JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1994 VOL. 9 61 Dissociation of Analyte Oxide Ions in Inductively Coupled Plasma Mass Spectrometry E. Poussel and Jean-Michel Mermet Laboratoire des Sciences Analytiques Unite de Recherche Associee au Centre National de la Recherche Scientifique (URA 435) Batiment 308 Universite Claude Bernard-Lyon I 69622 Villeurbanne Cedex France D. Deruaz Laboratoire d'Etudes Analytiques et Cinetiques du Medicament Universite Claude Bernard Lyon I 69373 Lyon Cedex 08 France Analyte oxide ions have been studied in inductively coupled plasma mass spectrometry as a function of the sampling position the carrier gas flow rate and the efficiency of the energy transfer from the plasma to the sample in order to support the hypothesis of analyte oxide formation in the plasma.Lanthanum was selected as the test species owing to its high oxide bond strength and its behaviour was compared to that of Pb as this element has a low oxide bond strength. A prototype ICP mass spectrometer was used. The Laof Laf ratio was found to be in the 0.2-13000°/~ range under the operating conditions used. Energy transfer was either degraded by adding a sheathing gas or improved by adding hydrogen. The LaOf:La+ ratio was minimized for low carrier gas flow rates and an efficient energy transfer. This was corroborated by temperature optical measurements. The results are in good agreement with those found in other recent work. In any instance the role of the sampling position was found to be crucial in studying analyte oxide ion behaviour and optimizing analytical performance.Keywords Inductively coupled plasma; mass spectrometry; analyte oxide ion; mixed-gas plasma Inductively coupled plasma mass spectrometry (ICP-MS) has gained general acceptance because of excellent limits of detec- tion for most elements of the Periodic Table and the capability of measuring isotopic abundances. However ICP-MS suffers from some limitations. Among them the presence of isobaric interferences,' particularly those produced by analyte oxide ions MO' and argon oxide ions ArO+ which is a severe Possible isobaric interferences due to oxide ions have been reported.6 To a large extent the formation of oxide ions is made possible because of the presence of oxygen resulting from the dissociation of water contained in the aerosol.The amount of oxide can be reduced by using a sample introduction system that does not introduce water into the plasma such as electrothermal vaporization ( ETV).31,32 Solvent loading can be reduced by using a water-cooled spray ~ h a m b e r ~ ' ~ ~ or minimized with the use of a desolvation ~ y s t e m . ~ ~ ~ ~ The origin of oxygen is clear but the processes of analyte oxide formation are still the object of debate. Several explanations have been suggested (i) analyte oxides are formed in the plasma and are not fully d i ~ s o c i a t e d ; ~ ' ~ ' ~ ' ~ ~ (ii) they are formed in the boundary l a ~ e r ; ~ ~ and (iii) they are formed in the interface region between the sampler and the skimmer.'6*'8 Because of the large orifices currently used for the sampler it does not seem that the formation of oxides in the boundary layer is a predominant p r o ~ e s s .~ From the work of Douglas and it is unlikely that the formation of oxide occurs in the interface region. It is clear that analyte oxides are already present in the plasma itself as it is common knowledge that injection of rare earth elements (REE) into the plasma leads to the emission of oxide molecular bands. Most of the REE exhibit emission in the visible part of the spectrum. Spatial distribution of MO bandheads has been d e ~ c r i b e d ~ ~ . ~ ~ using emission spectrometry. Therefore it is likely that the observation of MO' is the result of both easy formation of the oxide along with incomplete further dissociation of the oxide in the plasma.It has been demonstrated that the MO+:M+ ratio is dependent on the operating parameters of the plasma such as the the carrier gas flow rate4-9,14,15,20,22 and the sampling position defined as the distance between the load coil and the aperture of the ~ a m p l e r . ~ * ~ * ~ . ' ~ . ~ ~ Usually an increase in the power and a decrease in the carrier gas flow rate result in reduction of oxide ion intensities which indicates the influence of the energy transfer from the plasma to the sample. However most of the