AS 2 8( 6) 195

Forsubscriptions

and/or articlessee inside

front cover.

ASPND7 38(3) 37 – 75 (2017)ISSN 0195-5373

AtomicSpectroscopy

May/June 2017 Volume 38, No. 3

In This Issue:Direct Determination of Trace Lead in Seawater by Inductively Coupled PlasmaMass Spectrometry After Photochemical Vapor GenerationShuzhen Li, Ying Gao, Ying Yu, Hongyan He, Xiaorong Hu, Shijun Ni,Zeming Shi, Xiuhong Peng, and Rui Liu ............................................................... 37

Detection of Shigella flexneri DNA by ICP-MS Based on OligonucleotideHybridization and Labeling of Gold NanoparticlesXiuji Wang, Lanlan Jin, Wei Guo, Liuqin Huang, and Shenghong Hu ................ 44

Application of Supramolecular Microextraction and Flame Atomic AbsorptionSpectrometry for Ultra-trace Determination of Cadmium in Food andWater SamplesZ.A. ALOthman, M.A. Habila, Erkan Yilmaz, M. Soylak, and A.A. Alwarthan ..... 51

Magnetic Solid Phase Extraction of Trace Lead and Copper on Chromotrope FBImpregnated Magnetic Multiwalled Carbon Nanotubes From Cigarette andHair Samples for Measurement by Flame AASMustafa Soylak and Zeliha Erbas ........................................................................... 57

Determination of Cobalt, Iron, and Nickel in High-Purity Silicon by High-Resolution Continuum Source Graphite Furnace Atomic Absorption SpectrometryEmploying Solid Sample AnalysisMarcos André Bechlin, Ariane Isis Barros, Diego Victor Babos,Edilene Cristina Ferreira, and José Anchieta Gomes Neto .................................. 62

Feasibility of a Fast and Green Chemistry Sample Preparation Procedure forthe Determination of K and Na in Renewable Oilseed Sources by FlameAtomic Emission SpectrometryKamyla Cabolon Pengo, Vanessa Cruz Dias Peronico, Luiz CarlosFerreira de Souza, and Jorge Luiz Raposo, Jr. ....................................................... 68

PerkinElmer, ELAN, NexION, and Chromera are registered trademarks and AAnalyst, DRC,Dynamic Reaction Cell, Flat Plate, TemTip, PinAAcle, and Titan MPS are trademarks ofPerkinElmer, Inc.Dionex is a trademark and Nalgene is a registered trademark of Thermo Fisher Scientific, USA.Excelsa and Orion are trademarks of Fanem, Brazil.Extran is a registered trademark of Merck Millipore Corporation.CertiPUR and Suprapur are trademarks of Merck & Co., Germany.Microcal Oriinal Pro is a registered trademark of OriginLab Corporation.Microsoft and Office Excel are registered trademarks of Microsoft Corporation.Millipore, Milli-Q Academic, and Rios are registered trademarks of Millipore Corporation.Multiwave is a registered trademark of Anton Paar, Graz, Austria.SpecPure is a registered trademark of Johnson-Matthey Chemicals, Ltd., UK.Pyrex is a registered trademark of Corning Glass Works, USA.Registered names and trademarks used in this publication, even without specific indicationthereof, are not to be considered unprotected by law.

Printed in the United States and published six times a year by:

PerkinElmer, Inc.710 Bridgeport AvenueShelton, CT 06484-4794 USAhttp://www.perkinelmer.com

AtomicSpectroscopy

Vol. 38(3), May/June 2017

Guidelines for AuthorsAtomic Spectroscopy serves asa medium for the disseminationof general information togetherwith new applications andanalytical data in atomicabsorption spectrometry.The pages of Atomic

Spectroscopy are open to allworkers in the field of atomicspectroscopy. There is no chargefor publication of a manuscript.The journal has around 1100

subscribers on a worldwidebasis, and its success can beattributed to the excellentcontributions of its authors aswell as the technical guidanceof its reviewers and theTechnical Editors.The original of the manuscript

can be mailed to the editor inhard copy including electronicfile on disk or CD (or simply bye-mail) in the following manner:1. If mailed, provide text (dou-ble-spaced) and tables in hardcopy plus on disk or CD withtext and tables in .doc file; fig-ures in doc or tif files.3. Number the references in theorder they are cited in the text.

5. Consult a current copy ofAtomic Spectroscopy forformat.

6. Editor’s e-mail:[email protected]: [email protected]

All manuscripts are sent to tworeviewers. If there is disagreement,a third reviewer is consulted.Minor changes in style are made

in-house and submitted to theauthor for approval.If a revision of the manuscript

is required before publication canbe considered, the paper isreturned to the author(s) withthe reviewers’ comments.In the interest of speed of

publication, a pdf file of the type-set text is e-mailed to the corre-sponding author beforepublication for final approval.Once the issue has been printed,

each author receives a final pdf fileof the article including 50 compli-mentary copies of the article andseveral copies of the completeissue.Additional reprints can be

purchased, but the requestmust be made before printing.

PerkinElmer. Inc., holds copy-right to all material published inAtomic Spectroscopy unless other-wise noted on the first page of thearticle.

Anneliese LustEditor, Atomic SpectroscopyPerkinElmer, Inc.710 Bridgeport AvenueShelton, CT 06484-4794 USA

EDITORAnneliese LustE-mail:[email protected]@aol.com

TECHNICAL EDITORS FORAA, ICP-OES, ICP-MSWei GuoKenneth R. Neubauer

SUBSCRIPTION INFORMATION

Atomic SpectroscopyP.O. Box 3674Barrington, IL 60011 USAE-mail: [email protected]

2017 SUBSCRIPTION RATES

ORDER ON-LINE:• pdf only -1 yr - U.S. $60.00• 2 yr - U.S. $110.00• pdf and printed copy -

1 yr - U.S. $90.002 yr - U.S. $160.00

• printed copy only -1 yr - U.S. $70.002 yr - U.S. $130.00

PAYMENT BY CHECK:Check must be drawn on U.S.bank in U.S. funds made out to:“Atomic Spectroscopy”

BACK CLAIMS/ADDRESS CHANGES:• Subscriber claims for missingback issues will be honoredat no charge within 90 daysof issue mailing date.E-mail: [email protected]

Copyright © 2017PerkinElmer, Inc.All rights reserved.http://www.perkinelmer.com

ARTICLE INDEXING SERVICESThomson Reuters at:www.scientific.thomsonreuters.comElsevier BV at:www.elsevier.com

37

*Corresponding author.E-mail: [email protected].: 86-28-8407-8773

Atomic SpectroscopyVol. 38(3), May / June 2017

INTRODUCTION

The presence of heavy metalsposes a potential threat to marineecosystems due to their toxicity,resistance to degradation, and con-sequent tendency to bioaccumula-tion (1, 2). Among them, lead (Pb)is toxic to humans, aquatic plants,and animals at trace levels (3-5) andis widely regarded as a good indica-tor of heavy metal pollution inmarine environments caused byanthropogenic activities (6, 7). Elu-cidation of the distribution and con-centration of lead in seawater is ofgreat importance for identificationof the contamination source andcontamination prevention (1, 8, 9).

Atomic spectrometry-basedtechniques are the most popularmethods for the determination oftrace metals in environmental sam-ples. Inductively coupled plasmamass spectrometry (ICP-MS) is oneof the most powerful techniquesdue to its low detection limits,wide linear range, and high samplethroughput (10, 11). However,direct determination of lead in sea-water samples remains a challengedue to their extremely low concen-trations and high salt matrix (12-14). Therefore, preconcentrationand purification before detection isrequired for the analysis of lead inseawater. Solvent extraction andextraction onto chelating resincolumns followed by atomic spec-trometry detection were developed

ABSTRACT

A novel method is developedfor the determination of lead inseawater samples using nickel-assisted photochemical vaporgeneration (PVG) coupled withinductively coupled plasma massspectrometry (ICP-MS). Thesevere matrix effect of seawateron the suppression of the leadsignal was efficiently eliminatedby using a mixture of 15 µg g-1

Ni2+ and 10% (v/v) formic acidas a photochemical reductionmedium, making direct determi-nation of lead in seawater sam-ples feasible. A method detectionlimit of 0.003 ng g−1 based onexternal calibration wasobtained, and the sampling fre-quency of 20 h-1 was achievedwith a 30-second sample loadingtime and a 1.25-mL sample con-sumption. The relative standarddeviation of the measurementresults was 3.7% (RSD, n=7) inseawater spiked with 1 ng g−1

Pb2+ solution. The proposedmethod was successfully appliedfor the analysis of one standardreference material (Seawater-QC3163) and three seawatersamples (collected from Shang-hai, Haikou, and Sanya, P.R.China) with satisfactory results.

was also proposed for the determi-nation of lead in seawater (18).Nevertheless, it could result in highcontent of sea salt in a samplematrix which can clog the nebu-lizer and leave deposits on the ICP-MS cones.

Chemical vapor generation(CVG) is a widely used sampleintroduction method for tracemetal detection in atomicspectrometry which can greatlyincrease sample introduction effi-ciency and efficiently separate theanalyte from its troublesome sam-ple matrix (19). Conventional CVGmethods often involve using highblank oxidation reagents and unsta-ble reduction reagents, making itdifficult to determine trace/ultra-trace concentrations of Pb in envi-ronmental samples. Purification ofreagents before use or high purityreagents is always required (20,21). Photochemical vapor genera-tion (PVG), utilizing free radicalsgenerated by photoredox reactionsin the presence of low molecularweight organic compounds, notonly remains the major advantageof CVG but also provides simplerreactions (22-24). For lead determi-nation, metal ion-assisted PVG wasdeveloped for the analysis of sedi-ments, soils, and river waters (25).In the presence of Ni2+, a signifi-cant improvement in PVGefficiency of lead was achieved.The proposed method is simple,with low blanks compared to con-ventional CVG for lead, and avoidsunstable reagents. However, thesevere matrix effect is still a prob-lem for the direct determination of

(15-17). But these methods are usu-ally relatively labor-intensive (i.e.,reagent purification and sampleprocessing) and time-consuming asthey often require large sample vol-umes (hundreds of milliliter to literscale). A Mg(OH)2 co-precipatationmethod with low procedural blank

Direct Determination of Trace Lead in Seawater byInductively Coupled Plasma Mass Spectrometry

After Photochemical Vapor GenerationShuzhen Lia,c, Ying Gaoa,b,*, Ying Yub, Hongyan Heb, Xiaorong Huc, Shijun Nib,

Zeming Shib, Xiuhong Pengb, and Rui Liuc

a State Key Laboratory of Geohazard Prevention and Geoenvironment Protection,Chengdu University of Technology, Chengdu, Sichuan 610059, P.R. China

b College of Earth Sciences, Chengdu University of Technology, Chengdu, Sichuan 610059, P.R. Chinac College of Materials and Chemistry & Chemical Engineering,

Chengdu University of Technology, Chengdu 610059, P.R. China

µg g-1 were purchased from Envi-ronmental Express (South Carolina,U.S.). Lead working standard solu-tion was prepared by dilution ofthe stock solution with 0.1%HCOOH. High-purity NaCl (traceSELECT, >99.999%) was obtainedfrom Aladdin Industrial Corporation(Shanghai, P.R. China). A 2 M BrClsolution was prepared in a fumehood by dissolution of 27 g ofreagent grade KBr (Fisher Scientific,Ottawa, Canada) in 2.5 L of HCl ina glass container, followed by slowaddition of 38 g reagent gradeKBrO3 (Fisher Scientific, Ottawa,Canada). A rinse solution contain-ing 0.04 M BrCl was prepared bydilution of the 2 M BrCl with deion-ized water (DIW) to efficiently elim-inate any carry-over between thesamples. A certified reference mate-rial QC3163 Seawater was obtainedfrom Sigma Aldrich for method vali-dation.

Three seawater samples wereanalyzed for method validation,which were collected from Sanya(Hainan, P.R. China), Haikou(Hainan, P.R. China), and Shanghai(Chongmingdao, P.R. China),respectively.

Standard Additions SamplePreparation and AnalysisProcedure

For the determination of lead inseawater samples, standard additioncalibration was applied. Three repli-cate subsamples of 10 g of each sea-water sample were weighed intoprecleaned polyethylene bottles.A 10-g mass of a solution containing20% formic acid (FA) and 0.3 g ofa 1000-g g-1 Ni stock solution wereadded to each sample, resulting ina final concentration of 10% FA and15 µg g-1 Ni2+. Appropriate amountsof the standard solution of Pb2+

were added to each spiked seawa-ter sample to obtain approximatelya 1-, 2-, and 4-fold increase in thePb concentration, respectively. Theunspiked and spiked samples werethen diluted to 20 g with DIW.

lead in seawaters. The signal fromNi2+ was significantly suppressedin the presence of high concentra-tions of chloride, resulting in lowPVG efficiency of Ni as well as agreatly decreased PVG efficiencyof lead. Recently, Duan et al. (26)proposed a new method for thedetermination of trace lead by gen-erating volatile lead species in thepresence of 0.90% (v/v) acetic acidand 0.03% (v/v) hydrochloride acidunder UV irradiatinon. This tech-nique is simple, convenient, andhas low blank. But the suppressioneffect of high concentration inor-ganic anions is also a problem forthe direct analysis of seawater sam-ples.

The purpose of this work was todevelop a sensitive, simple, anddirect method for the accuratedetermination of trace lead in sea-water samples by nickel ion-assistedPVG coupled with ICP-MS detec-tion. The accuracy of the proposedmethod is demonstrated by success-ful analysis of three seawaters.

EXPERIMENTAL

Instrumentation

An ELAN® DRC™-e ICP-MS(PerkinElmer, Inc., Shelton, CT,USA), equipped with quartz torchand alumina sample injector tube,was used. The fitted GemTip™cross-flow nebulizer and a corro-sion-resistant double-pass Ryton®spray chamber mounted outsidethe torch box were replaced by aPVG-system for lead determination.A model FIA-3110 flow injectionpump system (Vital InstrumentsCo. Ltd., Beijing, P.R. China) wasused for the introduction of samplesolutions to the PVG system.A schematic of the UV-PVG pho-toreactor interfaced to the ICP-MSis similar to a previous report (25).The UV-PVG system consists of a19 W thin film flow-through lamp(Beijing Titan Instruments Co., Bei-jing, P.R. China) loosely coveredwith aluminum foil to shield the

operator from exposure to the UV.Argon carrier gas was introducedthrough a “T” connection betweenthe outlet of the photoreactor anda tandem set of two homemade gas-liquid separators (GLSs, ~2 mLinternal volume) maintained at 0 oCby immersion in an ice bath to min-imize any transport of liquiddroplets derived from condensationof water vapor to the ICP. The gen-erated analyte vapors were directedfrom the outlet of the last GLS tothe ICP-MS via a 0.25-m length ofTeflon® lined Tygon® tubing.Optimization of the ICP-MS parame-ters was performed as recommendedby the manufacturer. Optimumconditions for PVG were investi-gated independently. Typical oper-ating conditions are summarized inTable 1.

Reagents and StandardSolutions

All reagents were of analyticalreagent grade or better, and deion-ized water (DIW) was used through-out. ACS grade formic acid wasobtained from Aladdin IndustrialCorporation (Shanghai, P.R. China).Stock solutions of lead and nickelwith the concentration of 1000

38

TABLE IICP-MS Instrumental

Operating Parameters

Instrument Settings ICP-MS

RF power 1175 Wplasma gas flow 15 L min-1

Auxiliary gas flow 1.2 L min-1

Nebulizer gas flow 1.0 L min-1

Scanning mode Peak hoppingIsotope monitored 208PbResolution 0.7 amuDwell time 30 ms

Dead time 50 ns

UV-PVG

Sample flow rate 2.5 mL min-1

Ar carrier gas flowrate to PVG 1.0 L min-1

The signal response of 208Pbincreased with an increasing con-centration of formic acid from 1%to 10% (v/v), and then decreasedbeyond 10%. Obviously, the opti-mum concentration of formic acidwas higher than that of a previousreport (5%, v/v). It suggests thatreductive radicals arising from pho-todecomposition of formic acidwere largely consumed by poten-tially oxidative radicals generatedfrom the sample matrix, decreasingthe available reductive radicals forthe reduction of lead. Therefore, ahigher concentration of formic acidwas required for the efficient reduc-tion of lead in seawater comparedto that in a standard solution. But afurther increase in concentration offormic acid above 10% (v/v) maylose penetration depth of the UVradiation into the sample as theabsorption of the solutionincreases. In addition it may alsoincrease possible side reactions orphotodecomposition of the prod-uct and its redissolution in theaqueous phase. Consequently, 10%(v/v) formic acid was selected forsubsequent measurement.

Metal ions play important rolesfor the reduction of Pb in a PVGsystem. The presence of Ni2+ informic acid solution can signifi-cantly improve the PVG efficiencyof Pb (25). Recently, ferric ion-

Three sample blanks were preparedcontaining 10% FA, 15 µg g-1 Ni,and 1.75 % NaCl to match thematrix of the seawater samples.The sample solutions were sub-jected to UV-PVG-ICP-MS detectionas described earlier. Briefly, thesample solution was quickly deliv-ered to the PVG reactor at a sampleflow rate of 2.5 mL min-1 andremained in the reactor for 50 sec-onds by stopping the pump forphotochemical reduction of Pb intovolatile species. Then the irradiatedsample solution was transferred tothe GLSs from the PVG reactor foranalysis by ICP-MS at the same sam-ple flow rate. For the next analysis,a solution of 0.04 M BrCl was usedto efficiently rinse the system for50 seconds to eliminate any carry-over between the samples. Thepeak area of the 208Pb isotope wasused to construct a standard addi-tions calibration curve for eachsample.

RESULTS AND DISCUSSION

Serious suppression effects oftenoccur when detecting metal ions inseawater samples by the PVG-basedtechnique (27-29). According to theprevious report, volatile species ofPb can be efficiently generated in5% formic acid and 3 µg g-1 Ni2+

under UV irradiation (25). Duringinitial experiments, a 5-ng g-1 Pbstandard solution containing 5%formic acid, 3 µg g-1 Ni, and 1.75%NaCl (w/v) (matching matrix in theprepared seawater solutions) wasintroduced to the PVG reactor. Theresponse from the solution wasonly 10% of that obtained froma matrix-free standard solution,showing the severe suppressioneffect of the sample matrix on leaddetection in seawater samples. Itwas found that the PVG efficiencyof Ni2+ was also significantly sup-pressed in the presence of highconcentrations of chloride. It maybe that the oxidative radicals gener-ated from the sample matrix largelyconsumed the reductive radicals

arising from photodecompositionof formic acid, resulting in the lowPVG efficiencies of Ni and Pb.Increasing the concentration offormic acid could provide moreavailable reductive radicals for thephotochemical reduction of theanalytes in seawater samples andpotentially alleviate the effect of theseawater matrix. Due to the sup-pression effect of the samplematrix, spiked seawater instead ofstandard solution was used to inves-tigate optimum PVG conditions forthe determinationof Pb.

Optimization of UV-PVG system

The formic acid concentration,Ni2+ concentration, UV irradiationtime, and argon carrier gas flowrate comprised the four basic para-meters which determined the PVGefficiency of the analyte and thetransport efficiency to the ICP-MS.These parameters were investigatedto obtain optimized responses onlead detection.

PVG efficiencies highly dependon the type and concentration oforganic reductants used (28, 30-32).The effect of formic acid concentra-tion in the range of 1%–40% (v/v)on the response (peak area normal-ized to the maximum achieved) ofPb in seawater spiked with 5 ng g-1

Pb solution is shown in Figure 1.


39

Fig. 1. The effect of formic acid concentration on lead detection.

induced enhancement of PVG wasalso observed for As detection (33).The effect of added Ni2+ concentra-tion in the range of 0–30 µg g-1 onthe response (peak area normalizedto the maximum achieved) of Pb inseawater spiked with 5 ng g−1 Pbsolution is shown in Figure 2. Thepeak area of 208Pb increased withan increase in Ni2+ concentrationfrom 0 µg g-1 to 15 µg g-1, followedby a slight decrease at higher con-centrations. The optimal concen-tration of Ni2+ for the efficientgeneration of volatile Pb in seawa-ter was five times higher than thatobtained from standard solution,which was likely owing to thesevere suppression of the genera-tion of volatile Ni2+ in the presenceof high concentrations of Cl–.Although the mechanism of metal-ion-assisted PVG reaction was notclear, the generation of significantamounts of volatile Ni species wereneeded to enhance charge transferreactions of Pb and to facilitatethe reduction of Pb. Therefore,15 µg g-1 Ni2+ was selected for allsubsequent measurements.

The dosage of UV radiationreceived by sample solution mainlydetermines the efficiencies of radi-cal formation and the vapor genera-tion of lead. The effect of UVirradiation time over the range of30–150 seconds was investigated

using seawater spiked with 5 ng g−1

Pb containing 10% formic acids and15 µg g-1 Ni2+. As shown in Figure 3,the responses initially sharplyincreased as irradiation timeincreased from 30 to 50 seconds.Thereafter, a quick decreaseoccurred with a further increase inirradiation time. The relatively shortirradiation time leads to inefficientgeneration of volatile Pb species,and with increased irradiation timeleads to the reoxidation of volatilePb species by photo-decomposedoxidative radicals from the samplematrix. The optimal UV irradiationtime of Pb in seawater samples wasmuch less than reported previously(25). This probably is due to thecombined effect of sample matrixand the air temperature around theUV lamp. After UV irradiation with50 seconds at room temperature(25 oC), the temperature of thesample solution through the UVreactor was nearly 60 oC. When amini-fan was set under the UV reac-tor to accelerate circulation of airaround the UV lamp, the optimalUV irradiation time was delayed to100 seconds. Furthermore, the tem-perature of the sample solutionafter 50 seconds irradiation wasdown to 48 oC, which suggestedthat the air temperature around theUV lamp influenced the kinetics ofPVG for Pb. Therefore, an irradia-

tion time of 50 seconds (at 25 oC)was selected for all subsequentmeasurements to achieve a maxi-mum response and reasonable sam-ple throughput.

