Determination of bioavailable soluble arsenic and phosphates in mine tailings by spectrophotometric Sequential Injection Analysis

Talanta 78 (2009) 1069–1076

Contents lists available at ScienceDirect

Talanta journal homepage: www.elsevier.com/locate/talanta

Determination of bioavailable soluble arsenic and phosphates in mine tailings by spectrophotometric Sequential Injection Analysis ∗ ˜ Belén E. Ramírez Cordero, María P. Canizares-Macías Facultad de Química, Departamento de Química Analítica, Universidad Nacional Autónoma de México, México D.F., 04510, Mexico

a r t i c l e

i n f o

Article history: Received 15 October 2008 Received in revised form 13 January 2009 Accepted 13 January 2009 Available online 23 January 2009 Keywords: Sequential Injection Analysis Arsenic Speciation Phosphates Molybdenum blue reaction Mine tailings

a b s t r a c t By using a simple Sequential Injection Analysis (SIA) manifold and in base to the kinetic reaction of the molybdenum with As(V) and P(V) was possible to determine As(III), As(V) and P(V) in simple, binary and ternary samples. The activation energies for the reaction between molybdenum and As(V) and P(V) were of 70.90 kJ mol−1 and of 19.02 kJ mol−1 , respectively, therefore it was possible to determine both analytes in mixtures by using different reaction temperature. When the analyses were carried out at room temperature, only the P(V) supplied analytical signal; with increased temperature, the kinetics of reaction for As(V) also increased, and a signal was obtained, being 55 ◦ C the optimum temperature. In order to determine As(III), it was oxidized into As(V) with KIO3, and the reaction was carried out in the same way as for As(V). To resolve mixtures, an equations system from six calibration curves with different sequences of SIA at different temperature was performed. The lineal ranges were between 0.5 ␮g mL−1 and 10 ␮g mL−1 with a repeatability and reproducibility between 0.7% and 5.2% and detection limits between 0.36 ␮g mL−1 and 0.58 ␮g mL−1 . In binary mixtures of P(V)/As(V) the recoveries were close to 100% for both analytes at ratios lesser than 10:1. For As(V)/As(III) ratios between 1:1 and 5:1 the recoveries were ranged between 85% and 95%. The method was applied in mine tailings and in arsenopyrite. The results showed that the soluble arsenic was found oxidized as As(V). These results were compared with those obtained by atomic absorption spectrometry and both proved to be very close. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Tailings are solid residues produced during primary operations in the separation and concentration of minerals, therefore, environmental concerns make tailing control extremely important. The metals and metalates’ toxicity in tailings depends mostly on the extractable portion and not on the total concentration, due to the extractable portion causes environmental damage because of their mobility both water and soils. Also, the extractable portion is considered an indirect measurement for bioavailable fraction. Tailings contain residual metallic sulfides as arsenopyrite (FeAsS), galene (PbS), asphalerite (ZnS) and calcopyrite (CuFeS2 ), and their oxidation brings out acid drain and high concentrations of potentially toxic elements such As or Pb [1–4]. The transportation of these elements posses a serious environmental problem as they can pollute the soil and sediments, as well as superficial and underground water [5,6]. Therefore, it is necessary to carry out tests to identify the presence of potentially toxic elements under environmental conditions. Arsenic is a real cancer-causing agent, which can also damage human genetics [7–10]. The most toxic species of arsenic are arse-

∗ Corresponding author. Tel.: +52 55 56 22 37 88; fax: +52 55 56 22 37 88. ˜ E-mail address: [email protected] (M.P. Canizares-Macías). 0039-9140/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2009.01.024

nates and arsenites, and among them, As(III) is more toxic than As(V). Arsenic solubility depends principally on Mn, Fe, and Ca concentrations in water because of the formation non-soluble salts such as Mn3 (AsO4 )2 , FeAsO4 and Ca3 (AsO4 )2 . Therefore, the ions that compete with these metals change the mobility of the arsenic, as it is the case of phosphates, which at 0.003 mol L−1 avoid arsenic precipitation. Consequently, sometimes water exposed to arsenic with high phosphates concentrations has higher amounts of dissolved arsenic than water with fewer phosphates [11]. The most used analytical methods for the arsenic determination are: electrothermal atomic absorption (AA) [12], inductively couple plasma coupled with atomic fluorescence spectrometry (ICP–AES) [13,14] and ICP coupled with mass spectrometry (ICP–MS) [15,16]. However, although these techniques are precise and accurate, their cost can be enormous. Several colorimetric methods to determine arsenic are also used. Some of them are connected with the Gutzeit method, which is based on the generation of arsenine gas by reduction of arsenic under acid conditions, followed by the addition of zinc powder and the quantification of arsine gas on paper impregnated with mercuric bromide [17–19]. At present, the molybdenum blue method is the colorimetric method most used, and it is based on the reaction between the molybdate in acid medium and the As(V) and/or the

