Print 36 comment(s) - last by Major HooHaa.. on Oct 28 at 11:43 AM

The flying lens uses light emitted from plasmons. While this sounds complex, it basically operates sort of like a record arm to make ultra tiny circuits. The end result is circuits possibly as small as 5-nm and a possible replacement for Blu-Ray.  (Source: Liang Pan and Cheng Sun, UC Berkeley)

Patterns on a 4x4 array of lenses created these tiny features, with features 100-nm wide.  (Source: Xiang Zhang Lab, UC Berkeley)
New research from Berkeley could allow circuit shrinks to 5 nm

With the never-ending quest for greater processor power, the hardware industry's sharpest minds and biggest companies are pouring money, time, and effort at the challenge of extending the life of Moore's Law.  Moore's Law, which states that the maximum number of transistors on a given chip area doubles every one and a half years, is close to reaching its limit due to the problems with controlling light at ultra low nanometer resolution.  While Intel, AMD, IBM, and others race to 32 nm and beyond, this wall looms ahead.

Now new research from the University of California Berkeley could buy Moore's Law some more time, and could pave the way for the next generation of ultra-tiny transistors.  It could also pave the way for an optical drive replacement for Blu-ray.  The research was led by Xiang Zhang, UC Berkeley professor of mechanical engineering and David Bogy, UC Berkeley professor of mechanical engineering. Its new approach uses a metal arm similar to that on a record turntable or in a hard drive.  It also utilizes a tiny lens that literally flies above the surface of the chip wafer. 

With the new design, the chip makers were able to form chip designs 80 nm wide, which could easily be made much smaller.  Better yet, the wafer was spun at a rate of 12 meters per second, meaning that production would be very fast.  Professor Zhang explains an overview of the process, stating, "Utilizing this plasmonic nanolithography, we will be able to make current microprocessors more than 10 times smaller, but far more powerful.  This technology could also lead to ultra-high density disks that can hold 10 to 100 times more data than disks today."

The basic concept of photolithography is similar to film development for photography -- light triggers a reaction in a chemical layer.  In photolithography this reaction is typically a hardening.  Caustic baths such as acid can then wash away areas of semiconductor that are not protected by the hardened photolithographic mask.  Through multiple steps a circuit is built, which can contain various metal and semiconductor components.

"With optical lithography, or photolithography, you can instantly project a complex circuit design onto a silicon wafer.  However, the resolution possible with this technique is limited by the fundamental nature of light,” explained Liang Pan, a UC Berkeley graduate student working with Zhang and Bogy, and one of three co-lead authors on the associated paper. “To get a smaller feature size, you must use shorter and shorter light wavelengths, which dramatically increases the cost of manufacturing. Also, light has a diffraction limit restricting how small it can be focused. Currently, the minimum feature size with conventional photolithography is about 35 nanometers, but our technique is capable of a much higher resolution at a relatively low cost."

The new method uses a phenomenon where metal electrons vibrate when exposed to light.  These tiny vibrations are smaller than light's normal wavelength and are known as evanescent waves.  By exploiting these, light can be concentrated into patterns theoretically as small as 5 to 10 nanometers.  The test lens was 100 nm, and used a silver plasmonic lens with a concentric ring pattern.

During photolithography the lens flies above the surface of the substrate.  Similar to flying lenses developed by UC Berkeley's Computer Mechanics Laboratory, the flying lens contacts the surface similar to the needle on a record player arm.  However, unlike a record player the lens is not physically touching -- it uses light instead to "touch" the surface.  The tiny lens uses the same lift mechanics used by aircraft, except on a microscopic scale to stay at a constant 20 nm above the surface.

Professor Zhang added,”The speed and distances we're talking about here are equivalent to a Boeing 747 flying 2 millimeters above the ground.  Moreover, this distance is kept constant, even when the surface is not perfectly flat."

The scan speed is very fast -- approximately 4 to 12 m/s in the testing.  Further, up to 100,000 tiny patterned lenses could be packed into the arm, allowing many parallel writes.

