JOURNAL O F ANALYTICAL ATOMIC SPECTROMETRY JUNE 1994 VOL. 9 719 Signal Enhancement and Reduction of Interferences in Inductively Coupled Plasma Mass Spectrometry With an Argon-Trifluoromethane Mixed Aerosol Carrier Gas Isaac Platzner* Jose V. Sala Francis Mousty and Pier R. Trincherini Environment Institute Joint Research Centre 1-2 1020 lspra (VA) Italy Alberto L. Polettinit SEA Marconi Technologies Collegno (TO) Italy c/o Joint Research Centre T.P. 460 1-27020 lspra (VA) Italy The addition of trifluoromethane (CHFJ to the aerosol carrier gas in inductively coupled plasma mass spectrometry was assessed as a method of improving detection limits (DL) for elements such as As Se Cu and Zn in matrices containing interfering species. It was observed that in each case the analyte response was significantly increased with a coincident decrease in the blank signal.The improved DL for As in CI- and Ca2+ matrices (interferences 40Ar%I+ and 43Ca'60z were 0.02 and 0.04 ng ml-' compared with 0.65 and 0.28 ng ml-' respectively without CHF,; for 78Se (interference 40Af'8Ar+) 0.032 compared with 0.88 ng ml-'; for 63Cu in Na2S0 or Na,HPO matrices (interferences ,'AP3Na+ and 31P1602+) 0.022 and 0.089 compared with 0.35 and 0.53 ng ml-' respectively; and for 64Zn in an Na,HPO matrix (interferences H3'P1602+ and 31P160170+) 0.011 compared with 0.42 ng ml-' The reduction of the interference is attributed to competitive reactions between the matrix species and the CHF or species derived from it in the plasma. The analyte enhancement effect is not yet clear. It has been suggested that this effect is related to elements with ionization potential (IP) in the 9-11 eV region and is affected by organic compounds added to the aerosol carrier gas stream.Copper (IP 7.73 ev) Al (5.99 eV) Br (1 1.30 eV) and I (1 0.44 ev) are exceptions to this assumption. Analytical curves of the studied elements at low ppb and sub-ppb levels (in the interfering matrices) further demonstrated the advantage of adding CHF in trace elemental analysis. Keywords Inductively coupled plasma mass spectrometry; trifluoromethane addition; interference reduction; analyte signal enhancement; detection limits Intensive studies have been carried out in the past few years and are still under way at various laboratories on ways of improving the analytical performance of inductively coupled plasma mass spectrometry (ICP-MS).Particular attention has been given to improving the signal-to-blank ratio for different analytical applications in the range of up to 80 m/z where strong interferences are observed which originate from isobaric masses of the aerosol carrier (nebulizer) gas matrix compounds and their mutual interaction products. The experimental approaches to overcoming or partially reducing the interfering polyatomic species have recently been extensively reviewed by Evans and Giglio.' One of the simplest and most widely used methods is to add minor amounts of different gases such as hydrogen nitrogen oxygen air helium or xenon to one of the three argon gas flows mainly mixing with the aerosol carrier stream.Recently the addition of methane to the aerosol carrier flow was studied by Hill et al. who observed that interfering ions such as ArCl' ArO' ClO' and CeO' were reduced relative to the unmodified plasma. Allain et aL3 has shown that the addition of methane moderately enhanced the analyte response for As Se and Te. These workers also reported matrix enhancement effects on As Se Te Hg and Au in glycerol and glucose solutions. Signal enhancement through the addition of nitrogen to the argon outer flow was also observed by Lam and H ~ r l i c k . ~ In the present paper results are reported for the mutual effect of reducing the interfering blank and the signal enhance- ment achieved for several elements by mixing trifluoromethane with the argon aerosol carrier flow in a conventional ICP-MS instrument.Experiments with this aerosol medium were also performed by adding glycerol to several solutions and their blanks. To our knowledge this is the first report where reduction in interference accompanied by simultaneous *Visiting scientist from NRCN PO Box 9001 Beer-Sheva Israel. t To whom correspondence should be addressed. enhancement of the analyte signal has been observed upon mixing gases in the aerosol stream. Experimental Instrurnenta tion The experiments were performed on an inductively coupled plasma mass spectrometer Model PQ2 (VG Elemental Winsford Cheshire UK) equipped with standard Meinhard nebulizer. The addition of trifluoromethane (CHF,) to the argon aerosol carrier gas stream was controlled with a 0-5 ml min-' gas mass-flow controller Tylan General Model FC-260 (Swindon Wiltshire UK).It