Flashlamp Continuum AAS Time Resolved Spectra* HELMUT BECKER-ROSS STEFAN FLOREK REINHARD TISCHENDORF AND GISELA R. SCHMECHER Institut fiir Spektrochemie und angewandte Spektroskopie Laboratorium fiir spektroskopische Methoden der Umweltanalytik Rudower Chaussee 5 D-12489 Berlin Germany For flashlamp continuum atomic absorption spectrometry (FLACAAS) a continuum spectrum of high intensity throughout the entire UV-VIS spectral range is provided typically by xenon flashlamps. These lamps offer high intensity in the UV which is however strongly fluctuating. This drawback may be compensated for by using the simultaneously recorded spectral vicinity of the absorption line for correction. In order to accomplish this type of measurement a high resolution spectrometer and a multipixel photodetector are required.This paper describes such a combination of an Cchelle spectrometer and a linear-array charge coupled device (CCD) detector with a graphite furnace atomizer and a xenon flashlamp for AAS-measurements at a repetition rate of 10 Hz. The determination of thallium (absorption line 276.787 nm) in natural seawater in the presence of palladium (interfering line 276.309 nm) used as modifier demonstrates the effective background correction and the possibility for sensitive analysis of complex samples by time-resolved continuum AAS. The characteristic mass of thallium was determined to be 28 pg the limit of detection in seawater amounts to 20 pg I-'. Keywords Continuum-source atomic absorption spectrometry; xenon-Jlashlamp; kchelle spectrometer; charge coupled device detector; thallium Continuum-source atomic absorption spectrometry (AAS) has several advantages compared with conventional AAS,1-3 especially the coincident measurement of the background absorption the expanded dynamic range and the possibility of simultaneous measurements of several elements making it an attractive method for the analysis of complex samples.On the other hand the instrumentation used for continuum-source AAS must fulfil some special requirements a source emitting a continuum spectra of high intensity throughout the entire range from 190nm to 850nm; a spectrometer providing a resolution as high as R>50000 within this spectral range; a multipixel photodetector for simultaneous registration of one or several small spectral intervals each partial spectrum com- prising an absorption line with its spectral vicinity; and data collection and processing equipment for evaluating the recorded partial spectra and for determinating the background corrected absorbance. The above mentioned advantages can be exploited with the flash lamp continuum AAS (FLACAAS) developed in our laborat~ry.~?~ Xenon flashlamps provide the continuum spectra of high intensity that are needed.In the short-wavelength UV range between 190 and 250nm flashlamps offer higher emis- sion intensities in comparison to other continuum sources.6 However the repeatability of the spectral intensities from flash- to-flash is not good and intensity fluctuations of up to 30% can be encountered. This drawback may be compensated for * Presented at the XXVIII Colloquium Spectroscopium Internationale (CSI) York UK June 29-July 4 1993.Journal of Analytical Atomic Spectrometry by using the simultaneously recorded spectral neighbourhood of the absorption line for correction. In order to observe the temporal formation of absorption signals we measured sequences of 20 to 50 flashes at a repetition rate of about 10 Hz. EXPERIMENTAL Our FLACAAS device (Fig. 1) consists of the following compo- nents which were developed and built in our laboratory unless otherwise stated power supply and ignition triggering unit for the Xe-flas hlamp; Xe-flas hlamp modified for end-on o bser- vation (type QXA 18 Q-ARC Cambridge UK); AAS graphite furnace atomizer and autosampler (AAS-3 with EA-3 Carl Zeiss Jena); echelle spectrometer in tetrahedral mounting; linear-array charge-coupled device (CCD) detector 5 12 pixel pixel size 23 pnx480 pm (CCD L 172 Werk fur Fernschelektronic Berlin Germany); and computer-aided CCD controller.An Cchelle grating (75 