Best of all the new process is extremely cheap compared to conventional methods.  Technologies such as 45-nm lithography equipment are extremely expensive due to complex lens and mirror setups needed to concentrate the light.  With the new method less expensive slightly larger wavelength UV light can be used to excite the plasmonic lens.  With the only price component being the plasmonic lens setup, costs for shrinks would drop dramatically.

Professor Zhang is hoping the technology is put to use in commercial scale production within 3 to 5 years.  He states, "I expect in three to five years we could see industrial implementation of this technology.  This could be used in microelectronics manufacturing or for optical data storage and provide resolution that is 10 to 20 times higher than current Blu-ray technology."

The research will be reported in the December issue of Nature Nanotechnology.

Coauthors on the study are Werayut Srituravanich, a former Ph.D. student in Zhang's lab and currently a lecturer in mechanical engineering at Chulalongkorn University in Thailand, and Yuan Wang, a UC Berkeley graduate student in mechanical engineering, and Cheng Sun, a former graduate student in Zhang's lab and currently an assistant professor in mechanical engineering at Northwestern University.

The research was funded in part by a National Science Foundation Center for Scalable and Integrated Nano-Manufacturing grant.

Comments     Threshold

This article is over a month old, voting and posting comments is disabled

Research articles
By VaultDweller on 10/24/2008 8:02:09 AM , Rating: 2
I used to be very interested in articles about research breakthroughs, until I realized that the typical time between research being published and research being implemented far exceeds the length of time I can remember said articles.

So, anyone want to take bets on whether this technology will be put to practical use in the 3-5 year time frame that Zhang predicts?

RE: Research articles
By dflynchimp on 10/24/2008 8:31:00 AM , Rating: 3
haha, my stance is the same as yours. We're spectators, and can't really do sh!t with our enthusiasm. My take on this tech is that it'll be here when it gets here. Until then, I'm perfectly fine using what we already have. I just need to tell myself that, hey, 50 years ago they didn't even have personal computers, and I'm a much happier man XD.

RE: Research articles
By AlexWade on 10/24/2008 10:16:15 AM , Rating: 2
There is one key element that takes a product from design to production: money. If Sony makes this breakthrough, they have the money to promote it, to push it, and to make it. Universities don't have the funds to push a product. You will have to convince a company to actually produce and support a new technology, otherwise it goes nowhere. (The exception being if the military uses the technology.)

In this case, suppose I go to Sony and say "HEY! We developed this new technology to replace Blu-Ray! And we'll let you use it for royalty payments." Sony is going to shoot them down because Sony and other have millions invested in Blu-Ray and wanted the return on investment of royalties in Blu-Ray sales. When Blu-Ray has excellent market penetration and the cost of the players are cheap, then Sony will be willing to listen. But how long will that be? By then, something better may be along making this an afterthought.

RE: Research articles
By arrowspark on 10/24/2008 10:27:07 AM , Rating: 2
Money is not an issue for University research. Research at Universities is always funded. Since they are talking about lithography I'd imagine a semiconductor manufacturer is funding that research. Universities rarely "push products" based on their research findings. They transfer the knowledge to the research sponsors and the sponsors take it from there.

RE: Research articles
By theflux on 10/24/2008 11:24:19 AM , Rating: 2
My thoughts exactly. I used to enjoy Popular Science as a lad, but when I realized that nothing they ever talked about made it out of the lab in anything less than double the time frame they mention I kind of lost interest. "Breakthroughs" like this are always followed by "3 to 5 years", which in reality is much longer, if ever.

RE: Research articles
By Diesel Donkey on 10/24/2008 9:59:57 PM , Rating: 3
That's why it's called popular science as opposed to popular engineering.