experiments are performed at a fixed sampling position because of the difficulty of re-optimizing the lens voltages of the ion optics. Change in the carrier gas flow rate corresponds to both a variation in the amount of sample and the residence time.This results in a different optimum sampling position which is usually not taken into account. The MO+ M+ ratio can vary over a large range 0.001-0.3% for PbO+ Pb+,7.15916 0.63-360% for and 0.38-2% for UO+ U+.2,14,18*21 The MO+ M' ratio is also dependent on the oxide bond strength. Usually the greater the oxide bond strength the higher the ratio.",16,17,19,20'22 Addition of molecular gases35*39-43 with high thermal conduc- tivity such as N or H2 reduces the amount of oxide ions which is further evidence of the significant role of the energy transfer. To improve the understanding of analyte oxide formation and dissociation in the plasma investigations into the role of the efficiency of energy transfer have been undertaken. To obtain the spatial distribution of the various species scanning of the sampling position was performed for each value of the gas flow rates of concern. The efficiency of energy transfer was modified by varying the carrier gas flow rate.Furthermore the efficiency was either degraded by using a sheathing gas or improved by adding a molecular gas such as H,. This study was based on the use of a prototype TCP mass spectrometer designed to permit the user to vary easily the carrier gas flow rate and the sampling position. CeO+ Ce+,11,13,15,16,18,22 3.6-2400y0 for Lao+ La+,11.13.22 Experimental The prototype was based on the use of a 56 MHz ICP tuned- line generator and a Nermag R 1010 C mass spectrometer. The oscillator consisted of two parts (i) the first circuit where the high-frequency triode was set-up; and (ii) the second circuit where the coil was located.The transfer of energy between the two circuits was achieved by electromagnetic coupling through lines. There was therefore no grounding of the coil. The second circuit could move in front of the first one allowing62 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1994 40L. 9 the coil and consequently the torch to be positioned at various sampling positions through an x-y-z translation system. This ICP-MS system has been described p r e v i o u ~ l y . ~ ~ ~ ~ ~ The major features were (i) the use of a 6-turn load coil (26 mm long with an i.d. of 22mm) with a floating potential which resulted in the absence of any arcing at the sampler level; and (ii) very simple ion optics that consisted of a single 20 mm diameter high voltage extraction lens set up behind the skimmer and a quadrupole inlet ion.No photon stop was used. The extraction lens voltage was of the order of - 500 V. It was not found necessary to re-adjust the lens voltage for a variation in the carrier gas flow rate in the 0.5-1.0 1 min-' range. The sampler was made of nickel and had a 0.6mm diameter orifice. The skimmer was made of stainless steel and was 1 mm in diameter. The space between the sample and the skimmer was 8 mm. A cryogenic pump was used at the ion optics level (Edwards 1500 1 s-I). An analogue mode of detec- tion was used with a channeltron detector. The background was equivalent to 200 counts s-' if a digital readout was used. The operating parameters were a power of 1.2 kW an outer gas flow rate of 11 1 min-' a carrier gas flow rate in the 0.5-1 lmin-I range a sheathing gas flow rate in the 0-0.4 1 min-l range and a sampling position in the 2-35 mm range.When added hydrogen (0.02 1 min-') was mixed with the carrier gas. The Ar carrier and sheathing gas flow rates and the addition of H2 were controlled by Brooks 5878 mass flow controllers. A Meinhard type C pneumatic nebulizer was used in association with a double-pass spray chamber main- tained at room temperature to avoid drift of the oxide forma- tion. The inner diameter of the torch injector was 2 mm. A 1 m Jobin-Yvon THR monochromator equipped with a 3600 lines mm- holographic grating was used for temperature measurements by emission spectrometry. The practical reso- lution using slit-widths of 10pm and the second order of the grating was 1.7pm at 230nm.Side-on observation of the horizontal plasma was carried out through a Dove prism and a lens so that the axis of the plasma was vertically rotated and imaged on the vertical entrance slit of the monochromator. Use