Spectrochimica Acta Part B 101 (2014) 15–20

Contents lists available at ScienceDirect

Spectrochimica Acta Part B

j ourna l homepage: www.e lsev ie r .com/ locate /sab

Analytical Note

Multiwalled carbon nanotubes as a sorbent material for the solid phaseextraction of lead from urine and subsequent determination byelectrothermal atomic absorption spectrometry

Rosa M. Peña Crecente, Carlha Gutiérrez Lovera, Julia Barciela García, Jennifer Álvarez Méndez,Sagrario García Martín, Carlos Herrero Latorre ⁎Departamento de Química Analítica, Nutrición y Bromatología, Facultad de Ciencias, Universidad de Santiago de Compostela, Campus de Lugo, 27002 Lugo, Spain

⁎ Corresponding author. Tel.: +34 982824064; fax: +3E-mail address: carlos.herrero@usc.es (C.H. Latorre).

http://dx.doi.org/10.1016/j.sab.2014.07.0050584-8547/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 27 February 2014Accepted 14 July 2014Available online 23 July 2014

Keywords:SPECarbon nanotubesLeadUrineElectrothermal atomic absorption spectrometry

The determination of lead in urine is a way of monitoring the chemical exposure to this metal. In the presentpaper, a new method for the Pb determination by electrothermal atomic absorption spectrometry (ETAAS) inurine at low levels has been developed. Lead was separated from the undesirable urine matrix by means of asolid phase extraction (SPE) procedure. Oxidizedmultiwalled carbon nanotubes have been used as a sorbentma-terial. Lead from urine was retained at pH 4.0 andwas quantitatively eluted using a 0.7M nitric acid solution andwas subsequently measured by ETAAS. The effects of parameters that influence the adsorption–elution process(such as pH, eluent volume and concentration, sampling and elution flow rates) and the atomic spectrometryconditions have been studied by means of different factorial design strategies. Under the optimized conditions,the detection and quantification limits obtained were 0.08 and 0.26 μg Pb L−1, respectively. The results demon-strate the absence of a urine matrix effect and this is the consequence of the SPE process carried out. Therefore,the developedmethod is useful for the analysis of Pb at low levels in real samples without the influence of otherurine components. The proposedmethodwas applied to the determination of lead inurine samples of unexposedhealthy people and satisfactory results were obtained (in the range 3.64–22.9 μg Pb L−1).

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Lead is a toxic heavy metal that does not have any known positivephysiological role in the human body. Exposure to lead causes a varietyof diseases and undesired health consequences. An overview of the ad-verse effects that this metal has on the human organism is given in thereview by Needleman [1]. The main sources of Pb exposure includesmoking, contaminated air (mainly from tetraethyl lead used as amotor fuel additive), water, soil, food, and consumer products [2]. Occu-pational exposure in the workplace is the most common cause of leadpoisoning in adults, particularly in the smelting, scrapping and printingindustries, battery manufacture, pigment production, chemical andplastics industries, mining and metal work [3]. Since the lead level inurine reflects the amount of the element that has been recentlyabsorbed, the determination of Pb in urine is useful for assessingoccupational and environmental exposure [4]. Monitoring traceelements in urine is a difficult task due to the complexity of the matrixand the low concentration of analyte [5]. Therefore, the determination

4 982824001.

of Pb in urine requires a pretreatment step aswell as sensitive instrumen-tal techniques such as inductively coupled plasma-mass spectrometry(ICP-MS) and electrothermal atomization atomic absorption spectrome-try (ETAAS) [6,7].

ETAAS has been used for the determination of normal concentra-tions of this element in urine. However, the sample matrix representsthe main drawback for the measurements due to the considerableamount of numerous components in urine that might cause incompleteatomization, volatilization of lead, high background levels and the for-mation of carbonaceous residues in the atomizer [8]. Therefore, stabi-lized temperature-platform furnace (STPF) conditions, different matrixmodifiers and Zeeman background correction are essential to overcomethese problems [9–11]. However, these strategies are not enough tosuppress all of the problems in urine measurements and it may benecessary to remove the sample matrix [12]. For these reasons, priorto measurement, different techniques for the separation of lead fromurine have been applied. The most widely used technique for thisobjective was solid-phase extraction (SPE) [13–15]. SPE has proven tobe one of the suitable methods because of the following advantages:high enrichment factor, good recovery, rapidity, small quantities oforganic solvents, possibility of automation of the whole process andthe availability of a wide variety of sorbents.

Table 1Furnace heating programs for ETAAS determination of lead.

