The Silicon Carbide Race Begins(semiengineering.com)
semiengineering.com
The Silicon Carbide Race Begins
https://semiengineering.com/the-silicon-carbide-race-begins/
15 comments
Any compound semiconductor is much more difficult to purify than a semiconductor made of a single element, like silicon.
That is not a problem specific to silicon carbide.
The main reason why a long time was needed for developing methods to make silicon carbide crystals with few defects and methods for doping them, is that the methods that worked with silicon and with many other semiconductors would require extremely high temperatures to work with silicon carbide.
So either alternative methods using more manageable temperatures or techniques to handle higher temperatures had to be found.
The need for higher temperatures also delayed for some years the use of silicon compared to germanium, but the problems were much easier solved, because the temperatures needed for silicon processing, e.g. in epitaxy or diffusion furnaces are still relatively low, less than 1400 Celsius degrees.
The second problem of silicon carbide is shared with all the high-bandgap semiconductors, e.g. also with gallium nitride.
For silicon and other low-bandgap semiconductors there is a very wide choice of dopants that can convert the intrinsic semiconductor into either P or N semiconductor of any desired resistivity in a wide range.
For silicon carbide and the other high-bandgap semiconductors, most dopants are very inefficient, e.g. only a very small fraction of the dopant might be active, especially at low temperatures.
So it can be very difficult to make a doped semiconductor with low resistivities and it can also be very difficult to find a metal that will make good contact with the device, without exhibiting rectifying characteristics (like a diode).
Solving all of these material problems required a large number of costly experiments with various chemical elements so it is normal that commercial devices have appeared only after many years of research.
Moreover, an additional reason for the very long time-to-market for high-bandgap semiconductor devices is that nowadays the research is done with much higher secrecy than in the early years of the electronics industry, when the Silicon Valley was created.
In the beginning, essential patents, like the patent of the transistor, were licensed without restrictions to anyone who desired for small royalties. Now most patents do not have the purpose to gain money by licensing but to prevent the appearance of any competitors, so they either are not licensed at all or too high royalties are requested.
In the early years, until around 1970, all the results of semiconductor research were published, including very detailed recipes with what was tried in the lab. After reading a research article, you could go in your lab and try the same recipe and reproduce the published research.
So any innovation made by some company spread immediately to all the companies and the technology of germanium, then that of silicon advanced very fast.
Now, the research articles only seldom are detailed enough. They usually just report some successes without mentioning the more important parts, i.e. what was tried and failed and what is still wrong with their current approach and remains to be solved in order to be able to make usable devices.
So most companies had to duplicate all the research done by the others to be able to reach to the stage of being able to make commercial devices.
Few large electronics companies have spent money on the research for high-bandgap semiconductors, because it was a long-term project.
The research was done mostly by smaller companies dedicated to this, but most of those which have been successful have been bought by now by the big players.
That is not a problem specific to silicon carbide.
The main reason why a long time was needed for developing methods to make silicon carbide crystals with few defects and methods for doping them, is that the methods that worked with silicon and with many other semiconductors would require extremely high temperatures to work with silicon carbide.
So either alternative methods using more manageable temperatures or techniques to handle higher temperatures had to be found.
The need for higher temperatures also delayed for some years the use of silicon compared to germanium, but the problems were much easier solved, because the temperatures needed for silicon processing, e.g. in epitaxy or diffusion furnaces are still relatively low, less than 1400 Celsius degrees.
The second problem of silicon carbide is shared with all the high-bandgap semiconductors, e.g. also with gallium nitride.
For silicon and other low-bandgap semiconductors there is a very wide choice of dopants that can convert the intrinsic semiconductor into either P or N semiconductor of any desired resistivity in a wide range.
For silicon carbide and the other high-bandgap semiconductors, most dopants are very inefficient, e.g. only a very small fraction of the dopant might be active, especially at low temperatures.
So it can be very difficult to make a doped semiconductor with low resistivities and it can also be very difficult to find a metal that will make good contact with the device, without exhibiting rectifying characteristics (like a diode).
Solving all of these material problems required a large number of costly experiments with various chemical elements so it is normal that commercial devices have appeared only after many years of research.
Moreover, an additional reason for the very long time-to-market for high-bandgap semiconductor devices is that nowadays the research is done with much higher secrecy than in the early years of the electronics industry, when the Silicon Valley was created.
In the beginning, essential patents, like the patent of the transistor, were licensed without restrictions to anyone who desired for small royalties. Now most patents do not have the purpose to gain money by licensing but to prevent the appearance of any competitors, so they either are not licensed at all or too high royalties are requested.
In the early years, until around 1970, all the results of semiconductor research were published, including very detailed recipes with what was tried in the lab. After reading a research article, you could go in your lab and try the same recipe and reproduce the published research.
So any innovation made by some company spread immediately to all the companies and the technology of germanium, then that of silicon advanced very fast.
Now, the research articles only seldom are detailed enough. They usually just report some successes without mentioning the more important parts, i.e. what was tried and failed and what is still wrong with their current approach and remains to be solved in order to be able to make usable devices.
So most companies had to duplicate all the research done by the others to be able to reach to the stage of being able to make commercial devices.
