A separation system for lead fractionation in river water using electrothermal atomic absorption spectrometry

Abstract

This work proposes a separation system that uses the Sep-Pak C-18 cartridge to enable the determination of labile and total lead in natural water samples. This method is composed of two steps. First, the sample is percolated through the cartridge, where the lead that is bound to the organic compounds is retained and the quantification of labile lead is performed using the eluate from this separation. Afterwards, another sample enables the determination of total lead. Lead was determined through electrothermal atomic absorption spectrometry (ET-AAS) using aluminum as a chemical modifier under the following instrumental conditions: pyrolysis temperature of 800 °C and atomization temperature of 1800 °C. This method allows the determination of lead with limits of detection and quantification of 0.14 and 0.47 μg L⁻¹, respectively. The accuracy of this method for the determination of total lead was confirmed by analysing a standard reference material of trace elements in natural water supplied by NIST. Experiments using lead solutions in the presence of humic acid confirmed the efficiency of the Sep-Pak C-18 cartridge for the quantitative extraction of lead ions bound to organic complexing agents. This method was applied for the determination of labile and total lead in river water samples collected in Santo Amaro City, Bahia State, which is a region that is historically contaminated by lead as a result of a lead ore smelter that operated in the area from 1960 to 1993. For the five samples analyzed, the total lead concentrations ranged from 1.18 to 13.57 μg L⁻¹. The fraction of labile lead ranged from 48% to 80% in four samples. The fifth sample had a labile lead content of 11%. The total carbon content was also determined in all the analyzed samples. The obtained concentrations did not present a correlation with the results obtained for lead.

---

1 Introduction

Studies on the bioavailability of metals in waters are important because the biological response of organisms to trace metals in waters is proportional to the free-ion activity of the metals, not to their total or dissolved concentrations. In aqueous systems, trace amounts of chemical elements can appear in a variety of organic and inorganic forms, which range from simple hydrated molecules to large organic complexes. Therefore, the speciation analysis of metals in the environment is fundamental for predicting their impacts on aquatic biota. Metals in aquatic systems are generally complexed by soluble inorganic ions (e.g., fluoride, chloride, bicarbonate, sulfate, etc.) and organic (e.g., humic substances) ligands. Generally, complexation with organic ligands reduces the metal bioavailability because the majority of organic–metal complexes are not readily transported across the cell membranes. However, inorganic metal complexes (e.g., carbonates) typically rapidly dissociate to the free-metal form. Therefore, while the bioavailable fraction of metals includes both free-metal ions and kinetically labile metal complexes (i.e., those with rapid dissociation kinetics), the biological response is only proportional to the free-metal concentration.¹

Trace metal contamination of natural waters is a major concern due to the potential harmful effects both directly within the aquatic biota of the impacted systems and those associated with human health.² Furthermore, there is a permanent concern in urban environments related to trace metal loads and speciation.³ The effect of heavy metals, such as lead, cadmium, etc., on the biota in urban waters is considered to be a very important pollution matter worldwide.^4–6 Nevertheless, many of these previous studies were conducted in developed countries under temperate conditions. Unfortunately, less is known about the relationship between waterway conditions and land use in the tropical areas of developing countries. In developing countries such as Brazil, it is estimated that a considerable amount of sewage has been directly discharged into rivers, lakes, and coastal waters without any type of treatment. According to the latest National Survey on Basic Sanitation, the vast majority of Brazilian cities discharge their raw sewage into natural aquatic systems.

The traditional methods employed for the determination of total metal ions and labile metal ions utilize UV radiation for destroying organic matter and an electroanalytical technique for quantifying the metal content in two-steps. The labile metal ions are determined in the sample in their natural state, whereas the total metal ions are measured after UV irradiation of the sample. Thus, Abdel and Kureishy⁷ used UV radiation to destroy the organic matter in filtered water samples during a procedure proposed for the determination of the total dissolved and labile fractions of copper, lead and cadmium. During this procedure, cathodic stripping voltammetry was used as the analytical technique. Daniele et al.⁸ proposed a method that used stripping voltammetry with mercury microelectrodes for the determination of the labile and total fractions of cadmium, lead and copper in rain water samples. Scarponi et al.⁹ employed a method for the determination of labile and total lead in seawater using anodic stripping voltammetry (ASV). Magnier et al.¹⁰ determined the lability of dissolved copper, lead and zinc in fresh water samples collected from the Deule River, France. Lead and zinc were quantified using anodic stripping voltammetry. These methods are attractive because they provide high sensitivity, multi-element quantification and a low risk of sample contamination during analysis.