RE: Research articles
By TSS on 10/25/2008 7:34:22 AM , Rating: 3
i'm still interested in knowing what they've been able to do now, but i don't mind the timeframe at all.

even if it takes much longer, if it takes until 2021 to start producing 10 nm chips with this technology, that means we went from 10 micrometer on the 4004 to 10 nanometer in 50 years. which basicly means within a human lifetime we've managed to make something that's nigh impossible to see with the human eye, 1000 times smaller.

research/implementation right now is much much faster then any point in human history, and probably as fast as it'll ever get while money is still used on this planet. i'd say count your blessings and marvel :P

RE: Research articles
By Subzero0000 on 10/27/2008 2:11:53 AM , Rating: 2
"It will be built about 50 years after everyone stops laughing."

RE: Research articles
By Major HooHaa on 10/28/2008 11:43:36 AM , Rating: 2
3 to 5 years? That's okay then, I haven't even up-graded to Blue Ray yet and they are talking about its replacement.

evanescent waves
By flipsu5 on 10/24/2008 8:31:12 AM , Rating: 2
The article neglected to say the evanescent waves decay very fast away from the surface - within 100 nm. That means the lens must be held very close to the sample. Familiar? That is how X-ray lithography (proximity printing) was supposed to work. Except the gap is here even closer (X-ray proximity was on the order of microns).

RE: evanescent waves
By JediJeb on 10/24/2008 2:17:33 PM , Rating: 1
Actually microns are larger than nm, 1000nm = 1 micron or micrometer.

RE: evanescent waves
By geddarkstorm on 10/24/2008 6:07:01 PM , Rating: 3
There's also no lenses for x-rays, so they can't be focused or directed very easily. Instead, by using simple metal atom vibration wave lengths rather than light, these guys can theoretically make incredibly small, tens of angstroms, circuits; as they and the article says. Not going to be able to do that with unfocusable x-rays unless you have pinhole apertures cut to be only 5nms in diameter and then an incredibly powerful x-ray source to allow enough incident light to get through that small opening and still have the energy to harden the mask. That means massive costs for production, and added complexity, which this method bypasses.

RE: evanescent waves
By flipsu5 on 10/25/2008 11:53:09 PM , Rating: 2
Yeah I don't favor X-rays either, but the plasmonic resolution does depend strongly on the gap. Where are the pictures confirming their 5-10 nm theoretical resolution? It is quite hard to do. The photoresist used with their UV also may not support their resolution target.

I'd suggest them to first demonstrate the technology for high-speed microscopy (SNOM).

RE: evanescent waves
By flipsu5 on 10/25/2008 11:31:17 PM , Rating: 2

This physorg article has more data on the distance between the lens and the surface to be patterned. Since they are so close, there will be a strong constraint to keep the surface free from particles even 20 nm high.

X-ray proximity already couldn't tolerate microns...

RE: evanescent waves
By flipsu5 on 10/25/2008 11:56:00 PM , Rating: 2
Since they are so close, there will be a strong constraint to keep the surface free from particles even 20 nm high.

Actually, we don't want particles on the surface, period. But the 20 nm particle would not destroy a conventional photolithography lens (which is millimeters away), while it will certainly scratch the silver plasmonic lens.

Flying is not news
By skroh on 10/24/2008 9:45:18 AM , Rating: 2
Disks have used "flying" heads for at least 25 years. The method of forming the circuit patterns is new here, certainly, but the flying part doesn't need the gee-whiz emphasis it is getting in this story.

RE: Flying is not news
By ianweck on 10/24/2008 10:50:38 AM , Rating: 2
Maybe the size of the gap between the head and the disk is what they are emphasizing. 20nm is pretty small, anyone know what the gap is on a regular HDD?

RE: Flying is not news
By Chocobollz on 10/24/2008 3:14:39 PM , Rating: 2
Qoutes taken from Upgrading and Repairing PCs, 17th Edition by Scott Mueller:

For my example, I use an IBM Deskstar 75GXP drive, which is a 75GB (formatted capacity), 3 1/2" ATA (AT Attachment interface) drive. The head sliders (called pico sliders) in this drive are about 0.049" long, 0.039" wide, and 0.012" high. They float on a cushion of air about 15 nanometers (nm or billionths of a meter) over the surface of the disk while traveling at an average true speed of 53.55 miles per hour (figuring an average track diameter of about 2 1/2"). These heads read and write individual bits spaced only 2.56 micro-inches (millionths of an inch) apart, along tracks separated by only 35.27 micro-inches. The heads can move from one track to another in 8.5 milliseconds during an average seek.

