JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL. 6 199 Optimization of the Mass Scanning Rate for the Determination of Lead Isotope Ratios Using an Inductively Coupled Plasma Mass Spectrometer Naoki Furuta Division of Environmental Chemistv The National Institute for Environmental Studies 16-2 Onoga wa Tsukuba lbaraki 305 Japan High-speed motion pictures of a plasma and the noise-amplitude spectra of the emission signals enabled clarification that the inductively coupled plasma coupled with a mass spectrometer had a noise component at a fre- quency of 380 Hz under the experimental conditions used. In order to improve the precision of lead isotope ratio measurements the mass scanning rate was optimized. With an increase in the mass scanning rate the precision of lead isotope ratio measurements was improved because the low frequency noises could be eliminated.However the quadrupole mass analyser could not be operated at a speed that was sufficient to smooth out the audio frequency noise. Keywords Inductively coupled plasma mass spectrometry; lead isotope ratio; noise-amplitude spectra; high- speed motion pictures; dwell time Examination of lead isotope ratios has potential for use in dis- tinguishing the sources of environmental pollution. Thermal ionization mass spectrometry has been established as a method for the determination of lead isotope ratios. However recently inductively coupled plasma mass spectrometry (ICP-MS) has also been applied to the determination of lead isotope ratios.'-' When lead isotopic measurements are carried out improved results are obtained when the isotopes are measured simultane- ously in order to correct for the fluctuation of signals.However when a quadrupole mass analyser is used for ICP-MS isotopes cannot be measured simultaneously they must be measured se- quentially. If the measurement sequence is sufficiently fast to smooth out the signal fluctuation the effect can be eliminated. In the noise spectra of ICP-MS signals audio frequency (a.f.) noise was observed at 200400 Hz depending on the plasma op- erating I In this study an improvement of lead isotope ratio measurements was attempted by increasing the mass scanning rate of the quadrupole mass analyser. Experimental Instrumentation The ICP-MS instrument used in this study was a VG Plasma- Quad I1 (VG Elemental). The ICP was operated at 1.4 kW (27 MHz) and the outer intermediate and carrier gas flow- rates were 13.0 0.60 and 0.76 1 min-' respectively.The dis- tance from the load coil to the sampling orifice was fixed at 11 mm. The quadruple mass analyser was a Model 12-12s developed by VG Masslab. The mass resolution was set to 0.6 u defined as the peak width at 208 u at 5% of the peak height. Data acquisition was carried out in the range scanning mode. The mass range from 202.97 to 209.90 u was scanned and the data acquired were allocated to 5 12 channels of a mul- tichannel scaler. The dwell time for each channel and the number of sweeps were varied as listed in Table 1. By combi- nation of the dwell time and the number of sweeps an effort was made to fix the total acquisition time to 65.5 s for all measurements.Procedures for the Determination of Lead Isotope Ratios An ICP-MS spectrum for a National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 98 1 Lead Common Isotopic (a solution of 100 ng g-l) is shown in Fig. 1. Signals for _+ 0.4 u about the maximum peak were inte- Table 1 Various combinations of the dwell time and the number of sweeps Dwell t ime/ms Number of sweeps Per channel Per mass ( 1 channel) (73.88 channels) 0.005 0.0 10 0.020 0.040 0.080 0.160 0.320 0.640 1.28 2.56 5.12 10.24* 20.48* 40.96* 0.369 0.739 1.48 2.96 5.9 I 11.8 23.6 47.3 94.6 189 378 757 1513 3026 25600 12800 6400 3200 1600 800 400 200 100 50 25 13 7 4 * An effort was made to fix the total acquisition time at 65.5 s for all measurements.However as the dwell time could not be changed arbitrari- ly for the dwell times of 10.24 20.48 and 40.96 ms per channel the total acquisition times were 68.2 73.4 and 83.9 s respectively. grated