Precise Measurement of Ion Ratios in Solid Samples Using Laser Ablation With a Twin Quadrupole Inductively Coupled Plasma Mass Spectrometer LLOYD A. ALLEN, JAMES J. LEACH, HO-MING PANG AND R. S. HOUK* Ames L aboratory±US Department of Energy, Department of Chemistry, Iowa State University, Ames, IA 50011, USA Laser ablation (LA) is used with steel samples to assess the mode, a precision of 0.04% RSD was reported for the measureability of a twin quadrupole inductively coupled plasma mass ment of Pb and Mg isotope ratios.spectrometer to eliminate Øicker noise. Isotopic and internal A great deal of work has also focused on alternative mass standard ratios are measured in the Ærst part of this work. spectrometers for elemental analysis. Hieftje and co-workers4±8 Results indicate that Øicker noise cancels and signiÆcant have used a time of Øight (TOF) mass analyser for the very improvements in precision are possible. The isotope ratio rapid detection of an entire mass spectrum.This system has 52Cr+553Cr+ can be measured with a relative standard been used in both GDMS5,6 and ICP-MS4,7,8 experiments. deviation (RSD) of 0.06±0.1%, depending on the dwell time The best ratio precisions reported for this ICP-TOF-MS device and averaging method used. The level of noise above the shot are 0.46% RSD for steady state signals during solution nebuliznoise limit is greater in internal standard measurements than ation7 and 1.6% RSD for transient signals generated by LA when doing isotope ratio measurements.Nevertheless, RSDs with single shots.8 Koppenaal and co-workers9,10 introduced improve from 1.9% in the Cr+ signal to 0.12% for the ratio of an ion trap MS instrument in which ions from an ICP are 51V+ to 52Cr+ in a steel standard reference material (SRM). trapped and later detected. In the second part of this work, one mass spectrometer is Finally, Walder and co-workers11±13 described an ICP-MS scanned while the second channel measures an individual m/z instrument with magnetic sector mass analyser and multiple value.When the ratio of these two signals is calculated, the Faraday detectors. Lee and Halliday used this device to obtain peak shapes in the mass spectrum are improved signiÆcantly accurate relative atomic masses of some elements.14 This for a wide range of elements. This technique corrects for instrument is capable of simultaneous measurement of up to Øicker noise from the LA process while scanning a mass nine adjacent m/z values.Isotope ratios can be measured with spectrum for multi-element determinations. very high precision (0.01% RSD) even when a noisy sample introduction method such as LA is used. The ICP multicollec- Keywords: L aser ablation; inductively coupled plasma mass tor (MC) MS has recently been used with LA to measure spectrometry; internal standard; isotope ratio; solids analysis isotope ratios for Sr in feldspar15 and Hf in zircons.16 The precision of these isotope ratio measurements approaches that of TIMS.LA-ICP-MCMS has the additional advantages of ICP-MS has become a major force in elemental MS in the spatial resolution and minimal sample preparation. past several years. The simplicity of the spectrum and the These MC studies in ICP-MS and other ICP emission speed of analysis make it an attractive technique. ICP-MS studies with multichannel detection17 have shown that simul- does, however, have several limitations.The usual quadrupole taneous ratio measurements correct for most of the Øicker mass analyser is a sequential or scanning device, which can noise in the ICP. In the last two years, Houk and limit the precision of internal standard and isotope ratio co-workers18,19 have developed a twin quadrupole instrument measurements. It is desirable when doing these type of measure- that simultaneously detects ions produced from an ICP (Fig. 1). ments to use long dwell times in order to increase the total This instrument splits the ion beam into two parts.Each part number of ions measured, to reduce the effects of shot noise. is then sent to its own quadrupole mass analyser and detector. However, when using a typical quadrupole system, the mass The eventual objective of this project is to produce a device