Ultra-Dense Optical Storage on One Photon

Andreaskem submitted this story about researchers being able to encode an image into a photon and to later retrieve it intact. From the article: "It's analogous to the difference between snapping a picture with a single pixel and doing it with a camera — this is like a 6-megapixel camera... You can have a tremendous amount of information in a pulse of light, but normally if you try to buffer it, you can lose much of that information... We're showing it's possible to pull out an enormous amount of information with an extremely high signal-to-noise ratio even with very low light levels."

Comments

To Clarify

[logicnazi]

The Image is NOT encoded into one photon, at least not in a way that can be extracted again. Each individual photon is in a superposition of having gone all the possible paths and the set of those possible paths is the information to be extracted but when measured each photon will only reveal a small amount of information so it is only in the aggregate (by measuring lots of photons) that the initial image can be reproduced. At least this is what the article sounds like it is saying it wasn't very clear.

In fact it is probably best to think of this without quantum mechanics at all. What they did is pretty much like figuring out the shape of an object by shooting BBs at it and looking at which ones make it past the object.

The part that is supposedly new and interesting is the way they collected the photons at the other end. It didn't seem very clear on this but apparently by catching many of the photons in their device at one time it made it much easier to decode the image in the light.

Re:To Clarify

[drerwk]

Parent is right - article is not particlely clear.
By itself a photon can be described as having, location, and energy. Thanks to Heisenberg you can only know so much about location and energy at the same time. I don't recall what property fails to commute with spin, maybe time? But the total information in a single photon is at best 3 reals for location, a real for energy, and an imaginary for spin.

Re:To Clarify

[mastershake_phd]

In fact it is probably best to think of this without quantum mechanics at all. What they did is pretty much like figuring out the shape of an object by shooting BBs at it and looking at which ones make it past the object.

You mean like a X-Ray.

Re:To Clarify

[Andreaskem]

"To produce the UR image, Howell simply shone a beam of light through a stencil with the U and R etched out. Anyone who has made shadow puppets knows how this works, but Howell turned down the light so much that a single photon was all that passed through the stencil."

The article as a whole might not be very clear, but this line says that only a single photon passed through.

Re:To Clarify

[Teresita]

"Thanks to Heisenberg you can only know so much about location and energy at the same time."

Dern that Heisenberg. And you can also thank Einstein for the fact that it takes at least one year to travel one light-year.

Re:To Clarify

[Beryllium Sphere(tm)]

>don't recall what property fails to commute with spin, maybe time?

Spin in the non-measured axes.

Time pairs up with energy: if you look at a really fine time scale, energy is so uncertain that there's a sea of particles (m == E / c**2).

Re:To Clarify

[Myrv]

They simply say one photon passes through the mask at a time. They didn't say the entire image was reconstructed using that single photon.

This sounds very similar to the double slit experiment were you send single photons through a double slit and record where they land on a screen placed behind the slits. Each photon will only light up one spot on the screen but if you collect enough samples you see a pattern start to emerge that looks like the interference pattern you would expect if light passed through both slits simultaneously. Basically, each photon which passes through the slits interfers with itself to to form the interference pattern.

In the article they are simply firing the photon through a mask with a pattern in it instead of a double slit. The photon acts as if it passed through all parts of the mask at the same time. But to reconstruct the image they would have to sample many photons passing through mask.

From what I can gather the important part of the article is that they have been able to slow down each photon in order to buffer it. So you can send 100 photons through the mask (one after another) then buffer those photons for 100 ns and then pass them on to a detector that reassembles the image from the 100 or so photons. I'm also guessing they can't slow down multiple photons at a time (at least not reliably) so the ability to serialize the photons is important as well.

Re:To Clarify

[Anonymous Coward]

Sounds like the quantum bomb problem - detecting something by using less than a single photon.

Re:To Clarify

[CharlesEGrant]

rom what I can gather the important part of the article is that they have been able to slow down each photon in order to buffer it.

The original press release is very poorly writen. A better article is in the Washington Post. Also, the title of the actual peer-reviewed article is on Howell's publication page as "All-optical delay of images using slow light" Ryan M. Camacho, Curtis Broadbent, Irfan Ali Khan and John C. Howell, Phys. Rev. Lett (in press). As you say, the centeral acheivement is in their ability to slow down the photons. Unfortunately the actual paper doesn't yet seem to be available as the Phys Rev Letter website. I think the business of encoding an image on a single photon is a confabulation by the author of the press release.

Re:To Clarify

[mspohr]

Heisenberg might have been here.

Photo Storage

[Freestyling]

Does this mean I can now store my photos in a nice easy to carry cartridge or caesium gas? This is a great improvment on these clunky microSD cards I use now.

