4 ms·

If IBM can make a 150 GHz graphene transistor, why does Dr Phaedon Avouris, of IBM, say that it cannot replace silicon? Isn't clock frequency a measure of the t

If IBM can make a 150 GHz graphene transistor, why does Dr Phaedon Avouris, of IBM, say that it cannot replace silicon? Isn't clock frequency a measure of the transistor's ability to turn on and off quickly?

perlgeek 15y ago
It's not only important what it can do, but also if you can produce it in the way you want. Only time will show if dense packed transistors on a chip can be build cheap and reliably.
ori_b 15y ago
There's so much more than clock frequency. The place silicon shines isn't so much it's material properties for producing individual switches - they're good, but there's better out there. It's the raw manufacturability that silicon has. It's by far the easiest material to work with. It's abundant, cheap, and easy to obtain. It's got properties that lend it to easy junction doping and easy annealing. It's easy to deposit new layers on top of etched silicon. It's got a lattice that matches up nicely with the materials you'd want to use for contacts. In short, the advantages of silicon aren't in it's ability to make transistors. It's in it's ability to make transistors while on a high volume assembly line. It's going to be a while if we see whether graphene can match silicon for that.
- sliverstorm 15y ago
  Yup. It's always sobering to remember that there's already plenty of higher-performance materials than silicon- but silicon is so much cheaper to use, it simply outperforms everything else in performance-per-dollar by an incredible margin.
- redthrowaway 15y ago
  How much of the ease of using silicon is due to its inherent properties, and how much is due to the many billions that have been sunk into R&D?
  - Unseelie 15y ago
    I'm going to hazard that a lot is based on inherent properties. probably 'most' of the ease. They were mucking around with silicon computers with basically hobby chemistry sets back in the fifties and sixties, where today, with some of the best materials science we have, we're only able to manipulate graphene on the nanoscale..
baq 15y ago
Hz isn't the only important parameter of a transistor; amplification factor (or what's it called) is another, and IIRC graphene transistor didn't really perform in this area.
- hugh3 15y ago
  Not to mention size. How big is this graphene transistor? How densely can you pack 'em? How much energy does it use? And so forth. I've worked in the field of nanoscale electronic stuff, and I don't believe anything will replace silicon for a long time; at least not until we've completely run out of ways to do things better with silicon. It's silicon that's the true miracle material -- it's versatile and cheap and amazingly easy to build stuff out of. We've been talking about replacing silicon for a while, but we keep coming back to it. It's all about silicon for the next couple of decades at least.
Symmetry 15y ago
Making integrated devices out of graphene is problematic for all the reasons the other responses have pointed off, but it might prove a workable alternative eventually - well after Moore's law has run out in silicon due to atomic size issues.

perlgeek 15y ago

It's not only important what it can do, but also if you can produce it in the way you want. Only time will show if dense packed transistors on a chip can be build cheap and reliably.

ori_b 15y ago

There's so much more than clock frequency. The place silicon shines isn't so much it's material properties for producing individual switches - they're good, but there's better out there. It's the raw manufacturability that silicon has. It's by far the easiest material to work with. It's abundant, cheap, and easy to obtain. It's got properties that lend it to easy junction doping and easy annealing. It's easy to deposit new layers on top of etched silicon. It's got a lattice that matches up nicely with the materials you'd want to use for contacts. In short, the advantages of silicon aren't in it's ability to make transistors. It's in it's ability to make transistors while on a high volume assembly line. It's going to be a while if we see whether graphene can match silicon for that.

sliverstorm 15y ago
Yup. It's always sobering to remember that there's already plenty of higher-performance materials than silicon- but silicon is so much cheaper to use, it simply outperforms everything else in performance-per-dollar by an incredible margin.
redthrowaway 15y ago
How much of the ease of using silicon is due to its inherent properties, and how much is due to the many billions that have been sunk into R&D?
- Unseelie 15y ago
  I'm going to hazard that a lot is based on inherent properties. probably 'most' of the ease. They were mucking around with silicon computers with basically hobby chemistry sets back in the fifties and sixties, where today, with some of the best materials science we have, we're only able to manipulate graphene on the nanoscale..

baq 15y ago

Hz isn't the only important parameter of a transistor; amplification factor (or what's it called) is another, and IIRC graphene transistor didn't really perform in this area.

hugh3 15y ago
Not to mention size. How big is this graphene transistor? How densely can you pack 'em? How much energy does it use? And so forth. I've worked in the field of nanoscale electronic stuff, and I don't believe anything will replace silicon for a long time; at least not until we've completely run out of ways to do things better with silicon. It's silicon that's the true miracle material -- it's versatile and cheap and amazingly easy to build stuff out of. We've been talking about replacing silicon for a while, but we keep coming back to it. It's all about silicon for the next couple of decades at least.

Symmetry 15y ago

Making integrated devices out of graphene is problematic for all the reasons the other responses have pointed off, but it might prove a workable alternative eventually - well after Moore's law has run out in silicon due to atomic size issues.