The argon carrier gas flow ratethrough GLS influences the liquidgas separation efficiency of thevolatile Pb species as well as thedepth of sampling in the plasma.The effect of argon carrier gas flowrate on the response of Pb fromspiked seawater was optimized inthe range of 0.8–1.1 L min−1. Asshown in Figure 4, maximumresponse was obtained at the1.0 L min−1 gas flow rate. Lowercarrier gas flow rates may lead toinefficient stripping and transportof the analyte from the solution tothe ICP-MS; whereas higher flowrates will result in significant dilu-tion of the analyte in the carriergas. Therefore, the Ar carrier gasflow rate to the GLS of 1.0 L min−1

was selected for all subsequentmeasurements.

Figures of Merit

Under optimal experimental con-ditions, the signal response of leadfrom 1.75% NaCl (w/v) (matchingthe matrix in the prepared seawatersolution) was approximately 90% ofthat from the same concentrationof the matrix-free standard solution

40

Fig. 2. The effect of Ni2+ concentration on lead detection. Fig. 3. The effect of UV irradiation time and temperature onlead detection.

41


with 5% formic acid and 3 µg g-1

Ni2+. A relatively good precision of3.7% RSD was obtained from thereplicate measurements of the sea-water samples spiked with 1 ng g−1

Pb (n = 7). The sampling frequencyof 20 h−1 was achieved with a 30-second sample loading time and a1.25-mL sample consumption foreach analysis. Good linear responsewas obtained from Pb standardsolutions prepared in the range of0.02−200 ng g−1 in 1.75% NaCl.A method detection limit of 0.003ng g−1 was obtained using externalcalibration, based on three timesthe standard deviation of the ana-lyte concentration arising from themethod blank which comprises asolution of 1.75% NaCl, 10% formicacid, and 15 µg g-1 of Ni2+ in DIW(see Figure 5). Compared to photo-CVG-MC-ICP-MS determination, theLOD is slightly improved, which is0.005 ng g-1. It is probably due tothe better measurement precisionof MC-ICP-MS. But it is 30 timeslower than with the direct solutionnebulization method after 10-folddilution prior to analysis (0.09ng g-1) for the analysis of seawatersamples (Table II). Also, the pro-posed method is more sensitivethan most of the traditional meth-ods with the HG technique, and iscomparable with methods by ICP-MS detection after solid phase

extraction. In addition, the methodblank can be easily controlled asper the latest report for Pb detec-tion by the PVG method (26). Butthis method has great advantagesfor the direct analysis of seawatersamples.

Pb Determination in Seawater

To compensate for matrix effectsand achieve accurate results, stan-

dard additions were used for seawa-ter analysis. Three seawater sam-ples were analyzed under optimalexperimental conditions. Seawatersolutions with gravimetric additionsof approximately 1-, 2-, and 4-foldthe endogenous Pb content wereprepared, yielding a calibrationcurve with a correlation coefficientgreater than 0.999. The analyticalresults are listed in Table III.

Fig. 4. The effect of Ar carrier gas flow rate on lead detection. Fig. 5. PVG calibration curve generated from standard solutionscontaining 1.75% NaCl, 10% formic acid, and 15 µg g-1 Ni2+.

TABLE IIComparison of LODs for Pb Using Different Techniques

Chemical Method Precon- Samples LOD Ref.Reagents centration

- ICP-MS PVC-packed Seawater 7 pg g-1 (14)mini-column

- ICP-OES Ionic Seawater 1.88 ng g-1 (16)imprintedpolymer

0.28 mol L-1 H2O2,0.1 mol L-1 HNO3,and 1.5% (m/v) NaBH4 HG-ICP-MS - Seawater 0.024 ng g-1 (34)

1% (v/v) HCl, HG-ICP MS - Seawater 0.008 ng g-1 (21)2% (m/v) K3Mn(CN)6,and 2% (m/v) NaBH4

0.90% acetic acid (v/v) PVG-ICP-MS- Tap Water 0.0036 ng g-1 (26)and 0.030% HCl (v/v) and Lake

Water

15 µg g-1 Ni2+ and PVG-ICP-MS- Seawater 0.003 ng g-1 This10% (v/v) formic acid work

42

To further demonstrate the ana-lytical accuracy of this method, thespiked recovery test was applied.The recoveries of spiking 0.3 ng g−1

and 1.00 ng g−1 of Pb into the sea-water samples were in the range of95.2–108.9%. The developedmethod was also applied for theanalysis of a certified seawater sam-ple. The analytical result is shownin Table IV and is in agreementwith the certified values, confirm-ing the accuracy of the proposedmethodology.

CONCLUSION

A novel and sensitive approach isdeveloped for the direct determina-tion of Pb in seawaters using nickel-assisted PVG for sample introduc-tion with ICP-MS detection. Thesevere matrix effect from seawaterwas efficiently eliminated by using10% (v/v) formic acid and 15 µg g-1

Ni2+ as the photochemical reduc-tant medium, making direct deter-mination of Pb in seawater feasible.The method provides a detectionlimit of 0.003 ng g-1 for Pb usingexternal calibration, suitable forultratrace determinations of Pb inseawaters. The proposed methodhas promising application for thedetermination of trace Pb in seawa-ter samples.

ACKNOWLEDGMENT

Sichuan Youth Science andTechnology Foundation (No.2017JQ0043), China PostdoctoralScience Foundation (No.2016M590870), State Key Labora-tory of Geohazard Prevention andGeoenviroment ProtectionIndependent Research Project(SKLGP2016Z006), the EducationDepartment of Sichuan Province(Grant No. 17ZA0040) and the Cul-tivating Program of Middle-agedBackbone Teachers Program ofChengdu University of Technology(Grant No. KYGG201409) areacknowledged for their financialsupport. The authors declare nocompeting financial interest.

Received January 11, 2017.

REFERENCES

1. C. Christophoridis, D. Dedepsidis,and K. Fytianos, J. Hazard. Mater.168, 1082 (2009).

2. Y. Dong, R.K. Rosenbaum, and M.Z.Hauschild, Environ. Sci. & Technol.50, 269 (2016).

3. K. Chen, L. Huang, B. Yan, H. Li, H.Sun, and J. Bi, Environ. Sci. & Tech-nol. 48, 12930 (2014).

4. D. Absalon and B. Slesak, Science ofthe Total Environment 408, 4420(2010).

5. R.L. Boeckx, Anal. Chem. 58, 274A(1986).

6. Y. Li, R.J. Yang, A.B. Zhang, and S.R.Wang, Mar. Pollut. Bull. 85, 700(2014).

7. M. Komarek, V. Ettler, V. Chrastny,and M. Mihaljevic, Environ. Int. 34,562 (2008).

8. M.F. Soto-Jimenez, F. Paez-Osuna, G.Scelfo, S. Hibdon, R. Franks, J.Aggarawl, and A.R. Flegal, MarineEnvironmental Res. 66, 451 (2008).

9. J.F. Wu and E.A. Boyle, Geochim.Cosmochim. Acta 61, 3279 (1997).

10. S. Hill, Inductively Coupled PlasmaSpectrometry and its Application,Blackwell Publishing, Oxford, UK,98 (2007).

11. W. Xu, S.J. Ni, Y. Gao, and Z.M. Shi,Spectrosc. Lett. 48, 542 (2015).

12. J.M. Lee, E.A. Boyle, Y. Echegoyen-Sanz, J.N. Fitzsimmons, R.F. Zhang,and R.A. Kayser, Anal. Chim. Acta686, 93 (2011).

13. D.V. Biller and K.W. Bruland, Mar.Chem. 130, 12 (2012).

14. C.K. Su, T.W. Lee, and Y.C. Sun, J.Anal. At. Spectrom. 27, 1585(2012).

15. J.E. O'Sullivan, R.J. Watson, andE.C.V. Butler, Talanta 115, 999(2013).

16. N. Garcia-Otero, C. Teijeiro-Valino,J. Otero-Romani, E. Pena-Vazquez,A. Moreda-Pineiro, and P. Bermejo-Barrera, Anal. Bioanal. Chem. 395,1107 (2009).

17. T. Minami, W. Konagaya, L.J.Zheng, S. Takano, M. Sasaki, R.Murata, Y. Nakaguchi, and Y.Sohrin, Anal. Chim. Acta 854, 183(2015).

18. D. Weiss, E.A. Boyle, V. Chavagnac,M. Herwegh, and J.F. Wu, Spec-trochim. Acta Pt. B, At. Spectrosc.55, 363 (2000).

19. Y. Gao, R. Liu, and L. Yang, Chin.Sci. Bull. 58, 1980 (2013).

20. K. Huang, H. Xia, M. Li, Y. Gao,C. Zheng, and X. Hou, AnalyticalMethods 4, 4058 (2012).

TABLE IIIDetermination of Pb in Seawater

Sample and Method Determined Spiked Value Spike Recoveriesng g−1 ng g−1 (%)

(SD, n = 3) (SD, n = 3)

Sanya (Std. Add.) 0.26 ± 0.04 0.30 108.9 ± 8.0

Haikou (Std. Add.) 0.14 ± 0.01 0.30 106.2 ± 7.7

Shanghai (Std. Add.) 1.29 ± 0.02 1.00 95.2 ± 1.6

TABLE IVPVC-ICP-MS Determination of Pb in a Certified Seawater Sample

Sample Determined Certifiedng g−1 (SD, n = 3) ng g−1 (k=2)

QC3163 Seawater 691 ± 17.8 693 ± 6.4

43

Vol. 38(3),May/June 2017

21. V. Yilmaz, Z. Arslan, and L. Rose,Anal. Chim. Acta 761, 18 (2013).

22. X. Guo, R.E. Sturgeon, Z. Mester,and G.J. Gardner, Anal. Chem. 76,2401 (2004).

23. Y. Yin, J. Liu, and G. Jiang, Trac-Trends in Anal. Chem. 30, 1672(2011).

24. C. Zheng, Q. Ma, L. Wu, X. Hou,and R.E. Sturgeon, Microchem.Journal 95, 32 (2010).

25. Y. Gao, M. Xu, R.E. Sturgeon, Z.Mester, Z. Shi, R. Galea, P. Saull,and L. Yang, Anal. Chem. 87, 4495(2015).

26. H.L. Duan, N.N. Zhang, Z.B. Gong,W.F. Li, and W. Hang,Spectrochim. Acta Part B, At. Spec-trosc.120, 63 (2016).

27. X.M. Guo, R.E. Sturgeon, Z. Mester,and G.J. Gardner, Anal. Chem. 75,2092 (2003).

28. Y. Gao, R.E. Sturgeon, Z. Mester, X.Hon, C. Zheng, and L. Yang, Anal.Chem. 87, 7996 (2015).

29. Y. Gao, R.E. Sturgeon, Z. Mester, X.Hou, and L. Yang, Anal. Chim. Acta901, 34 (2015).

30. Y. Zeng, C. Zheng, X. Hou, and S.Wang, Microchem. Journal 117, 83(2014).

31. H.L. Duan, Z.B. Gong, and S.F.Yang, Journal of Anal. At.Spectrom. 30, 410 (2015).

32. R.E. Sturgeon, Anal. Chem. 87,3072 (2015).

33. Y.L. Wang, L.L. Lin, J.X. Liu, X.F.Mao, J.H. Wang, and D.Y. Qin,Analyst 141, 1530 (2016).

34. P.K. Petrov, G. Wibetoe, and D.L.Tsalev, Spectrochim. Acta Part B,At. Spectrosc. 61, 50 (2006).

44

Detection of Shigella flexneri DNA by ICP-MSBased on Oligonucleotide Hybridizationand Labeling of Gold Nanoparticles

Xiuji Wanga,b, Lanlan Jina, Wei Guoa, Liuqin Huanga, and Shenghong Hua,*a State Key Laboratory of Biogeology and Environmental Geology, School of Earth Sciences,

China University of Geosciences, Wuhan, 430074, P.R. Chinab Center of Analysis, Guangdong Medical University, Dongguan, 523808, P.R. China

*Corresponding author.E-mail: [email protected]

INTRODUCTION

The sensitive detection of DNAis very important in the fields ofdisease diagnosis, food safety, andenvironmental monitoring becausethe genus of bacteria, viruses, andother microorganisms contribute tohuman diseases. Various molecularmethods based on nucleic acidtechnologies (1, 2) have beendeveloped in which sequence-spe-cific DNA oligonucleotides (DNAprobes) are used directly inhybridization assays or as primersin amplification reactions to deter-mine the specific target DNArelated to pathogenic microorgan-isms. Polymerase chain reaction(PCR) based on the design ofprimers, especially real-time PCR,has been developed for DNA quan-tification and has shown high sensi-tivity and specificity because ofamplification (3–5). However, thecomplicated design of primers hasunavoidably limited the applicationof this kind of method (6). Thesequence-selective DNA detectionmethods based on hybridizationassays rely on thermolecular recog-nition abilities of the DNA probesthat hybridize with their comple-mentary sequences. Target nucleicacids are quantified on the basis ofthe chemical nature of tags attachedto DNA probes. Commonly, DNAprobes are labeled frequently withradioactive agents (7, 8), fluores-cent tags (9, 10), and chemilumi-nescence reagents (10, 11).However, some disadvantages,such as safety concerns anddisposal problems of radioactiveagents, sophisticated synthesis and

spectral overlap of fluorescent tags,significantly limit their applicationin DNA quantification.

Recently, inductively coupledplasma mass spectrometry (ICP-MS)based on element tag and hybridiza-tion has been used as a comple-

mentary technique for the targetedanalysis of nucleic acids. In thehybridization assays, the captureprobes and the reporter probeslabeled with element tags arespecifically complementary to thetarget DNA sequence. After thehybridization products areanchored to the surface ofmagnetic beads or microplates bycapture probes, elements from thereporter probes bound to thehybridization products can bequantified by ICP-MS. Two kinds ofelements, lanthanide chelates ormetal nanoparticles (gold or silver),are commonly used as tags to labelbiomolecules and the specific activ-ities of biomolecules can beretained after modification.Lanthanide elements can be taggedto biomolecules using the samestrategy because of their similarproperties, which favors multiplexanalysis (12, 13). Nanoparticles areanother useful elemental tag fortheir easy preparation and theenhanced signal from nanoparticlescontaining numerous metal atoms(14–16). Gold nanoparticle-basedprobes have been used in the deter-mination of DNA. Han et al. (17)presented an effective and ultrasen-sitive method for a one-step homo-geneous DNA assay using single-nanoparticle detection by ICP-MS.Multiplex quantification of DNAtargets associated with clinical dis-eases [human immunodeficiencyvirus (HIV), hepatitis A virus (HAV),and Hepatitis B virus (HBV)] werereported by labeling the probeswith different NPs (Au NPs, AgNPs, and Pt NPs) and detectinghybridizations using single-particle(SP) ICP-MS (18). In conclusion,methods based on ICP-MS for thedetermination of nucleic acids can-not only achieve multiplex analysis

ABSTRACT

A simple and sensitive detec-tion method to target DNA ofShigella flexneri was developedby quantification of goldnanoparticles (AuNPs) usinginductively coupled plasma massspectrometry (ICP-MS). A sand-wich hybridization assay wasestablished to identify the targetDNA using biotinylated captureprobes and AuNP-labeledreporter probes in microfugetubes. The hybridized productswere immobilized on micro-plates via biotin-streptavidinaffinity, and subsequently sepa-rated with excess non-hybridized probes by washing.Au, originating from thehybridized compounds, wasdetected by ICP-MS. Good linear-ity was obtained between theconcentration of the target DNAand 197Au signal intensity. Thelimit of detection (LOD, 3σ) forthe target single-stranded DNA(ssDNA) and the simulated targetdouble-stranded DNA (dsDNA)was calculated as 0.39 pM and39.17 pM. The relative standarddeviation (RSD, n = 3) for 10 pMtarget ssDNA was found to be5.8%. The advantages of the pro-posed method include conve-nient separation of thehybridization compounds by thebiotin-streptavidin system, easeof acquiring reporter probeslabeled with AuNPs, and highsensitivity as a DNA assay.


45


easily, but also quantify lower abun-dance DNA targets without PCRamplification because of its lowdetection limit for metal elements.This kind of method provides amore precise and sensitive signal ofthe mass-to-charge ratios from thelabeled element (especiallynanoparticles) of the DNA probeand alleviates the time-consumingand labor-intensive work (6, 19).

Infections caused by Shigellaflexneri are important agents ofdiarrhea and dysentery (20-22). Inthis study, we present a newapproach to quantify a 28-base con-served DNA sequence of Shigellaflexneri using a hybridization assaycombined with ICP-MS as the detec-tion technique. The synthesizedspecific target DNA associatingwith the Shigella flexneri was iden-tified by the reporter probeslabeled with AuNPs and the captureprobes in a sandwich hybridizationassay. The intensities of 197Au wereobtained using ICP-MS after thehybrid complexes were separatedby the biotin-streptavidin affinityreaction and dissolved in 5% HNO3.This is the first report of ICP-MSused to determine the specificgenomic sequence of Shigellaflexneri. Optimization of thehybridization assay, concentrationof streptavidin coated on themicrowell, specificity of hybridiza-tion, and the analytical perform-ance are also described.

EXPERIMENTAL

Instrumentation

A 7700x series ICP-MS (AgilentTechnologies Inc., Tokyo, Japan)was used in detecting the 197Au sig-nal. A solution of 100 ng g-1 103Rhinternal standard was spiked into allsamples. The optimized parametersfor ICP-MS are listed in Table I.A JEM-2010 transmission electronmicroscope (TEM) (Philips CM12TEM/STEM, The Netherlands) wasused to characterize the AuNPs.A UV-1750 spectrophotometer

(Shimadzu, Suzhou, P.R. China) wasused for recording the ultraviolet-visual (UV-Vis) absorption spectrumof the AuNPs and the reporterprobes labeled with the AuNPs ina 1-cm quartz cell.

Reagents and Materials

All oligonucleotides and DNAprobes were synthesized by Shang-hai Sangon Biotech Co., Ltd.(Shanghai, P.R. China). The sequen-ces of the oligonucleotides used inthis study are listed in Table II (seepage 48) where Sequence 3 wastarget ssDNA, Sequence 7 was thecomplementary strand to targetssDNA, and others were mismat-ched oligonucleotides. All of thesewere dissolved in ultrapure H2O.The capture probe (Sequence 1)and the reporter probe, functional-ized with the thiol groups (−SH) atthe 3’-end (Sequence 2), wasdiluted in the TE buffer.

The streptavidin was obtainedfrom Shanghai Sangon Biotech Co.,Ltd. Bovine serum albumin (BSA)was purchased from Wuhan ChuCheng Zheng Mao Science andTechnology Engineering Co. Ltd.(Wuhan, P.R. China). Chloroauricacid (HAuCl4·4H2O) for AuNPspreparation was acquired fromSinopharm Chemical Reagent Co.Ltd. (Shanghai, P.R. China). Tween

20 was purchased from Sigma–Aldrich Chemical Co. (St. Louis,MO, USA).

The buffers and solutions usedwere (a) phosphate buffered saline(PBS, pH 7.4); (b) coating buffer:(Na2CO3, NaHCO3, pH 9.6); (c)blocking buffer: 5% BSA in PBS; (d)hybridizing buffer (PBS, pH 7.8);(e) washing solution (PBS, Tween20 0.04%, pH 7.4); and (f) TEbuffer: 10 mM Tris-HCl, 1 mMEDTA, pH 8.0. All buffers were pre-pared using ultra-pure water (18.2MΩ cm−1).

Preparation of Reporter ProbesModified by AuNPs

Gold particles, approximately15 nm,. were synthesized accordingto previously published procedures(23, 24). For subsequent experi-ments, the solution was diluted to50 mL with ultrapure water. Theaverage particle diameter of about15 nm was measured using TEM,which is shown in Figure 1(a). Thefinal concentration of gold particlescould be estimated using Beer’sLaw by using the extinction coeffi-cient of about 108 M-1 cm-1 for15-nm diameter particles at around520 nm (24).

The preparation of reporterprobes modified by AuNPs was per-formed according to the literature(18, 25). The DNA probes, modi-fied with AuNPs, were adjusted toa stock concentration of 0.2 Msodium and 0.1 M phosphate buffer(pH=7.0) for subsequent use. TheUV–Vis spectra of the 15 nm AuNPsand the AuNPs conjugated withreporter probes are shown in Fig-ure 1(b). The results showed a5-nm red-shift. This was caused bythe increase in size of the AuNPswhen conjugated to the ssDNA,thus confirming the attachment ofthe ssDNA to the NPs (26).

