Determination of trace amounts of hexavalent

0 downloads 0 Views 757KB Size Report
Z. Bahadir a, V.N. Bulut b, M. Hidalgo c, M. Soylak d, E. Marguí c,⁎ a Department of Chemistry, Giresun University, Giresun, Turkey b Macka Vocational School, ...
Spectrochimica Acta Part B 107 (2015) 170–177

Contents lists available at ScienceDirect

Spectrochimica Acta Part B journal homepage: www.elsevier.com/locate/sab

Determination of trace amounts of hexavalent chromium in drinking waters by dispersive microsolid-phase extraction using modified multiwalled carbon nanotubes combined with total reflection X-ray fluorescence spectrometry Z. Bahadir a, V.N. Bulut b, M. Hidalgo c, M. Soylak d, E. Marguí c,⁎ a

Department of Chemistry, Giresun University, Giresun, Turkey Macka Vocational School, Karadeniz Technical University, Macka, Trabzon, Turkey c Department of Chemistry, Faculty of Sciences, University of Girona, Girona, Spain d Department of Chemistry, Faculty of Sciences, Erciyes University, Kayseri, Turkey b

a r t i c l e

i n f o

Article history: Received 18 December 2014 Accepted 12 March 2015 Available online 21 March 2015 Keywords: TXRF DMSPE MWCNTs Hexavalent chromium Drinking water

a b s t r a c t A methodology based on the combination of dispersive microsolid-phase extraction (DMSPE) with total reflection X-ray fluorescence (TXRF) spectrometry is proposed for the determination of hexavalent chromium in drinking waters. Multiwalled carbon nanotubes (MWCNTs) modified with the anionic exchanger tricaprylmethylammonium chloride (Aliquat 336) were used as solid sorbents. After the sorption process of Cr(VI) on the modified MWCNTs, the aqueous sample was separated by centrifugation and the loaded MWCNTs were suspended using a small volume of an internal standard solution and analyzed directly by a benchtop TXRF spectrometer, without any elution step. Parameters affecting the extraction process (pH and volume of the aqueous sample, amount of MWCNTs, extraction time) and TXRF analysis (volume of internal standard, volume of deposited suspension on the reflector, drying mode, and instrumental parameters) have been carefully evaluated to test the real capability of the developed methodology for the determination of Cr(VI) at trace levels. Using the best analytical conditions, it was found that the minimum Cr(VI) content that can be detected in an aqueous solution was 3 μg L−1. This value is almost 20 times lower than the maximum hexavalent chromium content permissible in drinking waters, according to the World Health Organization (WHO). Recoveries for spiked tap and mineral water samples were, in most cases, in the range of 101–108% which demonstrates the suitability of the TXRF methodology for monitoring Cr(VI) at trace levels in drinking water samples. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The determination of heavy metals ions at trace levels is very significant for control of environmental protection [1]. Chromium is one of the environmental contaminants for natural waters resulting from industrial activities such as metal manufacturing and electroplating [2]. Chromium exists predominantly in two oxidation forms in the environment, which are Cr(III) and Cr(VI). While chromium (III) is an essential micronutrient for living organisms, the hexavalent species of chromium have strong mutagenic properties [3,4]. For this reason, various international environmental agencies have presented tight policy with regard to the limit of hexavalent chromium concentration in drinking water. According to the World Health Organization (WHO) and the US Environmental Protection Agency (EPA), the maximum permissible content ⁎ Corresponding author. E-mail address: [email protected] (E. Marguí).

http://dx.doi.org/10.1016/j.sab.2015.03.010 0584-8547/© 2015 Elsevier B.V. All rights reserved.

of Cr(VI) are 0.05 μg mL−1 and 0.1 μg mL−1, respectively [5]. Therefore, owing to different properties of Cr(III) and Cr(VI), it is of significance the development of sensitive methods for chromium speciation in water samples. The determination of Cr(VI) species can be performed using ion chromatography with post-column derivatization of the hexavalent chromium with diphenylcarbazide and the detection of the colored complex at 530 nm. Another possibility is the use of atomic spectroscopic techniques such as flame atomic absorption spectrometry (FAAS), electrothermal atomic absorption spectrometry (ETAAS), inductively coupled plasma–mass spectrometry (ICP–MS) and total reflection Xray fluorescence spectrometry (TXRF) after a suitable sample treatment to extract and isolate Cr(VI) from water samples prior the analysis [6–8]. TXRF is a well-established analytical technique for multielement determination in various sample types, especially liquids and powdered or microsamples [9]. To perform analysis under total-reflection conditions samples must be provided as thin films, limiting the total dissolved salt

