2316
Anal. Chem. 1987, 59,2316-2320
collection speed would be commonly limited to 8 h, which will be followed by a fast measurement step, and total analysis time would be limited virtually by step 1. Since no complicated equipment and instruments would be kept occupied, sampling a t different sites or by different persons should be achieved in an inexpensive manner. Another advantage is the elimination of problems regarding breakthrough. In most solid sorbents, breakthrough behavior requires an effective control ( I ) , where in the cell proposed in this study, breakthrough occurs when all the path available is colored. Therefore, the analyst can easily observe breakthrough. Upon immediate observation of breakthrough, one may use shorter sampling times or flow rates or a less sensitive sampling cell with larger gas channels and/or higher Cd(I1) concentration on solid sorbent. Since dynamic range seems to be controlled by aforementioned variables, several cells with various sensitivities can be used at one site for improved dynamic range to avoid breakthrough, which would result in meaningless data. Further studies on this novel sampling device will help in better characterization of the process as well as in achieving better analytical performance. Registry No. CdCl,, 10108-64-2;H2S, 7783-06-4.
LITERATURE C I T E D (1) Melcher, Richard G.; Langner, Ralph R.; Kagel. Ronald 0. Am. I n d . Hyg. ASSOC. J . 1978, 3 9 , 349-361. ( 2 ) Vasireddy, Sivasankararao; Street, Kenneth W., Jr.; Mark, H. E., Jr. Anal. Chem. 1901, 53, 868-873. (3) Street, Kenneth W., Jr.; Mark, H. E., Jr.; Vasireddy, Sivasankararao; LaRue-Filio, Rebecca A.; Anderson, C. William; Fuller, Michael P.; Simon, Stephen J. Appl. Spectrosc. 1985, 3 9 , 68-72. (4) Wood, G. 0.; Anderson, R. G. Am. I n d , Hyg. Assoc. J . 1975, 3 6 , 538.
’On leave of absence from the Department of Chemistry, Middle East Technical University, Ankara, Turkey.
Rebecca LaRue 0. Yavuz Ataman’ Daniel P. H a u t m a n Gregory Gerhardt H a n s Zimmer H a r r y B. M a r k Jr.* Department of Chemistry and the Edison Sensor Technology Center University of Cincinnati Cincinnati, Ohio 45221
RECEIVED for review April 2, 1987. Accepted June 1, 1987. This research was supported in part by the Edison Sensor Technology Center.
AIDS FOR ANALYTICAL CHEMISTS Scintillation-Type Ion Detection for Inductively Coupled Plasma Mass Spectrometry Le-Qun Huang, S h i u h - J e n J i a n g , a n d R. S. Houk*
Ames Laboratory-US. Ames, Iowa 50011
Department of Energy and Department of Chemistry, Iowa State University,
Inductively coupled plasma mass spectrometry (ICP-MS) has become an important new technique for elemental and isotopic analysis (1-3). In all the ICP-MS instruments constructed to date, the ion signals are monitored by an electron multiplier, which is generally of the Channeltron variety ( 4 ) . Although good analytical performance can certainly be obtained with Channeltron electron multipliers, they do have several undesirable characteristics as detectors for ICP-MS. A Channeltron has a limited lifetime of approximately 1year under normal analytical use. The general use of pulse counting techniques, in which a high bias voltage is applied to the multiplier so that its gain is saturated, probably accelerates the rate of gain loss. The response of a Channeltron is linear up to count rates of approximately 1 x lo6 counts s-l; above this value, calibration curves tend to droop. Deviation from linear response a t high count rates particularly limits the concentration range over which either very large or very small isotope ratios may be determined (5) and can be a source of error when isotope dilution is employed for quantitation (6). Conceivably, the linear range could be improved by employing a segmented Channeltron that permits both analog current measurements and pulse counting a t the same time. There is some evidence that the detector gain can be degraded temporarily by scanning the mass analyzer across an intense peak, e.g., >lo7 counts s-l. If a peak for a trace analyte is monitored soon after the intense peak, this “fatigue”can affect the accuracy with which the trace constituent may be determined or limit the rate a t which the spectrum can be scanned (7). Although computer-controlled peak hopping routines can be devised to avoid exposing the detector to the
