Evaluation of an Inductively Coupled Air– Argon Plasma as an Ion Source for Mass Spectrometry† HIROSHI UCHIDA*a AND TETSUMASA ITOb aKanagawa Industrial Research Institute, 705–1, Shimoimaizumi, Ebina, Kanagawa 243–04, Japan bSeiko Instruments, 36–1, Takenoshita, Oyama-cho, Sunto-gun, Shizuoka 410–13, Japan An inductively coupled air–argon plasma (air–Ar ICP) has mass discrimination by Xiao and Beauchemin.10 The addition of N2 to the outer, intermediate and aerosol carrier Ar flows been developed using a modified torch and a 40.68 MHz generator (maximum rf power 4 kW), and evaluated as an ion has also been studied separately, mainly to reduce polyatomic ion interferences.11–13 Luie and Soo14 also investigated the source for mass spectrometry (MS).The outer and aerosol carrier Ar flows were completely replaced with air; however, addition of N2 and H2 to the main inlet for the three types of Ar flow and discussed its analytical characteristics. Further, 1.5 l min-1 of Ar was used as the intermediate gas to maintain a stable plasma discharge at low rf power.The optimized Lam and Horlick15 illustrated the eects of varying the sampler –skimmer spacing in a mixed gas ICP with N2 in the outer sampling depth and the aerosol carrier air flow rate for maximum analyte signals were found to be 10 mm above the flow. The addition of O2 to the aerosol carrier flow has frequently been used in order to decompose organic samples load coil and 0.9 l min-1, respectively.The analyte signal increases with rf power, but 2 kW was sucient for the stable in an Ar ICP, an example of which has recently been reported for the determination of lead by ICP-AES.16 The developments discharge and the acceptable analytical sensitivity. In the mass spectra obtained under the optimized conditions, N+, O+ and of molecular gas and mixed gas ICPs in AES have been comprehensively reviewed by Montaser.17 NO+ were clearly observed, but the signal for Ar+ was weak, which is similar to that for an N2 ICP.The ion signals for There have been a few reports on attempts to generate a molecular gas ICP for MS with the addition of a small amount N2+ and O2+ were relatively large, compared with N2 and O2 ICPs operated with Ar added to the outer gas. The analytical of Ar. Tanaka et al.18 reported N2 and O2 ICPs, where both the outer and aerosol carrier Ar flows were completely replaced sensitivity of the proposed air–Ar ICP is superior to an Ar ICP using the same equipment for elements with low first by N2 and O2, respectively.However, Ar remained in the intermediate flow. Another type of N219 or O220 ICP assisted ionization potentials (IP) of <6.5 eV, but inferior for elements with high first IPs (>6.5 eV). The secondary discharge by adding Ar to the outer gas has been achieved using a 40.68 MHz rf generator and a modified torch21 by Uchida and increases the average kinetic energy of the analyte ions, the distribution of which is wider in the air–Ar ICP than in the Ar Ito.Typical analytical characteristics of molecular gas ICP-MS were found and the analytical sensitivity was compared with ICP. The ratios of monoxide ion to singly charged ion remain almost constant along the plasma axis. Space charge eects Ar ICP-MS. In the present work, an attempt was made to generate an from co-existing elements are also discusssed. air–Ar ICP, because of the low running costs, using the same Keywords: Inductively coupled plasma mass spectrometry; air– equipment as for N219 and O220 ICPs, which was then evaluated argon plasma; analytical characteristics; comparison with as an ion source for MS.Mass spectra were obtained under other plasmas optimized plasma conditions, and the analytical sensitivity, kinetic energy distribution, ratios of doubly charged and monoxide ions and matrix eects