of a horizontal slit at the entrance slit level allowed a portion of the viewing zone to be selected. Temperature measurements were conducted based on the use of the Bolztmann plot of Ti I1 lines46 and the Doppler width of the Be I 234.86 nm line.47 Lead and La were selected as the test elements for this work based on the measurement of the 208 and 139 isotope signals respectively. They were selected owing to their large difference in oxide bond strength 4.32 eV and 8.24 eV respectively.48 Reactions leading to the formation of La' and Lao' and energies of dissociation of L a 0 and Lao' and ionization of La and L a 0 are summarized in Fig.l.48 Results Because of its low oxide bond strength PbO' is generally not observable in contrast to Lao' which is easily observed owing to its high bond strength. The behaviour of La+ Lao' and Pb' are therefore compared here. La0 La+O Lao' \ La+ +O Fig. 1 of ionization of La and La0 Energies (eV) of dissociation of L a 0 and Lao' and energies Effect of Carrier Gas Flow Rate and Sampling Position on Intensity of Pb' La' and Lao' An example of the effect of sampling position on the signal of Pb' for carrier gas flow rates in the 0.5-1.0 1 min-' range is shown in Fig. 2. For each flow rate there is an optimum of the sampling position leading to the maximum signal of Pb'.The location of the optimum is shifted away from the load coil as a function of the flow rate. The highest Pb' signal is obtained for a flow rate of 1.0 1 min-' and a sampling position of 8 mm. Even at 1.0 1 min-' the residence time is sufficiently long to fully dissociate PbO so that the maximum signal corresponds to the maximum amount of sample. The shift is explained by the higher velocity of the species in the plasma. However a different conclusion is observed for La+ where the formation of Lao' leads to a deterioration of the La' signal at high flow rates. The effect of sampling position on the signal of La' and Lao' for two extreme carrier gas flow rates i.e. 0.6 and 1.0 1 min-' is shown in Fig.3. In contrast to the results obtained for Pb' an increase in the amount of sample does not correspond to an increase in the La' signal owing to the presence of L a o + . Fig. 3 shows that the signal of Lao' is rather small at a carrier gas flow rate of 0.6 1 min-' whereas Lao' becomes the largest species at 1.0 1 min-'. Fig. 3 also shows that the formation zone of Lao' is closer to the injector than that of La'. From the same experiment it is possible to plot the L a o + La' ratio (Fig. 4) as a function of the sampling position for several carrier gas flow rates. The conditions for the observation of the optimum La' signal are also shown in Fig. 4. Under the operating conditions used for this experiment the Lao' :La+ ratio varies in the 0.2-13 000% range.The lowest La0':La' ratio is obtained for a carrier I 2 6 10 14 18 22 26 30 Sampling position/mm Fig. 2 Effect of the sampling position on the ion signal of Pb' for various carrier gas flow rates A 1.0; B 0.8; C 0.6; and D 0.5 1 min-' 1400 1200 - m .- u) .- 5 1000 a -5 ;s 800 C 5 I I \ \ \ 2 6 10 14 18 22 Sampling position/mm Fig.3 Effect of the sampling depth on the ion signal of La' (solid line) and Lao' (broken line) for two carrier gas flow rates D 0.6; and 0 1.01 min-lJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1994 VOL. 9 63 1 x105 1 x 1 0 - ' 1 I I I I Sampling position/mm 1 5 9 13 17 21 Fig. 4 Effect of the sampling depth on the L a o t :La' ratio (%) for various carrier gas flow rates A 1.0; B 0.9; C 0.8; D 0.7; and E 0.6 1 min-'.The broken line is used to indicate the location of the maximum La+ signal gas flow rate of 0.6 1 min- ' i.e. the longest residence time and a sampling position of 12 mm. These conditions are far from being optimum for the determination of Pb'. When different species with a large range of oxide bond strengths are to be determined it is difficult to define compromise conditions without sacrificing either species with low oxide bond strengths or those with high ones. Examples are given in Table 1. Under the operating conditions used to obtain the lowest L a o + La+ ratio 10 and 30% of the maximum signal of Pb' and La+ are observed respectively. The optimum conditions for the La' signal lead to 30% of the maximum signal of Pb' whereas those for the optimum of the Pb+ signal lead to 15% of the maximum signal of La'.Nevertheless