was introduced directly to the nebulized analyte through a modified glass port prior to the plasma torch. Materials All solutions of the elements were prepared from 1 mgml-' stock solutions (Aldrich Chemicals Milwaukee WI USA) diluted with 1 % HNO (Suprapur Merck Darmstadt Germany) in de-ionized water (MilliQ Millipore Bedford MA USA). The other chemicals were HC1 (Suprapur Merck); redistilled glycerol (pro analysi Merck); Na,S04 (Suprapur Merck); Na,HPO Ca(NO,) and NaCl (Carlo Erba Milano Italy) and CHF (Halocarbon or Freon 23 Sol Monza Italy). Procedures The concentration of all the elements was 100 ng ml-' in 1% nitric acid. The element solutions and their solutions in different matrices were studied without and with the addition of CHF3 over a range of 1-5 ml min-l.The matrix solution at variable flow rates of the CHF was measured as the blank in each experiment. Data collection started 3 min after gas mixing. All experiments were performed at a fixed distance between the plasma torch and the sampling interface.720 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1994 VOL. 9 The following isotopic species were studied in 1% HNO (unless otherwise stated) 'Be; 27Al; 63Cu and 65Cu (5% HNO,); 64Zn 66Zn 67Zn 68Zn and 70Zn (5% HNO,); 69Ga and "Ga; 72Ge and 74Ge; 75As; 78Se and 82Se; 79Br and 8'Br; '"In; 'I6Sn 'I7Sn 'I8Sn and "'Sn; "'Sb and lz3Sb; 124Te 126Te '"Te and More complex matrices were used for the following elements for Cu 5% HC1 0.02mol I-' Na2HP04 and 0.02 mol I-' Na2S04; for Zn 5% HCl 0.02moll-' Na2HP04 and 0.02 mol I-' Na2S04; for As 0.02 and 0.04 mol I-' NaCl 0.25 0.5 and 1.0 moll-' glycerol 0.01 moll-' Ca(NO,) and 0.02 moll-' Na2HP04; for Se 0.25 0.5 and 1.0 mol I-' glycerol; and for Te 0.25 0.5 and 1.0 moll-' glycerol.Blank measurements were also carried out at m/z values of 48 51 52 53 54 55 59 and 77. The linearity of the response in the interfering matrices for element concentration ranges from 5 to 50 (As) 1 to 10 (Zn) 0.5 to 5 (Se) and 0.25 to 1 ng ml-' (As and Cu) were tested at the optimal signal-to- blank ratio with addition of CHF,. Each result given is an average of at least three measure- ments. The operating conditions of ICP-MS instrument are summarized in Table 1.l 3 q e ; 1271; 202Hg; 208pb; and 2 3 8 ~ . Results and Discussion Arsenic Arsenic is a mono-isotopic element of mass 75. In most of the natural samples where chlorine is available (as C1- or organic chloride) a strong interference from 40Ar35Cl+ is always observed. Various procedures have been adopted to eliminate this interference hydride generation in the analysis of water;s modified dissolution procedures of marine sediment$ precipi- tation of chloride with Ag for protein sample^;^ matrix separ- ation by gel filtration for serum;8 liquid chromotographic separation (anion exchange) of As species in urine;g910 addition of nitrogen to argon for correction using elemental and mathematical calculations in l~bster,'~ oyster tissue bovine liver and kale.14 Interferences of 59C0160 + in nickel al10ys'~ and of iron oxides and hydroxides in steel on 75As have also been observed.16 The effect of addition of CHF to a 100 ng ml-' solution of As without and with 0.02 and 0.04 mol I-' NaCl and to the blank is shown in Table 2 and Fig.1. It can be seen that CHF enhances the "As+ signal in all the three experiments by a Table 1 Experimental and operating conditions ICP Nebulizer R.F. power/kW Outer gas flow rate/l min-' Intermediate gas flow rate/l min-' Aerosol carrier gas flow rate/l min-' CHF gas flow rate/ml min-' Solution uptake rate/ml min-' Spray chamber temperaturePC Mass spectrometer Instrument Sampler orifice (nickel)/mm Skimmer orifice (nickel)/mm Interface pressure/mbar* Analyser pressure/mbar Data acquisition parameters Mode Channels per m/z Interval between channels Dwell time/ms Acquisition time/s Measurements per sample Meinhard 13 1.35 0.5 0.70-0.90 0-5.0 2 0-10 VG/PQ2 1 0.5 1 1.8 x Peak jumping 3 5 10.24 3&60 3 * 1 mbar= 1 x 10' Pa. factor of 4-5 with a maximum at a flow of 2-3 ml min-'.The opposite effect is observed for the blank solutions the high blank counts in the presence of 0.02 and 0.04moll-' NaCl are reduced to a minimum at about the same CHF flow rate. It is not immediately obvious why the signal for As' increases but it is easier to understand the effect on the blank. Without CHF the blank counts are due to 40Ar35Clf ions. In the presence of CHF an efficient competitive reaction could take place increasing ArF+ relative to ArCl'. The net effect is that the relative enhancement for a 0.04 moll-' concentration NaCl is >9 times higher (37.4 to 4) than without NaCl (see Table 2).The 'relative enhancement' (RE) is defined as the