are also discussed, and Inductively coupled plasma mass spectrometry (ICP-MS) has been used to study the determination of ultratrace elements in compared with Ar19,22, N219 and O220 ICPs.various types of samples.1 The Ar ICP is the most useful ion source, because of its stable discharge at low rf power, high sensitivity for many elements and relative freedom from inter- EXPERIMENTAL ferences. However, mass spectral interferences resulting from Instrumentation and Operating Conditions the plasma-support gas Ar2,3 cannot be avoided, especially in the m/z range from 50 to 80.A high resolution mass spec- The system used consisted of a quadrupole mass spectrometer trometer has been developed to remove mass spectral inter- (SPQ8000A, Seiko Instruments) and a specially constructed ferences caused by polyatomic ions, but the initial cost is high 40.68 MHz rf generator with a maximum power of 4 kW and it requires a large space. Another problem for an Ar ICP (Nippon Koshuha). A description of the load coil, quartz torch is the high running costs because of the Ar consumption.and sampler has been given elsewhere,19,22 including details Mixed molecular gas ICPs have been investigated for the about the equipment used. A shielding system for reduction of reduction of spectroscopic interferences and other matrix the plasma potential23–25 could not be used, because the metal eects. The eects of the addition of N2 and O2 to the central plate inserted between the quartz torch and the load coil Ar flow have been discussed for the reduction of polyatomic melted at an rf power of more than 1.8 kW.ion interferences by Evans and Ebdon4,5 and the addition of The actual operating conditions for the proposed air–Ar N2 by Laborda et al.6 The addition of N2 to the outer Ar flow ICP are listed in Table 1. An Ar ICP was used as a starter for has been studied for the reduction of oxide interferences by the air–Ar ICP and also for a comparative study. The equip- Lam and McLaren,7 the elimination of the co-existing Na8 ment used was the same as for the air–Ar ICP, and the and K9 by Beauchemin and Craig, and for the reduction of operating conditions are also given in Table 1.Analytical characteristics of this Ar ICP are almost equivalent to those of a conventional Ar ICP; however, the detection limit (DL) † Presented at the 1997 European Winter Conference on Plasma Spectrochemistry, Gent, Belgium, January 12–17, 1997. is approximately one order of magnitude inferior.The ion Journal of Analytical Atomic Spectrometry, September 1997, Vol. 12 (913–918) 913Table 1 Operating conditions for the air–Ar and Ar ICPs. The same power in the proposed ICP having outer and aerosol carrier modified torch was used in both ICPs without shielding air flows. On the other hand, the intermediate Ar flow could be replaced by air when a certain amount of Ar was left in the Parameter Air–Ar ICP Ar ICP outer flow, which had previously been adopted in N218 and O219 ICPs using the same equipment.The flow rate of the Forward power/kW 2.0 1.0 Outer gas flow rate/l min-1 Air 20.0 Ar 16.0 intermediate or outer Ar flow in each case could be reduced Intermediate gas flow rate/l min-1 Ar 1.5 Ar 1.5 at high rf power, but the damage to the top of the sampler Aerosol carrier gas flow rate/l min-1 Air 0.9 Ar 0.65 was more pronounced than at low power. Consequently, a Sampling depth/mm above the work coil 10 13 flow of 1.5 l min-1 of Ar was used as the intermediate flow in subsequent experiments, because the Ar consumption was lower and the gas replacement was carried out more easily signal counts obtained from the Ar ICP using the present and rapidly, compared with the addition of Ar to the outer equipment were found to be less than those for conventional flow.19,20 use of an Ar ICP,22 because