 to 9,5 |
| Chemical stability | Is attacked by some strong oxidising agents and strong acids | Robust against acids, bases and solvents |
Amazing to use, very difficult to process
If changing the semiconductor from silicon to silicon carbide is so advantageous, why haven’t SiC-based converters been produced for a long time already?
In fact, SiC components have already existed for over 20 years. Infineon brought the first commercial SiC Schottkey diode onto the market in as early as 2001 [5]. Since then, researchers have continuously further developed the technology. However, silicon carbide also has disadvantages. Due to its higher Mohs hardness, which is close to values of diamond, it is extremely hard and brittle. This means that it is difficult to process. This drives up the manufacturing costs of SiC components. The high thermal conductivity and the high thermal expansion coefficient also prove to be problematic. The material stretches greatly when it is heated up, which the contact wires on the module need to compensate for. The aluminium wires otherwise used in Si modules can’t do this – also because of the high energy density – over the required service life of 35 years [3].
One solution in three steps
The newly developed SiC-based component is a metal-oxide-semiconductor field-effect transistor, also known as MOSFET with a vertical trench (trench MOSFET). Whereas, in planar MOSFETs, the current flow initially runs horizontally, trench MOSFETs offer purely vertical trenches. This results in a higher cell density per area, which, in turn, considerably reduces the losses in the chip during energy conversion, which increases the efficiency.
But how have the researchers solved the contact problem of the aluminium wires? Their approach was to use contact wires made of copper rather than aluminium. However, copper is now also used for doping silicon, occurs as its contamination and easily mixes with the material. This made it necessary to prevent copper atoms from penetrating into the semiconductor upon contact. To do this, Schraml and his colleagues developed a new bonding technology that prevents copper atoms from diffusing into the semiconductor material when in it comes into direct contact with the silicon carbide and from destroying the semiconductor material.
Schraml, the spokesperson for the research, summaries the team’s key successes, “in principle, we took three steps at once – designing a silicon-carbide-based ’high-voltage power transistor chip’, then taking the decision that we needed to switch from aluminium to copper due to the mechanical problems, and, finally, developing a bonding and the remaining module technology so that the copper doesn’t destroy the chip.” [3]
Digital light shines in 1st place at The Federal President’s Award for Technology and Innovation in 2024
For over 100 years, car headlights have lit up the streets. Nevertheless, this technology is still being developed. Three German researchers have now achieved a new milestone. Dr Norwin von Malm and Stefan Grötsch from ams-Osram and Dr Hermann Oppermann from the Fraunhofer Institute for Reliability and Microintegration IZM developed a highly efficient new headlight technology.
Their system works like a video projector with 25,600 individually controllable LEDs in a 320 x 80 matrix. Each LED can be switched on and off in a targeted manner. A corresponding control unit ensures that only those areas that need to be bright are lit – for example, ongoing vehicles remain dark and are not dazzled, while the driver can continue to drive with the high beam lights. This increases safety for everyone on the road.
What’s special about the new headlights is their high efficiency. Traditional systems first generate the full light and then filter out unwanted light, e.g. using blinds. The new system only switches on the required LEDs from the outset and therefore avoids wasting energy. This enables a particularly compact design without large cooling systems.
“The technology of the digital light solves the problem, can be defined as required and definable light distributions can be brought onto an LED chip using a digital signal,” summaries von Malm, the spokesperson for the research. In recognition of this achievement, von Malm and his research colleagues Grötsch and Oppermann were awarded the EUR 250,000 prize of The Federal President’s Award for Technology and Innovation in 2024.
Source: [6]
List of sources:
[1] Infineon: https://www.infineon.com/products/power/mosfet/silicon-carbide
[2] TU Chemnitz press release: Research team from Infineon and TU Chemnitz nominated for the Deutscher Zukunftspreis 2024: https://www.tu-chemnitz.de/tu/pressestelle/aktuell/12578
[3] German Future Prize: Team 3 2024: https://www.deutscher-zukunftspreis.de/de/team-3-2024
[4] SA Materials: Silicon Carbide vs Silicon – A Comparative Study: https://www.samaterials.de/silicon-carbide-vs-silicon-a-comparative-study-of-semiconductors-in-high-temperature-applications.html
[5] IEEE Xplore: https://ieeexplore.ieee.org/document/4633594
[6] German Future Prize: Digital Light – Team 1 2024: https://www.deutscher-zukunftspreis.de/de/team-1-2024