AS 2 8(1) 1-29 Spectroscopy 28(1).pdf · intestinal digestion and absorption. These analyses were made by ICP-MS and were compared to those by hydride generation atomic fluores-cence

Issues also available

electronically.

(see inside front cover)

ASPND7 28(1) 1–40 (2007)ISSN 0195-5373

AtomicSpectroscopy

January/February 2007 Volume 28, No. 1

In This Issue:Determination of As, Bi, Cd, Co, Cr, Ga, In, Mn, Ni, Pb, Sb, Se, Sn, Te, Tl, and V in Antihypertensive Drugs by Inductively Coupled Plasma Mass SpectrometryJosianne Nicácio Silveira, Paulo Celso Pereira Lara, Michelle Batista Dias, Judith Maria Gomes Matos, Júlio César José da Silva, Clésia Cristina Nascentes, Virgínia Sampaio Teixeira Ciminelli, José Bento Borba da Silva ............................... 1

Determination of Inorganic Constituents in Hemodialysis Water Samples Using Inductively Coupled Plasma Optical Emission Spectrometry With Axially and Radially Viewed ConfigurationsRoberta Eliane dos Santos Froes, Nilton de Oliveira Couto e Silva, Rita Lopes P. Naveira, Júlio César José da Silva, Virgínia Sampaio Teixeira Ciminelli, Cláudia Carvalhinho Windmöller, José Bento Borba da Silva .................................... 8

Comparison of Metallic and Ceramic Tubes as Atomization Cells for Tin Determination by TS-FF-AASFabiana Aparecida Lobo, Ana Cristina Villafranca, Adriana Paiva de Oliveira, and Mercedes de Moraes ............................................................................................... 17

Diaminobenzidine-coated Silica Gel Used for Solid Phase Extraction Preconcentration Separation of Cu(II) and Cd(II) in Soil Samples Prior to Flame AAS DeterminationP. Sreevani, P. Mamatha, C. Sivani, P. Gopi Krishna, and G.R.K. Naidu .................... 24

Determination of Trace Amounts of Silver in Various Samples by Electrothermal Atomic Absorption Spectrometry After Sample Preparation Using Cloud Point ExtractionZhefeng Fan and Fang Bai ............................................................................................. 30

Flow Injection On-line Solid Phase Extraction Using Multi-Walled Carbon Nanotubes as Sorbent for Cold Vapor Atomic Fluorescence Spectrometric Determination of Trace Mercury in Water SamplesXiao-Hong Shang ............................................................................................................ 35

PerkinElmer is a registered trademark, and DRC, Dynamic Reaction Cell, AAnalyst,and Optima are trademarks of PerkinElmer, Inc..SCIEX and ELAN are registered trademarks of MDS SCIEX, a division of MDS Inc.IBM is a registered trademark of International Business Machine Corp.Inconel is a registered trademark of Inco Alloys International, Inc.Milli-Q is a trademark of Millipore Corporation.Omegatite is a registered trademark of Mega Engineering, Inc.Suprapur is a registered trademark of Merck & Co.Titrisol is a registered trademark of DuPont Merck Pharmaceuticals Co.Tygon is a trademark of Norton Co.Visual Basic is a registered trademark of Microsoft Corporation.Registered names and trademarks used in this publication, even without specificindication thereof, are not to be considered unprotected by law.

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

PerkinElmer Life and Analytical Sciences710 Bridgeport AvenueShelton, CT 06484-4794 USAhttp://www.perkinelmer.com

AtomicSpectroscopy

Vol. 28(1) Jan./Feb. 2007

Guidelines for AuthorsAtomic Spectroscopy serves as a medium for the disseminationof general information togetherwith new applications and analytical 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 1500subscribers 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 manuscriptcan 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 the

order they are cited in the text.5. Consult a current copy of

Atomic Spectroscopy for format.

6. Editor’s e-mail:[email protected]

All manuscripts are sent to tworeviewers. If there is disagreement,a third reviewer is consulted.

Minor changes in style are madein-house and submitted to theauthor for approval.

If a revision of the manuscriptis required before publication canbe considered, the paper isreturned to the author(s) withthe reviewers’ comments.

In the interest of speed ofpublication, 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 bepurchased, 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 SpectroscopyPerkinElmerLife and Analytical Sciences710 Bridgeport AvenueShelton, CT 06484-4794 USA

EDITORAnneliese LustE-mail: [email protected]

TECHNICAL EDITORSLaura J. Thompson, AADennis Yates, ICPKenneth R. Neubauer, ICP-MS

SUBSCRIPTION INFORMATIONAtomic SpectroscopyP.O. Box 3674Barrington, IL 60011 USAFax: +1 (847) 304-6865E-mail: [email protected]

2006 Subscription Rates• U.S. $60.00 includes third-class

mail delivery worldwide; $20.00 extra for electronic file.

• U.S. $80.00 for airmail $20 extra for electronic file.

• U.S. $60.00 for electronic file only.• Payment by check (drawn on U.S.

bank in U.S. funds) made out to: “Atomic Spectroscopy”

Electronic FileSend request via e-mail to:[email protected]

Back Issues/Claims• Single back issues are available

at $15.00 each.• Subscriber claims for missing

back issues will be honoredat no charge within 90 daysof issue mailing date.

Address Changes to:Atomic SpectroscopyP.O. Box 3674Barrington, IL 60011 USA

Copyright © 2007PerkinElmer, Inc.All rights reserved.htpp//:www.perkinelmer.com

MicrofilmAtomic Spectroscopy issues are available from:University Microfilms International300 N. Zeeb RoadAnn Arbor, MI 48106 USATel: (800) 521-0600 (within the U.S.)+1 (313) 761-4700 (internationally)

1Atomic SpectroscopyVol. 28(1), Jan./Feb. 2007

Determination of As, Bi, Cd, Co, Cr, Ga, In, Mn, Ni, Pb,Sb, Se, Sn, Te, Tl, and V in Antihypertensive Drugs by

Inductively Coupled Plasma Mass SpectrometryJosianne Nicácio Silveiraa, Paulo Celso Pereira Larab, Michelle Batista Diasa, Judith Maria Gomes Matosc,

Júlio César José da Silvac, Clésia Cristina Nascentesb, Virgínia Sampaio Teixeira Ciminellic,

*José Bento Borba da Silvab

a Laboratório de Toxicologia Ocupacional, Faculdade de Farmácia, Universidade Federal de Minas Gerais b Departamento de Química, Universidade Federal de Minas Gerais,

Belo Horizonte, 31270-901, Belo Horizonte, MG. Brazil c Departamento de Engenharia Metalúrgica e Materiais, Universidade Federal de Minas Gerais

INTRODUCTION

Arterial hypertension consists ofblood pressure increasing to levelsabove the usual (140/90 mm Hg).This increase is due to narrowing ofthe arteries and occurs when theamount of blood to be pumpedthrough the body remains constantbut the arteries are narrowed. Highblood pressure can also occurwhen the body retains large quanti-ties of liquid which must flowthrough the arteries. Abnormallyhigh blood pressure can damageorgans, such as the brain, kidneys,and eyes (1).

Hypertensive people take med-ication daily which introduces met-als into the body, which is

Pharmacopoeia (JP), and EuropeanPharmacopoeia (EP). This methodallows specific detection and quan-tification for each of the elementsexpected to give rise to a positiveresponse in the compendial meth-ods: Ag, As, Bi, Cd, Hg, In, Mo, Pb,Pd, Pt, Ru, Sb, Se, Sn, and Ti. Soltyket al. (4) determined Ca, Cr, Cu,Fe, Mg, Mn, Mo, P, Se, and Zn inmultimineral and multivitaminpreparations and in pharmaceuticalraw material by ICP-MS and ETAAS.Murty et al. (5) determined tracemetals (Ti, Cr, Mn, Fe, Co, Ni, Cu,Zn, Cd, Hg, and Pb) in dicyclomine-HCl, ethambutol, pyrazinamide, andfurazolidone drugs using ICP-MS.Dolan et al. (6) determined As, Cd,Hg, and Pb in 95 dietary supplementproducts after microwave digestionand high-resolution inductivelycoupled plasma mass spectrometry.Wu et al. (7) investigated theamounts and characteristics of

*Corresponding author.E-mail: [email protected] Tel +55-331-34995750Fax: +55 31 34995700

particularly dangerous for thosewith a tendency to accumulate met-als. Along with other possible dailycontamination sources, the metalscontained in medicines can causeseveral health problems. Inductivelycoupled plasma mass spectrometry(ICP-MS) is a very sensitive and mul-tielement technique which is effi-cient in determining metal contentin different samples such as medi-cines.

Lozak et al. (2) determinedchromium, selenium, and molybde-num in a therapeutic diet by ICP-MSand ETAAS (electrothermal atomicabsorption spectrometry) after thesample was digested in a high-pres-sure microwave oven or by ashing.Lewen et al. (3) proposed a robustgeneral ICP-MS-based method as analternative to the wet chemicalheavy metals test prescribed in theUnited States Pharmacopoeia (USP),British Pharmacopoeia (BP), Japanese

ABSTRACT

A rapid procedure for thedetermination of 75As, 114Cd,59Co, 55Mn, 58Ni, 51V, 120Sn, 123Sb,209Bi, 69Ga, 115In, 78Se, 130Te, 205Tl,208Pb, and 52Cr in antihyperten-sive drugs by inductively coupledplasma mass spectrometry (ICP-MS) is proposed. For thesedeterminations the samples wereprepared by a simple, fast extrac-tion method with concentratednitric acid. The optimization ofplasma conditions, such asradiofrequency (RF) and nebu-lization gas flow rate, accordingto the manufacturer, was com-

pared to a multivariate design(MD) optimization. The valuesobtained for the slopes and thecorrelation coefficients in the MDfor all studied isotopes showedthat the central point (1000 Wand 1.0 L min–1) was the bestcondition and very close to theones obtained in the optimizationrecommended by the manufac-turer; therefore the followingstudies were made using the lat-ter optimization. Calibration wasperformed by matrix matchingwith an r2 higher than 0.99 for allisotopes investigated. The limitsof detection and quantificationranged from 0.01 to 1.41 µg L–1

for 205Tl and 209Bi, and from 0.03and 4.69 µg L–1 for 78Se. Intra-assay studies (n=7 for each level)on the samples spiked with 0.1,5.0, and 20.0 µg L–1 of the iso-topes of interest showed a coeffi-cient of variation ranging from0.3% to 10.9%. Recovery studieson the drug spiked with four lev-els (0.1, 2.0, 5.0, and 20.0 µg L–1;n=3 for each level) showedrecoveries between 96.5 ± 1.6%and 115.5 ± 7.5%. Seventy anti-hypertensive drug samples wereanalyzed and the results rangefrom 0.03 to 4087.6 ng per cap-sule or tablet.

2

toxic metals (As, Hg) in several Chi-nese medicinal materials under con-ditions simulating stomach andintestinal digestion and absorption.These analyses were made by ICP-MSand were compared to those byhydride generation atomic fluores-cence spectrometry (HG-AFS).Focused microwave-assisted extrac-tion and Soxhlet extraction was car-ried out and compared with theconventional sequential extractionmethod. Wang et al. (8) developeda fast and sensitive method whichwas validated for the determinationof tungsten in bulk drug substancesand intermediates using eitherinductively coupled plasma opticalemission spectrometry (ICP-OES)or ICP-MS. Sample preparation wasperformed by direct dissolutionwith a 80:20 (v/v) concentratednitric acid: deionized water mixturewhich avoids labor-intensive andpotentially hazardous digestiontechniques.

In this work, the levels of someelemental species in antihyperten-sive drugs were determined byICP-MS after a simple acid extrac-tion procedure and selecting a suit-able isotope. A factorial design wasemployed to optimize the ICP con-ditions; those conditions were com-pared with the recommendationsof the manufacturer. With the uni-variate optimization, the interac-tions between parameters are notevaluated, so that the origin ofsome matrix effects cannot beaddressed. Multivariate optimiza-tion seems to be more adequatewhen many variables are involved.The factorial design (9,10) is a goodand simple statistical tool that canbe used to verify the effects of vari-ables and their interactions usingonly a few experiments.

EXPERIMENTAL

Instrumentation

A Model ELAN 9000 quadrupoleinductively coupled plasma massspectrometer was used for all mea-

surements (PerkinElmer Sciex,Concord, Toronto, Canada). Theinstrument was equipped with across-flow nebulizer, a double pathspray chamber (Scott-type), and analumina injector with a 2.00-mminternal diameter (i.d.). Also usedwas 99.98% purity argon (Air Liquide,Contagem, Minas Gerais, Brazil),a Model MS1 tube stirrer (IKA,Germany) and an Excelsa Baby Icentrifuge (Fanem, Brazil).

Reagents and Solutions

All solutions were preparedusing deionized water from aMilli-Q™ system (Millipore, Bed-ford, MA, USA). Suprapur® nitricacid of 65% (v/v) (Merck,Darmstadt, Germany, Part No.1.00441.0250) was used. The fol-lowing solutions were used: Pbstock solution 1000 ± 0.002 mg L–1

(Tritisol®, Merck, Part No. 1.09969),and two multielementary NISTtraceable ICP-MS solutions (GFSChemical Inc., Columbus, OH,USA) 100 ± 1 mg L-1; I: Cd, Co, Cr,Cu, Fe, Mn, Ni, Ag, V, Zn (GFS PartNo. 1842) and II: Sb, As, Bi, Ga, In,Se, Te, Tl, Th, Sn,U (GFS Part No.1845), all in 5% (v/v) HNO3. For allstudies, the final solutions weremade to 4.0% (v/v) HNO3.

To clean and avoid analyteadsorption onto the surface of thecapillary, a 1% HNO3 (v/v) washsolution (Part No. N8122038 -PerkinElmer® Pure) was used.

All glassware was thoroughlywashed with Milli-Q water beforeuse and kept in a 20% (v/v) nitricacid bath for more than 24 hours.Before use, the glassware wasrinsed several times with Milli-Qwater.

Procedure

Sampling and Sample TreatmentSamples were obtained from 20

drugstores, while other industrialsamples were obtained from a localdrugstore. Two kinds of antihyper-tensive drugs were selected from

those most consumed by patientsin Belo Horizonte, MG, Brazil. Theactive pharmaceutical ingredientsof these two sample types wereenalapril maleate (EM) andamlodipine besilate (AB). The sam-ples consisted of 59 capsules: 36 EM,21 AB, 1 lozartan, and 1 atenolol.Eleven tablets were analyzed: 8 ABand 3 EM. Since different drugstoresuse different excipients, we had acapsule produced (reference sam-ple A) containing all the excipientsand the two active pharmaceuticalingredients as used and sold in thedrugstores that participated in thisproject.

Sample treatment is very simpleand fast. The capsules and tabletswere digested in a tube (withthreaded cover) in 1000 µL nitricacid, followed by one minute ofstirring. Then 9.0 mL of Milli-Qwater was added and again stirredfor one minute. This mixture wascentrifuged at 3000 rpm for 10 min-utes. The supernatant was trans-ferred into a 25-mL volumetric flaskand brought to volume with Milli-Qwater.

CalibrationFor preparing the matrix match-

ing curves, at each dilution pointone capsule (sample A) was trans-ferred to a 15-mL plastic tube,1000 µL HNO3 was added, and themixture stirred for one minute.Different volumes of the multiele-ment stock solutions (10.0 mg L–1,100.0 µg L–1, and 10.0 µg L–1) in 4%nitric acid were added to the tubesto obtain the matrix matching cali-bration standard solutions in con-centrations of 0.1; 1.0; 2.0; 5.0;10.0; 20.0, and 30.0 µg L–1. Thestandards were brought to 9-mLvolume with water. The solutionswere stirred with a tube stirrer forone minute and centrifuged at3000 rpm for 10 minutes. Thesupernatants were transferred into25-mL volumetric flasks whichwere brought to volume withwater. Concentration "zero" was

3

Vol. 28(1), Jan./Feb. 2007

obtained in the same way, but with-out addition of standard solution.This point was denominated assample blank A. The reagent blankwas prepared with 1000 µL Milli-Qwater and 1000 µL HNO3 followingthe same procedures. This blankwas used to eliminate the effects ofpossible metal contents in thewater, nitric acid, and flasks duringcalculation. Seven matrix matchingcalibration curves were made, andthe average of these was used toobtain the figures of merit.

Optimization of InstrumentalParameters

Reference sample A, spiked with0.1; 1.0; 2.0; 5.0; 10.0; 20.0, and30.0 µg L–1 of the analytes, wasused to obtain the most appropriateoperating conditions for the multi-variate method (r2 and slopes).

Table I shows the optimumoperating conditions obtainedaccording to the manufacturer’srecommendations.

Initially, several isotopes of eachelement were evaluated and,according to the r2 and sensitivities(slopes) obtained, the most appro-priate isotopes were selected.

Multivariate Optimization

Optimization of the multivariatemethod was accomplished by thesimultaneous variation of two para-meters: radio-frequency power andnebulization gas flow rates. Theradio frequency powers used were:500 W as the lowest level, 1500 Was the highest level, and 1000 W asa central point. The nebulizationgas flow rates analyzed were:0.5 L min–1 as the lowest level,1.5 L min–1 as the highest level, and1.0 L min–1 as a central point.Power and gas flow variations werecombined following the factorialplanning (22 + 1 experiments) pre-sented in Table II. The average ofthe counts per second (cps, n=3) ofeach element were analyzed withStatgraphics Plus (Statistical Graph-

ics Corp. Rockville, MD, USA)which supplied the best operatingconditions.

Figures of Merit

The limit of detection (LOD) andthe limit of quantification (LOQ)were calculated using the equationsLOD=3 x S0.1 and LOQ=10 x S0.1,where S0.1 is the standard deviationfor 10 measurements of the firstpoint of the matrix matching curve.

Recovery studies for the studiedelements in the drugs were alsomade in order to check for accu-racy.

For the recovery study, sample Awas spiked with 0.1, 2.0, 5.0, and20.0 µg L–1 of As, Cd, Co, Mn, Ni, V,Sn, Sb, Bi, Ga, In, Se, Te, Tl, Pb, andCr. Three samples containing eachof the concentrations were preparedand measured in triplicate.

For recovery studies of the cap-sules, references (Sample A) wereadded with the multielement stan-dard in four different concentrations:0.1, 2.0, 5.0, and 20.0 µg L–1 andmade in triplicate. Since the sam-ples were centrifuged, recoverywas then accomplished, allowingfor comparison of the metals beforeand after centrifugation.

Sample A capsules, spiked withthe metals before and after centrifu-gation, were analyzed by ICP-MS,and the recovery percentage calcu-lated considering the addition aftercentrifugation as 100%. The recov-ery was calculated as follows:

Ca × 100Recovery (%) =

Cb

where Ca = concentration of the metals(µg L–1) in the Sample A capsulesadded before centrifugation; and Cb = concentration of the metals(µg L–1) in the Sample A capsulesadded after the centrifugation.

The objective of this study wasto evaluate possible losses duringcentrifugation since only the super-natant was analyzed.

Precision was evaluated by intra-assay studies. The intra-assay coeffi-cient of variation was calculatedusing the standard deviationobtained for each concentration(0.1, 5.0, and 20.0 mg L–1) dividedby the average and multiplied by100. Seven samples (A) were spikedwith each of the concentrationsand the concentrations in eachwere measured in triplicate.

TABLE IICP-MS Operating Conditions

and Parameters for Data Acquisition

RF (40 MHz), Power (W)Forward 1000 WReflected < 5 W

Gas flows Plasma 15 L min–1

Auxiliary 1.2 L min–1

Nebulizer 0.94 L min–1

Measurements Peak HoppingSweeps 4Reading/Replicate 3Replicates 3Dwell Time 50 ms

Integration Time 600 ms

TABLE II Multivariate Optimization

Conditions

Step RF Power Neb. (W) (L min–1)a

1 500 0.52 500 1.53 1500 0.54 1500 1.5

5 1000 1.0

a Nebulization gas flow rate.