a 0.5 mm gap between the outer A bright discharge was clearly observed around the top of and intermediate tubes was adopted in the modified torch, and the Cu sampler, which seemed to be caused by a secondary a Cu sampler and skimmer with a thick top were used to discharge. In the MS spectra ( later shown in Fig. 4), 63Cu+ prevent damage caused by the secondary discharge. and 65Cu+ were strongly observed, and these ion signals could not be decreased in the present mass spectrometer without a Gases shielding system and with a load coil grounded at one end. The air used in this work was provided by a compressor with a maximum generation of 2.9 m3 min-1 (Hitachi HISCREW22 Distribution of Analyte Ions Along the Plasma Axis OSP-22E5AR II), and passed three times through an ordinary The spatial distribution of analyte ion signals along the plasma air filter and once through a charcoal filter (Nippon Pall axis should be discussed when optimizing sampling depth. The Filter), and then distributed at a pressure of 7.5 kgf cm-2 to optimized sampling position for the analyte signal was found each laboratory in Kanagawa IRI.The Ar gas used was of to be 6–7 mm above the load coil in N219 and O220 ICPs, and 99.99% purity (Nippon Sanso). 12–14 for the Ar ICP.22 The ion signal clearly decreased in the outer part of the plasma in the N219 and O220 ICPs, and Reagents might depend on cooling by the molecular gas. The results for Co+, Y+, Cd+ and Pb+ obtained using the Stock solutions (10 mg l-1) of the analytes and the matrix for air–Ar ICP are indicated in Fig. 1. The plasma discharge the discussion of co-existing Li, Co and Tl were prepared from became unstable in sampling positions greater than 15 mm commercially available 1000 mg l-1 solutions for AAS (Wako above the load coil. Each element has its own distribution, Pure Chemical or Kanto Chemicals). Working solutions were and the ion signals do not clearly decrease in the outer part diluted from the stocks using pure water from a Milli-Q system of the plasma.Maximum signals were found in a region of (Millipore) in a matrix of 3% ultrapure nitric acid from 9 to 12 mm. The maximum signal position and the (Tamapure-100, Tamakagaku Kogyo). Ultrapure water distribution obtained from the air–Ar ICP seem to be similar (Tamakagaku Kogyo), where the concentration of trace to those of the Ar ICP.22 A sampling position of 10 mm was impurities was less than 10 ng l-1 was used for the measureused subsequently. ments of mass spectra profiles and actual background ion counts.Eect of Rf Power on the Analyte Ion Signal RESULTS AND DISCUSSION An rf power of 2.5 kW was required in the N2 ICP,19 and the plasma discharge was maintained in a range of from 1.8 to Air–Ar Plasma Discharge 2.5 kW in the O2 ICP.20 In the air–Ar ICP, the plasma The Ar ICP was generated as a starter under the conditions discharge could be maintained at lower rf power, but became given in Table 1. After optimization of the ion lens parameters unstable below 1.8 kW.The eects of rf power on the analyte in the mass spectrometer, which were found to be almost ion signals are indicated in Fig. 2. The ion signals of almost equivalent to those for the proposed air–Ar ICP, the sampling all elements gradually increase with an increase in rf power. depth was set at 10 mm above the load coil. Before replacement with air, the outer Ar flow rate was increased to 20 l min-1 and then the rf power to 2.0 kW. The air was introduced to the outer flow with a reduction in the reflected power until the air flow rate reached 20 l min-1, and the Ar flow rate was thus decreased to zero. Shrinkage of the plasma was observed when the air was introduced to the outer flow, and the outer edge of the shrunken plasma was brilliant, which seemed to depend on