JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 483 Indirect Determination of Iodide as an Hgxl Complex by Electrothermal Atomic Absorption Spectrometry* P. Bermejo-Barrera A. Moreda-Pi Aeiro M. Aboal-Somoza J. Moreda-Pi Aeiro and A. Bermejo-Barrera Department of Analytical Chemistry Nutrition and Bromatology Faculty of Chemistry University of Santiago de Compostela 75706-Santiago de Compostela Spain A method for the indirect determination of trace amounts of iodide by electrothermal atomic absorption spectrometry through the measurement of the mercury signal generated when small amounts of iodide and mercury are heated in a graphite furnace is described. The measured absorbances are related to the values of the signals from an iodide-mercury complex. The pH required for the formation of the HgJ complex as well as other parameters involved in measurement of the signals are also determined.The limits of detection and quantification obtained were 3.0 and 10.1 pg I-' of iodide respectively and the characteristic mass was 38.8 pg of iodide. The relative standard deviations obtained were from 5.1 -8.9% (n = 7) depending on concentration. In the range 5-20 pg I-' recoveries were 94.8-104.4O/0. The method has been applied to a range of tap waters. Keywords Indirect method; iodide determination; electrothermal atomic absorption spectrometry The determination of anions by atomic absorption spec- trometry (AAS) has usually been performed by indirect methods. This is due to the fact that these species exhibit their main resonance lines in the vacuum ultraviolet region below 190 nm and therefore they cannot be determined directly with conventional instruments.'.' One indirect method is based on a chemical reaction between the anion to be determined and a metal. A reaction such as this is essential for the performance of the method.The absorbance of the metal that has reacted or remains after reaction is measured and related to the concentration of the anion.3 For iodide some indirect methods based on the formation of a metal-iodide complex have been reported."' For instance the reaction of iodide with cadmium as tris( l,lO-phenanthroline)cadmium(11) generates an ion pair that can be extracted before measurement of the cadmium is perf~rmed.~ Silver reacts with iodide to form a silver-iodide complex which when volatilized into a flame allows the measurement of iodide through the signal from atomic ~ilver.~ Iodide reacts with mercury the excess of which is then sorbed onto a cation exchanger the mercury-iodide complex that remains in the solution being measured in this instance.6 Other methods involve the reaction of mercury-iodide complexes with 2,T-dipyridyl which are extracted by ethyl acetate or isobutyl methyl However these indirect methods are often complicated because they involve extractions filtration etc.and hence the possibility of losing the complexes. Thus many workers have determined iodide through a decrease in the metal absorbance in the presence of the iodide. K~ldevere,~ Wifladt et a/.'' and Sun and Julshamn" have determined iodide by measuring the decrease in the mercury signal when a certain concentration of iodide is added. These workers used the cold vapour technique and the mercury-iodide complexes formed have a structure HgI (n = 1,2,3 .. .). In these methods the molar amounts of mercury and iodide used are 1+10. Therefore a high concentration of iodide relative to that of mercury is required to guarantee formation of the complexes. If the concentration of iodide in the sample is low the very small amounts of mercury that can be added give a very small mercury signal. Nomura and Karasawa" have described an indirect method based on the measurement of the mercury-iodide (HgI,) signal * Presented at the XXVIII Colloquium Spectroscopicum Inter- nationale (CSI) Post-Symposium on Graphite Atomizer Techniques in Analytical Spectroscopy Durham UK July 4-7 1993.generated when amounts of mercury and iodide are heated in a graphite furnace. They obtained two peaks for