A more suitable technique has recently been used for evaluating the bioavailability of metal ions in environmental matrices, which is the diffusive gradient in thin film (DGT) technique. This technique provides an in situ quantitative measure of labile species in aqueous systems.¹¹ Sherwood et al.¹² used DGT to assess environmental risk in marine waters. Schintu et al.¹³ employed macro-algae and DGT as indicators for available trace metals in marine coastal waters near a lead–zinc smelter. Agbenin and Welp¹⁴ evaluated the bioavailability of copper, cadmium, zinc, and lead in tropical savanna soils using DGT and ion exchange resin membranes. Other authors have proposed a method that employs DGT for the determination of dissolved aluminum species in waters.¹⁵ The advantages of DGT include in situ deployment, speciation capabilities, sensitivity, time-integrated signal and low-risk of contamination. In addition, the DGT device can be deployed and retrieved by minimally trained personnel at a relatively low-cost. The principal limitations of the DGT method include the limited functional pH range (5 to 9 for most metals), the limited applicability to certain metals/metalloids, and its unsuitability in waters with a very low cation concentration (<2 × 10⁻⁴ M).

Electrothermal atomic absorption spectrometry (ET-AAS) and inductively coupled plasma mass spectrometry (ICP-MS) are analytical techniques that are conventionally used for the determination of trace amounts of toxic metals in several matrices.^16–18 However, these techniques are not satisfactory for some environmental applications because interpretation of the results requires information on the concentrations of the inorganic metals and on the metals bound to organic compounds.

This study proposes a separation system that uses the Sep-Pak C-18 cartridge to determine labile and total lead in river water samples through ET-AAS. This method was applied to real samples collected from Santo Amaro da Purificação City, in Brazil. This area historically has a lead contamination issue.

2 Experimental

2.1 Instrumentation for lead quantification using ET-AAS

A ZEEnit 600 atomic absorption spectrometer (Analytik Jena AG, Jena, Germany) equipped with a transverse-heated graphite tube atomizer and correction for the Zeeman-effect background was used for the quantification of lead, and it employed transversely heated pyrolytic graphite-coated tubes with an integrated PIN platform. The lead hollow cathode lamp (Varian, Mulgrave, VA, Australia) was operated at 4 mA, and the instrumental conditions were as follows: wavelength of 283.3 nm and spectral bandwidth of 0.8 nm. Argon with a purity of 99.997% (White Martins, Salvador, Brazil) was used as the purge gas with an internal flow rate of 2.0 L min⁻¹ during all steps, except during atomization, when the internal flow was stopped. Analytical signals were measured as the integrated absorbance, A_int (peak area). Aluminum was used as a chemical modifier; details can be found in previous reports for the determination of lead in sugar cane¹⁹ and cadmium in rice.²⁰

The optimized temperature program for the graphite furnace used for the determinations is presented in Table 1.

Table 1 Temperature program used for determination of the lead content; an internal gas flow rate of 2.0 L min⁻¹ was used in all stages except during atomization, when the gas flow was turned off

Step	Temperature (°C)	Ramp (°C s^−1)	Hold time (s)
a Full power.
Drying	110	15	10
Drying	120	10	15
Drying	140	5	10
Pyrolysis	700	50	20
Atomization	1600	FP^a	4
Cleanout	2550	FP	5

---

2.2 Total carbon quantification using ICP-OES

The determination of total carbon was performed using a Varian Vista PRO Inductively Coupled Plasma Optical Emission Spectrometer (Mulgrave, Australia) with axial viewing under the following instrumental conditions: power (1.3 kW), plasma gas flow (15.0 L min⁻¹), auxiliary gas flow (1.5 L min⁻¹) and nebulizer gas flow (0.8 L min⁻¹). Carbon standard solutions with concentrations ranging between 50 and 500 mg L⁻¹ were prepared using citric acid (Merck, Darmstadt, Germany) diluted in deionized water. The emission line used was C (I) 193.027 nm, and yttrium was used as an internal standard for calibration.

2.3 Reagents and solutions

Solutions and standards were prepared using deionized water with a specific resistivity of 18 MΩ cm that was obtained from a Milli-Q system (Millipore, Bedford, USA). The employed nitric acid solutions were prepared by diluting concentrated nitric acid (Merck). The lead working solutions were prepared by serial dilutions from a 1000 mg L⁻¹ lead stock solution (Merck) in 0.05% nitric acid. These solutions were prepared daily before use. A 1000 mg L⁻¹ aluminum solution for ICP-MS (Merck), which was prepared in 0.50 mol L⁻¹ nitric acid, was used for the chemical modification.

