Evaluation of the Isotope Ratio Performance of an Axial Time-of-Flight

Reliable isotope ratio determination is of great importance in many application fields. For a number of elements (e.g., Sr, Os, and Pb), the isotopic ...
0 downloads 0 Views 83KB Size
Anal. Chem. 1999, 71, 3297-3303

Evaluation of the Isotope Ratio Performance of an Axial Time-of-Flight ICP Mass Spectrometer Frank Vanhaecke,* Luc Moens, and Richard Dams

Laboratory of Analytical Chemistry, Ghent University, Institute for Nuclear Sciences, Proeftuinstraat 86, B-9000 Ghent, Belgium Lloyd Allen and Stuart Georgitis

LECO Corporation, 3000 Lakeview Avenue, St Joseph, Michigan 49085

The isotope ratio performance of an axial time-of-flight ICP mass spectrometer (Renaissance TOF-ICPMS, LECO Corp.) was evaluated. The isotope ratio precision, expressed as the relative standard deviation (RSD) for 10 successive measurements, was evaluated using multielement standard solutions with analyte concentrations of 50-500 µg/L. The influence of the acquisition time per replicate measurement was studied by varying it between 0.5 and 300 s. For an acquisition time of 30 s per replicate and an elemental concentration of 500 µg/L, typical isotope ratio precisions of e0.05% RSD were obtained. The fact that this isotope ratio precision can be obtained for many ratios simultaneously is an especially attractive feature of TOF-ICPMS. In contrast to what was expected, increasing the acquisition time per replicate to values of >30 s resulted in a slightly deteriorated isotope ratio precision. At short acquisition times (30 s. For many ratios, the precision obtained at values of >30 s even seems to be somewhat deteriorated. The cause of this deterioration is not obvious. Possibly, when the acquisition time is extended, long-term instability and signal drift adversely affect the results obtained. Further studies of this deterioration are currently underway. The isotope ratio precision found is roughly similar to that established for single-collector sector field ICPMS instruments operated at the low-resolution setting.42,45,49 However, it is important to stress that, for the latter instrument type, RSDs of e0.05% can only be obtained when one pair of isotopes (one isotope ratio) only is monitored and when electric scanning (variation of the acceleration voltage) is used at a sufficiently high scanning or peak-hopping rate.42 As soon as the magnet setting (B-field) has to be changed, the isotope ratio precision is significantly deteriorates. With a TOF-based instrument on the other hand, many isotope ratios can be measured simultaneously over a broad mass range, since the number of isotopes monitored has no influence on the isotope ratio precision obtained. The influence of the signal intensity (concentration level) of the nuclides involved has been assessed by determining the isotope ratio precision attainable for several isotope ratios at elemental concentrations of 50, 100, 250, and 500 µg/L. This experiment was carried out using an acquisition time of 10 and of 30 s per replicate. The results obtained have been summarized in Table 5. As expected, an increase in the signal intensities resulted in an improved isotope ratio precision. From Table 5, it is clear that the results for 86Sr/88Sr, 135Ba/138Ba, and 137Ba/138Ba at concentrations of g250 µg/L do not follow this general trend (results in parentheses). Examination of the raw data revealed that for the more abundant isotope of these ratios (88Sr+ and 138Ba+, respectively) highly intense signals were observed (signal intensity, e.g., twice as high as for 208Pb+). Possibly, at the concentration levels used, the signal intensity for these highly abundant nuclides approaches or exceeds the upper limit of the detector’s dynamic range64 (range in which the output current is linearly proportional to the input current). The validity of this hypothesis seems to be confirmed by the fact that the isotope ratio results themselves are also affected, with the more abundant nuclide being discriminated relative to the less abundant one. When the results for those isotope ratios are discarded, the trend observed seems to be qualitatively and to some extent even quantitatively in agreement with the improvement expected on the basis of counting statistics. Theoretically a 3.2-fold (101/2) improvement is expected on increasing the concentration level (64) Kurz, E. A. Am. Lab. 1979, 67-82.

Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

3301

Table 3. Isotope Ratio Precision, Expressed as RSD (%) for n ) 10, Obtained Using Different Acquisition Times per Replicate (Elemental Concentrations of 50 µg/L) time (s)

