CHAPTER 9 |
 |
Spectrophotometric Flow-Injection System Using Multiwalled Carbon Nanotubes as Solid Preconcentrator for Copper Monitoring in Water Samples
Giovana de Fátima Lima*, Polyana Maria de Jesus Souza*, Mariana Gava Segatelli**, Pedro Orival Luccas* and César Ricardo Teixeira Tarley*
* Departamento de Ciências Exatas, Universidade Federal de Alfenas, Unifal-MG, Alfenas, Minas Gerais, Brazil
** Instituto de Química, Departamento de Química Inorgânica, Universidade Estadual de Campinas, Unicamp, Campinas, São Paulo, Brazil
Contents
2.4. Sorbent Flow Preconcentration System
3.2. Optimization Procedure Based on Fractional Factorial and Doehlert Designs
3.3. Effect of Foreign Ions as Potential Interferences
3.4. Analytical Features of the Proposed Method
3.5. Validation of the Proposed Method
1. Introduction
Copper is one of a relatively small group of metallic elements, which are essential for human, plants, and animal enzymes. Copper ions are introduced into human and animal bodies by food, water, air, etc. On the other hand, rain, snow, fertilizer, and irrigation water are the most common routes for copper absorption by plants [1]. Also, copper enters the marine environment primarily from riverine transport and aerosols. Anthropogenic and geothermal sources can also be included [2]. Depending on its concentration, copper may be a hazard to aquatic organisms [3]. Thus, although copper is essential to life, its determination in water samples is warranted by the narrow window of the concentration between essential and toxic [3]. In this way, copper levels in natural environments continue to be of interest, and many efforts to determine copper levels have been focused [4, 5]. In this context, one of the most difficult and complicated analytical tasks is the development of accurate and sensitive quantitative analysis of several kinds of samples especially at trace levels. At the same time, the demand for very simple and much cheaper analysis has been evident [6]. In this way, despite the importance of the highly sensitive atomic spectrometric techniques, such as inductively coupled plasma mass spectrometry or graphite furnace atomic absorption spectrometry is significant for metal determination at trace levels, flame atomic absorption spectrometry (FAAS), owing to simplicity and fairly low initial cost and running cost, still remains as a reliable method for metal determination. However, this technique suffers from lack of sensitivity for metal determination at trace levels, thus requiring usually preconcentration procedures prior determination [7]. UV–vis molecular spectrophotometry is another technique that also can be widely used for very simple and cheap analysis. Moreover, the existence of different chromogenic agents offers a widespread application of UV–vis molecular spectrophotometry for metal determination, enabling enhancements in the selectivity and attains lower detection limits [8]. In addition, several approaches can be found in literature for improving the sensitivity and selectivity of this technique, including formation of mixed aggregates with surfactants [9], measurements with long path length [10], and preconcentration procedures (liquid–liquid extraction, solid-phase extraction, and cloud point extraction) [11–13]. The preconcentration procedures using solid sorbent have gained special importance in analysis of complex matrices owing to some advantageous features, such as higher preconcentration factor (PF), simplicity, easy coupling between flow injection analysis (FIA), better repeatability, high sample throughput, and easy regeneration of solid phase [14]. Basically, a solid sorbent can be used as chelate-forming functional group or as support, and the properties of both components determine the features and the application of the respective material. Hence, the choice of the effective sorbent regarding selectivity and efficiency of the preconcentration in the analytical method is made by taking into account the nature of functional group and the physicochemical properties of the sorbent, such as mechanical and chemical stability, surface area, porous volumes, and kinetic characteristics [15]. The literature presents a variety of sorbents, and the most prominent among them are C18 silica, activated carbon, and polymeric resins [16]. Nevertheless, most of them have shown some drawbacks, for which the limited breakthrough volumes and the narrow pH stability range for C18 silica and poor selectivity of some polymeric resins can be pointed out [16]. Therefore, currently, there is interest in the development of new solid-phase extractors with high sorption capacity, selectivity, and PF. There are numerous attempts in this direction, and recently, nanomaterials have shown advantages over traditional sorbents, mainly owing to high chemical activity once their surface atoms are unsaturated and, therefore, can bind strongly with other atoms; large surface area; and high chemical stability. As examples, nanoparticles of titanium dioxide and alumina have been used as solid sorbents for the enrichment of metal ions following their determination by FAAS and inductively coupled plasma optical emission spectrophotometry (ICP-OES), respectively [17, 18]. Nanomaterials based on only saturated carbon atoms also known as carbon nanotubes, in both single-walled carbon nanotubes and multiwalled carbon nanotubes (MWCNTs), are becoming a potential sorbent for metal ions due to their unique electronic, mechanical, and chemical properties [19]. Literature survey showed that the MWCNTs were first used as sorbent in off-line preconcentration mode for cadmium, manganese, and nickel metals using ICP-OES as quantification technique [20]. Recently, we published two works that were based on sorbent flow preconcentration using MWCNT for cadmium and lead, using, respectively, thermospray flame furnace atomic absorption spectrometry and FAAS for metals determination [21, 22]. These works have emphasized the powerful performance of MWCNT when associated to an element-selective technique. However, there is no investigation with regard to selective performance of MWCNT for metal ions using flow injection preconcentration coupled to UV–vis molecular spectrophotometry, a technique naturally less selective and cheaper than atomic absorption spectrometry. Thus, the aim of this work was to develop a reliable flow injection preconcentration system coupled to UV–vis molecular spectrophotometry for copper determination using MWCNT as sorbent. Copper determination was accomplished after preconcentration followed by elution and complexation with diethyldithiocarbamate (DDTC), a complexant more broadly applicable to heavy metals [23]. The feasibility of the proposed method was assessed by its application for copper determination in water samples (mineral water, tap water, river water, and synthetic seawater) and reference material. Overall, chemometric tools based on factorial design and response surface methodology (RSM) have been used for the methodology optimization.
2. Experimental
2.1. Apparatus
A Femto UV–vis spectrophotometer (São Paulo, Brazil) model 482 equipped with flow cell of 1 cm optical length was used for FIA measurements. Data acquisition was carried out from an interfaced (Advantech) PCL 711S mode, and a computational program was developed in an EXCEL® spreadsheet, using macros of Visual Basic®. A Shimadzu UV–vis molecular absorption spectrophotometer (Tokyo, Japan) model AA-6800 with glass cell of 1 cm optical length was used for spectrum scan in the wavelength range from 400 to 700 nm. An Ismatec Model IPC peristaltic pump (Ismatec IPC-08, Glattzbrugg, Switzerland) furnished with Tygon® tubes was used to propel all sample and reagent solutions. The preconcentration/elution steps were selected by using a home-made injector commutator made of Teflon® (polytetrafluoroethylene). Adjustment of sample pH was done by using a Handylab 1 Schott pHmeter (Stafford, UK). A microwave oven (Milestone, Sorisole, Italy) was used for decomposition of reference material (beech leaves). The morphological characteristic of carbon nanotubes was evaluated using a JEOL JMT-300 scanning electron microscope (Tokyo, Japan) with an electron acceleration voltage of 20 kV. The samples were previously coated with a thin Au/Pd layer in a Bal-Tec MED 020 equipment.
2.2. Reagents and Solutions
The solutions were prepared with analytical grade chemical reagents, as well as with water obtained from a Milli-Q purification system (Millipore, Bedford, Massachusetts, USA). In order to prevent metal contamination from laboratory glassware, it was kept overnight in a 10% (v/v) HNO3 solution.
Copper standard solutions were prepared daily by appropriate dilution of 1000 mg L−1 copper solution (Biotec, Darmstadt, Germany).
DDTC solution [0.50% (w/v)] was prepared by solving 1.00 g of sodium DDTC in 25 mL of ethanol (Cetus, São Paulo, Brazil) and 175 mL of hot water and further stored in a freezer until use.
