JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 719 Simplex Optimization of Nitrogen-Argon Plasmas in Inductively Coupled Plasma Mass Spectrometry for the Removal of Chloride-based Interferences Steve J. Hill Michael J. Ford and Les Ebdon Plymouth Analytical Chemistry Research Unit Department of Environmental Sciences Polytechnic South West Drake Circus Plymouth Devon PL4 8AA UK In order to establish the effects of nitrogen addition to the coolant auxiliary and nebulizer gas flows of an Ar plasma polyatomic ions in a range of spiked and unspiked chloride reference materials (Rice Flour Citrus Leaves Dogfish Liver and Sea-water) and standard solution have been used. The response of a wide range of polyatomic ions was studied with particular emphasis on the chlorine-based interferences ArCI+ and CIO+.Simplex optimization was used to optimize the operating parameters of the spectrometer in order to facilitate the maximum possible removal of the ArCI+ when using nitrogen addition. The ArCI+ interference is shown to be successfully removed with the addition of nitrogen to the coolant and nebulizer gas flows in the presence of 1% chloride. Best results were obtained when 4.5% nitrogen was added to the nebulizer gas. These conditions have also greatly improved the determination of selenium and vanadium in the presence of chloride. In addition it has been found that nitrogen addition has some benefit in the reduction of MO+ and ArO+ interferences as well as the background response. The addition of nitrogen to the auxiliary gas offered little advantage for the removal of interference.Keywords Inductively coupled plasma mass spectrometry; mixed gases; nitrogen addition; chloride interfer- ence; simplex optimization An important limitation of inductively coupled plasma mass spectrometry (ICP-MS) is the formation of polya- tomic species particularly below mlz= 80 which interfere in the determination of elements in this mass These polyatomic ions typically come from precursors in the Ar support gas entrained atmospheric gases (N and 0) or from the sample matrix (0 OH C1 S and P) and can be reduced though not removed by careful setting of the instrumental parameters of the ICP mass ~pectrometer,~.~~~ of which the nebulizer gas flow and the forward power are most i m p ~ r t a n t . ~ Other methods of reducing the polya- tomic ions include cooling the spray chambefl to reduce the solvent loading and hence the O+ and OH+ levels hydride generation,' electrothermal vaporization,8 laser ablation9 and mathematical c~rrection.~ Also available though more expensive is high-resolution ICP-MS which uses a mag- netic sector mass spectrometer capable of separating the analytes and polyatomics with the same nominal mass.Io More recently attention has been turned towards the use of molecular and inert gases bled into or replacing one of the three gas flows of the ICP.Such mixed gas plasmas in ICP-MS have been utilized in three main areas attenuation of polyatomic interferences enhancement of analyte signal and the analysis of organic samples. The last two have been discussed elsewhere11J2 but have not been studied here.The investigations include coolant additions of air and oxygeni3 and nebulizer addition of 0~ygen.l~ Allain et af." studied the addition of methane to the nebulizer gas to enhance analyte responses. Smith et al.15 investigated the addition of xenon to the nebulizer gas flow and showed that it attenuated polyatomic interferences. In addition several workers have reported the use of nitrogen for the removal of polyatomic interferences. Houk et al. l6 studied the addition of nitrogen to the coolant gas investigating the formation and movement of background species such as N2+ and Ar+ within the plasma. Lam and Horlick13 reported work on the addition of nitrogen to the coolant and some limited studies on the addition of nitrogen to the nebulizer.They found nitrogen addition to the coolant to be useful in attenuating polyatomic interferences and also enhancing analyte sig- nals. Lam and McLareni7 investigated the use of nitrogen addition at 8% to the coolant gas for reduction of UO+ and ArO+. Evans and Ebdon14 examined the introduction of low flows of nitrogen into the nebulizer flow of an ICP mass spectrometer and were successful in reducing many poly- atomic interferences. More recently Beauchemin and Craig18 have investigated nitrogen addition to the coolant to improve the determination of iron and selenium and also reduce the effects of concomitant