compromise conditions can be defined (Table l) which lead to at least 50% of the maximum analyte ion signals without too large an Lao+ signal. As stated earlier the increase in the Pb' signal is directly related to the increase in the amount of aerosol up to a carrier gas flow rate of 11 min-' because of the lack of oxide formation. Fig. 5 shows that there is good agreement between the signal of Pb+ and the amount of aerosol measured by trapping the aerosol on silica gel. The signal was integrated over the 2-30mm sampling position range. In contrast the same agreement is obtained for the La-containing species only when summing the Lao' and La' signals (Fig. 5). Clearly an increase in the carrier gas flow rate i.e.a decrease in the residence time and a degradation of the energy transfer efficiency enhances the formation of analyte oxides with a high bond strength. It is also clear that the study of the influence of the carrier gas flow rates for fixed values of the sampling position can lead to misleading conclusions. From the previous results it can be demonstrated that for a variation of the carrier gas flow rate between 0.8 and 1.0 1 min-' Lao' increases and La+ decreases for a 7mm sampling position whereas both Lao' and La+ decrease for a 3 mm sampling position. Therefore the sampling position is a significant parameter although its variation is not always easy to perform 0 ' I I I 1 ' 0 0.5 0.6 0.7 0.8 0.9 1 Carrier gas flow rate/l min ' Fig. 5 Comparison of the amount of aerosol (A) the Pb+ signal (B) the La' signal (C) and the sum of the La+ and Lao' signals (D) as a function of the carrier gas flow rate.Signals are integrated over a sampling position range of 2-30 mm with commercially available systems owing to the need for re-optimization of the ion optics lens voltages. Effect of the Sheathing Gas To confirm the drastic role of the efficiency of energy transfer on analyte oxide observation a sheathing gas has been added at the exit of the spray chamber. The main use of the addition of a sheathing gas is to prevent the deposition of material on the inner part of the injector. However the addition of a sheathing gas results also in a degradation of the efficiency of energy transfer,49 owing to the formation of a layer between the plasma and the central channel. The effect of the sampling position on the La' and Lao+ signal for a carrier gas flow rate of 0.6 or 1.Olmin-' and a sheathing gas flow rate of 0.4 1 min-' added to a carrier gas flow rate of 0.6 1 min-l is illustrated in Figs. 6 and 7.The addition of 0.41 min-' sheathing gas to 0.6 1 min-' of carrier gas does not modify the amount of sample obtained at 0.6 1 min-l without sheathing gas. However the degradation of the energy transfer has a Q 5 60 C 0 .- 40 m > Q .- L .- > 20 + - Q CT 2 7 12 17 22 27 32 37 Sampling position/mm Fig. 6 Effect of the sampling position on the La' signal for various carrier gas flow rates A 0.6; B 1.0; and C 0.6 1 min-' +0.4 1 min-' of sheathing gas. Signals were normalized to the maximum La+ signal Table 1 to different optimizations Percentage values of the maximum signal of Pb+ La' and Lao+ for various carrier gas flow rates and sampling positions corresponding Carrier gas/ Sampling Optimization 1 min-' position/mm Max Pb' (%) Max La+ (YO) Max Lao+ (YO) Lowest Lao + La + 0.6 Optimum La+ 0.6 Compromise conditions 0.8 Optimum Pb' 1 .o 12 5.5 8 7 10 30 100 50 30 100 15 62.5 0.025 1 46 9.264 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1994 VOL.9 - 100 .- Ln t 2 80 -I a 5 + 60 C 0 .- .I- L m 40 .- m !? 20 .- +- m a a - 0 Sa m pl i ng posit io n/m m Fig.7 Effect of the sampling depth on the Lao' signal for various carrier gas flow rates A 0.6; B 1.0; and C 0.6 1 min-' +0.4 1 min-' of sheathing gas. Signals were normalized to the maximum La+ signal drastic effect on both La' and Lao' signals.The peak intensity of La' is strongly decreased and shifted to a larger sampling position (Fig. 6) in marked contrast to the behaviour of Lao' (Fig. 7) whose intensity is strongly enhanced. For the same residence time i.e. a total flow rate of 1 1 min-' the intensity of La' also decreases and shifts because of a smaller amount of sample and a degradation of the energy transfer. Therefore both the residence time and the efficiency of the transfer of energy to the sample have a significant influence on the observation of analyte oxide. Addition of Hydrogen Argon exhibits poor thermal conductivity. Among molecular gases such as nitrogen oxygen and hydrogen hydrogen exhib- its the highest thermal conductivity below 10 000 K because of its dissociation near 3700 K." Potential advantages of adding hydrogen to improve the thermal conductivity of the plasma and consequently to enhance the efficiency of energy transfer have been discussed recently to aid the volatilization of slurrie~.