4

matrix matching calibration. TableVIII presents the results obtained innanograms per capsule or tabletand in micrograms per liter,directly read by the ICP-MS. Thehigher concentration values (above40 µg L–1) were obtained for Cr andSe in the enalapril maleate capsulesand for Mn in the amlodipine besi-late tablets.

Considering the sum of the high-est concentrations of each analyzedelement, the result would be about10 µg per capsule or tablet. Assum-ing that the gastrointestinal absorp-tion is around 10% for most of theelements and that the time of bio-logical half-life varies from hours todays (15), we believe that a patientingesting one capsule per daywould not be at risk. For this kindof medicine, the concentrations ofthe elements were low; however,this does not invalidate the method,since it can be applied to othermedicines taken on a daily basis.

CONCLUSION

The experimental conditionsdetermined in this study and basedon the manufacturer’s recommen-dations were similar to thoseobtained by the multivariatemethod. Under these conditionsand with a simple and fast samplepreparation, it was possible todetermine the following elements:As, Bi, Cd, Co, Cr, Ga, In, Mn, Ni,Pb, Sb, Se, Sn, Te, Tl, and V in anti-hypertensive drugs. The resultsobtained with the figures of meritstudied indicate that the validatedmethod presents adequate sensitiv-ity, precision, and accuracy. The 70samples analyzed showed metal lev-els ranging from 0.03 to 4087.6 ngper capsule or tablet for Cd, Te andTl, and Mn, respectively. Eventhough the lowest levels for Cr andMn were below the LOQ, all of themetals studied were above theLOD.

RESULTS AND DISCUSSION

Establishment of OperatingParameters by the MultivariateMethod

The Pareto and main effectgraphs (not shown) provided byStatgraphics Plus indicate that thevariables of radio frequency powerand nebulization gas flow rate didnot contribute significantly to theoptimization of the ICP-MS analyit-cal parameters. On the other hand,the data analyzed by the software(Table III) clearly indicate, throughthe correlation coefficients and slopes,that the central point (1000 W and1.0 L min–1) is best for all isotopes.Exceptions were observed for 55Mnand 78Se, which presented r2 <0.99.However, in subsequent studies(Table IV), the r2 proved to beappropriate. In all results, the mostsuitable isotopes were the mostabundant ones, except for 123Sb and78Se. Since the analysis conditionsobtained by the multivariatemethod are very similar to the onesdetermined by the recommenda-tions of the manufacturer (1000 Wand 0.94 L min–1), we decided touse the latter (Table I).

Analytical Figures of Merit

The seven matrix matching cali-bration curves made in the range of0.1 to 30.0 µg L–1 gave an r2 greaterthan 0.99 for all of the analyzed iso-topes (see Table IV). The limits ofdetection ranged between 0.008and 1.41 µg L–1 for 205Tl and 78Se,respectively, while the limits ofquantification ranged between0.025 and 4.69 µg L–1 for the sameisotopes. In spite of the detectionlimits, some elements (As, Co, Mn,Ni, Sb, and Se) were higher thanthe first point of the calibrationcurves. In all cases the obtainedr2 was greater than 0.99. Table Vshows the limits of detectionand quantification for all analyzedisotopes.

Table VI shows the recoverieswith their respective deviations(n=3) for the four spike levels insample A. The minimum and maxi-mum recovery values obtained forthe studied isotopes were 96.5±1.6%for 75As and 130.0±9.4% for 208Pb,with more than 95% of the valuesbetween 80% and 120% (11,12).For 55Mn (120.9 ± 2.9%) and 52Cr(123.7±5.6%), the recovery valuesare close to the upper limit (120%)and could be considered as an errorin the preparation of the 5.0 µg L–1

spike since the recoveries for allelements in this spike are the high-est. Polyatomic interferences at m/z55 and 52 (40Ar14NH+ and 40Ar12C+)make measurement of Mn and Crdifficult. Considering the low ana-lyte concentrations and the com-plex nature of the matrix, theserecoveries indicate, in our opinion,an adequate accuracy (13).

Table VII presents the intra-assaycoefficients of variation, whichexpress the precision of the pro-posed analytical method.

The highest intra-assay coefficientof variation obtained in the studywas 10.9% for the 51V isotope at20.0 µg L–1, and the smallest onewas 0.3% for the 78Se isotope at0.1 µg L–1. Considering the criteriaestablished by The InternationalAssociation of Official AnalyticalChemists (AOAC) (14), the CV canvary from 15% to 30%, dependingon the range of working concentra-tions (100 µg/g and 1 µg/g). There-fore, the values obtained are withinthe acceptability range.

Quantification of TraceElements in AntihypertensiveDrug Samples

The As, Bi, Cd, Co, Cr, Ga, In,Mn, Ni, Pb, Sb, Se, Sn, Te, Tl, and Vconcentrations in different antihy-pertensive drugs were determinedusing the experimental conditionsdetermined according to the manu-facturer’s recommendations for

5

Vol. 28(1), Jan./Feb. 2007

TABLE IIICorrelation Coefficients and Slopes of Analyzed Isotopes

in Relation to Radio Frequency Power and Gas Flow Rate in the Multivariate Analysis (n=3)

Element Statistics Correlation Coefficients and Slopes500 W/ 500 W/ 500 W/ 1500 W/ 1000 W/0.5 L min–1 1.5 L min–1 0.5 L min–1 1.5 L min–1 1.0 L min–1

75As r2 0.9977 0.1231 0.1500 0.0001 0.9990slope 22.4301 0.2656 –42.7626 –0.0051 28.6690

114Cd r2 0.9997 0.1918 0.9983 0.0001 0.9997slope 200.1242 0.3938 62.8958 –0.0051 398.6162

59Co r2 0.9993 0.2582 0.9784 0.0053 0.9998slope 104.8237 0.5476 34.0637 0.0353 302.84320

63Cu r2 0.9918 0.2257 0.1692 0.0372 0.9968slope 55.2749 0.3597 –1404.039 –0.0935 122.6405

55Mn r2 0.9975 0.3522 0.6263 0.0002 0.9890slope 112.1468 0.5711 37.6648 –0.0063 389.9742

58Ni r2 0.9988 0.4529 0.9913 0.0452 0.9976slope 48.8872 0.5413 20.0771 0.1236 136.5048

51V r2 0.9996 0.0846 0.2354 0.0562 0.9996slope 78.1404 0.2418 –49.2342 0.1113 225.0988

120Sn r2 0.9989 0.1239 0.9997 0.0001 0.9993slope 48.0121 0.2831 19.4501 –0.0051 41.1758

123Sb r2 0.9996 0.5756 0.9650 0.1758 0.9948slope 24.9955 0.4412 10.3552 –0.1761 28.4229

209Bi r2 0.9995 0.1707 0.9999 0.1170 0.9990slope 364.1457 0.3715 198.0309 0.2346 375.6453

69Ga r2 0.9997 0.0764 0.8789 0.0372 0.9995slope 110.4691 0.1773 30.5182 –0.0935 2255.6296

115In r2 0.9990 0.2277 0.9946 0.0124 0.9992slope 34.4472 0.4291 11.2957 0.0714 56.1146

78Se r2 0.9175 0.2141 0.0964 0.0570 0.9790slope 4.2939 0.3124 1.6692 0.0766 6.8034

130Te r2 0.9990 0.2257 0.9977 0.0759 0.9990slope 25.4780 0.2762 17.5932 0.0884 40.8760

205Tl r2 0.9995 0.2471 0.9996 0.0001 0.9988slope 326.4899 0.4822 159.9701 –0.0051 407.2394

232Th r2 0.9995 0.0715 0.9998 0.0986 0.9995slope 690.8632 0.2246 314.2775 0.1474 317.2684

208Pb r2 0.9990 0.0833 0.9994 0.1542 0.9989slope 481.3646 0.1948 240.3894 0.1649 537.7516

52Cr r2 0.9930 0.1046 0.9351 0.2606 0.9926slope 59.9393 0.3419 25.8112 0.2476 230.1680

64Zn r2 0.9748 0.2388 0.2073 0.0053 0.9566slope 24.3538 0.4062 –157.9325 0.0353 65.0918

107Ag r2 0.7400 0.0501 0.7013 0.0262 0.8651slope 1.0989 0.1535 0.7865 0.0760 2.1633

63Cu r2 0.9918 0.2257 0.1692 0.0372 0.9968slope 55.2749 0.3597 –1404.0391 –0.0935 122.6405

238U r2 0.9993 0.0846 0.9999 0.1067 0.9985slope 790.0627 0.2418 324.1126 0.2240 617.0257

56Fe r2 0.7053 0.6152 0.4500 0.0200 0.7205slope 375.4891 0.4786 31.2500 –0.2320 1232.6830

6

TABLE IVRelationship of Correlation Coefficient

and Calibration Equation of Analyzed Isotopes (n=7)

Isotopes Correlation Coefficient Calibration Equation75As 0.9953 Y = 618.71X + 1796.13114Cd 0.9996 Y = 2627.93X + 1765.8159Co 0.9995 Y = 1999.92X + 27.6555Mn 0.9992 Y = 1964.09X + 23716.6658Ni 0.9996 Y = 914.98X + 3683.2251V 0.9996 Y = 1242.78X + 2393.24120Sn 0.9999 Y = 380.01X + 1570.81209Bi 0.9995 Y = 3773.53X + 1897.3469Ga 0.9996 Y = 1593.45X + 2208.62115In 0.9999 Y = 439.00X + 1558.1078Se 0.9987 Y = 43.55X + 2248.80130Te 0.9997 Y = 330.00X + 1325.91205Tl 0.9994 Y = 3900.18X + 2061.12208Pb 0.9989 Y = 5442.74X + 6168.6552Cr 0.9977 Y = 1188.77X + 20010.65

TABLE VLimits of Detection and

Quantification for Analyzed Isotopes in Sample A

Isotope LOD LOQ (µg L–1) (µg L–1)

75As 0.15 0.49114Cd 0.02 0.0559Co 0.15 0.5255Mn 0.32 1.0658Ni 0.23 0.7851V 0.06 0.20120Sn 0.06 0.20123Sb 0.16 0.52209Bi 0.01 0.0369Ga 0.04 0.14115In 0.06 0.1978Se 1.41 4.69130Te 0.10 0.33205Tl 0.01 0.03208Pb 0.02 0.0752Cr 0.86 2.86

TABLE VIMetal Recoveries in Sample A Under Optimized Conditions (n=3)

Isotope Recovery (%)0.1 µg L–1 2.0 µg L–1 5.0 µg L–1 20.0 µg L–1

75As 102.5 ± 0.3 104.4 ± 0.9 96.5 ± 1.6 100.2 ± 2.0114Cd 103.1 ± 0.8 103.1 ± 0.5 111.8 ± 5.6 99.3 ± 2.559Co 104.8 ± 3.3 104.3 ± 1.4 111.8 ± 3.2 99.9 ± 2.955Mn 111.4 ± 3.7 115.5 ± 7.5 120.9 ± 2.9 101.4 ± 3.058Ni 114.5 ± 8.9 113.0 ± 4.2 108.4 ± 2.3 100.4 ± 3.251V 102.9 ± 0.1 102.8 ± 0.9 104.7 ± 3.5 101.4 ± 3.2120Sn 103.0 ± 0.8 102.7 ± 1.7 102.5 ± 1.8 99.4 ± 2.4123Sb 102.6 ± 1.1 103.5 ± 1.2 97.1 ± 2.7 98.4 ± 1.3209Bi 102.4 ± 1.4 104.4 ± 5.3 108.8 ± 6.1 97.6 ± 3.269Ga 104.7 ± 0.9 106.1 ± 0.5 111.0 ± 5.3 100.4 ± 3.2115In 102.8 ± 0.3 103.2 ± 1.3 100.6 ± 2.4 97.8 ± 3.178Se 100.6 ± 0.9 102.3 ± 0.5 103.2 ± 1.5 99.1 ± 0.7130Te 101.3 ± 1.1 102.3 ± 0.8 98.5 ± 1.8 98.5 ± 2.9205Tl 102.6 ± 1.0 104.0 ± 2.1 110.0 ± 3.5 98.4 ± 2.1208Pb 130 ± 9.4 112.4 ± 5.8 110.3 ± 5.7 98.7 ± 2.552Cr 102.6 ± 2.7 104.2 ± 5.7 123.7 ± 5.6 104.3 ± 3.6

TABLE VIICoefficients of Variation (CV)

Intra-assay (n=7)

Isotope CV Intra-assay (%)0.1 5.0 20.0 µg L–1 µg L–1 µg L–1

75As 0.5 0.9 1.4114Cd 0.5 3.0 1.559Co 0.5 0.3 0.555Mn 1.6 1.4 1.658Ni 0.5 0.4 0.751V 10.8 4.0 10.9120Sn 8.6 6.1 3.0123Sb 0.8 3.1 3.7209Bi 1.0 3.6 1.369Ga 0.7 1.7 0.6115In 1.1 4.2 1.978Se 0.3 0.7 1.0130Te 0.8 3.1 1.6205Tl 0.6 3.6 1.0208Pb 9.7 5.8 1.952Cr 0.9 1.3 2.0

7

Vol. 28(1), Jan./Feb. 2007

REFERENCES

1. A. Malta, Jr., and L.A.C. Araújo,Pharm. Bras. 16, 80 (2004).

2. A. Lozak, K. Soltyk, P. Ostapczuk,and Z. Fijalek, Pharmazie. 59, 824(2004).

3. N. Lewen, S. Mathew, M.Schenkberger, and T. Raglione, J. Pharm. Biomed. Anal. 35, 739(2004).

4. K. Soltyk, A. Lozak, P. Ostapczuk,and Z. Fijalek, J. Pharm. Biomed.Anal. 32, 425 (2003).

5. A.S.R.K. Murty, U.C. Kulshresta, T.N.Rao, and M.V.N.K. Talluri, Indian J.of Chem. Tecnol. 12, 229 (2005).

6. S.P. Dolan, D.A. Nortrup, P.M. Bol-ger, and S.G. Capar, J. of Agric. andFood Chem. 51, 1307 (2003).

7. X.H. Wu, D.H. Sun, Z.X. Zhuang,X.R. Wang, H.F. Gong, J.X. Hong,and F.S.C. Lee, Anal. Chim. Acta453, 311 (2002).

8. T. Wang, Z. Ge, J. Wu, B. Li, and A.Liang, J. Pharm. Biomed. Anal. 19,937 (1999).

9. C. Bendicho and M.T.C. Loos-Volle-bregt, J. Anal. At. Spectrom. 6, 353(1991).

10. W. Slavin, D.C. Manning, and G.R.Carnrick, At. Spectrosc. 2, 137(1981).

11. J.M. Green, Anal. Chem. 1, 305A(1996).

12. Environmental Protection Agency -EPA. Guidance for Methods Devel-opment and Methods Validationfor the RCRA Program SW-846Methods (1992).

13. T.D. Saint’ Pierre, L.F. Dias, D. Poze-bon, R.Q. Aucélio, A J. Curtius, andB. Welz, Spectrochim. Acta Part B57, 1991 (2002).

14. AOAC, Peer-verified Methods Pro-gram, Manual on Policies and Pro-cedures, Arlington, VA, USA(Nov. 1993).

15. F.A. Azevedo and A.A.M. Chasin,Metais: Gerenciamento da Toxici-dade, Atheneu, São Paulo, Brazil(2003).

ACKNOWLEDGMENTS

The authors are thankful toConselho Nacional de Pesquisa eDesenvolvimento Tecnológico(CNPq), Fundação de Amparo àPesquisa do Estado de Minas Gerais(FAPEMIG), and to the Laboratóriode Toxicologia Ocupacional daFaculdade de Farmácia (LATO/FAFAR/UFMG). P.C.P. Lara andJ.C.J. Silva have scholarships fromCNPq. V.S.T. Ciminelli and J.B.B.Silva are grateful to CNPq for theresearch grants.

Received December 18, 2006.

The highest concentration values(above 40 µg L–1) were obtained forCr, Se (in EM capsules) and Mn (inAB tablets). Since the levels foundin the analyzed samples are low (onthe order of nanograms per capsuleor tablet), there would not be anyhealth risk to the patient. However,this does not invalidate the methodsince it can be applied to the analy-sis of other medicines.

TABLE VIIIQuantification of Trace Elements

in Antihypertensive Drug Samples by ICP-MS (n=2)

Analyte Concentration Concentration LOQ(ng per capsule (µg L–1)b (µg L–1)

or tablet)a

75As 1.3 – 327.8 4.7 – 14.9 0.49114Cd 0.03 – 20.2 0.4 – 1.3 0.0559Co 0.2 – 46.6 0.6 – 2.7 0.5255Mn 2.1 – 4087.6 0.6 – 164.5 1.0658Ni 0.1 – 697.6 1.1 – 29.1 0.7851V 0.4 – 137.2 1.2 – 6.7 0.20120Sn 0.1 – 145.7 3.4 – 10.7 0.20123Sb 0.2 – 264.5 4.3 – 13.4 0.52209Bi 0.1 – 21.2 0.3 – 1.5 0.0369Ga 0.8 – 64.0 0.7 – 3.2 0.14115In 0.3 – 127.8 2.7 – 8.8 0.1978Se 0.4 – 2112.4 33.0 – 91.0 4.69130Te 0.03 – 181.8 2.3 – 8.6 0.33205Tl 0.03 – 15.8 0.3 – 1.0 0.03208Pb 0.2 – 55.7 0.2 – 2.3 0.0752Cr 9.9 – 1224.7 2.2 – 49.0 2.86

a Subtracted from the blank and corrected by capsule or tablet.b Direct reading in ICP-MS.


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

Determination of Inorganic Constituents inHemodialysis Water Samples Using Inductively Coupled

Plasma Optical Emission Spectrometry With Axially and Radially Viewed ConfigurationsRoberta Eliane dos Santos Froesa, Nilton de Oliveira Couto e Silvab, Rita Lopes P. Naveirab,

Júlio César José da Silvac, Virgínia Sampaio Teixeira Ciminellic, Cláudia Carvalhinho Windmöllera,

*José Bento Borba da Silvaa

a Departamento de Química, Universidade Federal de Minas Gerais, Av. Antônio Carlos 6627, Belo Horizonte, MG, Brazil

b Fundação Ezequiel Diaz, FUNED, Belo Horizonte, MG, Brazilc Departamento de Engenharia Metalúrgica e de Materiais,

Universidade Federal de Minas Gerais, Belo Horizonte, MG, Brazil

ABSTRACT

Inductively coupled plasmaoptical emission spectrometry(ICP-OES) was used to determinemacro- (Ca, K, Mg, and Na) andmicroelements (Al, Ba, Cd, Cr,Cu, Fe, Mn, Pb, and Zn) inhemodialysis water samples withboth axially and radially viewedconfigurations. The analyticalperformance of ICP-OES wasevaluated in terms of theMg(II)/Mg(I) ratio. The figures ofmerit such as background equiv-alent concentration (BEC), signal-to-background ratio (SBR), limitsof detection (LODs), and accu-racy were evaluated for eachanalyte. The influence of easilyionized elements (EIEs) and highcalcium concentrations on differ-ent analytical lines were verified.

Robust conditions wereobtained employing a nebuliza-tion gas flow rate of 0.6 L min–1

and an applied power of 1.3 kW.Under robust conditions, thepresence of 10 mg L–1 of Ca, K,Na, and a mixture of these ele-ments combined, showed signifi-cant interferences on the Al(396.15 nm) and Pb (220.35 nm)lines. The use of Y as an internalstandard reduced the effectsobserved on the Al atomic line.Accuracy of the results was eval-uated by the analysis of samplesfrom an interlaboratory studyand by the recovery resultsobtained after addition of ana-lytes to the hemodialysis watersamples.