analyser must be scanned rapidly or peak hopped for multi- that is capable of high precision ratio measurements but is mass measurements to reduce the effects of Øicker noise much smaller and less expensive than the ICP-MCMS device introduced by the plasma and sample introduction system.and easier to use for measurements at widely different m/z Such fast scanning or peak hopping limits the amount of signal values. obtained and therefore runs the risk of propagating shot noise The present paper presents results for LA of two steel SRMs in the measurement. Begley and Sharp1 discussed these effects using the twin quadrupole ICP-MS instrument.The poor and described a detailed procedure for optimizing precision precision of LA-ICP-MS is one of its main limitations. Because for isotope ratio measurements. of the erratic nature of the ablation process, the noise level on There are presently two basic types of scanning, single- the signal is signiÆcant. For this reason, LA-ICP-MS with a channel ICP-MS devices: quadrupole or magnetic sector. Over conventional single quadrupole instrument is not commonly the last several years, the precision and stability of commercial used for isotope ratio measurements.quadrupole ICP-MS devices have been improved to the point The emphasis of much of the work that has been done using where isotope ratios can be measured with an RSD of 0.1% LA-ICP-MS for isotope ratio measurement is in the Æeld of (sometimes better), in cases where at least 106 ions are detected geochemistry. Lead isotope ratios have been measured in for the minor isotope.This level of precision is obtained for zircon grains with a precision of #0.5%.20 Using LA-ICP-MS, steady state signals resulting from continuous nebulization of a close correlation between the experimentally measured pre- solutions.1,2 Precise measurement of isotope ratios using a cision of the isotope ratio and the precision predicted from scanning, double-focusing magnetic sector instrument have recently been reported.3 Using the device in the low resolution counting statistics has been reported for the measurement of Journal of Analytical Atomic Spectrometry, February 1997, Vol. 12 (171±176) 171different signal levels while keeping shot noise to a reasonable limit. This does mean, however, that a calibration curve for accurate measurement of isotope and ion ratios would be required as is typical for isotope ratio measurements with commercial ICP-MS instruments. The absolute value of the split ratio can be altered by applying different voltages to the beam shift plates before the splitter.The ICP-MS operating conditions are listed in Table 1. Some care is taken to match the behavior of the two electron multipliers. The voltage applied to each detector is chosen such that the measured count rate does not change substantially as the detector voltage is altered slightly. This procedure ensures that small, independent Øuctuations in the voltage output of each power supply do not affect the measured ratios greatly.The sampling position, ion lens potentials and aerosol gas Øow rate were optimized to yield the best precision for 52Cr+552Cr+ for NIST SRM 1263 (Low Alloy Steel Cr-V) and 51V+551V+ for NIST SRM 1767 (Low Alloy Steel). The voltages on the beam shift plates were modiÆed slightly to increase the signal for minor isotopes or small concentrations in some cases. Fig. 1 Schematic diagram of the ICP twin quadrupole device. An Laser Ablation Conditions approximate scale is given and typical ion lens potentials, in volts, are listed.An Nd5YAG laser (Model NY 82-30, Continuum) was frequency doubled to yield a beam at 532 nm. The laser was pulsed at 15 Hz and a steady state signal was produced by the Pb isotope ratios in zircon.21 However, the precision obtained ICP-MS instrument. The pulse width was 8 ns and the energy using LA-ICP-MS was worse than similar analyses using was #70 mJ per pulse. Laser energy was measured using an solution nebulization ICP-MS,21 most probably because of the energy detector (Scientech Model PHF50) and radiometer noisy nature of the ablation process.