intense ion beam, this often requires prior knowledge of the major element composition of the sample and can be inconvenient. The scintillation device depicted by items J-L in Figure 1 is an alternate ion detector for mass spectrometry. Its operation is based upon the following sequence of events. Ions from the mass analyzer (G) strike the negative target (J)and generate secondary electrons. These electrons are accelerated from the target (ca. -5 kV) to a thin grounded metal film coated on the surface of a scintillator (K). The film thickness is chosen so that the secondary electrons can penetrate the film and deposit energy in the scintillator. Energy deposition promotes some of the scintillator molecules to excited electronic states. The excited molecules then emit photons at visible wavelengths that are sensed by the photomultiplier (L). This technique is often referred to as Daly detection after its inventor (8-10). A scintillation-type detector offers some potential advantages in the problem areas identified above for Channeltron electron multipliers. A t the very least, the gain of the scintillation detector should not deteriorate with time. In the present work, the analytical figures of merit of a scintillation-type detector are compared with those of a Channeltron for ICP-MS. EXPERIMENTAL SECTION The sampling interface and MS part of the apparatus are depicted in Figure 1. Instrumental components and operating conditions are identified in Table I. ICP-MS Apparatus. The basic features of the ultrasonic nebulizer, ICP, sampling interface, ion optics, mass filter, and vacuum system have been described previously; pertinent al-
0003-2700/87/0359-2316$01.50/0 0 1987 American Chemlcal Society
ANALYTICAL CHEMISTRY, VOL. 59, NO. 18, SEPTEMBER 15, 1987
Eiiik
2317
Table I. Instrumental Facilities
rn
m component
U
0
5cm
Figure 1. Schematic diagram of MS (ICP not shown): (A) sampler, skimmer, (C) port to rotary pump, (D) ion lens (voltage settings listed), (E) photon baffle, (F) port to diffusion pump (1600 L si),(0) quadrupole mass analyzer, (H) port to diffusion pump (800 L Si), (I) exit lens (voltage settings listed), (J) AI target (-5 kV), (K) AI-coated scintillator, (L) PMT, (M) turbomolecular pump (150 L Si). (B)
terations are noted below. The load coil was grounded at its downstream end to the shielding box as shown in Figure 2Y of ref 15. The maximum ion kinetic energy was lower (Le., approximately +10 V) and less strongly dependent on aerosol gas flow rate than was the case in ref 16, indicating that the residual discharge was less severe in the present work. A photon stop consisting of two metal cones (E, Figure 1) was added to the ion lens. These cones were kept at the same potential. Ions leaving the mass analyzer went through a stainless steel cylinder (2.5 cm long X 1.8cm inside diameter) with a wire grid at its downstream end. Separate potentials were applied to these ion optical elements to transmit ions into the detector. This assembly also served as the differential pumping orifice between the quadrupole and detector chambers. Scintillation Detector. The aluminum target was polished to a mirror finish. A power supply with a very high maximum voltage output (30 kV) was used to bias the target as suggested by previous studies with this detector (8). However, in the present work the background and background noise increased considerably as the magnitude of the target bias increased; the signal to noise ratio was optimum a t the relatively low target voltage shown in Table I. It is interesting to note that this optimum target bias (ca. -5 kV) was similar to the optimum voltage found for deflecting ions into a Channeltron (14). The scintillator was coated with A1 by glow discharge sputtering and vapor deposition. The A1 layer was approximately 20 nm thick and was grounded to the vacuum chamber. A scintillator material with a short decay time (2 ns) and an inexpensive photomultiplier tube (PMT)with a reasonably small pulse width (30 ns fwhm) were selected to minimize counting losses a t high count rates. The photomultiplier was kept in its own glass enclosure; i.e., it was not evacuated by the pump on the detector chamber. The rf leads to the mass filter did not pass through the detector chamber so that rf interference from the quadrupole power supply was not encountered. Other electrical leads and connections were also kept out of the detector chamber to prevent