Step Temperature (°C) Ramp (s) Hold (s)

Dry 90 30 25Ash 350 15 5Atomizationa 1550 1 3Clean 2400 1 2Cold 40 40 0

a Stop Ar flow.

16 R.M. Peña Crecente et al. / Spectrochimica Acta Part B 101 (2014) 15–20

Since the discovery of carbon nanotubes (CNTs) in 1991 by Ijima[16], it became clear that theymight be an excellentmaterial for SPE be-cause of their high surface area and inner volume, stability, mechanicalstrength and the possibility of establishing π–π interactions. Therefore,since 1995 the works using diverse types of CNTs as SPE sorbents formetal preconcentration purposes have increased significantly [17,18].In the case of Pb, different SPE procedures involving as-grown andoxidized CNTs have been proposed for separating and preconcentratingthe metal, either directly in the form of Pb(II) ions [19–21], or as Pb-chelates [22,23]. In addition, other methods employ functionalizedCNTs as SPE sorbents [24–26]. In all cases, after extraction, Pb wasmeasured by different atomic spectrometric techniques with adequateresults. However, the main focus of these methods was directed to asingle matrix such as diverse types of waters.

The aim of the present study was to develop a SPE–ETAAS methodfor the determination of Pb in urine using multiwalled carbon nano-tubes (MWCNTs) as SPE-sorbent. The SPE procedure achieved the elim-ination of the urine matrix and avoided the need for other samplepretreatment. Furthermore, Pb levels were determined with appropri-ate analytical figures of merit. The capability of the developed methodfor monitoring Pb exposure was demonstrated by measuring realurine samples.

2. Material and methods

2.1. Apparatus and statistical software

A Zeeman correction Varian-SpectrAA-600 atomic absorptionspectrometer (Varian Inc., Palo Alto, CA, USA) equipped with a VarianGTA-100 electrothermal atomizer linked to an automatic sample dis-penser was used for this work. Measurements were performed using aPb Varian hollow cathode lamp operating at 283.3 nmwith a current in-tensity of 10 mA. The bandwidth employed was 1.0 nm in all cases.Argonwas used as the inert gas at a flow rate of 300mLmin−1. Pyrolyticgraphite-coated atomization tubes with a pre-inserted Omega Platformwere used (Schunk Ibérica S.A., Madrid, Spain). A Gilson Minipuls-3peristaltic pump (Gilson Inc., Middleton, WI, USA) equipped withTygon pump tubing of 1.42 mm i.d. (Ismatec, Wertheim, Germany)was used to propel both the sample and reagents. Polytetrafluoro-ethylene (PTFE) connecting tubing of 0.5 mm i.d. and various end-fittings and connectors (Omnifit-Diba Industries Ltd, Cambridge, UK)were employed. Sample pH measurements were performed by using aHI221 Calibration CheckMicroprocessor pH meter (Hanna InstrumentsS.L., Spain).

Experimental designs as well as Pareto and surface contour plots forthe optimization of the developed method were carried out usingStatgraphics Centurion XVI ver. 16.1.15 (Rockville, MD, USA).

2.2. Reagents

Untreated multiwalled CNTs (purity N 95%, 20–30 nm o.d. and~30 μm length) prepared by Chemical Vapor Deposition (CVD) of acet-ylene in hydrogen flow were supplied by Chengdu Organic ChemicalsCo. Ltd. (Chengdu, China).

A standard lead stock solution (1.0 g L−1) was obtained fromPanreac (Barcelona, Spain) and diluted as necessary to obtain thework-ing standard solutions. Ammoniumdihydrogenphosphatewas obtainedfrom Fluka (Buchs, Switzerland). The pH adjustment of urine sampleswas carried out using different nitric acid solutions obtained fromultra-pure HNO3 (Merck, Darmstadt, Germany). All materials werewashed with 10% nitric acid (v/v) for a period of 24 h, rinsed withcopious amounts of pure water and shaken dry before being used. Thecleaning solution employed to wash the sampling capillary contained0.7% (w/v) HNO3 and 0.2% (v/v) Triton X-100 (Sigma-Aldrich, St.Louis, MO, USA).

2.3. MWCNT pretreatment and microcolumn preparation

The oxidation ofMWCNT introduces hydroxyl, carbonyl and carbox-yl groups on the nanotube surface and this enhances the solubility of thematerial and produces bonded surface oxygen-containing radicals thathave the ability to retain a variety of metal ions at the appropriate pH[18]. In the present work, MWCNTs were oxidized by a microwave-assisted procedure employing the mixture H2SO4/KMnO4 as oxidant.The procedure has been described in detail in a previous work [27].The surface functionalization obtained by oxidation was quantitativelydetermined using Boehm's titrationmethod [28]. A suitable total aciditysurface was achieved (3.02 mmol g−1). Oxidized multiwalled carbonnanotubes (45 mg) were loaded into a 35 mm × 4 mm (i.d.) PTFE-microcolumn plugged with a small portion of glass wool at both endsto avoid sorbent losses during the SPE. Prior to use, the column wascleaned and conditioned by passing through 2 mL of a 1.0 M solutionof nitric acid followed by 4 mL of pure water.