Z. Bahadir et al. / Spectrochimica Acta Part B 107 (2015) 170–177

concentration in the sample solution. For liquid samples, this is done by depositing 5–50 μL of sample on a reflective carrier with a subsequent drying of the drop. Even at such low sample amount the detection limits are mostly at μg L−1 level [10]. Last decades, most of the published TXRF analyses were performed using large-scaled instruments with highpower X-ray tubes, demanding water-cooling systems and liquidnitrogen cooled detectors. However, in recent years, the development and commercialization of benchtop TXRF instrumentation, which offer extreme simplicity of operation in a low-cost compact design, have substantially increased its application in industry as well as in research activities [11]. At present, there is an increasing interest in the development of environmentally friendly analytical preconcentration procedures according to the rules of green analytical chemistry [12,13]. Therefore, liquid-phase microextraction (LPME) and dispersive solid phase microextraction (DMSPE) have become the most valuable alternative techniques to classical liquid–liquid extraction (LLE) and solid phase extraction (SPE) procedures for preconcentration of metal ions in environmental waters [14,15]. In both cases, the use of a microanalytical technique is needed after the sample treatment process since only a few microliters of liquid or solid are obtained to perform the measurement. For that reason, ETAAS technique is usually employed [16]. Taking into account the microanalytical capabilities of TXRF, the use of this technique in combination with these microextraction procedures could be also a suitable analytical strategy for the determination of metal and inorganic species in water samples. In this sense, we recently published a couple of publications highlighting the potential use of LPME in combination with TXRF for the determination of trace amounts of cadmium and inorganic antimony in water samples [17,18]. In the case of DMSPE systems, a small amount of a solid sorbent (μg or mg range) is dispersed in the analyzed solution to extract the target analytes. The nature and properties of the solid sorbent are of prime importance and carbon nanotubes (CNTs) have desired characteristics to be used as solid sorbents because of their large specific surface area and their well-defined structure at atomic scale [19,20]. A large number of contributions reporting the use of such materials (raw of modified with organic groups) for the preconcentration of various metals, metalloids and inorganic species have been published last years. For instance, Zhang used renewed MWCNTs (Multiwalled carbon nanotubes) with thiosemicarbazide (TB) for extraction of Cd(II), Cu(II), and Pb(II) from water samples [21]. In a similar way, Samia et al. modified MWCNTs with 8-hydroxyquinoline (8-HQ) for removal of copper(II), lead(II), cadmium(II) and zinc(II) from aqueous solutions [22]. It is interesting to note that, in addition to the benefits of CTNs named above, the very small particle size CNTs made these solid sorbents also appropriate to prepare a representative suspension to be analyzed by TXRF, thus avoiding the elution step required when using other atomic spectroscopic techniques [23]. Nevertheless, in order to extract and isolate Cr(VI) species from water samples, a functionalization of the MWCNTs using an appropriate extractant is needed. In a former study, we demonstrated the potential use of the commercial anionic exchanger (Aliquat 336) for the selective adsorption of Cr(VI) over Cr(III) in preconcentration procedures based on the use of activated thin layers [24]. In the present paper we took benefit of this knowledge to develop a quite simple DMSPE method for preconcentration of chromium (VI) in drinking water samples using modified multiwalled carbon nanotubes with Aliquat 336. A similar approach was studied by some authors for hexavalent chromium determination by WDXRF instrumentation [25]. However, in that study, after the extraction step the direct analysis of the resulting suspension was not possible and an additional treatment of the sample (pellet preparation) was mandatory before the WDXRF analysis. In the present contribution we developed a DMSPE to be easily used in combination with TXRF without any sample treatment procedure after the exaction procedure. For that, parameters affecting the extraction process (pH and volume of the aqueous sample, amount of

171

MWCNTs, extraction time) and TXRF analysis (volume of internal standard, volume of deposited suspension on the reflector, drying mode, and instrumental parameters) have been carefully evaluated to test the real capability of the developed methodology for the determination of Cr(VI) at trace levels in tap and mineral waters. Additionally, we also demonstrated that purification of the MWCNTS before functionalization does not provide real benefits in terms of impurity removal, extractive capacity of chromium and repeatability, as previously reported by the aforementioned authors [25]. This fact reduces considerably the sample preparation time compared with other existing methodologies. 2. Methods 2.1. Reagents, materials and solutions Cr(VI) and Cr(III) stock solutions of 1000 mg L−1 were prepared by dissolving K2Cr2O7 (Panreac, Spain) and CrCI3·6H2O (Merck, Darmstadt, Germany) in water. Diluted Cr VI) and Cr(III) working standard solutions were prepared by adding water and the corresponding amount of 0.1 M hydrochloric acid (32%, Fluka, Switzerland) or 0.1 M sodium hydroxide (98%, Panreac, Spain) to adjust the pH. A standard solution of 2 mg L−1 of Y was used for internal standardization of samples. It was prepared by dilution of a commercial stock solution of Y of 1000 mg L−1 (Romil Pure Chemistry, Cambridge, UK) using 2 M nitric acid (diluted from the commercial nitric acid Suprapure, Merck, Darmstadt, Germany), a diluted solution of 1% of the commercial non-ionic detergent Triton® X-100 (purchased from Sigma-Aldrich (Spain)) or water. Multiwalled carbon nanotubes (MWCNTs) with diameters from 6 to 9 nm and lengths of approximately 5 μm were purchased from Sigma-Aldrich (Spain). The extractant, tricaprylmethylammonium chloride (Aliquat 336) was purchased from Sigma-Aldrich and methanol was used as the organic solvent (Suprapure, Merck, Darmstadt, Germany). All solutions were prepared using analytical reagent grade chemicals and ultrapure water, from a MilliQ Plus system (Millipore Corp., Bedford). For TXRF analysis, quartz glass discs with a diameter of 30 mm and a thickness of 3 mm ± 0.1 mm were used as sample holders. 2.2. Apparatus, instruments and operating conditions TXRF analysis was performed using a benchtop spectrometer S2 PICOFOX™ (Bruker AXS Microanalysis GmbH, Berlin, Germany). The spectrometer specifications and operating conditions used are summarized in Table 1. This instrument uses an air-cooled low-power X-ray tube and a Peltier cooled silicon drift detector and thus, no cooling media and gas consumption are required. The evaluation of TXRF spectra and calculation of analytes net peak areas was performed using the Table 1 Instrumental parameters and measuring conditions. S2 PICOFOX TXRF benchtop spectrometer X-Ray tube W Rating 50 kV, 1 mA (maximum power 50 W) Optics Multilayer Ni/C, 17.5 keV, 80% reflectivity Detector Si drift detector, 10 mm2, b160 eV resolution Mn–Kα Working environment Air Sample station Cassette changer for 25 samples Measurement time 2000 s Agilent 7500c ICP–MS spectrometer RF power 1500 W Plasma gas flow rate 15 L min−1 Nebulizer gas flow rate 1.12 L min−1 Sampling cone Ni, 1 mm aperture diameter Skimmer cone Ni, 0.4 mm aperture diameter Integration time for each isotope 0.1 s Readings per replicate 3 Signal measurement mode Three points per peak 53 Cr, 103Rh (as internal standard) Isotopes monitored Collision cell Pressurized: He 2 mL min−1

