JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 477 Direct Determination of Zinc in Sea-water Using Electrothermal Atomic Absorption Spectrometry With Zeeman-effect Background Correction Effects of Chemical and Spectral Interferences" J. Y. Cabon and A. Le Bihan Unite de Recherche Associee au CNRS No. 322 Universite de Brefagne Occidentale 6 Avenue Le Gorgeu 29275 Bfest-Cedex France The determination of zinc in sea-water using an electrothermal atomic absorption spectrometry system with Zeeman-effect background correction is presented. The influence of various chloride and nitrate salts on the atomization signal of zinc was examined. In chloride medium particularly the interference effect induced through losses of zinc chloride by the thermohydrolysis of magnesium chloride and simultaneous generation of HCI during the pyrolysis step is noted.In nitrate medium zinc is more stabilized by Mg>Ca>Na> NH:. The effect of various inorganic and organic acids used as chemical modifiers on the atomization of zinc and background absorption signals in sea-water were examined. In unmodified sea-water a Zeeman interference effect related to the vaporization of the chloride matrix leading to a systematic under- compensation and consequently to erroneous zinc concentration values was observed. In sea-water modified with 1 mol I-' nitric acid a spectral Zeeman interference effect induced by the Zeeman splitting of the absorption bands of NO molecules generated during the decomposition-reduction of nitrate was observed; the induced over-compensation is eliminated by selective pyrolysis at about 850 "C.The chemical interference effect (25%) is related to the simultaneous vaporization of zinc and sodium oxides; the detection limit (3a) being about 80 ng I-' for a 10 pI injected volume of sea-water. In sea-water modified with 0.7 mol I-' oxalic acid there is no significant interference effect and the detection limit in this medium is about 60 ng I-' for a 10 pl injected volume of sea-water. Keywords Atomic absorption spectrometry with Zeeman-effect background correction; zinc determination; sea-water; spectral and chemical interference; nitric and oxalic acid The determination of trace metals in the pgl-' range in sea- water necessitates very stringent precautions against contami- nation.Hence the direct determination of zinc in sea-water by electrothermal atomic absorption spectrometry (ETAAS) with- out a preconcentration step is of interest to the analyst. Owing to the high background absorption signal generated at 213.9 nm by the volatilization of the chloride matrix direct determination of zinc at low-level concentrations in sea-water remains difficult. In previous work zinc has been determined directly in sea-water without a rnodifier,'q2 or by using nitric acid ammonium nitrate4 and citric acid5 as modifiers. All these studies have been performed using deuterium-arc back- ground correction; no platforms; no simultaneous vizualization of specific and background signals leading to a loss of infor- mation; peak height measurements [chart recorder (no digitiz- ation of the signals)]; slow atomization ramps generally under gas flow; and empirical chemical conditions leading to various detection limits (30) of 2.55 (ref.2) 0.6 (ref. 4) and 0.27 pg 1-' (ref. 5 ) . In order to improve the analytical performance of ETAAS for direct determination of zinc in sea-water a more comprehensive study of the atomization of zinc in this medium has been made using a recent ETAAS instrument with a longitudinal-effect background correction system and iso- thermal atomization. In the first part of this work the influence of various chloride (including NaC1 CaC1 and MgC12 naturally present in sea- water) and nitrate salts [including Mg( N03)2 recommended as modifier by the manufacturer] on the atomization signal of zinc in water was studied. In the second part the influence of inorganic and organic acids used as modifiers on the atomization signal of zinc and the sea-water background absorption signal was studied.