Selenium and its redox speciation in rainwater from sites of Valparaı́so region in Chile, impacted by mining activities of copper ores

Water Research 36 (2002) 115–122

Selenium and its redox speciation in rainwater from sites of Valpara!ıso region in Chile, impacted by mining activities of copper ores Ida De Gregori*, Mar!ıa G. Lobos, Hugo Pinochet Instituto de Qu!ımica, Universidad Cato!lica de Valpara!ıso, P.O. Box 4059, Valpara!ıso, Chile Received 1 April 2000; received in revised form 1 April 2001; accepted 1 April 2001

Abstract The determination of the total concentration of selenium does not provide sufficient information about its toxicity and its bioavailability. The determination of its chemical forms is the basis for understanding the biogeochemical cycle in terrestrial and aquatic ecosystems and for detecting the species which might be toxic to biota. In this work we describe an analytical procedure to carry out the redox speciation of selenium present at ultratrace levels in rainwater from sites of Valpara!ıso region in Chile, impacted by mining activities of copper ores. A simple preconcentration step of the rainwater sample on a rotavapor system, in vacuum at low temperature permits the concentration of the different redox selenium species until levels quantifiable by sensitive techniques such as differential pulse cathodic stripping voltammetry or by spectrometric techniques, based on the hydride generation and detection by atomic absorption or atomic fluorescence spectrometry. These techniques coupled to redox chemical reactions allow the redox speciation of selenium. The results show that the open evaporation system can be used to concentrate water samples when the aim of the analysis is the determination of the total selenium concentration. On the contrary, to carry out its redox speciation only the preconcentration performed on rotavapor system, in vacuum can be used. When synthetic solutions containing different redox species of selenium, at ultratrace levels, were slowly evaporated on open system, Se(II) and Se(IV) were oxidized. The optimized procedure was then applied to the selenium determination and its redox speciation in rainwater samples collected in sites impacted by mining activities of copper ores. It was found that the amounts of total selenium in rainwater, as copper, from Puchuncavi valley decrease exponentially with the distance from the source, indicating that these elements in this region arise from the industrial complex Las Ventanas. In the redox speciation of selenium, Se(IV) and Se(VI) were the species found in all rainwater samples analyzed, providing selenium in species which are most favorable for their uptake by the vegetation grown in these soils. r 2001 Elsevier Science Ltd. All rights reserved. Keywords: Selenium; Speciation; Rainwater; Differential pulse cathodic stripping voltammetry; Hydride generation–atomic absorption spectrometry; Hydride generation–atomic fluorescence spectrometry

1. Introduction Selenium is an essential micronutrient for humans, animals and some plants. However, the safety margin between its nutrient and toxic doses is very narrow [1–4]. *Corresponding author. E-mail address: [email protected] (I. De Gregori).

Levels as low as 0.01 mg L
1 can cause deformation and death of wildfowl [5]. Se may be released to the environment by natural processes such as weathering of minerals or by anthropogenic activities such as fossil fuel combustion, industrial, agricultural and metallurgical processes, especially from mining activities of sulfide ores. In the environment, Se is generally found in metal sulfur deposits [6]. Inorganic Se can exist as selenide

0043-1354/01/$ - see front matter r 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 1 ) 0 0 2 4 0 - 8

