- Email: [email protected]
Voltammetric determination of Se(IV) and Se(VI) in saline samples—Studies with seawater, hydrothermal and hemodialysis fluids
Analytica Chimica Acta 648 (2009) 162–166
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Voltammetric determination of Se(IV) and Se(VI) in saline samples—Studies with seawater, hydrothermal and hemodialysis fluids Paulo C. do Nascimento a,∗ , Cristiane L. Jost a , Leandro M. de Carvalho a , Denise Bohrer a , Andrea Koschinsky b a b
Departamento de Química, Universidade Federal de Santa Maria, C.P. 5051, 97105-970 Santa Maria, RS, Brazil School of Engineering and Science, Geosciences and Astrophysics, Jacobs University Bremen GmbH, P.O. Box 750561, D-28725 Bremen, Germany
a r t i c l e
i n f o
Article history: Received 30 January 2009 Received in revised form 17 April 2009 Accepted 23 June 2009 Available online 30 June 2009 Keywords: Selenium Ionic strength Speciation analysis Cathodic stripping voltammetry UV-irradiation
a b s t r a c t Determination of Se(IV) and Se(VI) in high saline media was investigated by cathodic stripping voltammetry (CSV). The voltammetric method was applied to assay selenium in seawater, hydrothermal and hemodialysis fluids. The influence of ionic strength on selenium determination is discussed. The CSV method was based on the co-electrodeposition of Se(IV) with Cu(II) ions and Se(VI) determined by difference after sample UV-irradiation for photolytic selenium reduction. UV-irradiation was also used as sample pre-treatment for organic matter decomposition. Detection limit of 0.030 g L−1 (240 s deposition time) and relative standard deviation (RSD) of 6.19% (n = 5) for 5.0 g L−1 of Se(IV) were calculated. Linear calibration range for selenium was observed from 1.0 to 100.0 g L−1 . Concerning the pre-treatment step, best results were obtained by using 60 min UV-irradiation interval in H2 O2 /HCl medium. Se(VI) was reduced to the Se(IV) electroactive species with recoveries between 91.7% and 112.9%. Interferents were also investigated. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Selenium at low concentrations is known as an essential trace element for animals, however at higher concentrations, toxic effects can be observed. The range between the necessary and the toxic intake is narrow [1]. Additionally, there is a growing interest in the differentiation of selenium species in environmental matrices due to the well-known dependence of biological and toxicological effects on the chemical form. The so called ‘speciation analysis’ has gained attention during the past years. Some studies concerning the determination of selenium species have been published to understand its role in the environment and human health, as well [2–7]. Separation techniques, such as high performance liquid chromatography (HPLC) and capillary electrophoresis (CE) with inductively coupled plasma – mass spectrometry (ICP-MS) have been employed for selenium speciation studies [8–11]. The use of optical methods such as graphite furnace atomic absorption spectrometry (GF-AAS) with a complementary co-precipitation step has also been reported [12]. Concerning speciation of selenium in saline matrices, voltammetry can present better performances than other techniques since good sensitivity and selectivity can be reached in
∗ Corresponding author. Tel.: +55 55 3220 8870; fax: +55 55 3220 8870. E-mail address: [email protected] (P.C. do Nascimento). 0003-2670/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2009.06.057
such media. Additionally, low capital investment is required, and the equipment can be portable to be used in onboard-ship analysis [13]. Among the voltammetric techniques used for selenium determination, stripping analysis are uniquely suited for speciation because of their sensitivity to low metal concentrations and to the chemical forms of metals in solution, as well [14]. In alternative to mercury working electrodes, the use of bismuth-film and silver disk has been reported, as well [15,16]. With respect to selenium speciation, acidic electrolytes as HCl and some specific reagents such as copper have already been successfully used [17–23]. Copper ions play an important role for selenium determinations since the formation of an insoluble layer of Cu2 Se as a co-electrolysis step not only reduces the interferences from metal ions by forming stable selenides but also increases the sensitivity of the CSV method [2–4,18,24]. As selenite (Se4+ ) is the selenium species determined by CSV and selenium is usually found as Se(VI) or Se(IV), selective methods are required to convert it into selenite before the electrochemical determination. A well-established procedure for selenium reduction is the use of photochemical pre-treatment by UV-irradiation [2,3,5,25,26]. Furthermore, in samples with low levels of dissolved organic matter (DOM), the interfering organic material is removed with UV-irradiation and the selenium bound in DOM is transformed to inorganic species. Chemical reduction and heating with HCl (4–6 mol L−1 ) have been proposed, as well [27–29]. However, the use of drastic acidic conditions and sample heating is a time consuming procedure and may result in a
P.C. do Nascimento et al. / Analytica Chimica Acta 648 (2009) 162–166
selenium conversion to the metallic state associated to losses of analyte by volatilization [27] even when sample digestion is microwave assisted [6,30]. The purpose of this study was to optimize the CSV in the presence of added copper for selenium speciation concerning the ionic strength influence in order to assay the analytes in low and high saline matrices by using UV-irradiation as pre-treatment step. 2. Experimental 2.1. Instrumentation and apparatus 2.1.1. Voltammetry CSV was applied in ca. 0.1 mol L−1 HCl medium, with the addition of 1.0 mg L−1 Cu(II), according to previous published studies [4,6]. The electrodeposition step was performed under stirring (400 rpm) by using the adsorption potential of −300 mV for all media (60 s for 1.0–20.0 g L−1 and 10 s for 20.0–100.0 g L−1 ). After an equilibration time of 10 s, the differential pulse voltammetric scan was recorded from −300 to −750 mV (60 mV s−1 scan rate; −50 mV pulse amplitude) – the stripping peak occurs at ca. −610 mV. Voltammograms were collected by using a Metrohm 693 VA Processor with a 694 VA Stand operating with hanging mercury drop electrode (HMDE). A platinum wire was taken as the auxiliary electrode and all potentials were quoted against Ag/AgCl/KCl (3 mol L−1 ) reference electrode. The cell volume was 10 mL. The standard addition method was used to evaluate the analyte concentrations through the measurement of the peak height. 2.1.2. UV-irradiation The UV-reduction step was performed by irradiating 10 mL of the samples spiked with a known amount of Se(VI) and/or Se(IV) for 60 min at 85 ± 5 ◦ C (Hg-high pressure lamp, 500 W, Metrohm, Herisau, Switzerland) in order to test the method for total dissolved selenium determination. Each UV-irradiation was performed in 10 mL sample volume made of 0.1 mol L−1 HCl and added by 10 L of destilled H2 O2 30% (v/v), according to procedure published by Mattsson et al. [4]. 2.2. Reagents All chemicals were of analytical-reagent grade. The water used throughout was distilled, deionized and further purified by a MilliQ high purity water device (Millipore, Bedford, USA). The 1.0 mg L−1 Se(IV) standard reference material 3149 (batch F707118, National Institute of Standards and Technology (NIST), USA) was purchased from SpecSol (Brazil). The 1.0 g L−1 Se(VI) standard stock solution was prepared by dissolving 0.2392 g of sodium selenate (Fluka, Germany) in 100 mL 0.5 mol L−1 HNO3 instead of high purity water to help the maintenance of all selenium on the (VI) oxidation state. The solutions were stored in opaque PTFE flasks to protect against sunlight. Aliquots of these stock solutions were taken to prepare the working solutions. 100 mg L−1 Cu(II) aqueous solution was prepared from the chloride salt (Merck) in order to form the selenide complex while a 0.1 mol L−1 HCl solution was used as supporting electrolyte. The used hydrogen peroxide (30%, v/v) was purchased from Vetec (Brazil). Humic acid (Aldrich, Germany) was used to prepare 10.0 mg L−1 aqueous solutions.
