Speciation of Tl(III) and Tl(I) in hair samples by dispersive liquid–liquid microextraction based on solidification of floating organic droplet prior to flame atomic absorption spectrometry determination

2.1 Instrumentation

A SensAA GBC atomic absorption spectrometer (Dandenong, Australia) equipped with deuterium background correction and air–acetylene burner was used for absorbance measurements according to the instrument instruction. Tl hollow cathode lamp was used as light source at wavelength of 276.8 nm. The operating parameters of the element were set according to the manufacturer’s recommendation. The acetylene flow rate and burner height were adjusted in order to obtain the maximum absorbance signal, while aspirating the analyte solution. A Metrohm 692 pH meter (Herisau, Switzerland) was used for pH measurements. A Centurion scientific centrifuge model K241R (West Sussex, UK) was used to accelerate the phase separation.

2.2 Reagents and solutions

Deionized water was used throughout the experimental work. Stock solutions of Tl(I) and Tl(III) at a concentration of 1000.0 μg mL⁻¹ were prepared by dissolving appropriate amounts of TlNO₃ and Tl(NO₃)₃·3H₂O (Merck, Darmstadt, Germany) in 0.5 mol L⁻¹ HNO₃ solution. Working standard solutions were obtained by appropriate dilution of the stock standard solution. A solution of 1.2 mg mL⁻¹ 1-(2-pyridylazo)-2-naphthol (PAN) (Merck) in ethanol was prepared. The solutions of alkali metal salt (1% w v⁻¹) and various metal salts (0.1% w v⁻¹) were used to study the interference of anions and cations, respectively. The laboratory glassware was kept overnight in a 1.4 mol L⁻¹ HNO₃ solution and before using; all of the glasswares were washed with deionized water and dried.

2.3 DLLME-SFO procedure

All standards and samples were prepared for analysis according to the following procedure. Fifty milliliters of an aqueous solution containing 10.0 μg Tl(III) were placed in a test tube. To each test tube, 3.0 mL of 0.1 mol L⁻¹ phosphate buffer (pH 10.0), 0.2 mL of 1.5 mg mL⁻¹ PAN and 3.0 mL of 10% (w v⁻¹) NaCl were added. Then, 2.5 mL of ethanol containing 100.0 μL of 1-dodecanol (DOD) was injected rapidly into the sample solution. As a result, oil-in-water emulsions of DOD in water were formed. After centrifuging at 3000 rpm for 6.0 min, the organic solvent droplet due to lower density was floated on the surface of aqueous solution. Then, the test tube was put into an ice bath for 5.0 min; as a result, the floated solvent was solidified because of the low melting point (24 °C). Then, the solidified solvent was transferred to a conical test tube. The solidified solvent was washed by iced deionized water, and water in the vial was drawn out by a syringe. The solidified organic solvent was melted quickly in room temperature and 1.0 mL of 0.5 mol L⁻¹ HNO₃ in methanol was added to it prior to the determination of Tl by FAAS.

After the oxidation of Tl(I) to Tl(III) (Vogel, 2000, London), total Tl was determined as Tl(III) by the above described method and concentration of Tl(I) is calculated by the difference between total Tl and Tl(III) concentrations. For this purpose, 35.0 mL of an aqueous solution containing Tl(I) and Tl(III) was taken in a volumetric flask. Then, 5.0 mL of Ce(IV) 0.01 mol L⁻¹ and 5.0 mL of Mn(II) 0.01 mol L⁻¹ were added to it. The solution was heated at 70.0 °C for 30 min and cooled to room temperature and the final volume was completed up to 50.0 mL with deionized water.