matrix elements such as sodium. This paper gives details of investigations into the addi- tion of nitrogen to each of the three gas flows of the ICP.Experiments employing the addition of nitrogen to the nebulizer have involved the analysis of chloride-spiked reference materials one naturally high in chloride and standard solutions to define the limits of applicability of nitrogen addition. The effects of increased nitrogen in the nebulizer gas on response curves and for various interfering ions (MO+ M2+ and ArO+) are reported together with the effect of adding nitrogen to the intermediate and coolant gas flows. Simplex optimization aimed at removing the ArC1+ interference was undertaken for nitrogen addition in all of the gas flows and is compared with a simplex with no nitrogen. Experimental Instrumentation A standard ICP mass spectrometer (PQ2 VG Elemental Winsford Cheshire UK) was used although the sample introduction system was modified by the inclusion of a high solids nebulizer (Ebdon type PSA Sevenoaks Kent UK).Nitrogen was added to the nebulizer gas using either the standard mass flow controller of the instrument or a gas blender (Series 850 Signal Camberley Surrey UK). The gas blender was also used in the addition of nitrogen to the auxiliary gas flow. The addition of nitrogen to the coolant flow was achieved by using a T-piece before the torch the flow being regulated using a 1 dm3 air rotameter. Materials and Chemicals Four certified reference materials (CRMs) were analysed Rice Flour unpolished high level Cd [National Institute of720 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 Environmental Standards (NIES) No.1 Oc Onogawa 16-2 Tsukaba Ibaraki Japan] Citrus Leaves [National Bureau of Standards (NBS) Standard Reference Material (SRM) No. 1 5 12 National Institute of Standards and Technology Gaithersburg MD USA] Dogfish Liver (DOLT- 1 Na- tional Research Council of Canada Division of Chemistry Marine Analytical Chemistry Standards Programme Ot- tawa Canada) and Sea-water (NASS-2 National Research Council of Canada). Standard solutions were prepared from 1000 pg stock solutions of Sb As Co Mg Pb and Se [Merck (formerly BDH) Poole Dorset UK] Be and In (Aldrich Chemical Milwaukee WI USA) and Ce Ho La Tb U and V [prepared from compounds Ce2O3 Ho2O3 La203 Tb203 U02(N03)2-6H20 and NH,V03 respective- ly]. Internal standardization employing either Sb Co or In at lOOng was used in all experiments.Chloride spikes were prepared using either sodium chloride (Aristar BDH) or concentrated hydrochloric acid (Aristar BDH) and for sodium spikes sodium acetate (Aristar BDH). All standard solutions were made up in 2% nitric acid (Aristar BDH). Hydrogen peroxide (30% Aristar BDH) was used in the Dogfish Liver digestion. Sample Preparation The Rice Flour Citrus Leaves and Dogfish Liver were all digested using microwave bomb digestion. The procedure has been reported elsewhere,I9 although in this case 5 cm3 of nitric acid were used for the digestion of the Rice Flour and Citrus Leaves and nitric acid-hydrogen peroxide (3 + 2) was used to digest the Dogfish Liver. After digestion the solutions were quantitatively transferred in to calibrat.ed flasks (25 cm3 for Rice Flour and Citrus Leaves and 100 cm3 for Dogfish Liver) and made up to volume with distilled de-ionized water.Samples were spiked with increasing levels of chloride (0 100 1000 and 10000pgcm-3) as NaCI with indium added to a final concentration of 100 ng as an internal standard. Five replicates of each sample were prepared. Duplicate digestions of the Dogfish Liver were carried out and spiked with increasing levels of sodium (0,65,648 and 6480pg ~ m - ~ ) again using indium as an internal standard. The Sea-water was diluted (two- five- and ten-fold) and spiked with indium as above. Standard solutions and calibration solutions were pre- pared fresh from the stock solutions as required. Procedure Addition of nitrogen to the nebulizer gas flow Initial experiments of adding nitrogen to the nebulizer gas were aimed at assessing the limits of its applicability for the removal of chloride-based polyatomic interferences.The chloride spiked reference materials were all analysed with and without nitrogen addition (e0.035 cm3 min-I). Analy- sis of the 10 OOOpg cm -3 chloride-spiked samples employed higher nitrogen flows (0.045 and 0.06 cm3 min-*) and a ten- fold dilution of the samples. Dogfish Liver digests spiked with sodium were also analysed (0.035 dm3 min-l nitrogen) to assess the effect