the Ar in the intermediate or aerosol carrier flow.When the aerosol carrier Ar flow was replaced by air the plasma with the outer flow of air expanded again and the outer edge emission disappeared.The emission from the proposed air–Ar ICP, operated under the conditions shown in Table 1, was observed to be less brilliant than the Ar ICP. An entirely air plasma operated at 2.75 kW rf power with a 40.68 MHz generator was used as an emission source for an air pollution study,26 and a method for the approximation of the electron temperature was applied to the entirely air ICP at 2.2 kW using a 64 MHz generator.27 However, it was very Fig. 1 Distributions of the relative analyte ion signals along the axis: rf power, 2 kW; air carrier flow rate, 0.9 l min-1. dicult to replace the intermediate Ar with the air at low rf 914 Journal of Analytical Atomic Spectrometry, September 1997, Vol. 12of 0.9 l min-1 was used for the carrier gas flow in the air–Ar ICP. Air–Argon ICP Mass Spectra Mass spectra from the proposed air–Ar ICP, obtained under the optimized analytical conditions, are shown in Fig. 4 over the m/z range from 5 to 85.Fig. 4(a) and (b) was obtained by nebulization of ultrapure water with low sensitivity, where analyte ions were loosely focused on the ion detector with a low applied voltage. Signals for N+, O+ and NO+ were clearly observed, and the profile basically seems to be similar to that of an N2 ICP.19 The signal counts for O+, some of which might be caused by the introduced water, were larger than an N2 ICP. The signal counts for N2+ and O2+, clearly observed in Fig. 4(a), were 1.3 and 1.4% N+ and O+, respectively. These Fig. 2 Eect of the rf power on the relative ion signals: air carrier signal ratios of the polyatomic ion to the atomic ion are larger flow rate, 0.9 l min-1; sampling depth, 10 mm above the load coil. than in the N2 (0.25%)19 and O2 (0.19%)20 ICPs, and seem to indicate that the concentrations of N2 and O2 in the air–Ar ICP might be higher than in the N2 and O2 ICPs. The higher However, the top of the sampler was badly damaged by the concentration of molecules could depend on the lower rf power secondary discharge at higher power.The signal ratios of and also on the Ar in the intermediate gas. The Ar+ signal, as doubly charged ions to singly charged ions slightly increased shown in Fig. 4(b), is weakly observed, because the first IP of with increasing rf power, but the ratios of monoxide ions to Ar is relatively high (15.76 eV). The actual concentration of singly charged ions were kept almost constant.An rf power of Ar could be higher. 2.0 kW was used in the latter experiment, since the ion signals Fig. 4(c) was also obtained from nebulization of ultrapure obtained were sucient to detect trace amounts of analyte. water with relatively high sensitivity, where analyte ions were tightly focused on the ion detector with a high applied voltage. Eect of the Aerosol Carrier Gas Flow Rate on the Analyte Ion The detector was masked in the m/z range 14–18, 29–31, 40 Signal and 41 in order to avoid damaging the detector.Several types of polyatomic ions of the component elements of air, e.g., The aerosol carrier (central ) gas flow rate is one of the most N2O+ and NO2+, were observed. The sampler might be more significant factors aecting the analytical characteristics of an damaged by the secondary discharge than in an N2 ICP19, ICP. A higher carrier gas flow rate gives a higher eciency since CuN+ and CuO+ were observed with the addition of for the introduction of the sample solution into the plasma, strong Cu+ signals.but sometimes causes chemical and ionization interferences. 