for each isotope and the lead isotope ratios were calculated for ztmPb 204Pb 2caPb r07Pb and 2oRPb '@'Pb. The isotope ratio measurement was repeated ten times and the ac- curacy and precision were expressed by the average value and the relative standard deviation (RSD %) respectively. Plasma Pictures and Noise-Amplitude Spectra The ICP coupled with a mass spectrometer was photographed with a high-speed video camera (Kodak Model SP2000-C). The camera system enabled pictures of the plasma to be taken at a rate of 2000 frames per second.A small mono- chromator (HR-320 Jobin-Yvon; focal length 32 cm) was set at right angles to the mass spectrometer and the emission signals from the boundary region between the plasma and the sampling cone were measured by a photomultiplier tube (PMT). The PMT output was processed and noise-amplitude spectra were obtained by using the procedures described in a previous paper."200 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY. APRIL 1991. VOL. 6 204 205 206 207 208 209 mlz Fig. 1 5olution containing 100 ng g-' of lead ICP-MS \pectrum tor a lead iwtopic atandard (NIST SRM 98 I ) Reagents All sample preparations were carried out in a clean room (less than 1000 particles per ft'). De-ionized water was obtained from a Milli-Q system (Millipore) and the water collected was purified again by sub-boiling distillation.Nitric acid ( 13.36 rnol dm-') was of ultrapure-reagent grade purchased from Kanto Chemicals. The certified NIST SRM 981 was dis- solved in 2 mol dm-' nitric acid and the lead solution obtained was serially diluted. to give a concentration range from 0.1 to 500 ng g-I with 0.036 mol dm-I nitric acid. The lead isotopic standard solutions were used in all the experiments. Results and Discussion Plasma Pictures The ICP pictures taken with the high-speed video camera are shown in Fig. 2. In the pictures a sample injection tube can be seen at the right-hand side and a sampling cone at the left- hand side. Scrutiny of the boundary region between the plasma and the sampling cone shows the periodic fluctuation of the plasma.The plasma is expanded at first [Fig 2(a)]. but grad- ually shrinks [Fig. 2(h) and ((*)] and then expands again [Fig. 2(d)-(f)]. Starting from Fig. 2((r) one cycle of the plasma per- iodic motion terminates at Fig 2v); the next cycle starts from Fig. 2(g). The time interval between any two pictures is 0.5 ms thus one cycle corresponds to 2.5 ms that is an a.f. of about 400 Hz. Noise-Ampli t ude Spectra In order to clarify the noise characteristics of the ICP the emission intensity was measured by use of a small mono- chromator set at right angles to the mass spectrometer. Fig. 3 shows the noise-amplitude spectra for Ca I1 (393.4nm) emission in the frequency ranges from 0 to 500 Hz [Fig. 3(a)J and from 0 to 10 Hz [Fig. 3(h)]. The a.f. noise at 380 Hz corresponds to the fluctuation observed in the ICP pictures (Fig.2). In addition to the a.f. noise there are white and in- terference noises owing to the line source and the llf noise. In the low frequency range [Fig. 3(h)] the noise observed at 0.97 Hz and the harmonics of it were due to pulsations induced by the peristaltic pump used for sample introduc- tion. The general features of the noise spectra are very similar to those observed for ICP-MS signals."' I ' The Fig. 2 High-\peed motion picture\ of the ICP coupled with a ma$\ \pectrometerJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL. 6 20 1 0 2 4 6 8 10 Frequency/Hz ' Fig. 3 Noise-amplitude spectra for Ca I1 (393.4 nm) emission from the ICP coupled with a mass spectrometer. (a) Frequency range 0-500 Hz; and ( h ) frequency range 0- 10 Hz 301 30 1 5 V 0.1 1.0 10 100 Concentration of Pblng g-' Fig.4 Isotope ratios of A. z0xPb:2('Pb; B ""Pb "'Pb; and C zcmPb '07Pb as a function of the concentration of lead. ((I) The average value; and (h) the relative standard deviation (%) of ten measurements. The dwell time per channel was 0.32 ms a.f. noise shifted to a higher frequency with increasing radio- frequency power and outer gas flow-rate. However