So I guess it's 15 nm maybe?

RE: Flying is not news
By Silver2k7 on 10/26/2008 1:00:04 PM , Rating: 2
"For my example, I use an IBM Deskstar 75GXP drive, which is a 75GB (formatted capacity), 3 1/2" ATA (AT Attachment interface) drive. "

no way did they sell the drive with the formatted capacity!!
why don't they do that today!! >:(

today we have 1.5TB drives wich are really 1500GB (should be 1536GB)
wich becomes 1.36TB formatted..

Too much hype and self contradiction
By Shadowself on 10/24/2008 9:15:22 AM , Rating: 1
"Utilizing this plasmonic nanolithography, we will be able to make current microprocessors more than 10 times smaller, but far more powerful. This technology could also lead to ultra-high density disks that can hold 10 to 100 times more data than disks today."


This could be used in microelectronics manufacturing or for optical data storage and provide resolution that is 10 to 20 times higher than current Blu-ray technology."

So which is it? 100 times better or 20 times better?

If the best lithography today (not shipping chips, but best systems being tested today for the next generation chips expected to be shipping in 15 months) is 32 nm then 10 times smaller is 3.2 nm. At this size quantum effects become the overriding concern. At this size the underlying electronics need to change too or else the chips will become hopelessly unstable.

These guys, at the very least, need to get their story straight.

By ianweck on 10/24/2008 10:47:05 AM , Rating: 2
I agree with you regarding microprocessor production. I don't think the semiconductor industry is even ready for 3.2nm, materials wise. Even if this tech was ready to go now, the rest of the manufacturing process would be the bottleneck.
However if we only talk about data storage then this could easily be implemented. Printing pits and lands on a disk seems a lot less complicated than printing an IC pattern. I didn't notice in the article if they had a read head yet for something that small though. Would the write head work both ways then?

By JediJeb on 10/24/2008 12:11:41 PM , Rating: 3
I guess you have to think in three dimensions on that. If 32 reduces to 10 in one dimension then in three dimensions it is 32768 versus 1000 ( 32 cubed versus 10 cubed) that is a factor of over 32 in space savings. If taken only in two dimensions then it is 1024 versus 100 which is slightly over a factor of 10 in space savings.

Not sure how it works in lithography, but if there is a savings in design space other than the length of the component then there should be a large savings of space on the platter.

By pnyffeler on 10/24/2008 11:01:56 AM , Rating: 2
Any comment about electron leakage for 5 nm transistors? My understanding is that as transistors get smaller, leakage ends up being a huge problem, due to the quantum effect of tunneling, which allows electrons to pass right through the gate. I think some advances were made going to low-K silicon or something like that, but there has to be some limit.

At this size, would the transistors still function as gates or just make a really expensive heater?

RE: Leakage???
By hellokeith on 10/24/2008 1:41:12 PM , Rating: 2
Yes you are correct.

The 16nm node is likely the last node for photolithography, because features will be 8nm thick, which is very very close to the tunneling limit (~6nm).

This discovery may get us to the 16nm node faster and cheaper, but it doesn't solve tunneling issues.

By ViroMan on 10/24/2008 10:37:38 PM , Rating: 2
and tell them they can go to 5nm before Intel gets there!

AMD needs to pounce on this tech HARD! I mean, they are lagging behind in the die shrink and its showing. Shrinking the damn CPU makes your chip more efficient by using less energy and therefor making less heat. Not to forget smaller dies means you can pack iether more instructions or more cores. HELL with this tech AMD could put 12 cores on the damn chip PLUS 8 VIDEO cores!

lol Someone needs to hit up Sony too. Now introducing the new PS5... WITH 64 CORES!!!!

