Lasers perform their shortest act ever
By Glenn Zorpette
Redherring.com, February 02, 2001
herring.com
This article is from the February 13, 2001, issue of Red Herring magazine.
In a room the size of a handball court strewn with cables, hoses, mirrors, and gratings, Victor Yanofsky, a Ukrainian émigré with a bushy moustache and Birkenstock sandals, waves toward a laser that will offer us a window on the center of stars.
He hopes that by June, his "baby," aborning at the University of Michigan in Ann Arbor, will spit pulses that last no more than 25 quadrillionths of a second and peak at 100 terawatts -- 33 times the world's installed generating capacity. For one shining instant, it will rival the nuclear inferno that makes a star blaze. "This is like a bridge," he says, "from conventional optics to nuclear physics."
It will even heat your coffee -- by six thousandths of a degree. That is a measure of how fleeting the stupendous pulse will be.
IN A FEMTOSECOND
Now try to wrap your mind around this concept: 0.000000000000001 second, also known as a femtosecond. There are more femtoseconds in one second than there are letters in 224 million Bibles.
These metrics make palpable the briefest events ever created by human beings: laser pulses as short as 4.5 femtoseconds. These pulses are not as powerful as Mr. Yanofsky's will be. But even garden-variety femtosecond pulses are strong enough to focus themselves in air, ripping the electrons off molecules and producing a lightninglike effect.
Lightning-on-demand is the least of the anticipated uses for these lasers. In fact, some fields -- ophthalmology is one -- will be turned on their heads. Already doctors are starting to use ultrafast lasers as superfine scalpels for corrective vision surgery and for treating glaucoma, and to make sharper microscopic images of the retina and other tissue. Industrial engineers use them experimentally to drill and cut with mind-boggling precision. Chemists rely on them to study the twists and turns of chemical reactions that run their courses in hundreds of femtoseconds. Electrical engineers apply them in new schemes to bundle more communications channels into a single optical fiber and to analyze semiconductor materials and troubleshoot high-speed integrated circuits. "There is a new application every day," reports David Spence, a senior scientist at Spectra-Physics (Nasdaq: SPLI), a laser manufacturer in Mountain View, California.
It's clear why. Ultrashort laser pulses do things ordinary laser light cannot: to gauge something short, you need a measuring stick that's even shorter. In a typical chemical experiment, a laser pulse hits a molecule, breaking it apart; femtoseconds later, new pulses illuminate the fragments for a series of snapshots. Those snapshots reveal how the pieces flew apart and how they passed through intermediate chemical stages you never could have guessed from the final outcome.
Such information can be worth millions of dollars if the reaction is a critical step in the production of, say, a drug. The Nobel Prize committee liked the concept so much that it awarded the 1999 prize for chemistry to Ahmed Zewail of the California Institute of Technology, who pioneered the use of ultrafast lasers as femtosecond camera shutters.
Femtosecond lasers are also incredibly useful if you need to do great violence to tiny things. For instance, say you want to punch a very small hole with a precise shape and perfect edges. (Think of the hole Bugs Bunny leaves, ears and all, when he smashes through a wall.) Normal laser light can't cut that cleanly because a lot of the energy bleeds into the surrounding material, melting it. A femtosecond blast, however, instantly turns the material into a gas under such a high pressure that it flies away from the rest of the material at supersonic speed, carrying excess energy with it. No energy penetrates the surrounding area, which therefore remains unscathed.
Imagine a dentist's drill that would be painless, really and truly. When an ordinary drill bit excavates your tooth, the friction causes heat, which hurts when it hits a nerve. An ultrashort laser flash could dig the same hole without adding heat and causing pain. Alas, it would cost plenty. "You're looking at a half-million-dollar system," estimates a laser physicist. "A dental drill costs $3,000, and dentists are the cheapest people on the planet," he adds. (He asked not to be identified, fearing reprisals from his dentist.) "The pain will continue."
But the high price of femtosecond drills doesn't bother other potential users. Automakers are testing femtosecond lasers to see how well they can drill the tiny holes in fuel-injector nozzles, where more precision can mean better atomization of the fuel and a cleaner burn. And some biomedical engineers are investigating the possibility that ultrafast lasers can shape smoother, safer stents, the small cylindrical tubes that go into blood vessels to hold them open.
SINE LANGUAGE
Another, more subtle characteristic linked to the shortness of the pulses could lead to the most lucrative application of all: optical multiplexing. The subtlety has to do with the anatomy of a signal. Any signal -- the sound of a symphony orchestra, the X-ray emissions from a black hole, or the light flashing down a fiber-optic cable -- can be broken into a set of sine waves of different frequencies. Middle C on a piano, for example, might have hundreds of sine waves. By performing an operation called a Fourier transform, you can discover which sine waves went into the signal. If you transform a short pulse, you find that it is made up of many sine waves of different frequencies. The shorter the pulse, the more sine waves it contains. Femtosecond pulses have the most sine waves of all.
