A One Hundred Thousand-Fold Enhancement In the Nonlinearity of Silicon (phys.org) 28
An anonymous reader quotes a report from Phys.Org: A team of researchers led by Osaka University and National Taiwan University created a system of nanoscale silicon resonators that can act as logic gates for light pulses. ... [The scientists] have increased the nonlinearity of silicon 100,000 times by creating a nano-optical resonator, so that all-optical switches can be operated using a continuous low-power laser. They accomplished this by fabricating tiny resonators from blocks of silicon less than 200 nm in size. Laser light with a wavelength of 592 nm can become trapped inside and rapidly heat the blocks, based on the principle of Mie resonance. "A Mie resonance occurs when the size of a nanoparticle matches a multiple of the light wavelength," author Yusuke Nagasaki says.
With a nanoblock in a thermo-optically induced hot state, a second laser pulse at 543 nm can pass with almost no scattering, which is not the case when first laser is off. The block can cool with relaxation times measured in nanoseconds. This large and fast nonlinearity leads to potential applications for GHz all-optical control at the nanoscale. "Silicon is expected to remain the material of choice for optical integrated circuits and optical devices," senior author Junichi Takahara says. The current work allows for optical switches that take up much less space than previous attempts. This advance opens the way for direct on-chip integration as well as super-resolution imaging. The study has been published in the journal Nature Communications.
With a nanoblock in a thermo-optically induced hot state, a second laser pulse at 543 nm can pass with almost no scattering, which is not the case when first laser is off. The block can cool with relaxation times measured in nanoseconds. This large and fast nonlinearity leads to potential applications for GHz all-optical control at the nanoscale. "Silicon is expected to remain the material of choice for optical integrated circuits and optical devices," senior author Junichi Takahara says. The current work allows for optical switches that take up much less space than previous attempts. This advance opens the way for direct on-chip integration as well as super-resolution imaging. The study has been published in the journal Nature Communications.
Re: (Score:2, Flamebait)
Oh, that's right. You CAN'T. Because you are too STUPID.
Who cares if BeauHD understands this stuff or not. It's likely that a few Slashdotters do, and it just might lead to an interesting discussion.
I haven't a single clue about most of the things mentioned in TFS, but I do find discussions on these topics fascinating nonetheless.
Too bad you've got nothing of value to contribute other than a tired troll putting down the editors.
what's the operating temperature range? (Score:1)
Just how how narrow is the operating temperature range for these devices? Will it remain practical outside laboratory-controlled environments?
Re: (Score:2)
Re: (Score:1)
Boil this down a bit (Score:3)
My read is that they've invented a effective optical diode in silicon. Is that essentially correct?
Re:Boil this down a bit (Score:4, Insightful)
Re: (Score:2)
My read is that they've invented a effective optical diode in silicon.
More like an optical transistor, since it can function as a switch. But unlike a transistor, I don't see any mechanism for signal amplification.
An application for these devices may be in fiber optic switches and routers.
Re: (Score:2)
More like an optical transistor, since it can function as a switch.
Diodes function as switches.
But unlike a transistor, I don't see any mechanism for signal amplification.
Likewise, diodes are poor amplifiers.
Re: (Score:2)
Diodes function as switches.
Diodes are rectifiers, not switches.
To build logic circuits from diodes you also need either resistors or inverters.
An inverter is usually a transistor. A resistor isn't going to work at nanoscale.
The logic gate described in TFA is not a diode.
Re: Boil this down a bit (Score:2)
Re: (Score:2)
Diodes are rectifiers, not switches.
One function of diodes is to rectify, and when they are used as such they are called rectifiers. Diodes are also used as variable capacitors. Diodes are also used as switches. Diodes are also used to regulate voltage. These various functions are all distinct. It is not true to say that diodes are simply rectifiers.
Re: (Score:2)
That is my reading too, though I had to skim the article to make head or tails of anything. I and I still feel they were bullshitting too.
I believe what they mean by "nonlinearity" in this piece is "the magnitude of the nonlinear part of certain responses". Which is a horrible misuse of language even for a material physics subfield.
Re:Boil this down a bit (Score:4, Informative)
Transistors, diodes, etc., are referred to as "non linear devices" as their response is not a straight line through the active region of operation.
For example, a diode that's forward biased conducts no current until a threshold is reached, then it conducts a lot of current. Likewise, when reverse biased, you will see this behavior as well (also known as zener, avalanche or breakdown region). This is a non-linear V-I (voltage-current) curve, where the amount of current the device will carry is non-linear with respect to voltage.
Increasing non-linearity is a good thing for digital circuits - it means the difference between conduction and non-conduction is much sharper and allows for much faster devices. For this, it means when used as an optical interconnect, the diode is biased to below conduction. Then when light hits it, it will briefly conduct. The point is not amplification, but detection.
Re:10nm 592nm (Score:5, Insightful)
Right... because you know, computers before 1990 weren't useful in the slightest.
Despite computers at the time running at only a handful of MHz, and having fabrication technologies that numbered in the multiple microns, and the speculation which would hold for many years even since then that Moore's law might end at around the 0.1 micron point due to em interference between adjacent electrically conductive traces on the silicon. Even back then, optical computing was hypothesized as a possible workaround for this limitation as optical traces could in theory be arbitrarily close, or even literally intersect, and they would not interfere with eachother's operation.
The question you asked is an entirely reasonable one, but then you decide to make it rhetorical and answer it with your own assumption and draw conclusions from that,
Maybe it's a dead-end technology and this particular implementation will not be scalable, but absent any known physical laws which would literally preclude faster switching speeds or smaller fabrication scales, it seems to me that your conclusion is not logically valid.
I think that you need to stop and realize that "we do not know", or "we haven''t done it before" is not the same thing as probably not.
Re: 10nm 592nm (Score:3)
That 592nm is wave length, not amplitude or the node size. In the summary they said the component to be sub-200nm, which is in the ballpark of modern silicon chips. Naturally things need to improve, but we donâ(TM)t know exactly where this could lead to.
Re: 10nm 592nm (Score:3)
Considering that we already have silicon at 10nm
You have to be careful with this kind of claim, it doesn't really mean much anymore. IIRC it used to refer to the gate length of a transistor but it's turned into an essentially meaningless marketing term these days.
Energy/heat (Score:2)
Re: Energy/heat (Score:2)
Yes, this has been a big concern in the past.
Fast optical switching in silicon (Score:1)
Not mentioning "photothermal"... (Score:2)