You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
Scientists have found that a “superatomic” material is the fastest and most efficient semiconductor ever. Taking advantage of a tortoise-and-hare mechanism, the new material can transport energy much faster than silicon.
newatlas.com
I have no experience in this area and any opinions or thoughts where this material may impact or not impact the semi world would be appreciated, Thanks.
There is always a new material with special properties. 30 years from now it might become 10% of wafers. The best way to evaluate is see what real companies are looking at in research. even then it is 10-20 year from product. I would recommend going to IEDM and asking experts there for pros and cons of each material (there are 10-20 that are discussed by real companies. Its one of my interview questions I do at IEDM for my Analyst reports
Scientists have found that a “superatomic” material is the fastest and most efficient semiconductor ever. Taking advantage of a tortoise-and-hare mechanism, the new material can transport energy much faster than silicon.
newatlas.com
I have no experience in this area and any opinions or thoughts where this material may impact or not impact the semi world would be appreciated, Thanks.
I have the same view as MKW. Every year or two researchers trumpet new semiconductor materials that offer greater capabilities than silicon in one or more dimensions (greater mobility, greater/smaller band gap, direct band gap) as well as device structures that offer substantial advantages over the current ones. But bulk silicon (and to some degree SOI) has economies of scale, commercialization and simplicity advantages that make it the natural for high complexity / density electronics. When I graduated college, GaAs and sapphire-based CCD memories and Josephson junction transistors were the new industry darlings. GaAs has much higher mobility / speed than silicon, but never reached the cost / complexity threshold to compete with silicon, but is used in tons super high speed applications (Microwave and RF amplifiers, plus optical, LED and photonics usages). CCDs migrated away from memories into image sensing. And Josephson junctions kind of disappeared. We’re seeing a bunch of new materials, like SiC that leverage their larger band gap that are great for controlling high voltages. But trusty silicon remains the mainstay for VLSI integration and memories.
JJs show up as possible support logic around quantum, for example perhaps running the error correction logic fast and close. But JJs are physically large, microns, and do not scale down last I looked at them, and they do not have the spooky advantage of quantum, so no answer to the density of silicon.
JJs show up as possible support logic around quantum, for example perhaps running the error correction logic fast and close. But JJs are physically large, microns, and do not scale down last I looked at them, and they do not have the spooky advantage of quantum, so no answer to the density of silicon.
SiC that leverage their larger band gap that are great for controlling high voltages. But trusty silicon remains the mainstay for VLSI integration and memories.
SiC will not be the mainstay in power electronics any time soon as top tier silicon devices are still comparable at lower cost, except extremities of voltages, and switching speeds.
Switching speeds for large power supplies may actually go down as more efficient topologies are mainstreamed, which exploit passives more efficiently.
SiC will not be the mainstay in power electronics any time soon as top tier silicon devices are still comparable at lower cost, except extremities of voltages, and switching speeds.
/PRNewswire/ -- There is no denying that silicon carbide (SiC) MOSFETs are the story of the next decade for electric vehicle power electronics. The material...
/PRNewswire/ -- There is no denying that silicon carbide (SiC) MOSFETs are the story of the next decade for electric vehicle power electronics. The material...
I am very sure under 2 tonn EVs will feel just fine with 400V, and it's not a problem at all to find 800V capable silicon inverter switches.
The cream of the crop traction inverters in fact intentionally go for slower switching speeds, and very accurate zero current switching.
The trick is to so use very cheap switches, but then drive them as gently as possible, and as much within their efficient region as possible.
BMW went for much higher currents, but they also got a motor with much thicker windings to squeeze as much torque per kg from it as possible.
Electric motor peak power can be much higher than IC engines of similar size, if you go above 10000 RPM, but to reduce their RPM you will need an equally large gearbox, which also costs money, and takes space, and mass.
Advanced silicon power electronics will be there for a very long time, because it was already very competitive long before EVs were a thing.
I expect EV makers to soon pass the wheel reinvention phase. Once they will learn all inverter design tricks, they will find out that they can use much cheaper switches without making compromises.
The same goes for photonics. LiNbO3 is a superior material compare to silicon，but silicon based photonics will be there for a very long time，due to the maturity and cost competitiveness of manufacturing process based on silicon.
newatlas.com
www.prnewswire.com