grooves mm-' blaze angle 64") and a quartz prism (suprasil prism angle 25") are mounted within the Cchelle spectrometer in such a way that the directions of dispersion are standing perpendicular to each other. The resulting two-dimensional spectrum is focused by a spherical mirror (focal length 500 mm focal ratio f/5) into a focal area with a size of about 60 mm x 60 mm. The spectrum consists of 97 spectral orders (850 nm corresponds to order number 28 190nm corresponds to order number 125) where the length of the orders is varying with the wavelength (from 14mm to 64mm) and the width of the orders is determined by the slit height of the entrance slit (slit height 0.2 mm slit width 0.02mm).To record a selected partial spectrum the linear- array CCD detector is moved to the corresponding spectral position in the focal area by a suitable mechanical device. This way we are able to achieve sequential measurements of different elements. These measurements could also be performed simul- taneously if such a large area CCD image detector for the recording of the whole spectrum was available. As an example of the necessity for effective background correction we measured thallium in the presence of a high concentration of palladium. When measuring thallium and palladium at 276.787 and 276.309 nm respectively the CCD array covers a 1.74nm wide partial spectrum in the 87th Cchelle order resulting in a spectral bandwidth per pixel of about 3.3pm.The atomization step of the graphite furnace atomizer triggers the flash lamp to operate at lOHz with a partial spectrum recorded for each flash. RESULTS The partial spectra of the individual flashes keep their spectral shape within a spectral range as small as 1.74nm in spite of the fact that their intensities change from flash to flash. Therefore the spectra can be made practically congruent by multiplication with a numerical factor which depends on the actual flash intensity. Thus it is possible to correct for the flash-to-flash intensity variations as well as the differences of Journal of Analytical Atomic Spectrometry January 1995 Vol.10 61/ / Capacitor charging supply . and trigger - J I Xenon flashlamp polychromator CCD linear array 1 furnace monochromator Hollow cathode lamp I Temperature controller Fig. 1 Scheme of the FLACAAS device used in connection with the AAS-3 spectrometer sensitivity of the individual pixels of the array. These correc- /T--. tions are carried out by use of the average of the first 12 spectra measured before the absorption appears thus obtaining the so called intensity corrected absorption spectra shown in Fig. 2. Finally the absorbance spectra are baseline corrected by using the mean value of the spectral absorbance within intervals at both sides of the analytical line. These intervals have a width of 33 pm (10 pixels) and are located at a distance of 17 pm ( 5 pixels) from the line centre.The residual fluctu- ations of the absorbance spectra are predominantly due to shot noise as can be clearly seen from the averaged spectra in Fig. 3. Fig. 2 Intensity corrected absorption spectra for the determination of T1 in natural seawater diluted in de-ionized water by a factor of 20 and with 20 pg Pd as modifier time sequence of 34 partial spectra recorded with a repetition rate of 10Hz; and wavelength section of the recorded partial spectra (pixel range No. 101-412 wavelength range 276.085-277.155 nm respectively; spectral bandwidth 3.3 pm per pixel; pixel No. 166 Pd 276.309 nm; pixel No. 310 T1 276.787 nm) 8 N a N Fig. 3 Baseline corrected absorbance spectra in the spectral vicinity of T1 276.787 nm time sequence of 19 partial spectra recorded with a repetition rate of 10 Hz; wavelength section of the recorded partial spectra (pixel range No.290-330; spectral bandwidth 3.3 pm per pixel; and sample 20 pl of 1 + 19 diluted seawater (salinity 0.18%) prepared by adding T1 standard for a 4pg 1-' concentration with 20pg Pd as modifier. (a) Single sample and (b) average from a series of 9 equivalent samples 62 Journal of Analytical Atomic Spectrometry January 1995 Vol. 100.10 3.3 / .3 3.3 