Filaments of Light

Jérôme Kasparian. American Scientist. Volume 94, Issue 2. Mar/Apr 2006.

Next time you give a presentation about your research, take a close look at the laser pointer you’re holding in your hand. How big is the beam coming out of it? And how large is the spot that it forms? The answers will, of course, hinge on the particular laser pointer you’re wielding and the distance between podium and screen. Typical values might be a few millimeters for the beam as it exits the aperture of the pointer and a centimeter or so for the circle of light it casts across the auditorium. It takes only a smattering of physical intuition to guess the reason: Diffraction causes the beam to diverge. The actual cause may be a little more complicated, because some laser pointers include a lens that makes the light converge at a fixed distance from the tip, which leads the beam to spread out beyond this focal point-more so than if only diffraction had operated.

Imagine now that your laser pen packed a more powerful punch-say that the intensity of the beam was a whopping 1012 times that of a typical pointer. What then would the beam do as it crossed the room? (It’s clear enough what it’ll do when it hits the screen-quickly burn a hole). The answer, it turns out, is anything but intuitive. A laser of sufficient intensity traveling through air will-all by itself-engineer a narrow channel, one perhaps a tenth of a millimeter wide, over which light will propagate for tens or even hundreds of meters. Such filaments of laser light were first created a little more than a decade ago, and investigators are just now beginning to explore a variety of applications for them-mapping atmospheric pollutants, characterizing materials at a distance, perhaps one day even controlling lightning.

Let There Be (White) Light

Two separate physical phenomena account for the strange filamentary propagation of high-power laser light. The first is self-focusing, which comes about because the refractive index of air depends on the intensity of light passing through it, a phenomenon known as the optical Kerr effect. As a consequence, when one sends highpower laser light through the atmosphere, the center of the beam (where light intensity is highest) passes through gas that has a higher refractive index than the air located just off axis. The result is the same as if you had shot the beam through a convex lens, which has more glass (with its high refractive index) at the center than at the margins. Physicists refer to this configuration, sensibly enough, as a “Kerr lens.” For a laser with an 800-nanometer wavelength operating in air, a Kerr lens develops whenever the beam power exceeds a few gigawatts.

The more a laser is focused by such a Kerr lens, the higher the intensity becomes. And as the intensity rises, the focusing gets even stronger, boosting the intensity of light still further. Eventually, something has to give-and it does. When the light intensity reaches somewhere between 1013 and 1014 watts per square centimeter, a nonlinear process called multiphoton ionization comes into play. The oxygen and nitrogen molecules in air are then able to absorb many photons at once, stripping electrons from their parent atoms, forming a plasma.

Although Kerr-lens focusing and the ensuing creation of plasma could, in theory, be brought about using a laser that operates continuously at extreme power levels, in practice, it proves much easier to achieve the necessary oomph using short bursts-the shorter the better. Common sense explains why: For a laser pulse of a given energy, the more limited the duration, the higher the peak intensity. So with the laser’s energy concentrated in a brief pulse, the focusing effect is strong even though the average power in the beam is modest.

There’s a second reason to use very short bursts: The ability of a laser to ionize the air remains high, but the average density of electrons created is low, allowing the beam to propagate through them. (Electron density will be relatively low when the pulses are too short to shoot the released electrons into nearby gas molecules, releasing more electrons, which then would bash into other molecules and so forth in a process called cascade ionization.) Electrons are present in sufficient numbers, however, to decrease the refractive index of the air containing them, which results in the equivalent of a diverging lens and tends to defocus the beam.

Either phenomenon considered alone-the focusing of a Kerr lens or the defocusing induced by the electrons in a bleb of plasma-would prevent high-power laser pulses from propagating very far through the air. But it turns out that the two opposing effects can be made to balance, allowing the beam to travel over large distances without either diverging or collapsing. Instead, the energy is channeled along a narrow filament of light.

What happens when the intensity of laser light used is turned up higher than the critical value for filamentation to begin? You might guess that the light filament formed would become thicker and thicker, perhaps to the point of being better described as a “light rope.” But that is not what happens. Instead, several localized filaments emerge. That is, hiking the peak power of the laser pulses that are applied increases the number of filaments that result without notably influencing the individual intensity or the energy each filament carries.