The humic acid solution (Sigma-Aldrich, St Louis, MO, USA) was prepared by placing 0.01 g of a dried humic acid sample and 60.0 mL of ultrapure water into a 100 mL volumetric glass flask. The humic acid solution was homogenized for 30 min in an ultrasonic bath. After this procedure, the obtained suspension was filtered through 0.45 μm cellulose membranes and the final solution volume of 100 mL was obtained by adding ultrapure water.

Sep-Pak C18 cartridges (Millipore Waters) were pre-conditioned by elution with water (5 mL) and methanol (5 mL) before use.

River water samples were collected from Santo Amaro City, Bahia State, in October 2010. All samples were filtered using 0.45 μm cellulose acetate membranes and were analyzed immediately after sampling.

3 Results and discussion

3.1 Determination of lead by ET-AAS

Lead was determined using ET-AAS under the following instrumental conditions: pyrolysis temperature of 800 °C, pyrolysis time of 20 s and atomization temperature of 1800 °C, using 3 μg of aluminum as a chemical modifier following the conditions described previously.^19,20 The limits of detection and quantification were obtained according to the IUPAC recommendation,²¹ in which LOD = (3σ/S) and LOQ = (10σ/S), where σ is the standard deviation of a blank and S is the slope of the calibration curve. Under the optimized conditions, the LOD and LOQ were 0.14 and 0.47 μg L⁻¹, respectively. The analytical blank was a 0.05% nitric acid solution, which was also used for preparing the lead standard solutions for external calibration. The precision of the method, which is expressed as the relative standard deviation (%RSD), was 4.77 and 1.31% based on the analysis of two river water samples with total lead concentrations of 1.18 and 13.57 μg L⁻¹, respectively, for three replicates. For labile lead, the %RSD were 5.25 and 3.6% for samples with labile lead contents of 0.92 and 10.86 μg L⁻¹ (n = 3). The accuracy of the ET-AAS method was evaluated by analyzing a certified reference material (1643e) of natural water supplied by the National Institute of Standards and Technology (NIST). The determined lead content of 19.48 ± 0.41 μg L⁻¹ was in good agreement with the certified value of 19.63 ± 0.21 μg L⁻¹.

3.2 Efficiency of the separation column

To test the efficiency of the C-18 column for the retention of lead as organic complexes, several solutions were prepared with lead concentrations of 20 and 225 μg L⁻¹ and humic acid concentrations ranging from 100 to 2000 μg L⁻¹. These solutions were aged for 24 hours under stirring. During the experiments, 2.0 mL aliquots of these solutions were percolated by the C-18 column with a flow rate of 1.0 mL min⁻¹, and the lead was determined in the eluted fraction. The lead was also determined in the prepared solutions that were not eluted through the C-18 columns. All the experiments were performed in duplicate, and these results are presented in Table 2. An additional experiment involved the elution of an aliquot of the 1643e natural water CRM through the separation column. The obtained lead concentration was 19.36 ± 0.77 μg L⁻¹. This result confirms that inorganic lead is not retained in the C-18 column.

Table 2 Efficiency of the separation column^a

Experimental conditions	A int without elution on the column	A int after elution on the column
a HA – Humic acid; Aint – integrated absorbance.
Pb 20 μg L^−1 + HA 1000 μg L^−1	0.0306	0.00477
Pb 20 μg L^−1 + HA 2000 μg L^−1	0.0335	0.00287
Pb 225 μg L^−1 + HA 100 μg L^−1	0.291	0.0147
Pb 225 μg L^−1 + HA 300 μg L^−1	0.295	0.00345
Pb 225 μg L^−1 + HA 600 μg L^−1	0.315	0.00379
Pb 225 μg L^−1 + HA 1000 μg L^−1	0.310	0.00937
Pb 225 μg L^−1 + HA 2000 μg L^−1	0.310	0.0116

---

3.3 Application – determination of labile and total lead in river water samples

Santo Amaro da Purificacao City, Bahia State, Brazil, with a surface area of 500 km², has been contaminated by lead as a result of a lead ore smelter, which operated in the area from 1960 to 1993.^22,23 Several papers have reported the contamination by lead in this municipality.^24–26 Thus, the proposed separation procedure was used for the determination of labile and total lead in river water samples collected from this city. Five samples were analyzed for the quantification of labile and total lead. The obtained results expressed as a confidence interval²⁷ are presented in Table 3.