2.4 Sample preparation

The hair sample has been prepared by using the procedure given in the literature (Dadfarnia et al., 2007). At first, hair samples were rinsed with acetone, chloroform and deionized water, and then dried at 60.0 °C. Ten grams of each sample hair was weighted accurately into a beaker and 15.0 mL concentrated HNO₃ was added to it. The content of each beaker was heated on a hot plate (initially at 100 °C for 15 min and then at 150 °C for 15 min) after dissolution. Then the mixture was heated at 200 °C to dryness and a white residue was obtained. After this, 10.0 mL of 0.1 mol L⁻¹ HNO₃ was added to the beaker and heated at 100 °C for several minutes. After this, the solution was cooled to room temperature and the final volume was adjusted to 50.0 mL with deionized water and Tl(III) was determined. To parallel, 10.0 g of the hair sample was prepared (Dadfarnia et al., 2007), then 2.5 mL of Ce(IV) and Mn(II) 0.01 mol L⁻¹ were added to it. After this, 10.0 mL of 0.1 mol L⁻¹ HNO₃ was added to the beaker and heated at 100 °C for several minutes. The solution was heated for 30.0 min at 70 °C. Finally, the solution was cooled to room temperature and the volume was adjusted to 50.0 mL with deionized water. The total Tl was determined according to the proposed procedure.

One certified reference alloy was also analyzed. Approximately 0.5 g of BCR, no. 288 was weighed accurately into a teflon cup and dissolved in concentrated HNO₃ (∼5.0 mL) with heating on a water bath. Then, 15.0 mL deionized water was added to it and oxidation of Tl(I) to Tl (III) has been performed by using the given procedure in Section 2.3. Finally, the solution was cooled and diluted to 100.0 mL with deionized water in a calibrated flask.

3.1 Selection of type and volume of disperser solvent

Mixing of the disperser solvent with extraction solvent and sample solution is significant for choosing the disperser solvent. For this purpose, acetone, ethanol and methanol were selected as candidates of disperser solvent. For the selection of appropriate disperser solvent, a series of sample solutions containing Tl(III)–PAN complex were extracted using 2.5 mL of each mentioned disperser solvent containing 100.0 μL of DOD. The recoveries for acetone, ethanol and methanol were 98.4%, 97.6% and 99.3%, respectively. The higher recovery was obtained using methanol but ethanol has low toxicity and cost, therefore, it was selected as disperser solvent.

To obtain the optimized volume of ethanol, various experiments were performed using different volumes of ethanol (1.0–5.0 mL) containing 100.0 μL of DOD. The results showed that the recovery increased up to 2.0 mL and decreased in the higher volume of 3.5 mL. This trend can be probably contributed to that at low volume, ethanol could not disperse DOD properly and cloudy solution was not formed completely. Reversely, at high volumes, the solubility of the complex in water increased. Finally, 2.5 mL ethanol was chosen as the optimum volume.

3.2 Selection of type and volume of extraction solvent

The extraction solvent is important in the optimization of the DLLME-SFO process. For the selection of this solvent, its low solubility in water, high affinity to analytes, and lower melting point than room temperature and lower density than water are important. In this work, 1-undecanol and 1-dodecanol were investigated as extraction solvents, and their recovery were studied. The recoveries for 1-undecanol and DOD were 97.3% and 98.7%, respectively, and also 1-undecanol is more expensive than DOD. Thus DOD was selected as extraction solvent in the next experiments.

The effect of DOD volume on the recovery of Tl was also investigated over the range of 40.0–200.0 μL with 2.5 mL ethanol. The results showed that the recovery increased to 75.0 μL and remained constant. Based on these observations, 100.0 μL was selected for further experiments.

3.3 Influence of pH

The effect of pH on the recovery of Tl(III) ions was studied in the range of 3.0–12.0. The pH of sample solutions were adjusted by the addition of HNO₃ and NaOH solutions. The results are given in Fig. 1 and show that the highest recovery was obtained at a pH range of 9.5–12.0. Additional experiments on the volume of buffer showed that 2.0–8.0 mL of buffer solution (pH 10.0) gives the best results. Therefore, 3.0 mL of 0.1 mol L⁻¹ phosphate buffer solution was used in all subsequent experiments.

Close

Figure 1

Effect of the solution pH on the DLLME-SFO extraction of Tl(III). Conditions: sample, 50.0 mL Tl(III) 0.2 μg mL⁻¹; buffer, 3.0 mL; PAN 1.2 mg mL⁻¹, 0.2 mL; NaCl 10% (w v⁻¹), 3.0 mL and DOD, 100.0 μL.