Table 1 Test solutions matching the composition of real samples. SW (mol L−1 )
Species +
Na K+ Ca2+ Mg2+ Cl− SO4 2− HCO3 − CH3 COO−
0.048 0.010 0.010 0.054 0.56 0.028 0.002 –
The experiments were performed by using test solutions spiked with selenium having similar saline composition with seawater (SW), hydrothermal (HT) and hemodialysis (HD) fluids (Table 1). For DOM photo-oxidation experiments, 10.0 mg L−1 humic acid was added to SW or HT fluid samples. Non-saline (NS) test solutions
HT (mol L−1 ) 0.05–0.12 0.010 0.010–0.050 0–0.054 0.05–1.2 – 0.005–0.015 –
HD (mol L−1 ) 3.61 0.050–0.070 0.061 0.017–0.030 3.82 – – 0.105
SW: seawater; HT: hydrothermal fluid; HD: hemodialysis fluid.
were also prepared by adding selenium to pure water. The electrolyte (1.0 mg L−1 Cu(II) and 0.1 mol L−1 HCl) was added to all test solutions. Synthetic SW sample prepared according to IAPSO (International Association for the Physical Sciences of the Oceans) have the composition: As 1.7 g kg−1 ; Al 0.54 g kg−1 ; B 4.5 mg kg−1 ; Ba 14 g kg−1 ; C 27.6 mg kg−1 ; Ca 412 mg kg−1 ; Cd 0.08 g kg−1 ; Cl− 19,354 mg kg−1 ; Cu 0.25 g kg−1 ; Fe 0.055 g kg−1 ; K 399 mg kg−1 ; Mg 1290 mg kg−1 ; Mn 0.014 g kg−1 ; Na 10,770 mg kg−1 ; Ni 0.5 g kg−1 ; P 70 g kg−1 ; Pb 0.002 g kg−1 ; S 904 mg kg−1 ; Si 2.8 mg kg−1 ; U 3.3 g kg−1 ; Zn 0.4 g kg−1 . 2.4. Samples SW samples were collected in the vicinity of Helgoland Island, Germany, filtered through a 0.45 m membrane filter (Millipore, USA), acidified to pH 2.0 with HCl (37%, w/v) and then stored refrigerated in polyethylene bottles. A pool of HT fluid was made with samples from the Logatchev field (15◦ N, 45◦ W) and stored refrigerated without pre-treatment. HD fluid samples were taken directly from the original container without pre-treatment or dilution (HB Bic Salbego, Brazil). 3. Results and discussion The conventional CSV of selenium is based on the accumulation of HgSe at HMDE in acidic solutions [31–33]. The selenium peak obtained during the cathodic potential scan can also be attributed to the reduction of slightly soluble Se(IV) species like HgSeO3 and Hg3 (HSeO3 )2 (SeO3 )2 formed with mercury ions on the electrode surface [34]. However, in the presence of high chloride concentrations, calomel is also formed and can be responsible for some conversion of Se(IV) into Se(VI) and Se(0) together with an oxidation of Hg(I) to Hg(II) [35]. This way, the CSV method based on the co-electrodeposition of Se(IV) with Cu(II) ions [2–4,18,24] was more convenient considering the chloride concentration of the investigated samples. Additionally, the stripping peaks of accumulated Cu2 Se species on the HMDE electrode yields, in comparision with convetional CSV selenium determination, to better detection limits, higher linear range and low interference of metal ions [4]. The copper selenide stripping peak depends on the accumulation of Cu2 Se intermediate species at HMDE (Eq. (1)) with subsequent reduction during the cathodic scan (Eq. (2)). H2 SeO3 + 2Cu2+ + 4H+ + 8e− → Cu2 Se + 3H2 O +
2.3. Test solutions
163
−
Cu2 Se + 2H + 2e → H2 Se + 2Cu(Hg)
(1) (2)
The reaction pathway indicates that the reduction of Se(IV) to Se(−II) is pH-dependent. Previous results showed that both HCl and HClO4 were appropriate electrolytes. The best peak shapes were obtained at pH values between 1.6 and 2.0. Additionally, with increasing concentration of Cu(II) ions the height of Se(IV) stripping peaks is affected [34]. The effect of copper ions was inves-
164
P.C. do Nascimento et al. / Analytica Chimica Acta 648 (2009) 162–166
Fig. 1. Effect of deposition potential on the CSV peak current for 10.0 g L−1 Se(IV). Deposition time: 10 s. NS: non-saline sample; SW: seawater; HT: hydrothermal fluid; HD: hemodialysis fluid.
tigated for both SW and non-saline media. The 0.5–10.0 mg L−1 Cu(II) range was evaluated for the determination of 10.0 g L−1 Se(IV). In the present work, only for Cu(II) concentrations higher than 1.0 mg L−1 , the voltammetric Se(IV) signal was strictly a linear function of Se(IV) concentration and independent of small variations on Cu(II) ion concentration. The ionic strengths of the assayed samples (see below) are high and cover a wide variation range mainly due to their high chloride concentrations. A dependence between the stripping potential shift and the ionic strength of the investigated saline samples was observed. Comparing the CSV data obtained by measuring the test solutions NS ( = 0.10 mol L−1 ), SW ( = 0.70 mol L−1 ), HT ( = 0.75 mol L−1 ) and HD ( = 4.0 mol L−1 ), the selenium peak potential shifted from ca. −630 mV (NS) to −530 mV (HD). By increasing the ionic strength, the water activity increases and the hydration number of the metallic species decreases, shifting their reduction potential to more positive values [36,37]. This way, the Se(IV) peak potential depends in some extension on the composition and concentration of the saline sample. Conversely, in high chloride concentrations as existing in SW, HT and HD samples, a negative potential shift could be expected due to metal-chloro complex formation and mercury electrode surface competition between the analyte and those compounds [38]. However, this behavior was not observed for selenium measurements after the co-electrodeposition with Cu(II) ions. This fact may be related to the favorable formation of the slightly soluble Cu2 Se (Ks 3.8 × 10−9 ) monolayer on the electrode surface. 3.1. Effect of ionic strength The test solutions matching the composition of real samples (Table 1) were used to evaluate the voltammetric behavior of selenium after the co-electrodeposition of 10.0 g L−1 Se(IV) with Cu(II) ions. The influence of the ionic strength on the peak current of Se(IV) was investigated considering the deposition potential and time as well as the potential scan rate. The influence of DOM existing in SW and HT fluid was not considered here since its elimination was further investigated by photo-oxidation as cleanup step. 3.1.1. Ionic strength and deposition potential Variations on the deposition potential revealed that the Se (IV) peak height measured by CSV at −610 mV for a 10 g L−1 Se(IV) solution showed its highest values for deposition potentials ranging from −200 to −400 mV (Fig. 1) considering the whole ionic strength interval covered by the assayed samples (test solutions and real samples). The lowest peak currents were obtained for HD samples ( = 4.0 mol L−1 ) and no significant differences were observed among the other samples. Indeed, the stripping peaks remained almost the same for NS, SW and HT test solutions with no spe-
cific influence of ionic strengths ranging from 0.1 to 0.75 mol L−1 . This result suggests that the process responsible for the analyte deposition at the electrode surface is predominantly electrostatic. It is obvious that an deposition potential around −300 mV should be chosen for all samples. Regardless of these results, some positive deposition potentials were also tested because selenium was determined as copper complex in samples covering a very wide chloride concentration range. For electrolytes containing chloride ions, copper occurs predominantly as Cu(I) at deposition potentials ranging from ca. 0.2 to −0.15 V due to complex stabilization, although for positive deposition potentials, incomplete formation of Cu2 Se occurs since its reduction potential lies close to −50 mV [2]. On the other hand, decreases on peak height observed for deposition potentials more negative than −400 mV can be related to the charging process on mercury electrode surface, which changes from positive to negative in chloride medium, at potentials lower than ca. −500 mV [4]. 3.1.2. Ionic strength and deposition time The influence of the ionic strength on the peak current for selenium measurements by the CSV method was evaluated for deposition times ranging from 0 to 200 s. Two different profiles could be observed (Fig. 2). For the lowest ionic strength (0.10 mol L−1 ) the peak current increased slightly with increasing deposition time up to 100 s and more intensely for higher deposition times in the range of 100–140 s. This roughly S-shaped curve can be related to mercury electrode surface saturation [2] that seems to occur only for the lowest ionic strength. This behavior agrees with a predominantly electrostatic electrode process. The higher the ionic strength, the lower the analyte accumulation at the electrode surface as well as the sensitivity. In the present work, deposition time of 10 s was adequate to assay selenium in all samples in the concentration range of 20–100 g L−1 , and 60 s for the range of 1–20 g L−1 . 3.1.3. Ionic strength and scan rate Among the tested parameters, the voltammetric selenium peak showed not to be severely influenced by ionic strength with respect to the scan rate. It was observed that the height of selenium reduction peak increases linearly with the scan rate in the range of 4–120 mV s−1 (Fig. 3) with slopes fairly independent on the ionic strengths. This linearity suggests that Cu2 Se is deposited as a monolayer on the mercury electrode surface since a multilayered precipitate would exhibit a non-linear behavior as a result of stronger kinetic effects related to a possible heterogeneous nature of the electrode process. Since the peaks were well-shaped by using a 60 mV s−1 scan rate, this value was further considered as optimal value for all media.