1070

B.E.R. Cordero, M.P. Ca˜ nizares-Macías / Talanta 78 (2009) 1069–1076

P(V), in order to form the corresponding molybdenum heteropoly acid, which is yellow, has a type cage structure and is able to be reduced or oxidized [20]. As(III) does not form complexes with molybdenum, so that arsenic can only be quantified like As(V). To determine As(V), As(III) and P(V) in a mixture, some methods have been described, such as: Johnson and Pilson [21,22] determined the three analytes in sea water and fresh water with a good precision, but colour development took up to 90 min. In a more recent paper, Dhar et al. [23] decreased the reaction time by raising the concentration of potassium antimonyl tartrate and by changing KIO3 concentration from 0.25 mmol L−1 (the original conditions of Johnson and Pilson), to 2 mmol L−1 , which caused a more efficient oxidation for As(III). In 2002, Dasgupta et al. [24] determined As(III) and As(V) in drinking water, based on the same reaction of formation of arsenomolybdate, and by using a continuous flow manifold and an anion exchange resin column. They also used KBrO3 to oxidize As(III) into As(V) and l-cystine to reduce As(V) into As(III), so, it was possible to speciate both analytes. Other papers determining P(V) by using molybdenum reaction with flow injection analysis have been reported [25–27], and some of them have succeeded in decreasing the interference from silicates either by using a gas segmented flow system [28] or by a Sequential Injection Analysis (SIA) system [29]. In this paper, 0.25% oxalic acid was used to avoid the molybdosilic acid formation. Mexico has been a mining country since the Spanish colonial age and the wastes from the mines have been collected since then. Besides, Mexico produces 20% of the amount of arsenic in the whole world. Although storing and treatment methods of mine tailings have been actually improved, their study is, however, of an absolute necessity. In this paper, an analysis by sequential injection method to quantify As(V) and P(V) in mine tailings based on its kinetic reaction to form the molybdenum blue complex at different temperatures was carried out. The SIA system also allowed to determine As(III). 2. Experimental 2.1. Instruments and apparatus To construct the SIA manifold the following component were used: an automatic selection valve with six channels (SCIVEX), a peristaltic pump Gilson Minipuls 3 (France) which was controlled using software made in the laboratory, and Teflon and Tygon tubing of 0.5 mm I.D. One UV-VIS spectrophotometer Cary 3 (Varian, Sydney Australia) equipped with a flow cell of 118 ␮L inner volume measured the absorbance. A Lab-line thermostat water bath with an error range of ±0.2 ◦ C was used to control temperature during the arsenic reaction. The temperature was also measured with a Fisherbrand mercury thermometer (from −10 ◦ C to 260 ◦ C). An orbital stirrer (Lab-line) and a Millipore filtration system were used in the samples treatment. pH and potential from sample extracts were measured with a Metrohm 620 pH-meter and a Cole Parmer 05669-20 potentiometer. 2.2. Reagents and solutions All the used reagents were analytically graded, the solutions were prepared using distilled water and all glass materials were cleaned with phosphate-free detergent. A 500 mg L−1 P(V) stock solution was prepared from KH2 PO4 (99.5% of purity, Sigma). The stock solutions for As(V) and As(III) were prepared from Na2 HAsO4 7H2 O (99% of purity, Sigma) and NaAsO2 (100% purity, Sigma) at 500 mg L−1 each. Standard solutions were prepared from stock solutions to suitable concentrations.

A 6% (w/v) ascorbic acid (Sigma) solution, as well as a 0.4% (w/v) tetrahydrate ammonium molybdate (Sigma) solution prepared in 0.15 M H2 SO4 (Baker), were used to carry out the derivation reaction. To validate the method (in order to achieve reproducibility, repeatability and accuracy) certificated standards were used: phosphorous 1000 ± 3 ␮g mL−1 in 0.05% HNO3 , As(V) 1000 ± 3 ␮g mL−1 in 2% NaOH and Tr bromine, and As(III) 1000 ± 3 ␮g mL−1 in 2% HCl. All the standards were supplied by High Purity Standards. A 0.5% KIO3 (Baker) solution was used as As(III) to As(V) oxidation agent. 2.3. Sample preparation The proposed method was applied to mine tailings collected from three mining zones in Mexico (northern, central and southern), as well as to arsenopyrite samples. The extraction method was carried out in accordance with the ASTM D3987-85 method [30], used for lixiviation of tailings with water in equilibrium with atmospheric CO2 (pH 5.5). 2.3.1. Collection and preservation Tailings samples were collected in accordance with the Norma Oficial Mexicana NOM-141-SEMARNAT-2003 [31]. Each sample weighed approximately 2 kg, and each of one was made up of five different sub-samples. They were stored in plastic containers with no substance added to preserve them. Once the samples arrived at the laboratory, they were dried at room temperature and were stored in plastic bags until their analysis. 2.3.2. Determination of pH and potential 5 g of sample were weighed and 100 mL of distilled water were added to them. The mixture was stirred for 10 min, and then the pH and the potential were measured. 2.3.3. Extraction of arsenates from tailings 5 g of sample were weighed and placed into an Erlenmeyer flask, then 100 mL of distilled water were added. The mixture was stirred in an open flask, for 18 h at 200 rpm using an orbital stirrer [30]. Then, the mixture was filtered through a 45 ␮m porous size Nylon membrane by using a Millipore system, and the soluble arsenic was measured using the SIA system. 2.3.4. Extraction of arsenates from arsenopyrite 0.1 g of sample was weighed and 4 mL of a HCl/HNO3 (1:3) solution were added. The mixture was left to rest for 8 h. Then the sample was filtered and the extract was dissolved at 100 mL with distilled water. Later, a 1:100 dilution was carried out to measure the soluble arsenic by using the SIA system. The extraction of soluble compounds was carried out in duplicate for all samples. 2.4. Sequential injection procedure Fig. 1 shows the used SIA manifold to determine P(V), As(V) and As(III) in simple, binary and ternary solutions. The analysis sequence in Table 1 shows the sequence for single solutions. When these sequences are combined, it is also possible to resolve binary and ternary samples. The carrier was distilled water and the flowrate was 20 ␮L s−1 for steps A–L; for step M the flow-rate was changed into 15 ␮L s−1 in order to improve the oxidation reaction of the As(III). 2.4.1. Determination of P(V), As(V) and As(III) in single solutions Three calibration equations for each analyte were obtained by using the sequences showed in Table 1:

B.E.R. Cordero, M.P. Ca˜ nizares-Macías / Talanta 78 (2009) 1069–1076

1071

using Eq. (1). The As(V) concentration was calculated after the second analysis, by substituting Eqs. (1) and (4) in Eq. (5) and resolving CAs(V) : CAs(V) =

A55◦ C − CP(V) m4 − b4 − b2 m2

(6)

2.4.3. Determination of As(V) and As(III) in binary solutions Two analyses were carried out to resolve these mixtures: one by using sequence 2 to know the As(V) concentration, and another by using sequence 3. Consequently, with sequence 3 the dispersion changed and a new calibration curve was constructed for As(V) by using sequence 3: Fig. 1. SIA manifold for the determination of As(V), As(III) and P(V). HC: holding coil, RC: reactor coil, SV: selection valve, PSV: principal selection valve with six ways, PP: peristaltic pump, carrier: distilled water.

P(V) at room temperature (steps from A to D; sequence 1) A1 = CP(V) m1 + b1

(7)

In the binary solutions, the obtained signal with this sequence was the sum of absorbance given by the original As(V) and by the oxidized As(III): Aox = A3 + A7

(1)

where A is the absorbance signal, m is the slope and b is the intersection in zero for all the equations. As(V) (steps from E to H; sequence 2) A2 = CP(V) m2 + b2

A7 = AAs(V) m7 + b7

(8)

The As(III) concentration was calculated by substituting Eqs. (3) and (7) in Eq. (8) and resolving CAs(III) : CAs(III) =

Aox − CAs(V) m7 − b7 − b3 m3

(9)

(2)

As(III) (steps from I to M; sequence 3) A3 = CP(V) m3 + b3

(3)

2.4.2. Determination of P(V) and As(V) in binary solutions We need a third calibration curve of P(V) at 55 ◦ C by using sequence 2: A4 = CP(V) m4 + b4

(4)

Therefore, in binary solutions, when sequence 2 was used, the obtained absorbance (A55 ◦ C ) was the given absorbance by As(V) and by P(V) at 55 ◦ C. A55◦ C = A2 + A4

(5)

So, in order to resolve binary samples, two analyses were carried out: first using sequence 1, and second, using sequence 2. The P(V) concentration was directly obtained after the first analysis and

2.4.4. Determination of P(V), As(V) and As(III) in ternary solutions In this case, three analyses were carried out: one by using sequence 1, another by using sequence 2 and a third one by using sequence 3. It was necessary to build a new calibration curve for P(V) using sequence 3: AP(V) = AP(V) m10 + b10

(10)

Therefore, in the ternary solutions, the obtained signal with sequence 3 belongs to the three analytes: ATotal = A3 + A7 + A10

(11)

P(V) and As(V) concentrations were obtained with their corresponding equations (Eqs. (1) and (6), respectively). To determine the As(III) concentration, Eqs. (3), (7) and (6) were substituted in Eq. (11), and CAs(III) was resolved: CAs(III) =

ATotal − CP(V) m10 − b10 − CAs(V) m7 − b7 − b3 m3

(12)

Table 1 Sequence of analysis for P(V), As(V) and As(III) in single, binary and ternary samples by using the manifold shown in Fig. 1. Action

Time (s)a

Description

Determination of P(V): sequence 1 A 1 B 2 C 4 D 6

Aspiration Aspiration Aspiration Delivery

5 5 5 120

Ascorbic acid toward HC P(V) sample/standard toward HC (to select with SV1 ) Ammonium molybdate toward HC Delivery to the RC1 and detector (to select with SV2 )

Determination As(V): sequence 2 E 1 F 2 G 4 H 5

Aspiration Aspiration Aspiration Delivery

5 5 5 120

Ascorbic acid toward HC As(V) sample/standard toward HC (to select with SV1 ) Ammonium molybdate toward HC Delivery to the RC2 (55 ◦ C) and detector (to select with VS2 )

Determination of As(III): sequence 3 I 1 J 2 K 3 L 4 M 5

Aspiration Aspiration Aspiration Aspiration Delivery

5 5 2 5 130b

Ascorbic acid toward HC As(III) sample/standard toward HC (To select with SV1 ) KIO3 toward HC Ammonium molybdate toward HC Delivery to the RC2 (55 ◦ C) and detector (to select with SV2 )

Step

Selection valve position (PSV)

PSV: principal selection valve. a Flow-rate: 20 ␮L s−1 . b Flow-rate: 15 ␮L s−1 .