116

I. De Gregori et al. / Water Research 36 (2002) 115–122

Se(II), selenite Se(IV) and selenate Se(VI). The organic forms generally encountered are selenoaminoacids, methylated compounds and selenoproteins [7]. The atmosphere plays an important role in the biogeochemical cycle of Se. It can be emitted to the atmosphere in vapor form or as particles, especially from mining processes of seleniferous ores, or from the coal fire of the electric power plants, sources identified as responsible for trace metal emissions, including Se [8]. According to the meteorological condition, it can be transported over varying distances, polluting the terrestrial and aquatic ecosystem [9,10]. Usually, wet deposition through rain and snow is the prevailing deposition mode, except in the areas close to a pollution source [11]. Se in water is found at trace level, and in rainwater probably at ultratrace levels, hence the demand for selective, sensitive and accurate analytical methods to determine the total concentration. As the chemical species of Se have different toxicological characteristics and different behaviors in the environment [12,13], the ability to discriminate between its different species is an advantage. Currently, the determination of total Se concentration does not provide sufficient information to ascertain it’s polluting potential. Several electrochemical and spectroscopic methods have been developed to determine total Se and to carry out its speciation [7]. In the electrochemical determination, Se(IV) is preconcentrated as HgSe at a hanging mercury drop electrode, followed by a stripping step, applying to the electrode a cathodic potential scan. Se(VI) is not reducible at the mercury electrode [14–17]. Also, the spectroscopic techniques based on the hydride generation coupled to atomic absorption spectrometry (HG-AAS) or to atomic fluorescence spectrometry (HGAFS) allow the direct determination of Se(IV) only [18– 21]. Therefore, when both inorganic Se species are present, Se(IV) is determined directly and Se(VI) after a chemical reduction to Se(IV), the Se(VI) concentration is evaluated by difference. Reduction of Se(VI) to Se(IV) is generally achieved by heating the sample in 6 mol L
1 HCl [20,7,18,22]. In Chile many ecosystems have been and are susceptible to be contaminated by toxic trace elements produced by the mining and industrial processing of ores and metals. These activities are recognized as sources of contamination by toxic elements. More than 1600 tons of Se are produced annually in the world, from mine production, primarily as a by-product of copper refining [9]. The aims of this work were to elaborate an analytical methodology based on a preconcentration step and subsequent application of sensitive electrochemical or atomic spectrometry analytical techniques to determine the total Se concentration and to carry out its redox speciation in rainwater samples, collected between 1997 and 1999 in the vicinity of copper industries, located in

Puchuncavi and Catemu valleys, from Valpara!ıso region in Chile. In both zones, the concentrations of Se in the atmosphere may be important due to the emissions from mining industrial complexes.

2. Materials and methods 2.1. Sampling sites, rainwater collection and treatment Rainwater samples were collected at Puchuncav!ı valley, an agricultural zone (in the north of Valpara!ıso city), which receives the impact of the industrial complex ‘‘Las Ventanas’’, where a smelter and electrorefinery plant of copper ore as well as a coal-fired thermoelectric power plant are located. Se is obtained in this industry as a by-product of the copper electrolytic refining processes [23]. Catemu valley is located in the east of Valpara!ıso (in the Aconcagua river valley) which is also influenced by a smelter of copper ore (Chagres). The third zone was Casablanca valley, a rural and agricultural area, located in the south of Valpara!ıso, that is not particularly submitted to the impact of mining activities. Due to its similar characteristics with Puchuncav!ı valley it has been often selected as a reference area [24,25]. The choice of sites in both impacted zones was made taking into account the topography and the prevailing direction of winds [26]. The average annual rain deposition in this zone of Chile is 200–300 mm, and takes place only in the period MayOctober (autumn and winter seasons). Six sites were selected in Puchuncav!ı valley and three in Catemu valley, at different distances from the sources, and two in Casablanca valley All rainwater samples were collected in duplicate, in samplers made in home, which consist of a polyethylene funnel of 21.5 cm upper diameter joined to a polyethylene container of 4 L capacity, which were cleaned with 10% nitric acid and rinsed with deionized water. The samplers were stored in closed polyethylene bags and were manually opened and closed at the beginning and at the end of the rainfall, conditioning sample collection to wet deposition. The samplers were positioned on racks, 2 m above the ground to avoid interference of soil particles during rainfalls. Rainwater samples were collected on 17 May 1997 (the first rainfall in that year); during the only event in 1998 (10 September), one of the most dry years in our country; and two events in 1999 (16 July and 4–6 September). After collection, on the same day, the samplers were transported to the laboratory. After total volumes and pH measurements, the samples were immediately filtered through 0.45 mm pore and divided into four portions. To determine total Se, two aliquots were digested with HClO4+HNO3 (2 mL of each concentrated acid), and slowly evaporated until almost complete evaporation.