of sodium without chloride. Further experiments continued on the above theme using standard solutions. Solutions of arsenic selenium ( 100 n g ~ m - ~ ) and vanadium ( 1 10 and 100 n g ~ m - ~ ) were prepared and spiked with increasing levels of chloride (0 10 100 1000 and 10000pgcm-3). Sodium chloride was used for the selenium and vanadium and HCl for the arsenic (the spiking agent was changed because at high NaCl levels cone blockage became a problem).All the solutions were spiked with indium as internal standard to a final concentration of 100 ngcmA3 and the solutions were ana- lysed at varying nitrogen levels (O-8%). Finally the instrumental conditions were optimized for the removal of ArCl+ using a variable step size simplex optimization procedure with the ratio of Sb+ to ArCP (Sb+/ArCl+) as the criterion of merit. Antimony was selected in order to try and match the characteristics of arsenic i.e. a metalloid with a high first ionization energy. Although germanium would perhaps seem a better choice in terms of mass it has many isotopes and could potentially suffer from spectral interference from C12+.Solutions of 100 ng cm-3 of antimony and 10 OOOpg of chloride were used. Univariate searches were also performed at around the optimum parameters defined by the simplex and the optimum conditions were used to determine the detection limits for arsenic selenium and vanadium at increasing levels of chloride (0 100 1000 10000 and 33 000pg~m-~) as HCl. Investigations adding nitrogen to the nebulizer gas were concluded with experiments investigating the complete mass range. The mass response curves MO+ MN+ M2+ and ArO+ formation were examined together with random background with increasing nitrogen. This was performed using a solution of Be Ce Co Ho In La Pb Mg Tb and U ( 100 ng ~ m - ~ ) which was analysed at 0-8% nitrogen using 1% increments whilst all of the other instrumental para- meters were kept constant.Addition of nitrogen to the auxiliary gas flow Univariate searches were performed to assess the effect of nebulizer gas flow (from 0.6-1.2 dm3 min-I) and auxiliary gas flow (from 0.6 to 2.0dm3min-I) with 2% nitrogen in the auxiliary gas at 1600 W. The species examined were Be Ce Co La Pb and U; CeO La0 and UO; ClO+/NCl+ (m/z=51) ArN+ ArO+ ArCl+ and Ar2+. Searches were also performed to assess the effect of nebulizer gas at 1200 1400 and 1800 W on Be La and U; UO+; ClO+/NCl+ A m + and ArCl+. Analogous searches without any nitrogen addition were performed for comparison. Simplex optimization of the instrumental parameters was undertaken as above for addition to the nebulizer gas.The associated univariate searches were performed and detec- tion limits were determined. Addition of nitrogen to the coolant gas flow Univariate searches were performed to assess the effect of nebulizer (0.6-1.2 dm3 min-I) auxiliary (0.6-2.0 dm3 min-I) and coolant gas flow (13-15 dm3min-l) with 200 cm3 min-l nitrogen flow in the coolant gas at 1600 W on Be Ce Co and U; CeO; random background at m/z= 150 and 220; ClO+/NCl+ (m/z=5 I) ArO+ ArCl+ and Ar2+. Searches were also performed to assess the effect of nebulizer gas flow at 1400 and 1800 W on Be Co and U; CeO; and ClO+/NCl+ ArO+ and ArCl+. Again searches were repeated without any nitrogen the conditions optim- ized by simplex and the associated experiments performed.Finally for comparison one further simplex optimiza- tion was performed without any nitrogen in order to compare the best argon conditions with the best nitrogen- argon conditions for the removal of ArCl+. Results and Discussion Nitrogen Addition to the Nebulizer Gas Flow Analysis of chloride-spiked CRMs The apparent arsenic concentration with and without nitrogen addition in the three chloride-spiked CRMs is shown in Figs. 1-3. It can be seen that in all of the reference materials the ArCl+ interference is removed at up toJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 72 1 cn i E 8 2 20 60 5 0 - 40 5 30 E .