Fig. 2. Effect of deposition time on the CSV peak current for 10.0 g L−1 Se(IV). Deposition potential: −300 mV. NS: non-saline sample; SW: seawater; HT: hydrothermal fluid; HD: hemodialysis fluid.
P.C. do Nascimento et al. / Analytica Chimica Acta 648 (2009) 162–166
165
3.2. Selenium speciation Since Se(IV) is the voltammetrically active species, the use of a pre-treatment step to reduce the existing Se(VI) was required. UV-irradiation is a well-known and useful tool for photolytic conversions, which was used here. The reduction efficiency was checked by adding Se(VI) to SW samples where no Se(IV) was previously detected by CSV. The selenium speciation was achieved by the difference between the selenium concentration in irradiated and non-irradiated samples.
Fig. 3. Effect of scan rate on the CSV peak for 10.0 g L−1 Se(IV). Deposition potential: −300 mV; deposition time: 10 s. NS: non-saline sample; SW: seawater; HT: hydrothermal fluid; HD: hemodialysis fluid.
3.1.4. Calibrations Linear peak current responses were obtained from 1.0 to 100 g L−1 Se(IV). For concentrations up to 20.0 g L−1 , 60 s deposition time was necessary, while for the range of 20.0–100.0 g L−1 , a shorter deposition time (10 s) was already sufficient to produce well-shaped voltammetric peaks. This way, two different calibration plots could be used depending on the selenium level in the samples. Regression analysis showed for all test solutions r2 ≥ 0.996 and the following calibration functions for the 1.0–20.0 g L−1 concentration range (60 s deposition time): NS: y = 1.297x + 0.562; SW: y = 1.461x + 0.514; HT: y = 1.528x + 0.376; HD: y = 1.160x + 0.100, where x and y are concentration in g L−1 and current in nA, respectively. Similarly, for more concentrated Se(IV) test solutions and using 10 s deposition time the calibration functions NS: y = 0.396x + 0.795; SW: y = 0.389x + 0.187; HT: y = 0.406x + 0.003; HD: y = 0.307x + 0.185 were obtained. The relative standard deviation (RSD) was 6.19% for five independent measurements of 5.0 g L−1 Se(IV) solution. A detection limit of 0.030 g L−1 (240 s deposition time) was calculated considering the condition (3 ± BL ), where is the standard deviation (n = 10) of the background current (BL ) at −610 mV. Typical voltammograms of 7.5 g L−1 Se(IV) spiked SW, HT and HD samples are shown in Fig. 4.
Fig. 4. Voltammograms of 7.5 g L−1 selenium spiked samples. Deposition time: 60 s; deposition potential: −300 mV. I: hemodialysis fluid; II: hydrothermal fluid; III: seawater.
3.2.1. Selenium photolytic conversion The selenium reduction by the action of UV-irradiation was investigated in alkaline and acidic media in the test solutions using NS samples as control. UV-irradiation was performed after the addition of 5.0 g L−1 Se(VI) to SW and HT fluid test solutions. The use of alkaline conditions (NaOH 0.4 mol L−1 , pH 11.0) was previously reported for selenium speciation in SW [3]. By using this condition, the Se(IV) recovery lay around 115.0% for control samples. However, no analytic peak was observed for SW and HT fluid test solutions for irradiation interval up to 90 min. For a longer irradiation interval (180 min), recoveries about 88.0% for SW and 105.0% for HT fluid were calculated. For UV-irradiation in acidic media, it was showed elsewhere that the best results for selenium conversion were obtained with HCl compared to H2 SO4 and HClO4 [2–6,26,27]. Additionally, HCl was also employed in the present work as electrolyte for the voltammetric determinations. The use of 0.1 mol L−1 HCl with 60 min irradiation interval, reported by Mattsson et al. [4], was also used here for the selenium photolytic conversion. In fact, by using this medium, Se(IV) recoveries ranging from 91.1% to 98.4% were obtained. 3.2.2. Samples containing DOM Natural ligands present in the investigated samples (SW and HT fluids) can interfere with the selenium determination by voltammetry; therefore they must be inactivated before the analyses. To assay selenium in SW and HT fluids, UV-irradiation was used to overcome DOM interferences, besides helping on the selenium photolytic conversion process. Humic acid was added to test solutions to simulate DOM content of real SW and HT fluids. Recovery experiments using the selected experimental conditions were also carried out in real samples spiked with selenium. Among the tested cleanup/photolytic conversion conditions in test solutions containing humic acid (10.0 mg L−1 ), 60 min of UV-irradiation in presence of HCl/H2 O2 solutions produced good recoveries. By spiking SW and HT fluid test solutions at 5.0 g L−1 Se(VI) level, recoveries ranging from 104.6% to 113.8% were obtained for samples containing or not DOM. By using 30 or 90 min irradiation intervals, recoveries about of 85 and 125% were obtained, respectively. These results could be explained by either a redox equilibrium established between Se(IV) and Se(VI) species during the irradiation, or by a steady state created by opposing reduction and oxidation reactions [39]. In addition, photolytic conversion of Se(VI) can also be induced by direct absorption of wavelengths under 230 nm or from the action of carbon-containing radicals, generated by the photochemical degradation of natural DOM, which has reducing properties. On the other hand, highly oxidizing chlorine and hydroxyl radical, which are formed by the action of UV light with wavelengths <200 nm, may also be important, especially in the observed reoxidation of Se(IV) after long irradiation intervals. Perhaps these oxidizing radicals react with DOM until it is totally consumed. Then, they start to oxidize Se(IV). The adopted irradiation scheme consisted of DOM oxidation assisted by the addition of 30 L H2 O2 (30%, v/v) to the 0.1 mol L−1 HCl medium. An efficient photolytic degradation of DOM, preceding its complete photolytic conversion to carbon dioxide and water,
166
P.C. do Nascimento et al. / Analytica Chimica Acta 648 (2009) 162–166
Table 2 Determination and recovery for selenium in saline samples by cathodic stripping voltammetry. Sample
Se (g L−1 ) added
Se(IV) (g L−1 ) detected
Recovery (%)
HD
– 5.00 as Se(IV) 5.00 Se(IV) + 2.50 Se(VI) 2.00 as Se(IV)
– 99.80 91.73 110.00
HT + UVa
– 5.00 as Se(IV) 5.00 Se(IV) + 2.50 Se(VI) 2.00 as Se(IV)
– 110.20 112.13 105.50
SW + UVa
– 5.00 as Se(IV) 5.00 Se(IV) + 2.50 Se(VI) 2.00 as Se(IV)
– 99.20 112.93 100.50
SWb + UVa
– 5.00 as Se(IV) 5.00 Se(IV) + 2.50 Se(VI)
– 104.20 97.87
4. Conclusions The CSV method was investigated for selenium determinations in four different ionic strength media. Factors affecting the voltammetric behavior of selenium, related to sample composition and analytical parameters, were here evaluated and optimized in order to assay the analyte as Cu2 Se species in acidic medium. The UVirradiation pre-treatment step was employed for the reductively conversion of non-electroactive species into Se(IV) as well as for DOM photo-oxidation, which may interfere in voltammetric determinations. An efficient photolytic degradation of DOM, preceding its complete photo-oxidation to carbon dioxide and water, occurs in 60 min irradiation interval, which also ensures the selenium conversion in 0.1 mol L−1 HCl medium. The proposed methodology can be used for the selenium speciation analysis in low and high saline media with good accuracy. Acknowledgements
SW: seawater; HT: hydrothermal fluid; HD: hemodialysis fluid. a UV-irradiation. b Synthetic seawater (prepared according to IAPSO) spiked with Se(IV) standard reference material (3149, batch F707118, NIST, USA) and Se(VI).
