Polymers will replace Silicon; Polican Valley

[elpres05]

Its now looking good for conductive polymers to replace current silicon based electronics.
Not only its said that polymers will outperform silicon transistors, they will be cheaper to manufacture and will be even more heat resistant. (good, more OC)

The author is VP EnerTech and former NASA contractor, so all that he writes is bound to be correct.

Here is the link to the article
polymerismismism

 

[NeoCorn]

Just did a quick read through of the article, "Conducting Polymers and the Evolving Electronics Technology". Appears, at least to me, to be a prediction about the future of conducting polymers though it does include a list of applications which is inferred to be current technology. Anyone heard of specific work with conducting polymers in CPU construction?

Bashir A. Syed mentions that conducting polymers are cheaper to manufacture than existing technology, but he does not offer an existing technology nor does he cite which type of organic conducting polymer he is speaking of. "Compared to other existing technologies, conducting polymers are lightweight, take up less space, and are less expensive to manufacture." Can it be that metal conductors could be replaced with conducting polymers in the power grid?

Statements like "Polymers are a shiny material derived from acetylene, whose electrical conductivity surpassed those of conductors." make me wonder if I'm experiencing a language gap problem, or an editing problem.

 

[CycloWizard]

Originally posted by: NeoCorn
Statements like "Polymers are a shiny material derived from acetylene, whose electrical conductivity surpassed those of conductors." make me wonder if I'm experiencing a language gap problem, or an editing problem.

I agree. As soon as I read this, I knew this guy either didn't know what he was talking about or is dumbing this way down to reach a wider audience.

In reality, silicon transistors are conducting polymers. They're simply not organic (carbon-based) polymers. Such polymers can be made from many, many compounds, not just acetylene. My guess is that acetylene-based polymers are what he's targeting as the future of processors, though I'm not sure why he doesn't come out and say this.

He is right about a few things, though. Polymer transistors would be much, much easier to manufacture than silicon. Further, the products would be much more customizable, as small changes in the processing or chemistry can yield dramatic changes. You can only change the composition of silicon so much and still get a useful product.

 

[Loki726]

I haven't read the article yet, but just off hand it seems that it would be difficult for a polymer to take the place of a semiconductor. I was under the impression that the useful properties of silicon (constant bandgap, reaction to dopants, constant carrier mobility) were brought about by its crystal structure.

Unless I am totally wrong, the multi-atomic molecular nature of polymers makes it nearly impossible to form a crystal lattice. Also, what would we use for dopants? Do phosphorus and boron bond with these polymers? Even if they do, it seems unlikely to me that they could bond in a crystal lattice...

 

[CycloWizard]

Originally posted by: Loki726
I haven't read the article yet, but just off hand it seems that it would be difficult for a polymer to take the place of a semiconductor. I was under the impression that the useful properties of silicon (constant bandgap, reaction to dopants, constant carrier mobility) were brought about by its crystal structure.

Unless I am totally wrong, the multi-atomic molecular nature of polymers makes it nearly impossible to form a crystal lattice. Also, what would we use for dopants? Do phosphorus and boron bond with these polymers? Even if they do, it seems unlikely to me that they could bond in a crystal lattice...

I'm not all that familiar with transistor manufacturing or why they work. However, I am familiar with polymers and their manufacture. Basically, you can make just about anything out of a polymer by adjusting the chemistry and processing. The electronic properties can be manipulated in this same manner and have been studied extensively since probably the 70's. Polymers can also form crystalline structures, networks, or just about anything else you would want.

 

[elpres05]

Originally posted by: CycloWizard

Originally posted by: Loki726
I haven't read the article yet, but just off hand it seems that it would be difficult for a polymer to take the place of a semiconductor. I was under the impression that the useful properties of silicon (constant bandgap, reaction to dopants, constant carrier mobility) were brought about by its crystal structure.