By Garreye on 10/25/2008 7:47:06 AM , Rating: 2
Unfortunately, this tech is still in research phase. It is probably going to take several years to be able to commercialize this, and a lot of money, which, as we all know, AMD is lacking in. Also, seeing as AMD is going fabless it doesn't look like they're looking to invest in anything like this any time soon.

Two-years-old news: crossbar latches
By cete on 10/27/2008 5:41:44 AM , Rating: 2

SAN JOSE, Calif. (AP) -- Challenging a basic tenet of the semiconductor industry, researchers at Hewlett-Packard Co. have demonstrated a technology that could replace the transistor as the fundamental building block of all computers.
The smallest features of today's silicon-based transistors are about 90 nanometers long, a nanometer being roughly one hundred-thousandth the width of a human hair. The crossbar latch, by comparison, can work in a space of about 2 to 3 nanometers.

If they can do this by 2012, as they say, why bother with spending billion of dollars to shrink further than 32nm? New architectures/features on 45-32nm and better software is all we need to be happy...
This is a pointless race, IMO, do we need some monster cpu/gpu, built at extreme high cost, just to be abused by poor written software?

By goku on 10/27/2008 8:22:04 AM , Rating: 2
While I agree that software these days is hopelessly inefficient, some things really do need raw performance increases as no amount of optimization will lead to the desired goal, well unless time is on your side (i.e slow). I suppose the only thing you could hope for is for hardware to start slowing down in development speed so that software advancement is being written at or the same speed as hardware development..

By steve391 on 10/24/2008 4:34:28 PM , Rating: 1
"Caustic baths such as acid" this makes no sense. Caustics and acids are opposites.

RE: Contradiction
By menace on 10/24/2008 5:38:33 PM , Rating: 2
Not really. Caustic means burning and acids burn.
Although many typically thinks of caustic soda (NaOH) in regards to a theraputic caustic bath, I don't think that the poster intended that narrow meaning of the term.

By MrPoletski on 10/24/2008 7:49:40 AM , Rating: 2
so super fast and super cheap chips int he future then w00t...


I might pick up a copy of this magazine publishing the full details to find out how much of this is bull.

By dickeywang on 10/24/2008 9:15:42 AM , Rating: 2
If the lens need to fly around like this, I guess a simply drop on the ground are likely to damage the drive. You know, just like the current generation of HDDs. I guess they need to figure out a way to prevent this before these drives to be massively produced.

By swizeus on 10/24/2008 12:09:50 PM , Rating: 2
careful with scratches and dusts though, it'll make u lost 1 GB of data

Curved Plater
By EvL OnE on 10/25/2008 12:49:55 AM , Rating: 2
Curve the platter and rotate the lens, the platter is now stationary and the lens would rotate = increase in speed and more stable, the platter would look like a sphere cut in half with the 180` rotary lens in the center.

Just my $2c

By amanojaku on 10/25/2008 4:10:47 PM , Rating: 2
Keep flying.

By William Gaatjes on 10/26/2008 10:08:41 AM , Rating: 2
The new method uses a phenomenon where metal electrons vibrate when exposed to light. These tiny vibrations are smaller than light's normal wavelength and are known as evanescent waves.

Are these waves researched also with respect to super conducting ? I will have to read up on that. I find it very interesting.

"I mean, if you wanna break down someone's door, why don't you start with AT&T, for God sakes? They make your amazing phone unusable as a phone!" -- Jon Stewart on Apple and the iPhone

Most Popular ArticlesAre you ready for this ? HyperDrive Aircraft
September 24, 2016, 9:29 AM
Leaked – Samsung S8 is a Dream and a Dream 2
September 25, 2016, 8:00 AM
Inspiron Laptops & 2-in-1 PCs
September 25, 2016, 9:00 AM
Snapchat’s New Sunglasses are a Spectacle – No Pun Intended
September 24, 2016, 9:02 AM
Walmart may get "Robot Shopping Carts?"
September 17, 2016, 6:01 AM

Copyright 2016 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki