The supercontinuum gains momentum

R. R. ALFANO

Intense ultrafast broadband “white-light” pulses spanning the ultraviolet to the near-infrared are called a supercontinuum. First demonstrated 35 years ago, the phenomenon arises from the nonlinear interaction and propagation of ultrafast pulses focused into a transparent material.1, 2 A supercontinuum can be generated in various states of matter-condensed media (liquids and solids) and gases. The conversion of one color to white light is a dramatic and elegant effect (see Fig. 1). The supercontinuum pulses are intense, collimated, and coherent, and can be used in many diverse applications.

After its discovery in 1969, the supercontinuum (SC) found applications as a novel spectral tool in time-resolved pump-SC probe absorption and excitation spectroscopy to study the fundamental picosecond (10-12 s) and femtosecond (10-15 s) processes that occur in biology, chemistry, and solid-state physics, such as, for example, the primary events in photosynthesis and vision, the basic chemical steps in reactions and nonradiative processes in photo-excited molecules, and the underlying kinetics of how elementary excitations-optical phonons, polaritons, excitons, and carriers-behave and relax.

With the advent of the SC in microstructured fibers, there has been a rebirth and significant advancement in the SC field.3 Applications now include optical communication, optical coherence tomography, ultrashort pulse compression, broad-spectrum light detection and ranging (lidar), atmospheric science, frequency clocks, phase stabilization and control, timing, lighting control, and attosecond (10-18 s) pulse effects.

Supercontinuum from microstructures

Over the past several years, supercontinuum generation in microstructured photonic-crystal fibers by ultrashort-pulse propagation has become a subject of great interest worldwide. The main reasons are the low pulse energies (approximately 1 nJ) required to generate the SC, as well as its coherence and high brightness-making it the ideal white-light source. The most commonly used materials for SC generation include photonic-crystal fibers (PCFs), birefringent PCFs, tapered fibers, and hollow fiber arrays or “holey” fibers.

In microstructured fibers, the pulse duration of the ultrafast laser pump determines the operational mechanisms-for laser pulses below 10 ps, self-phase modulation and soliton generation dominate, while for pulses greater than 30 ps, stimulated Raman and four-wave mixing play a major role in extending the SC spectrum. Using 100-fs pulses with energy in the low tens of nanojoules, the SC spectrum can span more than an optical octave bandwidth spread from 400 to 1500 nm. This is important in controlling the phase of the carrier wave inside the pulse envelope of the modelocked pulse train. Using the first- and second-harmonic waves in the SC, the carrier-envelope offset phase can be detected using heterodyne beating between the high-frequency end of the SC with the doubled low-end frequency of the SC in an interferometer. These phase-controlling effects are important for maintaining accuracy of frequency combs for clocking and timing in metrology, high-intensity atomic studies, and controlling attosecond pulse jitter.4,5

The increasing worldwide demand for large-capacity optical communication systems needs to incorporate the wavelength, time, and phase control between pulse envelope and carrier wave for ultrashort pulses (less than 100 fs). The ultrabroad bandwidth and ultrshort pulses of the SC may be the enabling technology to produce a cost-effective super-dense wavelength-division multiplexer (more than 1000 wavelengths) and time-division multiplexer for the future terabits/s to pentabits/s communication systems and networks. Major advances in multiple-wavelength generation using the SC have been achieved by several Japanese groups.6

In addition, a femtosecond modelocked laser pulse train can produce a self-referencing phase-locked frequency comb. A phase-locked 50-fs SC pulse can be generated from a train of these phase-locked pulses using microstructured fibers. The use of two nonoverlapping intense femtosecond SC pulses separated in time by τ picoseconds can produce an interference pattern in the spectral domain at high modulation frequency (see Fig. 2). In this case, multiple-wavelength channels are produced for wavelength-division-multiplexing (WDM) applications in the 50-GHz range (see Fig. 3).7

Optical coherence tomography

The SC has been used to achieve ultra-high-resolution optical-coherence-tomography (OCT) imaging in biomedical areas. The images are formed from the ballistic light backscatter from refractive-index changes inside tissues, such as the colon, arteries, and eyes. The axial resolution of the OCT images is dependent on the bandwidth of the illumination source. Since the SC is much broader than even a superluminescent diode, it can achieve better resolution. Currently, longitudinal resolutions of 2.5 µm in air and approximately 2 µm in tissue have been achieved using the SC in OCT applications.8 Because the transmission zone for deep penetration of tissue is from 700 to 1300 nm, femtosecond laser sources ideal for central wavelengths in SC generation for OCT imaging in tissue are Ti:sapphire (centered at 800 nm), Nd:glass (centered at 1060 nm), and chromium-doped forsterite (Cr4+:forsterite; centered at 1300 nm) in microstructured fibers.