117

I. De Gregori et al. / Water Research 36 (2002) 115–122

The final volume was made to 10 mL with deionized water and stored at 41C in the refrigerator, until analysis (during the next 2–3 days). The redox speciation of Se was performed in the other aliquots. Rainwater samples were degassed with ultrapure nitrogen and were evaporated on rotavapor system, in vacuum, at 501C (Buchi RE 111-Buchi 461). Se(IV) was determined first in these solutions; the sum of Se(IV) and Se(VI) was determined in the solution obtained after the chemical reduction of Se(VI) to Se(IV), heating a 1 : 1 mixture of sample with concentrated HCl, in a closed tube, in a water bath at 901C, for 40 min, or in a microwave oven, for 2 min at 120 W (Micro digest 301, Prolabo). Se(VI) was calculated from the difference between these two determinations. Se(II) was evaluated from the difference between total Se and the sum of Se(IV) and Se(VI).

2.2. Analysis The Se concentrations were determined by differential pulse cathodic stripping voltammetry (DPCSV) and by hydride generation coupled to atomic absorption spectrometry (HG-AAS) or by hydride genera-

tion coupled to atomic fluorescence spectrometry (HG-AFS). 2.2.1. Electrochemical determination The determination of Se by DPCSV was made by the standard addition method, with a voltammetric analyzer (Princenton Applied Research (PAR), Model 264 A) in connection with the PAR Model 303A device hanging mercury drop electrode (HMDE). Data were displayed on a Rikadenki X-Y (t) recorder or on a personal computer. The experimental parameters used are summarized in Table 1. In these conditions, the detection limit calculated from the origin of the calibration curve, obtained with synthetic solutions was 60 ng L
1, for 120 s of preconcentration time. 2.2.2. Spectroscopy determinations In the HG-AAS method Se was determined by the standard addition method using a continuous flow hydride generator GBC model HG 3000, coupled to an atomic absorption spectrometer GBC model 905 AA, equipped with a Se GBC photron superlamp. The quantification of Se by HG-AFS was carried out by calibration curve, using a continuous flow hydride

Table 1 Instrumental and experimental conditions employed in the determination of selenium(IV) by electrochemical (DPCSV) and spectrometry techniques (HG-AAS and HG-AFS) Electrochemical technique

Spectroscopic techniques

DPCSV

HG-AAS

HG-AFS

Parameters

Conditions

Parameters

Conditions

Conditions

Supporting electrolyte (M) Sample volume (mL) Degassed (min) Deposition potential (Ed ) (V) Deposition time (td ) (s)

HCl 0.1 (5 mL)

Lamp

Se photron superlamp

Se photron superlamp

200–1000

Current lamp (mA)

18

25

8 (N2)
0.2 (Ag/AgCl/Cl) 120–240

196 2.2

196 6

2.1

3

7.5

6

Argon

Argon (0.3 mL min
1)

Rest period (s)

15

Potential scan rate (mV s
1) Pulse amplitude (mV) HMDE drop size LOD (mg L
1) Calibration range (mg L
1)

5

Wavelength (nm) 2 M HCl flow rate (mL min
1) NaBH4 0.6 %(m/V) (NaOH O.5%) flow rate (mL min
1) Sample flow rate (mL min
1) Carrier gas

25

Atomization

Quartz cell (Air–acetylene flame)

Hydrogen–Air flame (air 300 mL min
1)

LODa (mg L
1) Calibration range (mg L
1)

0.20 0.5–10

0.020 0.1–0.5

Medium 0.06 0.5–3.0

a The detection limit was calculated as LOD=3SD/m; where SD is the standard deviation of 10 measurements of a blank solution and m is the slope of the standard calibration curve.

