Analytica Chimica Acta 439 (2001) 229–238 Optimization of a flow injection hydride generation atomic absorption spectrometric method for the determination of arsenic, antimony and selenium in iron chloride/sulfate-based water treatment chemical Teemu Näykki a,∗ , Paavo Perämäki a , Jyrki Kujala b , Anna Mikkonen b a b Department of Chemistry, University of Oulu, P.O Box 3000, FIN-90014 Oulu, Finland Kemira Chemicals Oy, Oulu Research Centre, P.O Box 171, FIN-90101 Oulu, Finland Received 26 September 2000; received in revised form March 2001; accepted 19 March 2001 Abstract The flow injection hydride generation technique together with atomic absorption spectrometry was used for the determination of arsenic, antimony and selenium in the iron-based water treatment chemical FeClSO4 Thiourea, l-cysteine and potassium iodide–ascorbic acid were used as masking agents to diminish the interference caused by the very high iron concentrations in the samples These reagents act also as prereductants for As(V) and Sb(V) Thiourea and l-cysteine did not prevent the signal depression caused by such high iron content, but potassium iodide–ascorbic acid eliminated iron interference well even up to 2500 mg Fe l−1 The limits of detection (LODS) in aqueous solutions containing no iron were 0.037 g l−1 , 0.121 g l−1 and 0.131 g l−1 for As, Sb and Se, respectively The linear dynamic range was 0–10 g l−1 for As and 0–30 g l−1 for Sb and Se The precision relative standard deviation was expressed as 2.6% for As, 4.4% for Sb and 2.9% for Se The precision determinations were done on the FeClSO4 matrix at the level 0.5–0.8 g l−1 for the elements to be analyzed The accuracies of the methods were tested by using two standard reference materials (SRM 361, LA Steel and SRM 2074, river sediment) The concentrations obtained for As, Sb and Se were very close to the certified values © 2001 Elsevier Science B.V All rights reserved Keywords: Atomic absorption spectrometry; Hydride generation; Flow injection; Arsenic; Antimony; Selenium; Iron interference; Water treatment chemical Introduction It is commonly known that the transition metals interfere in the determination of hydride-forming elements when the hydride generation (HG) technique is used [1,2] Different mechanisms have been suggested ∗ Corresponding author Present address: Finnish Environment Institute (FEI), Research Laboratory, Hakuninmaantie 4-6, FIN-00430 Helsinki, Finland Fax: +358-9-4030-0890 E-mail address: teemu.naykki@vyh.fi (T Näykki) for the interference effects observed [3,4] The predominant mechanism is probably due to the reaction of the interfering transition metal ions with the NaBH4 reductant, and the precipitate which is formed is able to capture and catalytically decompose the evolved hydrides [3] For instance, Bax et al [4] and Bye [5] have stated that the precipitates are probably not elemental metals, but rather metal borides Lugowska and Brindle [6] investigated the redox processes occurring in transition metal solutions during reduction by NaBH4 Also they found, boron(III) as boride-like 0003-2670/01/$ – see front matter © 2001 Elsevier Science B.V All rights reserved PII: S 0 - ( ) 0 - 230 T Näykki et al / Analytica Chimica Acta 439 (2001) 229–238 species causing the suppression of the signal in hydride generation [6] Iron is a common transition metal that is present at high concentrations in many type of samples However, only a few studies have been published where the iron concentrations of the samples are as high as in this work Narsito et al [7] found that the interference caused by 200 mg l−1 iron was removed when 1% thiourea was used as a prereductant Wickström et al [8] studied hydride generation and the complexation reactions in an alkaline sample solution The hydrides formed were evolved by subsequent acidification of the sample solution They found that 0.3 mol l−1 tartrate completely removed iron interference at the 500 mg Fe l−1 level and when the iron concentration was 5000 mg l−1 , the signal was depressed by only 15% However, dissociation of the metal complex can occur during the acidification of the sample Thus, the formation of free metal ions and precipitation is possible Boampong et al [9] used 3% l-cystine in mol l−1 hydrochloric acid as a masking agent to prevent the interferences of very high concentrations of iron Welz and Sucmanova [10] proposed the use of l-cysteine instead of l-cystine as a masking agent Earlier Chen et al [11] had observed that l-cysteine removed iron interference up to 1000 mg l−1 even though the solution became turbid due to the formation of l-cystine when the iron(III) concentration was greater than 200 mg l−1 Generally l-cysteine has been proved to