Whether present singly or in bunches, these threadlike shafts of light exhibit another surprising property as well: Even though the laser used to create them produces essentially monochromatic light, each filament contains a broad range of wavelengths-what students of optics call “a white-light supercontinuum.” The transformation into white light is easy enough to understand once you realize that a pulsed laser doesn’t instantly switch on and off. Rather, the oscillatory electric and magnetic fields carried in each pulse gradually build to a maximum intensity and then diminish. That property alone explains some of the spectral broadening-basic physics dictating that the bandwidth of a pulse can be no less than the reciprocal of its duration. But that principle explains only a small part of the whitening effect. More important is the fact that the refractive index of the air containing the pulse is proportional to the intensity of the light. So where the intensity of light is highest (in the middle of the pulse), so is the refractive index, which causes the highest intensity light waves to be retarded with respect to the lower intensity waves that travel ahead and behind. The result is a distortion to the pulse envelope and the creation of light that contains both longer and shorter wavelengths than what the laser itself puts out. The range of different wavelengths that arise from this and other nonlinear effects makes the illumination essentially white.

As if the existence of narrow filaments and their ability to generate white light weren’t bizarre enough, another surprising phenomenon has been found to take place: A significant part of the white-light supercontinuum appears to be emitted backward! This back-directed light is the result of partial self-reflection of the forward-traveling beam, which experiences changes in refractive index along the axis of the filament as a result of the focusing and defocusing taking place. And just like with the beam of a flashlight shone on a double-glazed window, each change in refractive index produces a partial reflection.

Stepping Out

The laser-research consortium that I coordinate was established in 1999, when French teams led by Jean-Pierre Wolf at the Laboratoire de spectrométrie ionique et moléculaire (part of the Université Claude Bernard Lyon I) and André Mysyrowicz at Laboratoire d’optique appliquée (part of the École polytechnique in Palaiseau) joined forces with German groups led by Ludger Wôste at the Freie Universitàt Berlin and Roland Sauerbrey at the Friedrich Schiller Universitàt in Jena, Germany. Our aim was to create an experimental laser that could be brought into the field to study how light filaments propagate over greater distances than one can possibly arrange in the lab and to develop ways to use them for probing the atmosphere. A laser of this sort allows for the remote examination of gaseous or aerosol pollutants released, say, from automobiles or industrial installations. And it can be used to study the formation of water droplets in clouds.

It was clear early on that pursuing such investigations demands mobility, yet the high-power pulsed lasers then available took up most of a room-not something one could easily pack up and move. The solution was to install a laser of this type in a standard 20-footlong freight container, which could be carried by truck (or by ship) as needed anywhere in the world and operated even in adverse weather conditions. We call this portable terawatt laser system the “Teramobile.”

The laser we use is quite sophisticated. It sends out short pulses of infrared light (800-nanometer wavelength) 10 times per second. Each pulse is only 70 femtoseconds (70 millionths of a nanosecond) long when it exits the laser and carries 350 millijoules of energy. The peak power works out to 5 terawatts (5 × 1012 watts). My colleagues and I have been experimenting with this laser for several years, working mostly on schemes for measuring the composition of atmospheric trace gases as well as the abundance and nature of aerosol particles.

Several optical techniques for probing such properties of the atmosphere already exist, methods that go by such complicated names as “Fourier-transform infrared spectroscopy,” “differential optical absorption spectroscopy” and “light detection and ranging” (lidar). The Teramobile laser adds the possibility of carrying out such studies using one or more white-light laser filaments instead of the usual sources of light-ordinary (monochromatic) lasers or in some cases the natural illumination that the Sun or Moon provides.

Although laser filaments do not suffer the diminution in intensity that accompanies the spreading of a conventional laser beam, members of the Teramobile team were concerned at the outset of our investigations that these narrow channels of light might easily be blocked by raindrops or atmospheric dust. So we carefully studied the interaction of light filaments with such aerosol particles, introducing droplets of various sizes into the light path. It turned out that our worries were unjustified. We discovered that opaque droplets as large as 100 micrometers in diameter do not obstruct the propagation of a light filament, although they are about as large as the filament itself. At the same time we were doing these studies, see Leang Chin and his coworkers at the Université Lavai in Québec, found that a laser filament cannot be sent through a hole, even one that is several times the diameter of the filament.

This counterintuitive result is explained by the fact that a filament of light is not simply a tube through which all the photons flow; rather, it reflects a dynamic balance within the much more diffuse beam that surrounds it, something I like to call a “photon bath,” which acts as an energy reservoir feeding the filament when it encounters an obstacle. Thus, blocking the propagation of a filament in one place naturally spawns a new filament elsewhere within the wider beam. Numerical simulations by Jerome V. Moloney and his coworkers at the University of Arizona and by Luc Berge at Commissariat à l’énergie atomique (CEA, the French atomic energy agency) in Bruyères le Châtel show this effect well.