~' Hydrogen requires additional energy for dis- sociation but does not modify the resistivity of the plasma and does not add new information to the ion spectrum as it is already present in the plasma.Molecular gases can be mixed with the outer gas resulting in shrinkage of the discharge and a corresponding increase in the temperature of the central channel. Alternatively molecular gases can be added to either the sheathing gas or the aerosol carrier gas. It has been demonstrated in atomic emission spectrometry that the addition of a small amount of hydrogen to the carrier gas or the sheathing gas can improve the transfer of energy particu- larly when the transfer has been degraded by having a short residence time or with the use of a sheathing gas.52 Addition of 0.021min-' of hydrogen was sufficient to modify signifi- cantly the spatial distribution of La+ and Lao+.It can be seen in Fig. 8 that the addition of hydrogen increases the rate of oxide dissociation so that the optimum of the Lao'. signal corresponds to a sampling position of t 2 mm i.e. in a region where the sampler cannot be located. The optimum of the La' signal is also observed (Fig. 9) for a lower value of the sampling position. It is important to note that similarly to the addition of a sheathing gas addition of hydrogen requires a re- adjustment of the sampling position to obtain the optimum conditions.Otherwise false conclusions might be deduced about the influence of the addition of hydrogen. Temperature Measurements The magnitude of ion signals of analyte oxides with a high bond strength is related to the temperature of the plasma. In order to relate the temperature of the central channel to the - 100 07 v) .- + % 80 1 0 5 60 u- 0 C ,g 40 m m .- L ; 20 .- 4- - Q U 0 5 10 15 20 25 30 35 Sampling positionhnm Fig. 8 Effect of the sampling position on the Lao+ signal for various carrier gas flow rates A 1.0; and B 0.98 1 min-' +0.02 1 min-' of hydrogen 0 5 10 15 20 25 30 35 Sam pl i ng posit i on/m m Fig.9 Effect of the sampling depth on the La+ signal for various carrier gas flow rates A 1.0; and B 0.98 1 min-'+0.02 1 min-' of hydrogen operating conditions used to degrade or enhance the energy transfer two different types of temperatures were measured the excitation temperature via the Boltzmann plot of ionic lines of Ti and the kinetic temperature using the Doppler line broadening of the Be 1 234.86 nm line.It was first verified that the presence of the interface close to the plasma did not significantly modify the temperature. A variation of less than 100 K was observed at the location of the tip of the sampler which meant that the presence of the sampler did not modify the thermal properties of the plasma under the range of conditions used in this work. Over a 2-20 mm observation position range the change in either the excitation temperature or the Doppler temperature was small with a maximum variation of 200 K.It was also found that the kinetic temperature was higher than the excitation temperature by about 2500 K. Spatially resolved temperature measurements would certainly be useful to pro- vide more information on the kinetics of oxide dissociation. However such measurements should be conducted by using Ar Lao Lao+ and La' species. As the oxide dissociation occurs all along the central channel it was preferred therefore to indicate the global changes in the temperature as an indicator of the behaviour of the plasma instead of providing spatially resolved measurements. The maximum values of L a o + and La' over the 2-20mm sampling position range and the corresponding ratio were compared with the global changes in excitation temperature and Doppler temperature using an observation volume over the 2-20mm sampling position range.Results are given in Fig. 10. Although the excitation differs from the kinetic temperature the behaviour of both temperatures as a function of the operating conditions is similar. From Fig. 10 it can be seen that the highestJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1994 VOL. 9 65 I y0.6+0.4 4 0 . 