ratio of the sample counts to the blank counts when CHF is added (ISIb)cF divided by the ratio of the sample counts to the blank counts when CHFJ is absent (IsI,JoY [RE= (IsIb)CF/(lsIb)O]. The blank at m/z 59 (ArF') and at m/z 77 (40Ar37C1+) was measured with the addition of CHF3. A continuous increase of ArF+ from 150 counts s-' (at 0 ml min-') to 3470 (at 5 ml min-') was observed supporting the decrease in the amount of ArCl' owing to a competitive mechanism. The blank at m/z 77 followed exactly the behaviour of the blank at m/z 75 except that the signal intensities were only one third of those at m/z 75 proving that the blank is ArC1' (the isotopic ratio 35C1:37C1 is approximately 3:l).The matrix effect of NaCl which reduces the As' signal at 0.04 moll-' to 80% (at a CHF flow rate of 2 ml min-') is of minor importance. The same combined effect of As+ signal enhancement and parallel reduction in the blank signal was also observed for the 43ca1602 + interference originating from calcium nitrate. It is possible that atomic fluorine in the plasma reacts with Ca-containing species to yield different products. In this case the combined effect is a steady relative enhancement of about 10 for CHF flow rates of 2mlmin-' and above as shown in Table 2. Selenium and Tellurium The most abundant Se isotope 80Se (49.7%) is completely hidden by the strong 40Ar2+ interference the presence of chloride (40Ar37Cl+) interferes with 77Se (7.6%) 40Ar38Ar+ interferes with 78Se (23.6%) while considering 82Se (9.2%) there was no evidence of interference from krypton which could be an impurity in the argon used. The interferences in biological materials were eliminated by generation of selenium hydride and determination of the element by isotope dilution analy~is,'~ in coal by slurry nebulization'8 and in urine by addition of nitrogen to any of the three gas streams." Addition of CHF reduces the blank at m/z 78 has no effects (as expected) on the blank at m/z 82 and causes a 3-4-fold increase in the 78Se+ and "Se+ signal; consequently a relative enhance- ment of up to 10.8 is observed for 78Se and 2.7 for 82Se.The reduction in the blank at m/z 78 is consistent with the assumption of CHF reacting with argon.Thus it can be concluded that the addition of CHF solves the interference problems in quantitative determination of Se when using the second most abundant isotope (78Se). The method is compar- able to hydride generation in sensitivity and preferable in simplicity of experiment. The data are summarized in Table 3 and the enhancement for 78Se is shown in Fig. 2. It should be noted that when nebulizing a 1% solution of HNO in the presence of 2 ml min-' of CHF an increase in the blank is observed at m/z 77 relative to a zero flow rate of CHF,. The effect is more evident when increasing the concentration of HN03 to 5%. This indicates the formation of an inter- ference owing to the simultaneous presence of CHF and nitrogen in the plasma an interference that is attributed to Tellurium does not suffer from interferences except for its minor isotopes from "'Sn and 124Sn. The presence of xenon as an impurity in the argon used would interfere with the 12CF14N160 + 2 -JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1994 VOL.9 Table 2 Addition of CHF to As plus matrix and matrix only solutions 72 1 Parameter 100 ng ml-' 75As in 1% HNO,/counts s-' Blank (1% HNO,)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-' 75As in 0.02 rnol 1-' NaCl solution/counts s-' Blank (0.02 mol 1-' NaCl)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-' 75As in 0.04 moll-' NaCl solution/counts s-' Blank (0.04 mol 1-' NaCl)/counts s-' Signal- to-blank ratio Relative enhancement 100 ng ml-' "As in 0.01 mol 1-' Ca(NO,) solution/counts S - I Blank (0.01 rnol 1-' Ca(NO,),/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-' 75As in 0.02 mol 1-' Na,HPO solution/counts S-' Blank (0.02 mol 1-' Na,HPO,)/counts s-' Signal-to-blank ratio Relative enhancement CHF flow rate/ml min-' 18 1 0 37541 144 26 1 1 .o 3 1597 595 53 1 .o 26905 2043 13 1 .o 60823 4208 14 1 .o 39266 416 94 1 .o 1 93152 126 739 2.8 8098 1 229 3 54 6.7 82515 456 181 13.7 150149 1787 84 5.8 96340 675 143 1.5 2 167737 162 1035 4.0 147480 164 899 16.9 133620 27 1 493 37.4 206307 1522 136 9.4 211719 1589 133 1.4 3 168737 45 1 374 1.4 142546 249 572 10.8 130135 355 367 27.8 150258 1072 140 9.7 226500 1813 125 1.3 4 124150 589 21 1 0.8 112051 365 307 5.8 101451 520 195 14.8 929 18 64 1 145 10.0 193749 2119 91 1.6 5 86930 550 158 0.6 78213 457 171 3.2 72030 646 112 8.5 54083 403 134 9.3 161458 2476 65 0.7 x I /-A- - C - I I I I I 0 1 2 3 4 5 0 1 2 3 4 5 CHF flow rate/ml min-' Fig.1 Effect of addition of CHF (a) on 75As