We are grateful to the Brazilian foundations CNPq and CAPES (Probral N◦ 240/06) for scholarship support.
occurs in 60 min irradiation interval, which also ensures the selenium conversion in 0.1 mol L−1 HCl medium.
[1] R.J. Shamberger, Biochemistry of Selenium, Plenum, New York, USA, 1983. [2] C.M.G. van den Berg, S.H. Khan, Anal. Chim. Acta 231 (1990) 221. [3] L.M. de Carvalho, G. Schwedt, G. Henze, S. Sander, Analyst 124 (1999) 1803. [4] G. Mattsson, L. Nyholm, Å. Olin, U. Örnemark, Talanta 42 (6) (1995) 817. [5] P. Papoff, F. Bocci, F. Lanza, Microchem. J. 59 (1998) 50. [6] T. Ferri, S. Rossi, P. Sangiorgio, Anal. Chim. Acta 361 (1998) 113. [7] T. Ferri, G. Favero, Microchem. J. 85 (2) (2007) 222. [8] D. Tanzer, K. Heumann, Anal. Chem. 63 (1991) 1984. [9] C.Y. Kuo, S.Y. Jiang, J. Chromatogr. A 1181 (2008) 60. [10] R. Morales, J.F. López-Sánchez, Trends Anal. Chem. 27 (2) (2008) 183. [11] C. B’Hymer, J.A. Caruso, J. Chromatogr. A 1114 (1) (2006) 1. [12] L. Zhang, Y. Morita, A. Sakuragawa, A. Isozaki, Talanta 72 (2) (2007) 723. [13] S. Sander, A. Koschinsky, Mar. Chem. 71 (2000) 83. [14] W. Davison, M. Whitfield, J. Electroanal. Chem. 75 (1977) 763. [15] J. Long, Y. Nagaosa, Anal. Sci. 23 (2007) 1343. [16] T. Ishiyama, T. Tanaka, Anal. Chem. 68 (1996) 3789. [17] K. Ebhardt, F. Umland, Fresen. J. Anal. Chem. 310 (1982) 406. [18] G. Henze, P. Monks, G. Tölg, F. Umland, E. Wessling, Fresen. J. Anal. Chem. 295 (1979) 1. [19] S. Forbes, G. Bound, T. West, Talanta 26 (1979) 473. [20] C. Arlt, R. Naumann, Fresen. J. Anal. Chem. 282 (1976) 463. [21] G. Jarzabek, Z. Kublik, Anal. Chim. Acta 143 (1982) 121. [22] S. Abdeloju, A. Bond, Anal. Chim. Acta 148 (1983) 59. [23] R. Bin Ahmad, J. Hill, R. Magee, Analyst 108 (1983) 835. [24] G. Kaiser, G. Tölg, Fresen. J. Anal. Chem. 325 (1986) 32. [25] M. Ochsenkühn-Petropoulou, F. Tsopelas, Anal. Chim. Acta 467 (2002) 167. [26] C. Elleouet, F. Quentel, C. Madec, Water Res. 30 (1996) 909. [27] G. Cutter, Anal. Chim. Acta 98 (1978) 59. [28] J. Petterson, L. Hansson, U. Örnemark, A. Olin, Clin. Chem. 34 (1988) 1908. [29] J. Petterson, A. Olin, Talanta 38 (1991) 413. [30] T. Ferri, P. Sangiorgio, Anal. Chim. Acta 321 (1996) 185. [31] J. Lingane, L. Niedrach, J. Am. Chem. Soc. 70 (1948) 4115. [32] J. Lingane, L. Niedrach, J. Am. Chem. Soc. 71 (1949) 196. [33] G. Jarzabek, Z. Kublik, J. Electroanal. Chem. 137 (1982) 247. [34] P. Zuman, G. Somer, Talanta 51 (2000) 645. [35] Deutsche Chemische Gesellschaft (Ed.), Gmelins Handbuch der Anorganischen Chemie: System-Nr 10, Selen (B), Gmelin, Germany, 1949. [36] R.A. Durst, D.N. Hume, Anal. Chim. Acta 251 (1991) 3. [37] L.M. de Carvalho, P.C. do Nascimento, D. Bohrer, R. Stefanello, D. Bertagnolli, Anal. Chim. Acta 546 (2005) 79. [38] P.C. do Nascimento, D. Bohrer, L.M. de Carvalho, C. Caon, E. Pilau, Z. Vendrame, R. Stefanello, Talanta 65 (2005) 954. [39] Mattsson, L. Nyholm, Å. Olin, J. Electroanal. Chem. 377 (1994) 149.
3.3. Interferences Since the interference of DOM toward selenium determination in SW and HT fluid samples was overcame by the use of UVirradiation, the interference of trace elements was investigated here by adding potential interfering species in all investigated samples. This way, solutions of Fe(II)/(III), Zn(II), Mn(II), Al(III), Cr(VI), Sb(III), Pb(II), As(III), Cd(II), Te(IV) and S2− were chosen for the interference experiments. Only Cd(II) and Te(IV) at concentrations higher than 10 mg L−1 could interfere due to the voltammetric peaks at −650 and −870 mV for Cd(II) and Te(IV), respectively. However, this concentration is about of 103 times higher than the limit selenium concentration (100 g L−1 ) tested here, and in real samples, Cd(II) and Te(IV) concentrations are normally much lower (or even inexistent, e.g. Te(IV) in HD fluids). 3.4. Real sample analysis Since selenium was not found in the investigated real samples, the method accuracy was assessed by recovery experiments. Hydrothermal fluid and seawater samples were UV-irradiated for DOM photo-oxidation before the voltammetric measurements. Then, the SW, HT and HD samples were spiked with Se(IV) and/or Se(VI) in order to obtain the working samples. Recoveries between 99.8% and 110.2% were calculated for Se(IV), and the selenium photolytic conversion for Se(VI) spiked samples was satisfactory with recoveries between 91.7% and 112.9%. Additionally, a synthetic SW sample prepared according to IAPSO and spiked with selenium at the g L−1 level was also analyzed. Recoveries of 97.9 and 104.2% were obtained (Table 2). The method accuracy was also tested in the presence of organic matter with bovine liver CRM (0.73 g g−1 Se) NIST 15776. The determined selenium concentration was 0.69 g g−1 .