formation of an electrical discharge therein. The detector chamber was evacuated by a separate turbomolecular pump and was maintained at a slightly lower pressure than the chamber housing the mass analyzer (Table I). Ideally, the pressure in the detector chamber should be as low as possible because the background decreases and a higher target bias can be tolerated a t lower pressure. Data Acquisition. Data were obtained in three modes. Stability was evaluated by selected ion monitoring, Le., the mass analyzer transmitted only the m / z value of interest for the entire measurement cycle. Calibration curves and isotope ratios were determined in the multichannel scanning mode using the signal averager. The m / z range was restricted such that the full 4096 memory addresses spanned only the isotopes of interest. The dwell time was 100 ps address-' and 256 sweeps were averaged; this process took 105 s for each element. The individual peaks in the averaged spectra were integrated by summing the total counts in the range of memory addresses corresponding to each peak (3,5). The background was determined in each m/z interval
operating conditions, materials, or dimensions
nebulization continuous flow ultrasonic nebulizer (11)
desolvation temperature 80 "C solution uptake rate 2.5 mL min-'
ICP generator and torch (12)
operating conditions described in ref 13
ion extraction interface: Ames Laboratory construction
sampling position on center 10 mm from load coal sampling orifice 0.7 mm diameter skimmer orifice 1.2 mm diameter sampler-skimmer separation 8 mm
turbomolecular pump on detector chamber Model TMP/NT 150 Leybold-Heraeus
operating pressures (see Figure 1): expansion chamber 1 Torr second chamber -1 X
mass analyzer Model 270-9 with 012-15 rf head Extranuclear Laboratories (now Extrel)
mean rod bias -1 V dc
Torr
quadrupole chamber 2 X lo4 Torr detector chamber 1 X lo* Torr
scintillator Model NE104 EM1 PMT: Model 9924B EM1
PMT bias voltage, 800 V target bias voltage, 5000 V output pulse width 30 ns (fwhm)
counting electronics: maximum count rate Model 1763 capability 20 MHz pulse width 40 ns preamplifier TTL output Model 1762 amplifier-discriminator Photochemical Research Associates data acquisition Model 1170 signal averager Nicolet, Inc. LSI 11/23 based minicomputer Digital Equipment Corp.
maximum count rate capability 20 MHz
software and interfaces produced at Ames Laboratory (14)
of interest by analysis of an appropriate blank solution. The background was then subtracted from the gross integrated signal before the isotope ratio was calculated. Detection limits were determined by adding 10 scans across each major isotope in 10 s each scan. The peak height for each element was measured from this averaged spectrum, and the background level and standard deviation were determined by analyzing a blank solution and averaging the counts in adjacent m / z addresses spanning the center of each analyte peak. The detection limit represented the solution concentration necessary to yield a net peak height equivalent to 3 times the standard deviation of the background. Solutions and Standards. The solvent was 1% HN03 in distilled deionized H,O. Standard solutions were 1mg L-' of each element unless noted otherwise and were prepared by diluting aliquots of commercial stock solutions (Fisher).
RESULTS A N D DISCUSSION General Observations. One reason for evaluating the scintillation detector was the expectation that it would not
2318
ANALYTICAL CHEMISTRY, VOL. 59, NO. 18, SEPTEMBER 15, 1987
Table 11. Isotope Ratio Determinations element
solution concn, mg L-I
isotopes
cu Zn
1
3
63/65 64/66 64/67 64/68 85/87 140/142 160/162 1611162 1631162 1641162 2051203
Rb Ce
1 1
DY
2
T1
1
isotope ratios determined natural
re1 std dev, % determined” counting statistics
2.24 1.76 11.9 2.63 2.60 7.97 0.0899 0.74 0.987 1.10 2.39
2.15 1.73 9.76 2.59 2.45 7.66 0.0973 0.734 0.996 1.13 2.44
0.4 0.7 3.2 0.8 0.4
0.37 0.64 1.81 0.79 0.2 0.56 1.94 0.48 0.43 0.41 0.29
0.5
0.9 0.9 0.8 0.7 0.3
Determined from seven to eight seDarate measurements of each isotope ratio. Table 111. Ratio Measurements for Selected Gd Isotopes Gd isotoDe ratios for m l z solution 1 mg L-l Gd 1 mg L-I Gd
+ 15 mg L-’ Tb
natural
155/160 determined
RSD,” 70
natural
156/160 determined
0.673 0.673
0.836 0.834
2.3 1.7
0.934
1.07
0.934
1.06
RSD” 1.6 1.2
Relative standard deviation of five to six determinations.