and (b) on the blank at m/z 75 for different concentrations of NaCl A 1% HNO,; B 0.02 moll-' NaCl; and C 0.04 mol I-' NaC1. most abundant Te isotopes ("'Te and l3'Te) but this was not observed. The addition of CHF as shown in Table 3 has only a slight enhancement effect on this element without affecting the blank. Copper The behaviour of Cu towards the addition of CHF was studied in solutions of 5% HNO 0.02 moll-' Na2S04 0.02 moll-' Na2HP04 and 5% HC1. All of the results for Cu are summarized in Table4.In HNO no interferences were expected and the addition of CHF follows the pattern as with elements such as In Pb and U i.e. a slight increase followed by a strong decrease when the CHF flow rate is increased. The situation is different for 63Cu (69.2%) in an Na2S04 matrix. At a CHF flow rate of zero a strong interference of 40Ar23Naf is observed which disappears even with low flow rates of CHF,. The count rate for the element also increased consequently at a 1-2mlmin-' CHF flow rate the relative enhancement is between 28 and 30. The observed decrease in the blank signal in this case is consistent with the competition reactions of argon with the matrix elements uersus the reactivity with CHF,. The removal of the 40Ar23Na+ interference in samples of waste water was achieved by preconcentration and matrix removal using an iminoacetate resin,Ig in serum by size-exclusion separation followed by 63Cu 65Cu isotope ratio determination2' and by subtraction of a synthetic blank con- taining the matrix elements.21 The blank interferences at m/z 65 in this experiment are mostly ,,SI6O 2 + 32S'60170+ 32S33S+ and H32S1602+ which probably exhibit low reactivity with CHF except for a possible abstraction of H from the H32S1602+.As the element count rate increase with addition of CHF is very moderate the maximum enhancement effect for this isotope is only 2.9. In the Cu-Na2HP0 s stem there are two interferences on the blank at m/z 63 40Ar Na+ and 31P1602+. Comparing this blank with the blank in the Cu-Na2S04 experiment it is evident that about 1 x lo4 counts s-l are contributed by 40Ar23Na+ and about 4 x lo3 counts s-l by 31P1602+.The first interference disap- pears almost completely with a 1 ml min-' addition of CHF whereas PO2+ is apparently not affected. A relative enhance- ment of 9.5 is calculated for this system. The blank level at m/z 65 is almost constant and the 65Cu+ signal is enhanced only by a factor less than 4 so that the relative enhancement is about this value. The addition of CHF to Cu-HCl solutions provides an interesting case. The Cu+ signal is enhanced to a maximum at a 1-2mlmin-' flow rate of CHF,. The blank HCl solution at a CHF flow rate of zero has no interferences at 63 and 65 m/z22 but addition of CHF increases the blank counts maintaining a blank (63) blank (65) ratio very close to the 35C1:37C1 ratio. It was therefore assumed that these interferences are C03'Clf and C037Cl+ their intensity depending on the availability of carbon i.e.on the CHF flow rate. 7 3722 Table 3 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1994 VOL. 9 Addition of CHF3 to Se or Te in HNO and HNO solutions > r 10 G cr c CHF3 flow rate/ml min-' - 5 C F _ _ B t 1 I r3 " e Parameter 100 ng ml-l 78Se in 1% HNO,/counts s-' Blank (1% HNO,)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-1 82Se in 1% HNO,/counts s-' Blank (1% HNO,)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-' IZ4Te in 1% HNO,/counts s-l Blank (1% HNO,)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-l lZ6Te in 1% HNO,/counts s-' Blank ( 1 YO HNO,)/counts s - Signal-to-blank ratio Relative enhancement 100 ng ml-I IZ8Te in 1% HNO,/counts s-l Blank (1% HNO,)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-' 130Te in 1% HNO,/counts s-' Blank (1 % HNO,)/counts s-' Signal-to-blank ratio Relative enhancement 0 13601 3963 3.4 1 .o 3870 63 61.4 1 .o 4953 832 6.0 1 .o 16290 18 905.0 1.0 27856 49 568.5 1 .o 30351 59 514.4 1 .o 1 28237 2603 10.8 3.2 9567 100 95.7 1.6 11128 953 11.7 2.0 40574 23 1764.1 1.9 69259 87 796.1 1.4 75139 117 642.2 1.2 2 36372 985 36.9 10.8 13462 80 168.3 2.7 11763 69 5 16.9 2.8 43876 22 1994.0 2.2 75024 64 1172.3 2.1 8 1346 70 1162.1 2.3 3 26304 746 35.3 10.3 10003 114 87.7 1.4 7004 392 17.9 3.0 26466 12 2205.5 2.4 449 15 36 1247.6 2.2 49184 37 1329.3 2.6 4 18072 668 27.1 7.9 6826 96 71.1 1.2 3758 221 17.0 2.9 14026 13 1078.9 1.2 23980 28 856.4 1.5 25977 21 1237.0 2.4 5 13111 650 20.2 5.9 4802 125 38.4 0.6 2 143 126 17.0 2.9 8381 11 761.9 0.8 14210 18 789.4 1.4 15512 16 969.5 1.9 .