1072

B.E.R. Cordero, M.P. Ca˜ nizares-Macías / Talanta 78 (2009) 1069–1076

The detection limits were calculated by using the equation: YB + 3SB where YB is the blank signal and SB is the standard deviation from blank. SB was calculated from calibration graphs for each analyte at different conditions (temperature and SIA sequence). SB was calculated from the equation:



SB =

(yi − yˆ i )2

n−2

This equation utilizes the y-residuals, yi − yˆ i , where the yˆ i values are the points on the calculated regression line corresponding to the individual X-values. The calculated intercept for each calibration graph was used as an estimate of YB [32]. 3. Results and discussion First, the SIA manifold was optimized for phosphates. To optimize the reagent concentrations, the aspiration sequence as well as the injection volume, a 6 ␮g mL−1 P(V) solution was used. The optimum conditions for P(V) were the same as for As(V). However, for this analysis it was necessary to optimize the reaction temperature, because on flow conditions phosphates only form the complex with ammonium molybdate at room temperature. 3.1. Optimization of SIA manifold 3.1.1. Determination of P(V) and As(V) 3.1.1.1. Study of the reagents concentration. Ascorbic acid: The concentration used was 6% (w/v), in accordance with that mentioned by Linares et al. [33]. Ammonium molybdate: A range between 0.1% and 1% (w/v) was evaluated. All the tested solutions were dissolved in 0.6 M sulfuric acid. The optimum selected concentration was 0.4%. Reaction medium: Different acids were also evaluated as a reaction medium, such as: nitric acid, sulfuric acid and hydrochloric acid. The analytical signal was similar for the three of them, but the blank signal when HCl or HNO3 were used was higher than H2 SO4 and, therefore, the last one was selected to perform the analysis. 3.1.1.2. Study of the aspiration sequence. The first step of the molybdenum blue reaction is the formation of a yellow complex (heteropoly acid), and the second one is the reduction of this to Mo(V). With the aim of increasing the overlapping zone between the sample and the reagents, six aspiration sequences were studied: (a) (b) (c) (d) (e) (f)

Ammonium molybdate–sample/standard–ascorbic acid. Ascorbic acid–sample/standard–ammonium molybdate. Sample/standard–ammonium molybdate–ascorbic acid. Sample/standard–ascorbic acid–ammonium molybdate. Ascorbic acid–ammonium molybdate–sample/reagent. Ammonium molybdate–ascorbic acid–sample/reagent.

In sequence (a) double peaks were obtained although the RC length was increased; with the other sequences, only one peak was obtained. In sequence (b) the peaks were higher, so this sequence was selected as optimum for the analysis. In sequences (a and b), the sample was between the two reagents, allowing higher overlapping of the dispersed zones. But in these tests the order of the reagents influenced the reaction product: when the ascorbic acid was first aspirated (optimum sequence (b)), there was more contact time between it and the formed heteropoly acid, which was obtained from the reaction between the sample and the ammonium molybdate, causing a higher signal; on the contrary, when

Fig. 2. Study of the sulfuric acid concentration at 0.15 M and 0.6 M for As(V) and P(V), respectively. P(V): (*) 0.15 M, () 0.6 M; As(V): () 0.15 M, and () 0.6 M.

the ammonium molybdate was first aspirated (sequence (a)) double peaks were obtained possibly because of the refraction index gradient. 3.1.1.3. Study of the aspiration time and the flow-rate. A range between 8 ␮L s−1 and 33 ␮L s−1 to study the aspiration flow-rate for reagents and sample and the flow-rate for the complex formation was evaluated. The optimum flow-rate found was 20 ␮L s−1 . The aspiration time was established at 5 s equivalent to 100 ␮L. 3.1.1.4. Evaluation of the holding coil (HC) and the reaction coil (RC) lengths. The HC length was long enough to avoid the contamination of reagents and samples, and to assure the maximum overlapping between reagents and sample. The optimum length was 200 cm. In the case of RC the higher signal was obtained at 300 cm. 3.1.1.5. Optimization of the sulfuric acid concentration. Linares et al. [33] established that a high H2 SO4 concentration inhibited the formation of the arsenomolybdate but, in accordance with our results, the H2 SO4 was also the best acid for the reaction, therefore, we studied the effect of the sulfuric acid concentration. A concentration range between 0.1 M and 1 M was evaluated. Fig. 2 shows the obtained results when two concentrations of sulfuric acid (0.15 M and 0.6 M) are used to form the heteropolyacids of As(V) or of P(V) at different concentrations. The results show that, at lower sulfuric acid concentrations, the analytical signal for As(V) is higher and it is lower than P(V), which improves the determination of As(V). Therefore, a 0.15 M sulfuric acid concentration was selected as optimum, in order to carry out the analysis of As(V) in presence of P(V). 3.1.1.6. Evaluation of the temperature to quantify As(V) and P(V). When arsenates solutions were aspirated with the same concentrations as phosphates at room temperature, the reaction did not take place. Therefore, a study to evaluate the optimum reaction temperature, the rate velocity and the activation energy for As(V) and P(V) was performed. Fig. 3A shows the As(V) plots at several concentrations and at different temperatures. Fig. 3B shows the behaviour of P(V). To calculate the apparent reaction rate (K’) some concentrations absorbance at different temperatures for each analyte were measured. The measurement was carried out by using the SIA system and the reaction time was established in the maximum of the SIA peak corresponding to 60 s. The apparent activation energy (E’) was calculated after evaluating 1/T (◦ K) vs log K’. The reaction rate of each analyte at different temperatures and the activation energy are shown in Table 2. The results show that the reaction rate increases when the temperature also increases for both analytes. In the case of P(V) the kinetic of reaction is similar from room temperature to 45 ◦ C increasing to double at 55 ◦ C. On the other hand, for As(V) the K’ is much increased from 40 ◦ C. Moreover, at room temperature the complex is not formed at least in the time that is evaluated. Besides, the same table shows that for As(V) form the blue complex needs a higher activation energy than P(V), so P(V) can be quantified at