2.4. Urine samples

Different fresh urine samples from healthy people with no history ofexposure to lead were collected in acid-washed polypropylene bottles.Samples were prepared as follows. The appropriate amount of urinewas treated with a 0.1 M solution of HNO3 to pH 4.0 and it was dilutedto 40% (v/v) before solid phase extraction and ETAAS-measurement.When the analysis was not immediately carried out the urine sampleswere stored at 4 °C (the storage period was kept as short as possible).

2.5. SPE–ETAAS analytical procedure

A 4 mL aliquot of urine sample, treated as indicated in Section 2.4,was passed through the microcolumn with a peristaltic pump at aflow rate of 1.1 mL min−1 to achieve retention of the Pb ions by theMWCNTs. In the second step, 1.5 mL of a 0.1 M solution of HNO3 waspassed through the column at a flow rate of 0.6 mL min−1 to eliminatethe urine matrix. The retained Pb ions were subsequently eluted byemploying 1.5 mL of a more concentrated 0.7 M HNO3 solution (at aflow rate of 0.4 mL min−1). 20 μL of this eluted solution sample wasmixed with 2 μL of an aqueous matrix modifier solution containing0.005% (w/v) NH4H2PO4 and the resulting sample was subjected toETAAS under the optimized conditions indicated in Table 1. All measure-ments were made in integrated absorbance mode. Eight microcolumnswere used simultaneously for the SPE procedure.

3. Results and discussion

3.1. Optimization of the SPE

Different variables that influence the SPE retention of lead ions byMWCNTs packed in the microcolumn were studied. Since oxidized-carbon nanotubes lead to the formation of active hydroxyl, carbonyland carboxyl groups on the surface of the MWCNT, it is evident thatthe pH of the solution passed through the column will affect thesorbent's structure as well as the degree of ionization and speciationof the adsorbates. In the present case, the pH range investigated was

17R.M. Peña Crecente et al. / Spectrochimica Acta Part B 101 (2014) 15–20

between4.0 and 8.0 because: (i) at pH values lower than 4.0 active\OHand \COOH groups are protonated and this would cause unfavorableconditions for the lead ion retention, and (ii) at pH values higher than8, the lead ion precipitates as the hydroxide. The eluent and its concen-tration are other critical factors for SPE. In the present work, nitric acidwas selected, because it was shown that HNO3 produces a satisfactoryelution and a very low background in subsequent ETAAS analysis [29].Sampling and elution flow rates were also considered in this optimiza-tion step, taking into account the necessity of achieving a compromisebetween efficiency and time.

Given the above information outlined, an optimization study wasperformed in order to select the best values for those parameters thatinfluence the SPE procedure. The pH, sample flow rate, eluent flowrate, eluent volume and eluent concentration were initially evaluatedby means of a screening study using a half-fraction factorial design25−1 in 16 randomized experiments performed in triplicate (in theranges indicated in Table S1 on Appendix A). According to the Paretochart obtained from this design (see Fig. S1 in Appendix A), pH, elu-ent concentration and eluent volume are all factors that influence theSPE. Consequently, a more detailed optimization study for thesethree variables was carried out by employing a central composite de-sign 23+ star in 16 randomized experiments (the central point wasanalyzed twice). According to the response surfaces obtained (seeFig. S2 in Appendix A), the optimal values for the three variableswere as follows: sample pH 4.0, eluent volume 0.99 mL (approx.1.0 mL) and eluent concentration 0.7 M. However, under these con-ditions, certain problems concerning the reproducibility of Pb mea-surements were detected. Inadequate reproducibility could be dueto the low elution volume selected. For this reason, a univariatestudy of this parameter was performed. On the basis of the extractionefficiency obtained for different extraction volumes (in the range 0.5to 2.0 mL), it was demonstrated that the use of 1.5 mL of 0.7 M HNO3

produced a high extraction efficiency (similar to 1.0 mL) with betterresult for reproducibility (see Fig. S3). Therefore, considering thecompromise between extraction efficiency and reproducibility,1.5 mL of nitric acid solution was selected as the optimum elutionvolume for further measurements.

