"Many commercially available instruments are able to measure spectra on a timescale of tens to hundreds of milliseconds," said Peter R. Griffiths in his introductory remarks. Griffiths, a professor of chemistry at the University of Idaho, was an organizer of the symposium. If researchers need to record data on an even faster timescale, he added, then it can be done using special methods, but those require repeating a process or an experiment over and over again while measuring spectra. That approach isn't suited to probing events that occur only once or once in a while.

"In this session, we're discussing the most recent developments in instrumentation and application of time-resolved spectroscopy on the millisecond timescale for nonrepeatable events," Griffiths stressed. Such instruments can be used to study transient phenomena, chemical dynamics, and other processes that take place on short timescales.

	"It's hard to move things quickly in a perfect fashion."
	Manning

Christopher Manning, president of Manning Applied Technology, reported on advances in interferometer designs used in fast-scanning Fourier transform infrared (FTIR) spectrometers. At the heart of FTIR instruments is an interferometer that modulates radiation by splitting a beam of light from a source and then recombining the two beams in such a way that intensity variations in the combined beam are related to differences in the paths traveled by the two "halves." The path-length differences are commonly controlled by reflecting half of the light from a fixed mirror and the other half from a moving mirror, as is done, for example, in instruments based on the 100-year-old Michelson interferometer design.

Michelson interferometers work quite well, Manning noted, but they require precise motion of the moving mirror. They also require the moving mirror to remain perfectly perpendicular to the light source. Small imperfections in the alignment of the interferometer components detract from the device's optical performance and ultimately lead to photometric instability, weak spectral signals, and poor signal-to-noise ratios. And although instrument manufacturers have worked out various solutions to satisfy the stringent requirements, the common commercial designs may not perform well in rapid-scanning measurements.

"It's hard to move things quickly in a perfect fashion," Manning remarked. And Michelson interferometers don't tolerate imperfections in motion very well. For that reason, Manning and other instrument makers have designed so-called tilt-compensating interferometers, which do tolerate small tilts, wobbles, and other imperfections in the moving portion of interferometers.

AN OPTICAL ELEMENT known as a cube-corner reflector plays a key part in many tilt-compensating designs. Due to their forgiving optical properties, these types of reflectors can become slightly misaligned or tilted during rapid scanning, for example, and still continue to reflect radiation in a way that does not degrade an interferometer's ability to recombine light beams precisely and reproducibly. According to Manning, these same properties are responsible for the curious observation that at any viewing angle between an observer and one of these special reflectors, the observer's eye appears right in the center of the cube corner.

One type of tilt-compensating interferometer designed by Manning uses a rotating wedge-shaped disc-mirror to provide the back and forth motion of the moving mirror found in common FTIR instruments. Because the mirror is wedge-shaped and not uniform in thickness, as the disc is rotated, the mirror surface moves toward and away from the light source. In Manning's spectrometer, the disc rotates as fast as 30,000 rpm, enabling the instrument to record an IR spectrum in 1 millisecond.

Manning acknowledged that although spectrometers based on disc-mirror interferometers can be used to collect data very quickly, the design suffers from low throughput--meaning only a small fraction of IR light from a source reaches the detector. For many types of laboratory experiments, a low-throughput interferometer is adequate, he noted. But for applications characterized by weak signals--such as remote sensing, in which a spectrometer is used to measure IR emission far away from the source--low throughput leads to poor signal-to-noise ratios, and other types of instrument designs may be better suited to the job.

One such instrument described by Manning makes use of a nutating-prism interferometer, a design that features a sixfold increase in throughput compared with the disc-mirror instrument. In the higher throughput device, a prism is mounted on a rotation axis that's not quite perpendicular to the light source so that the prism rotates and wobbles simultaneously (nutates)--similarly to a poorly mounted automobile tire. The interferometer design can be used to record up to several hundred scans per second, Manning remarked.

DEMONSTRATING the capabilities of the nutating-prism design, Manning showed results from test measurements made with the company's recently built prototype spectrometer. In one simple test, a polystyrene film was scanned roughly 30 times per second. To the eye, the 30 or so overlaid spectra appeared sharp--demonstrating scan-to-scan reproducibility.

High-speed spectrometers capable of discerning weak signals, such as Manning's nutating-prism instrument, have use in military applications. These types of spectrometers can be used to scan a distant hillside, for example, in search of IR emission patterns that are the telltale signs of combustion of certain explosive materials, or to detect chemical warfare agents.