Table 3 Determination of total and labile lead in river water samples (N = 3)^a

Sample	Total lead (μg L^−1)	Labile lead (μg L^−1)	Labile lead (%)	Lead present in organic fraction (μg L^−1)	Total carbon^b (mg L^−1)
a Results expressed as a confidence interval at the 95% level; N = number of determinations. b Determination using ICP-OES.
1	13.57 ± 0.44	10.86 ± 0.97	80.02	2.71 ± 1.07	213 ± 2
2	4.14 ± 0.21	0.47 ± 0.15	11.35	3.67 ± 0.26	194 ± 2
3	2.30 ± 0.26	1.58 ± 0.23	68.70	0.72 ± 0.35	257 ± 3
4	2.09 ± 0.14	1.01 ± 0.11	48.32	1.08 ± 0.18	282 ± 3
5	1.18 ± 0.14	0.92 ± 0.12	77.97	0.26 ± 0.18	1795 ± 17

---

The determined concentrations for total lead are relatively similar, with the exception of sample 1, in which the lead content was greater than for the other samples. This sample was collected at the sampling point closest to the lead ore smelter. The content of labile lead in the five samples ranged from 11% to 80%. Sample 2, which had a labile lead content of 11%, was collected in a region with a low water flow rate.

Among the five samples analyzed in Table 3, only sample 1 exhibited total and labile lead concentrations that were greater than the maximum permissible levels specified by Brazilian legislation²⁸ for fresh water, which is 10 μg L⁻¹.

Another recent work²⁰ from our research group analyzed twenty drinking water samples that were collected from this city. The lead concentrations in these samples ranged from 2.9 to 7.17 μg L⁻¹.

In general, it is hypothesized that the presence of organic matter can significantly influence the lability of metallic species, such as lead, cadmium, copper and zinc, regardless of the method employed for the speciation analysis. Meylan et al.²⁹ compared in situ DGT measurements for copper and zinc with metal speciation assessed by competitive ligand exchange with voltammetric measurements, with quite similar results. Dos Anjos and co-workers³⁰ used DGT and anodic stripping voltammetry to assess the lability of copper, lead, cadmium, and zinc. Despite the differences among these methods, the obtained results also agreed well with predictions from an equilibrium model. In all the aforementioned cases, the presence of organic ligands contributed to the formation of inert metallic complexes. The availability to the biota can be moderated or even completely eliminated through the metal complexation with some functional groups present in the organic matter. However, the characteristics of the organic matter can drastically vary due to the diversity and variability of the original source of carbon. For instance, Sodré and Grassi³¹ have demonstrated that the complexation of copper by dissolved organic matter can be significantly dependent on the origin of the organic material. In their work, anthropogenic influences on the characteristics of aquatic dissolved organic matter (DOM) were evaluated using synchronous-scan fluorescence spectroscopy, and they collected evidence that the organic matter from both rivers exhibited simple organic structures, such as aromatic amino acids or conjugated aliphatics, as the major constituents during the rainy summer. However, during the dry winter, it became clear that raw sewage discharges from the urbanized area contributed to an increase of fluorescent organic matter that was unable to complex copper. In this study, it is quite probable that raw sewage discharges have provided a source of organic material that is unable to interact with the lead in the river water.

Additionally, the total carbon concentrations were similar (194–282 mg L⁻¹), with the exception of sample 5 (1795 mg L⁻¹), which was collected in a region of raw sewage discharges.

4 Conclusions

The bioavailability and toxicity of dissolved metallic ions in natural waters depends directly on their chemical speciation, which is partially controlled by their interactions with organic ligands. In this context, the proposed fractionation procedure is very opportune because the separation system is simple and the determination of lead can be achieved using ET-AAS or any other analytical technique.

Although five samples were analyzed, only one sample exhibited total and labile lead concentrations that were greater than the maximum limit allowed by Brazilian regulations for surface waters.

Acknowledgements

The authors are grateful to Fundação de Amparo a Pesquisa do Estado da Bahia (FAPESB), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for providing grants and fellowships and for the financial support.