[Owing to the difficulties in purifying acidic solutions of stabilizing modifiers (Pd and Pt) at the sea-water level concen- * Presented at the XVIII Colloquium Spectroscopicum Inter- nationale (CSI) Post-Symposium on Graphite Atomizer Techniques in Analytical Spectrometry Durham UK July 4-7 1993. trations the influence of such modifiers was not investigated in this study.] The experimental conditions for the determi- nation of zinc at low level concentrations in unmodified sea- water and in the presence of nitric and oxalic acids were optimized. Experimental Reagents Merck suprapure grade.Hydrochloric nitric sulfuric phosphoric and oxalic acids. Sodium magnesium and calcium nitrates. Merck pro analysi. Sodium magnesium and calcium chlorides. Merck pro analysi. Ethylenediaminetetraacetic acid (Na,EDTA) tetrasodium Zinc standard solution 1 g 1-' in 0.5 moll-' HNO,. Merck. Sea-water reference material for zinc determinations. NASS-3 (0.178+0.025 pg 1-' of Zn) and CASS-2 (1.97kO.12 pg I-' of Zn) from the National Research Council Ottawa Canada were used. In this case ultrapure HNO (Merck) was used. Ultrapure water from a Millipore Mro-MQ system was used. Blank determinations on 1 moll-' nitric and 0.7 mol I-' oxalic acids solutions were respectively 0.05 and 0.15 pg 1-' of Zn. salt. Aldrich. Instrumentation A Perkin-Elmer 4100ZL was used for all the atomic absorption measurements.Pyrolytic graphite coated graphite tubes equipped with pyrolytic graphite coated platforms (Perkin- Elmer) were used. Samples and modifier solutions were deliv- ered to the furnace using a Perkin-Elmer AS-70. The light source was a Perkin-Elmer hollow cathode lamp operating at 20 mA. The zinc resonance line at 213.9 nm was used with a 0.7 nm spectral slit-width. The inert gas was argon. Dilutions were carried out with calibrated Gilson Pipetman pneumatic syringes. Typical operating conditions were drying 120 "C t(s) 250 ml min-' argon; pyrolysis T("C) 60 s 250 ml min-' argon; cooling 100 "C 10 s 250 ml min-' argon; atomization,478 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 0.2 T("C) ramp step 0 gas flow interrupted; and cleaning 2500 "C 5 s 250 ml min-' argon.A C - I I A A Results and Discussion Water Medium Fig. 1 shows the evolution of the atomization signal of zinc with temperature in ultrapure water. As it appears the shape of the signal evolves from a single peak shape corresponding to the atomization mechanism:6 ZnO(s)eZnO (g)$Zn( g) + 1/20 to a two peak signal shape corresponding to a modification of the atomization mechanism and probably to a partial reduction of ZnO(s) to Zn by the graphite at the lowest temperature which also could explain the loss of zinc for relatively low calcination temperatures (about 450 "C). The maximum inte- grated absorbance value is reached for an atomization tempera- ture close to 1100°C [rn,=0.5 pg; at 1800 "C in the presence of Mg(N03) the rn reported by the manufacturer is 1.0 pg].The maximum peak height absorbance value is reached for temperatures above 1600 "C (rn = 1.0 pg]; the integration time being 10 times longer at 1100°C than at temperatures above 1600 "C. Theoretical rn values have been calculated from the model of L'vov7 with the use of the software of Berglund and Baxter.' Above 1 lOO"C there is good agreement between experimental and theoretical rn values. So atom formation is quantitative and atom removal is predominantly diffusion controlled. Under these conditions the calculated atomization efficiency &'A is close to 100% and not temperature dependent. Chloride Medium The influence of ammonium sodium magnesium and calcium chlorides at a 0.1 moll-' chloride concentration on the atomiz- ation signal of zinc was examined.Without pyrolysis and with the use of a 0 s atomization ramp step to 1600 "C no significant changes in the shape of the atomization signal of zinc were observed [Fig. 2(a)]. Interference effects on the atomization signal of zinc are not important and are essentially related to the stabilization of ZnC1 by the simultaneous vaporization of the chloride salts (consistent with boiling-point data"); the importance of the depressing effect being related to the amount of chloride in the vapour phase [Fig. 2(a) and (b)]. For this chloride concentration value NaCl and CaC1 have a small interference effect (about 10%); and NH4C1 and MgCl have a depression effect of about 20%. Moreover