respond at all to photons from the ICP. Therefore, we thought the detector could be positioned directly behind the exit lens of the mass analyzw, and no optical baffle would be necessary in the ion lens. This optical baffle is definitely a source of ion loss (14,17) and possibly is also a cause of m / z discrimination. However, the background was high and erratic without a photon baffle, so one was inserted into the ion lens (Figure 1). With the photon baffle present, the background was typically 500 f 25 counts s-l, using the scintillator detector. This latter level of background and standard deviation were similar to those obtained in earlier studies with a Channeltron detector with the photon baffle present. Possible reasons for the high background with the photon stop absent included ionization of excited neutral species (e.g., metastable Ar atoms) under the influence of the target voltage, a discharge caused by a high local density of species in the supersonic beam traveling between the target and the PMT, or emission of photoelectrons from bombardment of the target by plasma photons, particularly those in the vacuum ultraviolet region (18, 19). Precision of Isotope Ratio Measurements. Isotope ratios were determined repetitively for several elements. The results are listed in Table 11. The primary figure of merit in isotope ratio measurements is precision. With the scintillation detector the precision for each isotope ratio was better by a factor of 2 than that typically obtained with this ICP-MS device using a Channeltron detector. For example, in previous work with a Channeltron, isotope ratios near unity (Le., in the range 0.5-3) could be determined with a relative standard deviation of 1 to 2 % (14,20,21) compared to 0.3-1% using the scintillation detector (Table 11) for ion count rates and data acquisition times that were comparable for the two detectors. The precision obtained with the scintillator either was comparable to that expected from counting statistics or was worse than the counting statistics limit by up to a factor of 2 (e.g., Rb and Dy in Table 11). For comparable measurement times (- 100 s) with Channeltron detectors, the precision for isotope ratio measurements was generally worse than the counting statistics limit by a larger factor (typically 2 or 3) (22). Thus, at least for this ICP-MS device, the scintillator detector offered significant improvement in precision for isotope ratio determinations. Comparison of the determined ratios with the
expected natural abundance ratios indicated some bias; the extent of bias was similar to that seen with a Channeltron and was probably due to mass discrimination by the mass filter. The magnitude of the bias was not consistent from element to element in Table I1 because the resolution of the mass spectrometer was optimized separately for each element to minimize bias and to provide proper peak shapes. Detector “Fatigue”. The following experiment was performed to determine whether the scintillator detector suffered any gain loss while scanning over an intense peak. The mass analyzer was scanned in the range m J z 154-161. Gadolinium isotope ratios (mlz 155, 156, and 160) were measured for a sample solution containing only Gd; these data are shown in the first row of Table 111. Next, Gd isotope ratios were determined for a solution containing the same Gd concentration but with T b present at 15 mg L-l. Thus, the detector sensed the massive 159Tbpeak (- 1 x lo7 counts s-l) immediately before the ‘@Gd peak on each scan. As shown in the second row of Table 111,the Gd isotope ratios were unaffected by the presence of Tb. The Gd isotope ratio was also independent of the sweep rate even at the fastest rate used (5 p s address-’, 20 ms per full sweep), which was the fastest the mass analyzer could be swept with adequate resolution between T b and Gd peaks. This indicated that the