References
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Voltammetric determination of Se(IV) and Se(VI) in saline samples—Studies with seawater, hydrothermal and hemodialysis fluids Paulo C. do Nascimento a,∗ , Cristiane L. Jost a , Leandro M. de Carvalho a , Denise Bohrer a , Andrea Koschinsky b a b
Departamento de Química, Universidade Federal de Santa Maria, C.P. 5051, 97105-970 Santa Maria, RS, Brazil School of Engineering and Science, Geosciences and Astrophysics, Jacobs University Bremen GmbH, P.O. Box 750561, D-28725 Bremen, Germany
a r t i c l e
i n f o
Article history: Received 30 January 2009 Received in revised form 17 April 2009 Accepted 23 June 2009 Available online 30 June 2009 Keywords: Selenium Ionic strength Speciation analysis Cathodic stripping voltammetry UV-irradiation
a b s t r a c t Determination of Se(IV) and Se(VI) in high saline media was investigated by cathodic stripping voltammetry (CSV). The voltammetric method was applied to assay selenium in seawater, hydrothermal and hemodialysis fluids. The influence of ionic strength on selenium determination is discussed. The CSV method was based on the co-electrodeposition of Se(IV) with Cu(II) ions and Se(VI) determined by difference after sample UV-irradiation for photolytic selenium reduction. UV-irradiation was also used as sample pre-treatment for organic matter decomposition. Detection limit of 0.030 g L−1 (240 s deposition time) and relative standard deviation (RSD) of 6.19% (n = 5) for 5.0 g L−1 of Se(IV) were calculated. Linear calibration range for selenium was observed from 1.0 to 100.0 g L−1 . Concerning the pre-treatment step, best results were obtained by using 60 min UV-irradiation interval in H2 O2 /HCl medium. Se(VI) was reduced to the Se(IV) electroactive species with recoveries between 91.7% and 112.9%. Interferents were also investigated. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Selenium at low concentrations is known as an essential trace element for animals, however at higher concentrations, toxic effects can be observed. The range between the necessary and the toxic intake is narrow [1]. Additionally, there is a growing interest in the differentiation of selenium species in environmental matrices due to the well-known dependence of biological and toxicological effects on the chemical form. The so called ‘speciation analysis’ has gained attention during the past years. Some studies concerning the determination of selenium species have been published to understand its role in the environment and human health, as well [2–7]. Separation techniques, such as high performance liquid chromatography (HPLC) and capillary electrophoresis (CE) with inductively coupled plasma – mass spectrometry (ICP-MS) have been employed for selenium speciation studies [8–11]. The use of optical methods such as graphite furnace atomic absorption spectrometry (GF-AAS) with a complementary co-precipitation step has also been reported [12]. Concerning speciation of selenium in saline matrices, voltammetry can present better performances than other techniques since good sensitivity and selectivity can be reached in
∗ Corresponding author. Tel.: +55 55 3220 8870; fax: +55 55 3220 8870. E-mail address: [email protected] (P.C. do Nascimento). 0003-2670/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2009.06.057
such media. Additionally, low capital investment is required, and the equipment can be portable to be used in onboard-ship analysis [13]. Among the voltammetric techniques used for selenium determination, stripping analysis are uniquely suited for speciation because of their sensitivity to low metal concentrations and to the chemical forms of metals in solution, as well [14]. In alternative to mercury working electrodes, the use of bismuth-film and silver disk has been reported, as well [15,16]. With respect to selenium speciation, acidic electrolytes as HCl and some specific reagents such as copper have already been successfully used [17–23]. Copper ions play an important role for selenium determinations since the formation of an insoluble layer of Cu2 Se as a co-electrolysis step not only reduces the interferences from metal ions by forming stable selenides but also increases the sensitivity of the CSV method [2–4,18,24]. As selenite (Se4+ ) is the selenium species determined by CSV and selenium is usually found as Se(VI) or Se(IV), selective methods are required to convert it into selenite before the electrochemical determination. A well-established procedure for selenium reduction is the use of photochemical pre-treatment by UV-irradiation [2,3,5,25,26]. Furthermore, in samples with low levels of dissolved organic matter (DOM), the interfering organic material is removed with UV-irradiation and the selenium bound in DOM is transformed to inorganic species. Chemical reduction and heating with HCl (4–6 mol L−1 ) have been proposed, as well [27–29]. However, the use of drastic acidic conditions and sample heating is a time consuming procedure and may result in a
P.C. do Nascimento et al. / Analytica Chimica Acta 648 (2009) 162–166
selenium conversion to the metallic state associated to losses of analyte by volatilization [27] even when sample digestion is microwave assisted [6,30]. The purpose of this study was to optimize the CSV in the presence of added copper for selenium speciation concerning the ionic strength influence in order to assay the analytes in low and high saline matrices by using UV-irradiation as pre-treatment step. 2. Experimental 2.1. Instrumentation and apparatus 2.1.1. Voltammetry CSV was applied in ca. 0.1 mol L−1 HCl medium, with the addition of 1.0 mg L−1 Cu(II), according to previous published studies [4,6]. The electrodeposition step was performed under stirring (400 rpm) by using the adsorption potential of −300 mV for all media (60 s for 1.0–20.0 g L−1 and 10 s for 20.0–100.0 g L−1 ). After an equilibration time of 10 s, the differential pulse voltammetric scan was recorded from −300 to −750 mV (60 mV s−1 scan rate; −50 mV pulse amplitude) – the stripping peak occurs at ca. −610 mV. Voltammograms were collected by using a Metrohm 693 VA Processor with a 694 VA Stand operating with hanging mercury drop electrode (HMDE). A platinum wire was taken as the auxiliary electrode and all potentials were quoted against Ag/AgCl/KCl (3 mol L−1 ) reference electrode. The cell volume was 10 mL. The standard addition method was used to evaluate the analyte concentrations through the measurement of the peak height. 2.1.2. UV-irradiation The UV-reduction step was performed by irradiating 10 mL of the samples spiked with a known amount of Se(VI) and/or Se(IV) for 60 min at 85 ± 5 ◦ C (Hg-high pressure lamp, 500 W, Metrohm, Herisau, Switzerland) in order to test the method for total dissolved selenium determination. Each UV-irradiation was performed in 10 mL sample volume made of 0.1 mol L−1 HCl and added by 10 L of destilled H2 O2 30% (v/v), according to procedure published by Mattsson et al. [4]. 2.2. Reagents All chemicals were of analytical-reagent grade. The water used throughout was distilled, deionized and further purified by a MilliQ high purity water device (Millipore, Bedford, USA). The 1.0 mg L−1 Se(IV) standard reference material 3149 (batch F707118, National Institute of Standards and Technology (NIST), USA) was purchased from SpecSol (Brazil). The 1.0 g L−1 Se(VI) standard stock solution was prepared by dissolving 0.2392 g of sodium selenate (Fluka, Germany) in 100 mL 0.5 mol L−1 HNO3 instead of high purity water to help the maintenance of all selenium on the (VI) oxidation state. The solutions were stored in opaque PTFE flasks to protect against sunlight. Aliquots of these stock solutions were taken to prepare the working solutions. 100 mg L−1 Cu(II) aqueous solution was prepared from the chloride salt (Merck) in order to form the selenide complex while a 0.1 mol L−1 HCl solution was used as supporting electrolyte. The used hydrogen peroxide (30%, v/v) was purchased from Vetec (Brazil). Humic acid (Aldrich, Germany) was used to prepare 10.0 mg L−1 aqueous solutions.