28,29 Further, the analyte cannot acquire sucient energy for excitation or ionization at higher flow rates. The Sensitivity and Detection Limit maximum signals were found at 0.5–1.0 l min-1 in both Ar The sensitivities and DL of the air–Ar ICP were compared ICP-AES and ICP-MS, but higher carrier flow rates of 1.6 with the Ar ICP. The signals in both ICPs were measured and 1.4 l min-1 were adopted in N219 ICP and O220 ICP-MS, under the operating condition shown in Table 1, and using the respectively, in order to extend the hot region available for the same nebulizer, quartz torch, sampler, skimmer and mass ion source.The results obtained are shown in Fig. 3. Maximum signals were found at 0.9 l min-1 for Cd+, Co+ and Y+, and at 1.2 l min-1 for Pb+. These flow rates are a little higher than in an Ar-ICP,22 but lower than in N219 and O220 ICPs.The distributions are similar to those of an Ar ICP. The small hot region available for the ion source assumed in N2 and O2 ICPs does not seem to be distinct in the air–Ar ICP. A rate Fig. 4 Mass spectra of the air-Ar ICP under optimized conditions: (a) and (b) obtained under low sensitivity conditions; (c) obtained under Fig. 3 Eect of the air aerosol carrier flow rate on the relative ion high sensitivity conditions with masked detector in the m/z ranges 14–18, 29–31, and 40–41.signals: rf power, 2 kW; sampling depth, 10 mm above the load coil. Journal of Analytical Atomic Spectrometry, September 1997, Vol. 12 915spectrometer. The sensitivity was obtained as counts per second measurements for 56Fe+, 75As+ and 80Se+ could be easily carried out, because of the lower concentration of Ar in the per ppb in each ICP. The Ar ICP sensitivities were almost at the same level as previous work with N219 and O220 ICP-MS air–Ar ICP.On the other hand, signal detection for 77Se+ was dicult because of the formation of 77CuN+, as shown in arrangements. Ultrapure water was used for the measurement of the actual background signal for each element. The DL was Fig. 4(c). The DLs calculated for 30 elements were found to be approximately 1–2 orders of magnitude inferior to those in calculated as the concentration corresponding to a signal-tobackground noise ratio of three. Calculated DLs in the air–Ar the Ar ICP.ICP are listed in Table 2. The ratios of the sensitivity, background signal and calculated DL obtained from both ICPs Kinetic Energy Distribution of the Analyte Ions are also summarized in Table 2, together with first and second IP values30 and also oxide dissociation energy values.31 Ion kinetic energy measurements yielded interesting information about the ICP-MS characteristics.32,33 The ion kinetic The sensitivity seems to depend on the first IP,30 and to be higher than in the Ar ICP for elements with a low IP, energy can be discussed by use of the retarding dc quadrupole rod bias, since the ion lens system in the present mass particularly for Sr, Y, In, Ba and rare earth elements.However, the sensitivity was found to be lower for elements with a high spectrometer does not act as an energy filter. The normalized ion counts for Be+, Co+, Y+, Cd+ and Pb+ versus the IP, especially Pt, Au and Hg. These results are typical of molecular gas ICPs.19,20 Furthermore, the sensitivity seems to retarding voltage in the air–Ar ICP are shown in Fig. 5, together with that for Pb+ in the Ar ICP (broken line) for decrease slightly with increasing m/z values, which was observed in an O2 ICP with the same equipment.20 Although comparison. The energy distributions of the other elements in the Ar ICP were almost equivalent to that of Pb+. the oxide dissociation energy is high, the sensitivity for an element with a low first IP was found to be high, as shown The average energy values, obtained at a relative signal of 50 for five elements, were found to be 18–28 eV in the air–Ar for Ce+ and La+.The background signals were found to be higher in the