scintillation detector did not lose gain by scanning across the large ‘Yb+ peak. The precision of the Gd isotope ratios was similar in the presence or absence of Tb. In this experiment, the resolution of the mass analyzer had to be increased (relative to that used for the isotope ratios in Table 11)to properly separate the Gd and T b peaks. Operation a t higher resolution yielded (a) lower ion count rates, (b) more sharply pointed peaks that were less reproducible from scan to scan, and (c) more extensive discrimination against ions a t higher mlz. For these reasons, the accuracy and precision of the Gd isotope ratios were poorer than those for other elements with comparable ratios (e.g., Zn and Dy) in Table 11. Because the point of this experiment was to look for gain loss, no particular effort was expended to minimize the bias or optimize the precision of the determined isotope ratios. The resistance of the scintillation detector to long-term gain degradation was illustrated by the following observations. As a first step toward optimization, the plasma conditions, sam-
ANAL.YTICAL CHEMISTRY, VOL. 59, NO. 18, SEPTEMBER 15, 1987
2319
Table IV. Detection Limits (30)and Sensitivities
lXio7 , 4
i
element 52Cr 59cO
63Cu 64Zn 4
i 7
003
-
I
1
I
01
03
I
3
IO
CONCENTRATION i r n g L-')
Flgure 2. Calibration curve for "Vb'. extrapolation of the linear section.
The dotted line represents the
pling position, ion lens voltages, etc. were adjusted to yield ' O maximum signals for the major background ions (e.g., , H20+, Ar+, and ArH+). In this mode, the ion current was detected by analog means and the spectrum was displayed continuously on an oscilloscope. With a Channeltron, permanent gain loss occurred in 10 h or less unless the detector bias voltage was lowered from approximately -3 kV (i.e., a value useful for pulse counting) to -2.2 to -2.5 kV. In contrast, no loss of gain was apparent with the scintillator if the major ions were monitored a t the same target voltage and PMT bias voltage as would be used for pulse counting detection. I t is clear that the scintillation detector is more resistant to gain degradation than the Channeltron, although there may well be limits to the former's tolerance to intense ion beams not elucidated by the present work. L i n e a r Dynamic Range a n d Stability. A typical calibration curve obtained with the scintillation detector is shown in Figure 2. The curve was linear (correlation coefficient = 0.9998) up to a T b concentration that yielded 2 X lo6 counts s-l. With the same counting electronics, calibration curves obtained with a Channeltron begin to droop below the straight line at lower count rates (typically 1 X lo6counts s-') (22,23). Thus, the linear response for the scintillation detector extended to somewhat higher count rates than for a Channeltron. At higher count rates, the experimental curve (solid line, Figure 2) was below the linear extrapolation because of counting losses. For example, a t 3 X lo6 counts s-l, the experimental count rate was below the straight line by approximately 12%. This extent of counting loss corresponded to an overall dead time of (0.12)/(3 X lo6) = 40 ns (24), which agreed with the specifications for the P M T and counting electronics (Table I). Use of counting electronics and PMTs with still faster response times could further extend the linear range accessible with a scintillation detector. To evaluate the stability obtainable with the scintillation detector, the count rate at m / z 140 for a 0.5 mg L-' solution of Ce was measured every 2 s for 2 h. The relative standard deviation of these 3600 measurements was 5%. The average count rate for five adjacent measurements at the end of the 2-h period differed from that a t the beginning of the period by only -5% relative; Le., the sensitivity did not drift greatly over the 2-h period. With Channeltron detection, the sensitivity generally drifted by &lo% over 1 h with a relative standard deviation of 10%. With the scintillation detector, the figures of merit for stability were considerably better than we have achieved with Channeltron detection for this particular ICP-MS device; they also rival the stability performance of the present commercial instruments (25). The 5% relative standard deviation was still much worse than that expected from counting statistics