The best extraction efficiencies were achieved at low flow rates.However, the use of such low flows dramatically increased the analysistime; consequently, the appropriate balance between flow rate andtime must be attained. Sample and elution flow rates were evaluatedby means of univariate studies (under the previously optimized condi-tions for the other parameters) in the range from 0.4 to 1.4 mL min−1.Optimum values for extraction efficiency were obtained for a sampleflow rate of 1.1 mL min−1 and an elution flow rate of 0.4 mL min−1.

3.2. Effect of urine dilution, sample volume and sorbent mass

The use of diluted samples combined with the inclusion of aprecolumn with glass fiber has proven to be a good strategy to avoidcolumn blockage or overpressure. Several urine sample preparationswith different dilutions (from 20 to 50% v/v) were investigated, and40% dilution was selected as the optimum because this value preventsblockage of the column and does not compromise the sensitivity ofthe method.

The next parameter to be studied was sample volume. It is evidentthat the use of a high sample volume could improve the sensitivity.However, the use of such large volumes causes a significant increaseof the time required to pass the sample through the column. In addition,the requiredmass of MWCNTs to ensure the correct retention of the an-alyte is also higher. Bearing these considerations in mind, the samplevolume was evaluated in the range from 2 mL to 10 mL and theMWCNT mass in the SPE-column was also studied in the range from 5to 50 mg. A sample volume of 4 mL and a MWCNT mass of 45 mgwere selected as optimum values. These parameters provided enoughsensitivity for Pb determination in urine and a suitable extraction

time: 4mLof sample (at a flow rate of 1.1mLmin−1, established by tak-ing into account the dead volumes of PTFE tubes and connections) re-quired approximately 5 min.

3.3. Optimization of ETAAS conditions

The direct ETAAS determination of lead in urine suffers from somedrawbacks caused by the organic matrix components of urine samples.Volatilization of the organic phase conducts to an overestimation of thelead concentration, problemswith sample drop dispensation due to thesuperficial tension of the liquid, and the need to use non-aqueous stan-dards in some cases, amongst others [30]. In the present case, urinesamples were subjected to ETAAS measurements after SPE treatment,inwhichmatrix eliminationwas achieved. Therefore, ETAAS determina-tion was carried out using a single method that employed less harshconditions than those required in other methods without SPE. On thebasis of results previously reported [11,31,32], NH4H2PO4 was selectedas an adequate chemical modifier. The effects of mineralization and at-omization temperatures together with the concentration of modifier onthe lead integrated absorbance for eluted urine sampleswere studied inthe ranges indicated in Table S1 (Appendix A) using a Box–Behnkendesign in 15 randomized experiments. As it can be seen from Fig. S4(Appendix A), the optimum values obtained were 350 °C for minerali-zation temperature, 1550 °C for atomization temperature, and 0.005%for NH4H2PO4 concentration. A significant loss of analyte was observedfor atomization temperatures higher than 1650 °C and this gave rise to asignificant diminution in the absorbance signal. The small quantity ofmodifier necessary is a consequence of the sample matrix eliminationperformed in the SPE step. The rest of the furnace parameters, such astemperature ramps and holds, were also studied and optimized takinginto account the values obtained in previous studies and the character-istics of the sample [32]. The optimum furnace operation conditions,producing the highest analytical signal and also well-shaped peaks,are listed in Table 1.

3.4. Evaluation of interferences

The effect that the presence of other different ions with typical cat-ionic sorption has on the Pb(II) concentration was studied. Interferenceexperimentswere carried out under the optimized conditions describedin Section 2.5 with a urine sample solution containing 30 μg Pb L−1.Each interference was evaluated individually and was considered posi-tive if the foreign ions caused a change ofmore than±10% in the recov-ery of lead. The effects of typical transition metal ions, such as Fe(III),Zn(II), Cr(III), Ni(II), Se(IV), Co(II), Cu(II) and Cd(II), were examined inconcentrations up to 10 mg L−1. The results are shown in Table S2(Appendix A). The ions studiedwere tolerated at concentrations consid-erably higher than the levels present in urine samples.

3.5. Column reuse

Column reuse is an important factor to consider for the general ap-plicability of MWCNT-based SPE methods. The long-term stability andthe regeneration capability of the sorption systems for an elevated num-ber of cycles are highly desirable and constitute a useful indication ofapplicability. The stability of MWCNTs was examined by successivelypassing through the column urine samples of 4 mL prepared as indicat-ed in Section 2.4, with subsequent measurement under the optimumconditions established in Section 2.5. After each step, the column wasregenerated with 2 mL of 1 M HNO3 and 4 mL of Milli-Q pure water.The microcolumns allowed up to 50 adsorption–desorption cycleswhile maintaining their extraction efficiency without blockage oroverpressure.