- U Q1 0 I- C 1000 pg cm-3 of chloride but becomes significant at higher levels (1000O,~gcm-~) where a positive error is found in the apparent arsenic value.The data with no nitrogen addition for the three reference materials show that even without a chloride spike there is a positive bias in the apparent arsenic value thus demonstrating the value of nitrogen addition. The apparent arsenic values at the 10 000,~gcm'~ chloride spike are about 6 times too high for the Dogfish Liver,=20 times too high for the Citrus Leaves and ~ 4 0 times too high for the Rice Flour. Further investigation of the 10 0 0 0 p g ~ m - ~ chloride-spiked samples showed that increasing the percentage of nitrogen in the nebulizer gas to 5.03% or higher removed the ArCl+ interference on the Citrus Leaves but not on the other two reference materials. This might be because the digest level of arsenic in the Citrus Leaves was higher than the other two (=65 n g ~ m - ~ compared with ~ 2 2 n g ~ m - ~ for the Dogfish Liver and =3 n g ~ m - ~ for the Rice Flour) and that the residual ArC1+ was not significant.Analysis of 10-fold dilutions of the Dogfish Liver and Citrus Leaves digests resulted in values coincident with the certificate values. This indicated that it was an ArCl+ interference or suppression of the indium internal standard that was causing the positive bias in the apparent arsenic values rather than contamination from the NaCl. Analysis of Dogfish Liver digests spiked with sodium indicated that the sodium was having no effect on the determination of arsenic in the CRMs. It was also noted that when nitrogen was added to the nebulizer gas flow - - - - c 1 0 10 100 1000 10000 Chloride concentration/pg ~ r n - ~ Fig.1 Apparent arsenic concentration in chloride-spiked NIES CRM 1 Oc Rice Flour A with; and B without nitrogen addition to the nebulizer gas c 0 10 100 1000 10000 Chloride concentration/pg ~ r n - ~ Fig. 2 Apparent arsenic concentration in chloride-spiked NIST CRM 1572 Citrus Leaves A with; and B without nitrogen addition to the nebulizer gas there was a much less drastic decrease in the indium response than that found without the nitrogen for samples high in sodium. These observations are in agreement with the work of Beauchemin and Craig1* who added nitrogen to the coolant gas flow and found it beneficial in reducing the matrix effects of sodium. Analysis of standard arsenic selenium and vanadium solutions The apparent arsenic concentration for 100 n g ~ m - ~ arsenic solutions at increasing nitrogen and chloride levels is shown in Fig.4. As might be expected at higher levels of chloride the ArCl+ interference became worse and hence the apparent arsenic concentration increased but at higher nitrogen levels this interference is less severe. This would support the notion that nitrogen suppresses the interference through a competitive mechanism with the argon. The analysis of selenium solutions ( 100 ng cme3) was not totally successful; the values obtained can only be taken as indicative of the true values. The data whilst being poor still indicated a considerable improvement over the analy- sis without nitrogen where at 10000pg~m-~ chloride the apparent selenium value was =6000 n g ~ m - ~ .The poor quality of the data was attributed to the fact that selenium is poorly ionized and that both "Se and 78Se are low abundance isotopes typically yielding just 1-200 area counts per second (ACPS) for a 100 ngcmq3 selenium solution. Chloride concentration/pg ~ r n - ~ Fig. 3 Apparent arsenic concentration in chloride-spiked NRCC CRM DOLT-1 Dogfish Liver A with; and B without nitrogen addition to the nebulizer gas *) 10000 I cn \ 0 .- 5 1000 8 s U 0 .- = 100 3 L nr B 1 I I 1 0 10 100 1000 10000 Chloride concentration/pg ~ r n - ~ Fig. 4 Apparent arsenic concentration with increasing chloride levels (100 ng cm-j As) for increasing percentage nitrogen in the nebulizer gas A 0; B 0.5; C 1.0; D 3.0; E 5.0; and F 7.0%722 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL.7 I $ 0 10 100 1000 10000 Chloride concentration/pg ~ r n - ~ Fig. 5 Apparent vanadium concentration with increasing chloride levels (100 n g ~ m - ~ ) for increasing percentage of nitrogen in the nebulizer gas A 0; and B 3.4 1-8.1 I O/o For vanadium the 100 n g ~ m - ~ solutions gave values within 10% of that expected for all nitrogen ratios and levels of chloride (Fig. 5). These results are a little surprising as it was anticipated that the interference from l4N3’Cl + would become significant at higher nitrogen and chloride levels as had been seen in the CRMs and reported previou~ly.