Hybridization Assay

Before the hybridization assay,micro-well plates were coated with

TABLE IOperating Parameters for ICP-MS

Parameters Description

ICP RF power 1450 WPlasma gas flow rate 15 L min-1

Carrier gas flow rate 1.02 L min-1

Auxiliary gas flow rate 1.0 L min-1

Sample uptake rate 0.35 mL min-1

Integration time 1.5 sSampling depth 8 mmReplicates 3Isotope used 197Au

Internal standard used 103Rh

46

streptavidin. One hundred micro-liters of streptavidin (diluted 100xwith coating buffer) was added intoeach well and incubated at 4 °Covernight. After washing twicewith PBST, the coated micro-wellplate was blocked using a blockingbuffer at 37 oC for 2 hours. Theschematic diagram of the hybridiza-tion assay is shown in Figure 2. Anartificial 28-base oligonucleotide(Sequence 3) was used in its single-stranded form as the specific targetDNA which was a conservedsequence of Shigella flexneri. Ini-tially, a 10-µL solution of targetssDNA (Sequence 3), 10 µL ofbiotinylated capture DNA(Sequence 1), and 20 µL ofhybridizing buffer were mixed ina microfuge tube and hybridized for20 minutes at room temperature.Then, 5 µL of the reporter probes(Sequence 2) labeled with AuNPs

was added into each tube and incu-bated for another 20 minutes atroom temperature. The hybridizedproducts were transferred into thecoated microplates to be incubatedat 37 oC for 30 minutes beforebeing washed three times by PBSTand twice by ultra-pure water. A 5%(v/v) HNO3 solution was added torelease AuNPs from the hybridizedcomplexes, and the 197Au intensi-ties were determined using ICP-MS.External calibration was used forthe quantitative determination ofthe target ssDNA.

Simulated target dsDNAsequences were obtained by mix-ing the equimolar ratio of Sequence3 and Sequence 7 in water (27).The solution was first boiled at95 oC for 5 minutes to melt theduplex and then cooled slowly toroom temperature. A 10-microlitersolution of capture DNA probes(Sequence 1), 10 µL of reporterprobes functionalized with AuNPs,and 20 µL hybridization buffer wereadded to the dsDNA solution andincubated for 30 minutes. Subse-quent steps followed the same pro-cedure as that for the target ssDNA.All experiments were performed intriplicate.


Optimization of the Hybridiza-tion Assay Conditions

A sandwich hybridization is pre-sented in the schematic diagram(Figure 2). First, the excess captureDNA probes were reacted with thetarget ssDNA in hybridization 1.Then, the hybridized complexeswere reacted with the reporterprobes labeled with AuNPs inhybridization 2. The sensitivity ofthis assay is dependent on the num-ber of hybridized complexes thatare bound to the microplates viabiotin-streptavidin affinity. Moresites would be occupied by excess-free biotin-labeled capture probes,and fewer hybrids complexed withcapture probes would be immobi-lized by the microplate, whichmight result in poor detection sen-sitivity and reproducibility. In addi-tion, an excess of reporter probeslabeled with AuNPs in the reactionsystem is essential for hybridization;however, a high concentration ofreporter probes would lead to highbackground intensity (28). There-fore, the hybridization conditionsshould be optimized to increase thedetection sensitivity.

A series of capture probes(Sequence 1) were investigated toidentify an appropriate concentra-

Fig. 1 (a and b). (a) TEM image of 15-nmgold nanoparticles (AuNPs) and (b)Ultraviolet-visual spectrum of synthesizedAuNPs, and AuNPs conjugated withreporter probes (S2, Sequence 2).

Fig. 2. Schematic diagram of the assay to detect DNA based on oligonucleotideshybridization and the quantification of gold nanoparticles (AuNPs) by ICP-MS.

47


tion when the target ssDNA was10 pM and the dilution ratio of thereporter probes labeled withAuNPs was 1:10. The Au signalintensities were acquired whenthe concentration of Sequence 1changed from 0.01 to 5 nM. Asshown in Figure 3(a), the highestratio of Au signal/background inten-sity appeared when the concentra-tion was 0.5 nM. Therefore, 0.5 nMof Sequence 1 was used for subse-quent hybridizations.

The dilution ratio of reporterprobes labeled with AuNPs wasoptimized from 1:2 to 1:50 witha target ssDNA concentration of

10 pM. Figure 3(b) shows that afavorable 197Au signal/backgroundratio was obtained until the dilutionratio of the probes was reduced1:20. According to the results, ahigher dilution of the Au NPs-modi-fied probes when hybridizing withthe target DNA might result inpoorer detection sensitivity, while alower dilution ratio would probablyresult in a higher background andpoorer detection sensitivity. Conse-quently, a 1:20 dilution of AuNPs-modified probes was selected to beapplied in the hybridization assayconsidering both the effect ofdetection sensitivity and background.

Effect of Streptavidin Concen-tration on Au Signal/Back-ground Ratio

A convenient system for separat-ing hybridization complexes fromother reagents was carried out bymicroplates coated with strepta-vidin. In this system, streptavidinwas used to fix hybridization by thecapture DNA because streptavidinbinding to biotin is specific enoughto ensure that the binding isdirected only to the target of inter-est (29). The concentration ofstreptavidin coated on the micro-plate was a key factor for the Ausignal/background intensity ratio.The dilution ratio of streptavidin(1 mg mL-1) from 1:50 to 1:200 wasinvestigated when the target ssDNAwas in the range from 0.5 to 50 pM.Figure 4 indicates that the sameresults were obtained with a dilu-tion ratio of 1:50 and 1:100. Thelatter was appropriate for furtherstudies because less streptavidinwas consumed.

Fig. 3 (a and b). Effect of (a) concentration of captureDNA probes (S1, Sequence 1) and (b) dilution ratio ofreporter probes (S2, Sequence 2) on the Au signal/back-ground ratio. Target ssDNA was at a concentration of10 pM.

Fig. 4. Effects of concentration of streptavidin on the 197Ausignal/background ratio.

48

Specificity of Hybridization

In the detection of Shigellaflexneri in environmental or clini-cal samples, thousands of other mis-matched and random length DNAfragments are present in the DNAlysate together with the targetDNA. To verify specificity ofhybridization, three mismatchedssDNA sequences were also testedusing the same experimental proce-dure (Figure 5). Sequence 4 had a1-base mismatch oligonucleotide,and Sequence 5 had a 5-base mis-match oligonucleotide, respectively,while Sequence 6 was a randomoligonucleotide sequence (TableII). Compared to other mismatchedand the random DNA sequences,a higher intensity of 197Au from thecompletely matched target DNA(Sequence 3) was observed, whichindicated that the target DNA couldbe distinguished easily in the pres-ence of other irrelevant sequencesusing this method.

Analytical Performance

This method was based on theindirect detection of synthetic tar-get DNA sequences of Shigellaflexneri in a proof-of-principleexperiment. The artificial targetssDNA was studied first. Under theoptimized conditions, a series ofconcentrations of the target ssDNAsequence, from 0.5 to 50 pM, was

TABLE IISequences of Oligonucleotides Used in DNA Hybridization

Name Sequence (5’−3’)

Sequence 1 Biotin-gcagtCCGAAGTTAAGCTAC

Sequence 2 CTACTTCTTTTAC-(CH2)6-SH

Sequence 3 GTAAAAGAAGTAGGTAGCTTAACTTCGG

Sequence 4 GTAAAAGCAGTAGGTAGCTTAACTTCGG

Sequence 5 GTGGGAGAAGCCGGTAGCTTAACTTCGG

Sequence 6 ACATCTGCATTTCCGTTAAAGTCCCGTTCGTAAATGCTGTT-GCGGCTTGCTTTTCCGCG

Sequence 7 CCGAAGTTAAGCTACCTACTTCTTTTAC

Fig. 5. Specificity for the determination oftarget ssDNA (S3, Sequence 3) using theproposed hybridization (S4, Sequence 4;S5, Sequence 5; S6, Sequence 6). Alloligonucleotides were used at a concen-tration of 1 pM.

tested in the quantitative assay. Asshown in Figure 6(a), the signalintensity of 197Au correlated linearlywith the logarithm of the targetDNA concentration (correlationcoefficient of r2=0.9715). The pre-cision for concentration of the tar-get DNA at 10 pM with threereplicate measurements was 5.8%(RSD). The detection limit (LOD,3σ) of the proposed method for the28-base synthesized target ssDNAwas calculated to be 0.39 pM,where the σ is the standard devia-tion of repetitive measurements ofthe assay solution blank.

DNA is not readily available tobind complementary sequences inits native form because it is usuallypresent as a double-strand moleculein organisms. Thus, it is necessaryto convert dsDNA to ssDNA bymelting the duplex before hybridi-zation. In this process, the probesand the degraded complementarysequences were competitively com-bined with the target ssDNA (30).Compared with the spiked targetssDNA, quantifying the dsDNArequired the addition of excessreporter DNA probes (Sequence 2labeled with AuNPs) and captureDNA probes. The excess DNAstrands could be removed by wash-ing, which reduced the backgroundeffectively.

Simulated dsDNA in differentconcentrations was also determinedby the proposed method. The cali-bration curve of Figure 6(b) showsa good correlation between the197Au signal and the logarithm ofthe concentration of the targetdsDNA (correlation coefficient ofr2=0.9659) in the range from 102

pM to 104 pM. The precision forthe concentration of the targetDNA at 103 pM with three replicatemeasurements was 5.5% (RSD). Thedetection limit (LOD, 3σ) wasabout 39.17 pM.

Comparing the LODs of VariousDNA Hybridization Assays

The sensitivity could beenhanced using AuNP tags insteadof metal ions for the nanoparticlescontaining numerous metal atoms(28). Moreover, streptavidin hasfour binding sites for biotin, makingit possible to couple more tags perstreptavidin molecule which canfurther increase the sensitivity ofthe assay. The LOD of thedeveloped method for the synthe-sized 28-base ssDNA target wascompared with some valuesderived from the literature (6, 18,19, 31) for direct DNA hybridiza-tion assays with AuNPs tags. Asshown in Table III, the LODobtained by the present method isbetter than those of others.

49


CONCLUSION

A valid method based on a DNAhybridization assay and AuNPslabeling for the accurate ICP-MSdetermination of specific targetDNA related to Shigella flexneriwas studied. Highly efficienthybridization was verified by thesignal intensity from the hybridizedcomplexes. The convenient separa-tion process for hybridized com-plexes was accomplished by theeffective and low-cost system ofbiotin-streptavidin affinity. Thisstudy indicates that ICP-MS coupledwith a DNA hybridization assay is apromising method for the sensitivedetection of specific DNA ofShigella flexneri and provides newinsights into the quantification oflow abundance nucleic acids inbiological samples without amplifi-cation of PCR. It has great potentialfor the determination of variousbacterial pathogens by ICP-MS-based multiplex assay with multi-plex NP tags using oligonucleotidehybridization strategy in the future.

ACKNOWLEDGMENT

This work was supported by theChina Scholarship Council, theNational Nature Science Foundationof China (No. 41521001 and No.21207120), the Ministry of Scienceand Technology of China (No.2014DFA20720), and the ResearchProgram of the State Key Labora-tory of Biogeology and Environmen-tal Geology of China.

Received March 20, 2017.

REFERENCES

1 J. Shendure, and H.L. Ji, Nat.Biotechnol. 26, 1135 (2008).

2 D.J. Lockhart, H.L. Dong, M.C.Byrne, M.T. Follettie, M.V. Gallo,M.S. Chee, M. Mittmann, C.W.Wang, M. Kobayashi, H. Horton,and E.L. Brown, Nat. Biotechnol.14, 1675 (1996).

Fig. 6 (a and b). Correlation between the 197Au signal inten-sity and the logarithm of (a) target ssDNA (the concentrationvaried from 0.5 to 50 pM) and (b) simulated dsDNA (the con-centration varied from 102 to 104 pM). 100 ng g-1 103Rh wasspiked into the solution as an internal standard.

TABLE IIIComparison of Various DNA Assays Based on the AuNPs Labels

Detection Dynamic Range LOD ReferencesMethods

Colorimetry 0.25–50 nM 50 pM (31)

ICP-MS 1–100 pM 1 pM (18)

ICP-AES 350–3.5×106 pM 350 pM (19)

ETV-AAS 10–200 pM 3 pM (6)

ICP-MS 0.5–50 pM 0.39 pM This work

50

3 M.A. Nadkarni, F.E. Martin, N.A.Jacques, and N. Hunter, Microbiol-ogy-(UK) 148, 257 (2002).

4 M.J. Espy, J.R. Uhl, L.M. Sloan, S.P.Buckwalter, M.F. Jones, E.A. Vet-ter, J.D.C. Yao, N.L. Wengenack,J.E. Rosenblatt, F.R. Cockerill, andT.F. Smith, Clin. Microbiol. Rev.19, 165 (2006).

5 L. Gutierrez, M. Mauriat, J. Pelloux,C. Bellini, and O.V. Wuytswinkel,Plant Cell 20, 1734 (2008).

6 X.M. Xu, Y. Gao, S.X. Zhang, S.Z. Li,T. Bai, Y. Zhang, X.R. Hu, and R.Liu, Microchem J. 126, 302 (2016).

7 L. Chen, Y. Wang, D.F. Cheng, X.R.Liu, S.P. Dou, G.Z. Liu, D.J. Hna-towich, and M. Rusckowski,Bioorg. Med. Chem. 21, 6523(2013).

8 M.K. Dewanjee, A.K. Ghafouripour,R.K. Werner, A.N. Serafini, andG.N. Sfakianakis, Bioconj. Chem. 2,195 (1991).

9 H.A. Bassler, S.J.A. Flood, K.J. Livak,J. Marmaro, R. Knorr, and C.A.Batt, Appl. Environ. Microbiol. 61,3724 (1995).

10 E. Avanissaghajani, K. Jones, D.Chapman, and C. Brunk, Biotech-niques 17, 144 (1994).

11 J. Kwun, S. Yun, L. Park, and J.H.Lee, Talanta 119, 262 (2014).

12 G. Han, S. Zhang, Z. Xing, and X.Zhang, Angew. Chem. Int. Ed. 52,1466 (2013).

13 Y.C. Luo, X.W. Yan, Y.S. Huang,R.B. Wen, Z.X. Li, L.M. Yang, C.J.Yang, and Q.Q. Wang, Anal. Chem.85, 9428 (2013).

14 F. Li, Q. Zhao, C.A. Wang, X.F. Lu,X.F. Li, and X.C. Le, Anal. Chem.82, 3399 (2010).

15 R. Liu, P. Wu, L. Yang, X.D. Hou,and Y. Lv, Mass Spectrom. Rev. 33,373 (2014).

16 S. Hu, R. Liu, S. Zhang, Z. Huang, Z.Xing, and X. Zhang, J. Am. Soc.Mass. Spectrom. 20, 1096 (2009).

17 G. Han, Z. Xing, Y. Dong, S. Zhang,and X. Zhang, Angew. Chem. Int.Ed. 50, 3462 (2011).

18 S.X. Zhang, G.J. Han, Z. Xing, S.C.Zhang, and X.R. Zhang, Anal.Chem. 86, 3541 (2014).

19 L.L. Wu, L.W. Qiu, C.S. Shi, and J.Zhu, Biomacromolecules 8, 2795(2007).

20 K.L. Kotloff, J.P. Winickoff, B.Ivanoff, J.D. Clemens, D.L. Swerd-low, P.J. Sansonetti, G. Adak, andM. Levine, Bull. W.H.O. 77, 651(1999).

21 D.M. Deer, and K.A. Lampel, J.Food Prot. 73, 1618 (2010).

22 M.J. Hosseini, and A.R. Kaffashian,Arch. Iran. Med. 13, 413 (2010).

23 G. Frens, Nature Phys. Sc.i 241, 20(1973).

24 C.C. Huang, Y.F. Huang, Z.H. Cao,W.H. Tan, and H.T. Chang, Anal.Chem. 77, 5735 (2005).

25 R. Elghanian, J.J. Storhoff, R.C.Mucic, R.L. Letsinger, and C.A.Mirkin, Science 277, 1078 (1997).

26 J. Thavanathan, N.M. Huang, andK.L. Thong, Biosensors & Bioelec-tronics 55, 91 (2014).

27 K. Brueckner, K. Schwarz, S. Beck,and M.W. Linscheid, Anal. Chem.86, 585 (2014).

28 Q. He, Z.L. Zhu, L.L. Jin, L. Peng,W. Guo, and S.H. Hu, J. Anal. At.Spectrom. 29, 1477 (2014).

29 P.J. Ginel, J.M. Margarito, J.M.Molleda, R. Lopez, and M. Novales,Res. Vet. Sci. 60, 107 (1996).

30 B.S. Li, L.F. Zhao, C. Zhang, X.H.Hei, F. Li, X.B. Li, J. Shen, Y.Y. Li,Q. Huang, and S.Q. Xu, Anal. Sci.22, 1367 (2006).

31 X. Mao, H. Xu, Q.X. Zeng, L.W.Zeng, and G.D. Liu, Chem. Com-mun. 21, 3065 (2009).

51

*Corresponding author.:E-mai: [email protected]: +903524374933

Application of Supramolecular Microextraction andFlame Atomic Absorption Spectrometry for Ultra-traceDetermination of Cadmium in Food and Water Samples

Z.A. ALOthmana, M.A. Habilaa, Erkan Yilmazb, M. Soylakb, *, and A.A. Alwarthana

a Chemistry Department, College of Science, King Saud University, Riyadh 11451, Kingdom of Saudi Arabiab Erciyes University, Faculty of Sciences, Department of Chemistry, 38039 Kayseri, Turkey

ABSTRACT

In this work, a supramolecularmicroextraction procedure fortrace cadmium as 1,2,4 thiadia-zole-2,5 dithiol chelates has beenestablished. Flame atomic absorp-tion spectrometry was used forthe determination of cadmium.A rapid injection of 600 µLtetrahydrofuran and 200 µL of1-decanol in a sample solutioncontaining cadmium at pH 8leads to formation of a cloudysolution by ultrasonic waves.Then the supramolecular organicsolvent mixture, including thechelated cadmium, was separatedby centrifugation. Certified refer-ence materials (TMDA 64.2 andTMDA 53.3) were tested and thedetermined concentrations werein agreement with the certifiedvalues. Furthermore, the methodresulted in an LOD of 0.46 µg L-1,LOQ of 1.37 µg L-1, and RSD of5.1%. The supramolecularmicroextraction procedure wassuccessfully applied for the analy-sis of environmental samplesincluding wastewater, seawater,dam water, valley water, andblack pepper.

to real samples, etc. (30-33). Theuse of suitable new solvents inmicroextraction studies which areenvironmentally friendly and greenis a challenge that has beendiscussed by analytical chemists(34-36). Supramolecular solvents(SUPRAS) for microextraction area good choice (36-38). SUPRAS hasbeen introduced as an extractionmethod with many advantages suchas low use of organic solvents, highpre-concentration factors, and fastand simple procedure based on aself-assembly process of the supra-

molecular solvent. This procedureallows the different polarity sites inthese assembly constituents andprovides a suitable environ-ment toseparate and preconcentrate theanalytes (36-38).

A novel and simple supramole-cular microextraction procedurebased on the extraction of Cd(II) as1,2,4-thiadiazole-2,5-dithiol chelatesis presented in this work. The ana-lytical conditions for quantitativeextraction of Cd(II) were optimized.

EXPERIMENTAL

Instrumentation

A PerkinElmer® Model 3110flame atomic absorption spectrome-ter (PerkinElmer, Inc., Shelton, CT,USA) with an air–acetylene flameand hollow cathode lamp was usedto measure cadmium concentrations.The instrumental operating parame-ters and linear range for cadmiumare listed in Table I. The instrumen-tal parameters were adjusted as rec-ommended by the manufacturer.The samples were injected into theAAS with a Teflon® funnel using ahomemade micro-sample introduc-tion system. The absorbance signalwas measured according to peakheight in the continuous aspirationmode (39).

A Nel pH-900 (Ankara, Turkey)and Metrohm pH meter (Model691, Switzerland) with a combinedglass electrode were used for thepH measurements. An ALC PK 120Model centrifuge (Buckingham-shire, England) was used for cen-trifugation. A vortex mixer (VWRInternational LLC, USA) and anultrasonic bath (Sonorex, ModelNo. DT-255, Bandelin Co., Germany)were used.

INTRODUCTION

Metal ions at trace or ultratracelevels are generally problematic forhumans, plants, and animals (1-3).Lead (Pb) and cadmium (Cd) arevery toxict metals even at trace lev-els, whereas zinc (Zn) and copper(Cu) are toxic only when they arepresent at higher concentrations(4-7). Excessive use of fertilizerscan accumulate trace elements inthe soil, causing high toxicity tothe environment and plants. Cad-mium contamination in the foodchain can be harmful and causechronic health problems (8-13).

The determination of cadmiumand other metals at trace or ultra-trace levels by spectroscopic tech-niques are an important area ofchemistry, agriculture, medicine,biology, etc. (14-20). Due to thelow detection capabilities of thesetechniques, the matrix effects andlow levels of the analyte elementsin some samples are two importantproblems (21-25). Preconcentra-tion-separation systems, includingsolid phase extraction, cloud pointextraction, coprecipitation, mem-brane filtration, electrodepositionand flotation, are an important keyto solve these problems (26-29).

In the last decade, microextrac-tion techniques have become analternative for the preconcentra-tion-separation techniques due toadvantages such as limitedconsumption of organic solvents,the possibility of high preconcen-tration factors, and easy application


52

Reagents and StandardSolutions

All chemicals were of analyticalgrade. Distilled and deionized waterwith 18 MΩ resistivity were pre-pared using a Milli-Q® system (Milli-pore Corporation, USA). StockCd(II) solutions (1000 mg L−1) wereprepared by dissolving the nitratesalt in water (E. Merck. Darmstadt,Germany). Working standard solu-tions were obtained via serial dilu-tion of a stock standard solution.The 0.1% (w/v) 1,2,4-thiadiazole-2,5-dithiol (Sigma-Aldrich, St. Louis,MO, USA) solution was preparedusing deionized water.

