Spectral discrimination reveals band structures to identify analytes, but accessing spectral information has been challenging, as spectroscopists have been forced to balance the parameters of spectral range, sensitivity, resolution and acquisition time. Now scientists at JILA in Boulder, Colo., a joint institute of the National Institute of Standards and Technology and the University of Colorado, have developed a cavity ringdown technique that optimizes all of these features. Spectrometers usually must balance spectral range and resolution, sensitivity and acquisition time, trading low performance in one area for high performance in another. By matching a femtosecond laser frequency comb to the modes of a high-quality optical cavity, researchers at the National Institute of Standards and Technology have optimized all at once. This three-dimensional plot shows the high-resolution spectrum of water vapor over a 15-nm interval, acquired in a few microseconds. Courtesy of Jun Ye. Reprinted with permission from Science. ©2006 AAAS. Cavity ringdown spectroscopy involves the coupling of optical radiation into an optical cavity formed by two highly reflective mirrors. The radiation reflects between the mirrors thousands of times before exiting the cavity, so the intensity grows while it is incident and gradually decays when the radiation source is turned off. A gaseous sample in the cavity absorbs photons, changing the buildup power and the ringdown times at specific frequencies. Analyzing the exiting radiation identifies the missing photons, revealing the presence of the trace gas. Monochromatic radiation provides a single spectral response, while a broad incident spectrum carries detailed information about the sample. Coupling broad-spectrum radiation into a cavity, however, has been problematic. A cavity has well-defined modes, and a source that couples well over a few modes will be mismatched elsewhere. The new technique, developed by Jun Ye and colleagues at JILA, matches every component of an optical frequency comb of a femtosecond pulsed laser to a corresponding resonant mode of the cavity over the spectral bandwidth of the laser. (The pulse train of high-repetition-rate lasers can be regarded as many frequency components added together.) Ye’s team translates and tilts the laser cavity mirrors to match the frequency comb to the cavity. Low- or negative-dispersion mirrors, developed in conjunction with Advanced Thin Films LLC of Longmont, Colo., optimize the match for each sample. The pulses then add coherently across the 100-nm laser bandwidth, with the radiation in each cavity mode providing an independent spectral detection channel. The proof-of-concept device consists of a homemade 380-MHz, 10-fs-pulse, mode-locked Ti:sapphire laser, which is coupled into a cavity formed by mirrors with >99.9 percent reflectivity from 790 to 850 nm. A 0.25-m monochromator from CVI Laser LLC of Albuquerque, N.M., sends the output of the cavity to a scanning mirror that deflects the beam in the vertical direction and that directs it to a streak camera or to a 340-pixel-wide CCD for detection. Spectrally dispersed wavelengths are recorded horizontally, and the ringdown waveform in the time domain is recorded vertically, forming a three-dimensional spectral profile of the sample. A set of measurements of gas samples confirmed the researchers’ hopes: A 1-s acquisition time provides a sensitivity of 10–8 and a resolution of 0.01 cm–1 over a 100-nm wavelength range. An improved version of the system is under development. Lower-dispersion, more-broadband mirrors will improve the sensitivity, resolution and bandwidth. An improved spectral analytical instrument will further enhance the resolution, and a diode-array detector could reduce the acquisition time to nearly the few microseconds inherent in the cavity ringdown. Ye is particularly intrigued by the parallel information processing involved. “We suddenly have, in principle, a million detection channels, all working for us at the same time.” Science, March 17, 2006, pp. 1595-1599.