~~ Again the all argon plasma analysis showed the importance of nitrogen in removing poly- atomic ion interference.The data obtained by a similar analysis of 10 n g ~ m - ~ vanadium solutions are shown in Fig. 6 and demonstrate the NCl+ interference which adds to the residual C10+ at higher nitrogen and chloride levels. The 1 n g ~ m - ~ solutions showed similar trends to the 10 ng cm-3 solution. Simplex optimization of operating conditions for the re- moval of Arc/+ with nitrogen addition to the nebulizer gas The optimum parameters obtained from the simplex optimization to remove ArCl+ when using nitrogen addition to the nebulizer gas are shown in column one of Table 1. At these optimum conditions the plasma was seen to be dimmer than normal probably owing to the lower power. The central channel also appeared much wider than normal and consequently the coolant region of the plasma was smaller.The reflected power was about 40W. The large central channel is in practice very useful since the position of the torch is no longer critical. The diffuse central channel probably accounted for the loss of sensitivity for antimony (from about 80000 to about 10000 ACPS) however the effect was far less severe than the reduction in the ArCl+ response which fell to almost background levels (from about 50000 to about 100 ACPS). Univariate searches of the optimum parameters showed the power nebulizer and nitrogen values to lie at the conditions defined by the simplex whilst the coolant and auxiliary settings were seen to have little effect on the Sb+/ArCl+ ratio. This is in agreement with Gray and WilliamsS who when discussing optimization of all argon plasmas found power and nebulizer flow rate to be the most important parameters (if spray chamber temperature and sampling depth are con- stant) when trying to remove polyatomic ion interferences.They found that low power (1 100 W) and higher nebulizer flows (0.85 dm3 min-l for ArO+ and ClO+ and in excess of 1 dm3min-l for Ar2+) generally gave the best ratio of interference to cobalt and this compares well with the conditions shown in column one of Table 1. These conditions however also include the addition of nitrogen flow - I I I I 10 100 1000 10000 ii 1 ; 8 Chloride concentration/pg ~ r n - ~ Fig. 6 Apparent vanadium concentration with increasing chloride levels ( 10 ng ~ m - ~ vanadium) for increasing percentage of nitrogen in the nebulizer gas A 0; B 7; C 1 ; D 2; E and F 3 and 5; and G 4% to the nebulizer gas which further improves the ratio of Sb+/ArCl+ (from 10-20 without nitrogen to about 300 with nitrogen).Column one of Table 2 shows the detection limits determined at these conditions for arsenic selenium and vanadium at various chloride levels. The most important point of note is that the detection limit for arsenic (2.1 n g ~ m - ~ ) in the presence of 33 OOOpg cm-3 of chloride (1 0% HCl) was only a factor of four poorer than that normally obtained on our instrument even in the absence of chloride. The contribution of ArCl+ was equivalent to a background concentration of 5.7 n g ~ m - ~ . These conditions will allow the determination of arsenic in matrices that are extremely high in chloride.The analysis of the NASS-2 Sea-water was undertaken using both the unoptimized and the optimized conditions and both sets of data are presented in Table 3. Under unoptimized conditions values ranging from 5 to 30 times higher than the certificate value were found although these are considerably better than those obtained without nitro- gen which are 3000-4000 times too high. The data obtained with the simplex conditions are even better with a positive error of only 3-6 fold. The use of flow injection techniques could overcome the need to dilute the samples and hence facilitate accurate arsenic determination. Effects of increasing the percentage of nitrogen in the nebulizer gas flow It was found that the incremental increase of nitrogen to the nebulizer gas flow had no visible effect on the shape of the response curves.Metal oxide formation was found to be improved by the addition of nitrogen up to 2% but no further benefit was found above this level. The degree of oxide reduction ranged from a factor of 2.1 for uranium to 12.5 for holmium. These data indicate that even without any form of solvent reduction nitrogen addition could reduce the metal oxide levels. Lam and Horlick13 found that 20% nitrogen in the nebulizer removed almost all of the La0 signal which compares well with these findings. No metal nitrides were found even at the highest levels of nitrogen addition and the doubly charged species such as Ce2+ Ho2+ and U2+ were not detected. Random background and ArO+ were found to be reduced by the addition of nitrogen at 1-2% but once again further addition of nitrogen had no effect.Addition of Nitrogen to the Auxiliary Gas Flow Effects of nitrogen addition to the auxiliary gas flow It was found that the instrument could tolerate a maximum of just 3% v/v nitrogen