Light filaments sent into the sky can thus traverse a cloud so long as the accompanying photon bath makes it through. Small-scale laboratory tests had suggested that laser filaments should be able to pass through a typical cumulus or stratocumulus cloud without being visibly affected. My colleagues and I found similar results when we scaled up the experiment using the Teramobile beam and an open cloud chamber producing a 10-meter-long cloud of 1-micrometer droplets. Light filaments were visible exiting the fog, even for a concentration of almost 100,000 droplets per cubic centimeter, meaning that one filament must have hit an average of 2,000 droplets for each meter it traveled.

Up, Up and Away

To determine whether filaments of light could indeed penetrate high into the atmosphere, the scientists on the Teramobile project did the obvious: We tried it. After directing the Teramobile laser vertically upward, we studied the beam from the ground using the 2-meter-diameter astronomical telescope at Thuringer Landessternwarte in Tautenburg. Because the laser was located some distance from the telescope, we were able to obtain side-on images of the beam by virtue of Rayleigh scattering (the scattering of light off air molecules, which among other things causes the sunlit sky to appear blue). The pictures we took also revealed the pattern cast by the beam when it impinged on the bottom of clouds or layers of diffuse haze. These experiments, which were carried out in 2002, demonstrated for the first time an ability to bring light to a tight focus as far as 2 kilometers away from the laser source, at which point distinct filaments can propagate for hundred of meters. And although the reach of these high-intensity filaments is currently limited to such distances, the Teramobile laser is able to throw diffuse white light as high as 18 kilometers-that is, well into the stratosphere.

Having such a far reach holds great promise for probing the physical and chemical makeup of the atmosphere. Investigators have long applied lasers for this purpose, often using one or more refinements to the basic lidar technique, whereby a pulsed laser is directed into the air, and the backscattered light is measured as a function of time. Performing these measurements with a temporal resolution of, say, between one and ten nanoseconds provides a depth resolution of a few meters or less. Such observations, which are often obtained while sweeping the beam from side to side, allow for the construction of threedimensional maps of atmospheric aerosols or trace gases.

Currently, the most popular way to detect such gases (often pollutants) remotely is a technique called DIAL, shorthand for “Differential Absorption Lidar.” The strategy is to compare the lidar signals obtained at two slightly different wavelengths, one being set exactly to an absorption line in the spectrum of the pollutant under scrutiny. seeing a diminution in the amount of light returned at that wavelength but not at a slightly different wavelength attests to the presence of the targeted trace gas and rules out the possibility that something more mundane (say, clouds or haze) had obscured the light scattered back toward the observation station.

The problem with the DIAL method is that it can only be used to map trace gases that exhibit a narrow absorption line that is free of interference from the absorption spectra of other atmospheric components. This requirement limits its application severely. Worse, the need to tune the laser wavelength exactly to the absorption line makes it impossible to measure more than one pollutant at a time. And it makes DIAL blind to the presence of an unanticipated pollutant. Using the Teramobile laser or its equivalent for lidar should provide a better way to probe the sky, because the telescope can then gather light containing many wavelengths, not just one or two, and the resulting absorption spectra would reveal a wealth of information about the air this light passed through.

A similar tactic could one day be applied, for example, to characterize the nucleation of water droplets and their subsequent maturation in clouds. Measurements of droplet growth and density could allow meteorologists to forecast when rain or snow will form, or this information could be used to determine how much of the sunlight falling on a given cloud reflects back into space. Why use a ground-based laser for such investigations? Droplet nucleation and growth take place over just of a few tens of minutes, so making the required measurements from research aircraft is generally too expensive to consider, and weather-balloon soundings are typically too infrequent to provide helpful observations. Optical remote-sensing techniques are clearly the most straightforward avenue for conducting such research, and the capabilities of the Teramobile laser in its white-light lidar mode are quite promising in this regard.

Bug Zapper Extraordinaire

The fact that the Kerr effect can transform a high-power infrared laser into a remote source of white light opens the door to a number of exciting applications. For example, the tendency for some of the light to be reflected backward suggests that we could create an artificial “guide star” for use in adjusting astronomical telescopes equipped with adaptive optics. But there are other nonlinear optical effects of the Teramobile laser that can be exploited as well. One is something called multiphoton fluorescence.

In normal fluorescence, a substance, say the phosphor powder that coats the inside of a fluorescent lamp, absorbs high-energy photons (typically in the ultraviolet) and releases lower-energy photons (having, usually, visible-light wavelengths). In multiphoton fluorescence, two or more low-energy photons are absorbed simultaneously, raising an electron’s energy level enough to allow a single high-energy photon to be given off when the electron returns to its original state. But because the chance of an atom absorbing two photons at once is quite low, light of very high intensity (that is, containing a very large number of photons) is needed. The pulsed Teramobile laser provides just such light, which proves a great boon for remotely sensing certain compounds using the phenomenon of multiphoton fluorescence.