172

Z. Bahadir et al. / Spectrochimica Acta Part B 107 (2015) 170–177

Table 2 Evaluation of the purification treatment of the MWCNTs before functionalization with the anion exchanger Aliquat 336. Purification treatment description

Impurities (TXRF spectrum)

Cr/Y ratioa

RSD (%)a

MWCNTs were refluxed with concentrated nitric acid and washed with high-purity water until neutral pH was reached, and then dried at 100 °C MWCNTs were sonicated with diluted acid nitric (2 M) at room temperature for 30 min. Then they were filtered and washed with distilled water. To eliminate metal oxide catalysts the MWCNTs were then dispersed in diluted hydrochloric acid solution (6 M) for 30 min under ultrasonic agitation. Finally, MWNCNTs were washed with high-purity water until neutral pH was reached, and then dried at 100 °C No treatment before functionalization

Fe, Mo, and Ca peaks detected

0.66

5

Fe, Mo, and Ca peaks detected

0.79

0.9

Fe, Mo, and Ca peaks detected

0.99

0.8

a

Cr/Y and RSD (%) were are calculated for the analysis of three replicates of a standard solution containing 500 μg L−1 of Cr(VI).

software (Spectra Plus 5.3, Bruker AXS Microanalysis GmbH, Berlin, Germany) linked to the equipment [26]. For the quantification in TXRF analysis, the software applies a deconvolution routine which uses measured mono-element profiles for the evaluation of peak areas. In order to evaluate the extraction efficiency of the preconcentration procedure used, Cr content in liquid samples (before and after the DMSPE procedure) was determined by a quadrupole-based ICP–MS system (Agilent 7500c, Agilent Technologies, Tokyo, Japan) equipped with an octopole collision cell. Appropriate amounts of Rh internal reference solution were added to the aqueous samples of concern (final Rh concentration 50 μg L−1). The instrumental parameters employed are also summarized in Table 1.

The developed DMSPE procedure was as follows: an aliquot of 20 mL of sample was placed in a 30 mL polypropylene plastic tube and an appropriate volume of 0.1 M HCl or 0.1 M NaOH was added to adjust the pH (pH = 2). Next, 5 mg of the loaded MWCNTs was added and the resulting suspension was shaken in a rotary mixer for 20 min at 50 rpm. During this stage, Cr(VI) ions were extracted on to the Aliquat 336-MWCNTs. The mixture was then centrifuged for 20 min at 4000 rpm to obtain phase separation. After this process, the dispersed MWCNTs were sedimented at the bottom of the test tube and the supernatant was poured off carefully. Then, 0.3 mL of 2 mg L−1 Y in 2 M HNO3 (internal standard) was added to the solid phase. Finally, after homogenization using a Vortex device, an aliquot of 10 μL was transferred on a siliconized quartz glass sample carrier and dried under an IR heater for subsequent TXRF analysis.

2.3. Preconcentration procedure and TXRF analysis 3. Results and discussion Firstly, the anion-exchanger (Aliquat 336) was loaded on the raw MWCNTs. For this purpose, 0.025 g of raw MWCTNs was put in contact with 100 mL of 5% Aliquat 336 in methanol and the mixture was shaken overnight at room temperature. Then the suspension was filtered through a 0.20 μm filter membrane, washed with ultrapure water and dried in an oven at 110 °C.

3.1. Evaluation of Cr(VI) extraction using DMSPE with modified MWCNTs Because CNTs usually contain carbonaceous or metallic impurities, purification is an essential issue to be addressed. Acid treatment is one of the methods most commonly employed to remove these impurities

(B)

(A)

2 1

1.6

0.8

1.2

Cr/Y

Cr/Y

1.2

0.6 0.4

0.8 0.4

0.2 0

0 2

4

6

7

8

10

12

5

10

(C)

No good deposition (TXRF analysis)

Cr/Y

Cr/Y

0.8 0.6 0.4 0.2 0 5

10

amount MWCNTs (mg)

20

25

(D)

1

2

15

sample volume (mL)

pH

15

20

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

1 0.8 0.6 0.4 0.2 0 0

0

200

400

600

10

800

20

30

40

50

60

70

1000 1200 1400 1600

extraction time (min)

Fig. 1. Influence of some parameters on Cr(VI) extraction by DMSPE with modified MWCNTs: (A) Effect of sample pH (5 mg MWCNTs, 10 mL sample volume, 20 minute extraction time), (B) Effect of sample volume (5 mg MWCNTs, pH = 7.5, 20 minute extraction time), (C) Effect of amount of MWCNTs (10 mL sample volume, pH = 7.5, 20 minute extraction time), and (D) Effect of extraction time (5 mg MWCNTs, 10 mL sample volume, pH = 7.5). Error bars correspond to the standard deviation of duplicate sample analysis.

