Attosecond long laser pulse. What’s it good for?

[Michael Weber]

Well, I don’t rightly know or comprehend the explanation. But they’re in the news because of the Noble Prize award in Physics for work generating laser light pulses as short as that. An Attosecond is a billion billionth of a second. A very short pulse of light. What I found fascinating was the fact that an Attosecond is a time period compared to a second is shorter than a second is compared to the creation of the universe 13.7 billion years ago. Another fascinating thing I saw was in a MIT video shot 11 years featuring a short laser beam slowed down with a “camera” capable of 1 trillion frames per second. It turns out that speed has improved dramatically since this video was made. But what got my attention was the statement that in comparison a fired bullet slowed down the same amount would take 3 years to traverse that soft drink bottle.

[Jim Koepke]

An Attosecond is a billion billionth of a second.

At that speed of pulsation information sent through fiber optic strands could transmit a billion billion bits of data per second.

Higher speed data transfer is what drives the communication world and computer systems.

jtk

[Steve Demuth]

It would seem that way, but it's actually not so. First, the attosecond pulses being generated for physics experiments are not 1 attosecond in duration, but tens of attoseconds, and most typically around 100 as. So, two factors of ten lost right there. But there are more limits. First, a pulse of "light" that is on the order of tens of as long, is actually best understood as a single wavelength of very short wavelength light - which makes it in the very hard ultraviolet or even soft x-ray range. That is, not generally transmissible by optical fiber. And, the techniques for generating these pulses can't actually generate them a billion, billion times per second in succession - doing so would actually be a beam of x-rays. In reality, the fastest sources manage about 100,000 pulses per second.

[Bill Dufour]

maybe for ultra high speed filming. Watch atoms shake around.
Bill D

[Bert McMahan]

You can use pulsed lasers for material science as well. Imagine the difference between a propane torch and a heat gun. You could (theoretically...) cut through a piece of styrofoam with the handheld torch since the heat is directed and you can hit one small area and move quickly. If you tried to cut a piece of foam with a heat gun, you'd wind up with a much larger "cut" area since the heat would soak around the cut, affecting far more material than the handheld torch would.

Same with a pulsed laser. If you output, say, 1 joule of energy over one second, you'll gently heat one area, and during that one second the heat will soak into the surrounding areas. If you output a joule of energy over 100 attoseconds, then whatever is absorbing that energy would get a tremendous amount of heat applied to it all at once, but not for long enough to "soak" anywhere else. That one tiny area gets hot, but you have very limited heat affected zone.

Now, this can be used for things like cutting, but it can also be used for all sorts of physics/science things where you want a LOT of energy in a VERY small area. I'm no physicist but you could imagine changing the grain structure of metallics only at the surface of the material, or creating nano-scale structures on the surface of something. A "metamaterial" is a word that describes a material behaving the way it does because of its shape, rather than an inherent property of the material. For example, as I understand it a lot of the coloring on titanium parts isn't due to normal "absorb and reflect wavelengths of light" but of intricate patterns on the surface reflecting different wavelengths of light, despite all being the same "color".

Think of a hole in the side of a wooden box. Viewed from the outside, the hole looks black even though the inside of the box is a normal brown wooden color. If you could cover the box in "holes" then it would look completely black. This doesn't make much sense for the visible spectrum, but you could imagine how that might be helpful to do for something like radar absorption on a fighter plane.

[Mike Henderson]

At that speed of pulsation information sent through fiber optic strands could transmit a billion billion bits of data per second.

Higher speed data transfer is what drives the communication world and computer systems.

jtk

There's a limit to how fast you can send light pulses in optical fiber. The pulse tends to spread out as it travels down the fiber. If you go far enough, there's just light, no pluses. I haven't worked in the field for 20 years but back then, the max was about 40Gbs (OC-768).

There are ways of sending multiple channels in a fiber, on different wavelengths of light (wavelenght division multiplexing), and combining the data at the end to get more than 40Gbps. There are limits on how many wavelengths you can put into one fiber.

Mike

[Bill Howatt]

This is a complex topic but a point is that actual data bits per second transmitted, depending on encoding, is not necessarily the same as the basic bandwidth (frequency of pulses) of the channel.

[Mike Henderson]

This is a complex topic but a point is that actual data bits per second transmitted, depending on encoding, is not necessarily the same as the basic bandwidth (frequency of pulses) of the channel.

By encoding, I assume you mean something like QAM or even PAM. Back 20 years ago, there were no codecs that could operate at the rate of optical transmission (such as 10Gbps). It may be that such codecs have been developed by now.

All that could be detected back then was the presence or absence of an optical pulse (at those data rates).

Mike

[John Stankus]

Femtosecond laser spectroscopist here (I know it's a bit slower than the attosecond stuff but...)

Attosecond is 10^-18 seconds while femtosecond is 10^-15 second

attosecond pulses are basically good for physics. The problem you run into is the shorter you make a pulse of light the greater the frequency bandwidth you need to make that pulse. This is one of the consequences of the Heisenberg Uncertainty Principle. A short pulse in time will have a wide width in Frequency.

For telecom applications you need to look both at pulse duration and bandwidth. Capacity of a fiber link is not only due to using short pulses, but also the ability to use wavelength division multiplexing putting different datastreams at different wavelengths down the same fiber. If you run ultrashort pulses you may get adjacent channel overlap. You will also run into issues with running short pulses down a fiber since the index varies with frequency(dispersion) which would spread the pulse out as it propagates down the fiber (playing games with that fact allowed us to create femtosecond pulses back when I was in graduate school) TANSTAAFL (There ain't no such thing as a free lunch)

Also, likely the picture of a laser beam from MIT was giving a view of the length of a pulse in a pulsed laser system. At the speed of light, light travels about a foot a nanosecond (10^-12 seconds) so a nanosecond long pulse in duration, would be about a foot long in a vacuum (it would be shorter in a higher index material). I used to do time resolved ultrafast holographic experiments where you would have to get femtosecond (actually proabably around ~150 fs) pulses overlapping in space AND time. Those pulses were of the order of 50 micrometers long. Attosecond long pulses would be a factor of 1000 shorter.

John

[roger wiegand]

In answer to the original question, the utility of these ultra-short bursts is that they allow, among other things, the study of the motion of electrons in atoms for the first time. I wouldn't venture to speculate on so-called "practical" applications, but they will probably turn out to be different than whatever the headline writers were talking about last week. Virtually every advance in the precision of measurement over the last couple of millennia has made a difference in the real world, I doubt this will be different.