INTRODUCTION

Hemodialysis is the main treat-ment used to improve the quality oflife for patients with chronic renaldeficiency. Patients submitted tohemodialysis are exposed to vol-umes of water ranging from 18,000to 36,000 L per year. The use ofinappropriately purified water canresult in the loss of several chemi-cals and absorption of toxic andbacterial substances by the patient,resulting in intoxication and death(1,2).

The presence of metals in highconcentrations can produce manyproblems for the organs. The "hardwater syndrome" was the firsthemodialysis incident related towater quality. Symptoms like nau-sea, vomiting, lethargy, muscularfaintness, and arterial hypertensionare frequent during the hemodialy-sis process. These symptoms havebeen directly associated with thepresence of Ca and Mg in water(3–6).

Like other metals in hemodialy-sis water (7–14), aluminum canlead to intoxication and neurologi-cal disturbances such as dialysisdementia, encephalopathy, catato-nia, and death (3,15,16).

The elements in hemodialysiswater, dialysate, and/or in the dialy-sis solution have been analyticallydetermined. Bohrer et al. (17) usedseparation processes and precon-centration in the determination ofPb, Cd, Zn, and Cu in a high salineconcentration matrix (dialysis solu-tion) by electrothermal atomicabsorption spectrometry (ETAAS).Milacic and Benedik (18)determined Al, Cr, Cu, Mg, and Feby ETAAS and Zn by flame atomicabsorption spectrometry (FAAS) indialysate. Due to the complexity ofthe matrix (the presence of highconcentrations of proteins andsaline constituents), the instrumen-tal parameters were optimized foreach specific element by ETAAS.

Few methods have beenproposed in the literature to deter-mine metals in hemodialysis water.Hermann et al. (19) presented amethod using inductively coupledplasma optical emission spectrome-try (ICP-OES) with ultrasonic nebu-lization for the determination of Ag,Al, As, Be, Cd, Cr, Pb, Sb, Se, and Tlbut they did not determine many ofthe other officially regulated ele-ments.

Therefore, this study aims todevelop, optimize, and apply anICP-OES methodology for the fastand accurate determination of Al,Ba, Cu, Cd, Cr, K, Na, Mg, Mn, Fe,and Zn in hemodialysis water asestablished by ANVISA (Agência

9

Vol. 28(1), Jan./Feb. 2007

Nacional de Vigilância Sanitária,Brazil ) and AAMI (Association forthe Advancement of Medical Instru-mentation, Arlington, VA, USA)(14,20).

EXPERIMENTAL

Instrumentation

The experiments for this studywere performed using thePerkinElmer® Model Optima™2000 DV inductively coupledplasma optical emission spectrome-ter (Dual View ICP-OES) with axial(AX-ICP-OES) and radial configura-tions (RD-ICP-OES) (PerkinElmerLife and Analytical Sciences, Shel-ton, CT, USA). The instrument wasequipped with a radio frequencysource of 40 MHz which providesa power of 0.75–1.5 kW. It utilizesan Echelle grating (79 lines nm–1),a solid-state detector, a plasmatorch with a 2.0-mm internal diame-ter injection, and a cross-flow nebu-lizer coupled with a double-passScott-type spray chamber. Theequipment also has an interface(shear gas) that introduces a highair flow perpendicularly to thetorch to remove the low tempera-ture extremity of the plasma. Thesample introduction system wasautomated with the PerkinElmerModel 90 Plus autosampler. Theoperational parameters used arelisted in Table I. The wavelengthsused for the elements to be deter-mined eliminate spectral interfer-ences as described in the literature(Table II).

Reagents, Solutions, and Samples

A multi-element solution of Al,Ba, Ca, Cd, Cr, Cu, Fe, K, Mg, Mn,Na, Pb, and Zn (as analytes) and theCo, Li, Ni, and Y solutions (as inter-nal standards) were prepared byappropriate dilutions of 1000 mg L–1

Certipur™ICP stock standard solu-tions (Merck, Darmstadt, Germany).Nitric acid was from Vetec, Brazil(70%, v/v, p.a.). Ultrapure water

(resistivity of 18.2 MΩ cm) wasobtained from a Milli-Q™ system(Millipore, Bedford, MA, USA). TheNaCl, KCl, and CaCl2.2H2O solu-tions used in the interference stud-ies were from Vetec, p.a.

The reference materials usedwere from an interlaboratory assayby Rede Metrológica do Rio Grandedo Sul, Brazil, obtained from theMetallic Contaminants Laboratoryof FUNED (Fundação EzequielDias), Belo Horizonte, Brazil.

Procedure

The parameters such as nebuliza-tion gas flow rate in the range of0.3 to 1.1 L min–1 and appliedpower, ranging from 1.1 to 1.5 kW,were optimized based on therobustness of the plasma using theMg(II)/Mg(I) ratio as a diagnostictool (21).

The effect of 20, 100, and 500 mg L–1 of Ca, K, Na, and themixture of the three elements (EIEs and Ca), on the robustness ofthe plasma and the effect of 10 mg L–1

of the same elements on the analyti-cal lines of interest was assessedusing both axially and radiallyviewed configurations and robustand non-robust conditions.

The choice of an appropriateinternal standard was made by eval-uating the profile of the analyticalsignals of Al and Pb and of the ele-ments tested as the internal stan-dards (Li, Co, Ni, and Y) using thesame experimental conditions. Forevaluation of the internal standardperformance, 2.0 mg L–1 each ofthe analyte solutions and 1.0 mg L–1

of the internal standard were addedto the samples equivalent to eachpoint of the calibration curve.

The figures of merit such as BEC(background equivalent concentra-tion), SBR (signal-to-backgroundratio), and LODs (limits of detec-tion) were studied for each metaldetermined in the hemodialysiswater sample. Measurements weremade under robust and non-robustconditions using the axially andradially viewed ICP-OES configura-tions.

Accuracy of the method wasevaluated by quantifying four sam-ples of an interlaboratoryproficiency assay (Rede Metrológicado Rio Grande do Sul, Brazil, 2004and 2005), using the t-test and theaddition and recovery methods fora hemodialysis water sample (0.6 mg L–1). For calculation of theaccuracy of the method, 10 repli-

TABLE IICP-OES Operating Conditions

Parameters

Applied Power 1.1–1.5 kWa

Nebulization Gas Flow 0.3–1.1 L min–1a

Auxiliary Gas Flow 0.2 L min–1

Plasma Gas Flow 15 L min–1

Pumping Flow 1.0 mL min-1Injector Tube Diameter 2.0 mmRadially Viewed Height 14 mm

Interface Shear Gas

a Values optimized during the experi-ments.

TABLE IIWavelength (λ) and Ionization

Energy (E) of the ElementsDetermined

Element Line λ (nm) E (eV)

K I 766.49 5.96Na I 589.59 7.24Ba II 455.403 7.93Al I 396.153 9.13Cu I 324.75 11.55Mg I 285.213 11.99Mn II 257.61 12.25Cr II 267.716 12.95Fe II 238.204 13.07Ca I 317.933 13.15Cd II 214.44 14.77Pb II 220.35 14.79

Zn II 213.86 15.19

I = Atomic line; II = Ionic line.

10

cates were analyzed. All other sam-ples were analyzed in triplicate.


Evaluation of Plasma in Axiallyand Radially Viewed Configura-tions in Terms of Robustness

Effect of Nebulization Gas Flow In this step, the effect of varying

the nebulization gas flow rate(0.3–1.1 L min–1) was evaluated onthe Mg(II)/Mg(I) ratio with a fixedpower of 1.3 kW in nitric acid solu-tion.

As can be seen in Figure 1, fornebulization gas flow rates higherthan 0.8 L min–1 and lower than 0.4 L min–1, the Mg(II)/Mg(I) ratiowas significantly lower. Accordingto the literature (21–25), high nebu-lization gas flow promotes a moreefficient aerosol sample formation(26). However, due to theincreased amount of solvent thatreaches the plasma central channel,a higher amount of energy is neces-sary for the aerosol generation ofthe atomic cloud. This redistribu-tion of energy may cause a decreasein the efficiency of the atomizationand ionization processes. Another

important factor that contributes tothe low Mg(II)/Mg(I) ratio at highnebulization gas flow rates is thereduced residence time of thespecies of interest in the plasmacentral channel.

For nebulization gas flows lowerthan 0.4 L min–1, a significantincrease in the residence time ofthe species of interest in the plasmais observed. Consequently, it wouldbe expected to result in animprovement in the plasma excita-tion and ionization conditions.However, it was observed that theintensity of the analytical lines alsodecreases. According to Silva et al.(27), this effect could be associatedwith the phenomenon of self-absorption caused by the high resi-dence time. On the other hand, thelow gas velocity cannot provideenough kinetic energy to convertthe liquid sample into an aerosolwith the appropriate particle sizedroplets. Consequently, the globalaerosol formation and transportprocesses are negatively influenced(28).

In the axially viewed configura-tion, it takes more time to convertthe sample aerosol into atoms, ions,

excited ions, and molecular speciesgenerated by recombination due toits longer optical path and largerresidence time in relation to theradially viewed configuration mode.This leads to larger Mg(II)/Mg(I)ratios for AX-ICP-OES than for RD-ICP-OES for equal gas flow rates.

Effect of Applied PowerThe effect of the applied power

under plasma excitation and ioniza-tion conditions was investigatedusing fixed nebulization gas flowrates of 0.6 and 1.1 L min–1, androbust and non-robust ICP-OESoperating conditions, respectively.The power was varied from 1.1–1.5 kW.

Figure 2 shows the optimalenergy transfer characteristics ofthe plasma for both ICP-OES condi-tions for the species present in thesample aerosol and under high-applied power. The results indicatethe formation of a plasma underrobust operating conditions, that is,a Mg(II)/Mg(I) ratio of ≥8 (29–30).The ratios obtained were 9.0 and11.5 for the axially and radiallyviewed configurations, respectively.

Fig. 1. Effect of nebulization gas flow rate on Mg(II)/Mg(I)ratio (applied power 1.3 kW).

Fig. 2. Effect of applied power on Mg(II)/Mg(I) ratio (nebu-lization gas flow rate of 0.6 L min-1).

11

Vol. 28(1), Jan./Feb. 2007

In contrast, at a nebulization gasflow rate of 1.1 L min–1, the plasmawas operating at non-robust condi-tions regardless of the plasmaobservation mode (Figure 3). Underthese conditions, the maximum val-ues of the Mg(II)/Mg(I) ratio wereless than four. In accordance withFigure 3, the ionic line of 280.27nm for Mg(II) was dependent onthe plasma excitation and ioniza-tion conditions due to its high ion-ization energy (Table II). A plasmaoperating at these conditions isinadequate to guarantee an efficientenergy transfer between the excita-tion source and the magnesiumatoms found in the central chan-nel, as evidenced by the lowMg(II)/Mg(I) ratios. This behaviorcan be extended to other analyteswith similar ionization energies.

In accordance with the results inFigures 1–3, the following plasmaoperating conditions were estab-lished for all other experiments:Nebulization gas flow of 0.6 L min–1

and applied power of 1.3 kW. Innon-robust conditions, the plasmaoperates with a nebulization gasflow of 1.1 L min–1 and an appliedpower of 1.1 kW.

Effects of Easily IonizableElements (EIE) on Analytical Lines

Table III presents the effects ofthe easily ionizable elements (EIEs)and Ca on the different analyticallines. This study was carried out toevaluate the effects of these speciespresent as macro elements in dialysateon the species of interest, andbecause these concurrent elementscan occur at high concentrations inhemodialysis water samples (14,20).The data in Table III are shown aseither a positive or negative %enhancement of the analyte signals.

Using radially viewed and robustconditions, a slight positive interfer-ence was observed for most analyti-cal lines investigated. The degree ofinterference varied from 0.06% to3.0%. For Al, the effects were 11.7%(mixed solution) and 26% (K), and

for Pb, the effects were 3.6%, 7.2%,8.4%, and 9.4% for Na, Ca, K, andmixed solution, respectively.

Using the axially viewed configu-ration, a positive interferencebetween 0.05% and 4.6% wasobserved. Aluminum was again anexception with 7.8% (mixed) and14.2% (K).

Although most interferencescaused by the concurrent elementswere positive, some analytes suchas Cr, Al, Zn, and Mn presentednegative interferences. For Cr, theinterference was 0.3% in the pres-ence of K; while for the atomic lineof Al (396.15 nm), the effect was6.2% (Ca) and 7.4% (Na) in RD-ICP-OES, and 6.9% (Ca) and 8% (Na) inAX-ICP-OES. For Mn and Zn, theeffects were smaller than 0.4% inAX-ICP-OES. In general, the interfer-ence effects of the EIEs were smallor negligible with a plasma operat-ing under robust conditions.

Under non-robust conditionsand using AX-ICP-OES, a behaviorsimilar to that of RD-ICP OES wasobserved. The positive interferencesranged from 0.2% to 6.0%. The

largest interference was observedfor Al, which showed a positiveeffect of 5.7% and 28.5% and a sig-nal depression between 6.0% and8.3%.

With RD-ICP-OES, the positiveeffects ranged from 0.2–5.0%,except for Pb (220.35 nm) and Al(396.15 nm), which presented asignal increase ranging from7.0–8.0%, 3.7%, and 26.0%, respec-tively. Aluminum (396.15 nm) alsopresented a negative effect rangingfrom 9.0-9.8%.

Under robust conditions (TableIII), the effects caused by EIEs andCa with AX-ICP-OES were slightlylarger than those observed withRD-ICP-OES, which is in agreementwith the results obtained in otherworks (31,32). Under non-robustconditions, the effects in generalwere larger for the analytical linesinvestigated, but independent fromthe configurations used in thisstudy.

Under robust conditions, theeffect of the interferent was mini-mized on most analytical linesinvestigated, except for Al

Fig. 3. Effect of applied power on Mg(II)/Mg(I) ratio (nebu-lization gas flow rate of 1.1 L min–1).

12

TABLE IIIInterference Effect of Easily Ionizable Elements (EIEs) and Ca on Different Analytical Lines Using ICP-OES

(Results expressed in either a positive or negative percentage (%) enhancement of the analyte signal.)

Analyte λ (nm) Robust Conditionsa Robust Conditionsa

RD-ICP-OESc AX-ICP-OESd

K Ca Na Mixe K Ca Na Mixe

Al 396.15 26.0 –6.2 –7.4 11.7 14.2 –8.0 –6.9 7.8

Cu 324.75 1.0 0.09 2.1 2.7 1.3 –0.6 2.8 4.6

Mg 285.21 0.5 0.06 1.0 0.4 0.05 0.5 0.1 1.6

Ba 455.40 2.0 2.1 2.9 3.0 2.9 2.3 2.9 3.3

Mn 257.61 0.9 2.4 0.5 0.8 –0.3 0.8 1.2 1.4

Cr 267.72 –0.3 1.4 1.0 0.7 0.6 0.9 0.3 0.9

Fe 238.20 1.7 2.2 1.9 1.6 0.3 2.0 0.2 1.3

Cd 214.22 2.0 2.9 2.5 3.0 1.8 1.7 1.9 1.2

Pb 220.35 8.4 7.2 3.6 9.4 2.5 0.7 1.4 3.6

Zn 213.86 1.9 0.7 2.8 3.1 1.0 –0.4 2.6 2.6

Analyte λ (nm) Non-Robust Conditionsb Non-Robust Conditonsb

RD-ICP-OESc AX-ICP-OESd

K Ca Na Mixe K Ca Na Mixe

Al 396.15 26 –9.8 –9.0 3.7 28.5 –8.3 –6.0 5.7

Cu 324.75 0.2 –0.3 1.3 0.7 2.6 1.7 3.0 4.5

Mg 285.21 2 2 1.5 0.8 3 3 4.0 6.0

Ba 455.40 3.7 5.0 4.7 3.9 3.6 3.0 3.3 1.7

Mn 257.61 1.0 1.4 0.8 0.5 0.6 0.5 0.6 0.2

Cr 267.72 1.8 2.3 3.0 2.4 1.4 2.0 3.4 2.0

Fe 238.20 1.0 1.6 2.0 1.9 2.0 1.6 2.8 0.9

Cd 214.22 –0.3 1.6 3.0 1.3 1.8 1.8 3.3 1.0

Pb 220.35 7.0 8.0 7.2 7.7 2.5 0.8 4.3 2.8

Zn 213.86 1.9 0.5 3.0 1.3 2.7 1.5 4.0 3.8

a Nebulization gas flow rate: 0.6 L min–1; applied power: 1.3 kW.b Nebulization gas flow rate: 1.1 L min–1; applied power: 1.1 kW.c RD-ICP-OES = radially viewed ICP-OES.d AX-ICP-OES = axially viewed ICP-OES. e Solution containing 10 mg L–1 of each element.

13

(396.15 nm) and Pb (220.35 nm).The interference effect was gener-ally over 5.0% in the presence ofconcurrent species.

Internal Standardization (IS)

The analytical lines of Al (396.15 nm) and Pb (220.35 nm)suffer interferences even underrobust conditions due to the pres-ence of 10 mg L–1 Ca, K, Na, andthe mixture containing the concur-rent species (Mix) (Table IV). Analternative way to reduce theeffects caused by these elementswas the use of internal standards.Initially it was decided to studyLi (610.36 nm, 3.87 eV) and Y(371.029 nm, 3.52 eV) on Al(396.15 nm, 3.14 eV); and Y(371.029 nm, 9.9 eV), Co (228.61nm, 13.7 eV), and Ni (231.6 nm,14.03 eV) on Pb (220.35 nm, 14.8 eV)in order to verify which of theseelements has atomization and ion-ization profiles similar to those ofAl and Pb (33–35). In this study,previously established robust andnon-robust conditions and bothICP-OES observation modes wereevaluated. The results obtained forAl and Pb are shown in Figure 4.

It can be observed that for bothplasma observation modes (Figures4A and B), Y and Li presented pro-files very similar to that of Al whensubmitted to the same conditions.On the other hand, Ni was the ele-ment whose profile was most simi-lar to that of Pb. According to theresults, Y and Li were used as aninternal standard for correction ofthe matrix effects caused by K, Ca,and Na on the Al emission lines,while Ni and Y were used for Pb(see Table IV).

As can be observed in Table IV,the application of an internal stan-dard significantly reduced the inter-ference effect on the Al and Pbanalytical lines. For Al, both axiallyand radially viewed, the interfer-ence effect caused by K and Ca wasbetter corrected by using Li as aninternal standard and the interfer-

ence effect caused by the mixtureof the three elements wasimproved with Y as an internalstandard. The effect caused by Naon the Al line was not significantlycorrected by any of the twoelements tested. For Pb, Y was amore efficient internal standard inthe presence of K and the three-element (K, Ca, and Na) mixtureand Ni was the better choice in thepresence of Ca and Na.

Limits of Detection

The limits of detection (LODs)were determined in both axiallyand radially viewed configurationsunder robust and non-robust condi-tions. The values obtained are givenin Table V.

The LODs obtained in AX-ICP-OES under non-robust conditionswere better for atomic lines with1.62–7.93 eV, while for ionic linesbetween 12.2–15.2 eV energy, theLODs were better under robustconditions. The values of LODsobtained do not present any trendrelative to AX-ICP-OES or RD-ICP-OES). For some lines, the LODswere better in AX-ICP-OES, whileothers had better LODs in RD-ICP-OES, regardless of the type of line.

The LODs obtained for the ele-ments were below the maximumvalues allowed by Anvisa and AAMI,which makes ICP-OES an attractiveand feasible technique for thedetermination of trace elements Ca,K, Mg, Na, Al, Ba, Cr, Cu, Fe, Mn,and Zn in hemodialysis water sam-ples. The only exceptions are Aland Cd.