Procedure

A 20-mL sample solution contain-ing Cd(II) was placed into a 50-mLpolypropelene centrifuge tube andthe pH was adjusted to 8 using anammonia buffer solution. Then200 µL of 0.1% 1,2,4-thiadiazole-2,5-dithiol ligand was introduced.In this step, the chelation of Cd(II)with 1,2,4-thiadiazole-2,5-dithioloccurred. Then, a solution oftetrahydrofuran and 1-decanol wasrapidly injected into the mixturewhich leads to the formation ofsupramolecular solvent inside thetube. The mixture was exposed toultrasonic waves for 4 to 5 minutesand then shaken on the vortex forone minute. Finally, the phase wasseparated by using the centrifugeat 4000 rpm for 10 minutes. Afterremoving the bottom aqueousphase, the remaining supramolecu-lar solvents containing cadmium as1,2,4-thiadiazole-2,5-dithiol chelateswere dissolved in ethanol to com-plete the final volume to 500 µL.The Cd(II) concentration was deter-mined by injection of 50 µL intothe AAS with a Teflon® funnel

using a homemade micro-sampleintroduction system.

Application of Method to RealSamples

Water samples, including waste-water, seawater, dam water, andvalley water were obtained inTurkey. The samples were filteredwith a Millipore® cellulose mem-brane (0.45 micrometer), then theoptimized supramolecular microex-traction procedure was applied forCd(II) determination. Certified ref-erence materials for water such asTMDA 64.2 and TMDA 53.3(National Water Research Institute,Environment Canada, Burlington,Canada) were applied to evaluatethe process. In addition, black pep-per as a food sample from KayseriCity, Turkey, was collected,washed, dried, and digested asdescribed in the literature (31, 39).The proposed supramolecularmicroextraction steps were appliedto the food sample extract forCd(II) determination.


Optimization of AnalyticalParameters

The significant parameters con-trolling the supramolecular extrac-tion steps including pH, amount of1,2,4-thiadiazole-2,5-dithiol as lig-and, composition of supramolecu-lar mixture, and sample volume,were optimized.

The pH of the sample solutioncontaining Cd(II) was tested inacidic, neutral and basic mediumup to pH 9. Figure 1 shows theeffect of pH on the recovery ofCd(II). The quantitative recoverywas achieved at pH 8 which was

selected for further investigations.It was reported earlier that the pHof a sample solution significantlyaffects the recovery of heavy metals(14, 16, 18, 21).

The volume of 1,2,4-thiadiazole-2,5-dithiol as ligand was tested inthe range of 0–300 µL and the (%)recovery was calculated for eachcase. Figure 2 demonstrates thatthe (%) recovery values were verylow at 0, 25, and 50 µL, thenincreased at 100 and 150 µL. Quan-titative recoveries were obtainedbetween 200 to 300 µL. Theincrease in recoveries betweenincreasing volume of ligand solu-tion can be attributed to the factthat the microextraction proceduredepends on the formation of a com-plex between Cd(II) and 1,2,4-thia-diazole-2,5-dithiol as the ligand.This complex will leave the aque-ous phase and move to the organicsupramolecular solvent phasewhich separates later and is mea-sured. When using a limitedamount of 1,2,4-thiadiazole-2,5-dithiol ligand, it will not be enoughto chelate all Cd(II). Therefore,200 µL was chosen for furtherexperiments.

The component of the supramol-ecular solvent is reported to havesignificant effects on the recoveryof metals (39). Therefore, tetrahy-drofuran with 1-decanol, tetrahy-drofuran with undecanol, andtetrahydrofuran with decanoic acidwere tested as different supramole-cular solvents and the recovery was99%, 18%, 7%, respectively.Tetrahydrofuran was selected forfurther work due to quantitativerecoveries.

The volume of tetrahydrofuranand the amount of 1-decanol wereinvestigated in the range of 0–800µL for tetrahydrofuran and in therange of 0–250 µL for 1-decanol.The quantitative recoveries (>95%)were obtained for 600–800 µL oftetrahydrofuran and for 200–250 µLof 1-decanol. For further work,

TABLE IInstrumental Operating Conditions and Linear Range for Cadmium

Element Wavelength Slit Width Lamp Current Linear Range(nm) (nm) (mA) (µg mL-1)

Cd 228.8 0.7 4 0.1 – 2.0

53

Vol. 38(3),May/June 2017

600 µL of tetrahydrofuran and 200µL of 1-decanol were selected.

The effect of sample volume isa controlling parameter whichdirectly affects the preconcentra-tion factor (40–46) and was studiedin the range from 10 mL to 45 mL(Figure 3). The results show thatthe recoveries for Cd(II) werequantitative up to 20 mL. The pre-concentration factor calculated asthe ratio between the initial volumeof the Cd(II) sample and the lastvolume after supramolecularmicroextraction was 40, consider-ing that the last volume is 0.5 mL.

The optimum conditions for thequantitative recoveries of Cd(II)

as 1,2,4-thiadiazole-2,5-dithiolchelates for the presented supra-molecular extraction method aresummarized in Table II.

Matrix Effects

The studies on the effect of com-mon interfering ions on the recov-ery of analyte elements areimportant for the optimization ofseparation-preconcentration meth-ods (47-54). This was tested byapplying the supramolecular sepa-ration- preconcentration steps inthe presence of K+, Cl–, Mg2+, Ca2+,SO4

2–, F–, Ni2+, Cu2+, Fe3+, Zn2+,CO3

2–, NO3–, and Na+. Table III pre-

sents the % recovery of cadmiumfor each ion which was not less

than 96%, indicating that thesesupramolecular extraction stepscan be applied for different sampleswith various matrices.

Analytical Figures of Merit

The LOD for this process was0.46 µg L-1, the LOQwas 1.37 µg L-1,and the relative standard deviation(RSD) was 5.1%. The supramolecu-lar microextraction proceduredescribed in this work was com-pared with others from the litera-ture (55–61) and showedcomparable results (Table IV).

Application

Validation of the presentedsupramolecular microextraction

Fig 1. Evaluation of the effect of pH value on the recovery ofCd(II) (N=3).

Fig. 2. Evaluation of the effect of the quantity of ligand on the(%) recovery of Cd(II) (N=3).

Fig. 3. Effect of sample volume on the (%) recovery of Cd(II) (N=3)

TABLE IIOptimum Condition for Cd(II)

Supramolecular Microextraction

Parameters OptimumValues

pH value of sample solution 8

Amount of 0.1% 1, 2, 4thiadiazole-2, 5 dithiol 200 µL

Amount of tetrahydrofuran 600 µL

Amount of 1-decanol 200 µL

Sample volume 20 mL

54

procedure for Cd(II) was investi-gated by addition/recovery tests ofa tap water sample from ErciyesUniversity, Kayseri, Turkey (TableV). The procedure is applicable inthe presence of different concentra-tions of Cd(II) where the recoverieswere not less than 99%. For furtheroptimization, certified referencematerials TMDA 64.2 and TMDA53.3 Water were applied (TableVI). The concentrations found byusing the presented procedurewere in agreement with the certi-fied values.

Different samples such as waste-water, seawater, dam water, valley

water, and black pepper were usedfor the determination of Cd(II) con-tent. Table VII lists the obtainedresults which reveal that the devel-oped procedure is applicable andindependent of type of matrix.

CONCLUSION

A microextraction procedure,based on the application of tetrahy-drofuran with 1-decanol to form asupramolecular solvent to separateCd(II), was optimized. The 1,2,4-thiadiazole-2,5-dithiol as ligandplays an important role for Cd(II)separation and recovery. The maxi-mum operating conditions were a

pH of 8 and using 200 µL of 0.1%1,2,4-thiadiazole-2,5-dithiol, 600 µLof tetrahydrofuran, and 200 µL of1-decanol. The initial sample vol-umes can be increased to 20 mL toobtain quantitative recoveries. Theprocedure showed high efficiencycompared to other methods pro-vided in the literature.

TABLE IIIEffect of Some Common Interfering Ions

on the Recovery (%) of Cd(II) (N=3)

Ions Concentration Added as: Recovery(µg mL-1) (%)

K+, Cl– 2000 KCl 99.5±0.5Mg2+ 800 Mg(NO3)2.6H2O 98.5±0.3Ca2+ 800 CaCl2 99.5±0.4SO4

2– 500 Na2SO4 98.0±0.6F– 500 NaF 96.5±0.4Ni2+ 5 Ni (NO3)2.6H2O 98.0±0.6Cu2+ 10 Cu(NO3)2.3H2O 100.0±0.4Fe3+ 5 Fe(NO3)3.9H2O 96.0±0.4Zn2+ 5 Zn(NO3)2 97.0±0.8CO3

2– 2000 Na2CO3 97.0±0.4NO3

– 2000 KNO3 99.0±0.4

Na+ 8000 NaCl 96.0±0.3

TABLE IVComparison Between Proposed Procedure and Other Methods

Reported in the Literature for Cu Determination

Preconcentration Method LOD (µg L-1) Ref.

Dispersive liquid–liquid microextraction (DLLME) 1.0 (55)Dispersive liquid–liquid microextraction (DLLME) 0.4 (56)On-line solvent extraction 0.003 (57)Cloud Point Extraction (CPE) 0.006 (58)Cloud Point Extraction (CPE) 0.31 (59)Microprecipitation 0.25 (60)

Supramolecular microextration (SME) 0.46 Thisstudy

TABLE VAddition/Recovery Evaluationof Developed Procedures for

Cd(II) from a Tap Water Samplefrom Erciyes University (N=3)

Added Found Recovery(µg) (µg) (%)

0 0 -0.50 0.49±0.11 99

1.0 1.0±0.13 100

TABLE VIValidation of Developed

Microextraction ProcedureUsing TMDA 64.2 and

TMDA 53.3 Water CRMs (N=3)

CRM Certified Found RecoveryValue Value(µg L-1) (µg L-1) (%)

TMDA64.2 288 282±4 98

TMDA53.3 118 117±5 99

TABLE VIIApplication of Developed

SupramolecularMicroextraction Method for

Cd Determination inWater and Food Samples (N=3)

Sample Concentration

Wastewater 10.2±0. 5 µg L-1

Seawater 26.5±0.4 µg L-1

Dam water 3.9±0.2 µg L-1

Valley water 47.9±0.4 µg L-1

Black pepper 0.8±0.2 µg g-1

55


Haji Shabani, Quim. Nova 36, 63(2013).

33. R. Yousefi and F. Shemirani, Anal.Chim. Acta 669, 25 (2010).

34. C. Caballo, M.D. Sicilia, and S.Rubio, Talanta 119, 46 (2014).

35. C. Caballo, M. D. Sicilia, and S.Rubio, Anal Bioanal. Chem. 407,4721 (2015).

36. Q. Yang, W. Su, X. Jiang, and X.Chen, Int. J. Environ. Anal. Chem.94, 812 (2014).

37. M. Peyrovi and M. Hadjmohammadi,J. Chromatogr. 980B, 41 (2015).

38. F. Aydin, E. Yilmaz, and M. Soylak,RSC Adv. 5, 40422 (2015).

39. E. Yilmaz and M. Soylak, Talanta126, 191 (2014).

40. S. Saracoglu, M. Soylak, D.S.KPeker, L. Elci, W.N.L. dos Santos,V.A. Lemos, and S.L.C. Ferreira,Anal. Chim. Acta 575, 133 (2006).

41. M. Soylak, Y.E. Unsal, N. Kizil, andA. Aydin, Food Chem. Toxicol., 48,517 (2010).

42. M. Ghaedi, M. Montazerozohori,F. Marahel, M.N. Biyareh, and M.Soylak, Chin. J. Chem. 29, 2141(2011).

43. I.M.M. Kenawy, W.I. Mortada, Y.G.Abou El-Reash, and A.H. Hawwas,Can. J. Chem. 94, 221 (2016).

44. M. Soylak and E. Yilmaz, Desalina-tion 275, 297 (2011).

45. L. Hu, Y. Cai, and G. Jiang, Chemos-phere 156, 14 (2016).

46. M. Sheibani, F. Marahel, M. Ghaedi,M. Montazerozohori, and M. Soy-lak, Toxicol. Environ. Chem. 93,860 (2011).

47. M. Soleimani and Z.H. Siahpoosh, J.Applied Chem. Research 9, 7(2015).

48. M. Ghaedi, K. Niknam, E. Niknam,K. Mortazavi, K. Taheri, and M.Soylak, J. Chin. Chem. Soc. 57, 275(2010).

49. M. Firlak, S. Cubuk, E.K. Yetimoǧlu,and M.V. Kahraman, Chem. Pap.70, 757 (2016).

50. H. Erdogan, O. Yalcinkaya, and A.R.Turker, Turk. J. Chem., 40, 772(2016). 51. N. Khan, T.G. Kazi, M.Tuzen, and M. Soylak, Desalin.

ACKNOWLEDGMENTS

The authors extend their appre-ciation to the International Scien-tific Partnership Program ISPP atKing Saud University for fundingthis research work through ISPP#0029.

Received July 6, 2016.

REFERENCES

1. O. Turkoglu and, M. Soylak, and I.Belenli, Collect. Czech. Chem. C.68, 1233 (2003).

2. M. Soylak, and O. Turkoglu, J. TraceMicroprobe T., 17, 209 (1999).

3. T.Y. Gorgulu, O.D. Ozdemir, A.S.Kipcak, M.B. Piskin, and E.M.Deru, Appl. Biol. Chem. 59, 425(2016).

4. U.A. Barbosa, I.F. dos Santos, A.M.P.dos Santos, and S.L.C. Ferreira,Anal. Lett. 49, 799 (2016)

5. M. Soylak and L. Elci, J. Trace Micro-probe T., 18, 397 (2000).

6. M. Gebrelibanos, N. Megersa, andA.M. Taddesse, Int. J. Food. Con-tam. 3, 2 (2016). DOI10.1186/s40550-016-0025-7

7. I. Narin, M. Soylak, and L. Elci, Anal.Lett. 34, 1935 (2001).

8. V.G. Bizarro, E.J. Meurer, and F.R.P.Tatsch, Ciencia Rural 38, 247(2008).

9. A. Tumuklu, M. Çiflikli, and F.Z.Özgür, Asian J. Chem., 20, 6376(2008).

10. M. Kramarova, P. Massanyi, A.Jancova, R. Toman, J. Slamecka,F. Tataruch, J. Kovacik, J. Gasparik,P. Nad, M. Skalicka, B. Korenekova,R. Jurcik, J. Cubon, and P. Hascik,Bull. Vet. Inst. Pulawy 49, 465(2005).

11. U. Divrikli, M. Soylak, L. Elci, and M.Dogan, J. Trace Microprobe T. 21,713 (2003).

12. M.N. AlKathiri and A.F. AlAttar,Bull. Environ. Contam. Toxicol. 58,726 (1997).

13. A.A. AlWarthan and H.M.AlSwaidan, Arab. Gulf. J. Sci. Res.

J., 13, 453 (1995).

14. M. Tuzen and M. Soylak, J. Hazard.Mater. 164, 1428 (2009).

15. Y. Yang, F.S. Zhang, H.F. Li, andR.F. Jiang, J. Environ. Manage. 90,1117 (2009).

16. D. Mendil, O.F. Unal, M. Tuzen, andM. Soylak, Food Chem. Toxicol.48, 1383 (2010).

17. J.S. Barin, F.R. Bartz, V.L. Dressler,J.N.G. Paniz, and E.M.M. Flores,Anal. Chem. 80, 9369 (2008).

18. A. Shokrollahi, M. Ghaedi, O. Hos-saini, N. Khanjari, M. Soylak, J.Hazard. Mater. 160, 435 (2008).

19. M.A. Zazouli, M. Shorkzadeh, S.Fathi, and H. Izanlo, Afr. J. Biotech-nol. 7, 3689 (2008).

20. P. Nad, P. Massanyi, M. Skalicka, B.Korenekova, V. Cigankova, andV.Almasiova, J Environ. Sci. Heal. A42, 2017 (2007).

21. M. Soylak, L. Elci, and M. Dogan,Anal. Lett. 26, 1997 (1993).

22. M. Karimi, A.M.H. Shabani and S.Dadfarnia, J. Braz. Chem. Soc. 27,144 (2016).

23. M. Tuzen, M. Soylak, D. Citak, H.S.Ferreira, M.G.A. Korn, and M.A.Bezerra, J. Hazard. Mater. 162,1041 (2009).

24. S. Kaur, T.P.S. Walia, and R.K.Mahajan, J. Environ. Eng. Sci. 7, 83(2008).

25. A. Uzun, M. Soylak, L. Elci, and M.Dogan, Asian J. Chem., 14, 1277(2002).

26. S. Dadfarnia, A.M. Haji Shabani, andM. Amirkavei, Turk. J. Chem., 37,746 (2013).

27. U. Divrikli, A.A. Kartal, M. Soylak,and L. Elci, J. Hazard. Mater., 145,459 (2007).

28. A. Denizli, E. Piskin, and B. Salih,Turk. J. Chem. 19, 296 (1995).

29. H. Cesur, Turk. J. Chem., 27, 307(2003).

30. Y.E. Unsal, M. Tuzen, and M. Soy-lak, Turk. J. Chem., 38, 173 (2014).

31. Z.A. ALOthman, E. Yilmaz, M.Habila, and M. Soylak, Desalin.Water Treatmt. 51, 6770 (2013).

32. M. Amirkavei, S. Dadfarnia and A.M.

56

Water Treatmt. 55, 1088 (2015).

52. C. Polat, V. Eyupoglu, and O.N.Sara, AIP Conf. Proc. 1726, 020110(2016); DOIi: 10.1063/1.4945936

53. R. Gurkan and M. Eser, J. Iran.Chem. Soc. 13, 1579 (2016).

54. X. Xu, M. Zhang, L. Wang, S. Zhang,M. Liu, N. Long, X. Qi, Z. Cui, L.Zhang, Food Anal. Methods 9,1696 (2016).

55. A. Afkhami, T. Madrakian, H.Siampour, J. Hazard. Mater. 138,269 (2006).

56. S.Z. Mohammadi, Y.M. Baghelani, F.Mansori, T. Shamspur, D. Afzali.Quim. Nova 35, 198 (2012).

57. A.N. Anthemidis, G.A. Zachariadis,and J.A. Stratis, J. Anal. At. Spec-trom. 18, 1400 (2003).

58. X. Zhu, X. Zhu, and B. Wang,Microchim. Acta 154, 95 (2006).

59. J.L. Manzoori and G. K. Nezhad,Anal. Chim. Acta 521, 173 (2004).

60. Z.A. ALOthman, M.A. Habila, S.M.Alfadul, E. Yilmaz and M. Soylak.Anal. Methods 8, 3545 (2016).

61. M. Soylak, S. Saracoglu, L. Elci, andM. Dogan, Int. J. Environ. Anal.Chem. 82, 225 (2002).

57

INTRODUCTION

The levels of trace elements arean important component of safetyand quality of the environment.Trace elements such as lead (Pb),arsenic (As), and cadmium (Cd) aregenerally recognized as pollutantsand are harmful to the health ofhumans, plants, and animals (1-4).They are toxic when present inexcessive amounts (5-8). Metalsoccur naturally in soil, plants, andair. However, their concentrationsare usually increased due to anthro-pogenic activities such as fromindustry and traffic (9-12).

Lead is one of the toxic elementseven at ultra-trace levels (13, 14)and is absorbed in humans throughinhalation of air, dermal exposure,and ingestion of dust and soil.Other sources of lead are pollutedwater and lead-adulterated foods(15-18). Copper (Cu) is another crit-ical, important and necessary traceelement for humans, plants, andanimals. It is an essential part of sev-eral enzymes, essential for humannutrition (19-20), and necessary forthe synthesis of hemoglobin (21-23). Consequently, an analyticalmethod to preconcentrate, sepa-rate, and determine trace elementsin environmental samples includingnatural waters is very important.

Classical solvent extraction,micellar extraction, flotation, copre-cipitation and membrane filtrationare popular separation-preconcen-tration methods for trace elementsfrom environmental and food sam-

ples (24-28). Solid phase extractionis a popular separation-enrichmenttechnique using novel nanomateri-als with a high adsorption capacityand high surface area with resistiv-ity to concentrated acids and bases(29-31).

In the present work, Chromo-trope FB impregnated magneticmultiwalled carbon nanotubes havebeen used for the preconcentration-separation of Pb(II) and Cu(II) inhair and cigarette samples.

EXPERIMENTAL

Instrumentation

For this study, a Model 3110flame atomic absorption spectrome-

ter (PerkinElmer, Inc., Shelton, CT,USA) was used, equipped withair/acetylene flame with a 10 cmlong slot burner head and hollowcathode lamps of Cu and Pb. Thethe conditions for flame atomicabsorption spectrometty (FAAS)are given in Table I. The calibrationcurve equations for Cu and Pb forFAAS determinations, respectively,were (A = 0.0394 C (Cu) + 0.0056)and (A = 0.0133 C (Pb) + 0.0011),where A is the absorbance and Cis the concentration of the analyte.The correlation coefficients of theequations were 0.9992 and 0.99945for Cu and Pb, respectively.

The pH values were measuredusing a Sartorius PT–10 pH meter(Sartorius, Germany) with a glasselectrode. A neodymium magnetwas used for the phase separations.A vortex mixer (Wiggen Hauser,Malaysia) was used for thoroughlyvortexing and mixing of the solu-tions.