in the presence of MgC1 at the concentration level of sea-water (about 0.05 mol 1-I) for a slow atomization ramp step as is generally 0-'0° t I 0 co 0.350 2 n 0 2.5 Time/s 5 Fig.1 Influence of temperature on atomization signal of zinc in water ( 5 pl 10 pg 1-' of Zn) A 2000; B 1600; C 1400; D 1300; E 1200; F 1150; and G 1100°C $ 0 4.98 Fig. 2 (a) Zinc atomization signals ( 5 pl 10 pg 1-' of Zn) in the presence of A 0.1 mol I-' NH,Cl; B 0.1 moll-' NaCl; C 0.05 moll-' MgCl,; and D 0.05 moll-' CaCl,. (6) Background absorption signals for 5 pl of A B C and D used for the determination of zinc in sea-water a strong interference effect was observed depending on the heating rate. The influence of the pyrolysis temperature on the atomiz- ation signal of zinc in the presence of these different chlorides was examined.The variations of the integrated absorbance of zinc and of the background absorption signal are presented in Fig. 3(a) and (b). In the presence of MgCl a significant loss of Zn was observed for a pyrolysis temperature of >4OO"C with a corresponding decrease of the MgCl background absorption signal. For temperatures > 400 "C MgC1 is ther- mohydrolysed on the graphite to MgO (non-absorbing species under these experimental conditions) with the generation of hydrogen ~hloride,''-'~ zinc is lost during pyrolysis and as HC1 is simultaneously generated ZnCl is stabilized in the vapour phase owing to the high chloride concentration and blown out of the furnace. For CaC1 and NaCl higher pyrolysis temperatures can be used but Zn is less stabilized by CaC1 than by NaCl despite its higher volatilization temperature 0.20 v) -.8 0.16 n C (D 0.12 v) 0 0.08 4- E 0.04 4- - 0 (b) 0.8 0.6 100 300 500 700 960 1100 1 TemperatureK 1 DO Fig.3 (a) Influence of pyrolysis temperature on (a) integrated absorbance of zinc (5 pl 10 pg 1-' of Zn) in the presence of A 0.1 moll-' NH4C1; B 0.1 moll-' NaC1; C 0.05 MgC1,; and D 0.05moll-' CaC1,; and (b) on background absorption signal of 5 pl of A B C and DJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 479 m 0 f 4.96 0 4.96 V A A n7nn I 0 4.96 0 Timels 5.00 Fig. 4 Zinc atomization (solid line) and background absorption (broken line) signals in the presence of (a) 0.1 moll-' NaNO,; (b) 0.05 moll-' Mg(NO,),; (c) 0.05 moll-' Ca(NO,),; and ( d ) 0.1 moll-' NH,NO ( 5 p1 10 pg 1-' of Zn) [Fig.3 (b)]. This can be explained by a partial thermohydrolysis of CaCl to CaO which induces a loss of Zn as ZnCl species stabilized by the simultaneous generation of HCl. For NaCl the interference effect is not important and could be only due to the simultaneous vaporization of NaC1. In the presence of NH4C1 no loss of Zn was observed through the formation of ZnC1 or the ammonium bonded form during pyr~lysis,'~ but a stabilization of zinc on the graphite furnace was noted. Nitrate Medium The influence of ammonium sodium magnesium and calcium nitrates at a concentration of 0.1 moll-' on the atomization signal of zinc has been examined. As can be seen in Fig.4 without pyrolysis and using a 0 s atomization ramp the shape of the atomization signal of zinc is strongly modified therefore the integrated absorbance values are not very different (about 10%) from those obtained in water.The decomposition- reduction-volatilization step of the nitrate salts generates absorbing species (mainly NO) in the case of NH,NO NaNO Ca( and Mg( NO3) under these experimental conditions which induces an important spectral Zeeman interference leading to an over-compensation at low Zn concentrations.'6 Only sodium oxides resulting from sodium 0.20 VI -. 