Table 2Accuracy of the proposed method: recovery study.

Pb added (μg L−1) Pb found (μg L−1)a Pb recovery (%)

0 5.31 ± 0.44 –

5 10.59 ± 0.26 10610 15.41 ± 0.51 10230 34.31 ± 1.74 9760 65.11 ± 5.33 10090 97.21 ± 4.94 102

a Mean ± standard deviation (n = 3).

18 R.M. Peña Crecente et al. / Spectrochimica Acta Part B 101 (2014) 15–20

3.6. Analytical figures of merit

In order to evaluate the analytical characteristics of the developedmethod, the linearity, detection and quantification limits, aswell as pre-cision and accuracy were studied. The present method showed a linearresponse for urine samples from LOD up to 120 μg Pb L−1. The equationobtained for calibration in the SPE procedure coupled to ETAAS was:

Aint ¼ 1:10� 10–2 � 1:9� 10–4� �

Pb μg L–1� �h i

þ 8:4� 10–3 � 4:8� 10–3� �

r ¼ 0:999:

In order to verify successful matrix elimination, calibration and stan-dard addition calibration curves were prepared. An Anova-test was per-formed to compare the slopes of the two curves and this showed nosignificant differences between the slopes of the calibration and stan-dard addition graphs (95% confidence level). This fact was also con-firmed by observing the Pb background and peak profiles for both thepure urine sample and the eluted solution obtained after SPE (Fig. 1).It is clear that for urine samples, the peak obtained presented a signifi-cant background whereas the eluted solution gave a higher and well-shaped peak without background. As a result, the Pb measurements inurine were carried out using a calibration curve.

Detection (LOD) and quantification (LOQ) limits were calculated as3SD/m and 10SD/m, respectively, wherem is the slope of the calibrationgraph and SD the standard deviation of 10 consecutive measurementsof blank solutions. The LOD and LOQ values obtained were 0.08 and0.26 μg L−1, respectively, and these are acceptable for themeasurementof Pb at low levels in human urine. The LOD is better (or comparable)than those obtained for other ETAAS determination methods reported

Fig. 1. Lead and background peaks for a urine sample prepared as indicated in Section 2.4containing 6.95 μg L−1 of Pb. (a): without SPE procedure; (b): with SPE procedure. Leadpeaks ( ). Background peaks ( ).

in the literature, except in the case of the work by Sung & Huang [15](5 ng L−1). However, these authors applied a previous sample digestionand used a sophisticated laboratory-made flow injection on-linepreconcentration system.

The precision of the method was evaluated by measuring ten inde-pendent SPE cycles for urine samples spiked with 5, 10, 30, 60, 90 and120 μg L−1 of Pb. The relative standard deviations (RSD) were 4.3, 8.2,4.5, 5.8, 3.4 and 3.4%, respectively. Since the columns used in thiswork were prepared in the laboratory, the reproducibility of columnpreparation was therefore studied in order to estimate if the columnpreparation influences the results. Eight equivalent columns were pre-pared and packed as described in Section 2.3. Aliquots of the sameurine sample were subjected to SPE using the eight columns and thesamples were subsequently measured by ETAAS under the optimizedconditions. The results obtained for the Pb concentration using theeight columns prepared were compared applying an Anova test. Atthe 95% confidence level, the evaluated factor did not have a statisticallysignificant effect on the analytical signal, thus confirming the reproduc-ibility of the results obtained with the different columns.

The accuracy of the proposed method was evaluated by a recoverytest. A urine sample containing 5.31 μg L−1 of Pb, prepared as indicatedin Section 2.4, was supplemented with different quantities of Pb in therange 5–90 μg L−1 and theyweremeasured under the conditions previ-ously described in Section 2.5. From the results in Table 2, it can be seenthat values in the range 97–106%were obtained. Additionally, five urinesamples with low Pb concentration from non-smoker unexposed peo-ple were comparativelymeasured in triplicate: (i) by the presentmeth-od in our laboratory and (ii) by the ICP-MS method after samplemicrowave-assisted acid digestion described by Nakagawa et al. [33]in an external laboratory. The results are presented in Table 3. It canbe seen that good agreement was obtained between the levels deter-mined by both methods. A paired-test showed no statistical differences(at a 95% confidence level) for the results measured.