a mercury absorbance signal and related the first of these to the mercury and the second one to the mercury that was contained in the HgI that had formed. Therefore in their method they con- cluded that the molar ratio of iodide to mercury was two. As pointed out by Nomura and Karasawa if thermodynamic factors are considered for this complex to be formed it would be necessary for the molar ratio of iodide to mercury to be six. The aim of the present work was to develop an indirect method for the determination of iodide in the pgl-' range through the formation of an Hg,I complex by electrothermal atomic absorption spectrometry (ETAAS).Experimental Apparatus The absorbance of mercury was measured with a Perkin-Elmer 1100 B atomic absorption spectrophometer equipped with a deuterium lamp for background correction a graphite furnace atomizer Perkin-Elmer HGA-700 and an autosampler Perkin-Elmer AS-70. The radiation source was a mercury electrodeless discharge lamp (connected to its power supply) operated at 4 W which provided a 253.7 nm line. The band- width was 0.7 nm. Pyrolytic graphite coated graphite tubes and pyrolytic graphite (L'vov) platforms were used throughout ( Perkin-Elmer Uberlingen Germany). Reagents All solutions were prepared from analytical-reagent grade chemicals using ultrapure water resistivity 18 Mi2 cm-' which was obtained by means of a Milli-Q water purification system (Millipore).Potassium iodide stock standard solution 1.000 g I-'. Potassium iodide (Merck Darmstadt Germany) 0.327 g was dissolved in water and diluted to 250ml. This solution was diluted to obtain working standard solutions. Mercury(r1) nitrate stock standard solution 1.000 g 1- '. Panreac Barcelona Spain. Nitric acid. A solution containing 2 x moll-' of nitric acid was prepared from Suprapur acid (69-70.5% with a maximum mercury content of 0.001 pg ml-' BDH Chemicals Poole UK) by appropriate dilution with water and was used withou t s tandarization. Palladium stock standard solution 3.000 mg ml - I . Prepared by dissolving 300 mg of palladium (99.999% Aldrich Chemicals Milwaukee WI USA) in 1 ml of concentrated nitric acid and diluting to 100ml with ultrapure water.If the dissolution was incomplete 10 pl of hydrochloric acid484 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 (Suprapur 35.0% with a maximum mercury content of 0.001 pg ml-' BDH Chemicals) were added to the cold nitric acid and heated to gentle boiling in order to volatilize the excess of chloride. Suljde stock standard solution 0.1 g ml-'. Sodium sulfide monohydrate 98.0% ACS-reagent grade (Aldrich Chemie Steinheim Germany) 0.0749 g was dissolved in water and diluted to 100 ml. This solution was diluted to obtain working standard solutions. Argon. N50 purity argon (99.9990%) was used as a sheath gas for the atomizer and to purge internally. Synthetic air and oxygen. Used as gas phase chemical modifiers C45 purity (99.995 YO).Procedure For calibration 20 pl of standard solution containing iodide concentrations of between 0 and 30 pg 1-' in a medium con- taining 300 pg 1-' of mercury and 1 x lop4 mol I-' of nitric acid were injected into the atomizer. The sequential dry-atomi- zation-clean programme (Table 1) of the graphite furnace was run and the integrated absorbances recorded. Results and Discussion When amounts of mercury and iodide are heated in a graphite furnace two peaks are obtained (Fig. 1). An increase in the second peak and a decrease in the first was observed when higher concentrations of iodide were added. Therefore the first peak is related to mercury and appears at about 60"C and the second one at about 630 "C is related to a mercury-iodide complex.The indirect method described below is based on recording the second peak the one related to a mercury-iodide complex.'2 Table 1 Graphite furnace temperature programme and instrumentation Temperature/ Ar flow/ Step "C Ramp/s Hold/s ml min-l 0 (read) 50 30 20 300 Dry Atomization 900 35 15 Clean 2000 13 2 300 Hg electrodeless discharge lamp Wavelength 253.7 nm EDL