References

1. International Network for acid prevention, DGT Review, accessed 7 September 2012, http://www.inap.com.au/public_downloads/Research_Projects/Diffusive_Gradients_in_Thin-films.pdf.
2. F. F. Sodre and M. T. Grassi, Water, Air, Soil Pollut., 2006, 178, 103.
3. L. Herngren, A. Goonetilleke and G. A. Ayoko, J. Environ. Manage., 2005, 76, 149.
4. S. Ayrault, M. Roy-Barman, M. L. Cloarec, C. R. Priadi, P. Bonté and C. Göpel, Chemosphere, 2012, 87, 902.
5. M. Varol and B. Şen, Catena, 2012, 92, 1.
6. D. D. Loureiro, M. A. Fernandez, F. W. Herms and L. D. Lacerda, Environ. Monit. Assess., 2009, 159, 577.
7. M. AbdelMoati and T. W. Kureishy, Indian J. Mar. Sci., 1997, 26, 143.
8. M. E. Abdelsalam, G. Denuault and S. Daniele, Anal. Chim. Acta, 2002, 65, 452.
9. G. Scarponi, G. Capodaglio, G. Toscano, C. Barbante and P. Cescon, Microchem. J., 1995, 51, 214.
10. A. Magnier, G. Billon, Y. Louis, W. Baeyens and M. Elskens, Talanta, 2011, 86, 91.
11. H. Zhang and W. Davison, Anal. Chem., 1995, 67, 3391.
12. J. E. Sherwood, D. Barnett, N. W. Barnett, K. Dover, J. Howitt, H. Ii, P. Kew and J. Mondon, Anal. Chim. Acta, 2009, 652, 215.
13. M. Schintu, B. Marras, L. Durante, P. Meloni and A. Contu, Environ. Monit. Assess., 2010, 167, 653.
14. J. O. Agbenin and G. Welp, Environ. Monit. Assess., 2012, 184, 2275.
15. J. G. Panther, W. W. Bennett, P. R. Teasdale, D. T. Welsh and H. J. Zhao, Environ. Sci. Technol., 2012, 46, 2267.
16. O. T. Butler, W. R. L. Cairns, J. M. Cook and C. M. Davidson, J. Anal. At. Spectrom., 2011, 26, 250.
17. O. T. Butler, W. R. L. Cairns, J. M. Cook and C. M. Davidson, J. Anal. At. Spectrom., 2012, 27, 187.
18. S. L. C. Ferreira, H. S. Ferreira, G. D. Matos, D. S. Anunciaçao, W. N. L. dos Santos, V. P. Campos, J. B. de Andrade, B. Welz, R. G. O. Araujo, A. J. Curtius, P. C. Nascimento, L. Tormen and C. L. Jost, Appl. Spectrosc. Rev., 2012, 47, 633.
19. S. L. C. Ferreira, L. A. Portugal, G. D. Matos, D. G. da Silva, R. M. de Jesus and G. B. Brito, Anal. Methods, 2011, 3, 1168.
20. D. G. da Silva, M. M. S. Junior, L. O. B. Silva, L. A. Portugal, G. D. Matos and S. L. C. Ferreira, Anal. Methods, 2011, 3, 2495.
21. IUPAC Analytical Chemistry Division, Spectrochim. Acta, Part B, 1978, 33, 242.
22. F. M. Carvalho, M. L. Barreto, A. M. Silvany-Neto and T. M. Tavares, Sci. Total Environ., 1984, 35, 71.
23. A. M. Silvany-Neto, F. M. Carvalho, M. E. C. Chaves, A. M. Brandão and T. M. Tavares, Sci. Total Environ., 1989, 78, 179.
24. F. M. Carvalho, A. M. Silvany-Neto, A. M. C. Melo, M. E. C. Chaves, A. M. Brandão and T. M. Tavares, Sci. Total Environ., 1989, 84, 128.
25. W. N. L. dos Santos, J. V. S. Santos, L. O. B. Silva, A. S. Araujo, V. A. Lemos, M. Miro and S. L. C. Ferreira, Int. J. Environ. Anal. Chem., 2011, 91, 1425.
26. C. D. Guerra, R. F. Gerlach, N. G. V. Pinto, S. C. Cardoso, S. Moreira, A. P. de Almeida, I. T. A. Peixoto, C. H. Meloni, C. L. Mota, L. F. de Oliveira, D. Braz and R. C. Barroso, Appl. Radiat. Isotopes, 2010, 68, 71.
27. D. A. Skoog and J. J. Leary, Principles of Instrumental Analysis, Saunders College Publishing, Florida, 4th edn, pp. A-10–A-13.
28. Resolution 357 of the Brazilian government published in the Official Gazette on May 18, 2005.
29. S. Meylan, N. Odzak, R. Behara and L. Sigg, Anal. Chim. Acta, 2004, 510, 91.
30. V. E. dos Anjos, G. Abate and M. T. Grassi, Quim. Nova, 2010, 33, 1307.
31. F. F. Sodré and M. T. Grassi, J. Braz. Chem. Soc., 2007, 18, 1136.

---