Solutions and Reagents

Solutions were prepared withreverse osmosis purified waterusing a Milli-Q® system 18 MΩ−cmresistivity (Millipore Corporation,USA). All chemicals were of reagentgrade. Stock solutions of the analyteelements were prepared from theirnitrates as 1000 µg L-1, solutions in0.02 M HNO3. Multi-walled carbonnanotubes (MWNT) (Sigma no:636614-2G) were purchased fromAldrich, Milwaukee, WI, USA. Asolution of Chromotrope FB (% 0.1m/v) was prepared by dissolving ofChromo-trope FB No. B22328 (AlfaAesar, Germany) in small amountsof ethanol and diluting to 50 mLwith water. Certified referencematerial (CRM) NCS-DC73349 Bush

*Corresponding author.e-mail: [email protected] [email protected]. & Fax: +90 352 4374933

ABSTRACT

A novel and simple magnetic-solid phase extraction procedurefor the separation-preconcentra-tion of lead(II) and copper(II) attrace levels has been establishedby using Chromotrope FB impreg-nated magnetic multiwalled car-bon nanotubes as a newadsorbent. The influences of criti-cal analytical parameters like pH,flow rates, eluent, and samplevolume were optimized. Matrixeffects were also examined. Ana-lyte elements were quantitativelyrecovered at pH 6.0. The limit ofdetection values were 11.7 µg L-1

for lead and 3.7 µg L-1 for copper.The (%) RSD values were gener-ally found at < 7%. The validationwas performed by the analysis ofNCS-DC73349 Bush Branches andLeaves certified reference mater-ial. The method was applied tothe FAAS determination of leadand copper concentrations incigarette and hair samples.


Magnetic Solid Phase Extraction of Trace Lead andCopper on Chromotrope FB Impregnated Magnetic

Multiwalled Carbon Nanotubes From Cigarette and HairSamples for Measurement by Flame AAS

Mustafa Soylak* and Zeliha ErbasErciyes University, Faculty of Sciences, Department of Chemistry, 38039 Kayseri, Turkey

58

Branches and Leaves (NationalResearch Centre for Certified Refer-ence Materials, Beijing, P.R. China)were used in the experiments forvalidation of the procedure.

Phosphate buffer solution (0.1mol L-1) (pH 2.0–7.0), ammoniabuffer solution (0.1 mol L-1) (pH8.0), and phosphate buffer solution(0.1 mol L-1) (pH 9.0) were used inthe experiments.

Preparation of ChromotropeFB Impregnated Magnetic Multi-walled Carbon Nanotubes

Magnetic multiwalled carbonnanotubes were obtained accordingto the procedure given in the litera-ture (32) by using magnetite. Toimpregnate the Chromotrope FB tothe surface of the magnetic multi-walled carbon nanotubes, 1.0 gmagnetic multiwalled carbon nan-otubes was added to 2 mL of Chro-motrope FB solution and diluted to30 mL with reverse osmosis waterand stirring continuously for 20minutes. Afterwards, the Chromo-trope FB/magnetic multiwalled car-bon nanotubes were filtered off,washed with reverse osmosiswater, and dried overnight at 100oC. The amount of Chromotrope FBdeposited on the magnetic multi-walled carbon nanotubes was esti-mated by UV-Vis spectropho-tometry from the residual amountof a Chromotrope FB in the solu-tion. It was found that 66% of Chro-motrope FB was retained on theadsorbent.

Test Procedure

A 10-mL aqueous sample solu-tion containing Pb(II) and Cu(II)was adjusted to pH 6.0 using 2 mL

phosphate buffer solution. Then,75 mg of Chromotrope FB impreg-nated magnetic multiwalled carbonnanotubes were added to this solu-tion. After 5 minutes, it wasvortexed with a revolution speedof 4000 rpm. After completion ofthe adsorption of the analytes ontothe Chromotrope FB impregnatedmagnetic multiwalled carbon nan-otubes, the adsorbent wasseparated at the bottom of thetube by using a neodymium mag-net. The liquid phase was decanted.Then, 2 mL of 3.0 M HNO3 in 10%acetone was added to the adsor-bent for desorption of the analyteions. Then the Chromotrope FBimpregnated magnetic multiwalledcarbon nanotubes were separatedfrom the sample by using a neo-dymium magnet. An amount of100 µL was aspirated into the nebu-lizer of the FAAS for absorbancemeasurements using the micro-injection system.

Analysis of CRM and RealSamples

The method was also applied toCRM NCS DC73349 Bush Branchesand Leaves, and cigarette and hairsamples. A wet ash procedure was

used for this purpose. One gram ofa sample was transferred into abeaker and digested with 10 mLconcentrated HNO3 for 15 minutesat room temperature, then placedon a hot plate at 95 oC until a dryresidue was obtained. The residueof each beaker was again digestedwith a mixture of HNO3-H2O2 (2:1v/v) until dryness. The residues inthe beakers were then dissolved in15 mL distilled water and filteredwith a Whatman blue band filterpaper. Then the procedure givenin the “Test Procedure” wasapplied to these samples.


Optimization

The pH is an important factoraffecting quantitative adsorption ofthe analytes in the extraction stud-ies (33-38). The effects of the pHon the recoveries of Pb(II) andCu(II) ions on the Chromotrope FBimpregnated magnetic multiwalledcarbon nanotubes were tested atthe pH range of 2.0–9.0. The resultsare depicted in Figure 1. The recov-eries of Pb(II) and Cu(II) werefound to be quantitative in the pHrange of 3.0–8.0 and 5.0–7.0,respectively. For all further work,pH 6.0 was selected as optimal. ThepH adjustments for pH 6.0 weredone by using phosphate buffersolution.

The amounts of Chromotrope FBimpregnated magnetic multiwalledcarbon nanotubes were also tested

TABLE IFAAS Instrumental Parameters

Element Wavelength Slit Width Lamp Current Flame Conditions*(nm) (nm) (mA) (a) (b)

Cu 324.8 0.7 15 9.5 2.3

Pb 283.3 0.7 15 9.5 2.3

* (a) Air (L/min), (b) Acetylene (L/min).

Fig. 1. Relation between pH and (%) recovery for analyte ions.

59


in the range of 25–100 mg. Quanti-tative recoveries for lead and cop-per ions were obtained in the rangeof 50–100 mg of magnetic adsor-bent. An amount of 75 mg Chro-motrope FB impregnated magnetic

multiwalled carbon nanotubes wasused for all other experiments.

The elution of the adsorbed ana-lytes from the adsorbent is animportant critical parameter (38-43). Different eluent solutions (as)listed in Table II were used toelute the adsorbed analyte elementsfrom the Chromotrope FB impreg-nated magnetic multiwalled carbonnanotubes. Quantitative recoveriesfor both analyte ions were obtainedwhen 2.0 mL of 3.0 M HNO3 in 10%acetone was used as the eluent. Forother elution solutions, the recov-ery of one analyte generally wasquantitative; but the other ion wasnot recovered quantitatively. For allfurther studies, 2 mL of 3.0 MHNO3 in 10% acetone was used asthe eluent.

The volume of the samples isanother key parameter for solidphase extraction studies to obtain ahigh preconcentration factor (44-49). The effects of the sample vol-ume on the recoveries wereexamined by using 10–40 mL. Therecoveries were quantitative forboth Cu and Pb in the range of10–30 mL. The recovery valueswere not quantitative for samplevolumes higher than 30 mL. Thepreconcentration factor was30 with a final volume of 2 mL.

Matrix Effects

The negative and/or positiveeffects of alkali, alkaline earth met-als, and some metal ions at highconcentrations in real samples area big problem in the determinationof trace metals (50–56). The influ-ences of matrix components wereinvestigated, and the results arelisted in Table III. The model solu-tion containing analyte elementsand matrix components were pre-pared and then applied to thedeveloped procedure as describedin the Experimental section. Theresults clearly demonstrate that theinterfering ions did not interferewith the recovery values of theanalyte elements, and that the pro-posed method was fairly free frommatrix components as given inTable III.

Analytical Features

The detection limits (LODs) ofthe present method for lead andcopper, calculated based on 11determinations of the standard devi-ation of the blank, were 11.7 µg L-1

and 3.7 µg L-1, respectively. Therelative standard deviation (RSD)determined from 11 analyses wasbelow 5%.

Application to CRM Sample

The present method was appliedto CRM NCS-DC73349 BushBranches and Leaves to verify thevalidity of the method for copperand lead levels after the wet diges-tion procedure as given in theExperimental section. The resultsin Table IV show that the presentedsolid phase extraction method is ingood agreement with the certifiedvalues of the certified referencematerial.

The correctness of the presentprocedure was also verified by ana-lyzing the concentration after addi-tion of known amounts of Cu andPb onto a cigarette and a hair sam-ple. The results listed in Table V forthe cigarette sample and in Table

TABLE IIEffects of Various Eluents on the

Recoveries of the Analytes(Volume of Eluent: 2 mL; N=3)

Recovery (%)Eluent Type Pb(II) Cu(II)

0.25 M HNO3 92±2 81±40.5 M HNO3 96±3 82±41.0 M HNO3 91±2 84±32.0 M HNO3 94±1 88±23.0 M HNO3 96±3 90±10.25 M HCl 87±3 86±20.5 M HCl 98±3 91±31.0 M HCl 95±4 96±32.0 M HCl 96±2 90±13.0 M HCl 94±1 91±10.25 N HNO3

in 10% acetone 63±2 86±20.5 M HNO3

in 10% acetone 88±1 88±11.0 M HNO3

in 10% acetone 94±1 91±22.0 M HNO3

in 10% acetone 95±2 94±2

3.0 M HNO3in 10% acetone 97±1 97±1

TABLE IIIEffects of Some Matrix Ions on the Recoveries of Pb(II) and Cu(II)

(N=3)

Recovery (%)Ions Added as: Concentration Pb(II) Cu(II)

(µg/mL)

Na+ NaNO3 1000 97±0 99±1K+ KCl 1000 97±1 97±3SO4

2– Na2SO4 2500 94±1 95±2Fe3+ Fe(NO3)3.9H2O 5 91±2 92±1Zn2+ Zn(NO3)2.6H2O 10 98±1 100±2Mn2+ Mn(NO3)2.4H2O 10 100±2 102±1Cu2+ Cu(NO3)2.3H2O 10 100±2

Mg2+ Mg(NO3)2.6H2O 100 102±0 102±2

60

(from healthy subjects living in Kay-seri, Turkey) were determined byflame atomic absorption spectrome-try. The results are listed in TableVII.

Comparison of Proposed Proce-dure With PreconcentrationMethods in Literature

A comparison between thefigures of merit of the proposedmethod with other solid phaseextraction methods used for thepreconcentration/separation ofcopper and lead is given in TableVIII (55-62).

VI for the hair sample show thatgood agreement was obtainedbetween the added and found ana-lyte content using the present solidphase extraction system.

In addition, the analyte elementsof Cu and Pb in cigarette samples(purchased at a local market in Kay-seri, Turkey) and in hair samples

CONCLUSION

A simple and novel procedurebased on magnetic solid phaseextraction of lead and copper attrace levels on Chromotrope FBimpregnated magnetic multiwalledcarbon nanotubes in hair and ciga-rette samples from Kayseri, Turkey,has been established. The advan-tages of the presented magneticsolid phase extraction methodinclude time savings for adsorptionand elution steps, low use of use oforganic solvents, and the magneticadsorbent can be used at least fivetimes without loss of its adsorptionproperties.

Received November 8, 2016.

REFERENCES

1. X. H. Zhu, S. S. Lyu, P. P. Zhang, X.G. Chen, D. D. Wu, and Y. Ye, Env-iron. Earth. Sci. 75, 98 (2016). DOI10.1007/s12665-015-5036-9

2. S. Saracoglu, U. Divrikli, M. Soylak, L.Elci, and M. Dogan, J. Trace Micro-probe Techn. 21, 389 (2003).

3. G. Guleryuz, M. Arslan, B. Izgi, and S.Gucer, Z. Naturforsch. C 61, 357(2006).

4. O. Turkoglu, S. Saracoglu, M. Soylak,and L. Elci, Trace Elem. Electroly.21, 4 (2004).

TABLE IVApplication of Present Method

to CRM NCS-DC73349 BushBranches and Leaves for theDetermination of Pb and Cu

(N=3)

Lead Copper

Certifiedvalue (µg/g) 47 6.6

Found (µg/g) 45.2±3.0 6.1±0.8

Recovery (%) 96 93

TABLE VAddition–Recovery Tests forCigarette Sample (N=3) Using

Solid Phase Extraction


Pb 0 1.14±04 5.2±0.3 958 8.8±0.3 95

Cu 0 3.0±0.12 5.1±0.2 102

4 7.0±1.0 98

TABLE VIAddition-Recovery Tests for

Hair Sample (N=3) UsingSolid Phase Extraction


Pb 0 016 15.2±2 9524 22.7±1 94

Cu 0 0.4±0.18 8.2±0.1 97

12 12.2±1 98

TABLE VIIApplication of Present

Procedure Using Flame AASAnalysis of Cigarette and Hair

Samples (N=3)

Sample Concentration (µg/g)Pb Cu

Cigarette 2 BDL 13.2±0.3Cigarette 3 BDL 7.7±0.1Cigarette 1 BDL 10.2±0.4Hair 2 BDL 14.6±1.0

Hair 3 BDL 5.8±1.0

BDL: below the detection limit.

TABLE VIIIComparison of Proposed Method

With Some Methods in the Literature

Adsorbent PFa LODb (µg/L) Ref.

Magnetic allylamine modifiedgraphene oxide-poly(vinyl acetate-co-divinylbenzene) nanocomposite 40 Pb: 2.39, Cu: 2.34 32

Fe3O4/graphene magnetic nanoparticles - Pb: 0.87, Cu: 0.22 42

Core–Shell magnetic chitosan biopolymer 25 - 52

Oxidized single-walled carbon nanotubes 50 Pb:5.4, Cu: 3.9 57

Multiwalled carbon nanotubes 50 Pb:5.5, Cu:2.3 58

Chromotrope FB impregnated magneticmultiwalled carbon nanotubes 15 Pb: 11.7, Cu: 3.7 This

work

a PF: Preconcentration factor. b LOD: Limit of detection.

61


5. D.R. Chettle, J. Radioanal. Nucl. Ch.268, 653 (2006).

6. D. Mendil, O.F. Unal, M. Tuzen, andM. Soylak, Food Chem. Toxicol.48, 1383 (2010).

7. E. Chirila, S. Dobrinas, and V. Coatu,Rev. Chim-Bucharest. 55, 381(2004).

8. N. Alkan, A. Alkan, K. Gedik, and A.Fisher, Toxicol. Ind. Health 32,447 (2016).

9. S. Savasci, and M. Akcay, Turk. J.Chem. 20, 146 (1996).

10. F. Sipahi, and S. Uslu, Arab. J.Geosci. 9, 600 (2016). DOI10.1007/s12517-016-2620-6

11. R. Mehra, and M. Juneja, J. Ind.Chem. Soc. 81, 349 (2004).

12. N. Shaheen, N. Md. Irfan, I. N.Khan, S. Islam, Md. S. Islam, andMd.K. Ahmed, Chemosphere 152,431 (2016).

13. F. Yalcin, S. Kilic, D. G. Nyamsari,M. G. Yalcin, and M. Kilic, Pol. J.Environ. Stud. 25, 471 (2016).

14. U. Divrikli, A. A. Kartal, M. Soylak,and L. Elci, J. Hazard. Mater. 145,459 (2007).

15. D. S. Herman, M. Geraldine, C. C.Scott, and T. Venkatesh, Toxicol.Ind. Health 22, 249 (2006).

16. L. Vilizzi, and A.S. Tarkan, Environ.Monit. Assess. 188, 243 (2016).DOI 10.1007/s10661-016-5248-9

17. A. X. Wang, L. P. Guo, and D. M.Wu, Spectrosc. Spect. Anal. 26,1345 (2006).

18 M. Tuzen, S. Saracoglu, and M. Soy-lak, J. Food Nutr. Res. 47, 120(2008).

19. J. Semancikova, and D. Remeteiova,Holist. Approach Environ. 6, 55(2016).

20. M. Tuzen, M. Soylak, D. Citak, H.S.Ferreira, M.G.A. Korn, and M.A.Bezerra, J. Hazard. Mater. 162,1041 (2009).

21. M. Stanek, W. Andrzejewski, J.Mazurkiewicz, B. Janicki, D. Cygan-Szczegielniak, A. Roślewska, K.Stasiak, and I. Waszak, Pol. J. Envi-ron. Stud. 25, 301 (2016).

22. M. Ghaedi, R. Shabani, M. Montaze-rozohori, A. Shokrollahi, A.Sahraiean, H. Hossainian, and M.Soylak, Environ. Monit. Assess.

174, 171 (2011).

23. E. Pip, Environ. Health Persp. 108,863 (2000).

24. M. Tuzen, and M. Soylak, J. Hazard.Mater. 162, 724 (2009).

25. T. Stafilov, G. Pavlovska, and K.Cundeva, Turk. J. Chem. 24, 303(2000).

26. A. Shokrollahi, M. Ghaedi, O. Hos-saini, N. Khanjari, and M. Soylak, J.Hazard. Mater. 160, 435 (2008).

27. N. Ashouri, A. Mohammadi, R. Haji-aghaee, M. Shekarchi, and M.R.Khoshayan, Desalin. Water Treat.57, 14280 (2016).

28. S. Saracoglu, M. Soylak, D. Cabuk,Z. Topalak, and Y. Karagozlu, J.AOAC Int. 95, 892 (2012).

29. C.Q. Tu, and X.R. Wen, J. ChineseChem. Soc. 57, 356 (2010).

30. S.G. Ozcan, N. Satiroglu, and M.Soylak, Food Chem. Toxicol. 48,2401 (2010).

31. F. Raoufi, S. Bagheri, P. Nasehi, E.Niknam, K. Niknam, and H.R. Far-mani, Orient. J. Chem. 32, 575(2016).

32. M. Khan, E. Yilmaz, B. Sevinc, E.Sahmetlioglu, J. Shah, M.R. Jan,and M. Soylak, Talanta 146, 130(2016).

33. E. Aliyari, M. Alvand, and F. Shemi-rani, RSC Adv. 6, 64193 (2016).

34. R. Gao, and Z. Fan, At. Spectrosc.37, 195 (2016).

35. X. Lopez, and V.M. Castano, J.Nanosci. Nanotechn. 8, 5733(2008).

36. M. Soylak, L. Elci, and M. Dogan,Anal. Lett. 26, 1997 (1993).

37. E. Ghorbani-Kalhor, Microchim.Acta 183, 2639 (2016).

38. H.M. Marwani, A.E. Alsafrani, H.A.Al-Turaif, A.M. Asiri, and S.B. Khan,Bull. Mater. Sci. 39, 1011 (2016).

39. M. Soylak, Y.E. Unsal, N. Kizil, andA. Aydin, Food. Chem. Toxicol. 48,517 (2010).

40. S.Z. Mohammadi, and A. Seyedi,Toxicol. Environ. Chem. 98, 705(2016).

41. F.A. Aydin, and M. Soylak, J. Hazard.Mater. 173, 669 (2010).

42. Q. Yin, Y. Zhu, S. Ju, W. Liao, andY. Yang, Res. Chem. Intermed. 42,4985 (2016).

43. J.P.R. da Silva, G.F.B. Cruz, R.J. Cas-sella, and W.F. Pacheco, Quim.Nova, 39, 561 (2016).

44. M. Soylak, Anal. Lett. 37, 1203(2004).

45. M. Roushani, Y.M. Baghelani, S.Abbasi, S.Z. Mohammadi, M.Zahedifar, and M. Mavaei, Com-mun. Soil. Sci. Plan. Anal. 47, 1207(2016).

46. C. Duran, V.N. Bulut, D. Ozdes, A.Gundogdu, and M. Soylak, J. AOACInt. 92, 257 (2009).

47. N. Khorshidi, and A. Niazi, Separ.Sci. Technol. 51, 1675 (2016).

48. M. Soylak, Fresen. Environ. Bull. 7,383 (1998).

49. M. Khajeh, F. Saravani, M. Ghaffari-Moghaddam, and M. Bohlooli, Int.J. Vegetable Sci. 22, 266 (2016).

50. C. Maiti, R. Banerjee, and S. Maiti,D. Dhara, Des. Monomers. Polym.19, 669 (2016).

51. B. Bitirmis, D. Trak, Y. Arslan, andE. Kenduzler, Anal. Sci. 32, 667(2016).

52. M.H. Beyki, S. Miri, F. Shemirani, M.Bayat, and P.R. Ranjbar, Clean- SoilAir Soil Water 44, 710 (2016).

53. Z.A. AL Othman, E. Yilmaz, M.Habila, and M. Soylak, Desalin.Water Treat. 51, 6770 (2013).

54. S. Procházková, and R. Halko, Anal.Lett. 49, 1656 (2016).

55. M.H. Beyki, S. Miri, F. Shemirani, M.Bayat, and P.R. Ranjbar, Clean- SoilAir Soil Water 44, 710 (2016).

56. M. Soylak, and N. Kizil, At. Spec-trosc. 34, 216 (2013).

57. S. Chen, C. Liu, M. Yang, D. Lu, L.Zhu, and Z. Wang, J. Hazard.Mater. 170, 247 (2009).

58. E. Yilmaz, and M. Soylak, Environ.Monit. Assess. 186, 5461 (2014).

59. M. Soylak, I. Narin, L. Elci, and M.Dogan, Kuwait J. Sci. Eng. 28, 361(2001).

60. M. Soylak, and E. Yilmaz, Desalina-tion 275, 297 (2011).

61. S. Saracoglu, M. Soylak, and L. Elci,Trace Elem. Electroly. 18, 129(2001).

62. M. Ghaedi, M. Montazerozohori, F.Marahel, M.N. Biyareh, and M. Soy-lak, Chinese J. Chem. 29, 2141(2011).

62

* Corresponding author.:E-mail: [email protected]: +55 16 33019611

ABSTRACT

This paper describes thedevelopment of analytical meth-ods for the determination ofcobalt, iron, and nickel in puresilicon for photovoltaic and elec-tronics applications based onsolid sampling (SS) coupled to ahigh-resolution continuum sourcegraphite furnace atomic absorp-tion spectrometer (HR-CSGFAAS). Samples were also ana-lyzed by line-source flame atomicabsorption spectrometry (LSFAAS) as a comparative techniqueafter acid microwave-assisteddigestion. Cobalt, Fe, and Ni weredetermined in samples of solar-grade silicon (SoG-Si) and elec-tronic-grade silicon (EG-Si) bySS HR-CS GFAAS and LS FAAS. Apaired t-test at a 95% confidencelevel showed that the SS HR-CSGFAAS methods achieve similarresults to those obtained by LSFAAS. The relative standard devia-tions (n=12) for a sample contain-ing 7.80 mg kg-1 Co, 36.2 mg kg-1

Fe, and 228.5 mg kg-1 Ni were6.4% for Co, 6.1% for Fe, and1.0% for Ni for SS HR-CS GFAAS.For the SS HR-CS GFAAS, the lim-its of detection were 0.39 mg kg-1

Co, 1.14 mg kg-1 Fe, and 5.71 mgkg-1 Ni. Accuracy was alsochecked by the analysis of high-purity silica spiked with Co, Fe,and Ni, and the recoveries wereat 94.3–97.1% (Co), 86.7–109%(Fe), and 88.4–98.9% (Ni).