I5As
85Rb 88Sr 89Y 90zr
93Nb 98M~ 140Ce 14*Nd
15*Sm 153E~ ls9Tb le4Dy 205~1 aosPb
ionization energy, eV 6.77 7.86 7.73 9.39 9.81 4.18 5.70 6.38 6.84 6.88 7.10 5.47 5.49 5.63 5.67 5.85 5.95 6.11 7.42
sensitivity, counts per mg L-' norm to isotopic actual abundance 147000 170000 71000 26 000 28 000 124000 153000 240 000 85000 113000 38000 234 000 64000 82000 226 000 401 000 93000 126000 92000
175000 170000 103000 53 000 28 000 172000 185000 240 000 165000 113000 160000 264 000 236000 308000 434 000 401 000 330000 179000 178000
detection limit," pug L-' 1
0.9 2
6 5 1 1
0.6 2 1 4
0.6 2 2
0.7 0.4 2 1 2
aDetection limits quoted in terms of overall mass of element with mass analyzer monitoring isotope indicated. alone (approximately0.3% at these count rates). The stability performance of an ICP-MS device doubtless arises from the juxtaposition of various factors such as nebulizer fluctuations, deposition of material on the sampling interface and ion optics, and variability of the resolution and transmission characteristics of the mass analyzer with time. The present work indicates that the detector may contribute to drift and instability as well. Detection Limits a n d Sensitivities. These figures of merit are listed in Table IV. The sensitivities and detection limits obtained with the scintillation detector are similar to those obtained with this ICP-MS device with a Channeltron detector and with the photon stop (Figure 1) present. Naturally, the sensitivities described in Table IV are poorer (by factors of 5-10) than those obtained previously without a photon stop in this instrument (14, 17). Furthermore, the sensitivities in Table IV are comparable to those obtained with the commercial instruments; detection limits are poorer in the present work because the standard deviation of the background is higher. Although it was initially expected that the detection limits would be much better with the scintillation detector, this did not prove to be the case.
CONCLUSION The performance of the scintillator detector is either comparable or superior to that of a Channeltron electron multiplier with the particular ICP-MS instrument used. The initial cost of the scintillator detector is higher than that of the Channeltron, particularly if the scintillator requires an additional pump and vacuum chamber. However, the cost and inconvenience of replacing a dying Channeltron are not encountered with the scintillator, which compensates for the higher initial Cost of the latter detector. Additional improvements in the performance of the scintillator could be possible by further reduction of the background pressure or by offsetting both the target and P M T above or below the center line of the sampler, skimmer, and mass analyzer. ACKNOWLEDGMENT The authors thank C. Y. Ng and G. D. Flesch for advice and assistance in construction of the scintillator detector. Registry No. 52Cr,14092-98-9; 59C0,7440-48-4; 63Cu,1419184-5; 64Zn, 14378-32-6; 75As, 7440-38-2; 86Rb, 13982-12-2; 88Sr,
2320
ANALYTICAL CHEMISTRY, VOL. 59, NO. 18, SEPTEMBER 15, 1987
14119-10-9; 7440-65-5; wZr, 13982-15-5;93Nb,7440-03-1;%Mo, 14392-20-2;lace, 14191-73-2;14?Nd,14336-82-4;'%m, 14280-32-1; 1 5 3 E ~13982-02-0; , 159Tb,7440-27-9; 164Dy,13967-69-6; '05Tl, 14280-49-0;208Pb,13966-28-4;"Cu, 14119-06-3; %Zn,14378-33-7; 67Zn, 14378-34-8; @Zn, 14378-35-9; s7Rb, 13982-13-3; 14'-Ce, 14119-20-1;@ D l,y' 14683-25-1;161Dy,13967-68-5;163Dy,14391-36-7; 16*Dy,14834-85-6;'03Tl, 14280-48-9.