Export to PPT

3.4 Effect of the PAN concentration

Tl(I) cannot form the complex with PAN. The influence of PAN concentration on the recovery of Tl(III) ions was studied in the concentration range of 0.1–2.5 mg mL⁻¹. For this purpose, various experiments were performed using 0.2 mL of PAN. The results showed that the recovery increased to 0.5 mg mL⁻¹ and then remained constant. Therefore, a PAN concentration of 1.5 mg mL⁻¹ was selected in all further experiments to account for the other extractable species that potentially interfere with the assaying of Tl.

3.5 Effect of the ionic strength

In the extraction, the solubility of many analytes in aqueous solutions decreases with increasing ionic strength due to salting out effect (Mohammadi et al., 2009a,b). The influence of the ionic strength on the extraction of Tl(III) was studied with sodium chloride solution in the range of 0.05–1.0 g. The rest of the experimental conditions were kept constant. The recovery increased up to 0.2 g and for higher amounts it was constant. Therefore, 0.3 g of NaCl was used in all further experiments.

3.6 Effect of extraction time

In order to have a good precision, sensitivity and speed, it is necessary to select an extraction time that guarantees the achievement of equilibrium between aqueous and organic phase. In DLLME, extraction time is defined as the time interval between injection of the mixture of disperser and extraction solvents and starting of centrifuge. The effect of extraction time on the recovery was studied by varying the extraction time from 0.5 to 15.0 min while the other experimental conditions were kept constant. The results showed that the extraction time had no influence on the recovery of Tl(III) ions. This can be contributed to the increase in surface area between extraction solvent and aqueous phase after the formation of a cloudy solution. This behavior has also been seen in DLLME (Mohammadi et al., 2009b; Ma et al., 2009; Rezaei et al., 2008).

3.7 Calibration, precision and detection limit

Linearity, repeatability and detection limit were investigated under the optimized experimental conditions. For a sample volume of 50.0 mL, the calibration curve exhibited a linearity over the range of 6.0–900.0 ng mL⁻¹with a correlation coefficient of 0.9987 (A = 0.685C + 0.002, where A is the absorbance value and C is the concentration of Tl(III) (μg mL⁻¹)). Seven replicate determinations of 0.2 μg mL⁻¹ Tl(III) gave a mean absorbance of 0.134 with a relative standard deviation of 2.5%. Preconcentration of 50.0 mL of sample permitted the detection limit (3S_b) of 2.1 ng mL⁻¹ for Tl(III) ion.

The enhancement factor, defined as the ratio of the slope of the calibration curve for the DLLME-SFO procedure to that obtained without preconcentration (Ma et al., 2010), was 42.7.

3.8 Effect of diverse ions

In order to examine the effect of common coexisting ions on the extraction of Tl(III) different metal cations and anions were added individually to a sample solution containing 10.0 μg of analyte (Table 1). The tolerance limit was set as an amount of the interfering ion required to cause ±5% error. As can be seen in Table 1, most of the examined cations and anions did not interfere with the extraction and determination of Tl(III) even when present in 100-fold excess over Tl(III). This shows that the proposed method is suitable for the determination of Tl(III) in real samples.

3.9 Accuracy of the method

Accuracy of the proposed method has been evaluated for the determination of Tl(III) ion in BCR no. 288. The amount of Tl in BCR was found to be 2.26 ± 0.11 μg g⁻¹. It was found that there is no significant difference between the result obtained by the DLLME-SFO procedure and the certified result (2.3 ± 0.1 μg g⁻¹).

3.10 Analysis of real samples

The DLLME-SFO procedure has been applied for the determination of Tl content in hair samples. Furthermore, to evaluate the validity of the DLLME-SFO procedure for speciation of Tl, recovery of different amounts of T1(I) and Tl(III) ions added to hair samples were also studied. The results are given in Table 2. The results indicate that the added Tl(I) and Tl(III) ions can be quantitatively recovered from hair samples by the proposed procedure. These results demonstrate the applicability of the proposed procedure for Tl determination in hair samples.