The influence of changes in relative intensities of the hyperfine constituents of a line is clearly shown in Figure 8, A-C, which illustrates the effect of low, moderate, and high levels of self absorption on the profile of the thallium 377.6 nm line emitted by an EDL. As self-absorption increases, broadening of the line occurs, and the hyperfine components become of more equal intensity. These results should be compared to those observed with a Fabry-Perot interferometer for this line by Kirkbright and Sargent (26) using various sources. Profiles such as these show the importance of studies of hyperfine structure, particularly for atomic fluorescence spectrometry. Even at high levels of self-absorption, appreciable intensity will exist under the corresponding thallium profile. Figure 8, D and E, shows scans of the profiles of the 451.1 and 410.2 nm lines emitted by an indium EDL. I n this case too, a high level of self-absorption does not preclude the application of an indium EDL in AFS.
Table 111. Half Intensity Line Widths of EDL Lines Cadmium, 228.8 nm - Thallium, 535.0 nm Power, LineFidth, Power, Linewidth, watts A watts A 30 0.181 so 0.055 50 0.211 75 0.055 75 0.216 100 0.059 100 0.202 125 0.061 (Spectral slit width, 0.031A) (Spectral slit width, 0.025 A) Zinc, 213.9 nm Power, watts Linewidth. A 50
0.020
75
0.021 0,022 0.025
100
125
(Spectral slit width, 0.012 A)
ACKNOWLEDGMENT
of spectral line sources in analysis on a qualitative and semiquantitative basis. Figure 7, A-D, illustrates the effect of applied microwave power on the measured profile of the mercury resonance line at 253.7 nm emitted by an EDL. Evidence of hyperfine structure was visible at lower applied powers, but the relative intensity of the hyperfine components varied considerably at times for no apparent reason. Reversal is clearly indicated at 50 watts; the influence of this reversal and severe broadening at high applied powers on absorption sensitivity and on the shapes of fluorescence calibration curves is well established. The asymmetry of the profiles at high applied power is characteristic of collisional broadening and indicates that the assumption that Doppler broadening is the predominant feature of the spectral line profile from a n EDL is not always valid, especially at high values of applied microwave povter. Similar studies of the thallium 535.0 nm, cadmium 228.8 nm, and zinc 213.9 nm lines emitted by EDL's are shown in Figure 7, E-G, and in Table 111. The half intensity linewidths are uncorrected for instrumental broadening and. except in the case of thallium 535.0 nm, are considerably widened by the instrumental slit function. However, the trends shown should still be valid and indicate that, particularly for cadmium 228.8 nm and zinc 213.9 nm, there is very little line broadening, No self-reversal was observed at any time. It should be emphasized. however, that even minor variations in lamp manufacture, technique, and operating conditions can C'dUSe very substantial changes in the spectral characteristics of an individual EDL and it should not, therefore, be concluded, for example, that the cadmium 228.8 nm resonance line emitted by any cadmium EDL is never reversed under any conditions, although none was observed in this study.
We thank Michael Cichetti for the construction of some of the EDL's used in this study. RECEIVED for review October 19, 1970. Resubmitted May 30, 1972. Accepted September 5 , 1972. We thank SpectraMetrics, Inc., Burlington, Mass., for a research grant for the investigation of analytical techniques involving the use of the echelle spectrometer. -
-_
--
(26) G. F. Kirkbright and M. Sargent, Spectroclrirn. Acta. 2513, 577 (1970).
CORRECTION Analytical Study of the Phosphorescence of Purines in Aqueous Solution at 7 7 0K In this paper by J. J. Aaron and J. D. Winefordner [ANAL. CHEM., 44, 2127 (1972)], four lifetime values have been given by error in Table I and they should be corrected as follows: 0.09 sec (6-Methylmercaptopurine); 0.12 sec (6-Chloropurine); 0.0055 sec (6-Bromopurine) and 0.15 sec (2-Amino6-methylmercaptopurine). These values have been measured with the pulsed source, time-resolved phosphorimeter recently described [R. P. Fisher and J. D. Winefordner, ANAL.CHEM.,44, 948 (1972)l. Also, footnote / I of Table I should be omitted.
ANALYTICAL CHEMISTRY, VOL. 45, NO. 1, JANUARY 1973
117