Replacement for the silicon semiconductor -opinion needed

Page 1 of 1 [ 9 posts ] 

KevinLA
Veteran
Veteran

User avatar

Joined: 6 Nov 2007
Age: 54
Gender: Male
Posts: 744
Location: United States

12 Nov 2013, 5:34 pm

I am investing in a company that claims to have found a replacement for the silicon semiconductor.

Within the next decade the silicon semiconductor will no long will be viable due to Moore's law.

Someone on a message board expressed skepticism about the company's technology. I emailed investor relations for the company about the skepticism and received an esoteric response.

I was hoping someone could tell me whether the response addressed the skeptical opinion.

Here are the claims by the company.

Quote:
The POET (Planar Opto Electronic Technology) platform is a patented semiconductor fabrication process that uses gallium arsenide technology to combine electronic and optical elements on a single integrated circuit.

Utilizing completely standard CMOS fabrication equipment and process steps POET and PET will allow semiconductor manufacturers to make microchips that are far faster and more energy efficient than current silicon devices, and far less expensive to produce.

Key benefits of the POET platform include:

100x speed improvement over CMOS silicon (silicon hits a “power wall” at about 4 GHz that has limited circuit speeds to about 3.2 GHz over the last 10 years)

10-100x power efficiency improvement over CMOS silicon (depending on application)

Flexible application that can be applied to virtually any technical application, including memory, digital/mobile, sensor/laser and electro-optical, among many others

No retrofit or other modifications to existing silicon fabs required – Since POET/PET are CMOS technologies fabricated using standard lithography techniques; they are easily integrated into current semiconductor production facilities extending the profitable utilization of fabrication equipment and production lines that would otherwise be considered at the end of life.


I found a person that was skeptical of the company's technology. He stated the following:

Quote:
​​The POET guys claim way more gain than is realistic, given that they're only dinking with the I/O. Given that many PCs run substantially faster with SSDs, the I/Os are still a 2nd or 3rd order impact on overall system speed. I can see that there might be limited scenarios where POET would make lots of sense, like in a packet switching context. Their claims only address power in the white paper, running about 1/10th the power at the same datarate, but nothing that suggests that they can run 150 Gb/s, which is their primary claim of 10x speed improvement.


Here is the white paper.


http://www.poet-technologies.com/wp-con ... rcuits.pdf

I sent an email to investor relations about the opinion and received a response that was esoteric to me.

Quote:
​​That’s POET’s secret sauce. The platform is a shift from conventional technology in that they have modified the N channel speeds via their technology. The method is proprietary of course and this is why POET is a shift in conventional technology. You have green technology in that POET operates at a lower power in addition to speed gains via the N channel.
[url=http://www.poet-technologies.com/wp-content/uploads/2013/07/Optical-Interconntion-of-High-Speed-Circuits.pdf"]http://www.poet-technologies.com/wp-content/uploads/2013/07/Optical-Interconntion-of-High-Speed-Circuits.pdf]White paper[/url][url=http://www.poet-technologies.com/wp-content/uploads/2013/07/Optical-Interconntion-of-High-Speed-Circuits.pdf"]http://www.poet-technologies.com/wp-content/uploads/2013/07/Optical-Interconntion-of-High-Speed-Circuits.pdf]White paper[/url][url=http://www.poet-technologies.com/wp-content/uploads/2013/07/Optical-Interconntion-of-High-Speed-Circuits.pdf"]http://www.poet-technologies.com/wp-content/uploads/2013/07/Optical-Interconntion-of-High-Speed-Circuits.pdf]Web Page Name[/url]



Last edited by KevinLA on 12 Nov 2013, 5:45 pm, edited 2 times in total.

Thelibrarian
Veteran
Veteran

User avatar

Joined: 5 Aug 2012
Age: 62
Gender: Male
Posts: 1,948
Location: Deep in the heart of Texas

12 Nov 2013, 5:41 pm

"Within the next decade the silicon semiconductor will no long will be viable due to Moore's law."

