Y. Nagasaki1, K. Masutani2,*, 1School of Science,
Abstract
Introduction Over the last decade various kinds of time-resolved infrared spectrometers have been developed and applied to studies in a variety of research fields.[1-3] In 1992 Masutani et al.[4] proposed a time-resolving method for a conventional Fourier-transform infrared (FT-IR) spectrophotometer which does not require synchronization between the timing for time resolving and that for the sampling of the analog-to-digital (A/D) converter. The system was named as an asynchronous time-resolving system.[4] This method has the following advantages; (i) the signal-proceeding assembly for time-resolving measurements can be attached to any kinds of commercial FT-IR spectrometers and (ii) in principle, there is no shortest limit in time for transient phenomena to be measured. Therefore, asynchronous time-resolved FT-IR spectroscopy has been extensively employed to investigate repetitive fast phenomena such as the electric-field induced reorientation dynamics of a number of ferroelectric liquid crystals (FLCs).[5-12] Recently, Masutani et al.[13] reported a novel multichannel asynchronous time-resolvig system with 32 time-resolving channels. This new system reduces the measuring time substantially, decreasing the possible damage of a sample by the stimulus. In addition, simultaneous time-resolving by the multichannel system improves the reliability of the data. The purpose of this article is to report the principle, performance tests, and application of this novel time-resolving technique. Principle of Asynchronous Time-Resolved Before we explain the multichannel asynchronous time-resolving system, we describe the principle of asynchronous time-resolved FT-IR spectroscopy by using a single channel asynchronyous time-resolved system.
Figure 1 shows a typical set up for the asynchronous time-resolved measurement.[4] Figure 2 depicts shapes of the signals in processes A-H in Figure1 in alphabetical order: The main feature of this method is the use of a low- pass filter placed between the gate circuit and the A/D converter. The role of the low-pass filter is explained Figure 3 in more detail.[4]
Figures 3(A) and (B) show an analog interferogram for radiation from the light source and a corresponding spectrum obtained by inversely Fourier-transforming the interferogram. When a discrete interferogram (C) obtained by sampling this analog interferogram by the gate circuit is inversely Fourier transformed, the spectrum obtained (D) consists of the original component and others which appears at position determined by multiplying the sampling frequency 1/t by integers. Therefore, when this discrete interferogram is passed through a low-pass filter with an appropriate cutoff frequency ƒc (E), the high-frequency components are eliminated (F). The output spectrum (F) is identical to the spectrum (B) for the analog interferogram before sampling. Consequently, the analog interferogram (G) is obtained from the low-pass filter. Although the above Fourier transformations are not actually done by the low-pass filter, it has the function, in effect, as described above.[4] For a more detailed explanation, the readers should refer to ref. 4. Outline of the Multichannel Figure 4 shows a typical setup of the novel multichannel asynchronous time-resolving system.[13] The time-resolving module of the system has 32 time-resolving units, each of which consists of a gate circuit, a low-pass filter, and an A/D converter. In this system, time resolving is performed by setting each gate circuit at a different time delay. The sample is excited by repetitive stimuli at constant intervals, which do not depend upon the sampling timing of the FT-IR spectrometer. The sample excitations result in transient transmission differences, and thus the analog interferogram is modulated at very high frequency because of the repetitive excitation of the sample. The low-pass filters eliminate high-frequency components from the discrete time-resolved signals generated by the gate circuit and to convert them to analog signals.[4] The analog signals thus obtained are processed by the A/D converter at the sampling rate of the FT-IR spectrophotometer itself, as shown by the trigger signal for the sampling in Figure 4, which has no relation to the gate timing for time resolution.[4]
It should be also noted that the multichannel asynchronous time-resolving system described in this article applies only to time-resolving measurement with a continuous-wave probe, such as a globar, and does not apply to the pump/probe asynchronous sampling technique, which was also proposed in the original paper in 1992.[4] Accuracy and Precision Tests The accuracy and precision tests for this system were carried out using static samples. For this purpose, a JEOL JIR-WINSPEC50 Fourier-transform infrared spectrophotometer equipped with a HgCdTe detector was employed, to which the multichannel asynchronous time-resolving module was attached.[13] For the accuracy test an infrared spectrum of a polystyrene film was measured on the above spectrophotometer in the conventional data collection mode. Using the same material, another spectrum was obtained using the interferogram processed by a single channel of the multi-channel asynchronous time-resolving system. The measuring conditions for A and B are described n in the figure captions of Figure 5. These spectra were very close to each other over the entire spectral range, confirming that the present system allows accurate measurements for both the broad and also the sharp absorption bands.[13]