Unless I am totally wrong, the multi-atomic molecular nature of polymers makes it nearly impossible to form a crystal lattice. Also, what would we use for dopants? Do phosphorus and boron bond with these polymers? Even if they do, it seems unlikely to me that they could bond in a crystal lattice...

I'm not all that familiar with transistor manufacturing or why they work. However, I am familiar with polymers and their manufacture. Basically, you can make just about anything out of a polymer by adjusting the chemistry and processing. The electronic properties can be manipulated in this same manner and have been studied extensively since probably the 70's. Polymers can also form crystalline structures, networks, or just about anything else you would want.

Also, i think the electric conduction will have something to do with the double or tripple bonds. I don't remember who told me this but it makes some sense.

 

[CycloWizard]

Originally posted by: elpres05
Also, i think the electric conduction will have something to do with the double or tripple bonds. I don't remember who told me this but it makes some sense.

Right. The double and triple bonds are pi bonds, which allow the flow of electrons much more easily.

 

[elpres05]

Originally posted by: CycloWizard

Originally posted by: elpres05
Also, i think the electric conduction will have something to do with the double or tripple bonds. I don't remember who told me this but it makes some sense.

Right. The double and triple bonds are pi bonds, which allow the flow of electrons much more easily.

So if polymers are cheaper to manufacturer and more versatile, shouldn't they replace silicon based electronics?

 

[CycloWizard]

Originally posted by: elpres05
So if polymers are cheaper to manufacturer and more versatile, shouldn't they replace silicon based electronics?

Yes. However, first we need to make some that have the right properties.

 

[senseamp]

I'd like to see these polymers handle 100W/cm2 power density.

 

[pm]

From papers at conferences that I have read, and a fairly recent IEEE Spectrum article, all organic polymer-based circuitry that I have seen is vastly slower than silicon-based. Electron mobilities in doped silicon are around 1400cm2(V-S), electron mobilities in n-type organic polymer transistors are 0.07 cm2/Vs. I read through the article and never saw the author claim otherwise.

They have plenty of likely applications, and I see a very bright and profitable future for the organic polymer industry - particularly in the areas of RFID tags and displays - but I have a really hard time imagining them replacing silicon microprocessors barring a really massive breakthrough that makes them 10,000 times faster than they are.

 

[senseamp]

BTW, there is going to be a session on organic devices and circuits at the ISSCC on tuesday, mostly having to do with RFID.

 

[pm]

Originally posted by: senseamp
BTW, there is going to be a session on organic devices and circuits at the ISSCC on tuesday, mostly having to do with RFID.

Bummer. I decided not to attend ISSCC this year. I've gone most years - and presented last year - but this year I have my eye set on a conference later in the year and decided to forego ISSCC in hopes that I can make the other.

If the session is any good, please report back with anything interesting. The proceedings won't be out on IEEE Xplore for a while if past history is any judge.

 

[senseamp]

Originally posted by: pm

Originally posted by: senseamp
BTW, there is going to be a session on organic devices and circuits at the ISSCC on tuesday, mostly having to do with RFID.

Bummer. I decided not to attend ISSCC this year. I've gone most years - and presented last year - but this year I have my eye set on a conference later in the year and decided to forego ISSCC in hopes that I can make the other.

If the session is any good, please report back with anything interesting. The proceedings won't be out on IEEE Xplore for a while if past history is any judge.

You aint missing much, in terms of circuit design. I went today and I don't have much interest in going back. It seems like everyone is busy undoing all the advancements in circuit design over the last decade, because they either don't scale with technology, are too power hungry, or too complicated. So basically nothing to look at, aside from a few obvious power saving schemes here and there. Looks like LVS lasted for one design node at Intel, and now it's back to old school 2 phase domino, which was probably the most interesting paper, to give you an idea. Wouldn't be surprised at all if everything is static standard cell once it's at 45nm, and then there won't be much for circuit designers to talk about. I like static design, but it's mostly logic and architecture, not much circuit wizardry. And considering how much admission is, you don't get much out of it.