INTRODUCTION

Renewable energy and energyefficiency may be considered thepillars of sustainable energy, andthey are now entering global mar-kets as a result of research anddevelopment (1,2). The productionof pure silicon used as a raw mater-ial by the solar photovoltaic andmicroelectronic industries hasincreased in order to supply therapid-growth markets (3). The pres-ence of impurities in solar-gradesilicon (SoG-Si) and electronic-grade silicon (EG-Si) may signifi-cantly reduce the efficiency ofsemiconductor devices, solar cells,integrated circuits, etc. (4).Depending on the concentrationof Co, Ni, Cr, and Fe, theseelements may generate recombina-tion centers, which will consumeelectrons produced by the photo-voltaic effect in solar cells (5, 6).These devices may have their effi-ciency reduced by about 50% dueto the presence of these elements(4). Thus, rigid quality control ofthe SoG-Si and EG-Si materials isvery important.

Among the main spectroscopictechniques employed for elementaldeterminations, most use conven-tional sample introduction systemsfor the solutions. The conversion ofSoG-Si and EG-Si solid materials intosolutions is usually done by wetdigestions involving mixtures ofhydrofluoric acid with strong acidsat high temperature (7). These

energy, use large quantities of haz-ardous reagents, and generate sub-stantial amounts of waste. Thesedisadvantages may be avoided byusing direct solid sample analysis,an environmentally friendly methodoperating according to the princi-ple of green chemistry (8).

Solid sampling (SS) has beenexplored for the analysis of high-purity silicon by total reflectionX-ray fluorescence spectroscopy(9), laser-induced breakdown spec-troscopy (10), glow discharge massspectrometry (11), neutron activa-tion analysis (12), secondary ionmass spectrometry (13), and laser-ablation inductively coupledplasma mass spectrometry (14).These techniques are fast, sensitiveand accurate, but acquisition andmaintenance costs are relativelyhigher than other procedures suchas graphite furnace atomic absorp-tion (GFAAS).

High-resolution continuumsource graphite furnace atomicabsorption spectrometry coupledto solid sampling (SS HR-CS GFAAS)brought new possibilities for ele-mental spectrometric analyses (15)due to the ability to correct struc-tured background using least-squares background correction(LSBC), to measure the spectralenvironment in high resolution,furnish higher signal-to-noise ratiosthan those provided by linesources, to sum the absorbancemeasurement around the line core,or to sum the absorbance for differ-ent lines, to extend the linear work-ing range by measuring at alter-native lines, and to integrate auto-

procedures are arduous, time-con-suming, vulnerable to sample cont-amination and analyte losses,involve high consumption of


Determination of Cobalt, Iron, and Nickel inHigh-Purity Silicon by High-Resolution Continuum

Source Graphite Furnace Atomic AbsorptionSpectrometry Employing Solid Sample Analysis

Marcos André Bechlina, Ariane Isis Barrosa, Diego Victor Babosa, Edilene Cristina Ferreiraa,and José Anchieta Gomes Netoa,*

a UNESP – Univ. Estadual Paulista, Analytical Chemistry DepartmentRua Prof. Francisco Degni, 55, Quitandinha, 14800-060, Araraquara - SP, Brazil

63


matic solid sampler and microbal-ance which may reduce errorscaused by manual operations (16,17). The SS HR-CS GFAAS methodis efficiently employed for theanalysis of difficult samples such aspolymers (18), glass (19), coal (20),automotive catalyst (21), and oth-ers. However, little attention hasbeen given to the solid samplingmethod to analyze SoG-Si and EG-Simaterials.

Considering the above, thiswork deals with the developmentof analytical methods for the deter-mination of Co, Fe, and Ni contami-nants in So-GSi and EG-Si.

EXPERIMENTAL

Instrumentation

Atomic absorption measure-ments were carried out using anAnalytik Jena ContrAA 700 graphitefurnace atomic absorptionspectrometer (Jena, Germany),equipped with a short-arc Xenonlamp operated in “hot-spot” modeas a continuum radiation source,a high-resolution double monochro-mator, and a charge-coupled device(CCD) detector. Measurementswere done at the 252.136 nm (Co),305.760 nm (Ni), and 305.909 nm(Fe) lines. The lines of Ni and Fewere within the same spectral win-dow, so they were measured simul-taneously. The integrated peakabsorbance (peak volume selectedabsorbance) was equivalent to 3pixels (CP ± 1; CP: central pixelcorresponding to the center of theline). Solid sampling graphite tubeswithout a dosing hole, pyrolyticallycoated, and transversely heatedwere used throughout this work.The samples were weighed directlyonto solid sampling platforms usinga WZ2PW microbalance (Sartorius,Göttingen, Germany) with an accu-racy of 0.001 mg. The platformswere inserted into the graphitetubes using a SSA 600 solidautosampler (Analytik Jena). Highpurity argon (99.999%, White Mar-

tins, Sertãozinho, Brazil) was usedas the purge and protection gas.The graphite furnace heating pro-grams for the determination of Co,Ni, and Fe are presented in Table I.

A PerkinElmer® AAnalyst™100line source flame atomic absorptionspectrometer (PerkinElmer, Inc.,Shelton, CT, USA), equipped with adeuterium-lamp background correc-tion system, was used as the com-parative technique. Hollow cathodelamps for Co, Fe, and Ni were oper-ated at the 25, 30, and 25 mA cur-rent, respectively. Absorbancemeasurements at 240.7 nm (Co),248.3 nm (Fe), and 232.0 nm (Ni)using a 3-s integration time and0.2 nm spectral bandwidth wereused. Air and acetylene flow ratesfor all elements were established ata proportion of 4:2 oxidant/fuel.The aspiration flow rate and obser-vation height were fixed at 5.0 mLmin-1 and 7 mm, respectively.

An Anton Paar Multiwave®

microwave oven (Graz, Austria),equipped with six Teflon® flasks,was used for sample digestion.

Reagents, Standards, andSamples

All solutions were preparedusing high purity water (18.2 MΩ-cm resistivity) obtained from a Rios5® reverse osmosis system and aMilli-Q® Academic® deionizer (Milli-

pore Corporation, Bedford, NY,USA). Suprapur® nitric acid (MerckDarmstadt, Germany) was used forpreparing the solutions. Hydrofluo-ric acid (Merck, Darmstadt,Germany) and hydrogen peroxide(Panreac, Barcelona, Spain) wereused for wet digestion.

Working standard solutions con-taining 0.1 mg L-1 Co, 1.0 mg L-1 Feand Ni were prepared by properdilution of 1000 mg L-1 Co, Fe, andNi single stock standards (Specsol,São Paulo, Brazil) and acidified to0.1% (v/v) HNO3.

For LS FAAS analysis, calibrationcurves in the range of 0.1–4.0mg L-1 Co, Fe, and Ni wereprepared daily by properly dilutionof the standard solutions.

Solar- and electronic-grade sili-cons were furnished by the labora-tory of metallurgical processes ofthe Instituto de Pesquisas Tecnológ-icas do Estado de São Paulo (IPT)(São Paulo, Brazil). The sampleswere pulverized by scraping thesilicon piece with a silicon carbidespatula. Accuracy was assessed byaddition/recovery tests with 99%SiO2, 0.5-10 µm (Sigma-Aldrich,USA).

For comparative purposes, theSoG-Si and EG-Si samples weredigested according to the followingprocedure: a mass of 0.2 g of sam-

TABLE IOptimized Heating Program for the

Determination of Co, Fe, and Ni in SoG-Si by HR-CS GFAAS

Step Temperature Ramp Hold Time ArgonFlow Rate

(°C) (°C s−1) (s) (L min−1)

Drying 1 110 10 10 2.0Drying 2 130 5 10 2.0Pyrolysis 1600a, 1300b 50 15 2.0Auto-zero* 1600a, 1300b 0 5 0Atomization 2400a, 2650b 2400a, 3000b 6a, 10b 0

Cleaning 2600a, 2700b 500 5 2.0

* Step to ensure that the atomization starts without the presence of argon.a Conditions for Co; b Conditions for Fe and Ni determinations.

64

ples was accurately weighed andtransferred to the microwave flasks,followed by addition of 3.0 mL ofconcentrated nitric acid, 2.0 mLof concentrated hydrofluoric acid,and 1.0 mL of hydrogen peroxide.Then the optimized program ofthe microwave oven was run using(a) 15 minutes from 0 to 700 W;(b) 15 minutes from 700 to 900 W;(c) 15 minutes at 900 W; and (d)20 minutes at 0 W (ventilation).After digestion, the resulting solu-tions were transferred to digestiontubes and heated for 3 hours at 150oC in the digestion heating block inorder to eliminate the hydrofluoricacid. After cooling, the sampledigests were transferred to 25-mLvolumetric flasks and made up tovolume using distilled deionizedwater. The samples were digestedin triplicate.

All glasses and polypropylenevessels were washed with Extran®

detergent, soaked in 10% (v/v)HNO3 for 24 hours, then rinsedabundantly with deionized waterbefore use.

Analytical Procedure

The thermal behavior of the ana-lytes was evaluated in a multivariateexperiment using a factorial design2n, where n is the number of vari-ables. In this work, a full 23 factor-ial design was employed, thevariables evaluated were the pyroly-sis and atomization temperaturesand the sample size. Thesevariables were evaluated in two lev-els, minimum (–) and maximum(+). Table II lists the experimentalmatrix used for optimization,together with the minimum andmaximum levels of each evaluatedvariable for Co, Fe, and Ni.

The experimental data obtainedfrom the factorial design wereprocessed using the software Stat-graphics Centurion XVI (StatpointTechnologies, Warrenton, VA,USA), generating a mathematicalfunction whose maximization cor-

responded to the ideal analyticalconditions for each analyte (or forthe joint monitoring of theanalytes). The values of the threevariables optimized by the factorialdesign were used in the graphitefurnace heating program in subse-quent experiments.

After establishing the heatingprograms, analytical curves wereconstructed in the 0.0–2.0 ng Co,0.0–15.0 ng Fe, and 0.0–3.5 ng Nimass ranges, which were obtainedby delivering different aliquots of0.1 mg L-1 Co, 1.0 mg L-1 Fe and Nistandard solutions. The first pointof the calibration curves,corresponding to 0.0 ng (blank),was obtained by measuring theabsorbance with an empty platform(concept of zero mass).

The possibility of using calibra-tion with aqueous standards fordirect solid sample analysis wasevaluated by means of multipleeffects of matrices by comparingthe characteristic masses (m0) andslopes of the calibration curvesbuilt up in 0.1% (v/v) HNO3 andSiO2 media spiked with theanalytes.

Accuracy was evaluated bymeans of addition/recovery tests.Masses of 0.15 mg SiO2 were spiked

with 5.0 and 10.0 µL of 0.1 mg L-1

Co, 3.0 and 5.0 µL of 1.0 mg L-1 Feand Ni standard solutions in orderto obtain final concentrations of3.3 mg kg-1 Co, 6.7 mg kg-1 Co, and20 mg kg-1 and 33 mg kg-1 Fe andNi, respectively. The precision wasevaluated in terms of relative stan-dard deviation (%RSD) obtained forthree successive measurements ofeach sample.

The limits of detection (LOD)and quantification (LOQ) were cal-culated according to the IUPAC rec-ommendation: 3 x SDblank/b (LOD),and 10 x SDblank/b (LOQ), where SDis the standard deviation for 10measurements of the blank (usingempty platform) and b is the slopeof the calibration curve.

After optimizing the calibrationconditions, methods were appliedto the determination of Co, Fe, andNi in solar- and electronic-grade sili-con samples. Aliquots of aqueousstandards were manually injectedonto the SSA 600 platform usingmicropipettes. Sample amounts inthe range of 0.1– 0.2 mg were man-ually transferred to the graphiteplatforms, weighed and introducedinto the atomization compartmentusing the automated solid samplingaccessory. All measurements were

TABLE IIFactorial Design (23) Used for Optimization

of the Experimental Conditions

Experi- Pyrolysis Atomization Samplement* Temperature Temperature Mass

(°C) (°C) (mg)

1 + (1600a,1500b) + (2600a, 2650b) + (0.3–0.4a, 0.2–0.3b)2 + (1600a,1500b) + (2600a, 2650b) - (0.1–0.2)3 + (1600a,1500b) - (2400a, 2500b) + (0.3–0.4a, 0.2–0.3b)4 + (1600a,1500b) - (2400a, 2500b) - (0.1–0.2)5 - (1400a,1300b) + (2600a, 2650b) + (0.3–0.4a, 0.2–0.3b)6 - (1400a,1300b) + (2600a, 2650b) - (0.1–0.2)7 - (1400a,1300b) - (2400a, 2500b) + (0.3–0.4a, 0.2–0.3b)

8 - (1400a,1300b) - (2400a, 2500b) - (0.1–0.2)

* n=3.a Conditions for Co.b Conditions for Fe and Ni.

65


dow of the spectrometer. Shown inFigure 1 are the spectra recordedfor Co (Figure 1a) and Fe and Ni(Figure 1b) in a sample of solar-grade silicon.

After selecting the suitable ana-lytical lines, the heating program ofthe graphite furnace was optimizedin order to maximize the matrixremoval and to calibrate with aque-ous standards for solid sampleanalysis. For the evaluation of thethermal behavior of Co, Fe, and Ni,the maximum and minimum levelsof the multivariate experimentwere chosen after considering thedata published in earlier publica-tions (22–24) and using the condi-tions as recommended by themanufacturer.

The use of the factorial designmay reduce the number of experi-ments required for establishing theoptimum conditions in order toobserve further interactions amongthe variables which are not possiblein a univariate mode. The obtainedexperimental responses for eachcondition were used for generatinga mathematical model and evaluat-ing the influence of the variables onthe absorbance signals. For theselected intervals, no significantvariation in the Co absorbance wasobserved at the 95% confidence

carried out in triplicate (n=3), andthe integrated peak absorbancewas equivalent to 3 pixels.

For LS FAAS, the sample digestswere properly diluted in order toadjust the measured absorbance inthe linear working range. All mea-surements were carried out in trip-licate.


Thermal Behavior of Analytesand Heating Program ofAtomizer

The presence of metal contami-nants in solar-grade silicon andelectronic-grade silicon (e.g., Co,Fe, and Ni) may reduce theefficiency of silicon chips used inmost electronic equipment. First,measurements employing the mostsensitive lines of the analytesshowed absorbance exceeding thelinear working ranges, suggestingalternative lines with lower sensi-tivities should be used. The line ofCo at 252.136 nm (37% relativesensitivity) and the lines of Fe andNi at 305.909 nm (4% relative sensi-tivity) and 305.764 nm (3.3% rela-tive sensitivity) were selected forfurther experiments. It should benoted that Fe and Ni were simulta-neously monitored because theirlines are in the same spectral win-

level when the pyrolysis and atom-ization temperatures were varied.Higher analytical signals wereobserved for increased samplesizes. The optimum conditionsfor the pyrolysis and atomizationtemperatures were 1600 oC and2400 oC, respectively; and forsample mass it was 0.1–0.2 mg.

For Fe and Ni, the simultaneousmonitoring of both elements wasconsidered for selecting the opti-mum analytical signals. A significantincrease on the Fe response wasobserved by increasing the atomiza-tion and pyrolysis temperatures,and also for the interactions: (a)pyrolysis and atomization tempera-tures; (b) pyrolysis temperature andsample size; (c) pyrolysis tempera-ture, atomization temperature andsample size. Similarly to Co, no sig-nificant variation in the Niabsorbance was observed at the95% confidence level in the inter-vals evaluated for variables. Thismay be due to the more refractorycharacteristics of these elements incomparison to Fe. The optimumconditions found were 1300 oCpyrolysis temperature, 2650 oCatomization temperature, and0.1- 0.2 mg sample size. These con-ditions were employed for furtherexperiments.

Fig. 1. Time-resolved spectra of Co in the vicinity of line at 252.136 nm (a) and (b) Ni (305.909 nm) and Fe (305.760 nm)lines in SoG-Si.

66

samples. The analysis of the sam-ples showed concentrations in therange of 7.8–16.5 mg kg-1 for Co,28.0–36.2 mg kg-1 for Fe, and228–355 mg kg-1 for Ni. The rela-tive standard deviation (n=12) fora sample containing 7.80 mg kg-1

Co, 36.2 mg kg-1 Fe, and 228.5mg kg-1 Ni was 6.4% for Co, 6.1%for Fe, and 1.0% for Ni. The mainfigures of merit for the determina-tion of Co, Fe, and Ni in pure sili-con by the proposed SS HR-CSGFAAS methods are depicted inTable IV.

Matrix Effects

Matrix effects were evaluated bycomparing the slopes of the calibra-tion curves (0.0–2.0 ng Co,0.0– 15.0 ng Fe, and 0.0–3.5 ng Ni)built up in aqueous solution andsolid media (SoG-Si). For thissequence, good linear correlationcoefficients (r) were obtained forCo (0.9999, 0.9968), Fe (0.9959,0.9967), and for Ni (0.9977,0.9955); and the slopes of the ana-lytical curves were 0.22745,0.21372 (Co), 0.03791, 0.03812(Fe), and 0.0217and 0.0206 (Ni) foraqueous and solid calibration,respectively. These data expecterrors of < 6.0%, 1.6%, and 5.3% forCo, Fe, and Ni, respectively, whenaqueous standard calibration isused for analysis of the solid sam-ples. The characteristic masses cal-culated for aqueous and solidmedia, respectively, were 0.0195pg and 0.0199 pg (Co); 0.094 pgand 0.086 pg (Fe), and 0.53 pg and0.54 pg (Ni). Matrix matching oranalyte addition calibrations wouldminimize matrix effects. However,taking into consideration thaterrors of < 6% are acceptable forsolid sample analysis, aqueous cali-bration is straightforward and wasadopted for further studies.

Analytical Performance

The main figures of merit relat-ing to the direct determination ofCo, Fe, and Ni in solar- and elec-tronic-grade silicon were evaluated.A 0.10–0.20 mg quantity of samplewas weighed and assayed. The cal-culated characteristic masses calcu-lated for aqueous and solid mediawere 0.0195 pg and 0.0199 pg(Co); 0.094 pg and 0.086 pg (Fe),and 0.53 pg and 0.54 pg (Ni). Thelimit of detection, calculatedaccording to IUPAC recommenda-tion, was 0.39 mg kg-1, 1.14 mg kg-1,and 5.71 mg kg-1 for Co, Fe, and Ni,respectively. For comparison pur-poses, the samples were digestedand analyzed by LS FAAS. The val-ues for Co, Fe, and Ni obtained by

both techniques (Table III) were inagreement at the 95% confidencelevel (paired t-test). Also, accuracywas checked by addition/recoverytests. Cobalt, Fe, and Ni were deter-mined in high-purity silica spikedwith 3.3 and 6.6 mg kg-1 Co, 20 and33 mg kg-1 Fe and Ni. Recoverieswere in the range of 94–97% (Co),87–109% (Fe), and 88–99% (Ni).The found values were in agree-ment with the spike values for Co,Fe, and Ni at the 95% confidencelevel. The method was then appliedfor Co, Fe, and Ni determination insolar- and electronic-grade silicon

TABLE IIIResults (mean ± standard deviation) for Co, Fe, and Ni (mg kg-1)

Determined (n=3) in SoG-Si and EG-Siby Proposed DSS HR-CS GFAAS and LS FAAS Comparatve Technique

Analyte SoG-Si EG-SiDSS HR-CS LS DSS HR-CS LSGFAAS FAAS GFAAS FAAS

Co 7.80 ± 0.50 8.58 ± 0.62 16.5 ± 1.2 16.4 ± 1.3Fe 36.2 ± 2.2 37.9 ± 1.9 28.0 ± 2.7 27.5 ± 0.8

Ni 228.5 ± 2.2 233.7 ± 5.9 355.0 ± 28.8 329.1 ± 23.0

TABLE IVMain Figures of Merit for the Determination of Co, Fe, and Ni

in Pure Silicon by Proposed Method Based on DSS HR-CS GFAAS

Parameters Co Fe Ni

LOD (mg kg-1) 0.39 1.14 5.71

LOQ (mg kg-1) 1.30 3.80 19.0

m0 (pg) 0.0195 (aqueous); 0.094 (aqueous); 0.53 (aqueous);0.0199 (solid) 0.086 (solid) 0.54 (solid)

RSD (%) 6.0 1.6 5.3

b 0.22745(aqueous); 0.03791(aqueous); 0.0217(aqueous);0.21372 (solid) 0.03812 (solid) 0.0206 (solid)

r 0.9999 (aqueous); 0.9959 (aqueous); 0.9977 (aqueous);0.9968 (solid) 0.9967 (solid) 0.9955 (solid)

PyrolysisTemperature (°C) 1600 1300 1300

AtomizationTemperature (°C) 2400 2650 2650

Sample Mass (mg) 0.1–0.2 0.1–0.2 0.1–0.2

Modifier No No No

LOD: limit of detection; LOQ: limit of quantification; m0: characteristic mass;RSD: relative standard derivation; b: slope; r: correlation coefficient.