Fig. 4 Baseline corrected absorbance spectra for different T1 concen- trations in the vicinity of T1 276.787 nm time sequence of 34 partial spectra recorded with a repetition rate of 10 Hz; wavelength section of the recorded partial spectra (pixel range No. 290-330; spectral bandwidth 3.3 pm per pixel); and sample 20 pl of 1 + 19 diluted sea- water (salinity 0.18%) prepared by adding T1 standards with 20 pg Pd as modifier.TI concentrations (a) 5 (b) 15 and (c) 25 pg 1-' In order to prove the effectiveness of the background correction we determined thallium in natural seawater the salinity of which was 3.5%. The seawater was diluted by a factor of 20 in deionized water and from this solution samples with different thallium concentrations were prepared by adding suitable thallium standards. A 20 pl portion of Pd (NO,) (Merck Ultrapure) Solution (1 g 1-l in 3% HNO,) equivalent 0- 5 10 15 20 25 30 Con cent ra t ion/pg I - ' Fig. 5 Calibration curve for T1 in diluted seawater using standard additions sample 20 pl of 1 + 19 diluted seawater (salinity 0.18%) prepared by adding T1 standard with 20 pg Pd as modifier.Absorbance data are temporal peak heights recorded by the pixel at the centre of T1 276.787 nm absorption line to 20 pg of Palladium was used as modifier and added to 20 pl of sample seawater in the furnace for thermal stabilization of thallium. The whole sample was dried ( 100°C 40 s) ashed (9OO0C 15 s) and atomized (full power 2700 "C 2 s) in the conventional way. In Figs. 2 and 4 the intensity corrected and baseline corrected absorption spectra of thallium are shown in the presence of the strong absorption of the neighbouring palladium line. From the calibration curve in Fig. 5 the characteristic mass of 28 pg/0.0044 A was determined. For comparison the characteristic mass for the determination of thallium for a commercial line source ETAAS spectrometer with wall atomization (Perkin Elmer HGA 6000) is given as 15 pg/0.0044 A. The calculation of the detection limit is based on the spectral noise of the absorbance spectra in the vicinity of the absorption line which is comparable to the time dependent baseline noise of the line source AAS if the used spectral vicinity is free from structured background absorption.The used vicinity around the thallium line includes 133 pm (40 pixel) with the exception of the central range of 56pm (17 pixel). Taking 3 times the standard deviation of the spectral absorbance in the vicinity of the absorption line the detection limit of thallium was determined to be 20pg absolute and 20 pg 1-1 for the determination of thallium in seawater.CONCLUSIONS The baseline corrected absorbance spectra shown in Fig. 3 and Fig. 4 as well as the calibration curve in Fig. 5 demonstrate that time-resolved continuum AAS can successfully be employed for sensitive analysis of complex samples. In spite of their intensity fluctuations flashlamps can be used provided the spectral vicinity of an absorption line is included in the evaluation. Further studies are being carried out comparing FLACAAS background correction with other methods. The authors wish to thank Dr T. Florek for his help in the evaluation and presentation of the spectra. The financial support by the Senatsverwaltung fur Wissenschaft und Forschung des Landes Berlin and the Bundesministerium fur Forschung und Technologie is gratefully acknowledged. REFERENCES 1 2 O'Haver T. C and Messmann J. D. Prog. anal. Spectrosc. 1986 9 483. Moulton G. P O'Haver T. C and Hardy J. M. J. Anal. At. Spectrom. 1990 5 145. Journal of Analytical Atomic Spectrometry January 1995 Vol. 10 633 Jones B. T. Mignardi M. A. Smith B. W. and Winfordner J. D. 4 Schmidt K. P. Becker-Ross H. and Florek S. Spectrochim. Acta Part B 1990 45 1203. 5 Becker-Ross H. Florek S. and Schmidt K. P. in 6. CAS 1991 ed. Welz B. Bodenseewerk Perkin-Elmer fiberlingen 1991 6 Gavrilov V. E and Gavrilova T. V. Opt. Spektrosk. 1991,70,511. Paper 4/05 389C Received September 5 1994 Accepted October 17 1994 J. Anal. At Spectrom. 1989 4 647. pp. 497-504. 64 Journal of Analytical Atomic Spectrometry January 1995 Kd. 10