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!
May be basic questions, i have some, pls share your views,
When the electronics industry will move to Graphene based ICs? Again is it going to be tough for small players to start the graphene based fabs? Will the history change? Is there any IC fabs sector which will take less money to startup the fab plant?
Graphene-based transistors are a wonderful research technology that hasn't quite escaped the laboratories and made it into high-volume production. There's an interesting article from January 2014 about IBM Research and graphene-based transistors.
The received wisdom (for example, talking to An Steegen of imec) is that we will push FinFET as far as we can, adding III/V materials, down to 7nm for sure. Then we need to go to gate-all-round silicon nanowires. Probably at 5nm. Graphene / Carbon nanowires may be next. The big challenge is how to manufacture them so that they are all semiconductors and are not destroyed by a tiny percentage of wires that are pure conductors.
...The big challenge is how to manufacture them so that they are all semiconductors and are not destroyed by a tiny percentage of wires that are pure conductors.
One could argue that's the problem with any really small devices, and that it'll be difficult to change to any new technology that has generally low yield in manufacturing or higher failure rates in use when we are used to putting down billions of transistors and having most of them work (for years). As such it's partly an ecosystem problem that requires more work on the design tool side, and creating software systems that are fault tolerant.
Noting that: if you can't dope the material the only devices you can build are (n-channel) depletion mode FETs, and for logic circuits you want N & P enhancement mode FETs so that you can store state with using much power (as in CMOS). I.e. electrons will normally whizz through the graphene, but you can stop them by applying a repulsive field (a negative gate voltage).
IBMs approach might be interesting in conjunction with Spintronics where you can use regular logic to steer where electrons go - separating data and control planes.
Yes, with only N-channel depletion mode FETs you are extremely limited to what you can design and build, because the power dissipation is just too great. Let's see if they can fabricate complementary, enhancement mode graphene transistors. Until such time, this is just a technical curiosity.
Well, coming back to the question, when Graphene based ICs will start and whether such fabs will be out of reach of small players?
Looking at it, it appears that the industry is geared to realize 10nm, 7nm, and then 5nm. Till then we do not hear Graphene based fab taking place. Upto 5nm, it may take until 2022? By then Graphene based chips do have potential to come out of labs. Since Graphene is capable of operating at ~500 GHz freq with bandgap suitable for analog ICs, I guess the analog ICs for RF applications could be well suited for such fabs. They may not cost huge, but given the sophistication required, should cost moderately. That's my guess, no data at this time.
Having been studying 2D materials with a materials engineer for over five years, I don't think the 2D material used will be grapheme, but will be a compound using 2D structure. I asked DR. Nalamasu the CTO/EVP of AMAT about 2D materials about 18 months ago and he said they were 5 to 6 years out. I think the first applications of 2D technology will be in some form or RRAM memory, that's my best guess. 2D materials will become pervasive for structure, electronics, filtration and many other areas. The first use of grapheme will probably be used in generating light as some processes have already been developed in this area. The production of grapheme is set to drop 99% as I have explored in previous posts. Do a search for grapheme.
Use implantation to do chemical synthesize? Ziegler's srim.org will bring some help. You know we need 100 times of current higher than high current implanter nowadays to reach the density. And the high dose bombardment is not going to amorphize the Ni/SiO2, (they are not crystal anyway) but to mess up everything, even sending some Ni downward to Si. Get a 2D Graphene by annealing? Thermodynamics helps to form ordered world, doesn't it? Really? On the long way of reaching 2D, we see huge achievement now. Great! Tomorrow we are sure that we will achieve more.
Yes, with only N-channel depletion mode FETs you are extremely limited to what you can design and build, because the power dissipation is just too great. Let's see if they can fabricate complementary, enhancement mode graphene transistors. Until such time, this is just a technical curiosity.
Use implantation to do chemical synthesize?