CTAB solution [6.0% (w/v)] was prepared by solving 6.00 g of cetyltrimethylammonium bromide in 50 mL of ethanol (Cetus, São Paulo, Brazil) and 150 mL of water.
Acetate buffer at 2.0 mol L−1 solution without further purification was prepared by dilution of 57.36 mL of acetic acid (Vetec, Rio de Janeiro, Brazil) in volumetric flask of 500 mL with water. The pH adjustment was carried out with sodium hydroxide at 1.0 mol L−1 (Vetec, Rio de Janeiro, Brazil).
HNO3 solution at 1.0 mol L−1 concentration used as eluent was prepared from concentrated nitric acid (Merck, Darmstadt, Germany).
MWCNTs were supplied by CNT Co., Ltd. (Yeonsu-Gu, Incheon, Korea) with >93% purity, diameter between 10 and 40 nm and length of 5–20 µm. Prior to use, MWCNTs were submitted to acid treatment to create carbonyl and carboxyl groups onto MWCNT surface which are potentially involved on metal sorption. For relevant details, the reader is referred to the article by Barbosa et al. [22] and Tarley et al. [21]. The structural changes in the MWCNT morphology induced during acid treatment can be seen in Fig. 9.1. The MWCNT diameter was open-ended, thus favoring a better mass transport of analyte toward MWCNT surface. After this treatment, 30 mg of MWCNT was packed into the minicolumn (6.0 × 1.0 cm i.d.) made of polyvinyl chlorine containing glass wool at both ends to prevent losses of sorbent during preconcentration/elution steps.
2.3. Sample Preparation
The methodology developed was applied for copper determination in mineral water, tap water, river water, and synthetic seawater. Mineral water samples, acquired from local supermarkets, were spiked with a known amount of copper followed by pH adjusting with buffer solution. Tap water sample was collected from Unifal-MG campus and analyzed immediately after pH adjustment. River water samples were collected in polypropylene flasks from a lake near Alfenas city and immediately acidified with HNO3 solution until a pH sample at 2.0 was attained. Afterwards, the samples were filtered through 0.45 µm cellulose acetate membranes under vacuum, and the pH was adjusted with buffer solution. Synthetic seawater, which was prepared according to literature [22], also was analyzed without filtering procedure. Beyond these samples, the feasibility of the methodology also was evaluated by analysis of reference material (beech leaves CRM No 100). Approximately 300 mg of the sample was placed into Teflon® flasks followed by addition of 10 mL concentrated HNO3 and 4 mL 30% (v/v) H2O2. The heating program of the microwave oven was accomplished using two steps. A potency of 400 W during 5 min was used with further increase to 700 W, which was maintained during 5 min. After that, digested samples were heated on a hot plate to near dryness, then cooled at room temperature, and dissolved in buffer solution. Any contamination source was checked by using blank solutions.
Figure 9.1 Scanning electron micrographs of raw MWCNT (amplified 30,000 times) (A) and MWCNT oxidized with concentrated HNO3 (amplified 50,000 times) (B).
2.4. Sorbent Flow Preconcentration System
The diagram of the flow preconcentration system is displayed in Fig. 9.2. At the preconcentration step position (Fig. 9.2A), the sample buffered at pH 5.2 was percolated through a minicolumn of MWCNT at 5.1 mL min−1 flow rate during 2 min and 47 s. After this stage, by switching the central part of the injector commutator (Fig. 9.2B), a stream of 1.0 mol L−1 HNO3 displaces the copper ions at 1.0 mL min−1 flow rate. Afterwards, copper ions eluated react with 0.15% (w/v) DDTC solution at 3.0 mL min−1 flow rate forming colloidal slurry that is further solubilized by 1.0% (w/v) CTAB surfactant at 0.6 mL min−1 flow rate. The Cu(DDTC)2 complex is driven to the spectrophotometer where the absorbance measurements are continuously made at 452 nm. All absorbance signals were taken as peak height.