power 4 W Spectral bandwidth 0.7 nm Read delay 24 s Integration time 10 s Peak-area measurements D lamp background corrector Pyrolytic graphite tubes and platforms (L'vov) Injection volume 20 pl 0 25.0 Timels 50.0 Fig. 1 Effect of I - on the absorbance signal of Hg (300 pg 1-' of Hg) recorded over 50 s A without I-; By with 20 pg 1-' of I-; and C with 40 pg 1-' of I - Optimization of the Graphite Furnace Temperature Programme The proposed programme was optimized by increasing the atomization temperature until the second peak due to the mercury-iodide complex was observed.This begins to appear at a temperature of 630 "C and the peak is recorded completely if the temperature is increased up to 900°C and remains at that temperature for a few seconds. Thus to reach this temperature a ramp time of 35 s and a hold time of 15 s were needed. If the integration is delayed 24s and the integration time is about 12 s the peak due to the mercury-iodide complex can be recorded (Fig. 2). The gas flow during atomization was also optimized; 0 10 and 20 ml min-' were tried using the temperature programme given above. The result was a substantial decrease in the signal (more than 50%) for a gas flow of lOmlmin-' and for 20 ml min-' the signal fell to zero.As the signal obtained for 0-50 pg 1-' of iodide is low (an integrated absorbance of 0.032 for 30 pg 1-I) the use of a gas flow of 10 ml min-' decreases the signal considerably and thus there is no advantage to be gained. In Table 1 the optimized programme and instrumental conditions are shown. Optimization of Amount of Nitric Acid As reported by several workers an acidic medium is required in order for the complexes between the mercury and iodide to be Nitric acid appears to be the most appr~priate.~ However the nitric acid acts as a chemical modifier on the mercury and thus can delay its atomization. Hence in the proposed method an absorbance signal from uncomplexed mercury could occur at the same temperature as the mercury complexed with iodide. A peak could therefore appear even in absence of iodide.To estimate this modifying effect a reagent blank solution of mercury (300 pg 1-') containing different concentrations of nitric acid to give concentrations of about 1 x and 1 x moll-' nitric acid (which is sufficient to guaran- tee formation of the mercury-iodide complexg) were measured. An increase in the signal for greater amounts of acid is observed (Table 2). The lowest integrated absorbance value is related to a concentration of 1 x moll-'. Therefore a concentration of nitric acid of 1 x mol I-' was chosen for which the integrated absorbance of the mercury was negligible (0.002 Table 2). 1 x 24.0 29.0 Time/s 34.0 Fig.2 Effect of I - on the absorbance signal of Hg (300 pg 1-' of Hg] delayed 24 s and recorded over 10 s A without I-; B with 20 pg 1- of I-; and C with 40 pg 1-' of I- Table 2 Absorbance values for a solution containing 300 pg 1-' of mercury and varying concentrations of nitric acid Nitric acid concentration/ mol 1-1 Integrated absorbance/s 1 x 10-4 0.002 1 x 10-3 0.005 1 x lo- 0.014JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 485 Calibration and Standard Additions Graphs Following the procedure described above the calibration graph was found to be linear over the range 0-30 pg 1-' of iodide. The standard additions method was used over the same range of concentrations for a sample of tap water. The equations obtained were as follows Calibration graph QA=0.015+ 1.14 x 10-3[1-] Standard additions graph QA = 0.0047 + 8.022 x [ I-] where QA is the integrated absorbance in s and [I-] the concentration expressed in pg 1-I.The absorbance of the blank was 0.015 s. Both graphs are shown in Fig. 3. As can be seen from the figure the aqueous calibration graph cannot be used to determine the results and calibration by the standard additions method must be used. Precision and Accuracy To study the repeatability of the measurements at different concentrations four solutions containing 0,5 10 and 20 pg 1-' of iodide were prepared