ICP and 24 eV for Pb+ in the Ar ICP. These higher values in both ICPs seem to depend on the secondary discharge, since air–Ar ICP than the Ar ICP, but the RSD values were calculated to be in the range 0.5–2%, which is almost equal an average of 6 eV was found in the Ar ICP with a shielding system.22 Even if the average kinetic energy values are almost to those of the Ar ICP.One of the reasons might be the increase in pressure in the ion lens chamber, when the Ar is at the same level in both ICPs, the width of the energy distribution for each ion, calculated as the dierence between replaced with air. The increase in the background signal does not seem to depend on contamination of the air used, since the two retarding voltage values at relative ion signals of 100 and 0, was found to be approximately 1.5 times larger in the the background signal was found to be almost at same level when commercially available purified air (Nippon Sanso) was air–Ar ICP than in the Ar ICP, as shown in Fig. 5. In addition, each ion distribution curve shifts to a higher voltage with used instead of the original air used in this study. Ion signal Table2 Comparison of the sensitivities, background and DLs for the air–Ar and Ar ICPs DL Oxide† Sensitivity‡ Background Abundance First* Second* dis.en./ ratio signal ratio Air–Ar ICP Ratio Element m/z (%) IP/eV IP/eV eV Air–Ar:Ar Air–Ar:Ar (ppt)§ Air–Ar:Ar Be 9 100 9.32 18.21 4.6 0.40 170 20 290 Ti 48 73.8 6.82 13.57 7.2 1.1 11 60 10 V 51 99.8 6.74 14.65 6.4 1.0 4.8 1 1.7 Cr 52 83.8 6.76 16.49 4.4 0.92 4.0 10 3.3 Mn 55 100 7.43 15.64 4.2 0.48 21 80 40 Fe 56 91.7 7.86 16.18 4.3 0.42 0.33 1000 0.50 57 5.8 0.45 0.14 70 0.78 Ni 58 67.8 7.63 18.15 <4.2 0.21 0.53 10 1.3 Co 59 100 7.86 17.05 – 0.29 130 20 220 Zn 64 27.9 9.39 17.96 2.8 0.15 3.1 100 20 Ga 69 60.1 6.00 20.51 3.0 1.1 20 1 10 As 75 100 9.81 18.63 4.9 0.092 13 40 40 Se 80 49.6 9.75 21.5 4.3 0.045 0.99 700 1.0 Sr 88 87.9 5.69 11.03 4.2 3.0 55 0.6 6.0 Y 89 100 6.38 12.33 7.31 2.0 400 4 50 Zr 90 51.4 6.84 13.13 7.8 0.85 10 2 5 Mo 98 24.1 7.10 16.15 5.0 0.35 150 50 100 Pd 106 27.3 8.33 19.42 – 0.17 61 20 200 Ag 107 51.8 7.57 21.48 2 0.25 140 20 100 Cd 114 28.7 8.99 16.90 <3.8 0.097 12 20 40 In 115 95.7 5.79 18.66 3.3 1.7 39 0.9 9.0 Ba 138 71.7 5.79 10.00 5.75 3.3 60 2 10 La 139 99.9 5.58 11.43 8.2 2.2 6.8 0.4 2.0 Ce 140 88.5 5.47 10.00 8.03 1.7 39 0.6 6.0 Pt 195 33.8 9.0 18.56 – 0.015 10 100 250 Au 197 100 9.22 20.5 – 0.0038 5.5 400 700 Hg 200 23.1 10.43 18.75 – 0.010 1.1 300 100 Pb 208 52.4 7.42 15.03 3.87 0.22 1.9 10 10 Bi 209 100 7.29 16.68 3.1 0.16 23 5 71 Th 232 100 6.95 – 8.5 0.13 15 3 60 U 238 99.3 6.1 – 7.8 0.081 5.9 4 40 * IP, ionization potential from ref. 30. † Oxide dis. en., oxide dissociation energy from ref. 31. ‡ Sensitivity was obtained as counts per second per ppb. § Ppt: parts per 1012. 916 Journal of Analytical Atomic Spectrometry, September 1997, Vol. 12Fig. 7 Signal ratio distributions of monoxide ions to singly charged Fig. 5 Distributions of the analytical ion kinetic energy in the air–Ar ions: continuous line, air–Ar ICP; broken line, Ar ICP. and Ar ICPs: continuous line, air–Ar ICP; broken line, Ar ICP.Signal Ratio Distribution of Monoxide Ions increasing m/z values, which was observed in the Ar ICP with The distributions of the signal ratios of monoxide ions to the shielding system,22 but not clearly observed in the Ar, N2 singly charged ions are shown in Fig. 7. The ratio seems to and O2 ICPs without shielding.19,20 depend on the oxide dissociation energy value,31 also listed in Table 2, in the air–Ar and Ar ICPs. The ratio increases with Signal Ratio Distribution of Doubly Charged Ions decreasing sampling position above the load coil in the Ar ICP, which could depend on the eect of the introduced water.The concentrations