Z. Bahadir et al. / Spectrochimica Acta Part B 107 (2015) 170–177

173

Fig. 2. Influence of some parameters to carry out TXRF measurements: (A) Effect of the volume of internal standard, (B) Effect of the type of solution used to prepare the suspension, (C) Effect of deposition volume on the reflector and (D) Effect of measurement time. Error bars correspond to the standard deviation of duplicate sample analysis.

introduced by the preparation processes and also to increase the possibility of further modification and functionalization of the graphite surface. However, one of the main drawbacks of acid-oxidation methods is CNT fragmentation and defect generator of the graphitic network [27]. Therefore, before the functionalization of the MWCNTs with the anion-exchanger, two different purification systems were tested and the results obtained for a standard solution of 500 μg L−1 Cr(VI) were compared with those obtained without purification of the MWCNTs. A summary of the purification treatments and the results obtained is summarized in Table 2. According to the obtained results, purification of the MWCNTs appears not to provide real benefits in terms of impurity removal, extractive capacity of chromium and repeatability. For this reason, raw MWCNTs (without purification) were used in further experiments. In order to obtain high Cr(VI) preconcentration rates, after modification of the raw MWCNTs with the extractant Aliquat 336, the effect of different parameters affecting hexavalent chromium extraction was carefully evaluated. One variable at a time optimization was used to obtain the most favorable conditions for the DMSPE procedure. In all experiments, a standard solution containing 500 μg·L−1 of Cr(VI) was used. Taking into account that chromium (VI) has different ion species in aqueous solution depending on the pH, the influence of this parameter was evaluated in the pH range of 2–12. As it is shown in Fig. 1A, similar results were found when working at the pH range of 4–10. Taking into account the pH of water samples, a pH ~7 was finally selected to carry out the experiments. Another parameter that can significantly affect the global sensitivity of the methodology is the sample volume. To evaluate this effect, different sample volumes in the range of 5–25 mL were tested and the results obtained are displayed in Fig. 1B. As it can be seen, a significant increase in Cr/Y ratio is found when using higher sample volumes until 20 mL (above this volume there is a negative deviation). Taking into account these findings a sample volume of 20 mL was selected for further experiments.

The amount of MWCNTs used for preconcentration can have an essential effect on the extraction of metal species from the aqueous solution but also on the later TXRF analysis. As stated in the introduction section, to perform analysis under total reflection conditions, analyzed samples must be deposited as thin layers on a reflective carrier. Therefore, the thickness of the deposited samples, which is related to the amount of the MWCNTs, can have influence on the final determination of Cr (IV). In Fig. 1C, the results obtained using MWCNTs in a range of 2–20 mg are shown. When using higher amounts of MWCNTs the ratio Cr/Y is not enlarged and instead a higher dispersion of the results is found. Moreover, when using 20 mg of MWCNTs the resulting sample on the reflector is too thick and the analysis cannot be performed under total-reflection conditions. In view of these findings and also taking into account the repeatability of the obtained results, an amount of 5 mg of functionalized MWCNTs was used in the developed preconcentration procedure. Finally, the effect of extraction time was also evaluated in the range of 5 min to 24 h and the obtained results are displayed in Fig. 1D. As it is shown, the large contact area between dispersed MWCNTs and the aqueous sample provides a fast achievement of the equilibrium state. The adsorption nearly occurs immediately and the extraction time has

Table 3 Analytical parameters for the DMSPE–TXRF system. Sample pH

Linear range (μg L−1)

R2

RSD (%)a

LoDb

Extraction (%)c

(A)

7.5

10–500

0.9973

2

88 ± 1

(B)

2

10–3000

0.9988

9.5 (10) 3.5 (50) 0.5 (500) 11.9 (20) 5.4 (50) 0.8 (3000)

3

46 ± 6

a b c

RSD(%): between brackets Cr(VI) concentration level in μg L−1 (n = 2). LoD: Limit of detection calculated according to the 3σ approach [9]. Extraction (%) determined by ICP–MS (n = 2, 50 μg L−1 Cr(VI) initial solution).

174

Z. Bahadir et al. / Spectrochimica Acta Part B 107 (2015) 170–177

Fig. 3. Effect of sample matrix (tap water) on Cr(VI) determination using the DMSPE–TXRF system. Error bars correspond to the standard deviation of duplicate sample analysis.

no influence on the adsorption of the Cr(VI) on the loaded MWCNTs. For practical reasons, we chose a stirring time of 20 min. 3.2. Selection of analytical parameters for TXRF measurements Parameters affecting the sample deposition step as well as operating conditions for TXRF measurements were also evaluated to obtain the best sensitivity for Cr determination. Experimental tests were performed using the extraction conditions selected above and using a standard solution containing 500 μg L−1 of Cr(VI). As discussed in the previous section, to perform the analysis under total reflection conditions, the target sample has to be deposited as a thin layer on a reflective carrier. Therefore, in addition to the amount of MWCNTs, the volume and the type of the liquid used to prepare the suspension as well as the sample deposition volume are critical parameters which have to be evaluated. In Fig. 2A, the volume effect of the internal standard solution used to prepare the suspension on Cr relative area (Cr/Y) is evaluated. In parenthesis the value of the neat peak area for Cr peak is also displayed. As it is shown, better results in terms of Cr/Y ratio are found when preparing the suspension using a volume of 0.1 mL of Y 2 mg L−1. However, in this case, the area of the Cr peak is the lowest one. Therefore, it was considered appropriate to use 0.3 mL of Y 2 mg L−1 to prepare the suspensions in further experiments. Taking into account that the amount of MWCNTs was previously established at 5 mg, the concentration of the resulting suspension was around 16 mg mL− 1 (5 mg MWCNTS in 0.3 mL of solution). In the present study, three dispersant agents (MilliQ water, 1% Triton® X-100 in water and 2 M nitric acid) were also tested to prepare the solution of internal standard used to prepare the suspension. As it can be seen in Fig. 2B, the best sensitivity using a ratio 5 mg MWCNTs/0.3 mL suspension agent was obtained by using a solution of 2 M nitric acid. As reported in the literature, non-ionic detergents such as Triton® X-100 could