The enhancement of ETAAS sensitivity due to the simplification ofthe matrix (achieved by inserting the SPE step with MWCNTs) wasevidenced by calculating the signal enhancement factor (SEF). This pa-rameter was calculated as the ratio between the slopes of analyticalcurves obtained with and without the SPE step, with values expressedas a percentage. The SEF achieved was 640%. The entire extraction/elution cycle for one sample requires approximately 25 min. However,it would be straightforward to improve the operation time by using sev-eral columns simultaneously to extract different urine samples with asimple multichannel SPE. In the present case, eight microcolumns

Table 3Accuracy of the proposed method: comparative results between the proposed methodand ICP-MS (after sample digestion).

Sample(n = 3)

Pb found (μg L−1)a

Proposed methodPb found (μg L−1)a

ICP-MS

1 3.64 ± 0.29 3.48 ± 0.412 4.06 ± 0.45 4.32 ± 0.223 5.31 ± 0.44 5.05 ± 0.264 6.02 ± 0.11 6.79 ± 0.935 4.49 ± 0.44 4.79 ± 0.47

a Mean ± standard deviation (n = 3).

Table 4Comparison of the published methods for ETAAS Pb determination with the proposed method in this work.

Techniquea Sample dilution/enrichment procedurea LOD(μg L−1)

Accuracy(%)

RSD(%)

Ref.

ETAAS –/CPE:DDTP 0.04 96–110 – [5]THFA-ETAAS 1:5–1:10/Carbon fiber–tungsten collector 0.2 74–100 ≤2.5 [9]EC-THGA SIMAAS 1:4/– 0.57 – 1.7 [11]ETAAS –/SPE: Ag-nanoparticles + Morin 0.068 97–101 b4.1 [14]FI THGA-SIMAAS Digested sample/SPE:Muromac A-1 resine 0.0045 91–97 3.8 [15]THGA-ETAAS 1:9/USLE 0.7 – 5.7 [30]FI ETAAS 1:2/SPE:DETATA 0.2 97 – [34]ETAAS 1:2/– 0.83 100.16 2.98 [35]ETAAS 2:5/SPE:carbon nanotubes 0.08 96.8–105.6 3.4–8.2 Proposed

a THFA: Transverse heated filter atomized; EC-THGA: End capped transversely heated graphite atomizer; SIMAAS: Simultaneous electrothermal atomic absorption spectrometry; FI:Flow injection; CPE: Cloud point extraction; DDTP: ammonium O,O-diethyl dithiophosphate; USLE: Ultrasonic assisted solid–liquid extraction; DETATA: polystyrene–divinylbenzenecopolymer.

19R.M. Peña Crecente et al. / Spectrochimica Acta Part B 101 (2014) 15–20

were used simultaneously for the analysis of real samples. This modifi-cation allowed 19–20 samples to be processed per hour. The analyticalcharacteristics of the proposed method have been favorably comparedwith other published in the literature as it can be seen in Table 4.

3.7. Analysis of real urine samples

Themethod developed in this workwas applied for determining thePb content in human urine samples fromhealthy peoplewith nohistoryof exposure to lead. Samples were collected in standard plastic urinaly-sis containers according to the recommendations of the NationalCommittee for Clinical Laboratory Standards [6]. Two groups of smokersand non-smokers were established. Urine samples were analyzedaccording to the proposed method. For the non-smoker group, the Pbconcentrations were in the range 3.64 to 15.0 μg L−1, while the concen-trations found in the urine of smokers were higher, ranging between11.6 and 22.9 μg L−1.

4. Conclusions

A SPE procedure usingMWCNTs as sorbentmaterial has been devel-oped for the extraction of Pb from human urine samples prior to ETAASdetermination. Pb(II) ions were retained on oxidized MWCNTs andwere successfully desorbed with a nitric acid solution (carryover wasnot observed in the subsequent analysis). Samples were measured byETAAS and satisfactory analytical figures of merit were obtained. Thenewly developed method has several advantages compared to othermethods for the determination of Pb in urine: (i) the elimination ofthe matrix permitted the measurement of lead in the urine sampleswithout other pretreatment procedures; (ii) the absence of a matrixeffect allowed direct measurement with a calibration graph instead ofa standard addition procedure; and (iii) the low limit of detection oftheproposedmethod (LOD=0.08 μg L−1) allowed the possibility of de-termining Pb in urine at low levels. The remarkable capability ofMWCNTs to be used as SPE sorbents, for the elimination of matrixeffects in ETAAS-based methods for lead determination at ultratracelevels, has been confirmed. The present method has proven to be usefulfor the assessment of occupational exposure to lead.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.sab.2014.07.005.

