Improvement in Data Acquisition for a Step-Scan Fourier
Transform Spectrometer
THOU-LONG CHIN and KING-CHUEN LIN *
Department of Chemistry, National Taiwan University, and Institute of Atomic and M olecular Sciences, Academia Sinica, Taipei 106, Taiwan, Republic of China
We have designed a controller circuit to be incorporated into a comm ercial step-scan Fourier transform spectrom eter (Bruker IFS 88 FTS). Alternatively, time-resolved data m ay be acquired with a transient digitizer provided by the user. W e have dem onstrated the versatility of the controller with a Lecroy 9450A transient digitizer as the reco rder in the time-resolved em ission detection of the CH radical populated in the A2D state and in the absorption m easure-ment of the ambient air. The controller, which may be conveniently coupled to any data acquisition system, makes the use of the FTS more ¯ exible and less expensive.
Index Headings: Step-scan Fourier transform spectrom eter;
Inter-ferogram; Data acquisition; CH emission spectru m.
INTRODUCTION
Since the ® rst step-scan Fourier transform spectrom-eter (FT S) from Bruker (IFS 88 Model) was commer-cialized in 1987, the instrument has been increasingly used to study dynamic phenomena.1±3 With the advent of
the step-scan FTS, it becomes possible for time-resolved FT-infrared (FT-IR) to achieve time resolution ranging from nanoseconds to milliseconds. In contrast to the con-ventional rapid-scan method, the moving mirror in the step-scan FTS is moved step by step rather than in a continuous fashion. Feedback from the reference HeNe laser interference pattern is used to control the mirror position rather than the mirror velocity. In the operation of the step-scan technique, while the mirror is moved to a ® xed position, a repeatable phenomenum is initiated and the time evolution of the event is recorded at that position. The procedure is repeated at the next mirror position. After the moving mirror has completed a whole run of retardation positions, the collected interferograms corresponding to different time slices are Fourier trans-formed to yield time-resolved spectra.1±3 The subsequent
time resolution essentially depends only on the signal strength, the detection sensitivity, and the response speed of the detector and the data acquisition electronics.
Thus far, the step-scan FTS has been widely applied in different research such as investigations of the reori-entation of polymers and liquid crystals, associated with mechanical or electric ® eld perturbation,1,4 ±6and dynamic
behavior of a ligand binding to a transition metal element or a heme.1 Gerwert and co-workers employed step-scan
FT-IR difference spectroscopy to study membrane protein bacteriorhodopsin and resolved for the ® rst time the bi-phasic rise of bacteriorhodopsin’ s intermediate in the in-frared spectral region.7 Yang investigated the
effective-ness of various metal catalysts in anaerobic composition
Received 16 June 1998; accepted 21 September 1998. * Author to whom correspondence should be sent.
as a function of time.8 Photoacoustic (PA) signal
mea-surement performed with an FT-IR spectrometer has been widely accepted. In the conventional rapid-scan mode, however, a problem arises in the interpretation of the depth information from PA measurem ents.1,9 With
step-scan FT-IR, Palmer and co-workers demonstrated the ca-pability of mapping depth pro® les from different layers in terms of different phase characteristics between these substances.1,10 Later, Budsvska and Manning found that
tim e-r esolved im p ulse pho to acoustic sp ectra m ay straightfor wardly provide depth-pro® ling inform ation for solid samples.9 The step-scan technique has also been
ex-tended to the study of gas-phase kinetics and dynamics. Time-resolved FT emissions or absorptions in the infra-red or visible range have been measuinfra-red in order to look into the photofragmentation process, single collision re-actions, energy transfer events, and dynamics and kinetics of radical±radical reactions initiated by a photolysis
la-ser.2,3,11±16 The research activities involving the step-scan
FTS continue to increase.