(Vector S200). The ablation system was similar to the one With the twin quadrupole device, most of the Øicker noise depicted previously19 with two improvements. Firstly, the can be removed when the ratio of a signal between an isotope quartz window into the cell was Ætted at a 45° angle, which of the same element or that of an internal standard element is minimized back reØection.Secondly, the aerosol gas was added taken.18 The precision of these measurements more closely tangentially at the base of the ablation cell and exited tangen- follows counting statistics when an isotope ratio is measured. tially slightly above the sample. This tangential Øow of aerosol The device also provides signiÆcant improvements in precision gas minimized deposition of particles in the cell and improved for the ratios of signals for ions of different elements (i.e., for particle transport efficiency from the cell.internal standardization), although the precision of internal Steel samples were obtained from NIST. The samples were standard ratios is generally poorer than that of isotope ratios cut to Æt the dimensions of the cell and smoothed. No other of the same element. Results are also presented in which one sample preparation was required. The sample was held on the mass analyser is scanned over a mass range while the other stage of a stepper motor (AMSI Model 301SM) and rotated remains on a single m/z value.When the ratio of these two at #30 rev min-1. The laser was focused onto the sample signals is taken, peak shapes in the mass spectrum improve slightly off-center with a quartz lens ( f=10 cm). In this fashion signiÆcantly for a wide range of elements. This feature is a circular track was made in the sample during ablation. The investigated as a correction for Øicker noise during mass scans. lens was positioned to yield a maximum metal ion signal which was obtained with the focal point slightly below the surface of the sample.The ablated particles were transported to the ICP EXPERIMENTAL through Tygon tubing (#1 m long, 6.4 mm i.d.). ICP-MS Device and Conditions The ICP-MS system used in the present work has been RESULTS AND DISCUSSION described previously18 and is depicted in Fig. 1. The ion beam Isotope Ratio Measurements from the ICP is extracted in the normal fashion.22 The beam then passes through a set of beam shift plates in front of the A plot of count rate versus time is given in Fig. 2 for 52Cr+ and 53Cr+ in NIST SRM 1263 for a dwell time of 0.5 s. One beam splitter. The splitter divides the beam into two parts. Each part of the beam is then sent to its own quadrupole mass mass analyser transmits only m/z=52 and the other transmits m/z=53 for the entire experiment. The certiÆed Cr concen- analyser and detector.In this fashion, signals at two m/z values can be measured simultaneously. tration in this sample is 1.31% m/m. The RSD of the individual signals is 3%, typical of LA-ICP-MS. However, when the ratio ModiÆcations to the system include a decrease in the sampler ±skimmer separation from 10 to 8 mm. This change of the two signals is calculated, the RSD improves to 0.54%. If Æve consecutive ratios are averaged, the RSD of Æve such decreased the background from #40 to #10 counts s-1.The sensitivity of the instrument also improved #ten-fold. For averaged ratios is 0.24%. A longer dwell time (2 s) (Fig. 3) improves the RSD of the ratio to 0.29%. This is due to the elements with ionization energies below #7 eV, the total sensitivity of the device (i.e., the sum of the sensitivity of both accumulation of more counts and a decrease in the shot noise limit, as has been described previously.18 Again the RSD can channels) is #4×106 counts s-1 per ppm when ultrasonic nebulization and desolvation are used.be improved by averaging Æve consecutive ratios. In this case the RSD of Æve averaged ratios is 0.058%. The effect of Next, the voltages on the beam shift plates prior to the splitter were offset to give a split in the ion beam of about increasing dwell time on precision for the Cr isotope ratio measurements and the precision predicted from counting 451. This enabled simultaneous measurement of ions of very 172 Journal of Analytical Atomic Spectrometry, February 1997, Vol. 12Table 1 Instrumental components