the a.f. noise did not shift at all when the carrier gas flow-rate was changed. These trends are exactly the same as those which have already been reported in the literature."L'' Although several suggestions have been made regarding the origin of 1.095 c -/ 1.085 1 C I I 0 100 200 300 400 500 Concentration of Pblng mi-' 1.080 Fig.5 A First trial; B second trial; and C third trial Effect of concentration on the isotopic ratio value of 2MPb *07Pb. this noise it is likely that this a.f. noise is caused by vortices formed at the boundary region between the plasma and the surrounding air. ' * . I 3 As the ICP fluctuates at the a.f. there is a possibility that this fluctuation might have some effect when the isotopic measurement is carried out by the quadrupole mass spectrome- ter. Because the quadrupole mass analyser cannot measure iso- topes simultaneously they have to be measured alternately. If the frequency of the measurements occurs in the same time scale as the a.f. noise the precision of the isotopic measure- ments will not be improved. Therefore the mass scanning rate was optimized by changing the dwell time i.e.the mass scan- ning rate was chosen based on observation of the noise- amp1 i tude spectra. Lead Isotope Ratios Versus Concentration The linearity of the instrument was initially investigated. The instrument provides good linearity over a wide range of con- centrations (from 0.1 to 500 ng g-') with a precision of 3%. The detection limit defined as the concentration which gives a signal equal to three times the standard deviation of the background noise was about 10 pg g-' for lead. The isotopic ratio values for 2cbPb ?"Pb '"Pb *07Pb and *OnPb *04Pb and the precision data for these ratios were plotted against the lead concentration in Fig. 4. As expected the precision was poor at the lower concentrations but with increasing concentration the precision improved.When the concentration was greater than 50 ng g-I the precision becomes constant. On the other hand the ratio values appeared to be constant over a wide range of concentrations. However when the 206Pb:207Pb ratios were plotted on an expanded scale for a region of higher con- centration it was observed that the isotopic ratio values were dependent on the concentration (Fig. 5). The experiment was repeated three times and the same trend was observed in each instance. The sample used in this instance was the lead isotop- ic standard (NIST SRM 981) and the zcKPb:z07Pb ratio is certified as 1.0933 with a precision of 0.02%. The ?'"Pb ?07Pb ratio shows a significant increase with increasing concentra- tion from 50 to 500 ng g-I. This means that the concentration of the lead isotopic standard should be approximately the same as that of the sample in order to normalize the instru- ment.Lead Isotope Ratios Versus Dwell Time Fig. 6 shows the ICP-MS spectra obtained by changing the dwell time from 0.64 to 0.005 ms per channel. When the dwell time was too short e.g. 0.01 and 0.005 ms [Fig. 6(g) and ( h ) ] the ICP-MS spectra were shifted and distorted. In such instances the mass scanning rates were too fast for the202 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY APRIL 1991 VOL. 6 t 11.8 ms l- .g - C I 5.9 ms -r I 1.5 ms -I- 204 205 206 207 208 209 mlz Fig. 6 ICP-MS spectra obtained by using different dwell times. The sample was a lead isotopic standard (NIST SRM 98 I ) solution containing 50 ng g-' of lead.The dwell time per channel was ( a ) 0.64; ( h ) 0.32; ((') 0.16; ( d ) 0.08; ( e ) 0.04; v) 0.02; (g) 0.01; and ( h ) 0.0@5 ms. The ion intensity full scale is I .S x 10' counts s-' instrument to acquire signals properly. The figure written in the centre of each frame represents the dwell time for one mass. When the dwell time was set to 0.04 ms per channel the dwell time per mass became 2.96ms (see Table I); this was very close to the time scale of the a.f. noise. The isotop- ic ratio values for 2cmPb ?Pb 2'mPb 207Pb and 20nPb ""Pb and the precision data for these ratios are plotted I'ersus the dwell time in Fig. 7. As is clear