Few large electronics companies have spent money on the research for high-bandgap semiconductors, because it was a long-term project.
The research was done mostly by smaller companies dedicated to this, but most of those which have been successful have been bought by now by the big players.
Moreover, an additional reason for the very long time-to-market for high-bandgap semiconductor devices is that nowadays the research is done with much higher secrecy than in the early years of the electronics industry, when the Silicon Valley was created. In the beginning, essential patents, like the patent of the transistor, were licensed without restrictions to anyone who desired for small royalties.
Another price we pay for weak patents. Now everything new is secret.
Another price we pay for weak patents. Now everything new is secret.
I'm not sure why do you think weak patents are the cause, or what do you exactly mean by weak patents. Especially considering mentioned patents were licensed cheaply.
Seems to me, people's behavior has become much more materialistic / self-centered / greedy, along with the means and societal approval to keep it that way. Most companies will lock and not-share everything they can in almost any way they can.
Seems to me, people's behavior has become much more materialistic / self-centered / greedy, along with the means and societal approval to keep it that way. Most companies will lock and not-share everything they can in almost any way they can.
And this drives demand for new testing gear. My retirement fund thanks you AEHR!
Gamechanger in automotive? “These compelling benefits are leading to mass SiC adoption in BEVs, which brings SiC manufacturing cost reductions due to economies of scale”
Tesla leading the way. https://www.systemplus.fr/wp-content/uploads/2018/06/SP18413...
No. For 800+ volt EV platforms Lucid, Hyundai, and Volkswagen are leading the way. Tesla is still on a 400 volt platform.
For people (like me) who don't appreciate the advantages of 800v vs 400v, here's an article about why this is useful:
https://eandt.theiet.org/content/articles/2021/02/800v-syste...
> Such systems enable greatly reduced charging times, as long as they are using fast chargers capable of working at up to 270kW.
> It also allows for weight reduction, as less copper is needed in the electric system. Electric motors are much simpler than combustion engines in construction and at their core they have a rotor, which is driven by a changing magnetic field. To induce this field, electrical systems often use up to four times the amount of copper found in combustion engines; higher voltages reduces the need for large amounts of copper. This permits smaller motors, freeing space in the vehicle for batteries.
https://eandt.theiet.org/content/articles/2021/02/800v-syste...
> Such systems enable greatly reduced charging times, as long as they are using fast chargers capable of working at up to 270kW.
> It also allows for weight reduction, as less copper is needed in the electric system. Electric motors are much simpler than combustion engines in construction and at their core they have a rotor, which is driven by a changing magnetic field. To induce this field, electrical systems often use up to four times the amount of copper found in combustion engines; higher voltages reduces the need for large amounts of copper. This permits smaller motors, freeing space in the vehicle for batteries.
From your linked article:
> If the charger provides 800V and a minimum of 300A, the Taycan can charge from five to 80 per cent in 22.5 minutes.
Wow! That's an insanely huge amount of current. Obviously the cables would have to be very thick to reduce transmission losses, but then you'd also have to have sufficient insulation/protection at 800V!
> If the charger provides 800V and a minimum of 300A, the Taycan can charge from five to 80 per cent in 22.5 minutes.
Wow! That's an insanely huge amount of current. Obviously the cables would have to be very thick to reduce transmission losses, but then you'd also have to have sufficient insulation/protection at 800V!
The fastest CCS chargers you'll typically find are 350 kW. They use liquid cooled cables. Ionity's chargers are 920 volts and 500 amps but deliver a maximum of 350 kW:
https://ionity.eu/en/design-and-tech.html
Some CCS chargers are 400 kW:
https://insideevs.com/news/375020/repsol-most-powerful-charg...
Cars like the Lucid Air, Porsche Taycan, Porsche Taycan Cross Turismo, Audi e-tron GT, Hyundai Ioniq 5, and the Kia EV6 can take advantage of the higher voltage to get high peak charging rates.
The Lucid Air's peak charging rate is supposed to be above 300 kW. It'll be interesting to see what the full charge curve is like for the Air.
https://ionity.eu/en/design-and-tech.html
Some CCS chargers are 400 kW:
https://insideevs.com/news/375020/repsol-most-powerful-charg...
Cars like the Lucid Air, Porsche Taycan, Porsche Taycan Cross Turismo, Audi e-tron GT, Hyundai Ioniq 5, and the Kia EV6 can take advantage of the higher voltage to get high peak charging rates.
The Lucid Air's peak charging rate is supposed to be above 300 kW. It'll be interesting to see what the full charge curve is like for the Air.
Chademo 3.0 is 0.75-1MW
The CCS Megawatt Charging System is up to 3.75 megawatts:
https://www.charin.global/technology/mcs/
https://www.charin.global/technology/mcs/
Rivian will have an 800+V platform in a few years, too.
Rivian will have an 800V platform in a few years, too.
It took close to 30 years to get commercially relevant SiC crystals in quantity.
SiC has a much higher affinity for impurities than pure silicon.
Another example of a discipline within semiconductor research when just takes a lot of physics PhDs, and engineers to keep banging their heads against the wall for decades until the problem yields.