and operating conditions Operating conditions Component ICP– RF generator Plasma Therm Forward power 1.25 kW (now RF Plasma Products), ReØected power <8W Model HFP-2000D RF Plasma Products torchbox Aerosol gas Øow 0.8 l min-1 (modiÆed in-house for horizontal operation Outer gas Øow 16 l min-1 with laboratory-made copper shielding box) Intermediate gas Øow 0.8 l min-1 Ion extraction interface23– Ames Laboratory construction Sampler position 8 mm from load coil on center Sampler oriÆce 1 mm diameter Skimmer oriÆce 1 mm diameter Sampler±skimmer separation 8 mm Vacuum System18– Three stages differentially pumped Differential pumping oriÆce 1.5 mm diameter Welded stainless steel Operating pressure/Torr* Ames Laboratory construction Expansion chamber 1.1 Second (ion lens) chamber 4×10-4 Third (quadrupole) chamber 5×10-6 Mass analysers– From VG PlasmaQuad Mean rod bias 0 V Model SXP 300 with rf-only pre-Ælters Model SXP 603 controllers and rf generators Electron multiplier– Galileo Model 4870, pulse counting mode Bias voltage -2800 V Counting electronics– EG&G ORTEC Model 660 dual 5 kV bias supply Model 9302 ampliÆer/discriminator Model 994 dual counter/timer * 1 Torr=133.322 Pa.Fig. 2 Plot of count rate versus time during 15 Hz laser ablation of Fig. 3 Plot of count rate versus time during 15 Hz laser ablation of steel SRM 1263, dwell time=0.5 s.The RSD of each signal is 3.2% SRM 1263, dwell time=2.0 s. The RSD of each signal is 1.3% while while the RSD of the ratio is 0.54%. The concentration of Cr= the RSD of the ratio is 0.29%. The concentration of Cr=1.31% m/m. 1.31% m/m. Table 2 Effect of dwell time on precision of Cr isotope ratio statistics are shown in Table 2. The procedure for calculating the RSD expected from counting statistics is described in Precision of 53Cr+552Cr+ ratio, eqn. (1) of ref. 18. Application of the F-test24 indicates that for RSD (%) dwell times up to 2.5 s, the measured RSD and the RSD predicted from counting statistics are not signiÆcantly different. Mean ratio* Counting Five averaged Dwell time/s 53Cr+552Cr+ Measured statistics ratios Beyond 2.5 s precision deteriorates, for reasons that are unclear at this time. 0.5 0.530 0.543 0.50 0.240 It is also shown in Table 2 that the precision can be improved 1.0 0.525 0.338 0.34 0.090 1.5 0.523 0.358 0.28 0.228 by averaging every Æve ratios, as noted earlier.This averaging 2.0 0.522 0.288 0.24 0.058 procedure could be considered to be an `artiÆcial' way to 2.5 0.521 0.253 0.21 0.226 extend the dwell time and reduce the precision limits imposed 3.5 0.520 0.255 0.18 0.186 by counting statistics. 4.5 0.578 1.20 0.18 1.14 The signals for 206Pb+ and 208Pb+ from this sample (NIST 7.7 0.555 0.563 0.13 0.478 SRM 1263) are shown in Fig. 4. The dwell time is 2 s, and Pb is present at 22 ppm.The resulting signals for each isotope are * Accepted natural abundance ratio=0.113. Journal of Analytical Atomic Spectrometry, February 1997, Vol. 12 1730.12%. In this case application of the F-test indicates that the measured RSD is signiÆcantly higher than that predicted from counting statistics. It should be noted that the measurement of 51V+552Cr+ represents the easiest internal standard measurement for different elements. The masses of V and Cr are very close (Dm=1 u) and their ionization energies (IE) are low and nearly the same (6.74 and 6.77 eV, respectively), so they should behave similarly in the ionization, ion extraction and beam splitting processes.A more difficult case is demonstrated in Fig. 6. A plot of count rate versus time for 184W+ and 52Cr+ is shown again using NIST SRM 1263. The concentrations of these two elements are 0.046 and 1.31% m/m, respectively. The large difference in mass (Dm=132 u) and ionization energy (DIE= 1.21 eV) should make for a poor internal standard pair.At a Fig. 4 Plot of count rate versus time during laser ablation of SRM dwell time of 3.0 s the RSD of the individual signals is 3%. 