67


CONCLUSION

The determination of contami-nants in a silicon matrix is an ana-lytical challenge due to thedifficulties and complexities of sam-ple preparation. The heating pro-gram and sample size wereoptimized by using a factorialdesign which reduces the numberof experiments and enables theconsideration of interactionsbetween the variables. Sampleconcentrations ranged between7.8–16.5 mg kg-1 (Co); 28.0–36.2mg kg-1 (Fe), and 228–355 mg kg-1

(Ni). Cobalt, Fe, and Ni were pre-cisely and accurately determined.The employment of HR-CS GFAAScombined with solid samplingallows the determination of the ana-lytes in high-purity silicon withoutany previous sample preparationand addition of reagents.

ACKNOWLEDGMENT

The authors would like to thankthe Fundação de Amparo à Pesquisado Estado de São Paulo for financialsupport of this work (Grant#2014/12595-1). The authors arealso grateful to the Coordenação deAperfeiçoamento de Pessoal deNível Superior (CAPES) for fellow-ships to D.V.B. and to the ConselhoNacional de Desenvolvimento Cien-tífico e Tecnológico (CNPq) for fel-lowships to A.I.B. and M.A.B., andresearchship to J.A.G.N.

Received Novemvee 7, 2016.

REFERENCES

1. K.H. Solangi, M.R. Islam, R. Saidur,N.A. Rahim, H. Fayaz, Renew. Sus-tain. Energy Rev. 15, 2149 (2011).

2. G.R. Timilsina, L. Kurdgelashvili,P.A. Narbel, Renew. Sustain.Energy Rev. 16, 449 (2012).

3. M. Hystad, C. Modanese, M. DiSabatino, L. Arnberg, Sol. EnergyMater. Sol. Cells 103, 140 (2012).

4. Y. Delannoy, J. Cryst. Growth 7,

360 (2012).

5. G. Coletti, P.C.P. Bronsveld, G.Hahn, W. Warta, D. Macdonald, B.Ceccaroli, Adv. Funct. Mater. 21,879 (2011).

6. S. Meyer, S. Wahl, A. Molchanov,K. Neckermann, C. Moller, K.Lauer, Sol. Energy Mater. Sol. Cells130, 668 (2014).

7. G. Kolosovska, A. Viksna, G. Chik-vaidze, A. Osite, A. Opalais, Mater.Sci. Eng. 38, 1 (2012).

8. C. Bendicho, I. Lavilla, F. Pena-Pereira, V. Romero, J. Anal. At.Spectrom. 27, 1831 (2012).

9. J. Rip, K. Wostyn, P. Mertens, S.De Gendt, M. Claes, EnergyProced. 27, 154 (2012).

10. S. Darwiche, M. Benmansour, N.Eliezer, D. Morvan, Prog Photo-voltaics Res. Appl. 20, 463 (2012).

11. M. Di Sabatino, A.L. Dons, J. Hin-richs, L. Arnberg, Spectrochim.Acta Part B 66, 144 (2011).

12. J. Hampel, F.M. Boldt, H. Gersten-berg, G. Hampel, J.V. Kratz, S.Reber, Appl. Radiat. Isto. 69, 1365(2011).

13. P. Peres, A. Merkulov, F. Desse, M.Schuhmacher, Surf. InterfaceAnal. 43, 643 (2011).

14. P.L. Buldini, A. Mevoli, J.L. Sharma,Talanta 47, 203 (1998).

15. M.A. Belarra, M. Resano, F. Van-haecke, L. Moens, TrAC TrendsAnal. Chem. 21, 828 (2002).

16. B. Welz, Anal. Bioanal. Chem. 381,69 (2005).

17. M. Resano, M. Aramendía, M.A.Belarra, J. Anal. At. Spectrom. 29,2229 (2014).

18. A.T. Duarte, M.B. Dessuy, M.G.R.Vale, B. Welz, Anal. Methods 5,6941 (2013).

19. S. Kelestemur, M. Özcan,Microchem. J. 118, 55 (2015).

20. A.F. da Silva, D.L.G. Borges, F.G.Lepri, B. Welz, A.J. Curtius, U.Heitmann, Anal. Bioanal. Chem.382, 1835 (2005).

21 M. Resano, M.D.R. Flórez, I. Quer-alt, E. Marguí, Spectrochim. ActaPart B 105, 38 (2015).

22. B. Gómez-Nieto, M.J. Gismera,

M.T. Sevilla, J.R. Procopio, Talanta116, 860 (2013).

23. M. Resano, E. Bolea-Fernández, E.Mozas, M.R. Flórez, P. Grinberg,R.E. Sturgeon, J. Anal. At.Spectrom. 28, 657 (2013).

24. A.S. Ribeiro, M.A. Vieira, A.F. daSilva, D.L.G. Borges, B. Welz, U.Heitmann, Spectrochim. Acta PartB 60, 693 (2005).

68Atomic SpectroscopyVol. 38(3), May / June 2017

Feasibility of a Fast and Green Chemistry SamplePreparation Procedure for the Determinationof K and Na in Renewable Oilseed Sourcesby Flame Atomic Emission Spectrometry

Kamyla Cabolon Pengoa, Vanessa Cruz Dias Peronicoa, Luiz Carlos Ferreira de Souzab,and Jorge Luiz Raposo, Jr.a*

a Federal University of Grande Dourados, School of Exact and Technology Science,PO Box 364, 79804-970 Dourados, MS, Brazil

b Federal University of Grande Dourados, School of Agronomic Science,PO Box 364, 79804-970 Dourados, MS, Brazil

Corresponding author.E-mail: [email protected].: +55 67 34102092

INTRODUCTION

Agriculture is a modern, prosper-ous, and highly competitive sectorin Brazil, and it is considered as thepropelling agent of the nationaleconomy. Among various sectorsof agriculture, soybeans and sugar-cane are the most important, andtogether with coffee and livestockthey are currently the pillars ofBrazil’s economy (1). Although soy-beans are one of the main productsused in the Brazilian culture and are

ABSTRACT

This work describes a fast andgreen chemistry procedure todetermine potassium (K) andsodium (Na) in alternative oilseedcrops by flame atomic emissionspectrometry (FAES) using anultrasound (US) system for sam-ple preparation. The use of 10 mLof a 0.12 mol L-1 HCl solution, 10minutes of extraction, and 25 ºCallowed the use of ≈0.1000 g ofsamples to extract the mineralcontent of the samples. The mainand secondary atomic lines wereevaluated, but only the secondary(404.4 nm for K and 330.3 nmfor Na) provides a satisfactory(1.00–150.00 mg L-1 K and1.00–120.00 mg L-1 Na) analyticalcalibration range for the determi-nation of K and Na in a single runwithout need of further dilution

Potassium and sodium are themost important elements for vege-tative growth, and constitutearound 10% of its dry matter (2-4).These elements are required formany metabolic processes such asosmoregulation, protein synthesis,photosynthesis, opening and clos-ing of the stomata, soil and waterabsorption, enzymatic activity, andtherefore monitoring of the qualityof agricultural crops is essential(5-7).

The elemental determinationof K and/or Na is frequently per-formed by spectrometric tech-niques (8-15). Most of thesemethods involve a sample pretreat-ment step, which can be done byconverting the sample into an aque-ous solution using mineral acidsand thermal or radiant energy fororganic matter decomposition (16-19). This is a critical step in routineanalysis, but is very time-consum-ing, results in incomplete solubiliza-tion of the matrices, analyte lossesby volatilization, contamination inthe handling processes due to theinteraction between analyte andbottles, and contamination of thesolutions by the reagents used(16-18, 20, 21).

From this perspective, non-destructive sample pretreatmentprocedures, such as those employ-ing ultrasonic waves, are an alterna-tive to circumvent acid digestions,and can be used for extraction, sol-ubilization, and digestion processesusing an ultrasonic bath or a probe(22, 23) with diluted acids and is

raw materials used for the produc-tion of vegetable oil and/orbiodiesel, alternative and renew-able oilseed sources are often evalu-ated as another effective option forthese purposes. Oilseed cropsrequire no new investment of agri-cultural implements, contributes toimproving the crop rotationsystem, and can increase thefarmer's income by offering com-petitive profits. Since there is a lackof research done about the alterna-tive uses of oilseed crops, an inves-tigation related to the mineralcontent of these oilseed cropsshould be performed.

of the samples. Recoveries of Kand Na added to samples variedfrom 93.7–99.1% with the preci-sion better than 3.5%. Five sam-ples of renewable oilseeds wereanalyzed by FAES with the pro-posed sample preparation proce-dure using a closed-vesselmicrowave-assisted acid digestionfor comparative purposes. Theresults obtained using an ultra-sound sample preparation proce-dure were in agreement at the95% confidence level (pairedt-test), with those obtained bymicrowave-assisted digestion.The found concentrations were6.02±0.11 – 7.75±0.47 mg g-1 Kand 1.41±0.05 – 2.01±0.10 mg g-1

Na, with a precision better than5.3%. The limit of detection was52.45 and 73.83 µg g-1 for K andNa, respectively.

69


performed at room temperature(23-25). The most important advan-tages associated with the ultrasonicwaves includes: (a) reduced timerequired for sample preparation,(b) reduced consumption of con-centrated reagents, and (c) samplepreparation is simple and relativelylow cost (23, 25, 26).

In this sense, this work describesthe first development of a simpleand robust sample preparation pro-cedure using an ultrasonic extrac-tion system for the determinationof K and Na in renewable oilseedsources by flame emission absorp-tion spectrometry.

EXPERIMENTAL

Instrumentation

The measurements were carriedout using a Varian 240FS flameatomic absorption spectrometer(Agilent Technologies®, Mulgrave,Victoria, Australia) operating inemission mode. The instrumentaloperating parameters are listed inTable I. High-purity acetylene(99.7% White Martins, Dourados,Brazil) was used as fuel, and an air-acetylene flame was used for atom-ization of K and Na.

A 515 Orion (Fanem®, São Paulo,SP, Brazil) forced air oven and a TE-361 (Tecnal®, Piracicaba, SP,Brazil) stainless steel mill were usedto dry and powder the oilseed sam-ples, respectively. A Unique USC-14004 (Unique®, Indaiatuba, SP,Brazil) ultrasonic bath, operating at40 kHz, was used to leach the ele-ment from the oilseed samples. AnExcelsa™ 206 BL (Fanem®, SãoPaulo, SP, Brazil) centrifuge wasused to separate the solid and liq-uid phases. A Multiwave® 3000microwave oven (Anton Paar, Graz,Austria) was used as a comparativesample preparation procedurewhich employed acid digestion ofthe samples.

Reagents and AnalyticalSolutions

High purity deionized waterobtained from a Milli-Q® Plusreverse osmosis system (resistivity18.2 MΩ-cm, Millipore Corporation,Bedford, MA, USA) was used to pre-pare all solutions. Nitric acid [65%(v/v), Sigma-Aldrich®, St. Louis,MO, USA] was used to prepare allanalytical solutions, to optimize theextracting acid solutions, and forthe microwave-assisted acid diges-tion. Hydrochloric acid [37% (v/v),Sigma-Aldrich®, St. Louis, MO, USA]was used as the extracting acidsolution. A 150.000 mg L-1 Cs solu-tion was prepared by dissolving9.56 g CsCl [99.5% purity, Sigma-Aldrich®, St. Louis, MO, USA) withdeionized water, followed byadding 1.5 mL HNO3, and dilutionto 50 mL using deionized water.

Different dilute hydrochloricacid solutions (0.012, 0.120, and1.200 mol L-1) were prepared forthe evaluation of the effect of theacid concentration of the extrac-tion solution on the ultrasonic-assisted extraction.

Analytical solutions containing1.00 – 150.0 mg L-1 K and 1.00 –120.0 mg L-1 Na were prepareddaily using appropriate dilution of1000 mg L-1 single-element standardstock solutions (SpecSol®, SRM-682,USA) in 1.0% (v/v) nitric acid and1000 mg L-1 Cs media.

All of the solutions were storedin high-density polypropylene bot-tles (Nalgene®, Rochester, NY,USA). Plastic bottles and glasswarewere cleaned by soaking in 10%(v/v) HNO3 for at least 24 hours,followed by thorough rinsing withdeionized water before use.

Sampling and SamplePreparation

A 1000-g amount of four-monthsold seeds of the genus Crambeabyssinica Hochst, Guizotiaabyssinica, Brassica napus var.oleífera, Carthamus tinctorius L.,and Raphanus sativus L. var.oleiferus Metzg renewable oilseedsources were randomly collectedduring the 2014 harvest time fromfive different areas (15 x 30 m)located in the Experimental Farm ofthe Agricultural Sciences (22º 14’Sand 54º 49’W), which is situated8 km from the Federal University ofGrande Dourados (Dourados, MS,Brazil). The samples of each specieswere stored in individually labeledpaper bags, and then transported tothe laboratory. The seeds were sep-arately and thoroughly washed withtap water and then with deionizedwater. A 10.0-g subsample of eachoilseed sample was dried at 80 ºCfor 120 hours in a forced air oven,ground in a stainless steel mill, andthen sieved using a No. 20 sieve(0.84 mm opening size).

TABLE IFAES Instrumental Operations Conditions

for the Determination of Na and K

Instrumental Conditions K NaWorking range (mg L-1) 1.00 – 150.00 1.00 – 120.00

Wavelength (nm) 404.4 330.3

Burner head (mm) 100Air flow rate (L min-1) 13.0Acetylene flow rate (L min-1) 2.0

Slit width (nm) 0.5

70

A 0.1000-g portion of pretreatedsample was accurately weighed(± 0.0001 g) and transferred to a50-mL polypropylene flask,followed by addition of 15 mL acidleaching solution. The mixtureswere subjected to an ultrasoundenergy corresponding to 40 kHz for30 seconds to leach the elementsfrom the seeds into the acid solu-tion. After sonication, the acidextracts obtained were separatedfrom the remaining solid materialsusing centrifugation for 5 minutesat 4000 rpm, followed by filtrationinto 25-mL polypropylene flasks. Allof the acid leaching solutions (sam-ples and blanks) were prepared in1000 mg L-1 Cs. Three replicates ofeach alternative oilseed sample andblank were used to optimize theanalytical parameters of the extrac-tion procedure. All of the measure-ments were performed using fivereplicates.

For comparative purposes, thealternative oilseed materials werealso mineralized in a closed-vesselmicrowave-assisted digestion proce-dure. For microwave digestion, anaccurately weighed 0.1000-g sam-ple was transferred into a micro-wave flask, followed by additionof 6.0 mL of HNO3 and 2.0 mL ofdeionized water. The optimizedheating program is listed in TableII. After digestion and cooling, theresulting solutions obtained weretransferred into 25-mL volumetricflasks, and brought to volume withdeionized water in 1000 mg L-1 Csmedia. Three sample replicateswere used for microwave-assistedacid decompositions.

Ultrasonic (US) ExtractionConditions and MeasurementProcedure

The influence of the extractingsolution (acid, concentration, andvolume), sample mass, and sonica-tion time on the sensitivity for Kand Na was investigated usingPIATV 02/2010 Soybean referencematerial (0.84 mm particle size)

obtained from Embrapa Agrope-cuária Oeste (Dourados, MS, Brazil)by varying the acid solution [HNO3,HCl or HNO3:HCl (1:1, v/v)], con-centration (0.012, 0.120 and 1.200mol L-1), the volume of the extract-ing solution (7.5, 10.0 and 15.0mL), sample mass (0.1000, 0.2500and 0.5000 g), and the sonicationtime (30, 60, 120 and 180 seconds).Due to the absence of Na in PIATV02/2010, an aliquot of 1000 mg L-1

Na standard solution was added tothe selected soybean referencematerial in order to achieve a 10.0-mg g-1 Na content.

At a 5.0-mL min-1 sample flowrate, the integrated emission inten-sity for blanks, working standardsolutions, reference material extrac-tion solution, and the sample solu-tions were measured at lesssensitive atomic lines for K at404.4 nm and Na at 330.3 nmunder optimal instrumental condi-tions to obtain the calibration curvewithin the 1.00 – 150.0 mg L-1 Kand 1.00 – 120.0 mg L-1 Na ranges.All measurements were carried outin five replicates.

The matrix effect was checkedusing recovery tests for spiked sam-ples performed at two levels byadding aliquots of the 1000 mg L-1

single-element standard stock solu-tions to all alternative oilseed sam-ples before the extraction pro-cedure to obtain extracts with 25.0and 50.0 mg L-1 of K and Na. Thelimit of detection (LOD) and limitof quantification (LOQ) were calcu-lated according to the IUPAC rec-ommendation (27). Statistical tests

used in the data processing (mean,standard deviation, and precision)were done using the MicrocalOriginPro® 8.0 program andMicrosoft® Office Excel® 2007.


Analytical Features forDetermining K and Na

Whereas the content of macro-nutrients in plants is typically in theorder of mg g-1 (14, 28), the deter-mination of K or Na by line sourceflame atomic emission spectrome-try (LS-FAES) in one run is not feasi-ble if the most sensitive analyticallines (766.5 nm for K or 589.0 nmfor Na) are used due to the limitedlinear calibration interval (29). ForK and/or Na concentrations higherthan the upper limit of the linearresponse of the calibration plotsbuilt up using the most sensitiveanalytical lines, the secondary linefor K at 404.4 nm or Na at 330.3nm might be employed to circum-vent this problem. The use of a lesssensitive atomic line is a better wayto reduce sensitivity for determin-ing major elements and to avoidfurther dilution of the sample solu-tions. At 5.0 mL min-1 of sampleflow rate and 100 mm burner open-ing (standard air/acetylene burnerhead), the influence of the ratio ofair-acetylene flow rates (0.138,0.146, 0.154, 0.162, and 0.169)on emission intensity of K and Nawas evaluated using 1.0 mg L-1 at766.5 nm (K) and 589.0 nm (Na),and 10.0 mg L-1 at 404.4 nm (K)and 330.3 nm (Na). For this, differ-ent fuel-oxidant ratios wereobtained by changing the flow rate

TABLE IIMicrowave-Assisted Digestion Heating Program

for Renewable Oilseed Samples

Step Tramp (min) Thold (min) Power (W) T (°C)

1 10 5 600 120

2 15 10 1000 200

3 0 15 0 Ventilation

71


of acetylene from 1.8 to 2.2 L min-1

and fixing the air flow rate at 13.0L min-1. The best situationsachieved for flame compositionfor all wavelengths studied was2.0 L min-1 of acetylene. Under theestablished conditions of the 240FSequipment, the linear workingrange at 766.5 and 404.4 nm for Kand 589.0 and 330.3 nm for Na wasevaluated by plotting curves ofemission intensity versus K orNa concentration within the 0.10 –150.00 mg L-1 intervals. The calibra-tion plots in the 0.10 – 150.00mg L-1 intervals provide calibrationcurves only up to 4.00 mg L-1 at766.5 nm for K, and up to 2.00mg L-1 at 589.0 nm for Na.However, the less sensitive analyti-cal atomic lines provide calibrationplots up to 150 mg L-1 K and 120mg L-1 Na, with typical linear corre-lation coefficients better than0.9980 for K and 0.9992 for Na.The main figures of merit for theK and Na atomic lines are shownin Table III.

Analysis of Table III reveals thatthe highest sensitivity, as seen bythe slopes, are 0.2463 K and 0.1718Na. Lower limits of detection werepbtained with the main atomic line,however, with a narrow (0.50–4.00mg L-1 K and 0.50–2.00 mg L-1 Na)linear working range and perhapsinsufficient for determining highconcentrations of K and Na. In thiscase, where the content of the ana-lyte in the sample digests and/orextracts is naturally above the lin-ear range of calibration for 766.5nm (K) and 589.0 nm (Na), thealternate atomic line at 404.4 nm(K) and 330.3 nm (Na) can be usedto determine K and Na in theselected samples. The use of sec-ondary and less sensitive (0.0063-Kand 0.0061-Na slopes) atomic linesallowed extending the linear cali-bration range up to 150.0 mg L-1 Kand 120.0 mg L-1 Na with a satis-factory limit of detection (0.2098mg L-1 K and 0.2956 mg L-1 Na) andlow relative standard deviation

cases the relative standard devia-tions (%RSD) were < 4.6%. Only for1.00 mol L-1 HCl solution the recov-ery (89% for K and 83% for Na) canbe acceptable and produced RSD of< 2.5%. It is important to note thatfor a 1.00-mol L-1 HCl solution, therecovery was 89% for K (6.50 ±0.14 mg g-1) and 83% for Na (8.30 ±0.21 mg g-1). This value is statisti-cally different at the 95% con-fidence level (paired t-test) withthose obtained for 1.00 mol L-1

HNO3 or HCl:HNO3 solutions. Bythe way, alternative acid mixtures,such as aqua regia [3:1 (v/v)]HCl:HNO3 and HCl:HNO3 [1:3(v/v)] at 1.00 mol L-1 were alsoinvestigated for the extraction pro-cedure. In both cases the resultswere no better than those obtainedusing 1.00 mol L-1 HCl:HNO3 [1:1(v/v)] solution. According to theresults presented in Table IV, theremaining study was performedusing HCl solution.