Ultrafast pulse compression

The SC can be used to generate ultrashort pulses using pulse-compression techniques. Pulses as short as 3.8 fs with energies up to 15 µJ have been generated using two cascaded hollow fibers with adaptive pulse compression.9 The SC spans from 530 to 1000 nm. At sub-10 fs the pulses must be reverse-chirped to compensate for time-broadening effects in materials such as optical lenses, prisms, plates, and coated dielectric mirrors, including air itself, to improve the time resolution of pumping. With spectrally wider SC pulses, compression down to 1 fs and possibly even into attosecond regions may be possible over the next decade using shorter wavelength pulses, on the order of 200 nm.

Atmospheric remote science

Recently, ionization tracks have been obtained by propagation of intense ultrafast laser pulses in air that self-guide in the form of channels and produce SC “white-light” filaments over distances greater than 20 m.10 The spatial stabilization of filaments is attributed to the balance between defocusing processes of diffraction and Kerr self-focusing from the ionized plasma created in the air by the ultrafast laser pulse. An intense ultrashort laser pulse of about 10 to 350 mJ of about 70-fs duration self-focuses to about 80 µm to produce the SC in air in the form of long filaments. The production of well-defined ionized trails in air at a distance has potential use in controlling lightning around structures and airports.

The SC pulses can be transmitted through clouds for free-space wireless optical communications, for remote sensing of pollutants and aerosols, and for range finding. The use of differential SC absorption lidar for remote sensing of atmospheric species has been shown to be a viable experimental technique for the detection and identification of a wide range of molecular constituents. A multiwavelength source with built-in calibrated power spectrum provides the possibility of a more accurate determination of pollutants. The differential spectral reflectance error can be reduced significantly by using three or more frequencies. Furthermore, Mie scattering and Rayleigh scattering, which contribute to the background extinction coefficient, can be factored out, and the tail of the water vapor absorption contribution can be expanded around the molecular absorption line. Another advantage of the SC is the virtual elimination of the time interval (one nanosecond) between the different probing wavelengths. In these measurements, this interval should be less than 10-2 s to eliminate temporal fluctuations in atmospheric parameters.

Metrology

One of the newest uses of the SC in conjunction with a modelocked laser is in metrology to achieve better, more accurate clocks and timing,11 and possibly even determine the spatial variation of gravity. State-of-the-art standards, based on atoms, ions, and molecules, exhibit excellent stability to achieve ultrahigh reproducibility and accuracy for clocks. A new concept being explored is based on “femtosecond optical-frequency comb generation” over an octave bandwidth with regularly spaced sharp lines at well-defined frequencies. The pulses from a stabilized modelocked laser are ideal for a “frequency comb.” The mode spectrum of the femtosecond laser with the SC can generate optical pulses spanning over two octaves in bandwidth. Optical heterodyning of a portion of the SC has produced stability for frequency measurement with fractional frequency noise on the order of 6 × 10-16 in one second of averaging over the 300 THz bandwidth.12 This results in an ideal source for accurate frequency measurements, timing in optical computers, and possibly in detection of differences in gravity.

One of the most dramatic and elegant effects in optical physics, the SC has become an important nonlinear optical process in science and engineering since the advent of shorter and more stable ultrafast laser pulses. Thirty-five years after its discovery, the SC continues to be a phenomenon with applications yet to be discovered.

ACKNOWLEDGMENT

I would like to thank Lauren Gohara and Kestutis Sutkus from The City College of New York for their help in the preparation of this article. Additional details can be found in Reference 3.

REFERENCES

1. R. R. Alfano and S. L. Shapiro, Phys. Rev. Lett. 24, 592 (1970).
2. R. R. Alfano and S. L. Shapiro, Phys. Rev. Lett. 24, 584 (1970).
3. R. R. Alfano, The Supercontinuum Laser Source, Springer-Verlag, New York, NY (1989), second edit. (2005).
4. S.T. Cundiff and J. Ye, Rev. Modern Physics 75, 325 (2003).
5. F. W. Helbing et al., IEEE JSTQE 9, 1030 (2003).
6. H. Takara et al., Elect. Lett. 39, 1078 (2003).
7. I. Zeylikovich et al., J. Optical Soc. of America B, accepted for publication in 2005.
8. I. Hartl et al., Optics Lett. 26, 608 (2001).
9. B. Schenkel et al., Optics Lett. 28, 1987 (2003).
10. J. Kasparian et al., Science 301, 61 (2003).
11. Th. Udem et al., Nature 416, 233 (2002).
12. S. A. Diddams et al., IEEE Trans. Instr. Meas. 50, 552 (2001).

R. R. ALFANO is distinguished professor of science and engineering at the Institute for Ultrafast Spectroscopy and Lasers in the Department of Physics at The City College of The City University of New York, New York, NY 10031; e-mail: [email protected].