The data acquisition system in the step-scan FTS from Bruker (IFS 88 Model) depends mainly on two transient recorder boards. One is a 200 kHz, 16 bit analog-to-dig-ital converter (ADC) and the other is a 200 MHz, 8 bit PAD 82 board.17 For events in the microsecond time
do-main, the tiny signal may be effectively acquired by using the 16 bit ADC recorder. However, as the phenomena are short in the nanosecond region, the provision of the 8 bit, 1 V range PAD 82 recorder is sometimes insuf® cient to record a trace signal without distortion. When the strength of the analog signal is less than 1/256 voltage, the response is essentially lost in the analog-to-digital conversion. Such a small signal may often be found in retardation positions away from the centerburst of an in-terferogram. As a result, the spectral resolution is sacri-® ced, since the interference pattern beyond that position contributes insigni® cantly. Another disadvantage for the current Bruker IFS 88 is that the generated interferogram in the step-scan mode cannot be retrieved again. The FTS operated in the step-scan mode may generate much more data than in the rapid-scan mode, since an additional time coordinate is included. It becomes a heavy burden to save all the interferograms. Nevertheless, these data may be manipulated to produce different outcomes. They are also useful in examining the quality of the spectra prior to Fourier transform ation.
This work is intended to provide a controller circuit that will enable users to select their own data acquisition system as an alternative to the present commercial one. This circuitry can retrieve the interference signal in each mirror position in a way that is synchronous with the
FIG. 1. Schematic diagram of controller circuit. The notations are described in the text.
commercial time-resolved spectrometer (T RS). The data are acquired alternatively with a Lecroy 9450A transient digitizer or any other available recorder. The signal pro-cessing can be made more ¯ exible. The Fourier trans-form ed spectra are even better in terms of the spectral resolution and the signal-to-noise ratio than is the case with the current TRS provided in the IFS 88 apparatus. With the above controller and recorder, two experiments are demonstrated. One is the time-resolved FT emission spectrum of the CH radical populated in the A(2D ) state,
and the other is the absorption spectrum of ambient air.
EXP ERIM ENTAL
Design of Controller. In the operation of the
step-scan mode of the IFS 88 instrument, when the moving mirror is in the process of moving, a TKDA pin (on a CDP board) sends a negative pulse with a 20 ms duration, causing the system to stop taking data.17 Thus the action
triggering the transient recorder for data acquisition is ignored. In contrast, when the mirror has been moved to a ® xed position and stabilized, the TKDA sends out a 0.1 ms pulse, notifying the system to start taking data. There-fore, monitoring the TKDA signals is an effective way to determine the state of the step-scan FTS. The control-ler is designed on the basis of this concept.
As illustrated in Fig. 1, the circuitry is assembled with 8253 (program able timer/counter) and 8255 (program able peripheral interface) chips, which are mounted on the PC I/O bo ard , N O R g ate (7 4L S0 2) , D - ty pe ¯ ip -¯ o p (74LS374), transistor (9013), and solid-state relay (SSR). The total cost of the controller is less than $100. The operation procedure follows. The TKDA negative pulse is ® rst received from the IFS 88 instrument and then in-verted to a positive pulse with a NOR gate (74LS02).
The pulse and a clock generated from a function gener-ator are input to an 8253 chip, which is set to mode 5, the so-called hardware triggered strobe.18The function of
the 8253 chip is to recognize the status of the moving mirror. When the 20 ms TKDA pulse is met, a clock inside the 8253 chip begins to count down from n to 0, then gives a pulse (or clock) within the 20 ms period. The n value should be large enough so that the counting action will not ® nish during the duration of the 0.1 ms TKDA pulse. In this manner, the 8253 chip may ignore the signal of the 0.1 ms pulse but respond with a pulse (or clock) when the 20 ms pulse is encountered. The pulse sent from the 8253 chip can be inverted through the NOR gate, since outport pB0 of the 8255 chip is at logic 0. The 8255 chip is set at mode 0. That is, the ports pA, pB, and pC may be used as inport and outport in-dependently.18 If the pB1 is at logic 1, then the outport
Q1 of the D-type ¯ ip-¯ op (74LS374), while activated by the input clock, responds with logic 1. The output voltage is then ampli® ed by a transistor (9013) to an extent that is suf® cient to turn off the SSR. Accordingly, the output of a Si diode fails to simultaneously activate the transient digitizer (L ecroy 9450A) provided externally and the TRS recorders in the IFS 88. The Si diode is used to monitor the laser pulse (or other external perturbers), ser ving as a triggering source for the transient digitizers to take data. A polling line connected to the outport Q1 is fed back to the 8255 chip. Once the polling line re-ceives the signal of logic 1, a command is sent to the host PC (provided by the users) to start to read and save the previous data left on the Lecroy 9450A. The data transfer is executed through an IEEE 488 interface. Right after completion of the data transfer, the outport pB1 changes the logic from 1 to 0, and then outport pB0
gen-FIG. 2. The time sequence of the controller operation.