At Lucent Technologies (NYSE: LU)'s Bell Laboratories, researcher Wayne Knox and his group are eyeing that mother lode of sine waves. Mr. Knox built his first ultrafast laser at age 17. Now, 26 years later, he is pursuing an idea that could help realize the decades-old dream of getting fiber-optic lines to your doorstep. It concerns wavelength-division multiplexing (WDM), a standard technique in communications that combines many high-speed data streams.
With WDM, you impress each stream of data bits onto a wave -- that is, you modulate the wave -- and you use a different wavelength for each stream. All the beams are sent down a single fiber, at the end of which the wavelengths are separated, along with their data. The scheme uses up to 100 lasers, each with its own modulator.
Mr. Knox, Martin Nuss, and others at Bell Labs had another idea. Because femtosecond lasers emit pulses that have a huge number of wavelengths, a single such laser and one modulator could replace the many lasers and modulators in a conventional WDM system. As a side benefit, the femtosecond approach supports far more channels -- so far, 1,021. "Initially, people thought femtosecond lasers would be useless in telecommunications," Mr. Knox recalls. "They didn't think of the most important thing: how do you get the data into the system?"
Mr. Knox and company took a 100-femtosecond pulse and sent it down an optical fiber. The fiber "chirped" the pulse, stretching it 300,000-fold by shifting the longer wavelengths forward and the shorter ones backward. The result was a big, fat pulse all of 30 nanoseconds long. As this chubby pulse came out the end of the fiber, it was fed into a modulator, which basically chopped it into hundreds of segments and stamped each with data. Because the pulse had been chirped, its longest wavelengths came out first, followed by the medium ones, with the shortest bringing up the rear. So as the modulator separated the fat pulses into segments and put data on each one, it was also multiplexing them in a practical order.
These femtosecond multiplexers do not blow away conventional WDMs in terms of overall throughput. Their real breakthrough is the relatively huge number of channels they can pump full of data, which you need if you're going to supply every household with its own fiber link. What does it matter if you can put only 45 megabits on each channel? "That's a T3 line," notes Mr. Knox. "The Washington Post uses one to connect its servers to the world, and they came close to filling its capacity only once, after the paper broke the Monica Lewinsky scandal."
Mr. Knox envisions systems that handle 15,000 different channels, each carrying 1 Gbps. Running fiber-optic cable to individual homes is still years away, but when it eventually does happen on a large scale, it is a good bet that all those fibers will carry data from femtosecond-laser WDMs. It's a thought that gets Mr. Knox out of bed some mornings. "Everybody should have their own wavelength," he says.
Although it will be a while before you get your own wavelength, a much quicker payoff should come from another advantage of femtosecond lasers: their ability to zap tiny volumes inside a piece of material. The stuff being zapped must transmit the laser's wavelengths -- glass works; so does the cornea of the eye.
OPTICAL WONDER
The optics are set up so that the femtosecond pulses come to a focus below the surface of the material -- say, glass. The peak power of the pulses is set so that only at that focus point will the intensity be enough to achieve the desired effect on the glass. Adjust the focus or power, and you can etch a tiny light-channel, called a wave guide, much like a bit of optical fiber. Other structures already fashioned include signal splitters and couplers, which carry data to and from bundles of optical fiber.
This technique seems to have advantages over the lithographic method now used to make optical components. For one thing, you can put several devices, in different planes, inside a single block of glass. "We are the only company in this space that can actually fabricate telecommunications devices in three dimensions," says William Clark, CEO of Clark-MXR, a maker of industrial femtosecond lasers. The company, one of a bumper crop of startups spun off from the University of Michigan's Center for Ultrafast Optical Science, has secured a $1 million round of funding from Avalon Investments and is now looking for a second round to finance an expansion into optical devices.
The same idea, applied to the cornea, promises to transform the $3-billion-a-year LASIK surgery business. Conventional "flap and zap" LASIK (laser-assisted in situ keratomileusis) usually starts with a tiny motorized metal blade, a miniature version of what a merchant uses to make an imprint of a credit card. The blade cuts a thin flap of tissue at the front of the cornea, nanosecond pulses from an excimer laser sculpt the underlying tissue, and the flap is put back. Anywhere from 1 to 3 percent of cases have complications, many of them linked to that little metal guillotine. Problems include infections; ingrowths; and complete cuts that remove the flap from the cornea, rather than leaving it dangling and easy to replace when the operation is over.
So Intralase, another spin-off from the Michigan Center, has pioneered the use of femtosecond lasers to cut the cornea. Because nothing comes in contact with the eye, there is less risk of complications. Although the company has not yet started selling the lasers, it already has U.S. Food and Drug Administration approval and has prototype units performing the cuts in two LASIK centers, in San Diego and Houston. Greg Spooner, a University of Michigan research scientist who consults for Intralase, says the next step would be to replace the excimer laser with a femtosecond laser. That way, ophthalmologists would need only one laser rather than two, saving hundreds of thousands of dollars.