1263, dwell time=2.0 s. The RSD of each signal is 17% while the RSD When the ratio of the two signals is taken, the RSD improves of the ratio is 2.6%. The Pb concentration is 22 ppm. to 0.63%. This value is higher than the counting statistics limit (0.39%) by a factor of only 1.6.very noisy, with an RSD of #17%. The precision of the Values of RSD for the signal ratios 51V+552Cr+ and isotope ratio 206Pb+5208Pb+ is 2.6%. If every Æve ratios are 184W+552Cr+ at various dwell times, along with the counting averaged, the RSD of the averaged ratios improves to 0.68%. statistics limit and the RSD value obtained when using the This value is essentially the same as the counting statistics Æve ratio averaging technique, are given in Table 3. The limit of 0.66%.best RSDs obtained are #0.12% for 51V+552Cr+ and 0.35% Note that the signal for 206Pb+ is higher than that for for 184W+552Cr+. These NIST steel SRMs are generally 208Pb+ in Fig. 4. In contrast, the actual sample contains roughly half as much 206Pb as 208Pb. A similar bias occurs in the measured ratios for 53Cr+552Cr+ shown in Table 2. As described under Experimental, the signals for the less abundant isotopes (206Pb+ and 52Cr+) have been enhanced artiÆcially by adjusting the voltages applied to the deØection plates (Fig. 1) to send most of the ion beam through the appropriate channel. This ability to enhance the signal for the minor isotope actually helps improve the contribution made by the counting statistics to the precision. Internal Standardization A plot of count rate versus time for 51V+ and 52Cr+ is shown in Fig. 5 for a dwell time of 7.7 s using NIST SRM 1263. The certiÆed concentration of V is 0.31%, and Cr is present at 1.31% m/m. Each signal has an RSD of 1.9%.This RSD value Fig. 6 Plot of count rate versus time for 52Cr+ and 184W+ during is fairly good for LA-ICP-MS. However, the RSD improves 15 Hz laser ablation of SRM 1263, dwell time=3.0 s. The RSD of each signal is 3% while the RSD of the ratio is 0.63%. The left axis to 0.24% when the ratio of the two signals is taken. The corresponds to the 52Cr+ signal and the right axis corresponds to the counting statistics limit for this measurement is 0.12%. Again 184W+ signal. The concentration of Cr=1.31% and W=0.046% m/m.the precision can be improved further by averaging every Æve ratio measurements. The RSD of Æve such averaged ratios is Table 3 Effect of dwell time on precision using internal standard elements RSD of ratio (%) RSD of Æve Counting averaged Dwell time/s Measured statistics ratios (%) 51V +552Cr+– 0.5 0.882 0.440 0.350 1.5 0.493 0.255 0.348 2.5 0.374 0.200 0.226 3.5 0.326 0.171 0.149 4.5 0.479 0.153 0.416 7.7 0.241 0.119 0.119 10 0.369 0.107 0.294 184W +552Cr+– 0.5 1.41 0.900 0.477 1.5 0.817 0.526 0.354 Fig. 5 Plot of count rate versus time for 52Cr+ and 51V+ during 2.0 0.964 0.470 0.721 15 Hz laser ablation of SRM 1263, dwell time=7.7 s. The RSD of 3.0 0.632 0.390 0.427 each signal is 1.9% while the ratio is 0.24%. The concentration of 5.0 1.85 0.317 1.72 Cr=1.31% and V=0.31% m/m. 174 Journal of Analytical Atomic Spectrometry, February 1997, Vol. 12considered to be homogeneous. Thus spatial changes in sample relative error of 10.4%.The value obtained for the same ratio using the corrected mass spectrum is 1.060 with a much lower composition probably do not contribute to the differences between measured precision and that expected from counting relative error (1.14%). This value represents the mass bias in just one side of the beam splitter and one of the mass analysers statistics. The precision obtained for signal ratios of different elements is generally poorer than that obtained for isotope and is comparable to the bias seen with conventional single quadrupole instruments.25 This moderate value for mass bias, ratios of the same element, regardless of whether the sample is introduced by LA or as a solution aerosol.18,19 The RSD #1% per mass unit, is somewhat surprising considering the fact that the beam splitter could act as an energy analyser and values reported above represent substantial improvements over those obtained in early work with this device.18,19 cause more mass bias than usual.A similar mass spectrum and count rate versus time plot for several elements is shown in Fig. 8. One quadrupole is scanned Improvements in Peak Shape and Precision During Scans over the region m/z=44±53 while the other measures 51V+. The dwell