Analytical Chemistry, July 1, 2001 Volume 73, No. 13, pp 351

News In Brief

> Rainbow Stars: A spectrum of possibilities

Analytical chemistry can take you in directions you never expected--perhaps even across the "attosecond barrier".

Sandra Katzman, with contributions from Elizabeth Zubritsky

It's a mild winter afternoon on Kyushu Island (Japan), and graduate student Shuichi Kawasaki is hard at work on the near-IR laser that he is developing as a tunable source for thermal lens spectroscopy. He cranks up the laser's output power, hoping to obtain better data. The beam is supposed to be monochromatic; but once again, he notices bright, multicolored spots. His adviser, Totaro Imasaka of Kyushu University, previously chided him that these spots were the result of carelessness. But Kawasaki is not being sloppy, and his adviser will later realize it. Imasaka will also realize that his student has stumbled on a novel technique for generating ultrashort pulses that may revolutionize data transmission.

It's 1987. Communications companies worldwide know that new methods of data transmission will soon be needed, but efforts to generate ultrashort pulses are stymied in physics labs. None of this concerns Kawasaki, because his goal is to develop a laser for spectroscopy. But he does wonder why the strange, colorful spots have reappeared. The first time he saw them, they were faint. Now they are numerous and bright. He calls to Imasaka.

The beautiful sight--like beads of a rainbow in long straight lines--captivates adviser and student. Unlike the continuous spectrum produced by sunlight refracting through water drop lets, these are discrete spots, each a different color, arranged in rainbow order. The spots change color when the laser's wavelength is adjusted. The yellow is too intense to stare at. All of the spots twinkle, so the researchers call them "rainbow stars". In technical terms, they are "two-color stimulated Raman emissions" because the phenomenon requires two light beams of different colors.

The researchers see no immediate application for rainbow stars, but they are curious enough to investigate the phenomenon. They learn that rainbow stars require high power: The colorful spots appear only when the laser's wavelength is adjusted 6_7 nm away from maximum power. Thinking that rainbow stars are a nonlinear optical effect, the researchers expect the spots to become brighter when the laser is set to the maximum. Instead, the spots disappear.

Optical communications

In 2001, new methods of data transmission are still needed. Despite the widespread use of optical fibers and other technological advances, "the current capability of data transmission is almost at the maximum," says Imasaka. Data transmission rates of 10_20 GHz are possible right now. Future communications equipment will support 40- to 50-GHz transmission, but even that wonÅft suffice for long at the rate communications traffic is growing, so researchers continue to investigate new transmission methods. Japan wants to have a system that allows broadband data transmission as soon as possible, so the Japanese government is funding a three- to five-year initiative for developing new data transmission technology, Imasaka says. This is probably the largest national project in Japan, and it's just as important in many other countries, he adds. But a substantial improvement in transmission capacity won't be easy.

One requirement for a high transmission capacity is achieving a high repetition rate--many pulses per second, Imasaka says. The shorter the pulse width, the more possible pulses, and the higher the data transmission capacity. In theory, researchers could keep shortening the pulses indefinitely (assuming that they could develop suitable sources). That is why researchers want to break through the attosecond barrier. If they could generate a pulse in the attosecond range (less than 1 fs in duration), the capacity for data transmission would be enormous.

To date, pulse widths as short as 4.5 fs have been generated in the lab, but such short durations are not practical in current optical fibers because dispersion always occurs. Dispersion affects all pulses but is especially problematic for short ones; a 4.5-fs pulse, for example, might be stretched to 1 ps after passing through ~1 mm of optical fiber, Imasaka says.

Reliable pulse generation is also important for a high transmission capacity because a mistimed or missing pulse can cause transmission errors. Even in the best systems, regular pulse generation can be guaranteed only if the pulses are narrow and equally spaced and only for a limited duration, which is determined by the coherence length. The longer the coherence length, the more pulses can be sent without interruption.

As Imasaka and Kawasaki eventually discovered, rainbow stars can be useful for achieving high repetition rates and reliable pulse generation. Indeed, Imasaka believes that rainbow stars can meet both criteria better than other approaches. If he is correct, data transmission capacities could skyrocket to 17 THz.

Made possible by mistake

Despite the apparent utility of rainbow stars, they exist only by chance. Imasaka arrived at that conclusion after his exultation at seeing them subsided. He believes that rainbow stars are the accidental byproducts of attempts to reduce an undesirable laser effect called amplified spontaneous emission (ASE).