Accuracy

Accuracy was determined by thequantification of several elementsin water samples in the interlabora-tory proficiency assays and by addi-tion and recovery experimentsusing both viewing configurationsand robust operating conditions.The results are presented in TablesVI and VII, respectively. The values

determined agree with the certifiedvalues of the assay. The t-test wasused to compare the resultsobtained. Most results agreed at theconfidence interval evaluated(P=0.05), and the recoveries wereacceptable at an interval of ±10% ofthe real value, except for Na inAX-ICP-OES.

CONCLUSION

Optimization of the ICP-OESexperimental parameters was usedto establish optimal plasma operat-ing conditions to determine macro-(Ca, K, Mg and Na) and microele-ments (Al, Ba, Cd, Cr, Cu, Fe, Mn,Pb and Zn) in hemodialysis watersamples with both axially and radi-ally viewed configurations. Evenin the presence of 20, 100, and500 mg L–1 K, Ca, Na, and theirmixture, the plasma conditionsremained practically unaltered[Mg(II)/Mg(I) ratio of ≥ 8], with aninterference effect lower than 14%.Other authors found that AX-ICPOES gave higher Mg(II)/Mg(I) ratiosand a more profound interferencein the presence of easily ionizedelements (EIEs). However, ourstudy showed that even usingrobust operating conditions in thepresence of 10 mg L–1 each of K,Ca, Na, and their mixture, a signifi-cant interference effect wasobserved on the Al and Pb lines.However, this was sufficientlyreduced by the use of adequateinternal standards Y and Li for thecorrection of the matrix effectscaused by K, Ca, and Na on the Alemission lines, while Ni and Y wereused for Pb. The detection limitsobtained were below the maximumvalue established by ANVISA (Agên-cia Nacional de Vigilância Sanitária,Brazil) and AAMI (Association forthe Advancement of Medical Instru-mentation, Arlington, VA, USA).The precision of the proposedmethod was good which was veri-fied by the results obtained of theanalyte addition and recovery testsand the interlaboratory proficiencyassays.

Vol. 28(1), Jan./Feb. 2007

14

TABLE IVStudy of the Interference Effect (%) Caused by the Presence of EIEs(10 mg L–1) on Al (396.15 nm) and Pb (220.35 nm) analytical Lines.

Internal Standards: Li: 610.65 nm (1 mg L–1), Y: 371.03 nm (1 mg L–1),and Ni: 231.6 nm (1 mg L–1)

Inter- Aluminum Leadferent Radial View Axial View Radial View

– Li Y – Li Y – Ni Y

K 26 0.5 –3 14.3 1.4 –3.0 8.4 –5.0 –0.1

Ca –6.2 –2.3 –4.3 –8.0 –1.5 –4.0 7.2 –2.5 3.3

Na –7.4 –6.9 –4.3 –6.9 -5.7 –2.1 3.6 –0.6 –2.6

K, Ca, Na (Mix) 11.7 -2.3 2.1 7.8 –1.0 0.3 9.4 1.6 0.8

Fig. 4. Response profile of internal standard elements (Li, Y, Co, Ni) tested relative to the analyte elements Al (306.15 nm) andPb (220.35 nm):

(A) Li and Y relative to Al using RD-ICP-OES.(B) Li and Y relative to Al using AX-ICP-OES.(C) Co, Y, and Ni relative to Pb using RD-ICP-OES.(D) Co, Y, and Ni relative to Pb using AX-ICP-OES.

(A) (B)

(C) (D)

15

Vol. 28(1), Jan./Feb. 2007

TABLE VLimit of Detection (µg L–1) in Hemodialysis Water Under DifferentExperimental Conditions by AX-ICP-OES and RD-ICP-OES and the

Maximum Limit Allowed by Anvisa (MVA)

Element λ (nm) E (eV) Robust Non-Robust MVAConditions Conditions

Radial Axial Radial Axial

K (I) 766.49 1.62 38.0 40.7 3.9 4.3 8000Na (I) 589.59 2.1 9.9 0.05 2.3 0.03 70,000Al (I) 396.153 3.14 1.9 10.8 3.8 238.7 10Cu (I) 324.75 3.82 1.9 2.1 0.5 0.3 100Mg (I) 285.213 4.34 2.1 0.4 0.9 0.3 4Ca (I) 317.933 7.04 6.1 7.9 8.7 6.4 2000Ba (II) 455.403 7.93 0.2 0.0016 0.04 0.002 100Mn (II) 257.61 12.25 0.5 0.1 0.5 0.3 --Cr (II) 267.716 12.95 2.8 0.4 4.9 1.4 14Fe (II) 238.204 13.07 3.3 4.4 3.6 82.4 --Cd (II) 214.44 14.77 2.5 2.8 18.5 24.6 1Pb (II) 220.35 14.79 28.3 9.5 86.0 24.7 5

Zn (II) 213.86 15.19 1.9 0.6 7.1 1.2 100

(I) = Atomic Line. (II) = Ionic Line.Note: The data underlined represent LODs and do not meet the limits allowed by

Anvisa.

TABLE VIElement Concentration in Water Sample of Interlaboratory

Proficiency Assays (µg L–1)Measurements using robust conditions for AX-ICP-OES and

RD-ICP-OES configurations. Average values ± SD (n=3).

Element λ(nm) Sample 1 Sample 1 ExpectedRadial View Axial View

Mn 257.61 0.100 ±<0.001 0.101 ± 0.0004 0.103 ± 0.010

Fe 238.20 0.106 ± 0.0012 0.106 ± 0.002 0.121 ± 0.021

Cd 214.44 0.008± 0.005 0.001 ±<0.001 0.012 ± 0.003

Element λ(nm) Sample 2 Sample 2 ExpectedRadial View Axial View

K 766.49 14.78 ± 0.0013 17.67 ± 0.09 14.9 ± 1.5

Al 396.15 0.079 ± 0.003 0.808 ± 0.001 0.92 ± 0.08

Cu 324.75 0.944 ± 0.011 0.947 ± 0.014 0.904 ± 0.066

Pb 220.35 0.481 ± 0.0012 0.431 ± 0.001 0.437 ± 0.061

Zn 213.86 0.956 ±<0.001 0.965 ± 0.002 0.92 ± 0.07

ACKNOWLEDGMENTS

The authors are thankful to Con-selho Nacional de Pesquisas eDesenvolvimento Tecnológico(CNPq) for financial support andscholarships.

Received October 17, 2006.

REFERENCES

1. A.M.M. Silva, C.T.B. Martins, R.Ferraboli, V. Jorgetti, and J.E.R.Junior, J. Bras. of Nefrol. 18, 180(1996).

2. CVS (Centro de VigilânciaSanitária),Manual de boas práticasem Terapia Renal Substitutiva, Sec-retaria de vigilância sanitária, Riode Janeiro, Brazil (2004).http;//www.saude.rj.gov.br.

3. ADRNP (Associação dos doentesrenais do norte de Portugal)(2003). http://www.adrnp-sede.rcts.pt.

4. ABC – SAÚDE. Orientações parapacientes com insuficiência renalcrônica. Santa Catarina, Brazil(2001). http://www.abcdasaude.com.br

5. RM. Freeman, R. Hawton, and M.A.Chamberlain, Engl. J. Med. 276,1113 (1987).

6. M.C.M CASTRO, J. Bras. Nefrol 23,108 (2001). http://www.sbn.org.br

7. UTD (Thecnical University of Delft,The Netherlands), Tabla Period-ica.Universidade de Delft(1993).http://www.lenntech.com/espanol/tabla-periodica/Ba.htm

8. WHO (World Health Organization).IPCS: Enviromental Healthy Crite-ria 134- Cadmium. Geneva,Switzerland (1992).

9. J.R. Cannata, J. L. Martín, C.G.Alonso, and M. Serrano, Nefrol.V.X, 3, 28 (1990).

10. R.A. Goyer, New York: PergamonPress, 5 ed., pg. 691 (1996).

16

TABLE VIIRecovery of Elements in Hemodialysis Water Samples

Spiked With 0.6 mg L–1 of Element; (RSD, n = 3)

Radial View Axial ViewElement Obtained RSD Recovery Obtained RSD Recovery

(mg L–1) (%) (%) (mg L–1) (%) (%)

Al 0.595 0.08 99.2 0.586 1.08 97.7Ba 0.58 0.47 96.7 0.559 1.79 93.2Cu 0.635 0.23 105.8 0.624 1.16 104.0Cd 0.614 0.92 102.3 0.618 0.69 103.0Cr 0.612 0.34 102.0 0.618 0.58 103.0Mn 0.626 1.09 104.3 0.621 1.46 103.5Mg 0.594 0.41 99.0 0.599 0.87 99.8Fe 0.615 1.15 102.5 0.625 0.03 104.2Zn 0.607 0.85 101.2 0.618 0.38 103.0Na 0.563 0.61 93.8 0.523 0.31 87.2K 0.581 2.17 96.8 0.543 1.09 90.5Ca 0.556 1.02 92.7 0.563 0.12 93.8

Pb - - - 0.594 1.25 99.0

11. TBP (The Biology Project), Kidneysand Metals Problem Set. Universi-dade do Arizona, USA (2004).http://www.biology.arizona.edu/chh/problem_sets.html

12. D.G. Barceloux, Clin. Toxicol. 37,217 (1999).

13. WHO (World Health Organization).IPCS: Enviromental Healthy Crite-ria – 184 Aluminium. Geneva(1997).

14. AAMI (Association for the Advance-ment of Medical Instrumentation)Standards and Recommended Prac-tices. Arlington, VA, USA, v3. 4thed. (1995).

15. F.A. Azevedo and A.A.M. Chansin,São Paulo: Atheneu, 554 (2003).

16. R.V.V. Calderaro, 2002. 214f. Tese(Doutorado em Veterinária)- Uni-versidade Federal de Minas Gerais,Belo Horizonte, Brazil (2002).

17. D. Bohrer, P.C. Nascimento, M.Guterres, M. Trevisan, and E. Seib-ert, The Analyst, 124, 1345 (1999).

18. R. Milacic and M. Benedik, J.Pharm. Biomed. Anal. 18, 1029(1999).

19. A.B. Herrmann, E.J. Santos, L.A.Pegoraro, and F. Leite, ENQA(Encontro Nacional de QuímicaAnalítica, Brazil), 13º, Niteroi,A155 (2005).

20. ANVISA, RDC Nº 154, Brasil. Legis-lação em Vigilância Sanitária, Res-olução -RDC Nº 154, de 15 dejunho de 2004 (June 15, 2004).http://www.anvisa.gov.br

21. I.B.Brenner and A.T. Zander, Spec-trochim. Acta Part B. 55, 1195(2000).

22. J.C.J. Silva, D.M. Santos, N. Baccan,S. E. Cadore, and J.A. Nóbrega,Microchem. J., 77, 185 (2004).

23. V. Thonsen, G. Roberts, and K.Burgess, Spectroscopy 15, 33(2000).

24. I.B.Brenner, A. Zander, M. Cole,and A. Wiseman, J. Anal. At. Spec-trom. 12, 897 (1997).

25. C. Dubuisson, E. Poussel, and J.M.Mermet, J. Anal. Atom. Spectrom.12, 281 (1997).

26. R.F. Browner, A. Canals, and V. Her-nandis, Spectrochim. Acta 47B,659 (1992).

27. J.C.J Silva, N. Baccan, and J.A.Nobrega, J. Braz. Chem.Soc. 14,330 (2003).

28. S.J. Hill, 1st ed., Sheffield Acade-mic Press (England), pg. 370(1999).

29. J. M. Mermet, Anal. Chim. Acta 250,85 (1991).

30. F.V. Silva, L.C. Trevisan, C.S. Silva,A.R.A. Nogueira, and A.J. Nobrega,Spectrochim. Acta Part B. 57, 1905(2002).

31. I.B.Brenner, A. Zander, M. Cole, andA. Wiseman, J. Anal. At. Spectrom.12, 897 (1997).

32. I.B.Brenner, A.L. Marchand, C.Daraed, and L. Chauvet,Microchem. J. 63, 344 (1999).

33. C. Dubuisson, E. Poussel, and J.M.Mermet, J. Anal. At. Spectrom., 13,1265 (1998).

34. J.C. Ivaldi and J.F. Tyson,Spectrochim. Acta Part B 51, 1443(1996).

35. R.K.Winge, V.J. Peterson, and V.A.Fassel, Applied Spectroscopy 33,257 (1979).

17

*Corresponding author.E-mail: [email protected]: +55 16 3301-6692

Comparison of Metallic and Ceramic Tubes asAtomization Cells for Tin Determination by TS-FF-AAS

*Fabiana Aparecida Loboa, Ana Cristina Villafrancaa, Adriana Paiva de Oliveirab, and Mercedes de Moraesa

a Departamento de Química Analítica, UNESP-Universidade Estadual Paulista, 14801-970, P.O. Box 355, Araraquara-SP, Brazil

b Departamento de Química, UFMT – Universidade Federal do Mato Grosso, 78060-900, Cuiabá–MT, Brazil

Atomic SpectroscopyVol. 28(1), Jan./Feb. 2007

ABSTRACT

This work describes the devel-opment of an analytical proce-dure for on-line tin determinationusing thermospray flame furnaceatomic absorption spectrometry(TS-FF-AAS). Two tubes wereevaluated as atomization cells: a metallic tube (Ni-Cr, principalcomponents composition:73.95% Ni and 16.05% Cr) and a ceramic tube (99.8% Al2O3,).The use of air as the carrier wasmade by employing a Rheodynevalve to inject the samples, allow-ing an analytical frequency of 90 h-1

and avoiding sample dispersion.The carrier flow rate (air), sam-ple volume injected, and acidconcentration (HCl) were evalu-ated for the optimization of theTS-FF-AAS system. The sensitivityfor 50 mL of analytical solutionwith TS-FF-AAS was 2 and 5 timeshigher (to metallic and ceramictube, respectively) than using anacetylene-nitrous oxide flamewith pneumatic aspiration(requiring a sample volume ofapproximately 20 times higher).

oxide flame is the recommendedAOAC (3) official method forcanned foods. The flame tempera-ture reaches approximately 3000 ºC,thus minimizing possible interfer-ences in the analyte atomizationprocess. However, this high tem-perature can deteriorate the equip-ment and its connections.Moreover, the cost of an oxidant isvery high in comparison to com-pressed air (4). Other systems weredeveloped for tin determinationsuch as hydride generation atomicabsorption spectrometry (HGAAS).However, this method results indouble and false signals, memory

INTRODUCTION

Tinplate has been used for pre-serving food for well over a hun-dred years. Today it provides arobust form of packaging by allow-ing minimization of headspace oxy-gen and sterilization of thefoodstuff within the hermeticallysealed can. It also provides a long,safe, and ambient shelf life with noor minimal use of preservatives (1).However, the use of tinplate couldresult in contamination since sometin could dissolve into the food,particularly when plain uncoatedinternal surfaces are used. Monitor-ing this process is, therefore, veryimportant to ensure quality stan-dard and industrial control.

Atomic absorption spectrometry(AAS) is employed in the quantita-tive determination of manyelements (metals and semi-metals)in a large variety of food, biological,environmental, geological, andother samples. The principle of thetechnique is based on measuringthe absorbance of electromagneticradiation (originating from a radia-tion source) by gaseous atoms inthe fundamental state. Thesegaseous atoms are obtainedthrough sample atomization byflame, electrothermal, or a specificchemical reaction. Flame atomicabsorption spectrometry (FAAS) isthe most widely used technique forelemental determination at the mgL–1 levels (2).

When tin is determined byFAAS, use of acetylene-nitrous

effects, tin deposition in the quartzcell, re-volatilization and tin adsorp-tion on the surface of the quartzcell, elevated time of purging of thehydride, and other adverse effects(5). The interesting alternative fortin determination is thermosprayflame furnace atomic absorptionspectrometry (TS-FF-AAS) (4,6).

Thermospray (TS) was firstdeveloped by Vestal et al. in 1978as an interface between liquid chro-matography and mass spectrome-try. In atomic spectrometry,heating of the tube was performedelectrically which limited the use ofthe technique. In contrast, Gásparand Berndt proposed TS-FF-AAS (7)where a metallic tube is placed onthe burner head of the atomicabsorption spectrometer, function-ing as a reactor. The liquid is trans-ported via a metallic capillary intoa tube, both of which are simulta-neously heated by the flame. As theliquid reaches the end of theheated capillary, it vaporizes par-tially and an aerosol is produced.Finally, the aerosol is vaporized intothe tube, generating an atomiccloud that absorbs the radiationoriginating from the lamp (7). Asa result of this work, many applica-tions employing TS-FF-AAS havesince been developed andpublished (8–18).

The aim of this paper is thedevelopment of a system for on-linetin determination by thermosprayflame furnace atomic absorptionspectrometry supplied with acety-lene-air, where air is used as thecarrier. In addition, metallic andceramic tubes were evaluated foruse as the atomization cells. Theresults are compared with FAASusing acetylene-nitrous oxide.

18

EXPERIMENTAL

Instrumentation

A PerkinElmer® AAnalyst™ 100flame atomic absorption spectrome-ter, equipped with deuterium lampbackground correction and a hol-low cathode lamp (HCL) of Sn(λ=224.6 nm, slit=0.2 nm andi=20 mA) was used (PerkinElmerLife and Analytical Sciences, Shel-ton, CT, USA). Acetylene-air (4:2)flame was used for all experiments.

A Sartorius BL 2105 analyticalbalance, two thermocouples ModelK (Chromel vs. Alumel) with arange of –200 °C to 1370 °C, and aModel 51 K/J thermometer (Fluke)were used.

The TS-FF-AAS System

The thermospray flame furnaceatomic absorption spectrometry(TS-FF-AAS) system consisted of an8-channel peristaltic pump(Ismatec-ICP 8) with Tygon® pumptubing, a Rheodyne RE9725 samplevalve introduction system, andPEEK sample loops. The thermo-spray flame furnace unit consistedof an OMEGA TRX-164116 ceramiccapillary (OMEGATITE® 450, 0.4 mm. i.d., 1.6 mm o.d., 100 mmlong) and either an Inconel® metal-lic tube (Ni-Cr, principal compo-nents composition: 73.95% Ni and16.05% Cr) or the ceramic tube(Al2O3, 99.8%), both with 10.0 mmi.d., 12.0 mm o.d., and 100 mmlength. The tube was attached tothe burner head on the lab-mademetal alloy support with fourceramic pins. The tube was laid onthese pins. To adjust the atomiza-tion temperature for Sn inside thetube, six holes of 2.5 mm diameterwere drilled into the bottom part ofthe tube. Another orifice of 2.0 mmdiameter was drilled at 90º to thebottom orifices where the thermo-spray capillary was inserted. Thecapillary tip was pushed about 1 mm inside the tube furnace. The manifold was assembled with0.5 mm i.d. PTFE tubing.

Data Evaluation

This spectrometer does not havea suitable data/signal display. How-ever, using a separate personalcomputer and custom-written soft-ware, all signals could bevisualized, recorded, and stored.The data evaluation system (Micro-química Model MQ196) was man-aged by an IBM® PC/AT-486microcomputer using a programwritten in Visual Basic® software.The absorbance values were propor-tional to transient signal height.

Reagents and Solutions

All solutions were preparedusing deionized water (Permutiondeionizer systems, resistivity 18.2 MΩ cm) and HCl (Merck,Darmstadt, Germany). Glasswareand propylene flasks were washedwith Extran (Merck), soaked in 10%(v/v) HNO3 and rinsed with deion-ized water prior to use.

A stock solution containing1000 mg L–1 Sn was preparedweekly by weighing metallic Sn(99%, Carlo Erba) into a small glassbeaker, then dissolving in concen-trated HCl, and diluting with deion-ized water in a volumetric vessel.

The tin solution was standardizedby back-titration. An excess of stan-dard EDTA (Merck, Darmstadt,Germany) solution was added, theresulting solution was buffered atpH 6, and excess reagent was back-titrated with a standard zinc(Merck, Darmstadt, Germany) solu-tion (4,19). The working standardsolutions of tin (3.0–100 mg L–1)were prepared daily by successivedilutions of the 1000 mg L–1 Snstock solution.