Incorrect summary

[forand]

Both the poster's summary and the news release are incorrect. You cannot encode more information than quantum numbers on any quanta, it is not possible. I believe that another poster has a plausible explanation for what is actually going on: that they measure many photons and reconstruct the information by knowing the possible paths which do the encoding of information.

Re:Tomorrow's news

[Teresita]

"RIAA filed a suit against University of Rochester and all of its students for "Helping those damn, dirty pirates infringe on our copyrights!!"

They turned Britney Spears' "Oops I Did It Again" into a giant single number, and imprinted that number on the photon, thus making an illegal photon.

Reminds me of a short story...

[fredklein]

This reminds me of a short story (by Clarke or Asimov, I think). It's the far future, and increasingly dense data storage (the terms "notched quark" and "nudged quark" are used) means all of Humanities knowledge fits into a single file cabinet-sized drawer. All the rest of the world-wide internet-like system consists of indexes, indexes of indexes, and indexes of indexes of indexes of... well, you get the idea. One day a worker comes across an error, and forwards it to his boss. It keeps getting sent up the chain of command until a Master Troubleshooter realizes that to fix it, he needs to refer to the original datastore location. He enters the command to find the physical location of the datastore... and gets the same error.

Uhh-oh. :-)

A photon carries a lot of information

[Anonymous Coward]

Any photon has a frequency (wavelength, energy, whatever). The frequency is not quantified and can assume infinite values. By generating a photon with the correct energy, I have encoded, in theory at least, a vast amount of information. Of course your ability to encode and decode very much information is limited by the available technology and the noise environment. :-)

Re:A photon carries a lot of information

[unchiujar]

Not so, :) A single number can store a huge amount of information. Your hardrive is one single very long binary number. If you define a way of retrieving information you can store images as numbers (binary, hex, octal,decimal or otherwise).

Links to the researchers

[Beryllium Sphere(tm)]

Howell's home page
Boyd's home page

The article isn't a good match with any project listed there.

The idea of storage by slowing something down goes back to a comically ancient technology, which was converting bits to sound waves and sending them through tubes of mercury to be detected electrically milliseconds later.

Re:Links to the researchers

[Richard Kirk]

In about 1968, IBM had an optical memory where about 2 Km of optical path was folded into something the size of a filing cabinet using mirrors, and 1 bit was circulating endlessly. Optical fibres transparent enought to do this did not happen for years. This geta a brief mention in... http://www-03.ibm.com/ibm/history/history/year_196 8.html

How long to travel a light year

[benhocking]

Actually, you can travel a light year in significantly less than a year, depending on how one defines "light year" and "year". For example, if you accelerated at 1 g towards Alpha Centauri (fun fact: 1 g is just over 1 ly/yr^2!), you would reach Alpha Centauri in about 2.25 years. Of course, looking back the original distance of 4 light years would now be shortened (thanks to that fella Lorentz). Bonus fact: as you pass Alpha Centauri, you will be covering 5 light years (as measured in the Earth frame of reference) per year (as measured in your own frame of reference)!

See, Einstein wasn't so mean after all.

Better coverage ...

[CharlesEGrant]

I suspect the original press release and the articles on Science Daily and PhysOrg are FUBAR. I think an article in the Washington Post is probably more accurate. Unfortunately the Phys. Rev. Letter web site doesn't seem to have the actual paper publicly available yet.

If this were true, one profound implication...

[RyanFenton]

If this information-encoding method were true (single photon carrying megabytes of information), then there would a profound implication:

Because a computer of a given mass could then theoretically be used to completely store information of a physical structure of real objects (position and properties of each atom), these systems could then completely simulate/emulate these real objects of a mass larger than the mass of the computer, even if not in realtime. That enables a large variety of applications IF it is additionally possible to acceptably scan the data of the makeup of real objects. You could theoretically have a simulation of our physical universe, without having to use the mass of the universe to make that simulation!

Major roadblocks would be the depredation of data on the light over time, and requirements of isolating the data - if the properly shielded case for a 'light hard drive' needed to be heavy enough, or the energy needed to maintain the data were enough, it could make production impractical, even if it could do what we wanted.

Very interesting research, if the data 'storage' ends up being what they think it is.

Ryan Fenton

Re:Over simplification

[Dunbal]

Using just ONE photon to produce an image?

      Ahh, but what they fail to mention is that the image is of.... tadaa, the photon!

The Same News Story Twice on /. In Two Days

[Pooua]

I just wondered if anyone noticed that this news story is exactly the same as the one /. posted under the heading, Slow Light = Fast Computing, on January 19?