63Cu/65Cu

64Zn/66Zn

66Zn/68Zn

86Sr/87Sr

86Sr/88Sr

107Ag/109Ag

135Ba/138Ba

137Ba/138Ba

204Pb/208Pb

206Pb/207Pb

206Pb/208Pb

207Pb/208Pb

0.5 1 5 10 30 60 120 300

1.4 0.77 0.30 0.29 0.12 0.19 0.19 0.16

1.9 0.90 0.46 0.48 0.15 0.20 0.70 1.5

2.0 1.9 0.54 0.67 0.21 0.37 1.7 1.1

1.3 0.69 0.51 0.31 0.30 0.31 0.20 0.14

1.2 0.81 0.50 0.85 0.26 0.51 0.30 0.24

0.55 0.30 0.11 0.28 0.040 0.058 0.046 0.10

1.31 0.86 0.49 0.36 0.28 0.74 0.51 0.46

1.1 0.62 0.31 0.39 0.14 0.40 0.45 0.26

5.5 5.2 1.7 4.6 2.5 2.4 1.4 1.3

0.70 0.42 0.14 0.18 0.090 0.17 0.071 0.059

0.62 0.40 0.17 0.30 0.075 0.21 0.21 0.25

0.81 0.39 0.23 0.34 0.11 0.34 0.22 0.23

Table 4. Isotope Ratio Precision, Expressed as RSD (%) for n ) 10, Obtained Using Different Acquisition Times per Replicate (Elemental Concentrations of 500 µg/L) time (s)

63Cu/65Cu

64Zn/66Zn

66Zn/68Zn

86Sr/87Sr

86Sr/88Sr

107Ag/109Ag

135Ba/138Ba

137Ba/138Ba

204Pb/208Pb

206Pb/207Pb

206Pb/208Pb

207Pb/208Pb

0.5 1 5 10 30 60 120 300

0.27 0.20 0.11 0.034 0.031 0.049 0.039 0.046

0.27 0.28 0.11 0.12 0.056 0.082 0.048 0.058

0.56 0.29 0.14 0.13 0.063 0.057 0.049 0.062

0.56 0.30 0.14 0.051 0.044 0.072 0.030 0.047

0.21 0.15 0.081 0.082 0.038 0.074 0.040 0.11

0.16 0.12 0.066 0.073 0.033 0.044 0.070 0.062

0.38 0.28 0.10 0.13 0.069 0.15 0.11 0.18

0.24 0.15 0.054 0.065 0.051 0.081 0.066 0.12

0.78 0.61 0.19 0.21 0.069 0.54 0.20 0.24

0.33 0.13 0.089 0.042 0.023 0.022 0.022 0.033

0.24 0.13 0.067 0.045 0.041 0.070 0.086 0.060

0.28 0.14 0.11 0.036 0.032 0.054 0.075 0.031

206Pb/207Pb

206Pb/208Pb

207Pb/208Pb

Table 5. Isotope Ratio Precision, Expressed as RSD (%) for n ) 10, Obtained for Different Elemental Concentrationsa conc (µg/L)

63Cu/65Cu

64Zn/66Zn

50 100 250 500

0.20 0.093 0.067 0.051

0.37 0.26 0.14 0.082

50 100 250 500

0.095 0.071 0.080 0.24

0.11 0.10 0.085 0.037

66Zn/68Zn

86Sr/87Sr

86Sr/88Sr

107Ag/109Ag

135Ba/138Ba

137Ba/138Ba

204Pb/208Pb

0.49 0.21 0.19 0.16

0.28 0.15 0.092 0.076

0.42 0.10 (0.18) (0.59)

0.14 0.059 0.053 0.034

0.21 0.078 0.10 (0.48)

0.11 0.096 0.046 (0.64)

0.55 0.25 0.20 0.14

0.13 0.12 0.090 0.047

0.11 0.15 0.071 0.035

0.14 0.086 0.047 0.031

0.22 0.16 0.12 0.030

0.15 0.094 0.058 0.094

0.13 0.047 (0.54) (0.18)

0.080 0.036 0.025 0.050

0.082 0.060 0.088 (0.44)

0.052 0.061 0.043 (0.27)

0.24 0.22 0.14 0.078

0.11 0.051 0.038 0.037

0.10 0.042 0.028 0.041

0.087 0.024 0.041 0.014

a Acquisition time per replicate: upper part of table, 10 s; lower part of table, 30 s. Results in parentheses are adversely affected by the excessively high signal intensity of the major isotope, which may have been beyond the detector’s dynamic range (see text).

from 50 to 500 µg/L, while in practice, improvement factors of 3.7 and of 3.2 were observed for an acquisition time of 10 and 30 s, respectively. Also when the influence of the acquisition time was being evaluated, a concentration level of 500 µg/L was used. In this case, however, no deterioration of the isotope ratio precision was observed for 86Sr/88Sr, 135Ba/138Ba, and 137Ba/138Ba (Table 4). Hence, also for these measurements, the raw data were examined and it was established that, in this case (different measuring day), the signal intensities obtained were a factor of ∼4 lower than those obtained during the measurements corresponding to Table 5. On one hand, this difference in sensitivity is somewhat surprising in view of the good day-to-day stability normally observed with the instrument used. However, on the other hand, this specific instrument, being a prototype, did not have a computer-controlled axial torch positioning system, such that some degree of inconsistent torch positioning may be expected. Hence, most probably the difference in sensitivity observed has to be attributed to a different position of the torch with respect to the sampling cone. For a concentration level of 500 µg/L, the isotope ratio precision observed at an acquisition time as short as 5 or even 1 3302 Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