118

I. De Gregori et al. / Water Research 36 (2002) 115–122

generator (PS Analytical 10.003), fitted with a Permapure drier membrane (PS Analytical system) in connection with an atomic fluorescence system (Excalibur, PS Analytical). The instruments were under computer control, using the PS Analytical software. The instrumental and experimental conditions employed in both spectrometry techniques are summarized in Table 1. All results are reported as mean7confidence limit at 95% of probability. The analytical techniques employed were assessed analyzing the certified reference material (CRM) San Joaquin soil, 2709 (NIST); Irish soil and white clover 402 (BCR), after an acid digestion procedure applied to these matrices (HNO3/H2O2/HClO4 mixture, at high pressure, in a closed PTF vessel, at 1701C).

were prepared daily by serial dilution.

3. Results and discussion 3.1. Quality control

2.2.3. Preconcentration step To demonstrate whether loss of Se compounds or changes in its oxidation states are produced during the concentration step two evaporation procedures were assayed with synthetic solutions, containing Se(VI), Se(IV) and selenomethionine, at variables ultratrace levels (ng L
1). Aliquots of these solutions were submitted to a slow evaporation on an open system, other aliquots were previously degassed with nitrogen and maintained at natural pH, and then concentrated on a rotavapor system, in vacuum, at 501C. To evaluate the total Se recovery all Se species present in other aliquots of synthetic solution were oxidized before the evaporation step, adding 2 mL L
1 of concentrated HNO3 and HClO4 and then submitted to the same evaporation procedures.

The application of quality assurance requires the analysis of certified reference materials (CRM). In order to confirm the validity of the analytical techniques employed in this study and not having rainwater with certified Se concentration, the accuracy of the analytical techniques was assessed by analyzing different CRM. The results obtained by the different techniques are summarized in Table 2. As can be seen, the experimental values are in agreement with the certified values. For best quality control, blank samples and rainwater samples spiked with different Se species were analyzed by the different techniques. Total Se was determined after oxidation and subsequent reduction of samples. The efficiency of these reactions was studied with synthetic mixtures containing 10 mg L
1 Se(VI), 7.5 mg L
1 Se(IV) and 2.5 mg L
1 selenomethionine. The recovery for total Se was 9875% (CV 4%), and for the redox Se species the recoveries ranged between 93 and 108% (average 97%). However, in spite of analytical performances, these techniques cannot be applied directly to rainwater analysis, where Se was present at concentration levels close to or below the detection limits, so it was necessary to include one preconcentration step to allow total Se and its redox species, until quantifiable levels.

2.3. Reagents

3.2. Preconcentration step

All solutions were prepared with ultrapure water (Nanopure purifying water system Barnstead, USA). HCl, HNO3, HClO4 acids and H2O2 were of suprapure quality (Merck or Baker instra analyze grade). A standard solution of Se(IV) 1000 mg L
1 was prepared from a titrisol solution Merck. The stock Se(VI) solution was prepared from sodium selenate (Sigma) and Se(II) solution from DL-Selenomethionine (Sigma), chosen as the organic form of Se. Solutions of lower concentration

Between the different methods applied to concentrate water samples, slow evaporation on an open system or on a rotavapor system, in vacuum and low temperature, are undoubtedly the most simple systems that can be used, due to their great capacity to reduce the total volume. As the aim of this work was not only to determine total Se in rainwater but also to carry out its redox speciation, the criterion considered to determine total Se was to avoid the loss of Se compounds that are

Table 2 Results obtained by the different techniques for the certified reference materials analyzed CRM