B.E.R. Cordero, M.P. Ca˜ nizares-Macías / Talanta 78 (2009) 1069–1076

1073

3.1.2.1. Sequence of aspiration. Although the number of reagents was increased for this analysis, we studied only three sequences, as the KIO3 and the sample were aspirated one by one to make the oxidation easier. Ammonium molybdate was aspirated at the end, in accordance with the optimization for As(V). The studied sequences were: (a) Sample–KIO3 –ascorbic acid–ammonium molybdate. (b) KIO3 –sample–ascorbic acid–ammonium molybdate. (c) Ascorbic acid–sample–KIO3 –ammonium molybdate. As the highest analytical signal was in sequence (c), this one was selected to carry out the other analyses (Table 1, sequence 3). 3.1.2.2. KIO3 concentration. The evaluated range of concentrations was from 0.2% to 1.3%. The optimum concentration was 0.5%. The aspiration volume was also evaluated, being the optimum value 40 ␮L.

Fig. 3. (A) Temperature evaluation at different concentrations of As(V). () 0.15 ␮g mL−1 , () 1.5 ␮g mL−1 , () 3 ␮g mL−1 , (×) 7.5 ␮g mL−1 , and (*) 15 ␮g mL−1 . (B) Temperature evaluation at different concentrations of P(V). () 0.5 ␮g mL−1 , ()1.5 ␮g mL−1 , () 3 ␮g mL−1 , (×) 5 ␮g mL−1 , (*) 10 ␮g mL−1 , (•) 15 ␮g mL−1 . Table 2 Apparent reaction rate (K’) and apparent activation energy (E’) for As(V) and P(V). Tempertature (◦ C)

K’ (s−1 ) P(V)

As(V) −4

4.45 × 10 4.9 × 10−4 n.d. 5.1 × 10−4 9.7 × 10−4 19.02 kJ mol−1

25 35 40 45 55 Activation energy (E’)

– 5 × 10−5 6.8 × 10−5 7.2 × 10−5 2.7 × 10−4 70.90 kJ mol−1

n.d: not determined.

room temperature with no interference of As(V) and it is showed the influence of the temperature on the reaction kinetic. These graphics also show that the signal for As(V) begins to increase from 35 ◦ C but, although higher temperatures than 55 ◦ C gave higher peaks, these were not reproducible because bubbles into SIA system appeared. Consequently, 55 ◦ C was selected as the temperature analysis for As(V). 3.1.2. Determination of As(III) To determine As(III) using the molybdenum blue method, it is necessary to oxidize it into As(V). With this aim, the concentration of KIO3 , the aspiration SIA sequences, as well as the injection volume, were optimized.

3.1.2.3. Aspiration and reaction flow-rates. The aspiration flow-rate was of 20 ␮L s−1 and the reaction flow-rate (delivery step) was of 15 ␮L s−1 to increase the reaction time with the KIO3 . 3.2. Calibration graphs Six calibration curves were made by using the sequences of Table 1 to obtain Eqs. (1–4), (7) and (10). The slope and intersection in zero, the correlation coefficient, the lineal range, and the detection limit are shown in Table 3. With these equations data, and substituting them in Eqs. (5), (6), (8), (9), (11) or (12), the concentrations of As(V) and of As(III) in binary and ternary samples were obtained. P(V) was always obtained by using Eq. (1) and sequence 1. The results show that the determination of P(V) at 55 ◦ C has a sensibility 2.5 times higher than P(V) at room temperature, and 1.25 times higher than As(V) at 55 ◦ C. However, the linear ranges for the six curves are practically the same. When sequence 3 was used, the slope was lower, as the system had one reagent more. 3.3. Validation of the method Certificated standards of P(V), As(V) and As(III) were used to validate the method. Its precision, within-laboratory reproducibility as well as its repeatability were evaluated in a single experimental set-up in duplicate. 3.3.1. Simple solutions The experiments were carried out by using two concentrations of P(V), As(V) and As(III) (2 mg L−1 and 6 mg L−1 ), and using the suitable SIA sequence. To carry out the study, two daily measurements for each concentration were performed for seven days. Table 4 shows the results of reproducibility and repeatability, as well as the

Table 3 Characteristics of the method for the determination of P(V), As(V) and As(III) in single standard using different SIA sequences. Equationsa (1) (2) (3) (4) (7) (10) a b c d

Analyte

Sequenceb

Slope

Intercept

Linear range (␮g mL−1 )

r

D.L. (␮g mL−1 )c

P(V) As(V) As(III) P(V) As(V) P(V)

1 2 3 2 3 3

m1 = 0.033 ± 0.001d m2 = 0.0649 ± 0.002 m3 = 0.0127 ± 0.001 m4 = 0.0817 ± 0.004 m7 = 0.0255 ± 0.002 m10 = 0.0497 ± 0.003

b1 = −0.002 ± 0.007d b2 = 0.0085 ± 0.013 b3 = 0.0069 ± 0.004 b4 = 0.0133 ± 0.020 b7 = 0.0287 ± 0.010 b10 = 0.0065 ± 0.021

0.5–10 0.5–10 0.6–10 0.6–10 1–10 1–10

0.9995 0.9995 0.9990 0.9992 0.9990 0.9990

0.37 0.36 0.52 0.46 0.55 0.58

In accordance with the Sections 2.4.1–2.4.4. See Table 1. Detection limit. Confidence limits.