s per replicate is similar tosor for some isotope ratios even better thansthe best values ever reported for commercially available quadrupole-based instrumentation (Table 4). Hence, TOF-ICPMS also shows great promise for isotopic analysis involving transient signals of short duration (sample introduction by means of flow injection, electrothermal vaporization or laser ablation, or coupling with chromatographic separation techniques). To test this possibility, transient signals with a duration of 6 s fwhm were generated using a discrete solution injection (flow injection) device to introduce 10-µL aliquots of the multielement standard solution at a concentration level of either 50 or 500 µg/L. The isotope ratio precision for five successive measurements was determined (Table 6). As the RSDs were observed to be typically ∼0.5% at 50 µg/L and e0.1% at 500 µg/L, they confirmed the aforementioned expectation. As in all previously described experiments the integration region was selected manually for each peak, the corresponding results permitted no conclusions to be drawn concerning the extent of mass discrimination. Therefore, a full peak integration was performed for 112Cd and 114Cd and the experimental value obtained for 112Cd/114Cd was compared to the corresponding true

Figure 3. Evaluation of the isotope ratio accuracy attainable with axial TOF-ICPMS. Comparison of Pb isotopic results obtained for a “natural” standard solution obtained by TOF-ICPMS and single-collector TIMS. Uncertainties indicated are 95% confidence intervals. Table 6. Isotope Ratio Precision, Expressed as RSD (%) for n ) 5, Obtained for Transient Signals (fwhm, 6 s)a 63Cu/ 64Zn/ 86Sr/ 86Sr/ 107Ag/ 137Ba/ 206Pb/ 206Pb/ 207Pb/ 65Cu

66Zn

87Sr

88Sr

109Ag

138Ba

207Pb

208Pb

208Pb

50 µg/L 0.21 0.66 0.47 2.47 500 µg/L 0.12 0.072 0.15 0.23

0.23 0.043

0.31 0.10

0.36 0.12

0.48 0.044

0.48 0.10

a Elemental concentration: upper part of table, 50 µg/L; lower part of table, 500 µg/L.

value, calculated on the basis of IUPAC tabulated values.65 This comparison showed that, at midmass, the mass discrimination is ∼1% per mass unit. This is similar to the mass discrimination observed for quadrupole-based and sector field ICPMS instruments.49 Mahoney et al.52 have observed substantially larger mass discrimination with a TOF-ICPMS instrument. However, in that case, orthogonal acceleration was used to introduce the ions pulsewise into the flight tube of the mass analyzer, while the instrument used in this study applies axial acceleration for this purpose. The difference between both approaches in terms of mass discrimination has been comprehensively discussed before. The isotope ratio accuracy attainable by TOF-ICPMS was tested by means of determination of the 206Pb/207Pb and 206Pb/ 208Pb isotope ratios for a Pb standard solution, for which the isotopic composition was previously determined using singlecollector TIMS. To correct for mass discrimination, a solution of NIST 981 (Common lead isotopic standard) was used as an external standard. Of course, the isotope ratios of interest were measured simultaneously. The final TOF-ICPMS results are based on three subsequent determinations. Each determination consisted of five replicate measurements (30 s) of the “sample” solution and five replicate measurements (30 s) of the NIST standard solution. Hence, the values reported and graphically presented in Figure 3 (65) Rosman, K. J. R.; Taylor, P. D. P. J. Anal. At. Spectrom. 1998, 13, 45N55N.

are average values for n ) 3. The uncertainty on the results, expressed as the 95% confidence interval, was calculated by taking into account the uncertainty (standard deviation) on the experimental results for both the sample and the NIST standard and the uncertainty on the certified values for the latter. The TIMS results are average values based on five independent determinations. Also, in this case, the uncertainty has been expressed as the 95% confidence interval. As can be seen from Figure 3, the results obtained by TOF-ICPMS agree within the experimental uncertainty with the values obtained by TIMS. CONCLUSIONS It has been shown that, at sufficiently high signal intensities and for a moderate acquisition time, axial TOF-ICPMS yields a typical isotope ratio precision of e0.05% RSD. This isotope ratio precision can be obtained for many ratios simultaneously. The mass discrimination is similar to that observed with quadrupolebased and sector field instruments, and when appropriately corrected for, accurate results can be obtained. Of course, a complete evaluation of the isotope ratio performance of axial TOFICPMS requires further research, in which also the long-term isotope ratio stability, the dynamic range, the influence of the matrix, and the stability of the mass discrimination and its constancy across the mass range will have to be studied. ACKNOWLEDGMENT F.V. acknowledges Prof. Dr. K. G. Heumann and J. Diemer of the Laboratory of Inorganic Chemistry and Analytical Chemistry of the Johannes Gutenberg University in Mainz, Germany, for the opportunity to use their TIMS equipment and the assistance during the measurements, respectively. Received for review January 11, 1999. Accepted April 23, 1999. AC990016B

Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

3303