of doubly charged and monoxide ions are However, the ratios for three elements remain constant along also interesting with respect to analytical characteristics, in the plasma axis in the air–Ar ICP. The ratios are almost at spite of the interferences encountered in ultratrace element the same level in both ICPs for Ce, which has a higher oxide determinations by ICP-MS.The distributions of the signal dissociation energy. The ratio for Pb, with a lower dissociation ratio of doubly charged ion to singly charged ion for Ba, Y energy, is much higher in the air–Ar ICP than the Ar ICP, as and Pb in the air–Ar and Ar ICPs are shown in Fig. 6. The shown in Fig. 7. values of the second IPs30 are also listed in Table 2. The ratio in the Ar ICP clearly increases with increasing sampling depth above the load coil, however, it seems to be relatively constant Matrix Eects along the axis of the plasma in the air–Ar ICP.The ratio for Matrix eects on the analyte ion signals were investigated with the three elements clearly depends on the second IP values in co-existing Li (m/z=9, first IP=5.39 eV), Ag (listed in Table 2) the Ar ICP, but the ratio for Pb (second IP=15.03 eV) seems and Tl (200, 6.11 eV). Results obtained for Co+ and Pb+ as to be almost the same as that of Y (12.23 eV) in the air–Ar ICP.the analytes are shown in Fig. 8. As a 10 mg l-1 solution of Co The ratio for Ba and Y, whose second IP values are relatively and Pb was used, the matrix concentration was 1000–100 000 low, is lower in the air–Ar ICP than in the Ar ICP. In contrast, times that of the analyte. The analyte signal decreases with an the ratio is higher in the air–Ar ICP for Pb, which has a increase in the matrix concentration. As shown in Fig. 8, relatively higher second IP value.These phenomena were also greater signal suppression with co-existing Li was found, observed in N219 and O220 ICPs. As shown in Fig. 5, the compared with Ag or Tl. The signal for Pb+ was suppressed kinetic energy distribution in the air–Ar ICP is wider than in more than Co+ with co-existing Li, Ag and Tl. the Ar ICP. A high ratio of doubly to singly charged ions A decrease in the sample introduction rate to the plasma, could be linked to a high concentration of higher energy ion caused by an increase in the matrix concentration, will prob- species.Similar results were observed for Mo which also has ably not aect the signal suppression significantly, since the a higher second IP (16.15 eV). emission intensity obtained was reduced less when the same Fig. 6 Signal ratio distributions of doubly charged ion to singly Fig. 8 Eects of the co-existing elements, Li, Ag and Tl, on the ion signals of Co+ and Pb+; elements in parentheses, co-existing.charged ion: continuous line, air-Ar ICP; broken line, Ar ICP. Journal of Analytical Atomic Spectrometry, September 1997, Vol. 12 9177 Lam, J. W., and McLaren, J. W., J. Anal. At. Spectrom., 1990, introduction system was used in ICP-AES. Another inter- 5, 419. ference is the disturbance of the ion beam path through the 8 Beauchemin, D., and Craig, J. M., Spectrochim. Acta, Part B, ion optics and mass spectrometer, which is known as the space 1991, 46, 603. charge eect.It has been reported that this eect is readily 9 Craig, J. M., and Beauchemin, D., J. Anal. At. Spectrom., 1992, observed in the determination of light elements or in the 7, 937. 10 Xiao, G., and Beauchemin, D., J. Anal. At. Spectrom., 1994, 7, 509. presence of co-existing heavy elements.34 Suppression with 11 Lam, J. W., and Horlick, G., Spectrochim. Acta, Part B, 1990, co-existing Ag and Tl seems to be almost at the same level as 45, 1313. Li or less, if mass concentration is converted into atom 12 Hill, S.J., Ford, M. J., and Ebdon, L., J. Anal. At. Spectrom., concentration. Mass concentrations of 800 mg l-1 of Li, Ag 1992, 7, 719. and Tl correspond to atom concentrations of 6.9×1019, 13 Wang, J., Evans, E. H., and Caruso, J. A., J. Anal. At. Spectrom., 4.5×1018 and 2.4×1018 l-1, respectively. As the atom concen- 1992, 7, 929. 