Selenium (mg g
1) Certified values

San Joaquin soil NIST 2709 Irish soil BCR White clover BCR 402

1.5770.08 5.970.6 6.7070.25

Experimental values HG-ASS

HG-AFS

DPCSV

1.6070.07 6.1870.09 6.770.4

1.5970.03 6.070.1 6.670.2

1.670.1 5.7870.05 6.870.4

119

I. De Gregori et al. / Water Research 36 (2002) 115–122

volatile, and when the speciation of Se is performed any change in the oxidation states of different Se species could occur. The results obtained show that both evaporation systems can be used to concentrate water samples when the aim of the analysis is the determination of total Se, the recovery of total Se in samples preconcentrated by both methods ranged between 95 and 107%. On the contrary, to carry out the redox speciation of Se only the preconcentration procedure on rotavapor system, in vacuum can be used. Taking into account the ultratrace levels of Se species present in the synthetic solutions, the recovery for all species can be considered quantitative (98–110%) independent of the analytical technique employed. When the synthetic solutions were slowly evaporated on an open system, Se(II) was quantitatively oxidized and 50–78% of Se(IV) was also oxidized to Se(VI). Therefore, this system cannot be used to concentrate Se solutions for redox speciation analysis. 3.3. Analysis of rainwater samples In rainwater samples collected from different sites during the first event in 1997, only total Se (and copper) concentrations were determined. These analyses were made to demonstrate whether the total concentrations of these elements were functions of the distance from the possible anthropogenic sources, especially from the copper industries. This behavior can be expected since in general, it has been observed that Se concentration exhibits reasonable geographic patterns, its observation is at remote sites rather than local sources. In Table 3 are listed, simultaneously, the sites where the rainwater was sampled, with the respective distances from the sources and the total volumes of water collected in the first 1997 rainfall. As can be seen, the

total volumes of waterfall were very different at the diverse sampling zones. The lowest volumes correspond to sites from Puchuncav!ı valley, a zone located near the Pacific coast; Casablanca presented intermediate values and the highest volumes were collected at Catemu valley, a zone located in the proximity of the Andes mountains. On the other hand, from the pH values measured all samples must be considered as acid rains, with the exception of rain from Nogales. Rainwater from Puchuncav!ı valley shows a clear decrease in acidity with distance from the smelter. These rainwater samples having pH from 4.3 to 4.7 can leach and dissolve heavy metals from the dust particles deposited on the surface soils, during the dry periods. Comparing the results obtained from the different zones, it can be remarked that significantly larger, but also varying more widely as a function of location, are the Se amounts deposited in Puchuncav!ı valley compared to those from Catemu valley. However, the Se amount present in rainwater from Casablanca was similar to or higher than those from Catemu valley, or from sites in Puchuncav!ı valley, located at long distances from the source. The Se levels at Casablanca appear higher than some of those where mining activities are developed, suggesting the presence of another unidentified anthropogenic source. The different behavior observed for the rainwater from both impacted zones can be explained by the actions taken at Chagres smelter (Catemu valley) in order to decrease the total emissions, especially the airborne particulate. Currently, at Chagres smelter the fumes and dusts emissions from the chimney are practically not observable. At Las Ventanas only in 1998 the emissions to the atmosphere from the smelter were drastically decreased. But at Las Ventanas, besides the smelter are also located a coal-fired electric power plant, other anthropogenic sources of contamination by toxic elements. Analysis of

Table 3 Total volume and some chemical characteristics of rainwater samples collected from different sites of Valpara!ıso region (V) valley, (C) city, (N) north, (S) south and (E) east Valley

Puchuncavi

Catemu

Casablanca

Site

La Greda Los Maitenes Puchuncav!ı-V Puchuncav!ı-C Nogales-S Nogales-N Catemu San Jos!e Sta. Margarita Casablanca-V Casablanca-C

Distance (km)

2 (E) 2.6 (E) 8 (NE) 9 (NE) 26 (NE) 28 (NE) 4 (N) 5.5 (NE) 7 (NE) F F

Volume (mL)