be used to adjust the viscosity of solutions and enhance the homogeneity of the analyzed samples [11]. However, as it is shown in the obtained results, an improvement of the results when using such reagent as a diluting agent did not occur. The diameter of the sample spot on the sample carrier has to be within the X-ray beam size (10 mm) to ensure complete exposure of the drop to the X-ray beam. For that, only a few microliters (5–20 μL) of suspension solution should be deposited on the sample carrier. As it can be seen in Fig. 2C, the influence of the sample deposition volume was evaluated in the range of 5 to 20 μL. In this later case, instead of a single deposition of 20 μL of the sample, two sequential depositions of 10 μL in the same location with time (allowing droplet drying between successive depositions) were performed. As it is shown, no statistically significant differences in analyte response were obtained for the studied sample deposition volumes. In view of the obtained results a volume of 10 μL was established for TXRF analysis since using this volume the sample was provided as a centered-thin film on the reflector and moreover a better repeatability was obtained compared to that achieved when using lower or higher deposition volumes. After the deposition procedure on the reflector, the micro-droplet must be properly dried to perform the TXRF analysis. For that, different drying modes were tested in order to provide the achievement of a centered-thin film on the quartz reflector: (i) drying under and IRlamp, (ii) drying on a hot plate set at ~70 °C and (iii) drying under a vacuum chamber. No significant differences in terms of Cr signal were found between the tested drying modes. For practical reasons we selected drying the sample using an IR-lamp. Finally, operation conditions of the TXRF system were also selected to obtain the best instrumental sensitivity for Cr determination. The rate of kV/mA of the X-ray tube was selected to work under conditions of maximum efficiency of excitation (50 kV, 1 mA, max. power 50 W). Concerning the measurement time it was selected in respect to the lowest relative standard deviation (RSD) obtainable as detailed in Fig. 2D. Each experimental point represents the RSD value calculated from the analysis of five replicates of a deposited preconcentrated sample containing 50 ng mL−1 of Cr(VI) (maximum permissible content according to the WHO). As expected, the higher the measurement times the lower standard deviation. From these results, a measurement time of 2000 s was fixed (RSD ~5–6%). 3.3. Analytical figures of merit of the DMSPE–TXRF system Using the best analytical conditions selected in Sections 3.1 and 3.2, we evaluate the analytical figures of merit of the developed DMSPE– TXRF system. A summary of the obtained results is presented in Table 3A. For calibration purposes, 12 aqueous samples containing Cr(VI) concentrations in the range of 10–5000 μg L−1 were taken throughout the whole preconcentration and detection procedure described above. Relative peak area intensities of Cr (Cr/Y) obtained under the selected

Fig. 4. Influence of MWCNTs amount in the recovery of Cr(VI) at a concentration level of 50 μg L−1. Experimental conditions: (A) 5% Aliquat 336 and (B) 10 mL sample volume. Error bars correspond to the standard deviation of duplicate sample analysis.

Z. Bahadir et al. / Spectrochimica Acta Part B 107 (2015) 170–177

175

Fig. 5. Effect of carbonate amount on Cr(VI) determination: (A) pH = 7.5 and (B) pH = 2. Experimental conditions: 50 μg L−1 Cr(VI), 5 mg MWCNTs, 20 mL sample volume, (5% Aliquat 336). Error bars correspond to the standard deviation of duplicate sample analysis.