and each solution was measured seven times. The relative standard deviations (RSD) obtained are shown in Table 3. The accuracy of the method was studied through the recov- ery. For iodide concentrations of 5 10 and 2Opgl-' the recoveries were 104.4 94.8 and 101.05% respectively.Sensitivity The sensitivity was studied using three parameters the limit of detection (LOD) the limit of quantification (LOQ) and the characteristic mass (mo) which are defined as follows 3 x S D cs 10 x SD cs LOD=- LOQ=- v x c x 0.0044 mo = QA,-QAb 0 4 8 12 16 20 Concentration of l-/pg I-' Fig. 3 A Calibration and B standard additions graphs Table 3 Repeatability of measurements at different iodide concen- trations n = 7 0 5 10 20 where SD is the standard deviation of the measurements of the blank; CS is the slope of the calibration graph; C is sample concentration expressed in pgl-' of iodide; V is the volume of the sample in pl; and QA and QAb are the integrated absorbances of the sample and blank respectively. The results obtained were LOD = 3.0 pg 1-' of iodide; LOQ=10.1 pgl-' of iodide; and mo=38.3 pg of iodide.Effect of Various Chemical Modifiers Although the method offers a low background signal the possibilities of using chemical modifiers such as palladium sulfide synthetic air and oxygen were studied. The results obtained for palladium and sulfide were unsatisfactory. When small amounts of both species were added the mercury-iodide complex signal increased owing to overlap with the signal from the decomposing mercury-palladium or sulfide species. In Figs. 4 and 5 it can be seen that when palladium or sulfide are present the first peak due to mercury(II) decreases and the second one due to the mercury-iodide complex increases.For a palladium concentration of 50 pg 1-' it can be observed that the first peak fell to zero. When sulfide was added similar results were obtained. If a synthetic air or oxygen flow was used an increased signal was observed and the background was completely removed. In order to employ these gas-phase chemical modi- fiers a step was added to the original graphite furnace tempera- ture programme (Table 1). This new step was ashing at 400°C with a ramp time of 20 s (Table 4). Synthetic air and oxygen flows of 50 100 200 and 300 ml min-' were used during this ashing step. The results are shown in the Table 5. As can be observed when synthetic air was used the mercury-iodide complex signal increases and this signal remained constant for flows greater than 200 ml min-'.For oxygen the increase of the signal was lower than for synthetic air and the signal was constant for flow rates greater than 50 ml min-I. Although the new programme has an ashing step (up to 400"C) it does not affect the temperature ramp from 50 to 900 "C so atomization of the mercury-iodide complex does not differ from the pre- vious programme that is the one given in Table 1. To verify 0 25.0 Time/s 50.0 Fig. 4 Effect of Pd on the absorbance signal of 40 pg 1-' of I- A without Pd; B with 30 pg 1-' of Pd; and C with 50 pg I-' of Pd Iodide concentration/pg 1- ' RSD (%) 8.9 7.0 5.8 5.1 25.0 Time/s 50.0 Fig. 5 Effect of S2- on the absorbance signal of 40 pg I-' of I- A without S2-; B with 5 pg 1-' of S2-; and C with 10 pg 1-' of S2-486 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 Table 4 Graphite furnace temperature programme for gas-phase chemical modifiers (oxygen and synthetic air). Other conditions are as shown in Table 1 Temperature/ Ar flow/ Step "C Ramp/s Hold/s ml min-' 50 30 20 300 -* Dry Ash 400 20 0 Atomization 900 15 15 0 (read) Clean 2000 13 2 300 0.15 A I * 200 ml min-' or 50 ml min-' when synthetic air or oxygen were used respectively. n L I Table 5 Absorbance values obtained for gas-phase chemical modifiers (oxygen and synthetic air). With the original programme (Table l) the value obtained was 0.054 Flow/ml min- ' 0 2 4 6 8 10 12 Concentration of anionhg I - ' Fig. 6 Interfering anions in a solution containing 40 pg 1-' of I- A S2-; B CN-; C S2032-; and D SCN- (see text for details) Gas 