be very useful for preventing iron interferences, but nearly all previous publications have focused on iron concentrations below 1000 mg l−1 Finally, it is worth noting that iron(III) has also been used for minimizing interference effects of the other transition metals in the HG technique The high positive potential of the reduction of iron(III) to iron(II) suggests a preferential reduction of this species More seriously interfering metal ions, such nickel(II) and copper(II), will be reduced to the metals and precipitated only after all the iron has been reduced [12,13] The extent of the transition metal interference depends very much on the HG system used Severe interference was observed when we earlier used the batch type hydride generation system described by Siemer and Hagemann [14] The flow injection (FI) technique is less prone to transition metal interference [10,15] There are at least two main reasons for this observation When using the FI system instead of the batch system, the concentration of the reductant is usually lower and formation of the interfering precipitates, e.g borides, is decreased Another reason can be called kinetic discrimination The reduction of the hydride-forming elements is fast and the reaction is completed before the reduction of the transition metal ion to the interfering species Also the separation of the hydrides from the sample matrix is very fast in a gas–liquid separator In this study, the parameters for the determination of As, Sb and Se from the PIX-110TM chemical (FeClSO4 ) using flow-injection hydride generation (FI-HG) were optimized The PIX-110TM water coagulant is a FeClSO4 solution manufactured by Kemira Chemicals Oy, containing approximately 12.0 wt.% (180 g l−1 ) of Fe3+ , 22.3% of SO4 2− and 7.5% of Cl− Since As, Sb and Se are toxic, the European Committee for Standardization (CEN) has laid down the maximum limits for these elements and therefore it is essential to monitor their low amounts in water treatment chemicals [16] Experimental 2.1 Instrumentation A Perkin-Elmer model 5100 Zeeman atomic absorption spectrometer equipped with a Perkin-Elmer model FIAS-400 flow injection system and an AS90 autosampler, controlled by Perkin-Elmer AA WinLab version 2.61 software, was used for the measurements Silicone pump tubes (NaBH4 : 1.14 mm i.d.; carrier HCl: 1.52 mm i.d.; sample loading and waste: 3.17 mm i.d.) were used throughout this study, and all other tubing was mm i.d PTFE An electrically heated quartz tube was used as an atomizer Electrodeless discharge lamps, operated from an external power supply (Perkin-Elmer EDL system 2) were used for the measurements The instrumental parameters are shown in Tables and 2; they were mainly chosen according to the manufacturer’s recommendations 2.2 Reagents and standard solutions All reagents were of analytical-reagent grade unless otherwise stated Ultrapure water (18 M cm−1 ), T Näykki et al / Analytica Chimica Acta 439 (2001) 229–238 Table AA spectrometer settings used for the hydride generationAAS-measurements Wavelength (nm) Lamp current (mA) Measured signal Measuring time (s) Slit width (nm) Quartz cell temperature (◦ C) Quartz cell length/i.d (mm/mm) Reaction coil length (mm) Sample volume (l) Background correction As: 193.7 Sb: 217.6 Se: 196.0 As and Sb: 400 Se: 280 Integrated absorbance As: 15 Sb and Se: 18 0.7 900 180/7.5 300 500 No prepared with a USF Elga Maxima purification system, was used throughout All glassware and plastic containers were soaked in (1 + 1) nitric acid overnight and rinsed five times with de-ionized water and five times with ultrapure water prior to use The working standard solutions were prepared daily by diluting the 1000 mg l−1 standard stock solution of As(V) (Merck) and the 4000 mg l−1 Sb(III) solution (Merck) To prepare the selenium standard stock solution (1000 mg l−1 ) appropriate amounts of selenium salts NaHSeO3 (Merck, for Se(IV)) or Na2 SeO4 ·10H2 O (Merck, for Se(VI)) were dissolved in ultrapure water A few drops of concentrated hydrochloric acid were added before dilution The iron stock solution was prepared from PIX-110TM , because it was observed to contain less arsenic than the commercial iron(III) chloride standard The 20,000 mg l−1 stock solution of Fe(III) was prepared by diluting 16,5289 g (1000 g ≈ 0.667 ml) of PIX-110TM solution (purity not certified) to 100 ml with 4.7 mol l−1 HCl Table FIAS-400 flow injection program generation-AAS-measurements used for the hydride Flow injection program Step Time (s) Speed of pump (rpm) Speed of pump (rpm) Prefill Fill Injection 15 10 15 100 100 120 120 120 231 Sodium tetrahydroborate solution (0.1–0.3% (w/v)) was prepared daily by dissolving the appropriate amount