added to the auxiliary gas since723 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 Table 1 Simplex optimized conditions for the removal of ArCl+ with nitrogen addition Parameter Gas flow/dm3 min-' Coolant Auxiliary Nebulizer Nitrogen addition Forward power/W *Value given as %. tValue given in cm3 min-I. Added to the Added to the Added to the nebulizer gas auxiliary gas coolant gas 13 13 15 1 .o 0.5 0.5 0.9 1.2 1.2 4.50* 2.50* 300t 1300 1350 1300 Table 2 Detection limits determined for arsenic selenium and vanadium at increasing chloride concentrations using simplex optimized conditions for nitrogen addition Detection limitshg ~ m - ~ Addition to the nebulizer gas Chloride level/ p g ~ m - ~ V As 17Se ?%e 0 0.21 0.78 4.9 6.2 100 0.12 0.48 6.5 6.5 1000 0.48 0.69 3.9 42* 1 0 000 1.5 1.1 23 31t 33 000 13 2.1 25 200t *Either two-point or poor three-point calibration.tone-point calibration. Addition to the auxiliary gas V As 17Se ?*Se 0.06 1.5 23 8.5 0.06 1.5 ll*. 40t 0.51 1.5 8* 80t 2.9 3. I 15 19 16 4.9* 33t - Addition to the coolant gas V As 17Se 18Se 0.09 1.1 4* 4* 0.12 0.24 20t 10* 0.24 1.1 11 6.7 3.9 0.96 13 8.4 - 26 - 201- Table 3 Apparent arsenic concentration (average 5 error of duplicates) in NASS-2 CRM Sea-water at different dilutions and nitrogen ratios for unoptimized conditions and different dilutions for simplex optimized conditions Arsenic concentration*/pg ~ r n - ~ Dilution Unoptimized conditions- 0% Nitrogen? 3.4 1% Nitrogen 5.03% Nitrogen 6.59% Nitrogen l o x 7.42 -I- 0.43 0.008 rf 0.004 0.01 1 f 0.002 0.01 2 +- 0.00 1 5x 7.95 k 0.32 0.0 1 1 5 0.00 1 0.01 1 LO.001 0.01 4 5 0.001 2 x 5.56 s+_ 0.3 1 0.053 f 0.00 1 0.022 k 0.001 0.026 +- 0.001 Optimized condilions- Replicate 1 Replicate 2 10 x <0.005$ < 0.00 5$ 5 x 0.006 -I- 0.002 0.010~0.002 2 x 0.0 12 5 0.00 1 0.0 13 -t 0.00 1 *Expected value 0.001 65 pgcm-j.to% Nitrogen values are extrapolated. $Based on 30 of the blank. even at this low level the reflected power was prohibitively high (50 W at an auxiliary flow of 1 dm3min-l).The nitrogen was found to cause the plasma to shrink away from both the sampler cone and the torch. This thermal pinch effect was possibly due to the proximity of the nitrogen to the induction region. Effect of nebulizer and auxiliary gas flow adjustment on argon and nitrogen-argon plasmas A 2% addition of nitrogen to the auxiliary gas had very little effect on analyte response metal oxide formation and polyatomic ion interferences (ArCl+ ClO+ ArO+ and AT2+) compared with an all argon plasma. The optimum nebulizer gas flow was unaltered by the addition of nitrogen to the auxiliary gas unlike the optimum nebulizer gas flow when nitrogen is added to the coolant gas which in- c ~ e a s e s . ~ ~ J ~ Considering the physical similarity of the two types of plasma it might have been expected that the optimum nebulizer gas flow would be similar.A 50% loss in sensitivity was noted when the nitrogen was added to the auxiliary gas flow. Variation in the nebulizer gas flow was compared at four power levels (1200 1400 1600 and 1800 W) for the two plasma types; the data being similar in each case. In agreement with Gray and Williamss the polyatomic ions were least intense at the lower power levels and nitrogen addition offered no improvement. Simplex optimization of the operating parameters for the removal of ArCl+ with nitrogen addition to the auxiliary gas This study was carried out as described above and the optimum conditions are shown in column 2 of Table 1. The plasma was seen to be smaller than usual and most importantly had a wide central channel similar to that observed when the nitrogen was added to the nebulizer gas.The simplex optimized conditions for auxiliary gas nitrogen addition whilst showing effective removal of the ArCl+ (absolute response of about 100 ACPS) were in terms of the Sb+/ArCl+ ratio at least five times poorer than those found previously. As with the previous simplex a series of univariate searches were carried out on these conditions in each case flow724 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 seem that the other parameters particularly the power and the nebulizer gas flow result in most of the ArCl+ loss and that the nitrogen is merely removing a few additional ArCl+ ions. This is discussed below but the data indicated that the addition of nitrogen to the auxiliary gas is of little practical value. Once again the optimum conditions were tested for their analytical performance by determining the detection limits