Schematic of TS-FF-AAS System

The schematic of the TS-FF-AASsystem is shown in Figure 1 (4).Theliterature recommends attachmentof the atomization tube to theburner head only after igniting theflame in order to avoid the dangerof explosion (7,10). Nevertheless,in this work, the TS-FF-AAS systemwas first assembled and the tubewas inserted onto the burner headafter turning on the gas (acetylene/air), then the flame of the spectrome-ter was lit immediately to avoid anexplosion inside the tube due thegas accumulation. This assemblyand the heating steps made this pro-cedure simple and safe (4,6,15,18).

Fig. 1. Schematic of TS-FF-AAS system.

19

Vol. 28(1), Jan./Feb. 2007

In the TS-FF-AAS system, thesample volume was injected intothe carrier stream (air) in order toavoid sample dilution. According tothe literature, use of solution carri-ers can increase the dispersion ofthe sample (4,6,9,15,18).

Introduction of the sample intothe TS-FF-AAS system (Figure 1) isperformed manually using a Rheo-dyne valve where the carrier (air)transports the sample to theceramic capillary. As the ceramiccapillary is heated simultaneouslywith the ceramic tube (reactor),the liquid is partially vaporized,obtaining the thermospray. Atom-ization occurs when the thermo-spray reaches the ceramic tube,producing transient signals. Thetransient signals were recorded andthe information stored. The tran-sient signal heights were used asthe analytical parameters of themeasurements.

Optimization of Carrier FlowRate, Sample Volume, and HClConcentration

Influence of carrier flow rate(0.6–6.0 mL min–1), sample volume(20, 50, 100, and 200 mL), and HClconcentration (2–10%, v/v) wasevaluated for both the metallic andceramic tubes. Standard solutionscontaining 30 mg L–1 Sn were usedfor these studies.

Construction of AnalyticalCurve

After system optimization for themetallic and ceramic tubes, analyti-cal curves were obtained in the3.0–30.0 mg L–1 Sn range, and thelimits of detection (LOD) weredetermined by the concentrationcorresponding to three times thestandard deviation of 12 measure-ments of the blank divided by theslope of the analytical curves.


Influence of Carrier Flow Rate(Air)

Influence of the carrier flow rate(air) was studied in the 0.6–6.0 mLmin–1 range for the absorbance val-ues of a 50-mL solution containing30 mg L–1 Sn in 2.0% (v/v) HCl.The results obtained are shown inFigure 2a.

The metallic tube verified thatlow carrier flow rates reduce theabsorbance value because the sam-ple enters the atomizer slowly,while the measurement time is con-siderably increased, resulting in anerratic and uncontrolled vaporiza-tion. It was observed that highercarrier flow rates provide higherabsorbance values due to a homo-geneous sample vaporization(7,11). This increase occurs up to a3.0-mL min–1 flow rate resulting inmaximum absorbance value. Withcarrier flow rates higher than 3.0 mL min–1, the residence time ofthe flowing liquid in the heated tipof the capillary significantlydecreases. This also results in adecrease in evaporation time andthe sample does not achieve thespray/aerosol state, but remains asa continuous liquid flow. The inter-nal temperature of the tube alsodecreases (the color of the tubechanges from ruby-red to white),droplets form on the outside of theatomizer tube, an irregular tran-sient signals is produced, and thestandard deviation increases. Thus,a carrier flow rate of 3.0 mL min-1

was selected as the best flow ratedue to the high absorbance valueand low standard deviation.

With the ceramic tube, low car-rier flow rates showed irregularabsorbance values. For flow ratesfrom 3.0 to 3.6 mL min-1, theabsorbance values increased due tosample vaporization and a morehomogeneous atomization.

On the other hand, flow rateshigher than 4.2 mL min–1 decreased

the residence time of the liquidflow in the heated tip of the capil-lary and reduced the time for vapor-ization. In that case, the sample didnot achieve a spray/aerosol state,but remained as a continuouslyflowing liquid. The internal temper-ature of the tube decreased (thecolor of the tube changes fromruby-red to opaque white). In addi-tion, droplets were observed on theoutside of the atomizer tube, result-ing in irregular transient signal andincreasing the standard deviation.

Therefore, for flow rates higherthan 3.0 mL min–1, the medianabsorbance values did not vary, butflow rates higher than 3.6 mL min–1

showed high standard deviationvalues. Thus, a carrier flow rate of3.0 mL min-1 was selected as thebest flow rate, despite the higherabsorbance values.

Sample Volume

The sample volume was variedfrom 20 to 200 µL using a flow rateof 3.0 mL min–1; the results areshown in Figure 2b.

Use of Metallic Tube

Using the metallic tube verifiedthat 20 µL of sample showed signifi-cant absorbance values; however,this value is low due to the smallsample volume. The use of 50 µL ofsample increased the intensity ofthe absorbance signal significantly,because the atom population in theabsorption volume is greater (8). Forthese two volumes (20 and 50 µL),the standard deviation was relativelylow. Since atomization of the ana-lyte occurs homogeneously, thereis good repeatability in the experi-mental measurements. When a100-µL sample volume was used,there was a cooling of the ceramiccapillary and, consequently, of themetallic tube and the absorbancevalues did not increase proportion-ally. For a 200-µL sample volume,the ceramic capillary and metallictube were cooled and provided nohomogeneous formation of

20

spray/aerosol; instead, an irregularformation of droplets occurs whichdisperses the radiation. Therefore,the analytical signals are not repeti-tive and show high standard devia-tion. For this study, a samplevolume of 50 µL was selected dueto high atomization homogeneity,significant absorbance values, andlow standard deviation.

Use of Ceramic Tube

Use of the ceramic tube verifiedthat a 20-µL sample showed signifi-cant absorbance values; however,this value is low due to the smallsample volume. A 50-µL sample vol-ume increased the intensity of theabsorbance signal significantly,because the atom population in theabsorption volume is greater. For

both the 20- and 50-µL sample vol-umes, the standard deviations wererelatively low and showed goodrepeatability since atomization ofthe analyte occurs homogeneously.

Using a 100-µL sample volumeresulted in cooling of the ceramiccapillary and, consequently, of theceramic tube. Droplets wereformed which irregularly dispersethe radiation, thus significantlydecreasing the efficiency of theatomization and the analytical sig-nal and resulting in high standarddeviation.

For a 200-µL sample, significantcooling of the tube is even morepronounced. There is also no

homogeneous spray/aerosol forma-tion, because an irregular formationof droplets occurs thus dispersingthe radiation. Consequently, theanalytical signals are not repetitiveand for that reason also show highstandard deviation.

Air Used as Carrier

Air as the carrier was used for a 50-µL sample volume since itachieved high atomization homo-geneity, high absorbance signals forthe standard, and low standarddeviation. These characteristicsindicate repeatability of the experi-mental measurements for this sam-ple volume.

Fig. 2. Otimization. (a) Influence of the carrier (air) in the absorbance values. Sample volume: 50 µL, 30 mg L–1 Sn in 2% (v/v)HCl. (b) Effect of sample volume on the absorbance values. Flow rate: 3.0 mL min–1; 30 mg L–1 Sn in 2% (v/v). (c) Effect of acid-ity on the absorbance values. 30 mg L–1 Sn and (d) Transient signals for TS-FF-AAS system using different carriers and differenttubes as atomization cells.

21

Vol. 28(1), Jan./Feb. 2007

Figure 2c shows the influence of2–10% (v/v) HCl concentration onthe absorbance signals for a 50-µLsolution containing 30 mg L–1 Snand using a carrier flow rate (air) of3.0 mL min–1.

Metallic Tube

Figure 2c illustrates that theabsorbance values for the metallictube decrease considerably withHCl concentrations higher than3.0% (v/v), probably due to the lossin atomization because the analyteis susceptible to higher volatiliza-tion. An acidity of 2.0 % (v/v) wasselected due to the high absorbancevalues and low standard deviation.

Ceramic Tube

For the ceramic tube, theabsorbance signal varied becausethe acidity was too low. Since highabsorbance values were obtainedwith 4.0% (v/v) HCl, for furtherstudies 2.0% (v/v) HCl was selectedbecause it results in low standarddeviation.

Absorbance Values WithCeramic and Metallic Tubes

The absorbance values obtainedwith the ceramic tube were higherthan with the metallic tube. Tostudy the reasons for this discrep-ancy, the temperature inside thetubes was measured by using twothermocouples: Model K (Chromelvs. Alumel) in the temperaturerange of –200 to 1370 °C (51 K/JThermometer, Fluke). In the ceramictube, half of the tube, close to theflame, remained ruby-red and theinside temperature varied from 980 to 1000 °C. For the metallictube, the inner temperature variedfrom 1030 to 1060 °C and the tuberemained completely ruby-red onthe flame (4,6,18).

When 50 µL of HNO3 (~0.1 mol L–1)was injected, there was a decreasein temperature of around 50 °C and150 °C for both the metallic andceramic tubes, respectively, due tothe cooling of the tube by the solu-

tion added. However, the tempera-ture increased rapidly up to themaximum of 1030 to 1060 °C and980–1000°C for the metallic andceramic tubes, respectively(4,6,18).

The temperatures in the ceramictube were found to be lower thanin the metallic tube for all experi-ments. This can be explained bythe fact that a rapid analyte conden-sation occurs in the upper part ofthe ceramic tube, allowing a pre-concentration and rapid tin sorp-tion on the wall of the tube.

Figure 2d shows that when allliquid is evaporated, the solid in theceramic tube is vaporized and agradual analyte atomization takesplace, causing an increase in theabsorbance signals. This procedureresembles the Atom-TrappingAtomic Absorption Spectrometry(ATAAS) method as reported byMetcalfe (20). The pre-concentra-tion occurs outside the tube, thesolution is aspirated by pneumaticnebulization, and cooling of theinside of the tube is due to thewater added. The analyte is concen-trated on the outside of the tubeand, when it passes the air streaminside the tube, the sample isvaporized (20).

Another explanation for highersensitivity in the ceramic tube vs.the metallic tube can be that theceramic tube (Al2O3) hasplaces/holes that help in the ana-lyte atomization efficiency, and pos-sibly by catalytic effect in the tinatomization process (4,6).

According to Figure 2d, when2.0% (v/v) HCl was injected, thelinear base increased andsubsequently decreased due to thecooling of the metallic tube. Thecooling of the tube was significantbecause an injection of 30 mg L–1

Sn did not result in a transient sig-nal since the temperature was notsufficient for atomization.However, when air was used as the

carrier, repetitive signals and a sta-ble linear base for the two-atomiza-tion cells were observed. The useof air (or gas) as the carrier avoidsdilution and dispersion of the sam-ple, decreases analytical costs andresidues generated, and providesgood analytical results in the tindetermination with the TS-FF-AASsystem.

Figure 3 shows the resultsobtained for tin determinationusing the metallic tube, ceramictube, and FAAS. The analyticalcurves were linear, and theincrease in sensitivity with the TS-FF-AAS system using 50 µL of ana-lytical solution and air as the carrierwas 2 and 5 times higher with themetallic and ceramic tubes, respec-tively, than with the acetylene-nitrous oxide flame.

The analytical characteristics fortin determination by TS-FF-AASemploying the metallic and ceramictubes as atomization cells are listedin Table I.

It can be seen that the relativestandard deviations and limits ofdetection obtained with TS-FF-AASwere lower than by conventionalFAAS. The sensitivity obtained byTS-FF-AAS with air as the carrierwas higher than the resultsachieved by FAAS with acetylene-nitrous oxide.

CONCLUSION

Optimization of the thermosprayflame furnace atomic absorptionspectrometry (TS-FF-AAS) systemusing ceramic and metallic tubes asatomization cells for the determina-tion of tin permitted observation ofthe parameters that influenced theatomization process such as samplevolume, carrier flow rate, and acid-ity, and allowed selection of theexperimental conditions. The TS-FF-AAS system developed is simple,promising, requires minimal samplevolume (only 50 µL), uses lessreagents and generates less

22

Fig. 3. Analytical curves obtained for: Ceramic tube: A=0.0062 + 0.0069 (Sn).Metallic tube: A=–0.0016 + 0.0032 (Sn); Pneumatic aspiration: A=0.016 + 0.0014 (Sn).

residues, thus decreasing analyticalcosts. This analytical procedure forthe determination of tin insolutions is easily operated and hasgood analytical frequency (90 h–1).The ceramic tube together with airas the sample carrier showed thebest results for use as an atomiza-tion cell, because it providedhigher sensitivity than the acety-lene-nitrous oxide flame.

ACKNOWLEDGMENTS

The authors thank the Fundaçãode Amparo à Pesquisa do Estado deSão Paulo (FAPESP) for financiallysupporting this work. The authorsare also grateful to ConselhoNacional de Desenvolvimento Cien-tífico e Tecnológico (CNPq) for fel-lowship to F.A.L.

Received August 4, 2006.

REFERENCES

1. S. Blunden and T. Wallace, FoodChem. Toxicol. 41, 1651 (2003).

2. A.P. Oliveira, J.A. Gomes Neto, M.Moraes, and E.C. Lima, At. Spec-trosc. 23, 190 (2002).

3. P. Cunniff (Ed.), Official Methods ofAnalysis of AOAC International,16th ed., Arlington, VA, USA:AOAC (1995).

4. F.A. Lobo, Desenvolvimento de sis-temas de injeção em fluxo para adeterminação de estanho porespectrometria de absorçãoatômica com forno aquecido nachama, em amostras alimentíciasenlatadas, Dissertação (Mestradoem Química) – Instituto deQuímica, Universidade EstadualPaulista, Araraquara, Brazil (2005).

5. J. Dëdina and D.L.A. Tsalev, HydrideGeneration Atomic AbsorptionSpectrometry, John Wiley & Sons,New York (1995).

6. F.A. Lobo, A.C. Villafranca, A.P.Oliveira, and M. Moraes, Ecl.Quím. 30 (2), 61 (2005).

7. A. Gaspar and H. Berndt,Spectrochim Acta - Part B. 55, 587(2000).

TABLE IAnalytical Characteristics for Tin Determination by TS-FF-AAS,

10–80 mg L–1 Range in 2% (v/v) HCl, Using Metallic and CeramicTubes as Atomization Cells, and FAAS-C2H2-N2O

Analytical TS-FF-AAS TS-FF-AAS FAASCharacteristics (metallic tube) (ceramic tube) C2H2-N2O

Relative Standard Deviation(rsd, n=12, 1.0 mg L–1) ≤ 2.7 ≤ 8.5 ≤ 13.3

Characteristic Concentration(mg L–1) 1.4 0.64 3.1

Analytical Frequency (h–1) 90 90 360

Analytical Equation A = – 0.0016 A= 0.0062 A = 0.016 + 0.0032 (Sn) + 0.0069 (Sn) + 0.0014 (Sn)

Correlation Coefficient (r) 0.9998 0.9993 0.9995

Sample Volume (µL) 50 50 ~1000

Carrier Flow Rate(mL min–1) 3.0 (carrier, air) 3.0 (carrier, air) 7.6 (nebulizer)

LOD (mg L–1) 1.7 0.8 9.6

23

Vol. 28(1), Jan./Feb. 2007

8. E.R. Pereira-Filho, H. Berndt, andM.A.Z. Arruda, J. Anal. At. Spec-trom. 17, 1308 (2002).

9. A. Gáspár, É. Széles, and H. Berndt,Anal. Bioanal. Chem. 372, 136(2002).

10. J. Davies and H. Berndt, Anal. Chim.Acta 479, 215 (2003).

11. C. Nascentes, M.A.Z. Arruda, A.R.A.Nogueira, and J.A. Nóbrega,Talanta 64, 912 (2004).

12. M.G. Pereira, E.R. Pereira-Filho, H.Berndt, and M.A.Z. Arruda, Spec-trochim. Acta, Part B, 59, 515(2004).

13. E. Gonzáles, R. Ahumada, and V.Medina, Quim. Nova 27, 873(2004).

14. C.R.T. Tarley and M.A.Z. Arruda,Anal. Sci. 20, 96 (2004).

15. A.C. Villafranca, M. Moraes, and J.Neira, Resumos: XVII CongressoLatino Americano de Química e27a Reunião Anual da SociedadeBrasileira de Química, QA 261,Salvador , BA, Brazil (2004).

16. E. Ivanova, A. Berndt, and E.Pulver-macher, J. Anal. At. Spectrom. 19,1507 (2004).

17. C.C. Nascentes, M.Y. Kamogawa,K.G. Fernandes, M.A.Z. Arruda,A.R A. Nogueira, and J.A. Nóbrega,Spectrochim. Acta, Part B, 60, 749(2005).

18. A.C. Villafranca, Avaliação da espec-trometria de absorção atômicacom nebulização térmica em tuboaquecido em chama (TS-FF-AAS)para a determinação de Cd, Cu, Pbe Zn em álcool combustível e óleodieseL, Tese (Doutorado emQuímica) – Instituto de Química,Universidade Estadual Paulista,Araraquara, Brazil (2004).

19. G.H. Jeffery; J. Basset, J. Mendham,and R.C. Denney. VOGEL - AnáliseQuímica Quantitativa, 5 ed.: LivrosTécnicos e Científicos Editora S A.,Rio de Janeiro, Brazil (1992).

20. E. Metcalfe, Atomic Absorption andEmission Spectroscopy, John Wiley& Sons, Chichester, England(1987).


*Corresponding author. E-mail: [email protected]: +91- 877–223259391

Diaminobenzidine-coated Silica Gel Used for Solid PhaseExtraction Preconcentration Separation of Cu(II) and

Cd(II) in Soil Samples Prior to Flame AAS DeterminationP. Sreevani, P. Mamatha, C. Sivani, P. Gopi Krishna, and *G.R.K. Naidu

Department of Environmental Sciences, S.V. University, Tirupati – 517 502, India

INTRODUCTION

Various techniques have beenemployed for the determination oftoxic heavy metals at trace levels incomplex real samples. Direct analy-sis of these samples is difficult asthey occur at low concentrations(near or below the limit of detec-tion of the instrument) and are usu-ally present in significant amountsin complex matrices (1). Suitablepreconcentration separation isrequired to address these problemsin the analysis of complex real sam-ples containing trace/ultratrace lev-els of toxic heavy metals (2).Among the various preconcentra-tion methods, solid phase extrac-tion (SPE) is preferred due to itssimplicity, rapidity, and ability toattain a high enrichment factor; inaddition, SPE is environmentallyfriendly. A number of sorbents suchas polyurethane foam, cellulose,activated carbon, silica gel, alumina,chelating, or neutral or ionexchange resins have been widelyused in SPE preconcentration sepa-ration (3-5). The application of suit-able inorganic supports has severaladvantages over organic supports,i.e., good selectivity, mechanicalstability, and rapid sorption ofmetal ions (6).

Silica gel is a very good adsor-bent since it does not swell, andhas good mechanical strength andthermal stability. However, formost metal ions, the interactionswith the silica surface are ratherweak because of the low acidity ofthe silanol groups (SI-OH) as well asthe less pronounced donor proper-

ties of the surface oxygen atoms.Therefore, silica gel modified withdifferent chelating groups findsincreasing application in differentareas of chemistry. The preconcen-tration separation proceduresdeveloped since 1990 for heavymetals using silica gel as sorbent inSPE are summarized in Table I(7–14). This paper discusses thedevelopment of off-line preconcen-tration separation of admixtures ofCu(II) and Cd(II) in soil samplesprior to flame atomic absorptionspectrometry (FAAS) quantification.

EXPERIMENTAL

Instrumentation

Measurements were performedwith a PerkinElmer® Model 2380flame atomic absorption spectrome-ter (PerkinElmer Life and AnalyticalSciences, Shelton, CT, USA) and theinstrument settings were accordingto the manufacturer’s recommenda-tions. A LI-120 digital pH meter(Elico, India) was used for the pHmeasurements.