1074

B.E.R. Cordero, M.P. Ca˜ nizares-Macías / Talanta 78 (2009) 1069–1076

Table 4 Results of the accuracy and precision in the determination of As(V), P(V) and As(III) in single, binary and ternary solutions. Equationa

Analyte

Concentration (␮g mL−1 )

Reproducibility (r.s.d.)

Repeatability (r.s.d.)

Accuracy (%)

1

P(V)

2 6

2.2 3.3

4.4 3.9

101.4 102

2

As(V)

2 6

2.8 2.0

3.5 1.7

106.2 101.3

3

As(III)

2 6

4.6 4.2

5.2 4.5

105.2 98.9

4

P(V) at 55 ◦ C

2 6

1.9 0.7

2.3 1.4

100.4 102.1

7

As(V)b

2 6

2.8 2.6

2.6 3.2

99.8 104.6

10

P(V)b

2 6

1.6 1.3

2.9 2.1

97.3 102.4

1, 6

P(V)/As(V)

1.5/4.5 4.5/1.5

1.8/1.64 1.3/4.03

5.0/2.4 1.7/5.3

102.6/101.5 100.4/102.1

2, 9

As(V)/As(III)

1.5/4.5 4.5/1.5

1.9/1.74 2.2/3.59

2.3/2.8 2.8/4.9

101.2/98.8 98.6/96.2

1, 6, 12

P(V)/As(V)/As(III)

1.5/1.5/4.5 1.5/4.5/1.5

1.3/1.8/2.1 1.5/1.6/2.6

1.8/2.8/2.5 1.8/2.1/3.2

101.0/99.6/103.1 100.2/101.0/96.8

a b

Used equation to determine the concentration of each analyte. Using sequence 3.

accuracy, in determining the three analytes when found in simple samples. 3.3.2. Binary and ternary solutions In Table 4 are also shown the results of the study for two rates of P(V)/As(V), As(V)/As(III) and P(V)/As(V)/As(III). These results show a bigger error when the As(III) concentration is lower than As(V), because the slope for As(III) is lower than for As(V). Even though, the results also show that the precision and the accuracy are adequate for all the analyses. Another study on recoveries was also performed in order to evaluate the interference between the analytes. Some ratios of P(V), As(V) and As(III) in binary and ternary mixtures were measured. The ratios were selected in accordance with the lineal range from the different calibration graphs. So, the total concentration of the mixtures could not be higher than 9 ␮g mL−1 and the minimum concentration used for any analyte had to be higher than 1 ␮g mL−1 because at concentrations lower than 1 ␮g mL−1 the confidence limits from calibrations graphs are higher. In Table 5, these results Table 5 Recoveries study in binary and ternary mixtures. Mixture

Rate

Recovery (%)

P(V)/As(V)

1:1 1:3 1:5 3:1 5:1

As(V)/As(III)

1:1 1:3 1:5 3:1 5:1

P(V)/As(V)/As(III)

1:1:1 1:1:2 1:1:5 3:3:1 1:5:1 5:1:1

P(V)

As(V)

102.4 ± 1.0 104.2 ± 0.5 95.7 ± 0.8 101.8 ± 0.6 98.8 ± 0.3

101.0 ± 0.4 102.0 ± 0.8 100.2 ± 0.5 106.6 ± 0.3 108.6 ± 0.2

106.1 ± 0.2 107.1 ± 0.6 102.1 ± 1.4 99.2 ± 0.3 105.1 ± 0.5 97.6 ± 0.6

As(III)

101.2 ± 1.1 99.5 ± 0.5 96.2 ± 0.2 102.2 ± 0.3 97.2 ± 0.1

94.3 ± 0.2 104.8 ± 0.6 101.6 ± 0.1 90.5 ± 1.0 85.6 ± 0.6

102.5 ± 0.2 104.9 ± 0.6 98.5 ± 0.2 104.2 ± 1.1 102.3 ± 0.5 98.3 ± 0.3

93.3 ± 0.3 107.6 ± 0.6 108.4 ± 0.3 91.6 ± 0.2 85.3 ± 0.3 95.1 ± 0.3

show that when As(V)/As(III) rate is 3:1 or 5:1, the recoveries for As(III) are lower than As(V) and decrease down to 85%. Therefore, although the quantification of As(III) in binary or ternary samples is acceptable, it improves when the concentration is higher than As(V). On the other hand, the determination of P(V) and of As(V) in binary and ternary samples has a higher precision. Therefore, the SIA manifold allows the analysis of the three analytes in mixtures with a sample throughput of 27 determinations h−1 . A few calibration graphs during the tests of validation were built to evaluate their stability. Standards of P(V), As(III) and As(V) at different concentrations were evaluated every week to assure the precision and accurate from the graphs. The values were always between 96% and 103% for simple solutions and between 87% and 110% for binary and ternary mixtures for ratios between 1:1 and 5:1. The slope and the intercept values from calibration equations were always into the precision range of them. 3.4. Analysis of samples Samples of mine tailings and arsenopyrite were analyzed to measure bioavailable soluble arsenic. Five tailings from three mining zones in Mexico were analyzed: two samples from the southern zone (S), one sample from the northern (N) and two samples from the central one (C). All the samples were analyzed in duplicate, and 1.5 ␮g mL−1 As(V) and P(V) and 2 ␮g mL−1 As(III) were added for the recovery studies. Table 6 shows the concentrations found for soluble arsenic in the analyzed tailings samples, as well as for the arsenopyrite. The pH, the colour and the redox potential of the studied tailings are also reported and, also in the same table, the obtained results by atomic absorption are shown. As silicates are also interferences in the molybdenum blue reaction, some tests adding 0.25% oxalic acid [29] to the extracts were performed. The signals were the same when oxalic acid was not added because the silicates were not dissolved when the extraction took place. Tailings from northern and central zones were brown and grey, and the redox potential was close to 350 mV (reported with regards to standard hydrogen electrode, HNE) with a pH between 5 and