operating conditions, were plotted versus known concentrations of Cr(VI) in the solution and a straight line was fitted to measured points by the least-square method (confidence limit 95%). A good linearity was found (R2 N 0.99) from 10 to 500 μg L−1 Cr(VI). This fact demonstrated that, in this concentration range, the extraction efficiency is independent of the initial Cr(VI) concentration and thus the metal loaded in the functionalized MWCNTs is related with the initial hexavalent chromium in the aqueous solution. The extraction efficiency was calculated from the difference between the chromium concentration at the beginning of the experiment and the concentration after application of the developed preconcentration procedure, determined by the analysis of the aqueous samples by means of ICP–MS. It was found that the extraction efficiency of the DMSPE system was ~ 90% when using a sample pH ~7 (see Table 3A). The precision of the methodology was evaluated in terms of relative standard deviation (RSD) of replicate analysis of standard solutions and at different initial Cr(VI) concentrations (see Table 3A). In all cases, relative standard deviations calculated were less than 10%. Limits of detection were calculated according to the 3σ approach11 and a value of approximately 2 μg L−1 Cr(VI) was obtained. This result is more than 20 times lower than the maximum Cr(VI) content permissible in drinking waters according to the WHO. Moreover, the detection limit obtained in this work is about two orders of magnitude lower than those attained for conventional XRF instrumentation [24] and it is also competitive with those determined using other preconcentration methodologies in combination with spectrometric techniques [28,29]. In order to test the suitability of the method when dealing with more complex matrices as well as the influence of interfering anions, the determination of Cr(VI) in spiked tap water samples at concentration levels of Cr(VI) in the range of 50–500 μg L−1 was undertaken. The anionic composition of the water sample (determined by ionic chromatography) was: 100 mg L−1 chloride, 150 mg L−1 bicarbonate, 13 mg L− 1 sulfate, 100 mg L−1 chloride, 1.04 mg L− 1 nitrate, 0.01 mg L− 1 phosphate, 0.1 mg L− 1 chlorate, b 0.1 mg L−1 fluoride, and b 0.02 mg L−1 bromide. As it is shown in Fig. 3, Cr(VI) extraction was also constant on the studied concentration range but it was significantly lower than that obtained for Cr(VI) standard solutions. Therefore, for quantification purposes in tap water samples, it is not possible the use of external calibration with Cr(VI) standards and a standard addition procedure would be necessary. The reason for that could be the competition of Cr(VI) with other ions found in tap water samples for the anionic-exchanger Aliquat 336. Therefore, in order to decrease the effect of other interfering anions and to improve the extraction efficiency for Cr(VI) in this kind of water samples, three different parameters regarding the DMSPE procedure were modified: (i) Increase of the percentage of Aliquat 336 when functionalizing the MWCNTs (from 5% to 10%), (ii) Increase of the amount of MWCNTs used for preconcentration (from 5 mg to 20 mg) and (iii) Decrease of the initial sample volume

(from 20 to 10 mL). All the experiments were performed analyzing, in duplicate, a tap water sample spiked at the level of 50 μg L−1 Cr(VI). A summary of the obtained results is shown in Fig. 4. In general, when increasing the amount of MWCNTs used for preconcentration, the recovery of Cr(VI) was slightly improved but the standard deviation of obtained results (error bars in Fig. 4) was considerably higher. Nor the decrease of the volume of the initial aqueous sample or the increase of the amount of Aliquat 336 added to the MWCNTs lead to a quantitative recovery of Cr(VI) in the target spiked tap water sample. In view of these findings and taking into account the content of bicarbonates in tap water samples, it was considered appropriate to study in more detail the effect of this anion on the extraction of Cr(VI). For that, five standard solutions containing Cr(VI) at the level of 50 μg L−1 and different concentrations of carbonate, in the range of 0 to 500 mg L−1, were adjusted at pH = 7 and pH = 2 and were analyzed using the developed DMSPE–TXRF methodology. As it is shown in Fig. 5, the effect of carbonate on Cr(VI) extraction is significant when adjusting the aqueous sample at pH = 7 but when working in acidic conditions (pH = 2) the interference of carbonates is avoided. To corroborate this trend, two spiked tap water samples at the levels of 50 μg L−1 and 100 μg L−1 of Cr(VI) (Corresponding to the maximum permissible contents according to the WHO and EPA, respectively) were adjusted at pH = 2 and analyzed using the DMSPE–TXRF system. Then the results (Cr/Y ratios) were compared with those obtained for the analysis of two standard solutions at the same concentration levels. As it can be seen in Table 4, no statistical significant differences at the 95% confidence level were found. According to these results, a pH = 2 was finally selected for the analysis of tap and mineral water samples. However, it has been demonstrated previously that the extraction efficiency of Cr(VI) was significantly different depending on the sample pH (see Fig. 1A). Therefore, to get accurate results for the analysis of tap and mineral water samples, a new set of Cr(VI) standards, adjusted at pH = 2, was prepared and analyzed. As it is shown in Table 3B, a good linearity was found from 10 to 3000 μg L−1 Cr(VI). Despite the fact that extraction efficiency for Cr(VI) at pH = 2 is half the value obtained for pH = 7, the limit of detection for hexavalent chromium determination

Table 4 Comparison of mean Cr/Y ratios in standard solutions of Cr(VI) at 50 μg L−1 and 100 μg L−1 with tap water samples spiked at the same concentration levels. Cr/Y ratio a Added

Standard solution (MilliQ water)

Tap water

tcalc b

50 100

0.1250 ± 0.0145 0.2331 ± 0.0126

0.1253 ± 0.0130 0.2140 ± 0.0110

0.02 1.61

a b

Mean values ± standard deviations (n = 2). Students “t” test. tcalc = calculated absolute value, tcritic = 4.30 (P = 0.05).

176

Z. Bahadir et al. / Spectrochimica Acta Part B 107 (2015) 170–177 Fe -Kα

Table 5 Cr(VI) concentrations determined at different concentration ratios of Cr(VI) and Cr(III) using the system DMSPE + TXRF system. Cr(VI)/Cr(III) μg L−1

Cr determined by DMSPE + TXRF method a μg L−1

0/50 25/25 50/0

n.d. 26 ± 3 50 ± 3

Fe -K β Ca-K α

Mo-K α

Bremsstrahlung X-ray tube

n.d.: not detected. a Mean values ± standard deviations (n = 2). Cr -Kα

is still significantly below the maximum permissible content according to the WHO and EPA. The precision of the methodology at pH = 2, estimated by the RSD values obtained from replicate analysis of standard solutions and at different initial Cr(VI) concentrations (see Table 3B), was very similar to the one obtained at pH = 7. Finally, the selectivity of the method for Cr(VI) determination in the presence of Cr(III) was also evaluated. For that, three aqueous standards containing different concentration ratios of Cr(VI) and Cr(III) species were prepared and analyzed using the developed DMSPE–TXRF system (see Table 5). Results obtained indicated that the presence of Cr(III) in the aqueous sample did not affect the quantitative determination of Cr(VI). This fact can be explained taking into account that Cr(III) is not extracted under the working conditions since it is present as a cationic species.