Figure 9.2 Schematic diagram of the sorbent flow preconcentration system of copper ions onto MWCNT with spectrophotometry determination: (A) preconcentration stage and (B) elution stage. C, minicolumn packed with MWCNT. For more details, see text.
2.5. Optimization Procedure
The optimization procedure was performed using 26–2 fractional factorial design followed by Doehlert design for three factors [24], to attain the best sensitivity and to reduce the number of essays. All essays in duplicate were carried out by percolating 10 mL of 254 µg L−1 copper solution through minicolumn of MWCNT. The total of six factors was investigated, and their respective levels are shown in Table 9.1. Doehlert design was then used to establish the final optimization of those significant factors, using the RSM. It is important to point out that other factors that exert paramount importance in a sorbent flow preconcentration system, such mass of sorbent, type, and flow rate of the eluent, were fixed according to previous study [21]. Thus, 30 mg of MWCNT and HNO3 1.0 mol L−1 as eluent at 1.0 mL min−1 flow rate were adopted in this study. In addition, the complexant flow rate and buffer concentration were also fixed at 3 mL min−1 and 0.1 mol L−1. The analyses of experimental essays were processed using the STATISTICA software package (StatSoft, Tulsa, USA).
Table 9.1 Factors and experimental domain used in the 26–2 fractional factorial design for copper preconcentration onto MWCNT
3. Results and Discussion
3.1. Absorption Spectra
The DDTC complexes have been studied thoroughly, and detailed information on their behavior is readily accessible in the literature, and one of them is that DDTC form chelates with metals even at high acidities [23]. This is a favorable aspect once copper elution from MWCNT is commonly carried out with strong mineral acids, mainly nitric acid solution. In order to confirm the best sample pH of the reaction between DDTC and copper ions and to achieve the maximum wavelength of absorption, a scan spectrum was carried out from 400 to 700 nm, against a reagent blank. The absorption spectra were taken using 50 mg L−1 copper solution previously buffered at different pH range 2.5–3.5–5.5–7.0 and 9.0 containing CTAB at 1.0% (w/v) concentration. The absorbance values of Cu(DDTC)2 complexes at different pH values did not show significant differences, and a maximum absorbance was attained at 452 nm.
3.2. Optimization Procedure Based on Fractional
Factorial and Doehlert Designs
The screening of factors involved in the sorbent flow preconcentration system was made from a 26–2 fractional factorial design which enables investigation of the main estimated effect of each factor using only 16 essays (Table 9.2), against 64 essays in regard to a full factorial design. The fractional factorial design was built by using the two generator matrices I = pH • SFR • C and I = pH • SFR • SR • SC. Thus, the signals of surfactant coil (C) column were achieved by multiplying the signals of pH and surfactant flow rate (SFR), while the signals of surfactant concentration (SC) were obtained by multiplying the signals of pH, SFR, and sampling flow rate (SR). The main estimated effects of each factor in the system studied were calculated taking into account the height peak as experimental response. Analysis of variance (ANOVA) was successfully used for establishing the significance of estimated effects, which was represented from Pareto Chart (Fig. 9.3). This chart leads to an interpretation with a 95% confidence interval of those significant factors, whereas the horizontal bars higher than vertical line demonstrate statistically significant factors. Analyzing the Pareto Chart, it can be observed that the majority of factors were statistically significant [SFR, SR, C, sample pH, and CC (complexant complexation)], except the SC. So, the low level [1.0% (w/v)] of this latter factor was adopted in this study. In addition, it can be seen from Pareto Chart that some interaction effects were also significant, but they were not considered once their effects present serious confounding pattern, which stem from random experimental errors taking in consideration that a 26–2 fractional factorial design (1/4 of a full factorial design) was used.
Table 9.2 Experimental matrix of the 26–2 fractional factorial design and responses (height peak) for copper preconcentration onto MWCNT
Figure 9.3 Pareto Chart obtained from 26−2 fractional factorial design.