There is no easy way to get rid of ASE, Imasaka says. It tends to increase if you focus the pump beam tightly to the dye solution of the dye laser. However, you can't leave the pump beam too loosely focused, because that reduces the efficiency from the pump laser to the dye laser. In practice, this means that a user must judiciously adjust the laser every day. To combat the problem, one manufacturer designed a laser in which the diffraction grating directed the ASE away from the normal optical path. However, a prism reflected the ASE so that it could interact with the oscillating beam. Such a situation does not necessarily produce rainbow stars. But Imasaka began to suspect that this particular case satisfied the criteria for four-wave Raman mixing, a nonlinear optical phenomenon in which three photons interact to produce a fourth photon. In this case, two photons from the oscillating beam excite hydrogen--the Raman medium in this system--to an imaginary level. (There isn't a true energy level here, but this is a useful way to think about the phenomenon, Imasaka says.) A third photon from the ASE stimulates emission of the fourth photon. The trick is that the oscillating beam and the ASE must be separated by exactly 587 cm_1, which is the rotational Raman shift frequency of hydrogen. As it happened, Kawasaki's unlikely adjustment of the laser 6_7 nm away from the gain maximum created exactly this separation.

Ultrashort pulses

To confirm the four-wave Raman mixing hypothesis, Imasaka uses two dye lasers, which he can adjust independently, to recreate and thoroughly characterize the phenomenon. When Imasaka gets white light from the system, he passes it through a prism and counts three series of monochromatic, equally spaced emission lines in the visible region. He soon recognizes these lines as the additional products of four-wave Raman mixing. They are part of a cascade of low- and high-frequency emissions spanning the UV_vis_IR region. Imasaka realizes that the emissions satisfy two critical conditions for reliable ultrashort pulse generation: They are equally spaced over a wide spectral range and have narrow spectral lines generated in phase, which would correspond to an ample coherence length. He writes a paper explaining how to generate this multicolor laser light and mentions the possibility of generating subfemtosecond pulses (1).

The alternative

Other researchers are trying to generate ultrashort pulses with different methods such as the better-known high-order harmonics. In this approach, the laser beam is focused into a nonlinear medium that doubles, triples, or further multiplies the laser frequency, Imasaka explains. The generated laser emissions are referred to as high-order harmonics. For example, when a high-power Ti:sapphire laser is focused into an argon gas, a laser emitting at 800 nm is frequency-doubled (which means the new wavelength is 400 nm), frequency-tripled (266 nm), and so on down to the vacuum UV region. High-order harmonics look promising for data transmission because the laser emission covers a wide spectral range. Another possible application is biophotonics, which uses ultrashort pulses to study the motion of electrons, for example.

Although no strategy is a clear winner yet, high-order harmonics may be taking the lead. A 1999 paper by N. A. Papadogiannis and colleagues at the Foundation for Research and Technology_Hellas (Greece) got researchers excited about the possibility of attosecond laser pulses (2). A commentary in Nature, for example, proclaimed that the researchers might have broken the attosecond barrier, although some questions remained about the researchersÅf measurement technique (3). However, Imasaka says that rainbow stars may have the advantage because high-order harmonics require very high power to generate pulses. Rainbow stars do not. In addition, it has been difficult--at least until now--to verify the widths of pulses generated by high-order harmonics (4). Finally, the pulse spacing in rainbow stars is determined by a molecular constant of hydrogen, whereas systems that generate pulses directly from lasers must contend with a laser's tendency to drift in response to changes in temperature and other environmental factors, he explains. In particular, such drift can reduce the system's coherence length, meaning that fewer pulses will be sent before an interruption occurs. Nevertheless, he acknowledges, "A lot of research must still be done in our case."

A new global standard

The search for rainbow stars reminds Imasaka of the elusive double rainbow, but the benefit he envisions is very much in the moment. "I'm interested in making a global standard... a data clock," he says."Usually we use quartz--100 MHz for computers [and] 50 GHz for data communication [at] the research level." But an increase of at least 100 times is needed. Several approaches have the potential to boost the rate into the terahertz range, but he notes that a strategy based on rainbow stars would be the only method that is tied to a molecular constant.

In talking with other researchers, Imasaka has discovered that at least one other team has seen the same multicolor emission. Those researchers simply suppressed the effect so that they could achieve better efficiency at the emission lines of interest. "It was only our pure and innocent joy in playing with a colorful beam," Imasaka says, that led to the discovery of this "mistake" and an understanding of how useful it might be.

References

1. Yoshikawa, S.; Imasaka, T. Opt. Commun. 1993, 96, 94_98.

2. Papadogiannis, N. A.; Witzel, B.; Kalpouzos, C.;

Charalambidis, D. Phys. Rev. Lett. 1999, 83, 4289_4292.

3. Corkum, P. Nature 2000, 403, 845_846.

4. Paul, P. M.; Toma, E. S.; Breger, P.; et al. Science 2001, 292, 1698_1692.

Sandra Katzman is a freelance writer in Japan. Elizabeth Zubritsky is an associate editor of Analytical Chemistry.

Copyright American Chemical Society 2001