B.E.R. Cordero, M.P. Ca˜ nizares-Macías / Talanta 78 (2009) 1069–1076

1075

Table 6 Determination of soluble arsenic in tailings and arsenopyrite by the proposed SIA method. Sample

Colour

pH

E (mV)a

N1 N2 C1 C2 C3 C4 S1 S2 S3 S4 1d 2d

Brown Brown Grey Grey Brown Brown Yellow Yellow Yellow Yellow – –

8.0 7.6 5.3 5.6 6.2 6.5 2.4 2.7 3 2.9 – –

377 396 402 415 312 320 730 753 702 702 – –

a b c d e

AAc (mg kg−1 )

Concentration in the extractsb (␮g mL−1 )

Concentration in the tailings (mg kg−1 )

Recoveries (%)

P(V)

As(V)

P(V)

As(V)

P(V)

As(V)

As(III)

0.58 ± 0.02 0.46 ± 0.03 1.02 ± 0.13 0.96 ± 0.05 0.71 ± 0.07 0.59 ± 0.05 0.97 ± 0.04 0.96 ± 0.06 1.91 ± 0.10 1.99 ± 0.15

11.6 9.2 20.4 19.2 14.2 11.8 19.4 19.2 38.2 39.8 – –

– – – – – – 45.08 41.80 149.20 139.60 28.50e 28.40e

96.3 92.1 98.6 101.3 100.3 104.6 99.4 97.6 103.2 98.2 100.6 101.8

88.6 90.5 100.6 101.2 100.0 99.8 105.9 101.3 99.1 102.2 100.3 98.2

95.2 92.4 100.3 102.6 104.5 98.4 96.8 97.5 92.1 93.4 96.1 103.5

– – – – – – 44.20 43.00 146.00 140.00 30.00 29.00

Potential reported with regards to standard hydrogen electrode (ENH). Proposed method. The results are the average and standard deviation for three determinations by extract. Concentration in the extracts by absorption atomic: Varian SpectrAA 110, empty cathode lamp of arsenic, lengthwave: 193.7 nm, flame: air/acetylene. Arsenopyrite. Concentration (w/w).

8. These results indicated that tailings were fresh and, therefore, the soluble arsenic concentration should have been low. This was confirmed with the obtained results, where the soluble arsenic concentration was very low indeed. It was possible because the oxidation process was incipient, (because of the freshness of the tailings) and, besides, because they were stabilized by a high percentage of calcite. The tailings from the southern zone were old and their colour was yellow. The pH was lower than 3, and the redox potential close to 700 mV (HNE). Therefore, they were oxidized and the soluble arsenic concentration had turned higher. In oxidized tailings the oxidation of the arsenal pyrite is made easier from As(0) to As(V) [34]: 4FeAsS + 13O2 + 6H2 O ↔ 4Fe2+ + 4SO2− 4 + 4H3 AsO4 Authors as Foster et al. [35] analyzed mine tailings using an X-ray absorption near edge structure (XANES) spectroscopy, and found that the dominating oxidation state is the As(V) in oxidized mine tailings. The results show that soluble arsenic from extracts was found as As(V) and not as As(III), because the AAS results for total arsenic from extracts were similar to As(V), which also shows that the proposed method is suitable enough to determine bioavailable soluble arsenic in tailings. On the other hand, the extraction method also makes the oxidation from arsenic to As(V) easier, because the extraction process was carried out at room temperature in open conditions for 24 h. Moreover, when the extraction was finished, HNO3 was added to the extracts to avoid the arsenic precipitation, and therefore the oxidation of the As(III) was improved. With regard to the arsenopyrite mineral, the results showed an approximate arsenic value of 30%, which is very close to the percentage of arsenic in these minerals. In this case, the mineral was digested and, therefore, all the arsenic was oxidized into As(V). Besides, the S3 and S4 tailings were considered “hazardous residues” in accordance with the Official Mexican Norms NOM-052SEMARNAT-1993 [36] and the NOM-141-SEMARNAT-2003 [31], which establish that if the extract contains ≥5 mg L−1 of arsenic, the material may be considered a “hazardous residue due to its toxicity for the environment”. 4. Conclusions The obtained results of soluble arsenic in tailings using the proposed method were very similar to those obtained by AAS.

Therefore, it is possible to state that the molybdenum blue SIA method is an excellent choice to quantify soluble arsenic without P(V) interference. On the other hand, the obtained results about the validation of the SIA method proved to be precise and accurate and, besides, it was made possible to automatize all the analysis with a simple manifold. The values found for the apparent activation energies show the relation between the temperature and the complex formation for As(V) and P(V), which allows the quantification of these in mixtures when different temperatures are used. Although it is necessary to build some calibration curves to determine both three analytes with the proposed SIA method, it is stable and reproducible and the graphs are allowed to be used for a long time, besides the analysis is fast with a sample throughput of 27 h−1 . The method is not only suitable for other kinds of samples, where only one of the three analytes is measured, but also for more complex samples where one, two or the three analytes are present. Moreover, the proposed method is a good option for determining bioavailable arsenic in mine tailings, and at the same time minimizing pollution, due to its low levels of generated waste in comparison to other methods.