In sample preparation usingultrasound-assisted extraction,diluted acid solutions are widelyemployed as extracting solvent/solution to leach out as much ana-lyte content as possible withoutdestroying the sample matrix bymeans of minimum amount ofselected acids (26, 33). Using theacid solution chosen previously,different concentrations ofhydrochloric acid (0.012, 0.120,

(< 2.4%), suggesting that the sec-ondary line gives precise measure-ments. Then, the 404.4 nm for Kand 330.3 nm for Na atomic lineswere used to optimize the US pro-cedure, to validate the methodol-ogy, and to determine the elementsby FAES.

Optimization of Ultrasound-assisted Extraction Procedure

Nitric and/or hydrochloric acidare often reported in studies involv-ing extraction of the inorganicspecies using non-destructive sam-ple preparation procedures (30,31); however, acid mixtures havealso been reported (32). In thissense, different acid solutions [HCl,HNO3 and HCl:HNO3 (1:1)] at 1.00mol L-1 were evaluated to check theeffectiveness of these solutions forleaching K and Na from PIATV02/2010 reference material. Theinfluence of these different acidsolutions was determined using uni-variate analysis by fixing the solu-tion volume (10 mL), sonicationtime (180 seconds), sample mass(0.2500 g), and bath temperatureat 25 ºC. The results obtained areshown in Table IV.

It can be seen in Table IV thatunsatisfactory recoveries wereobtained with 1.00 mol L-1 HNO3

(67% for K and 62% for Na) orHCl:HNO3 (≈76% for K and ≈69%for Na) solutions. However, in both

TABLE IIIFigures of Merit of Main and Secondary Atomic Lines

for K and Na by FAES

Element Wavelength Calibration Slope Rc LODd RSD(nm) (mg L-1) (µg L-1) (%)

K 766.5a 0.50–4.00 0.2463 0.9989 9.61 3.5

404.4b 1.00–150.00 0.0063 0.9980 209.80 0.9

Na 589.0a 0.50–2.00 0.1718 0.9991 15.55 1.9

330.3b 1.00–120.00 0.0061 0.9992 295.30 2.4

a Main atomic line.b Secondary atomic line.c Linear correlation coefficient.d Limit of detection.

72

It is important to mention thatparticle size also plays an importantrole in the extraction process (38-41), and this was one of the para-meters that was evaluated toachieve the best conditions of theextraction procedure. Since PIATV02/2010 reference material wasground to 0.84 mm (20 mesh) parti-cle size, and the results obtainedwere 91% (K) and 84% (Na) recov-ery, it is possible to conclude thatparticle size does not have a signifi-cant influence on the extractionefficiency for these kinds of sam-ples.

and 1.200 mol L-1) were preparedto evaluate their performance. Theinfluence of these HCl solutions onthe K and Na recoveries were doneby fixing the solution volume (10mL), sonication time (180 seconds),sample mass (0.2500 g), and bathtemperature (25 °C). The resultsobtained for these solutions arelisted in Table IV.

A significant improvement wasobserved in the recoveries from the0.012 to 0.120 mol L-1 HCl. Whilethe 0.012 mol L-1 HCl solution canleach only 77% K (5.62 ± 0.15mg L-1 of 7.30 mg g-1) and 70% Na(7.00 ± 0.24 mg L-1 of 10.0 mg g-1)from the PIATV 02/2010 referencematerial, the 0.120 and 1.200mol L-1 HCl solutions obtained 91and 90% recoveries for K and 84and 86% recoveries for Na, respec-tively, with a relative standard devi-ation of < 2.7%. A paired t-test, atthe 95% confidence level, was per-formed and the results showed adifference between 0.120 and1.200 mol L-1 with the 0.012mol L-1 HCl solutions. Based on thesatisfactory recoveries obtainedusing 0.120 and 1.200 mol L-1 HCland due to a green chemistry prin-ciple, a 0.120 mol L-1 HCl wasadopted for further study.

Some studies report that smallsample masses (high dilutions)could produce low sensitivity andpoor precision/accuracy of the ana-lyte measures due to the inhomo-geneity of the sample mass of thesolid-liquid extracting solution (34-36). In this sense, the increase ofthe mass of solid materials (sample)can improve the limits of detectionand quantification due to the trans-ference of higher contents of theanalyte into the liquid phase (37).Taking this into account, portionsof 0.1000, 0.2500, and 0.5000 goilseed sample were evaluated todetermine the satisfactory samplemass that produces the best recov-ery of K and Na. Potassium and Narecoveries using 0.1000 – 0.5000 g

of the PIATV 02/2010 referencematerial were obtained by using0.120 mg L-1 HCl and fixing thesolution volume (10 mL), sonica-tion time (180 seconds), and bathtemperature (25 °C). The resultsdescribed in Table IV reveal no sig-nificant influence at 95% confi-dence level (paired t-test), ofsample mass up to 0.5000 g, and inall cases, the relative standard devi-ation was < 2.1%. Since no signifi-cant differences can be observedfor the determination of K and Nausing 0.1000 – 0.5000 g, an amountof 0.1000 g was used for furtherwork.

TABLE IVRecovery (Mean ± Standard Deviation, n= 3) of K and Na

from PIATV 02/2010 Reference Material on theOptimization of the Ultrasound-Assisted Extraction Conditions

K NaUS Extraction Conditions Recovery RSD Recovery RSD

(%) (%) (%) (%)

ExtractingSolution HCl 89 2.2 83 2.5

HNO3 67 3.7 62 4.6

HCl:HNO3 [1:1(v/v)] 76 3.0 69 4.0

HCl Solution(mol L-1) 0.012 77 2.7 70 3.4

0.120 91 1.9 84 2.7

1.200 90 1.6 86 2.3

Sample Mass(g) 0.1000 91 1.6 84 1.9

0.2500 93 1.7 85 2.1

0.5000 92 1.2 87 2.0

Volume(mL) 7.50 80 4.1 78 3.5

10.00 92 2.0 86 1.7

15.00 98 2.7 95 2.2

SonicationTime (s) 30 98 1.9 95 2.1

60 97 1.4 94 2.0120 99 0.9 94 1.9

180 99 1.1 95 1.5

K: 7.30 ± 0.67 mg g-1, Na: 10.0 mg g-1

73


For solid-liquid extraction ofmetals from solid matrices usingultrasound-assisted extraction, dif-ferent volumes of extracting solu-tions have been employed (39, 42,43). In this work, 7.5, 10.0, and15.0 mL of 1.0 mol L-1 HCl wereapplied to evaluate the recoveries.The analytical results are shown inTable IV. The influence of theextraction volume on extractionefficiency was significant from7.5 to 15.0 mL, and the recoveryobtained for 7.5 mL (80.0% for Kand 78.0% for Na) was statisticallydifferent, at a 95% confidence level(paired t-test), with those obtainedfor 15.0 mL (98% for K and 95.0%for Na). Thus, 15.0 mL of 0.120mol L-1 HCl was chosen for theoptimization parameters of theultrasound-assisted extractionprocedure.

Based on the results obtained inthis study (≈95% recovery), the pro-posed extraction procedure couldbe applied to determine K and Nausing FAES. However, the extrac-tion time efficiency depends onanalyte-matrix interaction, extract-ing solution composition, and theultrasonic system applied (26, 44).In this study, the extraction timewas evaluated to reduce the timeconsumption of the proposedmethod. For this, 0.1000 g ofPIATV 02/2010 reference materialwas subjected to 30, 60, 120, and180 seconds sonication time underthe previously optimized conditions.The results obtained are listed inTable IV.

Quantitative extractions of K(97 – 99%) and Na (94 – 95%) wereobtained using 30 – 180 secondssonication intervals with an ultra-sound energy corresponding to40 kHz. The results showed nosignificant difference at the 95%confidence level (paired t-test), andpresented relative standard devia-tions (%RSD) from 0.9 to 1.9% (K)and 1.5 to 2.1% (Na). It is importantto point out that 30 seconds of son-ication is a very short and satisfac-

tory time to extract K and Na fromoilseed samples.

By using 15 mL of 0.12 mol L-1

HCl solution, 0.84 mm particle size,0.1000 g sample mass, and 30 sec-onds sonication at 25 °C, betterconditions for extracting K and Nafrom the selected samples wereachieved. These values wereadopted to validate the proposedmethod and for analysis of the sam-ples.

Analysis of Oilseed Samples

Five different oilseed samples ofthe genus Crambe abyssinicaHochst, Guizotia abyssinica,Brassica napus var. oleífera,Carthamus tinctorius L., andRaphanus sativus L. var. oleiferusMetzg were submitted to the pro-posed extraction procedure andthe final extracts were analyzed byFAES. The K and Na content, deter-mined in the final extraction solu-tions, showed concentrationintervals within 22.12±0.32 –30.20±1.12 mg L-1 and 5.64±0.20 –8.04±0.20 mg L-1, respectively. Ascan be seen, only the secondaryatomic line could be used for the

determination of K and Na in theconcentration range describedabove in a single run without needof further dilution of the samples.The analytical results, expressed asaverage values ± standard deviation(n= 3) on a dry matter basis, for thedetermination of K and Na in fiverenewable oilseed samples arelisted in Table V. The analysis of thesamples revealed that the concen-tration ranges obtained using theestablished ultrasound-assistedextraction procedure were in the5.53 ± 0.08 – 7.55 ± 0.28 mg g-1 Kand in the 1.41 ± 0.05 – 2.01 ± 0.10mg g-1 Na intervals. To ensure theaccuracy of the developed method-ology by US, the strategies based onthe addition/recovery test (using alloilseed samples) and the analysis ofPIATV 02/2010 were performed.Recoveries of 97.4 ± 2.1 – 98.8 ±3.5% K and 93.7 ± 3.3 – 99.1 ±2.5% Na were obtained by addingan amount of K and Na as the inor-ganic standard at the beginning ofthe sample pretreatment procedurein order to achieve 25.0 and 50.0mg L-1 K and Na in the finalextracts.

TABLE VComparative Results (Mean ± Standard Deviation) of K and Na

(mg g-1) of the Five Renewable Oilseed Samples by FAESUsing the Proposed Ultrasound-Assisted Extraction Procedure

and Microwave-Assisted Acid Digestion

Oilseed Crops Sample PreparationUltrasound-assisted Microwave-assisted

Extraction Digestion

K Na K Na

Crambeabyssinica Hochst 7.55 ± 0.28 1.41 ± 0.05 7.73 ± 0.18 1.49 ± 0.07

Guizotiaabyssinica 5.53 ± 0.14 1.84 ± 0.08 5.64 ± 0.08 1.93 ± 0.06

Raphanussativus L. var.oleiferus Metzg 5.90 ± 0.21 1.58 ± 0.04 6.10 ± 0.15 1.55 ± 0.05

Brassica napusvar. oleífera 6.07 ± 0.32 1.52 ± 0.03 6.47 ± 0.27 1.47 ± 0.04

Carthamustinctorius L. 6.62 ± 0.16 2.01 ± 0.10 6.94 ± 0.21 2.10 ± 0.09

74

16. R. Anderson, Sample pretreatmentand separation, London, ENG, UK(1987).

17. F.J. Holler, S.R. Crouch, and D.A.Skoog, Principles of InstrumentalAnalysis, Philadelphia, PA, USA(2009).

18. F.J. Krug, Métodos de preparo deamostras fundamentos sobrepreparo de amostras orgânicas einorgânicas para análise elementar.Piracicaba, SP, Brazil (2008).

19. A.I. Vogel, and G.H. Jeffery, Vogel’stextbook of quantitative chemicalanalysis. New York, NY, USA(1989).

20. M. Hoenig, Talanta 54, 1021 (2001).

21. G. Knapp, TrAC-Trend. Anal. Chem3, 182 (1984).

22. E. de Oliveira, J. Brazil Chem. Soc.14, 174 (2003).

23. C.C. Nascentes, M. Korn, C.S.Sousa, and M.A.Z. Arruda, J. Brazil.Chem. Soc. 12, 57 (2001b).

24. N. Manutsewee, W. Aeungmaitre-pirom, P. Varanusupakul, and A.Imyim, Food Chem. 101, 817(2007).

25. F. Priego-Capote, and M.D. LuqueDe Castro, TrAC-Trend Anal.Chem. 23, 644 (2004).

26. F. Priego-Capote, and M.D. LuqueDe Castro, J. Biochem. Bioph.Meth. 70, 299 (2007).

27. L.A. Currie, Anal. Chim. Acta 391,105 (1999).

28. M.E. Farago, Plants and the chemi-cal elements: biochemistry,uptake, tolerance and toxicity.Weinheim, Germany (1994).

29. Agilent Technologies, Flame AtomicAbsorption spectrometry – Analyti-cal Methods, Mulgrave, VIC, Aus-tralia (2015).

30. J. Deng, X. Feng, and X.Qiu, Chem.Eng. J. 152, 177 (2009).

31. V.C.D. Peronico, and J.L. Raposo Jr,Food Chem. 196, 1287 (2016).

32. T.G. Kazi, M.K. Jamali, M.B. Arain,H.I. Afridi, N. Jalbani, R.A. Sarfraz,and R. Ansari, J. Hazard. Mater.161, 1391 (2009).

33. A.S.N. Trindade, A.F. Dantas, D.C.Lima, S.L.C. Ferreira, and L.S.G.

(CNPq, Process n. 479186/2013-8)for their financial support of thiswork.

Received October 24, 2016.

REFERENCES

1. Conab, http://www.conab.gov.br/conteudos.php?a=1253&t, Lastvisited: 10/10/2016.

2. M. Gierth and P. Mäser, FEBS Lett.581, 2348 (2007).

3. Z. Jianbin, H. Xiaoyan, Q. Xiaoyan,C. Shengguan, H. Yong, A.N.Umme,and Z. Guoping, JProteomics 126, 1 (2015).

4. R.A. Leigh and R.G. Wyn Jones, NewPhytol. 97, 1 (1984).

5. H. Marschner, Marschner’s MineralNutrition of Higher Plants, Acade-mic Press, Amsterdam, NED(2011).

6. W.T. Pettigrew, Physiol. Plantarum133, 670 (2008).

7. F. Yang, M. Du, X. Tian, A.E. Eneji, L.Duan, and Z. Li, Field Crop Res.163, 109 (2014).

8. M.A. Bechlin, F.M. Fortunato, R.M.Silva, E.C. Ferreira, and J.A. GomesNeto, Spectrochim. Acta B 101,240 (2014).

9. A.I. Barros, A.P. Oliveira, M.R.L. Mag-alhães, and R.D. Villa, Fuel 93, 381(2012).

10. S.E. Dancsak, S.G. Silva, J.A.Nóbrega, B.T. Jones, and G.LDonati, Anal. Chim. Acta 806, 85(2014).

11. A. Jesus, M.M. Silva, and M.G.R.Vale, Talanta 74, 1378 (2008).

12. A.P. Oliveira, R.D. Villa, K.C.P.Antunes, A. Magalhães, and E.C.Silva, Fuel 88, 764 (2009).

13. C.V.S. Ieggli, D. Bohrer, P.C. Nasci-mento, and L.M. Carvalho, FoodChem. 124, 1189 (2011).

14. S.R. Oliveira, J.L. Raposo Jr, and J.A.Gomes Neto, Spectrochim. Acta B64, 593 (2009).

15. S.R. Oliveira, J.A. Gomes Neto, J.A.Nóbrega, and B.T. Jones, Spectro-chim. Acta B 65, 316 (2010).

Finally, the results obtained forthe samples using the US methodwere compared with thoseobtained using total decompositionin a closed-vessel microwave oven.The results obtained (Table V)show no significant difference atthe 95% confidence level (pairedt-test) between the values obtainedfor the proposed methodology andacid digestion. These results indi-cated that the ultrasound extractionprocedure proposed in this studyis free of matrix interferences. Therelative standard deviation (n=12)for all measurements varied withinthe range of 1.8–5.3% for ultra-sound-assisted extraction and1.1–4.7% for microwave-assisteddigestion.

CONCLUSION

Use of a less sensitive atomic linein FAES for K and Na was feasiblefor the elemental determination ina wide range of concentrationswithout need of further dilutions ofthe sample extracts, and allowed toreduce the errors associated withexcessive sample handling. Themethodology described offers afast, accurate, and efficient samplepreparation for the determinationof K and Na in different oilseedcrop samples by flame atomic emis-sion spectrometry. In general, theproposed sample preparationmakes the ultrasonic bath systemadvantageous as compared to aciddigestion due to low time andreagent consumption, minimal gen-eration of wastes, and thereforecontributes to a green chemistryprocedure.

ACKNOWLEDGMENT

The authors would like to thankthe Fundação de Apoio ao Desen-volvimento do Ensino, Ciência eTecnologia do Estado de MatoGrosso do Sul (Fundect, Process23/200.665/2012), and theConselho Nacional de Desenvolvi-mento Científico e Tecnológico

75


Teixeira, Food Chem. 185, 145(2015).

34. B.L. Batista, J.L. Rodrigues, V.Oliveira, and F. Barbosa, ForensicSci. Int. 192, 88 (2009).

35. M. Costas, I. Lavilla, S. Gil, F. Pena,I. De La Calle, N. Cabaleiro, and C.Bendicho, Anal. Chim. Acta 679,49 (2010).

36. I. De La Calle, N. Cabaleiro, I. Lav-illa, and C. Bendicho, Spectrochim.Acta B 64, 874 (2009).

37. C.E.R. Paula, L.F.S. Caldas, D.M.Brum, and R.J. Cassella, J. Pharma-ceut. Biomed. 74, 284 (2013).

38. J.L. Capelo, C. Maduro, and C.Vilhena, Ultrason. Sonochem. 12,225 (2005).

39. A.V. Filgueiras, J.L. Capelo, I. Lav-illa, and C.Bendicho, Talanta 53,433 (2000).

40. J. Liao, B. Qu, D. Liu, and N. Zheng,Ultrason. Sonochem. 27, 110(2015).

41. C.C. Nascentes, M. Korn, andM.A.Z. Arruda, Microchem. J. 69,37 (2001a).

42. I. Lavilla, B. Perez-Cid, and C. Bendi-cho, Int J Environ Anal. Chem. 72,47 (1998).

43. H. Minami, T. Honjyo, and I. Atsuya,Spectrochim. Acta B 51, 211(1996).

44. S. Ohmori, J. Radioanal. Nucl.Chem. 84, 451 (1984).

Copy

right

©20

17Pe

rkin

Elm

er,I

nc.4

0036

0_03

Allr

ight

sre

serv

ed.P

erki

nElm

er®

isa

regi

ster

edtra

dem

ark

ofPe

rkin

Elm

er,I

nc.A

llot

hert

rade

mar

ksar

eth

epr

oper

tyof

thei

rres

pect

ive

owne

rs.

For more information, visit perkinelmer.com/NexION2000

Trace metals in food, nanomaterials in water, impurities in everything from pills to electronic components: ese are the sweet spot for the NexION® 2000 ICP-MS. Its sample introduction technology lets you run samples with up to 35% total dissolved solids. Plus, its interference removal capabilities give you the best detection limits for your

NexION 2000 ICP-MS: Triple quad power meets single quad versatility.application. And it delivers superior analysis times and single particle/cell detection capability – at least 10x faster than competitive systems. So the NexION 2000 ICP-MS is up to the most important challenge of all: Yours.

NexION 2000 ICP Mass Spectrometer

ANY MATRIXANY INTERFERENCEANY PARTICLE SIZE

AMAZINGLYCAPABLEREMARKABLYAFFORDABLE

Copy

right

©20

16Pe

rkin

Elm

er,I

nc.4

0035

3_06

Allr

ight

sre

serv

ed.P

erki

nElm

er®

isa

regi

ster

edtra

dem

ark

ofPe

rkin

Elm

er,I

nc.A

llot

hert

rade

mar

ksar

eth

epr

oper

tyof

thei

rres

pect

ive

owne

rs.

For more information, visit perkinelmer.com/avio200

It’s a very big breakthrough in a very compact, a ordable package: e all-new Avio™ 200 ICP-OES is capable of handling the most di cult high-matrix solutions without dilution, for a whole new level of performance and exibility. Its ease-of-use features and intuitive so ware make multi-

element analysis as easy as single-element. And it delivers a whole host of “ rsts” and “bests” – lowest argon consumption, fastest startup time (10 minutes from power on), superior sensitivity and resolution, and the widest linear range – in an amazingly small footprint. e Avio 200 system: It’s the complete – and a ordable – package.

Avio 200 ICP-OES

With the new Titan MPS™ sample preparation system, we’re delivering on the promise ofa simple, safe, cost-effective microwave. But what really sets it apart is everything else youget with it: It’s backed by PerkinElmer, the go-to resource and one-stop shop for all thingsatomic spectroscopy, so you can access the instruments, consumables, informatics, andexpertise you need to optimize your research and your business. The Titan MPS samplepreparation system: It’s the microwave that’s so much more.

www.perkinelmer.com/titanmps

©20

13Pe

rkin

Elm

er,I

nc.4

0027

8_01

.All

trade

mar

ksor

regi

ster

edtra

dem

arks

are

the

prop

erty

ofPe

rkin

Elm

er,I

nc.a

nd/o

rits

subs

idia

ries.

IT'S NOT JUSTTHE MICROWAVEIT'S EVERYTHINGBEHIND IT

013351_03

PerkinElmer, Inc.710 Bridgeport AvenueShelton, CT 06484-4794 USAPhone: (800) 762-4000 or(+1) 203-925-4602www.perkinelmer.com

Documents

AS 2 8( 6) 195