FIG. 3. Experimental setup for time-reso lved FT emission of the CH radical populated in the A(2D ) state. The data may be acquired by PAD 82
and Lecroy 9450A recorders simultaneously through the controller device interfaced to the IFS 88 instrument.
erates a clock. As the ¯ ip-¯ op (74LS374) is activated again by the positive transition of the clock, the Q1 changes the output from logic 1 to 0. At that moment, the signal of logic 0 may turn on the SSR, and then a pulse from the Si diode simultaneously triggers the
Lec-roy 9450A and TRS recorder of the FTS. These recorders start to take data from the new position of a moving mirror as soon as it is stabilized. The data recording may be averaged over a number of laser shots. While the mir-ror is moving to the next step, the TKDA emits a 20 ms
FIG. 4. The interference waveforms from CH A(2D ) emission as a function of retardation position recorded with the Lecroy 9450A transient
digitizer. The waveform ranges in the time domain from 0 to 1780 ns with a temporal resolution of 2.5 ns.
FIG. 5. The interferogram corresponding to a particular time slice given in Fig. 4.
pulse again. The above procedures then repeat. For clar-ity, the resultant time sequence for the controller opera-tion is depicted in Fig. 2.
Experimental Procedure. Time-Resolved FT
Emis-sion. The controller device has been successfully applied
to the following two experiments. One is for the time-resolved FT emission study of the CH A(2D ) state, from
which the rates of rotational and vibrational energy trans-fer may be evaluated.16 The experimental apparatus is
shown in Fig. 3. The CH radical in the A(2D ) state was
produced following photolysis of CHBr3 at 266 nm. The
dissociating laser pulse came from a 10 Hz, 5±8 ns Nd: YAG laser (Spectra-Physics, GCR3) operating in the fourth harm onic. The precursor ¯ owed through the re-action cell, so that emission from the fresh CH radical
was monitored. The emission spectrum from 420.2 to 434.8 nm of the CH A(2D ) state was collected through a
pair of lenses of 2 in. ( f /1.5) and 6 in. ( f /4) focal lengths onto the entrance of a step-scan Fourier transform spec-trometer. A fraction of the laser beam was sent to irra-diate a rodamine 6G dye contained in a cuvette. The re-sulting ¯ uorescence was monitored with a Si diode, which was connected to the external controller, as de-scribed above. The emission of the CH A(2D ) state
through a Michelson interferometer was detected by a photomultiplier tube (RCA, 1P28A). The dc-coupled sig-nal was ampli® ed before being fed into the PAD 82 and Lecroy 9450A transient digitizers simultaneously. The re-sulting data from either the PAD 82 or Lecroy 9450A were treated with a separate PC.
FIG. 6. The time-resolved FT emission spectra of the CH A(2D ) state corresponding to three time slices given in Fig. 4.
FIG. 8. (a) The interferogram for ambient air absorption acquired in
step-scan mode with the use of the Lecroy 9450A recorder with the voltage gain adjusted appropriately. (b) The enlarged region for the retardation points from 2000 to 2200.
FIG. 7. The interferogram for am bient air absorption acquired in step-scan mode with the use of the Lecroy 9450A recorder with the voltage gain ® xed.
Time-Resolved FT Absorption. The controller was also
applied to the time-resolved FT-IR absorption experiment on the ambient air. A standard setup for step-scan FT-IR was employed. A globar source irradiated through the Michelson interferom eter, and the resulting interference pattern after passing through the ambient air in the cham-ber was detected with a fast-response MCT (Model No. KMPV11-1-J2, Kolmar Technologies). The triggering source of the Si diode was replaced with a function gen-erator. The procedure for the controller operation was the same as in the emission experiment.