Reagents and Standards

All reagents used in this studywere of analytical reagent grade.The certified soil standardreference material IAEA SOIL-7 wasobtained from the InternationalAtomic Energy Agency, Vienna,Austria. Standard solutions of cop-per and cadmium were prepared bydissolving appropriate amounts ofAR grade copper sulfate and cad-mium sulfate in doubly distilled anddeionized water (18 MΩ).

ABSTRACTA new chelating solid phase

extractant is described (preparedby coating 3,3'-diaminobenzidineonto silica gel) and used for pre-concentration separation ofadmixtures of Cu(II) and Cd(II)prior to their determination byflame atomic absorption spec-trometry (FAAS). The effect ofpH, weight of the solid phaseextractant, time of stirring, aque-ous phase volume, and nature ofeluent on the trace determinationof Cu(II) and Cd(II) was studiedand optimized.

The calibration plots wererectilinear over the concentrationrange of 0.5–100 µg L–1 for cop-per and 0.1–10 µg L–1 for cad-mium with detection limits of 1.0and 0.5 µg L–1, respectively.These values are 200 times lowerthan by the direct flame atomicabsorption spectrometry (FAAS)method after preconcentration.The precision of the developedprocedure is considered goodwith relative standard deviationvalues of 2.29% and 1.64% duringfive replicate determinations of10 and 4 µg L–1 copper and cad-mium, respectively. Accuracy ofthe developed procedure wastested by analyzing soil certifiedreference material IAEA Soil-7 forcopper. The developed methodwas found to be suitable for spe-ciative analysis of soil samples.The analytical results obtainedfor real soil samples in the deter-mination of copper and cadmiumare comparable to the valuesobtained using neutron activationanalysis.

25

Vol. 28(1), Jan./Feb. 2007

was separated from the solution byusing vacuum filtration (filterpaper). It was first filtered and thenwashed. The prepared silica gelwas added after addition of theextraction solution to the beaker.The preconcentrated Cu(II) andCd(II) ions were stripped in batchmode with a 10-mL–1 mixture of 1:1CH3OH and HNO3, then determinedby FAAS.

Procedure for Soil Analysis

Soil samples were collected fromareas around industrial sites inIndia. They were subjected to themodified sequential extraction pro-cedure as seen in Figure 1 (15,16).Sample fractions I–IV were subjectedto preconcentration separation byfollowing the general proceduredescribed above and subjected toFAAS determination of Cu(II) andCd(II).


Examination of MainExperimental Variables

Effect of pHThe effect of pH on the precon-

centration of 10 µg each of Cu(II)and Cd(II) present in 1 L of solutionwas studied. The enrichment ofCu(II) and Cd(II) was found to beconstant and reached maximum inthe pH range of 7.0 to 10.0 (for Cu)and 8 to 10 (for Cd) (see Figure 2).The decrease in percent enrichmentat a pH of less than 7.0 is due toinsufficient formation of the metalDAB complex and caused by theprotonation of DAB. Hence, insubsequent studies the pH wasadjusted to ~8.0 after the additionof 10 mL–1 of 1M NH3-NH4Cl buffer.

Preparation of Diaminobenzi-dine-coated Silica Gel

A 2-g amount of activated silicagel was added to 100 mL–1 deion-ized water in a 250-mL–1 beaker.Then, 0.2 g of 3,3'-diaminobenzi-dine (DAB), dissolved in a minimumamount of methanol, was addeddropwise to the above solution andstirred for 3 hours using a magneticstirrer at room temperature, filteredrepeatedly, rinsed with water, anddried at room temperature (27 oC).

General Procedure

Sample solutions of ~1 L contain-ing 10 µg each of Cu(II) and Cd(II)were adjusted to a pH of 8.0±1.0after the addition of 10 mL–1 of 1MNH3-NH4Cl buffer. Diaminobenzi-dine-coated silica gel (0.1 g) wasadded to the above solution andstirred with a magnetic stirrer for5 minutes. The prepared silica gel

TABLE ISummary of Preconcentration Separation Procedures Using Silica Gel Sorbent for Toxic Heavy Metals

(as reported since 1990)

S. No. Support and Functional Groups Sorbed Elements Application Detection Techniques Ref.

1. Silica gel-thiourea derivatives Pd(II) Separation Aqueous Spectrophotometry 7Pd(II) Nonferrous solutions

2 Silica gel-N,N,N’N’-tetra, Co(II), Ni(II), Natural water Spectrophotometer 8(2-aminoethyl) ethylenedimine Cu(II), Zn(II),

Cd(II),Pb(II)

3 Kieselgur and silica gel-3-methyl- Cu(II), Mn(II), Aqueous solution - --- 9phenyl-5-pyrazolone Cd (II), Pd(II)

4 Silica gel-1,8 dihydroxy anthraquinone Pb(II), Zn(II),Cd(II) Water samples AAS 10

5 Silica gel-N-propyl salicyl aldimine Cd(II), Cu(II), Natural waters AAS 11Cr(III,IV), Mn(II), Pb(II)

6 CPG-N-propylsalicyl-aldimine Al(III), Ag(I), Hg(II) Water samples ICP-MS 12

7 Silica gel-dihydroxy- benzene (DHB) Pb(II), Zn(II), Water samples, AAS 13Cd(II), Co(II), , synthetic certified Ni(II), Cu(II)Fe(III) drugs

8 5-Formyl-3-(1,-carboxyphenylazol) Cd(II), Zn(II), Natural aqueous AAS 14salicylic acid silica gel Fe(III), Cu(II), system

Pb(II), Mn(II), Cr(III), Co(II), Ni(II)

26

Fig. 1. Schematic diagram of sequential extraction procedure to determine extractable Cu(II) and Cd(II) ions in soils.

27

Effect of Weight of Solid PhaseExtraction Material

The weight of the diaminobenzi-dine-coated silica gel sorbentranged from 25 to 150 mg in a totalvolume of 1 L containing 10 µgeach of Cu(II) and Cd(II). Eventhough a minimum of 75 mg of sor-bent was enough for the quantita-tive enrichment of Cu(II) andCd(II) (see Table II), 100 mg ofDAB-coated silica gel was employedin the subsequent studies.

Effect of Preconcentration andAqueous Phase Volume

The preconcentration timeranged from 5 to 60 minutes duringthe preconcentration of 10 µg eachof Cu(II) and Cd(II) from 1 L ofsolution. The results obtained andlisted in Table II indicate that 5 min-utes of stirring was enough for thequantitative enrichment of Cu(II)and Cd(II) on the SPE sorbent. Apreconcentration time of 5 minuteswas chosen for all subsequentinvestigations. The effect of aque-ous phase volume on the precon-centration of 10 µg each of Cu(II)and Cd(II) was studied in the rangeof 25 to 1000 mL. The results inTable II indicate that quantitativepreconcentration of Cu(II) andCd(II) took place up to 1000 mL.

Effect of Elution Time, ElutionVolume, and Nature of Eluent

Various eluents (such as 1NHNO3 : methanol, 1:1 HNO3-CH3OH,and 1N HCl) were tested for theelution of previously sorbed Cu(II)and Cd(II). A 1:1 HNO3-CH3OHratio was found to be effective forthe elution of Cu(II) and Cd(II)from SPE (see Table II). Again, 5 minutes of elution time wasrequired for the quantitative recov-ery of previously sorbed Cu(II) andCd(II). Furthermore, as low as 5 mLof eluent was enough for the quan-titative recovery of Cu(II) andCd(II) sorbed on the modified DAB,resulting in an enrichment factorof ~200.

Linearity, Sensitivity, and Precision

Under optimal conditions, thecalibration curve was linear overthe concentration range of 0.5–100 and 0.1–10 µg L–1 forCu(II) and Cd(II), respectively. Fivereplicate determinations of 10 µgCu(II) and 4 µg Cd(II) present in 1 L of solution resulted in meanabsorbances of 0.13 and 0.088,respectively, with a relative stan-dard deviation of 2.29% and 1.64%,respectively. The detection limits,corresponding to 3 times the stan-

dard deviation of the blank, werefound to be 1.0 and 0.5 µg L–1 ofCu(II) and Cd(II), respectively. Thelinear equation with regression (R)was as follows:

ACu = 0.001 + 0.013 x C

ACd = -0.005 + 0.023 x C

where A is peak height absorbanceand C is concentration in µg L–1 ofaqueous solution. The correlationcoefficients were 0.99989 and0.99965 for Cu(II) and Cd(II),respectively. All statistical calcula-tions are based on the average oftriplicate readings for each standardsolution in the given range.

Effect of Neutral ElectrolytesSample solutions containing

10 µg each of Cu(II) and Cd(II) inthe presence of 0.1M amounts ofneutral electrolytes and metal ionscoexisting with the chosen analytewere determined by following thegeneral procedure described in theExperimental section. It was foundthat none of the species has anydeleterious effect on the SPE pre-concentration in conjunction withflame AAS determination of Cu(II)and Cd(II) (see Table III).

Retention Capacity

The retention capacity ofdiaminobenzidine-coated silica gelfor Cu(II) and Cd(II) was determinedby stirring increasing amounts(0.02 g to 2.0 g) of Cu(II) andCd(II) into 100 mL of solution con-taining 10 mL of 1M (pH 8.0)NH3-NH4Cl buffer and stirring for30 minutes. The amount of Cu(II)and Cd(II) sorbed on the SPE sor-bent was eluted with a mixture of1:1 HNO3 : methanol and determinedby FAAS. The retention capacity ofdiaminobenzidine-sorbed silica gelincreases with increasing initialCu(II) and Cd(II) concentrationsand reaches constant and maximumvalues of 40 mg/g and 28 mg/g ofSPE, respectively.

Vol. 28(1), Jan./Feb. 2007

Fig. 2. Effect of pH on the % enrichment of 10 µg each of Cu(II) and Cd(II) ontoDAB-coated silica gel (weight of SPE sorbent = 0.1 g, volume of aqueous phase =1000 mL).

28

TABLE IIOptimization of Experimental Parameters During

SPE Preconcentration and Elution of 10 µg each of Cu(II) and Cd(II) Present in 1000 mL of Deionized Water

(Weight of SPE Sorbent = 0.1 g; Preconcentration and Elution Times = 5 min; Eluent Volume = 5 mL of 1:1 CH3OH:HNO3)

S.No. % Enrichment Chosen Cu Cd Condition

1. Weight of Sorbent (mg) 100 mg25 mg 74.68 66.8650 mg 77.21 72.6275 mg 99.99 99.98100 mg >99.9 >99.9150 mg >99.9 >99.9

2. Preconcentration Time (min) 5 min5 >99.9 >99.910 >99.9 >99.960 >99.9 >99.9

3. Eluting Agent 1:1 HNO3:CH3OH1M HNO3 62.00 48.72CH3OH 3.79 6.321:1 HNO3:CH3OH >99.9 >99.9

4. Elution Time (min) 5 min5 >99.9 >99.910 >99.9 >99.960 >99.9 >99.9

5. Aqueous Phase Volume (mL) 1000 mL500 >99.9 >99.9750 >99.9 >99.91000 >99.9 >99.9

6. Eluent Volume (mL) 5 mL5 >99.9 >99.910 >99.9 >99.920 >99.9 >99.9

TABLE IIIEffect of Neutral Electrolytes and Coexisting Metal Ions

on the SPE Preconcentration FAAS Determination of 10 µg Each of Cu and Cd present in 1 L of Deionized Water

S. No. Interferents Remarks

1. NaCl, NaNO3, NaCl, NaClO4,

CaCl2, Mg(NO3)2 No interference at 0.1M level

2. Fe(III), Pb(II), Cr(III), Co(II),

Ni(II), Mn(II), Fe(II), Zn(II) No interference at 1000-µg level

Determination of Copper inCertified Reference Material

The accuracy of the developedpreconcentration procedure wasverified by analyzing the certifiedreference material IAEA Soil-7.

The soil samples were broughtinto solution by treating 0.5 g ofsample with 5 mL–1 HF and 1 mL–1

concentrated HNO3 at 150 °C. Theprocess was repeated twice. Theresidue was cooled, dissolved in50 mL–1 water, and diluted to 100 mL–1

with doubly distilled water. Precon-centration and determination wascarried out by following the proce-dure described in the Experimentalsection. The results listed in Table IVshow that the amount of copper inIAEA Soil-7 and determined by thepresent method agrees well withthe certified values. Furthermore,the recoveries of copper added tothe CRM prior to dissolution, pre-concentration, and determinationwere found to be good. This indi-cates the suitability of the developedpreconcentration procedure for thedetermination of copper in soilsamples.

Analysis of Real Soil Samples

Soil samples, collected in andaround the industrial areas of Tiru-pati, India, were brought into solu-tion by following the proceduredescribed above for the soil CRM.The samples were preconcentratedand Cu(II) and Cd(II) determinedby flame AAS. Table V lists theCu(II) and Cd(II) concentrationsfound in four soil samples using amodified sequential extraction pro-cedure (see Figure 1). The resultsobtained compare favorably withINAA results (Table V).

CONCLUSION

The solid phase extraction (SPE)sorbent material synthesized in thiswork via sorption of 3,3'-diamino-benzidine onto silica gel is easy toprepare in comparison to manyother sorbents. This material allows

29

a rapid (~5 minutes of preconcen-tration and stripping time) and reli-able off-line preconcentration oftraces of Cu(II) and Cd(II). In addi-tion to a 100-fold enrichment, thisprocedure is useful in the separa-tion of traces of Cu(II) and Cd(II)from 0.1M levels of various neutralelectrolytes and 1000 µg of coexist-ing metal ions. Moreover, theimmobilized material offers a reten-tion capacity of 40 and 28 mg ofCu(II) and Cd(II) per gram, respec-tively. The accuracy of this precon-centration procedure was verifiedby determining total Cu(II) in certi-fied soil reference material IAEASoil-7. Unlike most preconcentra-tion procedures used for toxicheavy metals, the present enrich-

ment method allows for a rapid andreliable speciation of Cu(II) andCd(II) and their determination insoils by the simple and routinelyavailable flame atomic absorptionspectrometry technique.

Received September 5, 2006.

REFERENCES

1. A. Mizuike, Enrichment Tech-niques in Trace Analysis, SpringerVerlag, Berlin, Germany (1983).

2. P.K. Tewari and A.K. Sing,Talanta 56, 735 (2002).

3. J.S. Fritz, Introduction andPrinciples, Analytical Solid PhaseExtraction, First Edition, Wiley-VCH, New York (1999).

4. V. Camel, Spectrochim. Acta B58, 1177 (2003).

5. T. Prasada Rao and R. Kala,Talanta 63, 949 (2004).

6. N.V. Deorkar and L.L. Taval-rides, Ind. Eng. Chem. Res. 36,399(1997).

7. V.V. Skopenko, A.K. Trofim-chuk, and E.S. Janovskaya, Ukr.Khim. Zh. 50, 49 (1993).

8. D. Chambaz and W. Haerdi,J. Chromatogr. 600, 203 (1992).

9. O. Todorova, P.Vassileva, andL. Lakov, Analytical and Bioanalyti-cal Chemistry 346, 943 (1993).

10. K.S. Abou El Shirbini, I.M.M.Kenawy, M.A. Hamed, R.M. Issa,and R. El-Morsi, Talanta 58, 289(2002).

11. A.Goswami and A .K. Singh,Talanta 58, 669 (2002).

12. G.A.E. Mostafa, M.M. Hassa-neen, K.S. Abu El Shirbini, and V.Gorlitz, Anal. Sci. 19, 1151 (2003).

13. G. Venkatesh, A.K. Singh,and B. Venkataraman, Microchim.Acta 144, 233 (2004).

14. M.A.A. Akl, I.M.M. Kenawy,and R.R. Lasheen, Microchem. J. 78,143 (2004).

15. B. Marin, M. Valladen, M.Polve, and A. Monaco, Anal. Chim.Acta 342, 91 (1997).

16. K. Prasad, P. Gopikrishna, R.Kala, T. Prasada Rao, and G.R.K.Naidu, Talanta 69, 938 (2006).

Vol. 28(1), Jan./Feb. 2007

TABLE IVAnalysis of Certified Reference Material

Sample Cu Added Cu Presenta Certified Recovery (µg/g) (µg/g) Cu Value (µg/g) (%)

IAEA Soil-7b – 11.5±2.0 11.0±2.0 –10.0 21.5±2.5 100

20.0 31.5±3.0 100

a Average of 3 determinations.b Supplied by M/S. International Atomic Energy Agency (IAEA), Vienna, Austria.

TABLE VSpeciative Determination of Cu(II) and Cd(II) in Soil Samples

After Enrichment in Ciaminobenzidine-sorbed Silica Gel and Flame AAS Determination

Soil Sample Element Cu(II) and Cd(II) (µg/g)F I F II F III F IV Total NAA

1. Mango Garden Cu 16.92 153.81 54.23 38.23 263.19 269.81“ Cd N.D. N.D. N.D. N.D. N.D. 0.52

2. Near Madhu Cu 1.30 1.31 3.20 N.D. 5.81 5.90Industries Cd 2.16 N.D. 1.93 N.D. 4.09 4.12

3. SV Sugar Factory Cu 14.15 17.20 10.21 0.53 42.09 42.81G.Mandyam Cd N.D. N.D. 1.32 0.51 1.83 1.85

4. Pioneer Alloy Cu 0.40 0.41 4.56 N.D. 5.37 5.52Casting Ltd. Cd N.D. N.D. 0.31 2.12 2.43 2.45

5. Fertile Land Cu 0.48 1.20 3.31 N.D. 4.99 5.10“ Cd 0.78 N.D N.D 0.80 1.58 1.60

6. Amara Raja Cu 0.42 4.91 1.23 1.89 8.39 8.40

Batteries Cd N.D. N.D. 0.29 6.58 6.87 6.91

F = Fraction. N.D. = Not detectable.


Determination of Trace Amounts of Silver in Various Samples by Electrothermal Atomic Absorption

Spectrometry After Sample Preparation Using Cloud Point Extraction

*Zhefeng Fan and Fang BaiDepartment of Chemistry, Shanxi Normal University, Linfen 041004, P.R. China

INTRODUCTION

Silver is a metal of commercialimportance. It is used in electronicdevices, photographic material,mirrors, medicine, and jewelry(1–3). The increasing use of silverhas, therefore, also resulted in anincreased silver content in biologi-cal and environmental samples.Recent information about the inter-action of silver with essential nutri-ents, especially selenium, VitaminsE and B12, have focused attentionon its potential toxicity (4,5).Hence, a reliable and efficient ana-lytical technique is required forstudying and evaluating its potentialtoxicity.

Several analytical techniques,such as inductively coupled plasmaatomic emission spectrometry (6),inductively couple plasma massspectrometry (7,8), and atomicabsorption spectrometry (9,10)were applied for the determinationof silver in various samples. How-ever, these methods suffer fromsevere matrix effects, especially atthe ng g-1 or lower concentrationlevels of analytes in biological andenvironmental samples. Therefore,for the ultratrace determination ofsilver, a separation and preconcen-tration step is often necessary priorto analysis. The classical liquid-liq-uid extraction methods (11) areusually time-consuming and labor-intensive and require relativelylarge volumes of high-puritysolvents.

than extractions that use organicsolvents, and has a high capacity toconcentrate a wide variety of ana-lytes of varying nature with highrecoveries and high concentrationfactors. It has been used for theseparation and preconcentration oftrace metals after the formation ofwater-soluble complexes (12–19).Cloud point extraction is based onthe property of most non-ionic sur-factants in aqueous solutions toform micelles and become turbidwhen heated to the same tempera-ture as the cloud point temperature.Above the cloud point, the micellarsolution separates into a surfactant-rich phase of a small volume andinto a diluted aqueous phase. Anyhydrophobic species, such asmetallic chelates, remain preferen-tially in the surfactant-rich phase,which are then extracted or pre-concentrated.