0 50 100 200 300 0.056 1 Air 0.051 0.059 0.061 0.071 0.069 Oxygen 0.050 0.057 0.059 0.065 0.061 this a standard of 4Opg1-' of iodide was measured using both programmes but without a flow of synthetic air or oxygen.The values obtained by the simpler programme meas- uring the 40 pg 1-' iodide standard twice as can be seen in the Table 5 do not differ significantly. Interferences The effects of foreign ions on the determination of 40 pg I-' of iodide were studied and the results are shown in Figs. 4 and 5. Several cations including Cd2+ Co2+ Ni2+ Pb2+ and Zn2+ interfere with the measured signal. This is attributed to the reaction with the iodide. However cyanide sulfide thio- cyanate and thiosulfate also interfere in the method but these anions increase the signal because they react with the mer- cury(n) forming complexes that are atomized at the same atomization temperature as the mercury-iodide complex.Halide bromide fluoride and chloride ions also interfere in the measurements. The interfering effects of anions such as cyanide sulfide and thiosulfate are greater than those owing to cations and halide anions. For cations such as Cd2+ Co2+ and Ni2+ the interfer- ing effect begins to be important for concentrations of these cations greater than 500 pg 1-' for Pb2+ the interfering effect is observed for concentrations greater than 100 pg 1-' while for Zn2 + the interfering behaviour appears for concentrations lower than 100 pg 1-'. Bromide interferes at concentrations greater than 250 pg 1-' and fluoride and chloride at concen- trations greater than 1500 pg 1-' and 10 mg 1-' respectively. Anions such as cyanide and thiosulfate begin to interfere at concentrations of about 5 pg I-' and about 1 pg 1-l for sulfide.For thiocyanate no interfering behaviour is observed. The interfering effects of the anions and cations mentioned above are shown in Figs. 6 and 7. In both figures the value of integrated absorbance obtained without any interfering ions is shown in addition to the *lo% interval of that value. This interval was used as a limit to consider whether or not an absorbance value for a particular foreign ion was classed as an interference. To conclude some interferences have been observed but since the concentrations of such ions in the samples being analysed (that is the samples for which the method has been developed) are not that high it can be assumed that for many v) -.2 0.042 0 200 400 600 800 1000 Concentration of cation/pg I-' Fig. 7 Ni2+; B Cd2+; C Pb2+; D Zn2+; and E Co2+ (see text for details) Interfering cations in a solution containing 40 pg 1-' of I- A 'Table 6 Iodide levels measured in several tap waters Sample 1 2 3 4 5 6 7 8 9 10 11 12 13 Iodide concentration/pg 1-1 197.8 44.0 141.3 116.2 195.6 179.2 184.3 271.9 280.3 195.3 806.8 110.4 90.8 practical analytical applications there are no interferences in the method. Applications The method was applied to the determination of iodine in tap waters from several towns of Galicia (north-western Spain). A standard additions graph was used to determine the measure- ments. The iodide concentrations obtained (shown in the Table 6) varied between 44.0 and 806.8 pg 1-' of iodide. References 1 Kirkbright G. F. and Johnson H. N. Talanta 1973 21 433. 2 Manfield J. M. West T. S. and Dagnall R. M. Talanta 1974 21 787.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 487 3 Garcia-Vargas M. Milla M. and PCrez-Bustamante J. A. Analyst 1983 108 1417. 4 Kumamaru T. Bull. Chem. SOC. Jpn. 1969 42 956. 5 Fike R. S. and Frank C. W. Anal. Chem. 1978 50 1446. 6 Chuchalina L. S. Yudelvich I. G. and Chinankova A. A. Zh. Anal. Khim. 1981 36 920. 7 Chakraborty D. and Das A. K. At. Spectrosc. 1988 9 189. 8 Chakraborty D. and Das A. K. Talanta 1989 36 669. 9 Kuldvere A. Analyst 1982 107 1343. 10 Wifladt A. M. Lund W. and Bye R. Talanta 1989 36 395. 11 Sun F. and Julshamn K. Spectrochim. Acta Part B 1987,42,889. 12 Nomura T. and Karasawa I. Anal. Chim. Acta 1981 126 241. Paper 31046736 Received August 3rd 1993 Accepted September 20 1993