0.16 m + n 2 0.08 F 2 0.12 m 4- 0.04 e C - 0' I I 1 I 1 100 300 500 700 900 1100 1300 Tern perat u rePC Fig. 5 Influence of pyrolysis temperature on the atomization signal of zinc ( 5 p1 10 pg 1-' of Zn) in the presence of A 0.1 moll-' Mg(NO& B 0.1 moll-' NaN0,; C 0.05 moll-' Mg(N03)2; and D 0.05 moll-' Ca(NO,)* nitrate decomposition induce a background absorption signal simultaneously with the atomization signal of zinc; refractory oxides (CaO and MgO) being not yet volatilized.In Fig. 5 the variations of the integrated absorbance values with the pyrolysis temperature in the presence of the different metallic nitrates are presented. Zinc is stabilized by the nitrate salts which permit a higher pyrolysis temperature according to Mg( NO3) > Ca(N03) > NaN0 > NH,N03. The stabiliz- ation of ZnO is obtained through the adsorption of oxygen or oxygenated compounds on the graphite reductive active sites and through an occlusion process related to the respective vaporization temperatures of the thermally stable oxides.These stabilizing effects permit the elimination of the background absorption signal generated by the decomposition of the nitrate salts by selective pyrolysis and the suppression of the corre- sponding over-compensation. Sea-w a t er Medium In sea-water the atomization signal of zinc and the high background absorption signal generated by the volatilization of sea-water salts are not well separated and the integrated absorbance value is depressed by about 20% as compared with the integrated absorbance value obtained in water. Presence of Inorganic Acids The modification of the chloride matrix to a nitrate sulfate or phosphate matrix by using 1 moll-' HN03 H2S04 or H3PO4I7 does not lead to well separated atomization curves for the zinc and background absorption signals.The atomiz- ation signal of zinc is more complex and a delaying effect has been noted phosphate > sulfate > nitrate >chloride. Under these experimental conditions the integrated absorbance values are depressed by about 25% in a nitrate or sulfate medium and 40% in a phosphate medium. The use of 1 mol l-' HN03 as modifier leads to the smallest background absorption signal decreasing by about 10 times the sea-water background absorp- tion signal. Presence of Organic Acids In Fig. 6 the atomization signal for zinc and the background absorption signal obtained in sea-water in the presence of480 BG 2.000 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 (d 1 Fig. 6 Zinc atomization (solid line) and background absorption (broken line) signals in sea-water in the presence of (a) 0.1 moll-' ascorbic acid; (b) 0.1 moll-' citric acid; (c) 0.1 mol 1-' Na,EDTA; and ( d ) 1 0.1 moll-' oxalic acid and 2 0.7 moll-' oxalic acid ( 5 jd 10 pg 1-I) 0.1 moll-' citric and ascorbic acids and Na,EDTA and 0.1 and 0.7 moll-' oxalic acid are presented.The presence of the different acids leads to a complex atomization signal for zinc but the integrated absorbance values are only slightly smaller than in water (about 10%). The use of the organic acids facilitates the reduction of zinc in sea-water but does not permit a good separation of the atomization signal of zinc and the background absorption signal; zinc being partially seques- tered by the remaining chloride matrix or the corresponding oxides resulting from hydrolysis.It can be noted that for twice the citric acid concentration recommended by Guevremont,' only a partial separation of the atomization signal of zinc and the sea-water matrix absorption signal (which is not signifi- cantly reduced) was observed. Moreover the decompo- sition-volatilization of ascorbic acid citric acid and Na,EDTA generates a simultaneous background absorption signal which cannot be eliminated without losses of zinc. Oxalic acid does not generate any significant background absorption value. At a 0.7 moll-' concentration level by removing chloride as HCl and with conversion of salts into oxides (cf. HN03) it decreases the background absorption signal to a level not very different from that obtained in the presence of 1 moll-' HNO,; the remaining chemical interference effect being below 10%.Optimization of Experimental Conditions If in unmodified sea-water a single peak shape for the atomiz- ation signal of zinc at high concentrations not well separated from an important background signal is observed in the presence of the different inorganic and organic acids (used as modifiers) then the atomization signal of zinc must be complex corresponding to complex atomization mechanisms. The different modifiers studied do not lead to a good separation of the atomization signal of zinc and the background absorp- tion signal. The most important decrease of the background absorption signal is obtained with the use of nitric and oxalic acids so the best experimental conditions for determining zinc at low