Ultimately, femtosecond lasers will perform the operation by removing minute amounts of tissue below the surface. "You'd never open the cornea at all," Mr. Spooner says. Intralase has already tested the procedure on animals and, overseas, on blind or partially sighted human volunteers. The results are encouraging, but to do surgery in a reasonably short time, doctors will need higher pulse-repetition rates than are possible now. They will also need a fuller understanding of the biomechanics that govern how the corneal tissue relaxes after tiny bits are removed from under its surface. Mr. Spooner estimates that a commercial system will be ready in two years.
LASER MAGIC
There are loads of other possible applications, but most of them won't make commercial sense for a while, because femtosecond lasers aren't cheap. Still, the trend is good: a high-powered femtosecond laser costs hundreds of thousands of dollars, as opposed to millions just ten years ago. It can fit on a dining-room table or in a suitcase, instead of taking up an entire room. What's more, now it can be operated by a layman, whereas earlier versions had to be coddled by physics graduate students.
To understand how these devices work, you need to know a few basics of laser physics. A typical laser consists of a crystal between two partially reflective mirrors. When the crystal is "pumped" with light, energized electrons in its atoms jump to higher energy levels and then fall back, emitting photons of light. Those photons leave the crystal and bounce back and forth between the mirrors. On each trip, the photons go through the crystal, juicing electrons and stimulating the emission of yet more photons.
How do you turn a regular laser into a femtosecond laser? Start by using a material called titanium sapphire as your crystal. When its electrons give up energy, they emit photons with a huge number of different wavelengths. Recall that an ultrashort pulse consists of many wavelengths. Basically, the titanium sapphire gives you the raw materials for an ultrashort pulse.
Here's the first tricky part: to get that pulse to occur -- to turn a low level of continuous light into periodic pulses of stunning peak power -- you've got to force all those unruly, haphazard wavelengths to be, at one instant, at the exact same spot in their cycles. All those many waves with different frequencies, oscillating from crest to trough to crest to trough, must be forced to adjust so that every so often they all fall into sync. It doesn't matter where in the cycle they sync with each other -- at the top of a crest, the bottom of a trough, or anywhere in between -- as long as they are all there at the same time.
Now, the next tricky part: the Kerr effect. Titanium sapphire, like many other materials, transmits superintense light differently from dimmer light. If the central -- and brightest -- part of a beam is intense enough, it will travel more slowly through the crystal than will the dimmer edges. The beam becomes focused, as if it were passing through a lens.
To use the Kerr effect, physicists place aperture-like barriers in strategic locations in the laser. When the beam is intense enough to be Kerr focused, it slips through the apertures unobstructed. When the beam is not sufficiently intense, it does not focus, and some of the light is lost at the edges of the barrier, around the aperture. The only time the beam is intense enough to Kerr focus is when all its power is concentrated in short pulses.
Thus the laser can be stable in either of two states: emitting a low level of continuous light or emitting much higher levels of light in short, periodic pulses. Of the two, the laser is more efficient in the latter state, which it is said to "prefer."
How can an inanimate object like a laser prefer something? We could tell you, but then we'd have to kill you. Or at least bore you to tears. You can either accept the preposterous notion of a laser preferring to behave a certain way, or you can be stupefied with differential equations. The mathematics were "hashed out in the early '70s by people at Imperial College in London and at MIT in Cambridge," says Mr. Spence. Along with colleagues Peter Kean and Wilson Sibbett at the University of St. Andrews in Scotland, Mr. Spence was the first to exploit the Kerr effect, in the late '80s. The mathematics are "reasonably complicated," he says. "It's not something you'd teach to freshmen in a physics course."
Once the basic technique of making supershort pulses was established, a slew of other breakthroughs let researchers amplify and measure the pulses. Soon, refinements to the fundamental setup enabled laser physicists to get into a free-for-all to see who could make the shortest pulses. The current record, set in 1999 at the University of Groningen in the Netherlands, is a 4.5-femtosecond pulse.
Today, research is proceeding on several fronts. Some groups, like Mr. Yanofsky's at the University of Michigan, are pushing power levels to new heights. Those researchers were energized last autumn when physicists at the University of Osaka in Japan announced that they had used a femtosecond laser as a kind of spark plug to boost the energy yield of a laser-fusion experiment. It's far too soon to say whether femtosecond lasers may point the way at last to practical fusion power, but "if it works, it will have an enormous impact on fusion research," says Todd Ditmire, a laser physicist at the University of Texas at Austin.
At least a few researchers, meanwhile, are trying to do 2-femtosecond pulses, which would amount to single oscillations of lightwaves -- a crest and a trough, basically. And other physicists are working with X rays, which hold out the promise of attosecond pulses -- a thousand times shorter than femtosecond ones. (An attosecond is to the blink of an eye what the blink of an eye is to two billion years.)
Do we run out of funny prefixes then? No; there's still zepto and yocto. (There's even a proposal for harpo and groucho. Seriously.) But the advent of the atto age does make you wonder: Will the femtoweenies become attowonks? And will they change their rallying cry from "nanoseconds are forever" to "picoseconds are forever"? |