RESULTS AND DISCUSSIO N
CH(A) Em ission. As shown in Fig. 4, the interference
waveforms from the CH A(2D ) emission were recorded
in different retardation positions with the use of the
Lec-roy 9450A. Each pro® le was averaged over 10 laser shots to improve the signal-to-noise ratio. An example of the interferogram corresponding to a particular time slice (120 ns time delay) is given in Fig. 5. Some of the re-sulting spectra after Fourier transformation are shown in Fig. 6. The spectral and time resolution were adjusted to be 10 cm2 1 and 2.5 ns, respectively.
The Lecroy 9450A employed in this work has a better voltage resolution than the PAD 82, although both re-corders have the same resolution of 8 bits. The Lecroy 9450A has a selectable range of 0 to 20 V and selectable gain of 1, 2, 4, 10, 20, 40, 100, 200, 400, or 5000. With a voltage of 0 to 20 V and a gain of 5000, the smallest detectable change in voltage is 1.6 3 102 5 V, which
means that the Lecroy 9450A is theoretically capable of recording the signals down to 1.6 3 102 5 V. In contrast,
FIG. 9. Fourier transformed spectrum of ambient air corresponding to
the interferogram in Fig. 8.
the PAD 82 recorder has a voltage range of 0 to 1 V and a ® xed gain of 1; the smallest detectable change is there-fore 3.93 102 3 V. Thus, the theoretical resolution of one
bit for Lecroy 9450A is about two orders of magnitude better than that for the PAD 82 recorder. Accordingly, the time-resolved FT spectra recorded by the Lecroy 9450A are expected to have better quality than results from the PAD 82 recorder.
The current IFS 88 model does not save the data of interferogram s in the step-scan mode. Once the interfer-ograms are Fourier transformed, these data are deleted from the memory. Thus one cannot retrieve the corre-sponding interferogram to inspect the data quality or to further manipulate the data. In addition, if the spectral resolution is increased, one has to reduce the full range of spectral wavelength to keep the memory space unsat-urated. In contrast, with our external controller incorpo-rating the Lecroy 9450A, the acquired interference wave-form at a particular retardation position can be succes-sively rem oved to the PC during the period when the mirror is moving to the next position. Therefore, the space restriction may be solved with provision of a large RAM and hard disk. Let us estimate the memory space for taking the data given in Figs. 4 ±6. The Lecroy 9450A provides a temporal resolution of 2.5 ns. The pro® le (Fig. 4) spreads in the time domain from 0 to 1710 ns, con-taining 684 sampling points. The resulting data may oc-cupy a memor y space of 2 kbyte in the Lecroy 9450A. Each pro® le, averaged over 10 laser shots, takes only 5 ms to transfer to the host PC with a transmission rate of 400 kbyte s2 1. Given 1138 as the total retardation
posi-tions, corresponding to a spectral resolution 10 cm2 1, the
storage space needs to be about 2 Mbyte. Such space may easily be found in our PC. In comparison, we also esti-mate the memory space for the IFS 88 to accommodate the interference waveforms in Fig. 4. According to the form ula given in the OPUS manual, the evaluated mem-ory space corresponds to 7.8 Mbyte.17 This value is far
beyond the space allowance of 4 Mbyte RAM provided. On the other hand, because of no restrictions in the mem-ory space, a single Lecroy 9450A is able to measure the whole dynamic phenomena ranging from nanoseconds to milliseconds. Of course, that treatment is at the expense of more time spent in manipulating the data. To prune
the huge amount of data accumulated in the study of the reaction mechanism of membrane protein bacteriorho-dopsin, Ger wert and co-workers simultaneously em-ployed two transient recorders in a step-scan FT-IR dif-ference absorbance experiment.7 One 8 bit, 200 MHz
re-corder was used to measure the time domain from 100 ns to 20 m s, while the other 12 bit, 200 kHz recorder monitored the 5 m s to ; 50 ms time range.