Table 1 Test solutions matching the composition of real samples. SW (mol L−1 )
Species +
Na K+ Ca2+ Mg2+ Cl− SO4 2− HCO3 − CH3 COO−
0.048 0.010 0.010 0.054 0.56 0.028 0.002 –
The experiments were performed by using test solutions spiked with selenium having similar saline composition with seawater (SW), hydrothermal (HT) and hemodialysis (HD) fluids (Table 1). For DOM photo-oxidation experiments, 10.0 mg L−1 humic acid was added to SW or HT fluid samples. Non-saline (NS) test solutions
HT (mol L−1 ) 0.05–0.12 0.010 0.010–0.050 0–0.054 0.05–1.2 – 0.005–0.015 –
HD (mol L−1 ) 3.61 0.050–0.070 0.061 0.017–0.030 3.82 – – 0.105
SW: seawater; HT: hydrothermal fluid; HD: hemodialysis fluid.
were also prepared by adding selenium to pure water. The electrolyte (1.0 mg L−1 Cu(II) and 0.1 mol L−1 HCl) was added to all test solutions. Synthetic SW sample prepared according to IAPSO (International Association for the Physical Sciences of the Oceans) have the composition: As 1.7 g kg−1 ; Al 0.54 g kg−1 ; B 4.5 mg kg−1 ; Ba 14 g kg−1 ; C 27.6 mg kg−1 ; Ca 412 mg kg−1 ; Cd 0.08 g kg−1 ; Cl− 19,354 mg kg−1 ; Cu 0.25 g kg−1 ; Fe 0.055 g kg−1 ; K 399 mg kg−1 ; Mg 1290 mg kg−1 ; Mn 0.014 g kg−1 ; Na 10,770 mg kg−1 ; Ni 0.5 g kg−1 ; P 70 g kg−1 ; Pb 0.002 g kg−1 ; S 904 mg kg−1 ; Si 2.8 mg kg−1 ; U 3.3 g kg−1 ; Zn 0.4 g kg−1 . 2.4. Samples SW samples were collected in the vicinity of Helgoland Island, Germany, filtered through a 0.45 m membrane filter (Millipore, USA), acidified to pH 2.0 with HCl (37%, w/v) and then stored refrigerated in polyethylene bottles. A pool of HT fluid was made with samples from the Logatchev field (15◦ N, 45◦ W) and stored refrigerated without pre-treatment. HD fluid samples were taken directly from the original container without pre-treatment or dilution (HB Bic Salbego, Brazil). 3. Results and discussion The conventional CSV of selenium is based on the accumulation of HgSe at HMDE in acidic solutions [31–33]. The selenium peak obtained during the cathodic potential scan can also be attributed to the reduction of slightly soluble Se(IV) species like HgSeO3 and Hg3 (HSeO3 )2 (SeO3 )2 formed with mercury ions on the electrode surface [34]. However, in the presence of high chloride concentrations, calomel is also formed and can be responsible for some conversion of Se(IV) into Se(VI) and Se(0) together with an oxidation of Hg(I) to Hg(II) [35]. This way, the CSV method based on the co-electrodeposition of Se(IV) with Cu(II) ions [2–4,18,24] was more convenient considering the chloride concentration of the investigated samples. Additionally, the stripping peaks of accumulated Cu2 Se species on the HMDE electrode yields, in comparision with convetional CSV selenium determination, to better detection limits, higher linear range and low interference of metal ions [4]. The copper selenide stripping peak depends on the accumulation of Cu2 Se intermediate species at HMDE (Eq. (1)) with subsequent reduction during the cathodic scan (Eq. (2)). H2 SeO3 + 2Cu2+ + 4H+ + 8e− → Cu2 Se + 3H2 O +
2.3. Test solutions
163
−
Cu2 Se + 2H + 2e → H2 Se + 2Cu(Hg)
(1) (2)
The reaction pathway indicates that the reduction of Se(IV) to Se(−II) is pH-dependent. Previous results showed that both HCl and HClO4 were appropriate electrolytes. The best peak shapes were obtained at pH values between 1.6 and 2.0. Additionally, with increasing concentration of Cu(II) ions the height of Se(IV) stripping peaks is affected [34]. The effect of copper ions was inves-
164
P.C. do Nascimento et al. / Analytica Chimica Acta 648 (2009) 162–166
Fig. 1. Effect of deposition potential on the CSV peak current for 10.0 g L−1 Se(IV). Deposition time: 10 s. NS: non-saline sample; SW: seawater; HT: hydrothermal fluid; HD: hemodialysis fluid.
tigated for both SW and non-saline media. The 0.5–10.0 mg L−1 Cu(II) range was evaluated for the determination of 10.0 g L−1 Se(IV). In the present work, only for Cu(II) concentrations higher than 1.0 mg L−1 , the voltammetric Se(IV) signal was strictly a linear function of Se(IV) concentration and independent of small variations on Cu(II) ion concentration. The ionic strengths of the assayed samples (see below) are high and cover a wide variation range mainly due to their high chloride concentrations. A dependence between the stripping potential shift and the ionic strength of the investigated saline samples was observed. Comparing the CSV data obtained by measuring the test solutions NS ( = 0.10 mol L−1 ), SW ( = 0.70 mol L−1 ), HT ( = 0.75 mol L−1 ) and HD ( = 4.0 mol L−1 ), the selenium peak potential shifted from ca. −630 mV (NS) to −530 mV (HD). By increasing the ionic strength, the water activity increases and the hydration number of the metallic species decreases, shifting their reduction potential to more positive values [36,37]. This way, the Se(IV) peak potential depends in some extension on the composition and concentration of the saline sample. Conversely, in high chloride concentrations as existing in SW, HT and HD samples, a negative potential shift could be expected due to metal-chloro complex formation and mercury electrode surface competition between the analyte and those compounds [38]. However, this behavior was not observed for selenium measurements after the co-electrodeposition with Cu(II) ions. This fact may be related to the favorable formation of the slightly soluble Cu2 Se (Ks 3.8 × 10−9 ) monolayer on the electrode surface. 3.1. Effect of ionic strength The test solutions matching the composition of real samples (Table 1) were used to evaluate the voltammetric behavior of selenium after the co-electrodeposition of 10.0 g L−1 Se(IV) with Cu(II) ions. The influence of the ionic strength on the peak current of Se(IV) was investigated considering the deposition potential and time as well as the potential scan rate. The influence of DOM existing in SW and HT fluid was not considered here since its elimination was further investigated by photo-oxidation as cleanup step. 3.1.1. Ionic strength and deposition potential Variations on the deposition potential revealed that the Se (IV) peak height measured by CSV at −610 mV for a 10 g L−1 Se(IV) solution showed its highest values for deposition potentials ranging from −200 to −400 mV (Fig. 1) considering the whole ionic strength interval covered by the assayed samples (test solutions and real samples). The lowest peak currents were obtained for HD samples ( = 4.0 mol L−1 ) and no significant differences were observed among the other samples. Indeed, the stripping peaks remained almost the same for NS, SW and HT test solutions with no spe-
cific influence of ionic strengths ranging from 0.1 to 0.75 mol L−1 . This result suggests that the process responsible for the analyte deposition at the electrode surface is predominantly electrostatic. It is obvious that an deposition potential around −300 mV should be chosen for all samples. Regardless of these results, some positive deposition potentials were also tested because selenium was determined as copper complex in samples covering a very wide chloride concentration range. For electrolytes containing chloride ions, copper occurs predominantly as Cu(I) at deposition potentials ranging from ca. 0.2 to −0.15 V due to complex stabilization, although for positive deposition potentials, incomplete formation of Cu2 Se occurs since its reduction potential lies close to −50 mV [2]. On the other hand, decreases on peak height observed for deposition potentials more negative than −400 mV can be related to the charging process on mercury electrode surface, which changes from positive to negative in chloride medium, at potentials lower than ca. −500 mV [4]. 3.1.2. Ionic strength and deposition time The influence of the ionic strength on the peak current for selenium measurements by the CSV method was evaluated for deposition times ranging from 0 to 200 s. Two different profiles could be observed (Fig. 2). For the lowest ionic strength (0.10 mol L−1 ) the peak current increased slightly with increasing deposition time up to 100 s and more intensely for higher deposition times in the range of 100–140 s. This roughly S-shaped curve can be related to mercury electrode surface saturation [2] that seems to occur only for the lowest ionic strength. This behavior agrees with a predominantly electrostatic electrode process. The higher the ionic strength, the lower the analyte accumulation at the electrode surface as well as the sensitivity. In the present work, deposition time of 10 s was adequate to assay selenium in all samples in the concentration range of 20–100 g L−1 , and 60 s for the range of 1–20 g L−1 . 3.1.3. Ionic strength and scan rate Among the tested parameters, the voltammetric selenium peak showed not to be severely influenced by ionic strength with respect to the scan rate. It was observed that the height of selenium reduction peak increases linearly with the scan rate in the range of 4–120 mV s−1 (Fig. 3) with slopes fairly independent on the ionic strengths. This linearity suggests that Cu2 Se is deposited as a monolayer on the mercury electrode surface since a multilayered precipitate would exhibit a non-linear behavior as a result of stronger kinetic effects related to a possible heterogeneous nature of the electrode process. Since the peaks were well-shaped by using a 60 mV s−1 scan rate, this value was further considered as optimal value for all media.