in Fig. 7(h) when a longer dwell time was used the precision was poor. With decreas- ing dwell time the precision was improved from about 2 toJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY.APRIL 1991. VOL. 6 203 f I l o t 0 ' I I 1 I I I 41 10 2.56 0.64 0.16 0.04 0.01 Dwell time/ms Fig. 7 Isotope ratios of A 'OxPb:?'MPb; B 2cmPb:?"Pb; and C 206Pb:2"7Pb as a function of the dwell time per channel. ( a ) The average value and (h) the RSD (%) for ten measurements. The lead concentration is 100 ng g-' 1.095 n a 5 1.090 1.085 1.080 ' I 1 I 0.6 0.5 0.4 0.3 0.2 O.' t 0 ' 0.64 0.16 0.04 0.01 1 I I Dwel I ti me/ms Fig. 8 Effect of dwell time on the isotopic ratio value of 'lmPb ?'I7Pb. A. 50; B. 100. and C 500 ng g'. ( u ) The average value and ( h ) the RSD (%) for ten measurements 0.5%. On the other hand the isotopic ratio values are con- stant over a wide range of dwell times.However when the zcmPb:z07Pb ratios are plotted on an expanded scale using a shorter range of dwell times the isotopic ratio values are seen to be dependent on the dwell time [Fig. 8 ( a ) ] . The ex- perimental result was confirmed by using three different con- centrations. This result indicated that once the dwell time was decided the instrument had to be normalized at a fixed dwell time. The precision data for 2cmPb:207Pb ratios are also plotted on an expanded scale in Fig. 8(h). Occasionally a deterioration of the precision was observed at a dwell time of 0.04ms per channel; in such instances the measurement frequency was very close to that of the a.f. noise. However the deterioration was not serious and moreover it was not reproducible. As can be seen in Fig.3 the a.f. noise was ob- served within a narrow frequency range. Therefore if the frequency of the isotopic measurements was slightly different from that of the ICP fluctuation the a.f. noise did not have a serious effect on the isotope ratio measurements. From the data shown in Fig. 8(h) it was decided that the optimum dwell time for the instrument used was 0.08 ms per channel i.e. 6 ms per mass. Conclusions This study did not show clearly that the a.f. noise at 380 Hz caused a deterioration of lead isotope ratio measurements however shorter dwell times and higher concentrations enabled better precision to be obtained. The best precision that could be achieved was 0.13% which was limited by the statis- tical errors of the ion counts. There is a possibility that the pre- cision might be improved by using a longer integration time.It was also found that the isotopic ratio values were dependent on the dwell time and the concentration of lead. Therefore once the dwell time has been optimized the concentration of the lead isotopic standard should be approximately the same as that of the sample in order to normalize the instrument. By using an optimum dwell time of 0.08 ms and a lead concentra- tion of 100 ng g-I the practical precision of lead isotope ratios was about 0.3% for '(fiPb:207Pb and about 0.5% for '06Pb:?0"Pb and 'OXPb ?"Pb. I 2 3 4 5 6 7 8 9 10 1 1 12 13 References Sturges W. T. and Barrie L. A. Nature (London). 1987,329 144. Sturges W. T. and Barrie L. A. Atmos. Emiroti. 1989 23 25 13. Ting B. T. G. and Janghorbani. M. ./. Anal. At. Specrrom. 1988. 3 325. Hinners. T. A. Heithmar. E. M. Spittler T. M. and Henshaw J. M. A n d . Chem. 1987,59 2658. Longerich. H. P. Fryer B. J. and Strong D. F. Specn.ochim. A m . Purt B 1987.42.39. Russ. P. G. 111. and Bazan J. M. Spectrochim. Acru Part B 1987 42.49. Dean J. R. Ebdon. L. and Massey R. J . And. At. Spectrom. 1987 2 369. Campbell M. J. and Delves H. T. ./. Anul. At. Spectrom. 1989. 4 235. Mukai H.. and Ambe Y. Bunseki KupAic. 1990. 39 177. Crain. J. S. Houk. R. S.. and Eckels D. E.. Anul. Chem.. 1989 61 606. Furuta N. Monnig. C. A.. Yang. P.. and Hieftje. G. M.. Spec.fi.ocJiim. Acta Purr B 1989.44 649. Furuta N.. A m / . Sci. 1990.6,683. Winge R. K.. Eckels D. E.. DeKalb. E. L.. and Fassel V. A.. J . Aiial. At. Spc~i.trom.. 19x8. 3 849. Paper- Ol04393A Received Septenihei. 28th. 1990 Accepted No\vniher- 22nd 1990