- 40 1 I Fig.2 the blank in 1% HN03. Effect of addition of CHF on A 78Se in 1% HNO and B on Zinc This element was studied in matrices of 5% HCl 5% HNO 0.02 mol 1-1Na2S04 and 0.02 mol 1-lNa2PO4 The experi- ments with the 5% acid solutions which are slightly higher concentrations than normally used in ICP-MS applications were carried out to increase the formation probability of yet unknown interferences originating from HCl or HNO reacting with CHF in the 63-70 m/z range. Such interferences have been observed in the case of Cu with 5% HC1.For Zn isotopes no evidence for interferences of this type were observed. Only for 64Zn (the most abundant isotope 48.6%) in the Na2HPO4 solution has a relative enhancement owing to CHF been observed. There is an increase in the signal intensity for this element parallel to the decrease in the blank with an overall relative enhancement of about 9 for a CHF flow rate of between 2-3 ml min-l. The possible interferences at m/z 64 are H31P1602+ and 31P160170+. The results are presented in Table 5. Methods for reducing phosphate inter- ferences on 64Zn have not been considered in the literature. Concerning the Na2S04 matrix CHF did not affect the strong 32S1602+ and 34S1602+ interferences at m/z values of 64 and 66 and also no effect was observed on the 64Zn and %Zn signals.The various analytes the matrices used their concentrations and the consequent interferences for the elements discussed above are summarized in Table 6. Other Elements and Blanks The effect of addition of CHF on 100ngml-' solutions in 1% HNO of a series of elements listed under Experimental were studied. Taking the relative enhancement as a criteria for enhancement by CHF the following cases (apart from the already discussed As Se Te Cu and Zn) were identified. (i) An increase in the ion signal to a maximum at a CHF flow rate of about 2rnlmin-l followed by a decrease in the signal at higher flow rates. The blank intensity increased with increasing CHF flow rate. In this category are Be and Al. The sample counts for Be were 3 x lo5 4.1 x lo5 9.1 x lo5 6.2 x lo5 and 3.7 x lo5 counts s-l and the blank counts 100 200 900 1500 1700 and 1200 counts sK1 for the addition of 0 1 2 3 4 and 5 mlmin-1 CHF respectively. For A1 the sample counts were approximately the same as for Be but the blank increased from 5000 counts s-l at zero CHF flow rate to 55 000 counts s-l at 4 ml min-l CHF addition.The blank increase in both cases could be a result of stripping from the cones since both elements are present in the mass calibration solution. (ii) A continuous decrease in ion signal along the whole CHF flow rate range followed by a stable low or slightly increasing blank signal. In this category are Ga Ge Br In Sn Sb I Hg Pb and U. (iii). A strong increase in the blank counts in 1% HNO for 0-3mlminP1 additions of CHF were observed at m/z values of 48 (Ti) (800-2.2 x lo4 counts s-'); 51 (V) (200-2.4 x lo4 counts s-'); 52 (Cr) (1000-3 x lo5 counts s-'); 54 (Fe) (9000-9 x lo4 counts s-'); and 59 (Co) (500-104 counts s-I).Moderate blank increases under the same conditions were observed at m/z values of 53 (Cr) (100-4000 counts s-'); 55 (Mn) (700-3000 counts s-'); and 57 (Fe) (900-2000 counts s-I); The polyatomic interferencesJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1994 VOL. 9 723 Table 4 Addition of CHF to Cu plus matrix and matrix only solutions CHF flow rate/ml min-' Parameter 100 ng ml-' 63Cu in 5% HNO,/counts s-' Blank (5% HNO,)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-' 65Cu in 5% HNO,/counts s-' Blank ( 5 % HNO,)/counts s - ' Signal-to-blank ratio Relative enhancement 100 ng ml-' 63Cu in 5% HCl/counts s-' Blank (5% HCl)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-' 65Cu in 5% HCl/counts s-' Blank (5% HCl)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-' 63Cu in 0.02 mol 1-' Na2S0 solution/counts s-' Blank (0.02 mol 1-' Na,SO solution)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-' 65Cu in 0.02 moll-' Na,HPO solution/counts s-' Blank (0.02 mol 1-' Na2HP04 solution)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-' 63Cu in 0.02 mol I-' Na2S04 solution/counts s-' Blank (0.02 moll-' Na2S04 solution)/counts s-' Signal-to-blank ratio Relative enhancement 100 ng ml-I 65Cu in 0.02 mol I-' Na2S0 solution/counts s-' Blank (0.02 mol 1-' Na2S0 solution)/counts s-' Signal-to-blank ratio Relative enhancement 0 226698 524 433 1 .o 102373 294 348 1 .o 216578 619 350 1 .o 96393 273 353 1 .o 102360 14644 7.0 1 .o 44578 1013 44 1 .o 176163 10419 17 1 .o 88129 8467 10.4 1 .o 1 318670 599 532 1.2 140265 332 422 1.2 333522 1466 228 0.7 145161 533 272 0.8 155568 4350 35.8 5.1 703 17 860 82 1.9 260796 543 480 28.4 121691 9466 12.9 1.2 2 310320 603 515 1.2 134219 304 442 1.3 295976 2679 110 0.3 128320 968 133 0.4 382601 5731 66.8 9.6 164782 1176 140 3.2 237874 47 1 505 29.9 109819 3685 