The most significant factor, SFR, which presented a negative effect (−25.33), indicates a decrease in the analytical response when working with high flow rate (3.0 mL min−1). Bearing this in mind, we decided to continue the optimization using Doehlert design (Table 9.3) to obtain reagent consumption reduction and to attain better analytical signal. The SR also showed a negative effect (−23.35), thus revealing a slowly sorption process of copper ions onto MWCNT. Therefore, a better analytical response would be obtained at lower levels than 5.1 mL min−1, but experiments in this direction were not tested to guarantee a reasonable sample throughput. So, SR at 5.1 mL min−1 was maintained for subsequent experiments. As regard, the influence of C, the negative effect (−12.67), reveals that the inclusion of coil (250 µL) in the flow system leads to dispersion of sample zone and, as a consequence, a decrease in the analytical response. Hence, the flow system was operated in the absence of the coil.
The influence of the sample pH on the analytical signal studied within the experimental domain from 3.8 to 5.5 shows, according to positive effect (9.25), that copper sorption onto MWCNT increases with increasing pH. CC presented a similar behavior to sample pH, whereas when working with its highest level (0.1% w/v), the formation of the metal chelates is favored. Under some conditions previously fixed, SC at 1.0% w/v, SR at 5.1 mL min−1, and absence of coil in the flow system, Doehlert design for three factors was used as a great tool for the final simultaneous optimization of CC, sample pH, and SFR. Their levels used in the experimental design (Table 9.3) were chosen in accordance with those results based on 26–2 fractional factorial design. According to real data applied to Doehlert design, the following equation was achieved, which shows the relationship between the three factors and the analytical response:
Table 9.3 Doehlert matrix for three factors used for final optimization
Abs = – 5.21 + 3.29CC – 13.65CC2 + 0.67pH2 + 12.61SFR
– 9.37SFR2 – 0.21CC • pH + 2.96CC • SFR – 0.26pH • SFR
From this equation, three response surfaces were obtained and are illustrated in Fig. 9.4. This quadratic model significance was checked by using ANOVA (Table 9.4). From MSlof/MSpe ratio (2.7381), in which the result is smaller than that critical F3, 2 tabled value of 19.16, it is possible to conclude that quadratic model does not present lack of fit, thus being a suitable model for adjusting with experimental results. In addition, this equation has a high determination coefficient (R2) of 0.97745 revealing the residues of the experiment are very low. In order to check the presence of maximum points, i.e., whether in fact the critical point is the relative maximum, relative minimum, or saddle point, the Lagrange’s criterion was used, which is based on calculation of Hessian determinant [4]. There is no information about the quadratic model when Δ2 = 0; there is a maximum point when Δ1 < 0, Δ2 > 0, and Δ3 < 0; there is a minimum point when Δ1 > 0, Δ2 > 0, and Δ3 > 0, and there is a saddle point when none of the above situations are observed. The Hessian determinants of a function H(CC, pH, SFR) were calculated by using the following expressions:
Figure 9.4 Surface responses obtained from Doehlert design for three factors.
Table 9.4 ANOVA from the Doehlert design shown in Table 9.3
The values found for Δ1, Δ2, and Δ3 were −27.3, 2.4, and −42.2, respectively, confirming the presence of maximum points. These maximum points were obtained by solving the quadratic model as follows:
This way, the maximum values for CC, pH, and SFR were found to be 0.15% (w/v), 5.20, and 0.60 mL min−1, respectively, being these values fixed for the method.