Fig. 6. TXRF spectra of the Mineral water 2, spiked at the level of 30 μg L−1 of Cr(VI), obtained using the system DMSPE–TXRF.

3.4. Application to tap and mineral water samples

4. Conclusions

The proposed method for Cr(VI) determination by DMSPE–TXRF was applied to the analysis of tap and mineral water samples. As it is shown in Table 6, Cr(VI) concentration was below the detection limit for all of the analyzed samples. However, in order to validate the methodology, the target samples were spiked with 30 and 50 μg L−1 Cr(VI). The recovery of the spiked samples was, in most cases, in the range of 101–108% which demonstrates the suitability of the TXRF methodology for monitoring Cr(VI) at trace levels in tap and mineral water samples. For readers' information, the content of some other anions found in the analyzed mineral water samples (determined by ionic chromatography) is also presented in the last column of Table 6. The results dem− 2− onstrated that the effect of ions (HCO− 3 , SO4 , Cl ) commonly present in this kind of water samples has no obvious interference effect for the determination of hexavalent chromium. In Fig. 6, an example of the TXRF spectrum obtained for the analysis of Mineral water-2 is also

In this study the usefulness of the combination of DMSPE using modified MWCNTs with the anionic exchanger Aliquat 336 with benchtop TXRF instrumentation for the determination of trace amounts of hexavalent chromium in drinking waters has been demonstrated. The use of TXRF as detection system eliminates the need for an elution step after the adsorption process of Cr(VI) on the modified MWCNTs, simplifying the procedure and reducing the total analysis time. Additional advantages of the methodology are the consumption of low volumes of reagents and the low operating costs since the portable TXRF system employed does not require cooling media and gas for operation. The presented methodology has been successfully applied to the quantification of Cr(VI) in spiked tap and mineral water samples. However, the obtained results highlight that sample pH has a crucial role to achieve good recoveries. The achieved limit of detection for Cr(VI) is suitable according to WHO and EPA regulations for Cr(VI) in drinking waters.

Table 6 Determination of Cr(VI) in drinking water samples (results are expressed as mean values

Mo-Kβ

displayed. As it is shown, in addition to the characteristic peaks of Cr and Y (internal standard), noticeable peaks of Ca, Fe and Mo are also detected. As previously described (Table 2), these elements are arising from the MWCNTs but they are not potential spectral interferences for Cr determination by TXRF.

Acknowledgments

± standard deviations, n = 3). Sample

Cr(VI) added (μg L−1)

Cr(VI) found (μg L−1)

Mineral water-1

0 30 50 0 30 50 0 30 50 0 30 50 0 30 50 0 25 50

n.d. 30 ± 3 48 ± 8 n.d. 30 ± 4 50 ± 5 n.d. 35 ± 7 54 ± 4 n.d. 30 ± 10 51 ± 6 n.d. 33.6 ± 0.5 53 ± 4 n.d. 26 ± 3 50 ± 3

Mineral water-2

Mineral water-3

Tap water-1

Tap water-2

Tap water-3

n.d.: not detected. a Information displayed in the container label.

Recovery (%) 108.3 95.3 98.4 99.6 115.7 107.8 102.9 101.4 111.0 105.6 104.0 100.5

Anions (mg L−1)a HCO− 3 : 258 SO2− 4 : 6.11 − Cl :5.69 HCO− 3 : 147.2 SO2− 4 : 29.1 Cl−: 5.65 HCO− 3 : 114 SO2− 4 :11.2 Cl−:6.0

This work was supported by the Spanish National Research Program (Project ref. CGL2010-22168-C03-01). Zekeriyya Bahadir wishes to thank the Higher Education Institution (YOK) in Turkey for the International Postgraduate Research Scholarship and also his own sincere gratitude to Girona University for laboratory facilities. References [1] M.A. Bezerra, M.A.Z. Arruda, S.L.C. Ferreira, Cloud point extraction as a procedure of separation and pre-concentration for metal determination using spectroanalytical techniques: a review, Appl. Spectrosc. Rev. 40 (2005) 269–299. [2] F. Priego-Copote, M.D. Lugue de Castro, Speciation of chromium by in-capillary derivatization and electrophoretically mediated microanalysis, J. Chromatogr. A 1113 (2006) 244–250. [3] W. Skwara Chwastowska, E. Sterlińska, L. Pszonicki, Speciation of chromium in mineral waters and Salinas by solid-phase extraction and graphite furnace atomic absorption spectrometry, Talanta 66 (2005) 1345–2349. [4] J. Kotaś, Z. Stasicka, Chromium occurrence in the environment and methods of its speciation, Environ. Pollut. 107 (2000) 263–283. [5] W. Chen, G.P. Zhong, Z.D. Zhou, P. Wu, X.D. Hou, Automation of liquid–liquid extraction-spectrophotometry using prolonged pseudo-liquid drops and handheld CCD for speciation of Cr(IV) and Cr(III) in water samples, Anal. Sci. 21 (2005) 1189–1193.