29.8 2.9 3 183970 467 394 0.9 79458 210 378 1.1 166979 2750 61 0.2 73145 921 79 0.2 324351 5324 60.9 8.7 136422 896 152 3.5 145177 566 256 15.2 68041 2583 26.3 2.5 4 140840 379 372 0.9 45487 159 286 0.8 103656 2741 38 0.1 45348 930 49 0.1 196744 4295 45.8 6.6 84775 806 105 2.4 84264 878 96 5.7 40097 2714 14.8 1.4 5 69896 388 180 0.4 30424 166 183 0.5 73228 2605 28 0.1 32323 874 37 0.1 129092 3944 32.7 4.7 53502 968 55 1.3 52734 2918 18 1.1 25062 2664 9.4 0.9 Table 5 Addition of CHF to Zn-Na,HPO and Na2HP0 solutions CHF3 flow rate/ml min-' Parameter 0 1 2 3 4 5 100 ng ml-' 64Zn in 0.02 mol 1-' Na,HPO solution/counts s-' 39034 56640 125526 118492 83950 59523 Signal-to-blank ratio 11.1 62.2 96.7 100.3 82.1 63.1 Relative enhancement 1 .o 5.6 8.7 9.1 7.4 5.7 Blank (0.02 mol 1-' Na2HP04 solution)/counts s-l 3524 910 1298 1181 1023 943 which originate from CHF (in addition to those originating from HN0322) presumably are m/z 48 36Ar12C+; m/z 52 40Ar12C+; m/z 53 40Ar13C+; m/z 54 F2I60+; m/z 55 36ArF+; m/z 57 38ArF+; and m/z 59 40ArF+. Effect of Glycerol Following the results from Allain et al.on the enhancement effect of glycerol on As Se and Te sol~tions,~ glycerol solutions were studied with the addition of CHF,. No improvement of ion signal intensities was observed in the presence of CHF,. Analytical Applications of Addition of CHF It is immediately evident that the reduction of an interference and under the same conditions an increase in the analyte signal will improve the detection limits (DL). The DLs for As '%e 63Cu and 64Zn in various matrices are summarized in Table 7. The following expression was used DL=3 x (SD) x Cs/(Is-IB) where (SD) C Is and IB are the stan- dard deviation of the blank (in counts s-') concentration of the standard counts of the standard and counts of the blank (both in counts s-') respectively.It should be noted that the detection limits improved by factors of 7-38 except for As in 1% HNO where the factor is only 2. A further important point is the stability of the plasma with the addition of CHF,. The following cases have been studied at CHF flow rates of 0 and 2mlmin-I (a) As in the concentration range 5-50 ng m1-I; (b) As 0.25-1 ng m1-I; (c) 78Se and 82Se 0.5-5 ngml-I (total Se); (d) 63Cu 0.25-1 ng ml-1 (total Cu); and (e) 64Zn 1-5 ng ml-1 (total Zn). (a) Arsenic was determined in the National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 3172 Multielement Mix B Standard Solution which contains 200+ 1 jig ml-1 of As and nine other metallic elements (Ba Ca Co Cu Pb Se Ag Sr and Zn) in the range724 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1994 VOL.9 Table 6 Summary of analytes matrices and their interferences ~ Analyte 75As 77Se 78Se 82Se 63cu 65cu 66Zn Matrix and concentration NaCl(0-0.04 moll-') Ca(N03)2 (0.02 mol 1-' 1 Na2HP04 (0.02 mol 1- ') + CHF NaCl(0-0.04 moll-') HN03 ( 1 YO) + CHF - HCl(5Yo) + CHF3 Na2S04 (0.02 mol 1-') Na2HP04 (0.02 mol 1-l) Na,S04 (0.02 mol 1- ') HC1( 5%) + CHF3 Na2HP04 (0.02 mol 1 - '11 Na2S04 (0.02 mol I-') Na2HP04 (0.02 mol 1-':I Na2S04 (0.02 mol 1-') 1 Interference 40Ar35C1 + 43ca160 + 12c31p'602+ 40Ar37Cl + 40Ar38Ar + 12CF_14N1602+ - 12C16035c1+ 40Ar23Na + 4 0 ~ ~ 2 3 ~ ~ + 31p1602+ 2 ~ 1 6 0 3 7 ~ 1 + 12 c 18 0 35 c1+ 3 3 ~ 1 6 0 + 32 16 17 + ~ 3 2 ~ 1 6 0 ~ + 3 2 ~ 3 3 s + 31p160180f 9 2 s o o 32s'602+ ~31p160~+ 31p160170+ 3 4 ~ 1 6 0 + 2 Table7 Effect of addition of CHF on detection limits in different matrices; detection limits (3 x SD) are given in ng ml-' CHF3 flow rate/ml min-' Analyte and matrix As-1% HNO As-0.02 moll-' NaCl As-0.04 moll-' NaCl As-0.01 moll-' Ca(N03)2 78~e-1% HNO 63Cu-0.02 mol 1-' Na2S0 63Cu-0.02 mol 1-' Na2HP04 64Zn-0.02 moll- Na2HP04 0 0.032 0.23 0.65 0.28 0.88 0.35 0.53 0.42 2 0.014 0.01 0.02 0.04 0.032 0.022 0.089 0.01 1 10-500pgml-' in a 5% HNO matrix.The standard was diluted to yield a 18.93 ng ml-' of As solution in 1% HNO and 760 pg ml-' of NaCl were added to simulate potable water. The results are summarized in Table 8. The table reveals that the sensitivity in the presence of CHF is almost six times higher compared with the absence of CHF and the accuracy and precision are comparably or slightly improved 18.89 & 0.05 uersus 18.51 kO.19 ng ml-' of As (calculated 18.93 ng ml-' of As).It should be pointed out that other potable water samples (laboratory working standards) were also analysed. These samples contain variable amounts of As and other cations (Cu Table 8 Determination of As in NIST/SRM 3 172/B in the presence of 760 ppm of NaCl(1 YO HNO solution) CHF flow rate/ml min-' Parameter 0 2 Blank*/counts s-' 1130 & 30 258 f 1 Mean sensitivityt/counts s- ' per 343 1994 ng ml-' of As As concentration measured/ng ml-l 18.51 f0.19 