3.3. Effect of Foreign Ions as Potential Interferences
Although there are some studies demonstrating interference and studies concerning selectivity of preconcentration procedures based on copper sorption onto MWCNT [21, 22], they usually use atomic absorption spectrometric measurements an inherent selective technique. Hence, the interference may occur only by competition between foreign ions and analyte for the active sites of sorbent. In the proposed method, the presence of other metallic cations could cause loss or enhancement of analytical signal due to the competition for the active sites of sorbent as well as by reaction with the DDTC complexant. So, the selectivity of the method was assessed by copper preconcentration under optimized conditions in the presence of each metallic ion Mn(II), Zn(II), Fe(II), Co(II), Ni(II), Ba(II), Pb(II), Cd(II), Al(III), and Cr(III) in concentrations up to 10-fold higher than that of the analyte (100 µg L−1), while Ca(II) and Mg(II) were investigated at a concentration up to 100-fold higher. The interference was here established as a change higher or lower than 10% in the recovery of copper analytical signal with respect to the copper signal alone. As can be seen in Table 9.5, the interference was noted only for those metals that exhibit strong sorption on MWCNT [20–22] and that react likewise with DDTC [23]. Despite the interference, the ratio analyte/foreign ion tested was higher than that observed in real samples analyzed in this work. Therefore, it suggests that the proposed method can be successfully applied for copper determination in different kinds of samples (water samples and biological material) as will be demonstrated later.
Table 9.5 Effect of foreign ion on the recovery of copper analytical signal
3.4. Analytical Features of the Proposed Method
An external calibration curve was built ranging from 8.0 to 500.0 µg L−1, as shown in Fig. 9.5 (inset figure), resulting in the following linear equation: Abs = 0.0013 + 0.00217[Cu(II) (µg L−1)] (r > 0.999). The calibration curve built without using preconcentration procedure was Abs = −0.00768 + 0.00005709[Cu(II) (µg L−1)] with linear correlation coefficient r = 0.9999. PF was calculated as described by literature [24], using the ratio of the slopes of the analytical curve derived by the proposed method with that one built without preconcentration procedure. The PF achieved was found to be 38.0 taking into account a preconcentration time of 2 min and 47 s. The detection and quantification limits, calculated as three times the standard deviation of the black according to IUPAC recommendation [25], were 1.56 and 5.19 µg L−1, respectively. Precision values, in terms of repeatability expressed as relative standard deviation (n = 6), were 6.8 and 0.21% for copper solutions at 20 and 400 µg L−1 concentrations, respectively. The concentration efficiency [21], CI [22], and sample throughput were found to be 13.6 min−1, 0.37 mL, and 18 h−1, respectively. A brief comparison of the proposed preconcentration method with others previously published (Table 9.6) allowed us to develop an advantageous method, mainly owing to the lower limit of detection and low sample consumption, which leads to a high sample throughput.
Figure 9.5 Recorder of analytical signal (fiagram) for building external calibration (inset figure).
3.5. Validation of the Proposed Method
The usefulness of the sorbent flow preconcentration system was evaluated by analyzing water samples and reference material. Essays based on addition and recovery tests of known copper concentration were carried out, and the results are shown in Table 9.7. Values ranging from 90.9 up to 102.3%, obtained from external calibration, prove that it is possible to achieve excellent accuracy even for those complex samples, such as synthetic seawater. Moreover, average amounts of copper found in beech leaves CRM No 100 (12.4 ± 1.4 µg g−1) obtained by the current method agreed closely with those values derived from certified material (12.0 µg g−1) with 95% confidence level (test t-student), thereby also attesting its accuracy. Also, it is important to comment that the methodology was successfully applied for copper monitoring in river water without interference. According to National Council of Environment (CONAMA) [31] regularized by Brazilian legislation, the allowed maximum level of copper in sweet water is 13.0 µg L−1, a lower concentration than that determined. In addition, copper ions also can be determined in saline water without interference once the allowed maximum level in accordance with CONAMA is 7.8 µg L−1 for this kind of sample.