Recently, Jamshid et al. (20)developed a selective CPE methodand preconcentration of traceamounts of silver as a dithizonecomplex. The major problem withthe use of dithizone in CPE is thelack of conventional solvents forcomplete dissolution. O,O-diethyl-dithiophosphate (DDTP), whichhas a S atom as electron donor,behaves like a soft base, and shouldcomplex with silver that has a softacid character. In addition, DDTP isadvantageous because it formsquite stable chelates even in an acidmedium, which is the mediumresulting from most of the sampledissolution procedures (21,22).

The aim of the present work wasto apply CPE as a separation andpreconcentration step for silver*Corresponding author.

E-mail: [email protected]

Cloud point extraction (CPE), asan effective separation and precon-centration technique, has beendemonstrated to have the distinctmerits of a low-cost, simple, andspeedy procedure. This method isof lower toxicity to the environment

ABSTRACT

O,O-diethyl-dithiophosphate(DDTP) was used as a complex-ing agent in cloud point extrac-tion (CPE) and applied for theselective preconcentration oftrace amounts of silver. The ana-lytes in the initial aqueous solu-tion were acidified with acid andTriton X-114 was added as a sur-factant. The enriched analyte wasdetermined by electrothermalatomic absorption spectrometry(ETAAS). The parameters affect-ing the extraction efficiency,such as solution pH, concentra-tion of DDTP and Triton X-114,equilibration temperature andtime, were investigated in detail.Under the optimum conditions,a preconcentration factor of 35was obtained for the preconcen-tration of 10 mL of sample solu-tion. The limit of detection was7.8 ng L–1 and the relative stan-dard deviation (RSD) was 4.9%(n=7, c= 0.2 ng mL–1).

The proposed method wasapplied to the determination ofsilver in water, serum, andhuman hair samples. The recov-eries for the spiked samples were92.4–98%. To verify the accuracyof the method, silver was deter-mined in certified referencematerials, and the determinedvalues are in good agreementwith the certified values.

31

Vol. 28(1), Jan./Feb.. 2007

prior to electrothermal atomicabsorption spectrometric (ETAAS)determination. In the proposedmethod, DDTP was used as thechelating agent and Triton® X-114as the surfactant. Potential factorsaffecting the CPE separation andpreconcentration of silver wereinvestigated in detail. Thedeveloped method was applied tothe determination of silver in vari-ous samples.

EXPERIMENTAL

Instrumentation

A Model TAS-990 atomic absorp-tion spectrometer (Beijing PuxiInstrument Factory, Beijing, P.R.China), equipped with deuteriumlamp background correction and atransversely heated graphite atom-izer, was used as the detection unit.Pyrolytically coated graphite tubes(Beijing Puxi Instrument Factory,Beijing, P.R. China) were usedthroughout. A silver hollow cath-ode lamp was employed as the radi-ation source at 328.1 nm with a slitwidth of 0.4 nm. Argon 99.99% wasused as the protective and purgegas. The integrated absorbancemode was used throughout. A ther-mostated water bath, maintained atthe desired temperature, was usedfor the cloud point preconcentra-tion experiments; phase separationwas performed using a centrifugewith 15-mL centrifuge tubes. Thegraphite furnace temperature pro-gram for the determination of silveris given in Table I.

Standard Solutions andReagents

Stock standard solutions of 1.0 mg mL–1 silver were preparedby dissolving 0.1000 g silver in 100 mL 0.1 mol L–1 nitric acid.Working standard solutions wereobtained by appropriate dilution ofthe stock standard solution. Thenon-ionic surfactant Triton X-114was obtained from Sigma (St. Louis,MO, USA) and was used without

further purification. A solution of5.0% (m/v) DDTP (Aldrich, Milwau-kee, WI, USA) was prepared by dis-solving appropriate amounts of thisreagent in water. Doubly deionizedwater was used throughout thiswork. All reagents used were of atleast analytical grade. The certifiedreference materials used in thisstudy were GBW 09101 HumanHair, GBW 07403 Soil, and GBW07108 Rock (obtained from theNational Institute of Standards andTechnology, Beijing, P.R. China).

Sample Collection and Preparation

Water samples were collectedfrom Fen River, Linfen, P.R. China.The water samples were filteredthrough a 0.45-µm membrane filter(Tianjin, Jinteng InstrumentFactory, Tianjin, P.R. China), andacidified to a pH of about 1 withnitric acid prior to storage and forsubsequent use.

Portions of 0.2500 g of soil androck were weighed and transferredinto Teflon beakers with about 10 mL concentrated nitric acid and5 mL HF; then boiled for about 1 hour until the samples were com-pletely decomposed. After the solu-tions were heated to near dryness,the residue was dissolved with 0.15 mol L–1 of nitric acid.

Portions of 0.2500 g of humanhair and serum were weighed andtransferred into 10-mL PTFEbeakers. Then 2.0 mL of concen-trated nitric acid was added, and

the solution kept standingovernight. Then, 2.0 mL of H2O2

was added to the samples and thesolution heated to near dryness.The residue was dissolved in0.15 mol L–1 nitric acid.

Analytical Procedures

For cloud point extraction,aliquots of 10 mL of the sample orstandard solution containing theanalyte, 0.5% (m/v) DDTP, and0.2% (v/v) Triton X-114 were keptin a thermostated water bath at 45 oC for 15 minutes. Separationof the aqueous and surfactant-richphase was accomplished by cen-trifugation for 10 minutes at 3500 rpm. After cooling in an icebath, the surfactant-rich phasebecomes viscous. The supernatantaqueous phase was then separatedcompletely with a syringe centeredin the tube. To decrease the viscos-ity of the surfactant-rich phase andto facilitate sample handling, 100µL of methanol was added to thesurfactant-rich phase, and 20 µL of the resulting solution was intro-duced into the graphite furnace. A blank submitted to the same pro-cedure was always measured in parallel to the samples.


Effect of Acid Concentration

The complexation of most metalions with DDTP was found to bemore favorable in hydrochloric acidmedium than in nitric acid. How-ever, the presence of hydrochloric

TABLE IGraphite Furnace Temperature Program

for the Determination of Silver

Step Temp. Ramp Time Hold Time Ar Flow(oC) (s) (s) (mL min–1)

Drying I 80 10 10 250Drying II 120 10 10 250Pyrolysis 600 5 15 250Atomization 2500 0 3 0

Furnace Cleaning 2600 1 2 250

32

acid might cause the precipitationof silver. Therefore, nitric acid waschosen to adjust the acidity of thesample solutions. To assureefficient analyte extraction, theeffect of the nitric acid concentra-tion on the integrated absorbancesignal of silver was evaluated overthe concentration range of 0.001 to2.0 mol L–1. The results indicatethat the integrated absorbance sig-nal of silver increased as the con-centration of nitric acid increasedfrom 0.02 to 0.1 mol L–1 andremained constant between 0.1 and 0.2 mol L–1. In this work, 0.15 mol L–1 of nitric acid wasemployed.

Effect of DDTP Concentration

Cloud point extraction efficiencydepends on the hydrophobicity ofthe ligand and the complex formed,the apparent equilibrium constantsin the micelle medium, the kineticsof the complex formation, and thetransference between the phases.The effect of DDTP concentrationon the integrated absorbance signalof silver is shown in Figure 1. Ascan be seen, the integratedabsorbance signal of silver washigher at lower DDTP concentra-

tions, indicating that silver is moreefficiently extracted at these con-centrations. The complex of silverand DDTP has a 1:1 composition.For higher DDTP concentrations, acharged complex such as ML–

2 canform and decrease the extraction,while an uncharged complex ismore easily extracted (23). Thus,0.025% (m/v) of DDTP was chosenfor all further experiments.

Effect of Triton X-114 Concen-tration

Triton X-114 was chosen for theformation of the surfactant-richphase due to its low cloud pointtemperature and high density of thesurfactant-rich phase, which facili-tates phase separation by centrifu-gation. The effect of Triton X-114concentration in the range of 0.025to 0.2% (v/v) was investigated per-taining to the integrated absorbancesignal of silver; the results areshown in Figure 2. As can be seen,the integrated absorbance signal ofsilver increased from 0.025 to 0.05%(v/v) and remained constantbetween 0.05 and 0.1% (v/v). Thus,0.075% (v/v) of Triton X-114 wasemployed throughout this study.

Effect of Equilibration Temperature and Time

Optimal equilibration tempera-ture and incubation time are neces-sary to complete reactions and toachieve easy phase separation anda preconcentration as efficiently aspossible. The effect of equilibrationtemperature on the integratedabsorbance signal of silver wasinvestigated from 25 to 60 oC. Itwas found that the integratedabsorbance signal increased withan increase in equilibration temper-ature from 25 to 35 oC, and reachedmaximum between 35–60 oC. Forfurther study, an equilibration tem-perature of 45 oC was used. Thedependence of extractionefficiency on incubation time wasstudied in the range of 5 to 20 min-utes. An incubation time of 10 min-utes was found to be optimal forextraction.

Analytical Performance

A calibration curve wasconstructed by preconcentrating10 mL of sample standard solutionwith Triton X-114 and the graphs ofabsorbance versus silver concentra-tion were linear over the concentra-tion range of 0.08 to 0.6 ng mL-1 of

Fig. 1. Effect of concentration of DDTP on the cloud pointextraction of 0.2 ng mL–1 silver. Other cloud point extractionconditions: 0.075% (m/v) Triton X-114, equilibration temper-ature 45 oC, incubation time 10 minutes.

Fig. 2. Effect of concentration of Triton X-114 on the cloudpoint extraction of 0.2 ng mL–1 silver. Other cloud pointextraction conditions: 0.025% (m/v) DDTP, equilibrationtemperature 45 oC, incubation time 10 minutes.

33

Vol. 28(1), Jan./Feb.. 2007

tions, the detection limit of the present CPE-ETAAS method for thedetermination of silver was 7.8 ng L–1 which is lower thanmany previously published reports(24).

Received September 5, 2006.

REFERENCES

1. I.M. Kolthoff and P.J. Eiving, Treatiseon Analytical Chemistry,Interscience, New York (1996).

2. H. Renner, "Ulmanns Encyklopedieder Technischen Chemie," VerlagChemie, Weinheim, Germany(1982).

3. N.N. Greenwood and A. Earnshow,"Chemistry of the Elements," Perga-mon Press, New York (1989).

4. Environmental Protection Agency(EPA), Ambient Water Quality Cri-teria for Silver, EPA 4405-80-071,Office of Water Regulations, Wash-ington, D.C., USA (1980).

5. E. Meian, "Metals and TheirCompounds in the Environment,"VCH, New York (1991).

6. R.P. Singh and E.R. Pambid, Analyst115, 301 (1990).

7. M. Rehkampor and A. N. Halliday,Talanta 44, 663 (1997).

8. H. Jingyu, L. Zheng, and W. Haizhou,Anal. Chim. Acta 45, 329 (2002).

9. P. Anderson, C.M. Davidson, D. Littlejohn, M.A. Ure, C.A. Shand,and M.V. Cheshive, Anal. Chim.Acta 327, 53 (1996).

10. S. Dadfarnia, A.M.H. Shabani, andM. Gohari, Talanta 64, 682 (2004).

11. T. Koh and T. Sugimoto, Anal.Chim. Acta 333, 167 (1996).

12. C. D. Stalikas, Trends Anal. Chem.21, 343 (2002).

13. J. L. Manzoori and G. K. Nezhad,Anal. Sci. 19, 579 (2003).

14. M.A.M.S. da Veiga, V.L.A. Frescura,and A.J. Curtius, Spectrochim. Acta

results are in good agreement withthe certified values; no significantdifferences were observed (t-test,P=0.05).

Application to Real Samples

The proposed method was alsoemployed for the determination ofsilver in water, serum, and humanhair samples. The analytical resultsare given in Table III. The recover-ies for the spiked samples werefound to be in the acceptable range(92.4–98.0%).

CONCLUSION

In this work, an effective samplepreparation technique using CPEfor the determination of silver isdeveloped. The methodology issimple, sensitive, and inexpensive.It provides good enrichment factorsand an efficient separation and pre-concentration, which enables thedetermination of ultra-low levels ofsilver in complex matrices. Underthe optimized experimental condi-

silver. The equation of the calibra-tion graphs was A=0.4337C +0.00766 (where A is the absorbanceand C is the concentration) with acorrelation coefficient of 0.99987.The limit of detection (LOD) was7.8 ng L–1, calculated as the ratiobetween three times the standarddeviation of seven blank readingsand the slope of the calibrationcurve after preconcentration. Therelative standard deviation was4.9% (n=7, c=0.2 ng mL–1) and anenrichment factor of about 35obtained by preconcentrating a 10-mL sample volume (calculatedas the ratio of the slope of precon-centrated samples to that obtainedwithout preconcentration).

Validation of the Method

To verify the accuracy of themethod, silver was determined incertified reference materials GBW09101 Human Hair, GBW 07403Soil, and GBW 07108 Rock; the ana-lytical results are given in Table II.As can be seen, the analytical

TABLE IIAnalytical Results (Mean ±SD, n=3) for Silver in Certified Samples

Sample Certified Value Determined Value (µg g–1) (µg g–1)

GBW 09101 Human Hair 0.35±0.02 0.33±0.07

GBW 07403 Soil 0.091±0.011 0.087±0.009

GBW 07108 Rock 0.043±0.017 0.040±0.016

TABLE IIIAnalytical Results (Mean ±SD, n=3) for Determination of Silver

in Water and Biological Samples

Sample Added Found Recovery

(ng mL–1) (ng mL–1) (%)River Water 0 0.090±0.005 ---

0.2 0.286±0.013 98.0Serum 0 18.5±0.33 ---

25 42.3±4.3 95.2Human Hair (µg g–1) (µg g–1) (%)

0 0.0746±0.0023 ---

0.1 0.167±0.018 92.4

Results of recoveries of spiked samples.

34

17. J. Chen and K.C. Teo, Anal. Chim.Acta 450, 215 (2001).

18. K.C. Teo and J. Chen, Anal. Chim.Acta 434, 325 (2001).

19. J.L. Manzoori and A.B. Tabrizi, Anal.Chim. Acta 470, 215 (2002).

20. J.L. Manzoori and G.K. Nezhad,Anal. Chim. Acta 484, 155 (2003).

21. V.L. Dressler, D. Pozebon, and A.J.Curtius, Anal. Chim. Acta 379, 175(1999).

22. V.L. Dressler, D. Pozebon, A.J. Curtius, Spectrochim. Acta53B, 1527 (1998).

23. M.A.M. da Silva, V.L.A. Frescura,and A.J. Curtius, Spectrochim. Acta55B, 803 (2000).

24. S. Dadfarnia, A.M. Haji Shabani, andM.Gohari, Talanta 64, 682 (2004).

35

*Corresponding author.E-mail: [email protected]: +86-22-23506075

Flow Injection On-line Solid Phase Extraction Using Multi-Walled Carbon Nanotubes as Sorbent for

Cold Vapor Atomic Fluorescence SpectrometricDetermination of Trace Mercury in Water Samples

*Xiao-Hong ShangResearch Center for Analytical Sciences, College of Chemistry, Nankai University,

Tianjin 300071, P.R. China

Atomic SpectroscopyVol. 28(1), Jan./Feb. 2007

INTRODUCTION

Mercury is a heavy metal pollu-tant and is of great concernbecause of its toxicity, and itscumulative and persistent characterin the environment and biota. Thedetermination of mercury in theenvironment is very important forgeneral and control purposes. Thedevelopment of reliable methodsfor the determination of mercury attrace levels in environmental andbiological samples is of particularsignificance.

Cold vapor atomic fluorescencespectrometry (CV–AFS) is a power-ful method for mercury determina-tion owing to the high sensitivity,cost-effectiveness, and ease of theinstrumental operation (1–3). Itpermits efficient separation andpreconcentration of the analytesfrom the sample matrices, therebyreducing or eliminating the poten-tial chemical and/or spectral inter-ferences commonly encounteredwith direct solution analysis and bysignificantly increasing the signal ofmercury. Because mercury oftencannot be determined directly andaccurately due to its low concentra-tion levels in natural waters, a sam-ple preconcentration step prior toits determination is usually indis-pensable. Solid phase extraction(SPE) is the most promising methodfor this purpose. It has many advan-tages including a high enrichmentfactor, high recovery, short extrac-tion time, low analytical and instru-

mental cost, and low consumptionof solvents. Various types of materi-als have been utilized as SPE sor-bents for preconcentration ofmercury including C18 silica gel(4,5), immobilized anion exchangeresin (6), cigarette filter (7–9), andthe PTFE knotted reactor (10,11).Although a detection limit as low as0.07 ng L-1 was obtained with a2–mercaptobenzimidazol silica gelsorbent, an extremely toxic reagentof KCN had to be employed (3).

Recently, carbon nanotubes(CNTs), which were firstintroduced in 1991 by Iijima (12),

have attracted enormous interestdue to their exceptional mechani-cal and electrical properties, as wellas high chemical and thermal stabil-ity. CNTs can be described as asheet of graphite rolled into a tubeand divided into single–walled car-bon nanotubes (SWCNTs) ormulti–walled carbon nanotubes(MWCNTs) depending on the dif-ferent carbon atom layers in thewall of the nanotube. CNTs havebeen widely used in many fieldsincluding for hydrogen storage (13)and field emission (14), as well asfor catalyst supports (15) andchemical sensors (16). CNTs havealso found great potential applica-tion as sorbents in pollutantremoval owing to their large spe-cific surface area, and hollow andlayered structure. MWCNTs werefound to be a superior adsorptionmaterial for removing dioxin (17).The strong binding ability of dioxinto MWCNTs is attributed to theinteraction of two benzene ringsof dioxin and the surface of theMWCNTs. Further researchextended the applications to vari-ous contaminant removals anddetection such as sulfonylurea her-bicides (18), dichlorobenzenes(19), polycyclic aromatic hydrocar-bons (20), and metals including Zn(21), Pb (22), Cd (23), Am (24), andthe rare earth elements (25). CNTshave also been employed as SPEsorbents for preconcentration ofpollutants such as phenols (26),chlorobenzenes (27), and somemetals (28, 29). Although a numberof reports have discussed thepotential application of MWCNTsas promising sorbents for environ-mental protection studies, their use

ABSTRACT

A sensitive, simple, and rapidmethod was developed for thedetermination of trace dissolvedinorganic mercury in environ-mental water samples by on–linecoupling solid phase extraction(SPE) to cold vapor atomic fluo-rescence spectrometry (CV–AFS)using multi–walled carbon nan-otubes (MWCNTs) as sorbent.Potential factors affecting the SPEprocedure and cold vapor gener-ation as well as possible interfer-ences were investigated in detail.With a sample loading flow rateof 3.3 mL min–1 and a 120-secondpreconcentration time, anenhancement factor of 39 withan adsorption efficiency of 62%was obtained. The detection limit(3 σ) was 1.2 ng L–1 and the rela-tive standard deviation (RSD%,n=7) was 4.3% at the 0.1-µg L–1

Hg(II) level. The developedmethod was applied to the deter-mination of trace dissolved inor-ganic mercury in local watersamples.

36

was purchased from the NationalResearch Center for Certified Refer-ence Materials (NRCCRM, Beijing,P.R. China). Working standard solu-tions were freshly prepared by step-wise dilution of the stock solution.Doubly deionized water (DDW,18.2 MΩ cm–1) was obtained byusing a WaterPro water purificationsystem (Labconco Corporation,Kansas City, MO, USA).

All glassware was soaked in a10% (v/v) nitric acid solution for atleast 24 hours and rinsed withDDW before use.

SPE Microcolumn Preparation

A 6–mg amount of MWCNTs wasintroduced into a PTFE microcol-umn (6×3.0 mm i.d.), which wasplugged with a small portion ofglass wool at both ends. The newlypacked column was sequentiallywashed with the eluent and waterfor cleaning and conditioning.