18.89 f 0.05 Regression coefficient 0.9999 1 1 As concentration calculatedjng ml-' 18.93 ~~ ~ * Quoted errors are & 1 SD.7 Mean sensitivities are for blank subtracted analytical curves. Zn Cd Cr Hg Ni Pb and NH4+) and anions (S04-2 NO3- and P205) in a 760 pg ml-' C1- solution. The sensitivity of the blank subtracted analytical curve without CHF was 330+ 10 counts s-' per ng ml-' of As. Whenever this was sufficient excellent agreement was obtained for As between the present measurements and the mean values (as established by nine independent laboratories). When CHF was used to enhance the sensitivity the present results for As were 10-20% higher than the mean values. This erratic difference is attributed to an interference introduced at m/z 75 by one or more components in solution. Solutions containing each of the matrix components at the same concentrations as in the laboratory standard have been tested separately with and without CHF3 for increments in the blank counts at m/z 75.Increments in the blank signal have been observed only for a phosphate solution (0.02 moll-' Na2HP04) and the inter- ference was attributed to 12C3'P160 2'. The enhancement effect for a 100 ng ml-' As solution with 0.02 moll-' Na2HP04 was studied and the results are also given in Table 2. It is evident that As solutions can also be analysed in the presence of phosphates against an appropriate blank. (b) Solutions of 0.25,0.5 and 1 ng ml-' of As in 0.04 moll-' NaCl solution and 1% HNO produced without CHF a blank of 2683 counts s-' and non-linearly increasing readings between 2900 and 3400 counts s-' whereas with 2 ml min-l CHF the blank was reduced to 411 counts s-l and the blank subtracted analytical curve showed a sensitivity of 1250+38 counts s-' per ng ml-' of As.(c) For 0.5 1 and 5 ng ml-' solutions of Se in 1% HNO without CHF3 at m/z 78 a blank of 5004countss-' and solution readings randomly scattered between 5800 and 6350 counts s-' were observed. With CHF the blank was reduced to 818 counts s-l and the blank subtracted analytical curve yielded a sensitivity of 444 & 39 counts s-' per ng ml-' total of Se. For 82Se without CHF3 the blank was 94 counts s-' and the sensitivity 30+ 15 counts s-' per ng ml-l of total Se whereas with CHF the blank was 88countss-' but the sensitivity increased more than six times to 193 & 23 counts s-' per ng ml-' of total Se. (d) For 0.25-0.5 and 1 ngml-' of Cu in a 0.02moll-' Na2S04 solution and 1% HNO without CHF the blank at 63Cu was 8500 counts s-' and solution readings were randomly scattered between 8200 and 9700 counts s-' whereas with CHF the blank was greatly reduced to 308 countss-' and the blank subtracted analytical curve showed a sensitivity of 1888+25 counts s-l per ng ml-' of total Cu.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1994 VOL.9 725 Table 9 Calibration of As Se Cu and Zn in interfering matrices with a CHF3 flow rate of 2 ml min-'; quoted errors are f 1 SD Analyte (b) 75As+0.04 mol I-' Nucl- Blank 0.24 ng ml- ' As 0.47 ng ml - ' As 0.97 ng ml-' As (c) "Se- Blank 0.49 ng ml-' Se 0.99 ng ml-' Se 5.04 ng ml-' Se (c) "Se- Blank 0.49 ng ml-' Se 0.99 ng ml- ' Se 5.04 ng ml-' Se ( d ) 63Cu in 0.02 mol 1-' Nu2S04- Blank 0.25 ng ml-' Cu 0.50 ng ml-' Cu 1 ngml-' Cu (e) 64Zn in 0.02 mol I-' Na2HP04- Blank 1.01 ngml-' Zn 4.99 ng ml - ' Zn 10.05 ng ml - ' Zn Signal/ counts s-' 41 1 715 992 1602 Mean sensitivity* Regression coefficient 818 1020 1297 3023 Mean sensitivity Regression coefficient 88 179 304 977 Mean sensitivity Regression coefficient 308 778 1240 2225 Mean sensitivity Regression coefficient 2507 3735 7674 13131 Mean sensitivity Regression coefficient Blank subtracted signal/counts s - ' 0 304 58 1 1191 0 202 479 2505 0 91 216 889 0 470 934 1917 0 1228 5167 10624 Sensitivity/counts s - ' per ppb of element - 1294 1226 123 1 1250fi 38 0.99993 - 409 486 437 444k39 0.99976 - 185 219 176 193 1 2 3 0.99977 - 1880 1686 1917 1888 2 25 0.99981 - 1209 1036 1057 1 100 Ifr 94 0.99955 * Mean sensitivities are for blank subtracted analytical curves.