for arsenic selenium and vanadium with increasing levels of chloride.The results obtained are shown in column 2 of Table 2 and are similar to those found at the optimum conditions for the addition of nitrogen to the nebulizer gas. Sampling Almost invisible Torch Intense region cone region I \ I I I I I f r Addition of Nitrogen to the Coolant Gas Effects of nitrogen addition to the coolant gas flow The introduction of even the smallest amount of nitrogen into the coolant gas again caused the plasma to shrink away from both the torch and the sampling cone as discussed above.The addition of nitrogen resulted in a large increase in the reflected power 50 W at ( 5 0 cm3 min-l of nitrogen. The maximum nitrogen addition that the instrument could tolerate in this intance was 300~m~min-l of nitrogen in 15 dm3 min-' of argon (i.e. approximately 2%). It may be noted that the level of nitrogen able to be employed in this work was far less than reported by other workers. Lam and Horlick13 for example added nitrogen at 5-20% Lam and McLaren17 at 8% and Beauchemin and Craig18 at 0-10%. None of these workers reported a problem with the reflected power although all used instruments manufactured by a different company which might have different tuning characteristics. Eflect of nebulizer auxiliary and coolant gas flow adjust- ment on all argon and nitrogen-drgon plasmas When the analyte (beryllium cobalt cerium and uranium) responses are compared for an all argon and a nitrogen- argon plasma two points are of immediate interest.Firstly the maximum response increases from 0.9 dm3 min-' to the maximum flow of 1.2 dm3 min-I and secondly the maxi- mum response is about 50% higher in the nitrogen plasma. These two points are consistent with the findings of other workers.13-17 The increase in the nebulizer gas flow probably corrects for the shrinkage of the plasma which results in moving the sampling zone away from the cones. Oxide formation was found to be similar in both of the plasmas with the nitrogen-argon plasma giving better ratios (MO+/M+) only at the highest nebulizer gas flows. The interference (ArCl+ NCl+/C10+ ArO+ and AT2+) re- sponses were similar to those obtained for the analytes.The plots for the auxiliary and coolant searches for all of the analytes CeO+ and the interference ions are essentially flat with no noticeable difference between the two types of plasma. When the effect of the nebulizer gas flow at three powers was assessed for the two plasmas the results were again found to be very similar. Simplex optimization of the operating parameters for the removal of ArCP with nitrogen addition to the coolant gas flow This procedure was carried out as described above. The optimized conditions which are very similar to those found already are given in column 3 of Table 1. These conditions change the plasma considerably as represented in the schematic diagram in Fig.7. The effect of thermal pinch and also the separation of the plasma into three distinct regions may be noted. Once again the plasma had a dim and wide central channel as found in all of the simy;!ex optimizations. Outer reg& smaller Central chanAel wider than than normal normal and also fainter Fig. 7 Sch,ematic representation of the plasma with nitrogen addition to the coolant gas under the simplex optimized conditions As before univariate searches were undertaken to con- firm the optimum. The optimum nebulizer gas flow power and nitrogen addition were all shown to have correct optima and the coolant and auxiliary were shown not to be influencing the Sb+/ArCl+ ratio. The removal of ArCl+ was effectively complete under these conditions the absolute counts recorded being of the order of 50 ACPS.The detection limits (column 3 Table 2) were found to be similar to those given above. Simplex optimization of the operating parameters for the removal of A r W with an all argon plasma This experiment was undertaken to see how well the instrument could remove the ArCl+ interference without the presence of nitrogen. The optimum conditions are given in Table 4. In this case the plasma once optimized looked similar to that using the nitrogen addition conditions although the central channel was not as wide and the absence of nitrogen meant that the reflected power was less than 10 W. Table 4 Simplex optimized conditions for the removal of ArCI+ with an all argon plasma Parameter Value Gas flow/dm3 min-1 Coolant I8 Auxiliary 0.8 Nebulizer 1.2 Forward power/W 1300 Table 5 Detection limits determined for arsenic selenium and vanadium at increasing chloride concentrations using simplex optimized conditions for minimum interference for an all argon plasma Detection limit/~gcm-~ V As 17Se 78Se 100 0.78 - 21* - Chloride level/ ~ g c m - ~ 0 0.96 3.45 15* 14* 1000 1.14 10.5 30* 1 oot 20$ 303 90$ 60$ 1 ot 3.5* 221 26* 10 000 33 000 *Either two-point or poor three-point calibration.tone-poi n t Cali brat ion. $Doubtful data due to elevated blank level.