6 + 0.4 I \ O!B + 0.OJ \ 0.98+0.02 \ \ \ I ; 1 1 0.6f0.2 .\i o.8 \ 0.6 + 0.2 2' o.8 1 \ \ \ 0.6 \ 0.6 0 I I I 3000 4000 5000 6000 7000 Excitation and Doppler temperatures Fig. 10 Effect of the excitation temperature (left) and the Doppler temperature (right) on the ratio of Lao' Laf ion signals. The ion signals were obtained by integrating the signal over the sampling position. All values are given in 1 min- '; 0.6 + 0.4 (or 0.6 + 0.2) means that 0.4 1 min-' (or 0.2) was added as a sheathing gas 0.98 +0.02 means that 0.02 1 min-' of hydrogen was added Lao' :La+ ratio is obtained at the lowest temperature. For a total injector gas flow rate of 11 min-' an increase in the temperature and a corresponding decrease in the Lao' La' ratio is observed first when the sheathing gas is suppressed and later when hydrogen is added to the carrier gas.A similar conclusion (Fig. 10) is reached when the sheathing gas flow rate decreases from 0.4 to 0.2 1 min-' or when the carrier gas flow rate decreases from 1 to 0.6 1 min-'. There is therefore a direct relation between the magnitude of the Lao' signal and the temperature of the central channel irrespective of the way the temperature is measured.This is further evidence that the observation of analyte oxides is related to the physical properties of the plasma. Conclusion Analyte oxide ions are mainly formed in the plasma during the dissociation of the sample and owing to the presence of oxygen produced by the dissociation of water. Hobbs and O l e ~ i k ~ ~ have clearly demonstrated that large droplets are not fully evaporated in the ICP. A decrease in the local temperature is observed in the vicinity of the droplets which aids the formation of analyte oxides. Recently Tanner54 has measured the kinetic energy of analyte and analyte oxide ions.A lower temperature of the analyte oxides was deduced from the measurements which indicated that the analyte oxide ions were formed in a cooler part of the plasma. Moreover,54 the oxide ion ratio was independent of the pressure at the interface which is further evidence that oxides are not formed during the extraction process. These results are consistent with the conclusions of Hobbs and O l e ~ i k . ~ ~ More efficient evaporation of the droplets and therefore a lowering of the analyte oxide signal can be obtained using a long residence time and a high energy transfer. In the instance where these conditions are not fullfilled large oxide ion signals are observed. The high carrier gas flow rate of 1.2 1 min-' along with an injector id. of 1.5 mm used originally with the first commercially available ICP-MS systems explains why high oxide ion:ion ratios were observed at this time.Careful optimization of the operating conditions such as the carrier gas flow rate and the sampling position results in an Lao' Laf ratio of the order of 0.2% with only a loss of a factor of 3 of the La' signal. The value of 0.2% is lower than that commonly observed with commercially available ICP-MS systems which is typically of the order of 1%. The value of 0.2% was obtained although the temperature of the plasma was not very high as a result of the selection of a high frequency (56 MHz) which was not beneficial for oxide dissociation. The addition of a molecular gas such as hydrogen increases the rate of analyte oxide dissociation. The only limitation is the need to work with a short sampling position which is not always possible with most commercially available systems.In any instance adjustment of the sampling position was found to be crucial both to minimize the analyte oxide formation although the system is only optimized for elements with a high oxide bond strength and to define compromise conditions for multi-elements analysis. Moreover the re-adjustment of the sampling position when other parameters are modified prevents false conclusions being reached about the behaviour of the various ionic species. Formation of analyte oxides is easy to observe in ICP-MS. However similar poor conditions of energy transfer can also be experienced in ICP atomic emission spectrometry when using low power and a high carrier gas flow rate.Emission of elements having oxides with a high bond strength can be similarly degraded because of the formation of the oxide to the detriment of the ion and consequently to the emission of ionic lines. This is an interference effect that has not been dealt with in emission spectrometry. It will be possible to study this effect with the use of a combined emission-ICP-MS system. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 References Evans E. H. and Giglio J. J. J. Anal. At. Spectrom. 1993 8 1. Gray A. L. and Date A. R. Analyst 1983 108 1033. Date A. R. and Gray A. L. Spectrochim. Acta Part B 1983 38 29. Olivares J. A. and Houk R. S. Anal. Chem. 1985 57 2674. Horlick G. Tan S. H. Vaughan M.A. and Rose C. A. Spectrochim. Acta Part B 1985 40 1555. Vaughan M. A. and Horlick G. Appl. Spectrosc. 1986 40 434. Long S. E. and Brown R. M. Analyst 1986 111 901. Gray A. L. J. Anal. At. Spectrom. 1986 1 247. Gray A. L. Spectrochim. Acta Part B 1986 41 151. McLeod C. W. Date A. R. and Cheung Y. Y. Spectrochim. Acta Part B 1986 41 169. Longerich H. P. Strong D. F. and Kantipuly C . J. Can. J. Spectrosc. 1986 31 11 1. Beauchemin D. McLaren J. W. Mykytiuk A. P. and Berman S . S. Anal. Chem. 1987 59 778. Lichte F. E. Meier A. L. and Crock J. G. Anal. Chem. 1987 59 1150. Boomer D. W. and Powell M. J. Anal. Chem. 1987 59 2810. Zhu G. and Browner R. F. Appl. Spectrosc. 1987 41 349. Gray A. L. and Williams J. G. J. Anal. At. Spectrom. 1987,2 81. McLaren J. W. Beauchemin D.and Berman S. S. J. Anal. At. Spectrom. 1987 2 277. Wilson D. A. Vickers G. H. and Hieftje G. M. J. Anal. At. Spectrom. 1987 2 365. Hutton R. C. and Eaton A. N. J. Anal. At. Spectrom. 1987,2,595. Gray A. L. and Williams J. G. J. Anal. At. Spectrom. 1987,2,599. Price Russ 111 G. and Bazan J. M. Spectrochim. Acta Part B 1987 42 49. Longerich H. P. Fryer B. J. Strong D. F. and Kantipuly C. J. Spectrochim. Acta Part B 1987 42 75. Gordon J. S. van der Plas P. S. C. and De Galan L. Anal. Chem. 1988 60 372. Longerich H. P. J. Anal. At. Spectrom. 1989 4 491. Vaughan M.-A. and Horlick G. J. Anal. At. Spectrom. 1989 4 45. Vaughan M.-A. and Horlick G. Spectrochim. Acta Part B 1990 45 1289. Ross B. S. Yang P. and Hieftje G. Appl. Spectrosc. 1991,45 190. Lepla K. Vaughan M.A. and Horlick G. Spectrochim. Acta Part B 1991 46 967. Norman L. A. Munoz M. Myers D. P. Ross B. S. and Hieftje G. M. Appl. Spectrosc. 1992 46 3. Crain S. J. and Gallimore D. L. Appl. Spectrosc. 1992 46 547. Hall G. E. M. Pelchat J. C. Boomer D. W. and Powell M. J. Anal. At. Spectrom. 1988 3 791. Shibata N. Fudagawa N. and Kubota M. Anal. Chem. 1991 63 636. Zhu G.-X. and Browner R. F. J. Anal. At. Spectrom. 1988,3,781.66 34 35 36 37 38 39 40 41 42 43 44 45 JOURNAL OF ANALYTICAL Tsukahara R. and Kubota M. Spectrochim. Acta Part B 1990 45 581. Lam J. W. and McLaren J. W. J. Anal. At. Spectrom. 1990 5 419. Douglas D. J. and French J. B. J. Anal. At. Spectrom. 1988 Furuta N. Spectrochim. Acta Part B 1986 41 1115. Li. K. P. Dowling M. Fogg T. Yu T. Yeah K.S. Hwang J. D. and Winefordner J. D. Anal. Chem. 1988 60 1590. Vaughan M. A. and Horlick G. Spectrochim. Acta Part B 1990 45 1313. Beauchemin D. and Craig J. M. Spectrochim. Acta Part B 1991 46 603. Wang J . 3 Evans E. H. and Caruso J. A J. Anal. At. Spectrom. 1992 7 929. Craig J. M. and Beauchemin D. J. Anal. At. Spectrom. 1992 7 937. Louie H. and Soo S. Y.-P. J. Anal. At. Spectrom. 1992 7 557. Allain P. Mauras Y. Douge C. Jaunault L. Delaporte T. and Beaugrand C. Analyst 1990 115 813. Allain P. Jaunault L. Mauras Y. Mermet J. M. and Delaporte T. Anal. Chem. 1991 63 1497. 3 743. 46 47 48 49 50 51 52 53 54 A LTOMIC SPECTROMETRY FEBRUARY 1994 VOL. 9 Mermet J. M. in Inductively Coupled Plasma Emission Spectroscopy Part 11 Applications and Fundamentals ed. Boumans P. W. J. M. Wiley New York 1987 ch. 10 p. 353. Boumans P. W. J. M. and Vrakking J. J. A. M. Spectrochim. Acta Part B 1986 41 1235. Ackerman R. J. Rauh E. G. and Thorn R. J. J. Chem. Phys. 1976 3 1027. Murillo M. and Mermet J. M. Spectrochim. Acta Part B 1987 42 1151. Pateyron B. Elchinger M. F. Delluc G. and Fauchais P. Plasma Chem. Plasma Process. 1992 12 421. Ebdon L. and Goodall P. J. Anal. At. Spectrom. 1992 7 1111. Murillo M. and Mermet J. M. Spectrochim. Acta Part B 1989 44 359. Hobbs S. E. and Olesik J. W. Spectrochim. Acta Part B 1993 48 817. Tanner S. D. 76th Canadian Society for Chemistry Conference Sherbrooke 1993 paper 79. Paper 3/04207C Received July 19 1993 Accepted November 4 1993