367 353 433 479 550 505 1440 1344 1328 1060 1040

pH

4.39 4.47 4.58 4.64 6.30 6.77 4.81 4.63 4.38 4.61 4.68

Selenium mg

mg m
2

0.4870.15 0.1470.04 0.1270.01 0.0970.01 0.0470.01 0.00870.001 0.0370.01 0.0107 0.002 0.01370.003 0.04770.003 0.1470.06

1373 3.970.9 3.170.3 2.470.3 1.170.3 0.2270.02 0.870.3 0.2870.04 0.3670.07 1.370.1 3.870.9

120

I. De Gregori et al. / Water Research 36 (2002) 115–122

rainfall near the smelter sites, especially at Puchuncav!ı valley, shows a considerable enrichment of Se. Furthermore, the amounts of these elements found in rainwater from the different sites clearly decrease exponentially with the distance from the source; a good correlation index (r ¼
0:9025) was calculated from the relation ln Se ðmg m
2 Þ ¼ ð2:170:4Þ2ð0:1170:02Þ distance ðkmÞ These results are in accordance with those published by Haygarth et al. [9], who found that Se deposition is heavily influenced by geographical proximity to an emission source, with highest levels associated with industrial zones. A similar pattern was observed for copper amounts present in rainwaters from the different sites of Puchuncav!ı valley: ln Cu ðmg m
2 Þ ¼ ð7:1070:03Þ2ð0:0970:02Þ distance ðkmÞ; r ¼
0:9322 But the Cu amounts in rainwaters were higher than Se amount; they ranged between 17497141 and 83719 mg m–2 (values for La Greda and Nogales, respectively). Both elements had the same behavior, the slopes of both relations are similar, the amounts of Se and Cu in rainwaters from Puchuncav!ı valley were significantly correlated: Selenium ðmg m22 Þ ¼ 0:01Cu ðmg m22 Þ20:43;

r ¼ 0:9670:

The fact that the amounts of both elements in rainwater present the same pattern of distribution as a function of the distance indicates that these elements arise from the same anthropogenic source, the smelter of copper ore and/or the thermoelectric power plant.

3.3.1. Selenium speciation The redox speciation of Se was performed in rainwater samples, collected in 1998 and 1999 years, in those

sites where the total Se found in the rainwater of 1997 was the highest: La Greda and Los Maitenes, from Puchuncav!ı valley. The rainfall volumes from all samplers were joined obtaining, for each sampling site, only one composite sample. The sampling dates, the samplers number put in each sampling site and the total volume collected in all samplers are listed in Table 4. The determination of the total Se concentrations was performed in rainwater aliquots acidified before the concentration step, and also in aliquots concentrated on the rotavapor system, that were subsequently acidified and digested on open systems. The redox speciation of Se was done in the other aliquots. The results obtained for total Se, Se(VI) and Se(IV) amounts are listed in Table 4. Fig. 1 shows the percentage of each redox Se species found in the rainwater samples. The results shown that the total Se amount per unit of area (m2) found in the rainfall during one event, clearly decreases with the elapsed time between sampling, especially for those realized during 1999. This behavior is most evident when the value for the rainwater collected in 1998 is compared with those sampled in 1999, at the same season of the year, even more considering that the total volume of rainfall in 1999 is approximately 6-fold higher. This fact can be due to the several actions taken by the smelter and electrorefinery plant ‘‘Las Ventanas’’, in order to decrease the total emissions from the smelter, especially the airborne particulate. An industrial plant to produce sulfuric acid, from the SO2 fumes, was installed at ‘‘Las Ventanas’’, at the end of 1998. As can be seen in Table 4, the sum of Se(VI)+Se(IV) amounts was always statistically similar to the total Se amounts. Furthermore, Se(II) was never detected in all rainwater analyzed, which can be explained by the high temperature and oxidizing conditions present in the