To test the prototype instrument's usefulness for these types of applications, Manning used a sample holder to position explosive powders in front of the spectrometer's optics and burned them while recording their IR emission spectra. The tests showed that the short bursts of radiation emitted during combustion were readily recorded by the rapid-scanning instrument and could be used to distinguish among various types of energetic materials. For example, the spectrum of burning "flash powder," a sulfurous material, is characterized by vibrationally excited SO₂ bands. Manning showed that flash powder is easily distinguished from the nitrocellulose-nitroglycerin mixture known as "double base" and from other common explosive materials.

Moving the discussion to laboratory applications of fast-scanning spectrometers, Griffiths reported on work conducted by his research group using a disc-mirror instrument built by Manning, a former postdoctoral associate of his.

Griffiths explained that a key shortcoming of Michelson interferometers that makes them unsuited to very fast scanning is the reciprocating (back and forth) motion of the moving mirror. For moderate scan speeds, the classic interferometer design works well, he remarked. But as the moving mirror is scanned more and more quickly, "the momentum causes the time required to reverse the mirror direction and get it back up to speed for the next scan to increase," Griffiths said. The outcome is a lot of dead time during which data cannot be recorded. For that reason, the Idaho group uses rotating-mirror instruments to conduct time-resolved studies of short-timescale processes.

As an example of the type of experiment that can be conducted with fast rotating-mirror spectrometers, Griffiths discussed a study in which films of polyethylene terephthalate (PET) were examined while the material was stretched until it fractured. Christian Pellerin, a former visiting scientist in Griffiths' lab, probed trans and gauche conformational changes in the OCH₂CH₂O segment of PET molecules during the half-second experiment by monitoring the intensities of key vibrational bands.

Based on spectral analysis, Griffiths reported that as the films are stretched, the fraction of molecules in the gauche conformation decreases and the intensity of the band corresponding to the trans configuration increases, even though the film is thinning. Eventually, nearly all of the molecules adopt the trans conformation and roughly 100 milliseconds later, the film snaps.

TIME-DEPENDENT methods were also used by postdoctoral associate Husheng Yang and Benjamin A. Weinstock, a graduate student in Griffiths' group, to study adsorption of low-molecular-weight aldehydes on bare silica and silica modified with amino propyl-silyl groups. The work is used to evaluate materials for use in cigarette filters, in which pyrolysis gases interact with adsorbents on a millisecond timescale.

By delivering pulses of gas to the filter media while recording IR spectra and changes in the gas cell pressure, the researchers found that formaldehyde is adsorbed more efficiently on the modified silica gel than on bare silica. Just the opposite was observed with acetaldehyde, Griffiths noted. The results were used to develop a kinetic model of the adsorption process.

A radically different sort of IR spectrometer with microsecond scanning capability was described by John F. Rabolt, a professor of materials science and engineering at the University of Delaware. Working with D. Bruce Chase, a senior research fellow at DuPont, Rabolt's research group designed a spectrometer that uses a diffraction grating--as found in pre-Fourier transform instruments--but has no moving parts. Rather than scanning the grating to record a spectrum one wavelength (region) at a time, all wavelengths are focused on a large array of tiny photodetectors simultaneously. The spectral wavelengths are determined from the position of the detector elements.

"THE DESIGN gets rid of moving parts--making it rugged and useful outside of the laboratory," Rabolt said. He added that planar-array spectrometers could be useful, for example, in online chemical processing, homeland security, and other applications. Thus far, however, the Delaware group has focused on in-lab measurements to test the new design.

In one test, Rabolt and coworkers examined the growth of monolayer films of trichlorosilanes on glass. By monitoring frequency and intensity changes in CH₂ stretching bands, the group was able to discern conformational changes indicative of molecular ordering as the layers assembled. The group also found that hexane is a better solvent for deposition than benzene or toluene (Langmuir, published online March 11, http://dx.doi.org/10.1021/la0208878).

Rabolt pointed out that the detector, which is a 256 x 320 array of InSb dots, registers signals in roughly 50 microseconds, but the readout step takes some 17 milliseconds. The holdup is due to the relatively slow readout electronics, but commercially available electronics are constantly improving, he said. Now the group is working on design modifications that include other gratings and detectors in order to broaden the technique's spectral range. The researchers are also developing methods to probe film deposition in real time.

Griffiths summed up the symposium's purpose by emphasizing that with continuing development of fast instrumental techniques, "researchers will be able to derive time-resolved information about short-lived processes that cannot be repeated." Rabolt echoed that view, noting that whereas he used to say "science drives technology," nowadays he says it the other way around. "As we develop new tools, we can do experiments that we couldn't even conceive of just a few years ago."

Chemical & Engineering News
Copyright © 2003 American Chemical Society