level concentrations in unmodified sea-water and sea- water modified with nitric or oxalic acid were determined.Chloride Zeeman Interference Effect in Unmodified Sea-water The 'atomization' signals of zinc obtained in a purified sea-water [ammonium pyrrolidin-1-yldithioformate (APDC)- Freon extraction] and with addition of 1 pg 1-1 of Zn are 0 5.00 Time/s Fig. 7 (a) 'Atomization' signal of zinc in A purified sea-water; B spiked (1 pg 1-') unmodified sea-water; and C background ( 5 p1 Tat= 1800°C). (b) Atomization signal of zinc in A 0.5 moll-' NaCl solution; B spiked (1 pg 1-I) 0.5 moll-' NaCl solution; and C back- ground ( 5 PI Kt = 1800 "C) shown in Fig. 7(a). In the purified sea-water a residual 'specific' atomization signal is obtained in the purified sea-water which is delayed as compared with the atomization signal correspond- ing to the addition of 1 pgl-' of Zn; the shape of this signal is very similar to the shape of the background absorption signal.The variation of this residual 'specific' integrated absorbance with the pyrolysis (200-1500 "C) and atomization (850-1800 "C) temperatures is rather small and not analogous to the variation of the integrated absorbance of a spiked sea- water; this signal remains nearly constant as Zn is lost at pyrolysis temperatures of >6OO"C and not atomized at tem- peratures of < 1400 "C. From these observations it appears that this 'specific' atomization signal is not the atomization signal of zinc but a Zeeman interference effect related to the vaporization of the chloride sea-water matrix.This Zeeman interference has also been observed in a 0.5molI-1 NaCl medium [Fig. 7(b)]; a 'specific' atomization signal delayed from the atomization signal of a 1 pg1-l Zn spike being generated during the vaporization of the NaCl matrix. The Zeeman suppressed interference in sea-water appears to be mainly induced by the vaporization of the NaCl salt. In sea- water this Zeeman under-compensation occurring at the end of the Zn atomization signal could be equivalent to theJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 48 1 atomization signal of about 1 pg 1-l of Zn at 1800°C and consequently induces an important systematic error (depending on the electrothermal atomization programme) on the direct determination of zinc at low level concentrations in sea-water.Presence of Oxalic and Nitric Acids The maximum integrated absorbance is reached for an atomiz- ation temperature of 1300 "C in the presence of 0.7 mol I-' oxalic acid (m,=0.7 pg) and of only 1400 "C in the presence of 1 moll-' nitric acid (m = 0.9 pg). The chemical interference effect observed in nitric acid being probably due to the production of oxygen in the furnace related mainly to the vaporization of sodium oxide species simultaneously with zinc oxides. This interference is smaller in the presence of oxalic acid because of the lower atomization temperature of zinc in this more reducing medium. In sea-water modified with nitric acid [Fig. 8(a)] the pres- ence of a significant over-compensation is observed just before the appearance of the Zn atomization signal which induces an important systematic error. This Zeeman spectral inter- ference effect observed also in the case of selenium,16 has been attributed to the Zeeman splitting of the absorption bands of NO molecules produced during the decomposition-reduction step of the different metallic nitrates NaNO Mg(NO,) and Ca(NO,),.Therefore owing to the stabilization of ZnO by the metallic oxides mainly MgO (Fig. 5) this spectral inter- ference can be eliminated during pyrolysis at about 850°C without loss of zinc. In 0.7 moll-' oxalic acid [Fig. 8(b)] no systematic over-compensa tion is detected. For two injected sea-water volumes (5 and 10 pl) calibration graphs are linear for an atomization temperature of 1800°C in a 0.7 moll-' oxalic acid medium (pyrolysis 450 "C 60 s) and a 1 mol I-' nitric acid medium (pyrolysis 850 "C 60 s) and pass through the origin indicating no systematic errors.The concentration of Zn was determined in two certified sea-water reference materials NASS-3 and CASS-2 in the presence of 0.7 moll-' oxalic acid and 1 moll-' nitric acid. The values obtained for NASS-3 and CASS-2 (10 measure- ments) are in a good agreement with the respective certified values (Table 1). For a 10 pl injected volume of sea-water the detection limit obtained under these experimental conditions is about 80 ng 1-' in a nitric acid and 60 ng 1-' in an oxalic acid medium. Conclusion The direct determination of zinc in sea-water at the 1 pgl-' level concentration by ETAAS with Zeeman correction is difficult owing to the high background absorption generated $ AA0.050 2 BG 0.600 (b) B .. 