Table 4 Total Se, Se(IV) and Se(VI) amounts in rainwater samples collected at different dates from Puchuncavi valley Sampling date

Sampler number

Total volume (mL)

Total Se (mg m
2)

Se(VI) (mg m
2)

Se(IV) (mg m
2)

10/09/1998 16/07/1999 3–6/09/1999

10 6 12

2325 2380 15450

971 4.870.3 0.7570.05

7.170.1 2.570.2 0.3070.04

2.070.1 2.370.1 0.4270.06

Fig. 1. Percentage of redox selenium species present in rainwater samples, from Puchuncav!ı and La Greda, collected (A) 10/09/1998, (B) 16/07/1999 and (C) 3–6/09/1999.

I. De Gregori et al. / Water Research 36 (2002) 115–122

smelter process or by the reactivity of gaseous atmospheric Se(II) species. H2Se is a chemical species that is always absent from any atmospheric samples analyzed, since this substance is easily oxidized. It has been described that Se(II) species react with OH and NO3 radicals and with ozone (O3) present in the atmosphere [27,28]. The oxidation states (IV) and (VI) of Se found in the rainwater analyzed in this work are in accordance with those described by Cutter and Church [29], who studied the oxidation state of Se in rainwater from the Atlantic Coast of USA. The fact that Se in rainwater analyzed was present as Se(VI) and Se(IV) can help to understand the behavior of this element in agricultural soils from the same sites located near the copper industries, where it was observed that Se, in contrast to Cu, is not accumulated in these soils [30]. Se which is not fixed by the vegetation, due to its high mobility, can percolate to the lower soils profile.

Acknowledgements The authors gratefully acknowledge the financial support of Fondecyt (project 1/97/1289), the VRIEA de la Universidad Cato! lica de Valpara!ıso (project 125.728) and the Program ECOS-CONICYT (Scientific Co-operation between France and Chile), Action C96E04. M.G. Lobos also thanks CONICYT for the fellowships conceded.

References [1] N"eve J, Henry M, Peretz A, Mareschi JP. L’importance nutritionnelle du s!el!enium. Cah Nutri Di!et 1987; XXII(2):145–63. [2] Dubois F, Belleville F. S!el!enium: ro# le physiologique et int!er"et en pathologie humaine. Pathol Biol 1988;36:1017– 25. [3] Baruthio F. Toxicologie des e´le´ments trace essentiels. In: Chappuis P, editor. Les oligoe´le´ments en me´dicine et biologie. Paris: Editions Me´dicales Internationales, Lavoisier Tec & Doc, 1991. Ch 8, 261–8. [4] Oldfield JE. Efficacy of various forms of selenium for livestock: a review. Bulletin of Selenium–Tellurium Development Association (special issue) 1997. p. 1–8. [5] Ohlendorf HM. Bioaccumulation and effects of selenium in wildlife. In: Jacobs LW, editor. Selenium in agriculture and the environment, special publication 23. SSSA, Madison WI, 1989. p. 133–77. [6] Rosenfield J, Beath OA. Selenium. Geobotanyc, biochemistry, toxicity and nutrition. New York: Academic Press, 1964. [7] Mun* oz Olivas R, Donard OFX, Camara C, Quevauviller P. Analytical techniques applied to the speciation of selenium in environmental matrices. Anal Chim Acta 1994;286:357–70.