Ambient Air Absorption. The ambient air absorption
experiment in the step-scan mode is demonstrated to ex-amine the capability of measuring a broad band using an 8 bit transient recorder. The interferogram for a broad-band spectrum usually causes a sharp contrast in signal strength between the centerburst and the position away from it. The use of an 8 bit transient recorder may some-times cause the trace signal acquired away from the cen-terburst to be ignored. An example is shown in Fig. 7. The interferogram for the ambient air was measured with the Lecroy 9450A with the voltage gain purposely ® xed. The weak signal in the ® gure was seriously distorted. In comparison, Fig. 8 shows the corresponding interfero-gram from the same recorder with the voltage gain ad-justed appropriately. The weak signal was enlarged by changing the gain. Therefore, the interference pattern even at retardation points that are far away may be rec-ognizable. The resulting Fourier transform ed spectrum (Fig. 9) may maintain the spectral resolution as expected in the initial condition. In contrast to the Lecroy 9450A measurement, the 8 bit, 200 MHz PAD 82 recorder lacks ¯ exibility in the adjustment of the voltage gain. The re-sulting dif® culty in recognizing the whole interference pattern may reduce the spectral resolution.
CONCLUSION
In conclusion, we have designed a circuit controller to be incorporated into a step-scan FTS (IFS 88). The time-resolved data may be acquired alternatively with a transient digitizer provided by the user. By taking advan-tage of the Lecroy 9450A as a recorder, we have dem-onstrated the versatility of the controller in measuring time-resolved FT emission and absorption spectra. The controller, which may feasibly be coupled to any data acquisition system, makes the use of the FTS more ¯ ex-ible and less expensive.
ACKNO WLEDG MENTS
The authors wish to thank J. F. Nien, D. K. Liu, and Y. P. Chen for help in setting up the emission experiment. This work is supported by the National Science Council and Petroleum Company of the Republic of China under Contract No. NSC87-2119-M-002-001.
1. R. A. Palmer, J. L. Chao, R. M. Dittmar, V. G. Gregoriou, and S. E. Plunkett, Appl. Spectrosc. 47, 1297 (1993).
2. G. V. Hartland, W. Xie, and H. L. Dai, Rev. Sci. Instrum. 63, 3261 (1992).
3. T. J. Johnson, A. Simon, J. M. Weil, and G. W. Harris, Appl. Spec-trosc. 47, 1376 (1993).
4. K. Masutani, H. Sugisawa, A. Yokota, Y. Furukawa, and M. Tas-umi, Appl. Spectrosc. 46, 560 (1992).
5. M. A. Czarnecki, B. Jordanov, S. Okretic, and H. W. Siesler, Appl. Spectrosc. 51, 1698 (1997).
6. H. Wang, R. A. Palmer, and C. J. Manning, Appl. Spectrosc. 51, 1245 (1997).
7. R. Rammelsberg, B. Hebling, H. Chorongiewski, and K. Ger wert, Appl. Spectrosc. 51, 558 (1997).
8. D. B. Yang, Appl. Spectrosc. 47, 1425 (1993).
9. B. O. Budevska and C. J. Manning, Appl. Spectrosc. 50, 939 (1996).
10. R. M. Dittmar, J. L. Chao, and R. A. Palmer, Appl. Spectrosc. 45, 1104 (1991).
11. S. A. Rogers and S. R. Leone, Appl. Spectrosc. 47, 1430 (1993). 12. S. R. Leone, Acc. Chem. Res. 22, 139 (1989).
13. D. E. Heard, R. A. Brownsword, D. G. Weston, and G. Hancock, Appl. Spectrosc. 47, 1438 (1993).
14. G. Hancock and D. E. Heard, in Advances in Photochemistry (John Wiley and Sons, New York, 1993), Vol. 18, pp. 1±65.
15. J. Eberhard, P. S. Yeh, and Y. P. Lee, J. Chem. Phys. 107, 6499 (1997).
16. C. C. Wang, T. L. Chin, and K. C. Lin, J. Chem. Phys. 107, 10348 (1997).
17. ``Step-Scan Technical Manual for IFS 66, IFS 66v and IFS 88’ ’ (Bruker, Karl Sruhe, Germany).
18. Intel Microsystem Components Handbook Vol. II (Intel, Folsom, California).