Z. Bahadir et al. / Spectrochimica Acta Part B 107 (2015) 170–177 [6] C. Zeng, Y. Lin, N. Zhou, J. Zheng, W. Zhang, Room temperature ionic liquid fluids enhanced the speciation of Cr(VI) and Cr(III) by hollow fiber liquid phase microextraction combined with flame atomic absorption spectrometry, J. Hazard. Mater. 237–238 (2012) 365–370. [7] M.G. Pasias, I.N. Borova, V.L. Panara, A.K. Thomaidis, Comparative study of two different speciation methods for the determination of hexavalent chromium in water samples using electrothermal atomic absorption spectrometry, Curr. Anal. Chem. 9 (2013) 288–295. [8] K.C. Hsu, C.C. Sun, Y.C. Ling, S.J. Jiang, Y.L. Huang, An on-line microfluidic device coupled with inductively coupled plasma mass spectrometry for chromium speciation, J. Anal. At. Spectrom. 28 (2013) 1320–1326. [9] R. Klockenkämper, in winefordner, J.D. (series Ed.). Total-reflection X-ray fluorescence analysis. Volume 140 k in Chemical Analysis: A series of monographs on analytical chemistry and its applications. John wiley&sons, New york, 1997. [10] R. Fernández-Ruiz, M. Malki, A.I. Morato, I. Marin, Study of bioaccumulation kinetics of chromium (VI) in Acinetobacter beijerinckii type bacterium by total reflection X-ray fluorescence spectrometry, J. Anal. At. Spectrom. 26 (2011) 511–516. [11] H. Stosnach, Environmental trace-element analysis using a benchtop total reflection X-ray fluorescence spectrometer, Anal. Sci. 21 (2005) 873–876. [12] A. Sarafraz-Yazdi, A. Amiri, Liquid-phase microextraction, Trends Anal. Chem. 29 (2010) 1–14. [13] F. Pena-Pereira, I. Lavilla, C. Bendicho, Liquid-phase microextraction techniques within the framework of Green chemistry, Trends Anal. Chem. 29 (2010) 617–628. [14] B. Majidi, F. Shemirani, Salt-assisted liquid–liquid microextraction of Cr(VI) ion using an ionic liquid for preconcentration prior to its determination by flame atomic absorption spectrometry, Microchim. Acta 176 (2012) 143–151. [15] J. Chwastowska, W. Skwara, E. Sterlińska, L. Pszonicki, Speciation of chromium in mineral waters and Salinas by solid-phase extraction and graphite furnace atomic absorption spectrometry, 66 (2005) 1345–1349. [16] B. Hu, M. He, B. Chen, L. Xia, Liquid phase microextraction for the analysis of trace elements and their speciation, Spectrochim. Acta, Part B 86 (2013) 14–30. [17] E. Marguí, M. Sagué, I. Queralt, M. Hidalgo, Liquid phase microextraction strategies combined with total reflection X-ray spectrometry for the determination of low amounts of inorganic antimony species in waters, Anal. Chim. Acta 786 (2013) 8–15.

177

[18] E. Marguí, I. Queralt, M. Hidalgo, Determination of cadmium at ultratrace levels in environmental water samples by means of total reflection X-ray spectrometry after dispersive liquid–liquid microextraction, J. Anal. At. Spectrom. 28 (2013) 266–273. [19] K. Pyrzynska, Carbon nanostructures for separation, preconcentration and speciation of metal ions, Trends Anal. Chem. 29 (2010) 718–727. [20] M. Trojanowicz, Analytical applications of carbon nanotubes: a review, Trends Anal. Chem. 25 (2006) 480–489. [21] J. Zhang, Preparation, characterization and application of thiosemicarbazide grafted multiwalled carbon nanotubes for solid-phase extraction of Cd(II), Cu(II) and Pb(II) in environmental samples, J. Environ. Sci. 25 (2013) 2331–2337. [22] S.A. Kosa, G. Al-Zhrani, M.A. Salam, Removal of heavy metal from aqueous solutions by multi-walled carbon nanotubes modified with 8-hydroxyquinoline, Chem. Eng. J. 181–182 (2012) 159–168. [23] K. Kocot, B. Zawisza, E. Marguí, I. Queralt, M. Hidalgo, R. Sitko, Dispersive micro solid-phase extraction using multiwalled carbon nanotubes combined with portable total-reflection X-ray fluorescence spectrometry for the determination of trace amounts of Pb and cd in water samples, J. Anal. At. Spectrom. 28 (2013) 736–742. [24] C. Fontàs, I. Queralt, M. Hidalgo, Novel and selective procedure for Cr(VI) determination by X-ray fluorescence analysis after membrane concentration, Spectrochim. Acta, Part B 61 (2006) 407–413. [25] P.R. Aranda, S. Moyano, L.D. Martinez, I.R. De Vito, Determination of chromium (VI) in drinking water using X-ray fluorescence spectrometry after solid phase extraction, Anal. Bioanal.Chem. 398 (2010) 1043–1048. [26] S2 PICOFOXTM, User Manual, Bruker advanced X-ray solutions, Bruker AXS Microanalysis GmH, Berlin (Germany), 2007. [27] P.X. Hou, C. Liu, H.M. Cheng, Purification of carbon nanotubes, Carbon 46 (2006) 2003–2025. [28] M. Tuzen, M. Soylak, Multiwalled carbon nanotubes for speciation of chromium in environmental samples, J. Hazard. Mater. 147 (2007) 219–225. [29] M. Tuzen, M. Soylak, Chromium speciation in environmental samples by solid phase extraction on Chromosorb 108, J. Hazard. Mater. 129 (2006) 266–273.