121

[8] Agnithotre R, Chauk S, Mahuli S, Fan LS. Selenium removal using Ca-based sorbents: reaction kinetics. Environ Sci Technol 1998;32:1841–6. [9] Haygarth PM. Global importance and global cycling of selenium. In: Frankenberger WT, Benson S, editors. Selenium in the Environment. New York: Marcel Dekker, 1994. p. 1–27. [10] Haygarth PM. Determination of gaseous and particulate selenium over a rural and grassland in the UK. Atmos Environ 1994;28:3655–63. [11] Nu. renberg HW, Valenta P, Nguyen VN. The wet deposition of heavy metals from the atmosphere in the Federal Republic of Germany. International Conference on Heavy Metals in the Environment, vol. 1, Heidelberg, 1983. p. 115–25. [12] Lemly AD, Finger SE, Nelson MK. Sources and impacts of irrigation drainwater contaminants in arid wetlands. Environ Toxicol Chem 1993;12:2265–79. [13] Masscheleyn PH, Delaune RD, Patrick Jr. WH. Arsenic, selenium chemistry as affected by sediment redox potential, pH. J Environ Qual 1991;20:522–7. [14] Adeloju SB, Bond AM, Briggs MH, Hughes HC. Stripping voltammetric determination of Se in biological materials by direct calibration. Anal Chem 1993;55:2076–83. [15] Batley GE. Differentia lFpulse polarographic determination of selenium species in contaminated waters. Anal Chim Acta 1986;187:109–16. [16] Potin-Gautier M, S!eby F, Astruc M. Interference of humic substances on the speciation analysis of inorganic selenium in water and soils. Fresenius J Anal Chem 1994; 351:443–8. [17] Elleouet C, Quentel F, Madec C. Determination of inorganic and organic selenium species in natural waters by cathodic stripping voltammetry. Water Res 1996; 30(4):909–14. [18] Allabashi R, Rendl J, Grasserbauer M. Validation of three atomic absorption spectrometric methods for the determination of seleniumFa comparative evaluation of performance characteristics. Fresenius J Anal Chem 1997;357:1024–8. [19] Haygarth PM, Rowland AP, Stu. rup S, Jones KC. Comparison of instrumental methods for the determination of total selenium in environmental samples. Analyst 1993;118:1303–8. [20] Workman SM, Soltanpour PN. Importance of prereducing selenium(VI) to selenium(IV) and decomposing organic matter in soils extracts prior to determination of selenium using hydride generation. Soil Sci Soc Am J 1980; 44:1331–3. [21] Kos V, Veber M, Hudnik V. Determination of selenium in soil by hydride generation AAS. Fresenius J Anal Chem 1998;360:225–9. [22] Brunori C, de la Calle Guntin* as MB, Morabito R. Optimization of the reduction of Se(VI) to Se(IV) in microwave oven. Fresenius J Anal Chem 1998;360:26–30. [23] Miner!ıa Chilena. ENAMI en cifras, 1982–1988. Miner!ıa Chilena 1989;9(101):47–59. [24] Malman S, Sabatini F, Geisse G. El Trasfondo socioecon!omico del conflicto ambiental de Puchuncav!ı. Ambiente y Desarrollo 1995;XI(4):49–58.

122

I. De Gregori et al. / Water Research 36 (2002) 115–122

[25] Gonz!alez S, Bergovist E. El Impacto de emisiones de gases y otros productos desde chimeneas de fundiciones de minerales sobre las actividades agropecuarias. Ambiente y Desarrollo 1986;II(3):35. [26] En Ambiente Enami Divisio! n Corporativa Ambiental. An* o 2, No. 3–4, 1998. [27] Atkinson R, Aschmann SM, Hasegawa D, Thompson– Eagle ET, Frankenberger Jr WT. Kinetics of the atmospherically important reactions of dimethyl selenide. Environ Sci Technol 1990;24:1326–32.

[28] Haygarth PM, Jones KC, Harrison AF. Selenium cycling through agricultural grassland in the UK: budgesting the role of the atmosphere. Sci Total Environ 1991;103: 89–111. [29] Cutter GA, Church TM. Selenium in western Atlantic precipitation. Nature 1986;322:720–2. [30] De Gregori I, Lobos G, Lobos S, Pinochet H, PotinGautier M, Astruc M. Comparative study of copper and selenium in agricultural ecosystem from Valpara!ıso region, Chile. Environ Technol 2000;21:307–16.