Acknowledgements The Facultad de Química of the Universidad Nacional Autónoma de México and the “Programa de Apoyo a Proyectos de Investigación e Innovación Tecnológica, PAPIIT” (grant no. IN209707) of the Dirección General de Asuntos del Personal Académico are gratefully acknowledged for the financial support.

References [1] Z. Lin, Sci. Total Environ. 198 (1997) 13. [2] R.H. Johnson, D.W. Blowes, W.D. Robertson, J.L. Jambor, J. Contam. Hydrol. 41 (2000) 49. [3] C.B. Rousel, H.A. Fernández, J. Environ. Qual. 29 (2000) 182. [4] M.C. Moncur, C.J. Ptacek, D.W. Blowes, J.L. Jambor, Appl. Geochem. 20 (2005) 639. [5] J.G. Bain, D.W. Blowes, W.D. Robertson, E.O. Frind, J. Contam. Hydrol. 41 (2000) 23. ˜ [6] M.A. Armienta, G. Villasenor, R. Rodriguez, L.K. Onley, H. Mango, Environ. Geol. 40 (2001) 571. [7] H.G. Sieler, H. Sigel, Handbook of Toxicity of Inorganic Compounds, Marcel Dekker Inc., USA, 1988, p. 79. [8] D.Q.O. Hung, O. Nekrassova, R.G. Compton, Talanta 64 (2004) 269. [9] Organización Mundial de la Salud. Guías para la calidad del agua potable, 2◦ edición, vol. 1, Ginebra, 1995, pp. 37–43. [10] B.K. Mandal, K.T. Suzuki, Talanta 58 (2002) 201.

1076

B.E.R. Cordero, M.P. Ca˜ nizares-Macías / Talanta 78 (2009) 1069–1076

[11] M.P. Landeros Albores, Estudio de Arsénico y Selenio en Aguas Subterráneas, Tesis. [12] D.Q. Hung, O. Nekrassova, R.G. Compton, Talanta 64 (2004) 269. [13] K.J. Lamble, S.J. Hill, Anal. Chim. Acta 334 (1996) 261. [14] N.V. Semenova, F.M. Bauzá de Mirabó, R. Forteza, V. Cerdá, Anal. Chim. Acta 412 (2000) 169. [15] X. Wei, C.A. Brockhoff-Schwegel, J.T. Creed, JAAS 16 (2001) 12. [16] I. Havezov, A. Datcheva, J. Rendl, Michrochim. Acta 119 (1995) 147. [17] H. Gutzeit, Pharmaz. Zeitung 24 (1879) 263. [18] P. Lohmann, Pharmaz. Zeitung 36 (1891) 756. [19] D. Kinniburg, W. Kosmus, Talanta 58 (2002) 165. [20] M.T. Pope, Heteropoly and Isopoly Oxometalates, Spring-Verlag, Berlin, 1983. [21] D.L. Johnson, Environ. Sci. Technol. 5 (1971) 411. [22] D.L. Johnson, M.E.Q. Pilson, Anal. Chim. Acta 58 (1972) 289. [23] R.K. Dhar, Y. Zheng, J. Rubrnstone, A. Van Geen, Anal. Chim. Acta 526 (2004) 203. [24] P.K. Dasgupta, H. Huang, G. Zhang, G. Cobb, Talanta 58 (2002) 153. [25] J.L. Manzoori, A. Miyazaki, H. Tao, Analyst 115 (1990) 1055. [26] R.L. Benson, Y.B. Truong, I.D. McKelvie, B.T. Hart, Water Res. 30 (1996) 1959.

[27] K. Grudpan, P. Ampan, Y. Udnan, S. Jayasvati, S Lapanatnoppakhun, J. Jakmunee, G.D. Christian, J. Ruzicka, Talanta 58 (2002) 1319. [28] J.Z. Zhang, J. Charles Fischer, P.B. Ortner, Talanta 49 (1999) 293. [29] C.X. Galhardo, J.C. Masini, Anal. Chim. Acta 417 (2000) 191. [30] ASTM D 3987-85 American Society for Testing and Materials, Standard Test Method for Shake Extraction of Solid Waste with Water, ASTM, West Conshohocken, PA, 1992, 4 p. [31] Norma Oficial Mexicana NOM-141-SEMARNAT-2003. Requisitos para la caracterización del sitio, proyecto, construcción, operación y postoperación de presas de jales, Secretaria de Medio ambiente y Recursos Naturales, México, 2003. [32] J. Miller, J. Miller, Estadística y quimiometría para química analítica, 4a edición, ˜ 2002, 78–155. Prentice Hall, Madrid, Espana, [33] P. Linares, M.D. Luque de Castro, M. Valcárcel, Anal. Chem. 58 (1986) 120. [34] G. Morin, G. Calas, Elements 2 (2006) 97. [35] A.L. Foster, E.B. Gordon Jr., T.N. Tingle, G. Parks, Am. Minerol. 83 (1998) 553. [36] Norma Oficial Mexicana NOM-052-SEMARNAT-1993. Características de los residuos peligrosos y el listado de los mismos y los límites que hacen un residuo peligroso por su toxicidad al ambiente. Characteristics of dangerous residues and listed of them and the limits of toxicity to the environmental. Secretaría de Medio Ambiente y Recursos Naturales, México, 1993.