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1992 VOL. 7 725 Table 6 Antimony to ArCl+ ratios (Sb+/ArCl+) for the four simplex optimizations Maximum ratio Ratios at optimum Simplex in simplex during searches* Nitrogen addition to Nitrogen addition to Nitrogen addition to the nebulizer gas 492 300 (200-350) the auxiliary gas 50 40 (30-35) the coolant gas 105 200 (100-250)’ All argon 46 20 (15-25) *Approximate values from the search graphs.Univariate searches were performed and showed that the auxiliary gas did not affect the conditions. The optimum nebulizer gas coolant gas and power defined by the simplex were confirmed. The actual performance under these conditions was found to be poorer than in the nitrogen work with Sb+/ArCl+ ratios in the region of 20-40. Once again the detection limits were determined and are given in Table 5. The determination of detection limits did however indicate one limitation of the optimized conditions in that they produced a very large 76Ar2+ peak that interfered with the 7sA~+.At the higher chloride levels the calibration graphs were very poor as a result of this. The four simplex optimizations can be compared for their analytical capabilities by either considering the detec- tion limits or the ratios for Sb+ to ArCl+. The use of detection limits does not clearly show any one set to be better than any other. In Table 6 data on the various ratios obtained with each set of conditions are presented. It is apparent that the all argon conditions give the poorest ratios although interestingly the auxiliary gas conditions do not show much improvement. The nebulizer gas and coolant gas optimizatons however are considerably better. It can therefore be concluded that in the case of nitrogen addition to the auxiliary gas the removal of the vast majority of the ArC1+ is as a result of other parameters particularly the power and the nebulizer flow and that the nitrogen has only a marginal effect.Conclusions The initial work showed that the addition of nitrogen to a plasma under normal operating conditions could remove the polyatomic interference on arsenic and vanadium at up to 1000,ug cmw3 of chloride. Subsequent optimization ex- periments enabled the level of chloride to be increased to at least 3.3% without producing any significant interference on arsenic. Work is now continuing using other molecular gases such as the addition to methane to the ICP mass spectrometer with the aim of reducing a wider range of polyatomic interferences. The authors gratefully acknowledge the financial support of M.J.F. by the Science and Engineering Research Council and VG Elemental which has made this work possible. 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 References Tan S. H. and Horlick G. Appl. Spectrosc. 1986 40 445. Vaughan M. A. and Horlick G. Appl. Spectrosc. 1986 40 434. Munro S. Ebdon L. and McWeeny D. J. J. Anal. At. Spectrom. 1986 1 21 1. Gray A. L. Spectrochim. Acta Part B 1986 41 15 1. Gray A. L. and Williams J. G. J. Anal. At. Spectrom. 1987 2 599. Hutton R. C. and Eaton A. N. J. Anal. At. Spectrom. 1987 2 595. Branch S. Corns W. T. Ebdon L. Hill S. J. and O’Neill P. J. Anal. At. Spectrom. 1991 6 155. Whittaker P. G. Lind T. Williams J. G. and Gray A. L. Analyst 1989 114 675. Gray A. L. Analyst 1985 110 551. Bradshaw N. Hall E. F. H. and Sanderson N. E. J. Anal. At. Spectrom. 1989 4 801. Allain P. Jaunault L. Mauras Y. Mermet J. M. and Delaporte T. Anal. Chern. 1991 63 1497. Hutton R. C. J. Anal. At. Spectrom. 1986 1 259. Lam J. W. H. and Horlick G. Spectrochim. Acta Part B 1990,45 1313. Evans E. H. and Ebdon L. J. Anal. At. Spectrom. 1990 5 425. Smith F. G. Wiederin D. R. and Houk R. S. Anal. Chem. 1991,63 1458. Houk R. S. Montaser A. and Fassel V. A. Appl. Spectrosc. 1983 37 425. Lam J. W. H. and McLaren J. W. J. Anal. At. Spectrom. 1990 5 419. Beauchemin D. and Craig J. M. Spectrochim. Acta Part B 1991 46 603. Branch S. Ebdon L. Ford M. J. Foulkes M. E. and O’Neill P. J. Anal. At. Spectrom. 1991 6 151. Paper 2/00349J Received January 22 1992 Accepted April 6 I992