More properly, I think the above should read: Within the next decade, Moore Law will no longer be viable due to the silicon semiconductor. There are no guarantees that a successor will appear that is both practical and economically feasible. Therefore, betting that there will be a superior successor is just that--a bet.

Bottom line: As one who works hard for his money and doesn't like to gamble with it, I wouldn't touch an investment like this unless I was intimately familiar with this technology and this outfit's business plan.



cberg
Veteran
Veteran

User avatar

Joined: 31 Dec 2011
Gender: Male
Posts: 12,183
Location: A swiftly tilting planet

12 Nov 2013, 5:59 pm

Nothing has fully replaced it yet, but feasible alternatives are growing in terms of engineering applications. Optical memory & transistor gates are being comprehensively researched and quantum processors are already being tested in large organizations such as Google's. There is no gamble because semiconductors comprise the tools necessary to render themselves obsolete - of course, I could say the same about a lot of tech companies!


_________________
"Standing on a well-chilled cinder, we see the fading of the suns, and try to recall the vanished brilliance of the origin of the worlds."
-Georges Lemaitre
"I fly through hyperspace, in my green computer interface"
-Gem Tos :mrgreen:


1401b
Veteran
Veteran

User avatar

Joined: 21 May 2012
Age: 125
Gender: Male
Posts: 1,590

12 Nov 2013, 6:10 pm

The feel I get from their comments is that they're hyping hard.
Feels just like so many other hypes.


_________________
(14.01.b) cogito ergo sum confusus


drh1138
Velociraptor
Velociraptor

User avatar

Joined: 2 Dec 2012
Gender: Male
Posts: 498

12 Nov 2013, 6:41 pm

Rule of thumb: if you have to ask, don't risk it. Personally, as tech-literate as I generally am, I wouldn't put any money into these kinds of start-ups, but your mileage may vary.

Reading the opening paragraph to the whitepaper sets off my "BS alarm", but I tend to regard people who use buzzwords and pretentious, overcomplicated language like that as as*holes, due to personal experience.



Spudz76
Tufted Titmouse
Tufted Titmouse

User avatar

Joined: 18 Aug 2009
Age: 48
Gender: Male
Posts: 41

19 Nov 2013, 2:54 am

I agree, they seem to use far too many "hype words" and claim inordinate gains (likely theoretical maximum gains, which can not be fully realized). Also the quality of most silicon fabs is such that a wafer of cores may have a normal circuit error rate that is quite high. Imagine then using this same silicon fab technology with light or photons which are always more sensitive to imperfections than electrons. This is why an entire optical network can be knocked out simply by crushing the fiber optic cable causing internal diffusion and diffraction and creating an impassable signal-to-noise ratio. Take a copper wire cable and stomp on it all you like, it will still work fine (assuming you aren't a neutron star, or wearing ice skates).



Fogman
Veteran
Veteran

User avatar

Joined: 19 Jun 2005
Age: 58
Gender: Male
Posts: 3,986
Location: Frå Nord Dakota til Vermont

19 Nov 2013, 11:34 am

So.... whatever happened to using Gallium Arsenide as a Semiconductor?


_________________
When There's No There to get to, I'm so There!


ruveyn
Veteran
Veteran

User avatar

Joined: 21 Sep 2008
Age: 88
Gender: Male
Posts: 31,502
Location: New Jersey

19 Nov 2013, 11:46 am

Fogman wrote:
So.... whatever happened to using Gallium Arsenide as a Semiconductor?

We know from the basic physics that gallium arsenide is a semi-conductor. So the basic science is right. The question is what can these people produce and at which price?

ruveyn



Arran
Deinonychus
Deinonychus

User avatar

Joined: 20 Nov 2012
Gender: Male
Posts: 375

25 Nov 2013, 6:33 pm

GaAs is an expensive material originally intended for professional and military grade components. There has been a trend since the late 1990s to replace many GaAs MMICs by silicon RFICs in consumer and some professional products. A viable replacement for silicon semiconductors is not around the corner. Silicon has its problems but it's cheaper than other III-V compounds for most applications. Organic transistors are a possibility but can they compete effectively with silicon on size, frequency, lifespan, and production costs all at the same time?