time used is once again 1.2 s. Irregular peak shapes The previous results were all obtained with each mass analyser are also seen in Fig. 8(a), owing to Øicker noise in the plasma set at one m/z value.A unique feature of this instrument is the and sample introduction process. As indicated by the broken ability to scan with one quadrupole, while the other measures lines, these noise spikes in the peaks correspond to similar a single m/z value. Thus, an internal standard ion can be spikes on the 51V+ signal. The RSD of the 51V+ signal is 7.4%. measured at precisely the same time as each analyte ion to The shapes of the peaks in the mass spectrum improve correct for Øicker noise during multi-element determinations.substantially when the ratio of the two signals is calculated as Such an experiment is demonstrated in Fig. 7. NIST SRM shown in Fig. 8(b). 1767 is ablated and the resulting mass spectrum and count Suppose that V and Ti are homogeneously distributed and rate versus time plots are shown in Fig. 7(a). One quadrupole have natural isotopic abundances, that selective ablation is not measures 51V+ while the other scans over the region m/z= a problem and that both elements are 100% ionized in the 89±102.Each data point in the 51V+ signal versus time plot is plasma. Under these assumptions a comparison of accuracy measured at the same time as the point directly beneath it in can be made. The certiÆed value of the atomic abundance the scanned spectrum. Also, note that the dwell time for both ratio of 51V+548Ti+ is 3.813. Following background subtrac- the scan and single m/z measurement is 1.2 s.The use of this tion, the uncorrected mass spectrum gives a value of 4.204. relatively long dwell time allows for accumulation of sufficient The corrected mass spectrum yields 4.128. Again both meas- counts to reduce the importance of shot noise, which is not ured values are higher than the actual certiÆed value. However, removed by these simultaneous ratio measurements. the relative error using the uncorrected mass spectrum In Fig. 7(a) the RSD of the 51V+ signal is 7.0%.The broken (10.25%) is higher than the relative error using the corrected lines in Fig. 7(a) indicate several places where Øicker noise has mass spectrum (8.26%). This error is somewhat higher than perturbed the 51V+ measurement and the mass scan at the that expected from mass bias. Once again, it appears that same time. The ratio of the two signals is taken and the signals from different elements do not correlate as well as corrected mass spectrum is shown in Fig. 7(b).As can be seen, signals for different isotopes of the same element.18 irregularities in the peak shapes are corrected and a smoother The proposed method could be of value in two ways. Firstly, mass spectrum results. This result is consistent with the results it yields a more representative mass spectrum since Øicker presented above in that Øicker noise cancels for a wide range noise in the plasma and sample introduction process has been of elements. removed. Secondly, since a more accurate mass spectrum is The accepted abundance ratio for 96Mo595Mo is 1.048.The obtained, the presence of a spectral interference can be deduced measured ratio, obtained using peak height after subtraction and corrected more readily. As can be seen, the correction of the background and 96Zr+ signal, is 1.157. This value has a using two different elements is not as good as an isotope ratio Fig. 7 (a) Mass spectrum and count rate versus time plot for 51V+ Fig. 8 (a) Mass spectrum and count rate versus time plot for 51V+ during 15 Hz laser ablation of SRM 1767, dwell time=1.2 s.The broken lines indicate spikes due to Øicker noise in both the mass scan during 15 Hz laser ablation of SRM 1767, dwell time=1.2 s. The broken lines indicate spikes due to Øicker noise in both the mass scan and the 51V+ signal. (b) The ratio of the mass spectrum signal to the 51V+ signal in (a) multiplied by 1000. The value for 96Mo+595Mo+ is and the 51V+ signal. (b) The ratio of the mass spectrum signal to the 51V+ signal in (a) multiplied by 1000.The value for 51V+548Ti+ in the 1.157 in the uncorrected mass spectrum. The value for the same ratio is 1.060 in the corrected mass spectrum. The accepted value of the uncorrected mass spectrum is 4.204. The value of the ratio in the corrected mass spectrum is 4.128. The accepted value is 3.813. ratio is 1.048. Journal of Analytical Atomic Spectrometry, February 1997, Vol. 12 1754 Myers, D.P., Li, G., Yang, P., and Hieftje, G. M., J. Am. Soc. correction; however, a more accurate value results in both Mass Spectrom., 1994, 5, 1008. cases. 