(e) For 1 s and 10 ng ml-' of Zn the blank at 64Zn without CHF was 3700 counts s-' and the sensitivity 414k 38 counts s-'per ng ml-' of total Zn as compared with 2507 counts s-' and 1100+94 counts s-' per ng ml-' of total Zn with CHF respectively. The data with CHF for calibrations (b)-(e) are summarized in Table 9. General Comments For the cases where the net effect of addition of CHF is positive such as As Se Cu and Zn there is a plausible explanation for the decreasing intensity in the blank signal. It is most probable that the corresponding polyatomic interfering species 40Ar35Cl+ 40Ar38Ar+ and 40Ar23Na+ at m/z values of 75 78 63 are produced with a lower reaction rate owing to the competition of species originating from CHF with compo- nents in the matrices.For the blank at m/z 64 in an Na2HP04 solution two interferences were suggested 31P160170f which are considered as non-reactive (see Cu-Na,HPO,) and H31P160,+ which can be eliminated by CHF by abstraction of H thus the blank is significantly reduced. The interpretation of the element count rate increase in the presence of CHF is not straight forward. Allain et aL3 observed a signal enhancement in a 1 moll-' glycerol (and also glucose) solution for Hg (6OO%) Au (325%) Se (250%) As (110%) and Te (190%). For other elements such as Bi Co Eu Ho I In La Nb Ni Pb Pt Sn Sr T1 and U the change in the signal was between 90 and 110% (all relative to signal intensity without glycerol).In their work the effect of glycerol on the blank was not reported. Allain et al. also observed that addition of methane to the argon of up to 6% (v/v) increased the ion signals for As Se and Te. They tried to correlate the ionization enhancement with the IP of the element noting that the effect is observed for elements with IP values of between 9 and 11 eV. Iodine (10.44 eV) was an exception as was also Br (11.30 eV). Their argument was that addition of organic com- pounds modifies the ionization equilibrium over a limited range of energies. The IP for C (11.20 eV) is slightly above this range. In the present experiments no enhancement was observed for Hg (the element most affected in the work by Allain et al.) or for iodide and bromide.An effect on Zn (9.39 eV) and on Be (9.32 eV) which supports their hypothesis was noted but there was also an effect on Cu and Al which have IP values of only 7.73 and 5.99 eV respectively. A further parameter that affects the absolute intensities of the measured elemental and blank currents at the different flow rates of CHF is the relative plasma torch position along the x-axis (the axis of the quadrupole sampling cone and torch). On moving the torch along this axis variable counts for blank and for element signals were observed. This effect was not studied systematically because it was difficult to move the torch without shutting off the discharge. In the present work a 'relative enhancement' has been defined. Generally it was observed that this effect was maintained independent of726 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY JUNE 1994 VOL.9 the motion of the torch. As this parameter was not optimized it is fairly likely that larger enhancements could be observed by fine tuning the torch position. The authors express their thanks to Professor S. Facchetti Head of the Soil Water and Waste Unit Environment Institute J. R. C. Ispra for his helpful comments and to Dr. H. W. Muntau Head of the Chemistry of the Aquatic Systems Section at the above Unit for supplying the potable water laboratory working standards. References Evans E. H. and Giglio J. J. J. Anal. At. Spectrorn. 1993 8 1. Hill S. J. Ford M. J. and Ebdon L. J. Anal. At. Spectrom. 1992 7 1157. Allain P. Jaunault L. Mauras Y. Mermet M. and Delaporte T.Anal. Chem. 1991 63 1497. Lam J. W. H. and Horlick G. Spectrochim. Acta. Part B 1990 45 1313. Branch S. Corns W. T. Ebdon L. Hill S. and ONeill P. J. Anal. At. Spectrom. 1991 6 155. McLaren J. W. Beauchemin D. and Berman S . S . J. Anal. At. Spectrom. 1987 2 277. Lyon T. D. B. Fell G. S. Hutton R. C. and Eaton A. N. J. Anal. At. Spectrom. 1988 3 265. Lyon T. D. B. Fell G. S. Hutton R. C. and Eaton A. N. J. Anal. At. Spectrom. 1988 3 601. 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Heitkemper D. Creed J. Caruso J. A. and Fricke F. L. J. Anal. At. Spectrom. 1989 4 279. Sheppard B. S. Shen W. Caruso J. A. Heitkemper D. and Fricke F. L. J. Anal. At. Spectrorn. 1990 5 431. Branch S. Ebdon L. Ford M. Foulkes M. and O'Neill P. J. Anal. At. Spectrom. 1991 6 151. Wang J. Evans E. H. and Caruso J. A. J. Anal. At. Spectrom. 1992 7 9.29. Ridout P. S. Jones H. R. and Williams J. G. Analyst 1988 113 1383. Munro S. Ebdon L. and McWeeny D. J. J. Anal. At. Spectrom. 1986 1 211. McLeod C. W. Date A. R. and Cheung Y. Y. Spectrochim. Acta Part B 1986 41 169. Vaughan M. and Horlik G. J. Anal. At. Spectrom. 1989 4 45. Buckley W. T. Budac J. J. Godfrey D. V. and Koenig K. M. Anal. Chem. 1992 64 724. Ebdon L. Foulkes M. E. Parry H. G. M. and Tye C. J. Anal. At. Spectrom. 1988 3 753. Heithmar E. M. Hinners T. A. Rowan J. T. and Riviello J. M. Anal. Chem. 1990 62 857. Lyon T. r). B. and Fell G. S. J. Anal. At. Spectrom. 1990 5 135. Vanhoe H. Vandecasteele C. Versieck J. and Dams R. Anal. Chem. 1989 61 1851. Tan S. H. and Horlick G. Appl. Spectrosc. 1986 40 445. Paper 3 1074431 Received December 20 1993 Accepted March 16 1994