References

[1] H. Needleman, Lead poisoning, Annu. Rev. Med. 55 (2004) 209–222.[2] C. Zeng, X. Wen, Z. Tan, P. Cai, X. Hou, Hollow fiber supported liquid membrane

extraction for ultrasensitive determination of trace lead by portable tungsten coilelectrothermal atomic absorption spectrometry, Microchem. J. 96 (2010) 238–242.

[3] S. Hernberg, Lead poisoning in a historical perspective, Am. J. Ind. Med. 38 (2000)244–254.

[4] M. Yaman, Determination of cadmium and lead in human urine by STAT-FAAS afterenrichment on activated carbon, J. Anal. At. Spectrom. 14 (1999) 275–278.

[5] T.A. Maranhão, E. Martendal, D.L.G. Borges, E. Carasek, B. Welz, A.J. Curtius, Cloudpoint extraction for the determination of lead and cadmium in urine by graphitefurnace atomic absorption spectrometry with multivariate optimization usingBox–Behnken design, Spectrochim. Acta B 62 (2007) 1019–1027.

[6] P.J. Parsons, J.J. Chisolm, H.T. Delves, R. Griffin, E.W. Gunter, W. Slavin, N.V. Stanton,R. Vocke, C40-A: Analytical Procedures for the Determination of Lead in Blood andUrine; Approved Guideline, National Committee for Clinical Laboratory Standards,Wayne, PA, USA, 2001.

[7] WHO, Brief Guide to Analytical Methods for Measuring Lead in Blood, World HealthOrganization, Geneva, Switzerland, 2011. ((Available at: http://www.who.int/ipcs/assessment/public_health/lead_blood.pdf ). Accessed 30 January 2014).

[8] N. Campillo, P. Viñas, I. López-García, M. Hernández-Córdoba, Rapid determinationof lead and cadmium in biological fluids by eletrothermal atomic absorptionspectrometry using Zeeman correction, Anal. Chim. Acta. 390 (1999) 207–215.

[9] P. Ngobeni, C. Canario, D.A. Katskov, Y. Thomassen, Transverse heated filter atomiz-er: atomic absorption determination of Pb and Cd in urine, J. Anal. At. Spectrom. 18(2003) 762–768.

[10] I.L. Grinshtein, Y.A. Vilpan, A.V. Saraev, L.A. Vasilieva, Direct atomic absorption deter-mination of cadmium and lead in strongly interfering matrices by double vaporiza-tion with a two-step electrothermal atomizer, Spectrochim. Acta B 56 (2001)261–274.

[11] P.R.M. Correia, C.S. Nomura, P.V. Oliveira, Multielement determination of cadmiumand lead in urine by simultaneous electrothermal atomic absorption spectrometrywith an end-capped graphite tube, Anal. Sci. 19 (2003) 1519–1523.

[12] Z. Čánký, P. Rychloský, Z. Petrová, J.P. Matousek, A technique coupling the analyteelectrodeposition followed by in situ stripping with electrothermal atomic absorp-tion spectrometry for analysis of samples with high NaCl contents, Spectrochim.Acta B 62 (2007) 250–257.

[13] T.Q. Shi, P. Liang, J. Li, Z.C. Jiang, B. Hu, Nanometer TiO2 separation/preconcentrationand GFAAS determination of trace Pb in water samples, Chin. J. Anal. Chem. 32(2004) 1421–1423.

[14] M. Khajeh, E. Sanchooli, Silver nanoparticles as a new solid phase adsorbent and itsapplication to preconcentration and determination of lead from biological samples,Biol. Trace Elem. Res. 143 (2011) 1856–1864.

[15] Y.H. Sung, S.D. Huang, On-line preconcentration system coupled to electrothermalatomic absorption spectrometry for the simultaneous determination of bismuth,cadmium, and lead in urine, Anal. Chim. Acta. 495 (2003) 165–176.

[16] S. Ijima, Helical microtubules of graphitic carbon, Nature 354 (1991) 56–58.[17] A.V. Herrera-Herrera, M.A. González-Curbelo, J. Hernández-Borges, M.A. Rodríguez-

Delgado, Carbon nanotubes applications in separation science: a review, Anal. Chim.Acta. 734 (2012) 1–30.

[18] C. Herrero Latorre, J. Álvarez Méndez, J. Barciela García, S. García Martín, R.M. PeñaCrecente, Carbon nanotubes as solid-phase extraction sorbents prior to atomicspectrometric determination of metal species: a review, Anal. Chim. Acta. 749(2012) 16–35.

[19] A.F. Barbosa, M.G. Segatelli, A.C. Pereira, A.S. Santos, L.T. Kubota, P.O. Luccas, C.R.T.Tarley, Solid-phase extraction system for Pb(II) ions enrichment based on multiwallcarbon nanotubes coupled on-line to flame atomic absorption spectrometry, Talanta71 (2007) 1512–1519.