14 Louie, H., and Soo, S. Y.-P., J. Anal. At. Spectrom., 1992, 7, 557. tration of Co is also calculated to be 3.5 times that of Pb, the 15 Lam, J.W., and Horlick, G., Spectrochim. Acta, Part B, 1990, space charge eect should be more severely observed for Pb+ 45, 1327. in the presence of Li, Ag and Tl, as shown in Fig. 8. Ionization 16 Brenner, I. B., Zander, A., Kim, S., and Shkolnik, J., J. Anal. At. interference with co-existing Li seems to contribute less to the Spectrom., 1996, 11, 91. signal suppression, since the carrier flow rate of the air is 17 Montaser, A., and Gilightly, D.W., Inductively Coupled Plasmas in Analytical Atomic Spectrometry, VCH, New York, 1987. moderate in the air ICP.28,29 18 Tanaka, T., Yonemura, K., Obara, K., and Kawaguchi, H., Anal. Sci., 1993, 9, 765. 19 Uchida, H., and Ito, T., J. Anal. At. Spectrom., 1995, 10, 843. CONCLUSIONS 20 Uchida, H., and Ito, T., Anal. Sci., 1997, 13, 391. 21 Yang, P., Barnes, R. M., Vechiarelli, J., and Uden, P. C., Appl. An economical air–Ar mixed ICP (1.5 l min-1 of Ar in the Spectrosc., 1990, 44, 531.intermediate gas flow) was achieved at the relatively low rf 22 Uchida, H., and Ito, T., J. Anal. At. Spectrom., 1994, 9, 1001. power of 2 kw using a 40 MHz generator and a modified 23 Gray, A. L., J. Anal. At. Spectrom., 1986, 1, 247. torch. Analytical sensitivity depends on the IP value of the 24 Nonose, N. S., Matsuda, N., Fudakawa, N., and Kubota, M., analyte, and the DLs calculated were approximately 1–2 orders Spectrochim. Acta, Part B, 1994, 49, 955. 25 Sakata, K., and Kawabata, K., Spectrochim. Acta, Part B, 1994, of magnitude inferior to those of an Ar ICP.The analytical 49, 1027. characteristics are basically the same as those of molecular gas 26 Baldwin, D. P., Zamzow, D. S., and D’Silva, A. P., J. Air Waste ICPs, however, some characteristics were as for an Ar ICP, Manage. Assoc., 1995, 45, 789. i.e., the stable plasma discharge, analyte ion signal distribution, 27 Gomes, A.-M., Sarrette, J. P., Madon, L., and Epifanie, A., J. Anal. the eect of carrier gas flow rate and the mass spectra. The At. Spectrom., 1995, 10, 923. 28 Kalnicky, D. J., Fassel, V. A., and Kniseley, R. W., Appl. proposed plasma should be termed an air–Ar mixed ICP as Spectrosc., 1977, 31, 137. regards its analytical characteristics. A study of the application 29 Kosinsky, M. A., Uchida, H., and Winefordner, J. D., Anal. of an air–Ar ICP to practical analyses is in progress. Chem., 1983, 55, 688. 30 Moore, C. E., Atomic Energy L evels, Circular No. 467, National Bureau of Standards, Washington DC, 1949. REFERENCES 31 Gaydon, A. G., Dissociation Energy and Spectra of Diatomic Molecules, Chapman and Hall, London, 1968. 1 Date, A. R., and Gray, A. L., Application of Inductively Coupled 32 Olivares, J. A., and Houk, R. S., Appl. Spectrosc., 1985, 39, 1070. Plasma Mass Spectrometry, Blackie, Glasgow, 1989. 33 Gray, A. L., and Williams, J. G., J. Anal. At. Spectrom., 1987, 2, 599. 2 Tan, S. H., and Horlick, G., Appl. Spectrosc., 1986, 40, 445. 34 Tan, S. H., and Horlick, G., J. Anal. At. Spectrom., 1987, 2, 745. 3 Vaughan, M. A., and Horlick, G., Appl. Spectrosc., 1986, 40, 434. 4 Evans, E. H., and Ebdon, L., J. Anal. At. Spectrom., 1989, 4, 299. Paper 7/01269A 5 Evans, E. H., and Ebdon, L., J. Anal. At. Spectrom., 1990, 5, 425. Received February 24, 1997 6 Laborda, F., Baxter, M. J., Crews, H. M., and Dennis, J., J. Anal. At. Spectrom., 1994, 9, 727. Accepted June 10, 1997 918 Journal of Analytical Atomic Spectrometry, September 1997, Vol. 12