0 5.00 Time/s Fig. 8 Atomization and background absorption signals for zinc in sea-water (10 pl 1 pg 1-1 of Zn T 1800°C (a) in the presence of 1 mol I-' HNO A zinc without pyrolysis B background and C zinc with pyrolysis at 850 "C for 60 s; and (b) in the presence of 0.7 moll- ' oxalic acid A zinc and B background Table 1 reference materials in the presence of nitric and oxalic acids; n = 10 Comparison of zinc concentrations obtained using certified Zn concentration/pg 1-' Certified 1 mol I-' 0.7 mol 1-' Sample value/pg 1-' nitric acid oxalic acid NASS-3 0.178 0.025 0.189 & 0.016 0.16 & 0.020 CASS-2 1.97 k0.12 2.05 & 0.10 2.00*0.10 by the vaporization of the chloride matrix which moreover generates a Zeeman interference effect; the induced under- compensation leading to an important systematic error on the determination of low level concentrations of zinc in unmodified sea-water.It appears also that slow atomization ramps lead to loss of ZnC12 through hydrolysis of MgCl and simultaneous generation of HCl. The different inorganic or organic acids used as modifiers do not separate entirely the atomization signal of zinc and the sea-water background absorption signal. Oxalic and nitric acids which significantly reduce the background absorption signal appear to be the most interesting modifiers. In the presence of 1 moll-' HNO the presence of a Zeeman inter- ference effect was noted due to the Zeeman splitting of the absorption bands of the NO molecules generated during the decomposition-reduction step of metallic nitrates in the graph- ite furnace.This interference effect induces an over- compensation which can be eliminated in sea-water without loss of Zn by selective pyrolysis at about 850°C. Therefore it remains an interference effect (25%) related to the simultaneous vaporization of zinc and sodium oxides. In an oxalic acid medium no significant interference effect or systematic over- or under-compensation has been observed. So oxalic acid appears as the most interesting modifier at very low level Zn concentrations. Owing to its commercial availability in a higher purity grade and its higher solubility HNO can be used instead of oxalic acid for practical reasons.Using these experimental conditions the detection limit (30) obtained for an injection of 10 pl of sea-water is about 80 ng 1-l Zn in a nitric acid medium and 60 ngl-' of Zn in an oxalic acid medium leading to a precision of about 10% at the 1 pg 1 - 1 level concentration. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 References Burrell D. C. and Wood G. G. Anal. Chim. Acta 1969 48 45. Campbell W. C. and Ottaway J. M. Analyst 1977 102 495. Le Bihan A. and Courtot-Coupez J. Analusis 1975 3 59. Sturgeon R. E. Bennan S. S. Desaulniers A. and Russell D. S. Anal. Chem. 1979 51 2364. Guevremont R. Anal. Chem. 1981 53 911. L'vov B. V. and Ryabchuk G. N. Spectrochim. Acta Part B 1982 31 673. L'vov B. V. Spectrochim. Acta Part B 1990 45 633. Berglund M. and Baxter D. C. J. Anal. At. Spectrom. 1992 7 461. Frech W. and Baxter D. C. Spectrochim. Acta Part B 1990 45 867. Shekiro J. M. Skogerboe R. K. and Taylor H. E. Anal. Chem. 1988 60 2578. Erspamer J. P. and Niemczyk T. M. Anal. Chem. 1982,54,538. Kantor T. Bezur L. Pungor E. and Winefordner J. D. Spectrochim. Acta Part B 1983 38 581. Byrne J. P. Chakrabarti C. L. Gregoire D. C. Lamoureux M. and Ly T. J. Anal. At. Spectrom. 1992 7 371. Chaudry M. M. Mouillere D. Ottaway B. J. Littlejohn D. and Whitley J. E. J. Anal. At. Spectrom. 1992 7 701. Kantor T. Pungor E. Sztatisz T. and Bezur L. Talanta 1979 26 357. Le Bihan A. Cabon J. Y. and Elleouet C. Analusis 1992,20,601. Cabon J. Y. and Le Bihan A. Anal. Lett. 1986 19 755. Paper 3/0081 OJ Received February 10 1993 Accepted October 7 1993