Fig. 2. Effect of deposition time on the CSV peak current for 10.0 g L−1 Se(IV). Deposition potential: −300 mV. NS: non-saline sample; SW: seawater; HT: hydrothermal fluid; HD: hemodialysis fluid.
P.C. do Nascimento et al. / Analytica Chimica Acta 648 (2009) 162–166
165
3.2. Selenium speciation Since Se(IV) is the voltammetrically active species, the use of a pre-treatment step to reduce the existing Se(VI) was required. UV-irradiation is a well-known and useful tool for photolytic conversions, which was used here. The reduction efficiency was checked by adding Se(VI) to SW samples where no Se(IV) was previously detected by CSV. The selenium speciation was achieved by the difference between the selenium concentration in irradiated and non-irradiated samples.
Fig. 3. Effect of scan rate on the CSV peak for 10.0 g L−1 Se(IV). Deposition potential: −300 mV; deposition time: 10 s. NS: non-saline sample; SW: seawater; HT: hydrothermal fluid; HD: hemodialysis fluid.
3.1.4. Calibrations Linear peak current responses were obtained from 1.0 to 100 g L−1 Se(IV). For concentrations up to 20.0 g L−1 , 60 s deposition time was necessary, while for the range of 20.0–100.0 g L−1 , a shorter deposition time (10 s) was already sufficient to produce well-shaped voltammetric peaks. This way, two different calibration plots could be used depending on the selenium level in the samples. Regression analysis showed for all test solutions r2 ≥ 0.996 and the following calibration functions for the 1.0–20.0 g L−1 concentration range (60 s deposition time): NS: y = 1.297x + 0.562; SW: y = 1.461x + 0.514; HT: y = 1.528x + 0.376; HD: y = 1.160x + 0.100, where x and y are concentration in g L−1 and current in nA, respectively. Similarly, for more concentrated Se(IV) test solutions and using 10 s deposition time the calibration functions NS: y = 0.396x + 0.795; SW: y = 0.389x + 0.187; HT: y = 0.406x + 0.003; HD: y = 0.307x + 0.185 were obtained. The relative standard deviation (RSD) was 6.19% for five independent measurements of 5.0 g L−1 Se(IV) solution. A detection limit of 0.030 g L−1 (240 s deposition time) was calculated considering the condition (3 ± BL ), where is the standard deviation (n = 10) of the background current (BL ) at −610 mV. Typical voltammograms of 7.5 g L−1 Se(IV) spiked SW, HT and HD samples are shown in Fig. 4.
Fig. 4. Voltammograms of 7.5 g L−1 selenium spiked samples. Deposition time: 60 s; deposition potential: −300 mV. I: hemodialysis fluid; II: hydrothermal fluid; III: seawater.
3.2.1. Selenium photolytic conversion The selenium reduction by the action of UV-irradiation was investigated in alkaline and acidic media in the test solutions using NS samples as control. UV-irradiation was performed after the addition of 5.0 g L−1 Se(VI) to SW and HT fluid test solutions. The use of alkaline conditions (NaOH 0.4 mol L−1 , pH 11.0) was previously reported for selenium speciation in SW [3]. By using this condition, the Se(IV) recovery lay around 115.0% for control samples. However, no analytic peak was observed for SW and HT fluid test solutions for irradiation interval up to 90 min. For a longer irradiation interval (180 min), recoveries about 88.0% for SW and 105.0% for HT fluid were calculated. For UV-irradiation in acidic media, it was showed elsewhere that the best results for selenium conversion were obtained with HCl compared to H2 SO4 and HClO4 [2–6,26,27]. Additionally, HCl was also employed in the present work as electrolyte for the voltammetric determinations. The use of 0.1 mol L−1 HCl with 60 min irradiation interval, reported by Mattsson et al. [4], was also used here for the selenium photolytic conversion. In fact, by using this medium, Se(IV) recoveries ranging from 91.1% to 98.4% were obtained. 3.2.2. Samples containing DOM Natural ligands present in the investigated samples (SW and HT fluids) can interfere with the selenium determination by voltammetry; therefore they must be inactivated before the analyses. To assay selenium in SW and HT fluids, UV-irradiation was used to overcome DOM interferences, besides helping on the selenium photolytic conversion process. Humic acid was added to test solutions to simulate DOM content of real SW and HT fluids. Recovery experiments using the selected experimental conditions were also carried out in real samples spiked with selenium. Among the tested cleanup/photolytic conversion conditions in test solutions containing humic acid (10.0 mg L−1 ), 60 min of UV-irradiation in presence of HCl/H2 O2 solutions produced good recoveries. By spiking SW and HT fluid test solutions at 5.0 g L−1 Se(VI) level, recoveries ranging from 104.6% to 113.8% were obtained for samples containing or not DOM. By using 30 or 90 min irradiation intervals, recoveries about of 85 and 125% were obtained, respectively. These results could be explained by either a redox equilibrium established between Se(IV) and Se(VI) species during the irradiation, or by a steady state created by opposing reduction and oxidation reactions [39]. In addition, photolytic conversion of Se(VI) can also be induced by direct absorption of wavelengths under 230 nm or from the action of carbon-containing radicals, generated by the photochemical degradation of natural DOM, which has reducing properties. On the other hand, highly oxidizing chlorine and hydroxyl radical, which are formed by the action of UV light with wavelengths <200 nm, may also be important, especially in the observed reoxidation of Se(IV) after long irradiation intervals. Perhaps these oxidizing radicals react with DOM until it is totally consumed. Then, they start to oxidize Se(IV). The adopted irradiation scheme consisted of DOM oxidation assisted by the addition of 30 L H2 O2 (30%, v/v) to the 0.1 mol L−1 HCl medium. An efficient photolytic degradation of DOM, preceding its complete photolytic conversion to carbon dioxide and water,
166
P.C. do Nascimento et al. / Analytica Chimica Acta 648 (2009) 162–166
Table 2 Determination and recovery for selenium in saline samples by cathodic stripping voltammetry. Sample
Se (g L−1 ) added
Se(IV) (g L−1 ) detected
Recovery (%)
HD
– 5.00 as Se(IV) 5.00 Se(IV) + 2.50 Se(VI) 2.00 as Se(IV)
– 99.80 91.73 110.00
HT + UVa
– 5.00 as Se(IV) 5.00 Se(IV) + 2.50 Se(VI) 2.00 as Se(IV)
– 110.20 112.13 105.50
SW + UVa
– 5.00 as Se(IV) 5.00 Se(IV) + 2.50 Se(VI) 2.00 as Se(IV)
– 99.20 112.93 100.50
SWb + UVa
– 5.00 as Se(IV) 5.00 Se(IV) + 2.50 Se(VI)
– 104.20 97.87
4. Conclusions The CSV method was investigated for selenium determinations in four different ionic strength media. Factors affecting the voltammetric behavior of selenium, related to sample composition and analytical parameters, were here evaluated and optimized in order to assay the analyte as Cu2 Se species in acidic medium. The UVirradiation pre-treatment step was employed for the reductively conversion of non-electroactive species into Se(IV) as well as for DOM photo-oxidation, which may interfere in voltammetric determinations. An efficient photolytic degradation of DOM, preceding its complete photo-oxidation to carbon dioxide and water, occurs in 60 min irradiation interval, which also ensures the selenium conversion in 0.1 mol L−1 HCl medium. The proposed methodology can be used for the selenium speciation analysis in low and high saline media with good accuracy. Acknowledgements
SW: seawater; HT: hydrothermal fluid; HD: hemodialysis fluid. a UV-irradiation. b Synthetic seawater (prepared according to IAPSO) spiked with Se(IV) standard reference material (3149, batch F707118, NIST, USA) and Se(VI).