5 Myers, D. P., Heintz, M. J., Mahoney, P. P., Li, G., and Hieftje, G. M., Appl. Spectrosc., 1994, 49, 1337. 6 Heintz, M. J., Myers, D. P., Mahoney, P. P., Li, G., and Hieftje, CONCLUSION G. M., Appl. Spectrosc., 1995, 49, 945. 7 Myers, D. P., Li, G., Mahoney, P. P., and Hieftje, G. M., J. Am. Results obtained indicate that, using a twin quadrupole Soc.Mass Spectrom., 1995, 6, 411. ICP-MS instrument, precision can be improved by simul- 8 Mahoney, P. P., Li, G., and Hieftje, G. M., J. Anal. At. Spectrom., taneous measurement of ion signals in both mass scanning 1996, 11, 401. 9 Barinaga, C. J., and Koppenaal, D. W., Rapid Commun. Mass and selected ion monitoring modes. Flicker noise from laser Spectrom., 1994, 8, 71. ablation and from the plasma are correlated and can be 10 Koppenaal, D. W., Barinaga, C.J., and Smith, M. R., J. Anal. At. cancelled. For ratios of isotopes of the same elements, measured Spectrom., 1994, 9, 1053. values of precision are only slightly above the counting stat- 11 Walder, A. J., Abel, I. D., Platzner, I., and Freedman, P. A., istics limit. Internal standard ratios improve precision values Spectrochim. Acta, Part B, 1993, 48, 397. to about 1.6 times the counting statistics limit. Thus, Øicker 12 Walder, A. J., and Freedman, P. A., J. Anal.At. Spectrom., 1992, 7, 571. noise does not cancel as effectively when two different elements 13 Walder, A. J., Platzner, I., and Freedman, P. A., J. Anal. At. are measured, compared with an isotope ratio of the same Spectrom., 1993, 8, 19. element. Mass spectra from one channel are corrected using 14 Lee, D., and Halliday, A. N., Int. J. Mass Spectrom. Ion Processes, an internal standard from the other channel. The corrected 1995, 146/147, 35. mass spectrum gives more accurate results for isotope and 15 Christensen, J.N., Halliday, A. N., Lee, D., and Hall, C. M., Earth Planet. Sci. L ett., 1995, 136, 79. elemental ratios than an uncorrected spectrum. Interestingly, 16 Thirwall, M. F., and Walder, A. J., Chem. Geol. (Isot. Geosci. a similar principle, although without simultaneous detection, Sect.), 1995, 122, 241. could be used with a single channel instrument equipped with 17 Mermet, J.-M., and Ivaldi, J. C., J. Anal. At. Spectrom., 1993, 8, 795. the appropriate controlling software to hop or scan rapidly, 18 Warren, A. R., Allen, L. A., Pang, H., Houk, R. S., and thereby minimizing Øicker noise during multi-element Janghorbani, M., Appl. Spectrosc., 1994, 48, 1360. 19 Allen, L. A., Pang, H.-M., Warren, A. R., and Houk, R. S., J. Anal. determinations. At. Spectrom., 1995, 10, 267. 20 Feng, R., Machado, N., and Ludden, J., Geochim. Cosmochim. Ames Laboratory is operated for the US Department of Energy Acta, 1993, 57, 3479. by Iowa State University under Contract No. 21 Fryer, B. J., Jackson, S. E., and Longerich, H. P., Chem. Geol., 1993, 109, 1. W-7405-ENG-82. This work is supported by the US 22 Niu, H., and Houk, R. S., Spectrochim. Acta, Part B, 1996, 51, 779. Department of Energy, Environmental Remediation andWaste 23 Olivares, J., and Houk, R. S., Anal. Chem., 1985, 57, 2674. Management, Office of Technology Development. 24 Skoog, D. A., West, D. M., and Holler, F. J., Fundamentals of Analytical Chemistry, Saunders, New York, 5th edn., 1988, p. 38. 25 Jarvis, K. E., Gray, A. L., and Houk, R. S., in Handbook of ICPREFERENCES MS, Blackie, London, 1991, p. 331. 1 Begley, I. S., and Sharp, B. L., J. Anal. At. Spectrom., 1994, 9, 171. Paper 6/03310E 2 Koirtyohann, S. R., Spectrochim. Acta, Part B, 1994, 49, 1305. ReceivedMay 13, 1996 3 Vanhaecke, F., Moens, L., Dams, R., and Taylor, P., Anal. Chem., 1996, 68, 567. Accepted September 4, 1996 176 Journal of Analytical Atomic Spectrometry, February 1997, Vol. 12