In a 2002 experiment, my colleagues and I showed that the Teramobile beam and detection apparatus could sense biological aerosols at a distance. The motivation was to be able to map a cloud, say, of bacteria (perhaps given off during some industrial mishap or even a biological attack) and to identify potentially pathogenic agents among the various background atmospheric aerosols, among which may be more mundane organic particles such as soot or pollen.

Our test used water droplets sized to mimic bacteria and laced with the compound riboflavin, which fluoresces at visible wavelengths when it absorbs two infrared photons, producing a characteristic spectrum in the backscattered light. The experiment, carried out on a cloud located about 45 meters from the Teramobile laser, showed that it was easy to distinguish such a plume from a cloud of pure water droplets. With refinement, this technique could, potentially, be quite sensitive. We calculated that a laser tuned to excite two-photon fluorescence in the amino acid tryptophan would boost sensitivity by a factor of 10, allowing concentrations of as little as 10 bacteria per cubic centimeter to be detected 4 kilometers away. Although lidar systems based on normal fluorescence could also be used to probe for biological agents, the laser employed would have to operate at a shorter wavelength and thus be more prone to attenuation, limiting the distance over which it could function effectively.

The ability of laser filaments to deliver high-intensity light at substantial distances also opens the door to other very interesting applications. For example, it becomes possible to conduct elemental analyses of the surfaces of metals, plastics, minerals or liquids from an appreciable distance, using a variation of a technique called laserinduced breakdown spectroscopy. For that, a powerful laser is focused on the material of interest, causing some of it to be transformed into plasma. The emission spectrum of the glowing plasma can then be analyzed, revealing the nature of the substrate, with a detection limit that can be as little as a few parts per million for some elements. This method is currently used for such applications as the identification of highly radioactive nuclear waste and for monitoring the composition of molten alloys, because the tests can be performed without having to touch the sample. Imagine being able to do such probing from a large distance away! Normally, diffraction limits the intensity of light that can be focused on a remote target. But laser filaments can deliver intensities that are higher than the ablation threshold of many types of materials, at distances of hundreds of meters or even kilometers.

Another application under investigation may prove more spectacular yet-the control of lightning strikes. Lightning has always fascinated people, in part because of its unpredictable nature and destructive power-qualities that make these electrical discharges very difficult to study. Investigators from Electricité de France and CEA partially overcame those obstacles in the 1970s, when they developed a technique to trigger lightning on command using small rockets trailing thin wires. If shot upward at the right moment, the rockets and the wires they unspooled behind them served to initiate and channel the flow of electric current.

One outgrowth of this work was the idea of using a high-intensity laser to ionize air along the beam, thus forming a conducting channel of plasma that could replace the rockethoisted wires. The first attempts, mounted in the 1970s and ’80s, used lasers that produced nanosecond-long pulses. Those experiments were unsuccessful, however, because the plasma created by such lasers is largely opaque, which keeps the beam from extending a conductive path very far. But recently this field of research has seen renewed interest, because lasers can now provide higher intensities in shorter pulses, thereby avoiding the severe absorption that would otherwise occur. In particular, the team of Henri Pépin (Institut national de la recherche scientifique) and Hubert P. Mercure (Hydro-Québec) in Montreal have obtained quite promising results, using pulsed lasers to trigger and guide high-voltage discharges over several meters in the laboratory.

Spectacular experiments with the Teramobile system, installed in a highvoltage facility at the Technische Universität Berlin, showed that laser filaments can trigger and guide electric discharges over distances exceeding 4 meters. Moreover, the breakdown voltage is typically reduced by 30 percent. My colleagues and I have also shown that rain (or rather simulated rain) does not prevent the laser filaments from triggering these huge sparks. Research now focuses on the possibility of extending the lifetime of the plasma and increasing the length over which it is able to guide a discharge. Although the control of real lightning remains science fiction for the moment, recent progress in laser technology has brought this three-decade-old dream much closer to reality.

Over the past few years, the capabilities of terawatt-class lasers have improved markedly, while size and cost have come down. At the same time, physicists have made great strides in understanding the nonlinear propagation of these high-power laser pulses in air. The rapidity of this progress suggests that Teramobile-type lasers, or systems like it, might soon be used widely, not just by scientists in the course of their research but for any number of military, commercial or public-safety applications.