LITERATURE CITED (1) Douglas, D. J.; Houk, R. S.Prog. Anal. At. Spectrom. 1985, 8, 1-18. (2) Houk, R. S. Anal. Chem. 1886, 58, 97A-105A. (3) Gray, A. L. Spectrochim. Acta, Part8 1985, 4 0 8 , 1525-1537; 4 1 8 , 151-167. (4) Kurz, E. Am. Lab. (Faiffield, Conn .) 1979, I 7 (3), 67-82. (5) Chong, N. S.; Houk. R. S. Appl. Spectrosc. 1986, 4 7 , 66-74. (6) Paulsen, P. Pittsburgh Conference on Anawical Chemlstry and Applied Spectroscopy, Atlantic City, NJ, 1986; Paper No. 3. (7) Brown, R., personal communlcation, 1985. (8) Daly. N. R. Rev. Sci. Instrum. 1960, 3 1 , 264-267. (9) Gibbs, H. M ; Commins, E. D. Rev. Sci. Instrum. 1966, 37, 1385-1390. (10) Ridley, 6. W. Nuci. Instrum. Methods 1961, 1 4 , 231-236. (11) Bear, B. R.; Fassei, V. A. Spectrochim. Acta, Part 8 1986, 4 1 8 , 1089-1113. (12) Scott, R. H.; Fassel. V. A,; Kniseley, R. N.; Nixon, D. E. Anal. Chem. 1974. 4 6 , 75-80. (13) Jiang, S.-J.; Houk, R. S.Anal. Chem. 1986, 58, 1739-1743. (14) Olivares, J. A.; Houk, R. S.Anal. Chem. 1985, 5 7 , 2674-2679.
(15) Gray, A. L.; Houk, R. S.: Williams, J. G. J. Anal. At. Spectrom. 1987, 2. 13-20. (16) Oiivares, J. A.; Houk, R. S. Appl. Spectroc. 1985, 4 0 , 1070-1077. (17) Houk, R. S. I n Analyfical Chemistry in the Exploration, Mining and Processing of Materials; Butler, L. R. P., Ed.; Blackwell: Oxford, 1986; Chapter 3. (18) Houk, R. S.; Fassei, V. A.; LaFreniere, B. R. Appl. Spectrosc. 1986, 4 0 , 94-100. (19) LaFreniere, B. R.; Houk, R. S.;Fassel, V. A. Anal. Chem.. 1987, 59. 2276-2282. (20) Jiang, S.J.; Houk, R. S. Spectrochim. Acta, Part 8 1987, 428, 93.. 100. ... (21) Serfass. R. E.; Lindberg, G. L.; Olivares, J. A,; Houk, R. S. Proc. Soc. €XI). 8iol. Med.. in oress. (22) B&rn, A. W.; Fulford: J.; Douglas, D. J. Winter Conference on Plasma Soectrochem.. Kona. HI. 1986 Paoer No. 15. (23) Tan, S.H. Ph.D. Dissertation, Unive;sity of Alberta, Edmonton, Alberta, 1987. (24) Skoog, 0. A. Principles of Instrumental Analysis, 3rd ed.; Saunders College Publishing: Philadelphia, PA, 1985; p 510. (25) Thompson, J. J.; Houk, R. S. Appl. Spectrosc., in press.
RECEIVED for review December 16, 1986. Accepted June 8, 1987. The Ames Laboratory is operated for the U.S. Department of Energy by Iowa State University under Contract No. W-7405-Eng-82. This research was supported by the Director for Energy Research, Office of Basic Energy Sciences.
CORRECT1ONS Estimation of Absolute Analyte Number Densities in Atomic Emission, Absorption, and Fluorescence Using Line and Continuum Sources
M. J. Rutledge, B. W. Smith, J. D. Winefordner, and Nicolo Omenetto (Anal. Chem. 1987,59, 1794-1797). Nicolo Omenetto of CCR, European Community Center, Stabilimento di Ispra, Anal. Chem. Div., Bat. 29,21020 Ispra, Varese, Italy, was omitted in error from the list of authors of this manuscript.
Effect of Subambient Temperatures on Separation of Steroid Enantiomers by High-Performance Liquid Chromatography Sana U. Sheikh and Joseph C. Touchstone (Anal. Chem. 1987, 59, 1472-1473). In the title, the figure caption, and throughout the text, the word enantiomer should be replaced with the word diastereomer.