Table 9.6 Comparison of different sorbent flow preconcentration systems for copper determination using UV–vis molecular spectrophotometry with the proposed method
Table 9.7 Validation of the method for copper determination in spiked water samples
4. Conclusions
An attempt to coupling between flow preconcentration system and solid-phase extraction using MWCNT with spectrophotometric determination allowed a reliable and useful method for copper determination in water and biological samples. This system is the first one that demonstrates the powerful application of MWCNT in field of solid-phase preconcentration using spectrophotometric determination, a technique naturally less selective than atomic absorption spectrometry. The DDTC reagent, a nonselective complexant for copper determination, was used purposely so that the selectivity was attributed only in the solid-phase extraction, i.e., mainly owing to the high affinity of copper ions for the active sites of MWCNT. Besides the good analytical performance of such flow system including high PF, simplicity, low detection limit, and low cost of implementation, the MWCNT minicolumn has also shown excellent stability once the same packed minicolumn has been used in the development of two more methodologies [21, 22] without losses of sorption capacity. Overall, the results achieved in this work, based on accuracy and precision, also clearly showed the wide application of carbon nanotubes in the field of solid-phase preconcentration procedures for monitoring of copper ions in water samples.
Acknowledgments
The authors would like to thank the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP), Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG), and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for the financial support and fellowships. They also would like to thank Prof. Maria Antonieta Avarenga for language assistance.
References
[1] Agency for Toxic Substances and Disease Registry. Toxicological profile for copper draft for public comment. Atlanta (GA): U.S. Department of Health and Human Services, Public Health Service; 2002.
[2] Merian E, Anke M, Ihnat M, Stoeppler M. Elements and their compounds in the environment. 2nd ed. Weinheim: John Wiley; 2004.
[3] Fleming CA, Trevors JT. Copper toxicity and chemistry in the environment: a review. Water Air Soil Pollut 1989;44:143–58.
[4] Tarley CRT, Dos Santos WNL, Dos Santos CM, Ferreira SLC, Arruda MAZ. Factorial design and Doehlert matrix in optimization of flow system for preconcentration of copper on polyurethane foam loaded with 4-(2-pyridylazo)-resorcinol. Anal. Letters 2004;37:1437–55.
[5] de Peña YP, López W, Burguera JL, Burguera M, Gallignani M, Brunetto R, et al. Synthetic zeolites as sorbent material for on-line preconcentration of copper traces and its determination using flame atomic absorption spectrometry. Anal Chim Acta 2000;403:249–58.
[6] Sweileh JA. Sorption of trace metals on human hair and application for cadmium and lead pre-concentration with flame atomic absorption determination. Anal Bioanal Chem 2003;375:450–5.
[7] Anthemidis AN, Zachariadis GA, Farastelis CG, Stratis JA. On-line liquid-liquid extraction system using a new phase separator for flame atomic absorption spectrometric determination of ultra-trace cadmium in natural waters. Talanta 2004;62:437–43.
[8] Lemos VA, Santos ES, Santos MS, Yamaki RT. Thiazolylazo dyes and their application in analytical methods. Microchim Acta 2007;158:189–204.
[9] Akl MA. An improved colorimetric determination of lead(II) in the presence of non-ionic surfactant. Anal Sci 2006;22:1227–31.
[10] Bonifácio VG, Figueiredo-Filho LC, Marcolino Jr LH, Fatibello-Filho O. An improved flow system for chloride determination in natural waters exploiting solid-phase reactor and long pathlength spectrophotometry. Talanta 2007;72:663–7.
[11] Bashammakh AS, Bahaffi SO, Al-Sibaai AA, Al-Wael HO, El-Shahawi MS. Extractive liquid-liquid spectrophotometric procedure for the determination of thiocyanate ions employing the ion pair reagent amiloride monohydrochloride. Anal Chim Acta 2007;592:16–23.
[12] Figueiredo EC, Tarley CRT, Kubota LT, Rath S, Arruda MAZ. On-line molecularly imprinted solid phase extraction for the selective spectrophotometric determination of catechol. Microchem J 2007;85:290–6.
[13] Filik H, Tayman A. A new cloud-point preconcentration approach for the spectrophotometric determination of p-aminophenol in the presence of paracetamol with 2-(2-hydroxyphenyl)-lH-benzimidazole as a coupling reagent. J Anal Chem 2007;62:530–5.