We are grateful to the Brazilian foundations CNPq and CAPES (Probral N◦ 240/06) for scholarship support.
occurs in 60 min irradiation interval, which also ensures the selenium conversion in 0.1 mol L−1 HCl medium.
[1] R.J. Shamberger, Biochemistry of Selenium, Plenum, New York, USA, 1983. [2] C.M.G. van den Berg, S.H. Khan, Anal. Chim. Acta 231 (1990) 221. [3] L.M. de Carvalho, G. Schwedt, G. Henze, S. Sander, Analyst 124 (1999) 1803. [4] G. Mattsson, L. Nyholm, Å. Olin, U. Örnemark, Talanta 42 (6) (1995) 817. [5] P. Papoff, F. Bocci, F. Lanza, Microchem. J. 59 (1998) 50. [6] T. Ferri, S. Rossi, P. Sangiorgio, Anal. Chim. Acta 361 (1998) 113. [7] T. Ferri, G. Favero, Microchem. J. 85 (2) (2007) 222. [8] D. Tanzer, K. Heumann, Anal. Chem. 63 (1991) 1984. [9] C.Y. Kuo, S.Y. Jiang, J. Chromatogr. A 1181 (2008) 60. [10] R. Morales, J.F. López-Sánchez, Trends Anal. Chem. 27 (2) (2008) 183. [11] C. B’Hymer, J.A. Caruso, J. Chromatogr. A 1114 (1) (2006) 1. [12] L. Zhang, Y. Morita, A. Sakuragawa, A. Isozaki, Talanta 72 (2) (2007) 723. [13] S. Sander, A. Koschinsky, Mar. Chem. 71 (2000) 83. [14] W. Davison, M. Whitfield, J. Electroanal. Chem. 75 (1977) 763. [15] J. Long, Y. Nagaosa, Anal. Sci. 23 (2007) 1343. [16] T. Ishiyama, T. Tanaka, Anal. Chem. 68 (1996) 3789. [17] K. Ebhardt, F. Umland, Fresen. J. Anal. Chem. 310 (1982) 406. [18] G. Henze, P. Monks, G. Tölg, F. Umland, E. Wessling, Fresen. J. Anal. Chem. 295 (1979) 1. [19] S. Forbes, G. Bound, T. West, Talanta 26 (1979) 473. [20] C. Arlt, R. Naumann, Fresen. J. Anal. Chem. 282 (1976) 463. [21] G. Jarzabek, Z. Kublik, Anal. Chim. Acta 143 (1982) 121. [22] S. Abdeloju, A. Bond, Anal. Chim. Acta 148 (1983) 59. [23] R. Bin Ahmad, J. Hill, R. Magee, Analyst 108 (1983) 835. [24] G. Kaiser, G. Tölg, Fresen. J. Anal. Chem. 325 (1986) 32. [25] M. Ochsenkühn-Petropoulou, F. Tsopelas, Anal. Chim. Acta 467 (2002) 167. [26] C. Elleouet, F. Quentel, C. Madec, Water Res. 30 (1996) 909. [27] G. Cutter, Anal. Chim. Acta 98 (1978) 59. [28] J. Petterson, L. Hansson, U. Örnemark, A. Olin, Clin. Chem. 34 (1988) 1908. [29] J. Petterson, A. Olin, Talanta 38 (1991) 413. [30] T. Ferri, P. Sangiorgio, Anal. Chim. Acta 321 (1996) 185. [31] J. Lingane, L. Niedrach, J. Am. Chem. Soc. 70 (1948) 4115. [32] J. Lingane, L. Niedrach, J. Am. Chem. Soc. 71 (1949) 196. [33] G. Jarzabek, Z. Kublik, J. Electroanal. Chem. 137 (1982) 247. [34] P. Zuman, G. Somer, Talanta 51 (2000) 645. [35] Deutsche Chemische Gesellschaft (Ed.), Gmelins Handbuch der Anorganischen Chemie: System-Nr 10, Selen (B), Gmelin, Germany, 1949. [36] R.A. Durst, D.N. Hume, Anal. Chim. Acta 251 (1991) 3. [37] L.M. de Carvalho, P.C. do Nascimento, D. Bohrer, R. Stefanello, D. Bertagnolli, Anal. Chim. Acta 546 (2005) 79. [38] P.C. do Nascimento, D. Bohrer, L.M. de Carvalho, C. Caon, E. Pilau, Z. Vendrame, R. Stefanello, Talanta 65 (2005) 954. [39] Mattsson, L. Nyholm, Å. Olin, J. Electroanal. Chem. 377 (1994) 149.
3.3. Interferences Since the interference of DOM toward selenium determination in SW and HT fluid samples was overcame by the use of UVirradiation, the interference of trace elements was investigated here by adding potential interfering species in all investigated samples. This way, solutions of Fe(II)/(III), Zn(II), Mn(II), Al(III), Cr(VI), Sb(III), Pb(II), As(III), Cd(II), Te(IV) and S2− were chosen for the interference experiments. Only Cd(II) and Te(IV) at concentrations higher than 10 mg L−1 could interfere due to the voltammetric peaks at −650 and −870 mV for Cd(II) and Te(IV), respectively. However, this concentration is about of 103 times higher than the limit selenium concentration (100 g L−1 ) tested here, and in real samples, Cd(II) and Te(IV) concentrations are normally much lower (or even inexistent, e.g. Te(IV) in HD fluids). 3.4. Real sample analysis Since selenium was not found in the investigated real samples, the method accuracy was assessed by recovery experiments. Hydrothermal fluid and seawater samples were UV-irradiated for DOM photo-oxidation before the voltammetric measurements. Then, the SW, HT and HD samples were spiked with Se(IV) and/or Se(VI) in order to obtain the working samples. Recoveries between 99.8% and 110.2% were calculated for Se(IV), and the selenium photolytic conversion for Se(VI) spiked samples was satisfactory with recoveries between 91.7% and 112.9%. Additionally, a synthetic SW sample prepared according to IAPSO and spiked with selenium at the g L−1 level was also analyzed. Recoveries of 97.9 and 104.2% were obtained (Table 2). The method accuracy was also tested in the presence of organic matter with bovine liver CRM (0.73 g g−1 Se) NIST 15776. The determined selenium concentration was 0.69 g g−1 .
References