of NaBH4 powder (Fluka) in ultrapure water The solution was stabilized with NaOH (Merck) Hydrochloric acid (0.03–4.7 mol l−1 , Merck) was used in samples and as a carrier solution Different concentrations of HCl were used in the optimization of the operating conditions The stock solution of potassium iodide was prepared by dissolving g of potassium iodide (Merck) in 100 ml of ultrapure water The ascorbic acid stock solution was prepared by dissolving g of l(+)-ascorbic acid (Merck) in 100 ml of ultrapure water Thiourea stock solution (2.0 mol l−1 ) was prepared by dissolving 38 g of thiourea, (J.T Baker, purity > 99%) in 250 ml of 1.5–4.7 mol l−1 HCl l-Cysteine stock solution was prepared by dissolving g of l-cysteine (Merck, for biochemistry, purity > 99%) in 100 ml of 0.03–0.09 mol l−1 HCl Different concentrations of l-cysteine and/or HCl were used to optimize the operating conditions The standard reference materials (SRMs) 361 Low-alloy Steel and 2704 Buffalo River Sediment were obtained from the National Institute of Standards and Technology (NIST, Gaithersburg, MD) The samples were prepared by dissolving g of the SRMS in 30 ml of aqua regia (HCl:HNO3 , 3:1 (v/v)) 2.3 Procedure The FI system used consisted of two peristaltic pumps: one was used for pumping the carrier and reductant solutions (HCl and NaBH4 , respectively) and the other for pumping the sample solution (Fig 1) The carrier flow rate was ml−1 and the reductant flow rate was ml−1 The 500 l sample loop was used in all measurements The optimal instrumental parameters for As, Sb and Se are given in Tables and Potassium iodide (0.5% (w/v)) and ascorbic acid (1.0% (w/v)) were added as prereductants for arsenic and antimony prior to analysis The prereduction time used was 15 Selenium(VI) was reduced to selenium(IV) by heating the sample in a beaker with HCl (4.7 mol l−1 ) at 90◦ C for 20 The standard addition method was used when PIX-110TM was analyzed Dilutions of 1–225 ml (As) or 1–72 ml (Sb and Se) were made before analysis 232 T Näykki et al / Analytica Chimica Acta 439 (2001) 229–238 Fig Flow schematic of FIAS-400 Iron was added to calibration standards to match the iron content in the certified reference materials when the accuracy of the method was tested A more detailed description of the method optimization and validation is given in the next section Table Optimal experimental conditions for the measurement of As and Sba Parameter Optimal parameter values (mol l−1 ) HCl (carrier and samples) Carrier gas (Ar) flow rate (ml min−1 ) As Sb 4.7 ∼65 1.5 ∼100 a 0.5% (w/v) KI, 1.0% (w/v) ascorbic acid, 0.3% (w/v) NaBH in 0.05% (w/v) NaOH Table Optimal experimental condition for the measurement of Se Parameter Optimal parameter values HCl (carrier and samples) (mol l−1 ) Carrier gas (Ar) velocity (ml min−1 ) Prereduction temperature (◦ C) Prereduction time (min) NaBH4 4.7 ∼65 90 20 0.1% (w/v) in 0.05% (w/v) NaOH Results and discussion 3.1 Selection of the prereduction agent 3.1.1 Arsenic and antimony At the beginning of the study, the instrumental parameters were optimized using only aqueous As standards The same parameters were used for Sb because of the similar chemical nature of As and Sb Optimum levels for carrier gas flow, reductant concentration, carrier-HCl concentration and the concentrations of the prereductants were investigated using two-level factorial designs Measurements were performed without an interfering iron matrix The results were processed with Modde for Windows version 3.0 software (Umetri AB) [17] Later, iron was added to the standard solutions and the optimization was repeated using the parameters obtained from the previous stage The optimal concentrations of the various prereductants were studied using iron contents up to 5000 mg l−1 When optimizing the prereduction conditions, the sample matrix contained 1500 mg l−1 of iron The optimal concentration of the ascorbic acid was investigated first and then kept at a constant level (0.8% (w/v)) when optimizing the KI concentration (Fig 2A) When the sample matrix contained 5000 mg l−1 iron 1.0% (w/v) ascorbic acid gave the T Näykki et al / Analytica Chimica Acta 439 (2001) 229–238 233 Fig The effect of KI and ascorbic acid concentrations in the determination of As, when the iron concentration of the samples was: (A) 1500 mg l−1 and (B) 5000 mg l−1 Arsenic concentration in the samples was g l−1 HCl concentration of samples and carrier was 4.7 mol l−1 maximum sensitivity The concentration of ascorbic acid was kept at a constant level and the optimal KI concentration was found to be 0.8% (w/v) (Fig 2B) The sensitivity obtained with thiourea and l-cysteine was worse than that with KI Using an iron concentration,