[14] Gama EM, da Silva Lima A, Lemos VA. Preconcentration system for cadmium and lead determination in environmental samples using polyurethane foam/Me-BTANC. J Hazard Mater 2006;136:757–62.
[15] Bilba D, Bejan D, Tofan L. Chelating sorbents in inorganic chemical analysis. Croat Chem Acta 1998;71:155–78.
[16] Liu G, Wang J, Zhu Y, Zhang X. Application of multiwalled carbon nanotubes as a solid-phase extraction sorbent for chlorobenzenes. Anal Lett 2004;14:3085–104.
[17] Liang P, Shi T, Jing LI. Nanometer-size titanium dioxide separation/preconcentration and FAAS determination of trace Zn and Cd in water sample. Int J Environ Anal Chem 2004;84:315–21.
[18] Hang C, Hu B, Jiang Z, Zhang N. Simultaneous on-line preconcentration and determination of trace metals in environmental samples using a modified nanometersized alumina packed micro-column by flow injection combined with ICP-OES. Talanta 2007;71:1239–45.
[19] Chiang IW, Brinson BE, Smalley RE, Margrave JL, Hauge RH. Purification and characterization of single-wall carbon nanotubes. J Phys Chem B 2001;105:1157–61.
[20] Liang P, Liu Y, Guo L, Zeng J, Lu H. Multiwalled carbon nanotubes as solid-phase extraction adsorbent for the preconcentration of trace metal ions and their determination by inductively coupled plasma atomic emission spectrometry. J Anal At Spectrom 2004;19:1489–92.
[21] Tarley CRT, Barbosa AF, Segatelli MG, Figueiredo EC, Luccas PO. Highly improved sensitivity of TS-FF-AAS for Cd(II) determination at ng L–1 levels using a simple flow injection minicolumn preconcentration system with multiwall carbon nanotubes. J Anal At Spectrom 2006;11:1305–13.
[22] Barbosa AF, Segatelli MG, Pereira AC, Santos AS, Kubota LT, Luccas PO, et al. Solidphase extraction system of Pb (II) ions enrichment based on multiwall carbon nanotubes couple on-line to flame atomic absorption spectrometry. Talanta 2007;71:1512–9.
[23] Duran C, Gundogdu A, Bulut VN, Soylak M, Elci L, Sentürk HB, et al. Solid-phase extraction of Mn(II), Co(II), Ni(II), Cu(II), Cd(II) and Pb(II) ions from environmental samples by flame atomic absorption spectrometry (FAAS). J Hazard Mater 2007;146:347–55.
[24] dos Santos WNL, Santos CM, Ferreira SLC. Application of three-variables Doehlert matrix for optimisation of an on-line pre-concentration system for zinc determination in natural water samples by flame atomic absorption spectrometry. Microchem J 2003;75:211–21.
[25] Long GL. Winefordner JD. Limit of detection. Anal Chem 1983;55:712–5.
[26] Yamini Y, Tamaddon A. Solid-phase extraction and spectrophotometric determination of trace amounts of copper in water samples. Talanta 1999;49:119–24.
[27] de Moraes SVM, Brasil JL, Milcharek CD, Martins LC, Laranjo MT, Gallas MR, et al. Use of 1, 3-diaminepropane-3-propyl grafted onto a silica gel as a sorbent for flow-injection spectrophotometric determination of copper (II) in digests of biological materials and natural waters. Spectro. Chim. Acta A 2005;62:398–406.
[28] Castilho E, Cortina JL, Beltrán J-L, Prat M-D, Granados M. Simultaneous determination of Cd(II), Cu(II) and Pb(II) in surface waters by solid phase extraction and flow injection analysis with spectrophotometric detection. Analyst 2001;126:1149–53.
[29] Lee T, Choi H. Determination and preconcentration of Cu(II) after adsorption of its cupferron onto benzophenone. Bull Korean Chem Soc 2002;23:861–4
[30] CONAMA. National Council of Environment. Resolution number 357, may 9, 2005.