Figure 5 demonstrates the accuracy on each channel.[13] Each spectrum in Figure 5 is a differential one calculated by subtracting the transmission spectrum of the first channel from those of the others. All the spectra are flat, below the noise level over the entire spectral range, and do not show any spurious peaks. Note that even the bands due to water vapor and carbon dioxide are also almost perfectly removed. It can be seen from Figure 5 that each channel yields an accurate spectrum. In addition, Figure 6 reveals that an unexpected spectrum arising from disturbances in the interferometer, sample compartment, and detector is almost equal in all the channels of the time-resolving system and can be eliminated from the resultant data. Therefore, the reliability of time-resolved data obtained with the present system is increased substantially compared to the existing system.[13]
The experimental data for the precision test are presented in Figure 6.[13] These data were measured several times in the same channel. Figure 6A shows the infrared transmission spectra of the polystyrene film and 6B exhibits the difference spectra calculated by subtracting the transmission spectrum of the reference channel from those of the sample channel.[13] The difference spectra were used to eliminate the influence of the disturbances that occurred in the FT-IR spectrophotometer. The transmission spectra of both channels were obtained simultaneously by using the same interferograms. For the reference and the sample channels, arbitrary ones were chosen. It is noted in Figure 6 B that all the spectra are flat, below the noise level over the entire wavenumber range, and do not have any unexpected peaks. The bands attributed to water vapor and carbon dioxide are not detectable. These results reveal that this multichannel asynchronous time-resolving system is extremely stable as well. Performance Tests of the In our previous paper as the performance test of the multichannel asynchronous time-resolving system, we reported a study of reorientation dynamics in a ferroelectric liquid crystal the reorientation induced by the reversal of the external electric field.[13]In this article another example is described. The sample investingated was a ferroelectric liquid crystal having a naphtharene ring.[14] The structure and phase sequence of the sample are shown in Figure7 . A cell made by Fujitsu Laboratories Ltd. was used for this test. It consists of a pair of BaF2 windows with a space between them of 1.7 mm. The inner surface of the windows was coated with indium tin oxide serving as an electrode and with polyimide rubbed in one direction. After the sample was heated to the region of the isotropic-liquid phase (110°C), it was injected into the cell fitted into a heating unit. The cell was cooled down slowly at the rate of 1°C /min to 60°C to obtain a homogeneously oriented liquid crystal in the region of Sc*.
A JEOL JIR-6500 FT-IR spectrophotometer equipped with an infrared microscope incorporating a HgCdTe detector and an infrared polarizer was used in this study. This FT-IR spectrophotometer is capable of acquiring interferograms in bi-directional scans. The infrared polarizer was set at an angle of 45° to the orientation direction of the liquid crystal to obtain the maximum absorption change. The time-resolving tests were carried out for the following two cases. First is the case for which a boxcar integrator was used as the gate circuit with the existing (singlechannel) asynchronous time-resolving method. IN the second the multichannel asynchronous time-resolving system was employed . To compare both systems, test data were collected under almost the same conditions except for the number of scans. Rectangular electric pulses with ±40 V peak voltages and 200 ms period were applied to the liquid crystal cell. The wavenumber resolution was 4 cm-1, the mirror speed of the interferometer was 1.6 mm/sec, and the frequency of sampling was about 5 kHz.
Figure 8 compares the time-resolved polarized spectra in the 3200-2600 cm-1 region of the liquid crystal for the two cases measured at a delay time of 40µs. The ordinate represents the absorbance change from the spectrum of the shortest time delay. For the spectrum A, obtained by use of the existing (single-channel) asynchronous time-resolving system, the number of scans of the interferometer for each time-resolved spectrum was 200, but the number of accumulation of interferograms was 400 because the sampling was carried using both forward and reverse scans of the interferometer. It took over 6 minutes to measure the time-resolved spectrum. For spectrum B, measured by use of the multichannel system, the numbers of scans and accumulation were 2000, respectively. Since the sampling function at both forward and reverse scans had not yet been developed for the multichannel system, the number of the scans and accumulations were the same. It took only about 3 min to measure the spectrum B. Note that the multichannel system provides much higher signal-to-noise ratios in shorter times.
Figure 9 depicts the absorption change in time-resolved polarized infrared spectra measured by the multi-channel system. Though the spectral regions of the bands of the water vapour and the carbon dioxide are not shown in Figure 9, the bands are almost eliminated without any further modification. It is of note that the changes in transients are very smooth. This article clearly shows that the multi-channel asynchronous time-resolving system has great advantages in measuring the time-resolved infrared spectra of repetitive fast phenomena. This system not only enables us to reduce the measuring time but also to acquire more reliable spectral data. References
Received 31st October 1999, accepted 24th November 1999
|