[20] S. Chen, C. Liu, M. Yang, D. Lu, L. Zhu, Z. Wang, Solid-phase extraction of Cu, Co andPb on oxidized single-walled carbon nanotubes and their determination by induc-tively coupled plasma mass spectrometry, J. Hazard. Mater. 170 (2009) 247–251.

[21] M. Savio, B. Parodi, L.D. Martínez, P. Smichowski, R.A. Gil, On-line solid phase extrac-tion of Ni and Pb using carbon nanotubes and modified carbon nanotubes coupledto ETAAS, Talanta 85 (2011) 245–251.

[22] H. Tavallali, H. Asvad, Flame atomic absorption spectrometric determination of traceamounts of Pb(II) after solid phase extraction using multiwalled carbon nanotube,Int. J. Chem. Tech. Res. 3 (2011) 1635–1640.

[23] A. Duran, M. Tuzen, M. Soylak, Preconcentration of some trace elements via usingmultiwalled carbon nanotubes as solid phase extraction, J. Hazard. Mater. 169(2009) 466–471.

20 R.M. Peña Crecente et al. / Spectrochimica Acta Part B 101 (2014) 15–20

[24] R. Li, X. Chang, Z. Li, Z. Zang, Z. Hu, D. Li, Z. Tu, Multiwalled carbon nanotubesmodified with 2-aminobenzothiazole modified for uniquely selective solid-phaseextraction and determination of Pb(II) ion in water samples, Microchim. Acta 172(2011) 269–276.

[25] Y. Cui, S. Liu, Z.J. Hu, X.H. Liu, H.W. Gao, Solid-phase extraction of lead(II) ions usingmultiwalled carbon nanotubes grafted with tris(2-aminoethyl)amine, Microchim.Acta 174 (2011) 107–113.

[26] Z. Zang, Z. Hu, Z. Li, Q. He, X. Chang, Synthesis, characterization and application ofethylenediamine modified multiwalled carbon nanotubes for selective solid-phaseextraction and preconcentration of metal ions, J. Hazard. Mater. 172 (2009)958–963.

[27] J. Álvarez Méndez, J. Barciela García, R.M. Peña Crecente, S. García Martín, C. HerreroLatorre, A new flow injection preconcentration method based on multiwalled car-bon nanotubes for the ETA-AAS determination of Cd in urine, Talanta 85 (2011)2361–2367.

[28] H.P. Boehm, Some aspects of the surface chemistry of carbon blacks and othercarbons, Carbon 32 (1994) 759–769.

[29] E. Hosten, B. Welz, Evaluation of an immobilised macrocyclic material for on-linecolumn preconcentration and separation of cadmium, copper and lead for electro-thermal atomic absorption spectrometry, Anal. Chim. Acta. 392 (1999) 55–65.

[30] D. Guimarães, J.P. Santos, M.L. Carbalho, G. Vale, H.M. Santos, V. Geraldes, I. Rocha, J.L. Capelo, Ultrasonic energy as a tool to overcome some drawbacks in the determi-nation of lead in brain tissue and urine in rats, Talanta 86 (2011) 442–446.

[31] P.J. Parsons, W. Slavin, Electrothermal atomization atomic absorption spectrometryfor the determination of lead in urine: results of an interlaboratory study,Spectrochim. Acta B 54 (1999) 853–864.

[32] J.C. Rodríguez García, J. Barciela García, C. Herrero Latorre, S. García Martín, R.M.Peña Crecente, Comparison of palladium–magnesium nitrate and ammoniumdihydrogenphosphate modifiers for lead determination in honey by electrothermalatomic absorption spectrometry, Food Chem. 91 (2005) 435–442.

[33] J. Nakagawa, Y. Tsuchiya, Y. Yashima, M. Tezuka, Y. Fujimoto, Determination of tracelevels of elements in urine by inductively coupled plasma mass spectrometry, J.Health Sci. 50 (2004) 164–168.

[34] L.A. Vasil'eva, I.L. Grinshtein, S. Gucer, B. Izgi, Determination of lead and cadmium inurine by electrothermal atomization atomic spectrometry, J. Anal. Chem. 63 (2008)649–654.

[35] P. Olmedo, A. Pla, A.F. Hernández, O. López-Guarnido, L. Rodrigo, F. Gil, Validation ofa method to quantify chromium, cadmium, manganese, nickel and lead in humanwhole blood, urine, saliva and hair samples by electrothermal atomic absorptionspectrometry, Anal. Chim. Acta. 659 (2010) 60–67.