Sample Pretreatment

Three river water, two lakewater, and one tap water samplewas collected locally. Immediatelyafter sampling, the samples werefiltered through a 0.45-µm pore sizemembrane filter, adjusted to a pHof 6–7 with 2 mol L–1 HNO3, andanalyzed.

Procedures

The flow injection (FI) manifoldand its operational sequenceincludes two steps for the on–lineSPE. The operational sequences foron-line SPE preconcentration forCV-AFS determination of mercuryare listed in Table II. In step one,the injection valve was in the fillposition, and pump A was activatedso that the sample solution wasloaded onto the microcolumn. Instep two, pump B started to work,whereas pump A was stopped, and

as SPE packing material for on–linepreconcentration of toxic metals isstill limited in the literature.

The aim of this work was toassess the feasibility of MWCNTs asan adsorbent in an on-line systemfor CV-AFS determination of tracedissolved inorganic mercury in nat-ural water samples. The methodwas applied to the analysis of a cer-tified reference material and theresults showed good agreementwith the certified values.

EXPERIMENTAL

Instrumentation

A Model AFS–820 non-dispersiveatomic fluorescence spectrometer(Beijing Titan Instruments, Beijing,P.R. China), as used in previousworks (30–32), was employedthroughout. A high–intensity Hghollow cathode lamp (Beijing Insti-tute of Vacuum Electronics, Beijing,P.R. China) was used as the radia-tion source. The atomizer was anelectrically heated quartz tube(6-mm i.d., 65-mm length) intowhich the volatile species gener-ated in the reactor were swept outby an argon–carrier gas. A home-made gas–liquid separator was usedto isolate the gas from the liquid.The instrumental conditions aresummarized in Table I. A ModelFIA–3100 flow injection analyzer(Vital Instrument Co. Ltd., Beijing,P.R. China) was used for on–linepreconcentration of mercury.Tygon® peristaltic pump tubing wasemployed to propel the sample,eluent, and reagents. A minimumlength of PTFE tubing with a 0.5-mmi.d. was used for all connections.

Standards and Reagents

The MWCNTs used as the SPEsorbent were 40-60 nm in diameterand 5–15 µm in length (L-MWNT-4060, Shenzhen Nanotech Port Co.,Shenzhen, P.R. China). All reagentswere at least of analytical grade.Hg(II) stock solution (1000 mg L–1)

TABLE IExperimental Conditions for CV–AFS

Parameters Setting

Lamp Hg Hollow Cathode LampLamp Current 40 mAQuartz Furnace Height 11cmNegative High Voltage of Photomultiplier 290 VFlow Rate of Carrier Liquid 0.7 mL min–1

Flow Rate of 0.1% (m/v) KBH4 0.7 mL min–1

Carrier Gas (Argon) Flow Rate 400 mL min–1

Shield Gas (Argon) Flow Rate 900 mL min–1

Table IIOperational Sequences for the On–line SPE Preconcentration System

and CV–AFS Determination of Mercury

Function Time Pumped Flow Rate (s) Medium Pump Pump Valve

A B Position

1 Sample Loading 120 Sample 3.3 Off Fill

2 Analyte Elution 60 0.5 mol L–1HCl – Off 0.7 Injection1.5 mol L–1 HNO3 – 0.2%(m/v) Na2S2O80.1% (m/v) KBH4 0.7

Flow rate: (mL min–1)

37

Vol. 28(1), Jan./Feb.. 2007

the injection valve was turned tothe inject position to introduce aneluent solution of 0.5 mol L–1

HCl–1.5 mol L–1 HNO3–0.2% (m/v)Na2S2O8. The eluent also providesthe required acidic medium for sub-sequent cold vapor generation.Meanwhile, a flow of 0.1% (m/v)KBH4 solution was introduced tomerge with the eluent solution con-taining the eluted analyte to gener-ate mercury vapor just beforeentering the gas-liquid separator.The generated gaseous mixture wastransported to the atomizer with anargon flow at 400 mL min–1 foratomic fluorescence spectrometry(AFS) detection. The total timerequired for a single determinationwas 180 seconds.


Effect of Sample Acidity

Sample acidity plays an impor-tant role with respect to the adsorp-tion of metal ions on MWNTs.Figure 1 shows the effect of sampleacidity on the preconcentration of0.1 µg L–1 Hg(II) in the pH rangingfrom 2.5–8.5 using a citric acid–disodium hydrogen phosphate(C6H8O7–Na2HPO4) buffer solution.The fluorescence intensity of mer-cury increased from a pH of2.5–6.0, reached maximum in the

120 seconds was used consideringthe total time of the procedure.

Selection of Eluent

It is of great importance tochoose a suitable eluent for SPEperformance. When online SPE iscoupled to CV–AFS, a proper eluentshould not only have a sufficientlystrong elution ability, but also facili-tate the ensuing cold vapor genera-tion. Diluted hydrochloric acid ornitric acid is generally chosen foron–line elution of the retained ana-lyte from the column because theyare a suitable medium forsubsequent chemical vapor genera-tion. In this work it was found thata mixture of HCl and HNO3 waspreferable to using only one of theacids. For this work, the concentra-tion and volume ratio of HCl andHNO3 were optimized.

As shown in Figure 4, the fluo-rescence intensity increased as thetotal concentration of the mixedacid increased from 0.5 mol L–1 to2.0 mol L–1, then decreased slightlywith a further increase of the totalconcentration to 2.5 mol L–1. A totalconcentration of 2.0 mol L–1 mixedacid was used in this work.

Figure 5 shows the influence ofvolume ratio of 2.0 mol L–1 HNO3

to 2.0 mol L–1 HCl on the elution

pH range of 6.0–7.0, and decreasedfrom pH 7.0–8.5. The low mercuryadsorption at a low pH can beattributed in part to a competitionbetween the proton and Hg(II) onthe same sites. The decrease ofmercury adsorption at pH >8 maybe due to metal hydrolysis. Thus, inthis work, the pH of the samplesolutions was adjusted to 6.0–7.0before preconcentration.

Effects of Sample Loading FlowRate and Loading Time

The effects of sample loadingflow rate and sample loading timeon the fluorescence intensity ofmercury were investigated with astandard solution of 0.1 µg L–1. Asshown in Figure 2, the fluorescenceintensity increased linearly as thesample loading flow rate increasedfrom 1.8 to 3.5 mL min–1, and thenincreased slowly with a furtherincrease of the sample loading flowrate up to 3.8 mL min–1. A flow rateof 3.3 mL min–1 was selected forthis work.

The effect of sample loadingtime is shown in Figure 3. The fluo-rescence intensity increases linearlywith a sample loading time from 1 to 6 minutes, and then levels offwith a further increase of the sam-ple loading time. For further experi-ments, a sample loading time of

Fig. 1. Effect of acidity on the preconcentration of 0.1 µg L–1

Hg(II). All other conditions as in Tables I and II.

Fig. 2. The influence of sample loading flow rate on the deter-mination of 0.1 µg L–1 Hg(II). All other conditions as inTables I and II.

38

efficiency. The fluorescence inten-sity kept almost constant at therange of 3:1 to 5:1; and 3:1 wasselected for this work. Therefore,the concentration of HCl and HNO3

was 0.5 mol L–1 and 1.5 mol L–1,respectively.

It was also found that the addi-tion of Na2S2O8 in the eluent fur-ther increased the fluorescenceintensity. The possible reason wasa strong interaction between Hg(II)and MWCNTs besides a physicalattachment, which can be impairedby the strong oxidant Na2S2O8. Dif-ferent concentrations of Na2S2O8 inthe mixed acid were optimized toobtain high elution efficiency for

of the acid medium and KBH4 werestudied. As mentioned above, theacid medium of AFS was the sameas the eluent for on–line SPE. TheKBH4 solution served as the reduc-tant for cold vapor generation.Higher concentrations of KBH4

(>1% m/v) would cause seriouseffervescence and splashing of solu-tion droplets on the walls of thegas–liquid separator due to the fastreaction, resulting in a decrease insignal intensity. Lower concentra-tions of KBH4 (<0.02%, m/v) didnot completely reduce the analyte.The appropriate concentration ofKBH4 selected for this study was0.1% (m/v).

Hg(II). As shown in Figure 6, thebest concentration range forNa2S2O8 was from 0.2%–1.0%(m/v), which doubly increased thefluorescence intensity. Further addi-tion of Na2S2O8 consumed theamount of KBH4 and resulted in adecrease in fluorescence intensity.Therefore, for this work, 0.2%(m/v) Na2S2O8 was used and thefinal eluent of 0.5 mol L–1 HCl –1.5 mol L–1 HNO3 – 0.2 % (m/v)Na2S2O8.

Factors Affecting AFSPerformance

Factors affecting the AFS perfor-mance including the concentration

Fig. 3. The influence of sample loading time on the determination of 0.1 µg L–1 Hg(II). All other conditions as in Tables I and II.

Fig. 6. The influence of concentration of Na2S2O8 on the determination of 0.1 µg L–1 Hg(II). All other conditions as in Tables I and II.

Fig. 4. The influence of the total concentration of the mixedacid on the determination of 0.1 µg L–1 Hg(II). All other conditions as in Tables I and II.

Fig. 5. The influence of the volume ratio of 2.0 mol L–1 HNO3

to 2.0 mol L–1 HCl on the determination of 0.1 µg L–1 Hg(II).All other conditions as in Tables I and II.

39

Vol. 28(1), Jan./Feb.. 2007

Evaluation of Interferences

One alkali metal ion, seven typi-cal transition metal ions, and twohydride-forming elements wereselected to examine the interfer-ences on the determination of0.1 µg L–1 Hg(II) under the aboveoptimized conditions. The resultsare shown in Table III. The tolera-ble concentrations of As(III), Cd(II),

Co(II), Cu(II), Fe(III), Ni(II), Na(I),Pb(II), Se(IV), and Zn(II) for 0.1 µg L–1

Hg(II) were found to be 200, 100,1.5×103, 100, 750, 500, 1×105, 400,150, and 1.5×104 µg L–1, respectively,which are much higher than theirconcentrations usually found in nat-ural water. Therefore, it can bestated that the present systemallowed for the interference-freedetermination of trace dissolvedinorganic Hg in natural water sam-ples.

Analytical Figures of Merit

Under the above optimized con-ditions, effective enrichment andmatrix separation can be obtainedfor mercury. The analytical perfor-mance for the proposed methodwas assessed and the results aregiven in Table IV. With a consump-tion of 6.6 mL of sample solution,an enhancement factor of 39 wasobtained in comparison with thedirect injection of an aqueous solu-tion of mercury. The retention effi-ciency relative to the total mass ofthe analyte introduced to themicrocolumn was 62%. The repro-ducibility of the method given asthe relative standard deviation was4.3% at the 0.1-µg L–1 level (n=7).The limit of detection (3σ) was 1.2 ng L–1.

Validation of Developed Methodand Analysis of Water Samples

To test the accuracy of the devel-oped method for the determinationof trace inorganic mercury, certi-fied reference material GBW(E)080392 (simulated natural water)was analyzed. Several local naturalwater samples were also analyzed.As shown in Table V, thedetermined concentrations of mer-cury by the present method werein good agreement with the certi-fied values. This demonstrates theapplicability of on–line SPE (usingMWCNTs as the sorbent) coupledto CV–AFS for the determination oftrace dissolved inorganic mercuryin natural water samples. In addi-tion, the agreement between realamounts added and thosedetermined was satisfactory for allsamples. Therefore, it wasconcluded that the method devel-oped in this work is suitable for thedetermination of trace dissolvedinorganic mercury in natural water.

CONCLUSION

This work has demonstrated thefeasibility of multi-walled carbonnanotubes (MWCNTs) as solidphase extraction (SPE) sorbent foron-line preconcentration and sepa-ration coupled to cold vapor-atomicfluorescence spectrometry(CV–AFS) in the determination of

TABLE IIIEffect of Foreign Ions on the Determination

of 0.1 mg L–1 Hg(II)

Foreign Conc. RecoveryIon (µg L–1) (mean±σ,

n=3)(%)

As(III) 10 98 ± 2100 88 ± 4200 84 ± 4

Cd(II) 10 95 ± 250 89 ± 3

100 86 ± 3Co(II) 10 99 ± 4

1000 88 ± 51500 84 ± 4

Cu(II) 10 97 ± 4100 83 ± 5200 78 ± 4

Fe(III) 10 99 ± 3500 91 ± 3750 87 ± 3

Ni(II) 10 99 ± 4250 90 ± 3500 84 ± 3

Na(I) 1000 99 ± 210,000 95 ± 3

100,000 87 ± 4Pb(II) 10 99 ± 3

100 91 ± 4400 85 ± 3

Se(IV) 10 98 ± 3100 89 ± 4150 85 ± 3

Zn(II) 1000 96 ± 35000 92 ± 3

15,000 86 ± 4

TABLE IVAnalytical Performance of On–line SPE

for CV–AFS Determination of Inorganic Mercury

Preconcentration Time 120 sEnhancement Factor 39Sampling Frequency 15 h–1

Sample Consumption 6.6 mLRSD% (n = 7) 4.3%Detection Limit (3 σ) 1.2 ng L–1

Range of Calibration Graph 25–500 ng L–1

Calibration Function (n = 5) (F=Fluorescence Intensity, C=in ng L–1) F = -431+325C (C in ng L–1)

Correlation Coefficient 0.9994

40

trace inorganic mercury. With asample loading rate of 3.3 mL min–1

and 120 seconds preconcentrationtime, an enhancement factor of 39with an adsorption efficiency of62% was obtained. The detectionlimit (3s) was 1.2 ng L–1. The pro-posed method is rapid, simple, andsensitive, and can be applied to thedetermination of trace dissolvedinorganic mercury in natural watersamples.

Received October 2, 2006.

REFERENCES

1. M.J. Bloxham, S.J. Hill, and P.J. Wors-fold, J. Anal. At. Spectrom. 11, 511(1996).

2. L. Rahman, W.T. Corns, D.W. Bryce,and P.B. Stockwell, Talanta 52, 833(2000).

3. H. Bagheri, and A. Gholami, Talanta55, 1141 (2001).

4. L. Gámiz–Gracia and M.D. Luque deCastro, J. Anal. At.Spectrom. 14,1615 (1999).

5. E.L. Seibert, V.L. Dressler, D. Poze-bon, and A. J. Curtius,Spectrochim. Acta. Part B. 56,1963 (2001).

6. P. Cañada Rudner, A.Garcia de Tor-res, J.M.Cano Pavón, and F.Sanchez Rojas, Talanta 46, 109(1998).

7. L.-M. Dong, X.-P. Yan, Y. Li, Y. Jiang,S.-W. Wang, and D.-Q. Jiang, J.Chromatogr. A. 1036, 119 (2004).

8. Y. Li, Y. Jiang, X.-P. Yan, and Z.-M.Ni, Environ. Sci. Technol. 36, 4886(2002).

9. X.-P. Yan, Y. Li, and Y. Jiang, Anal.Chem. 75, 2251 (2003).

10. L.-P. Yu, and X.-P. Yan, At. Spec-trom. 25, 145 (2004).

11. X.-P. Yan and Y. Jiang, Trends Anal.Chem. 20, 552 (2001).

12. S. Iijima, Nature 354, 56 (1991).

13. C. Liu, Y.Y. Fan, M. Liu, H.T. Cong,H.M. Cheng, and M.S. Dresselhaus,Science 286, 1127 (1999).

14. Q.H. Wang, A.A. Setlur, J.M. Lauer-haas, J.Y. Dai, E.W. Seelig, andR.P.H. Chang, Appl. Phys. Lett. 72,2912 (1998).

15. J.M. Planeix, N. Coustel, B. Coq, V.Brotons, P.S. Kumbhar, R.Dutartre, P. Geneste, P. Bernier,and P.M. Ajayan, J. Am. Chem. Soc.116, 7935 (1994).

16. J. Kong, N.R. Franklin, C. Zhou,M.G. Chapline, S. Peng, K. Cho,and H. Dai, Science 287, 622(2000).

17. R.Q. Long, and R.T. Yang, J. Am.Chem. Soc. 123, 2058 (2001).

18. Q. Zhou, W. Wang, and J. Xiao,Anal. Chim.Acta. 559, 200 (2006).

19. X. Peng, Y. Li, Z. Luan, Z. Di, H.Wang, B. Tian, and Z. Jia, Chem.Phys. Lett. 376, 154 (2003).

20. K. Yang, L. Zhu, and B. Xing, Envi-ron. Sci. Technol. 40, 1855 (2006).

21. C. Lu, and H. Chiu, Chem. Eng. Sci.61, 1138 (2006).

22. Y.H. Li, S. Wang, J. Wei, X. Zhang,C. Xu, Z. Luan, D. Wu, and B. Wei,Chem. Phys. Lett. 357, 263 (2002).

23. Y.H. Li, S. Wang, Z. Luan, J. Ding, C.Xu, and D. Wu, Carbon 41, 1057(2003).

24. X. Wang, C. Chen, W. Hu, A. Ding,D. Xu, and X. Zhou, Environ. Sci.Technol. 39, 2856 (2005).

25. P. Liang, Q. Ding, and F. Song, J.Sep. Sci. 28, 2339 (2005).

26. Y.Q. Cai, G.B. Jiang, J.F.Liu, andQ.X. Zhou, Anal. Chem. 75, 2517(2003).

27. G. Liu, J. Wang, Y. Zhu, and X.Zhang, Anal. Lett. 37, 3085 (2004).

28. P. Liang, Y. Liu, and L. Guo, Spec-trochim. Acta. Part B. 60, 125(2005).

29. P. Liang, Y. Liu, L. Guo, J. Zeng andH. Lu, J. Anal. At. Spectrom. 19,1489 (2004)

30. L.-M. Dong and X.-P. Yan, Talanta65, 627 (2005).

31. C.-Y. Lu, X.-P. Yan, Z.-P. Zhang, Z.-P.Wang, and L.-W. Liu, J. Anal. At.Spectrom. 19, 277 (2004).

32. Y. Li, X.-P. Yan, L.-M. Dong, S.-W.Wang, Y. Jiang, and D.-Q. Jiang, J.Anal. At Spectrom. 20, 467 (2005).

TABLE VAnalytical Results for Determination of Inorganic Mercury

in Water Samples and a Certified Reference Material

Sample Hg Concn. in Original Sample Hg Added Recovery (%)(ng L–1) (ng L–1) (ng L–1) (mean±σ, n=3)

Certified Determined(mean±σ, n=3)

GBW(E) 080392 10.0±0.5a 9.9±0.3a – –Tap Water – – 25.0 96±2Lake Water 1 – 8.7±0.3 25.0 92±4Lake Water 2 – 6.5±0.3 25.0 90±4River Water 1 – 8.9±0.4 25.0 92±2River Water 2 – 20.4±0.4 25.0 93±4

River Water 3 – 12.8±0.4 25.0 91±3

a Concentration in µg L–1.

1. Concepts, Instrumentation and Techniques inAtomic Absorption SpectrophotometryAuthors: Richard D. Beaty and Jack D. KerberOrder No. AA-914C (free of charge)Ordering information: http://www.las.perkinelmer.com or contact your localPerkinElmer representative.

This book contains theoretical concepts and definitions of the science of atomic spectroscopy:atomic emission, atomic absorption, and atomic fluorescence. It also discusses high sensitivitysampling systems and the advantages and limitations of the cold vapor mercury, hydridegeneration, and graphite furnace atomic absorption techniques. Also discussed are spectral and non-spectral interferences, including the goals and use of thestabilized temperature platform furnace (STPF) system.

2. Analytical Graphite Furnace AAS – A Laboratory GuideAuthors: G. Schlemmer and B. RadziukOrder No. B051-1731 Ordering and price information: http://www.las.perkinelmer.com or contact your localPerkinElmer representative.

This book provides insight into the theoretical and practical aspect of graphite furnace AA making it the perfectreference resource for all laboratories wanting to use their graphite furnace more effectively.Using an easy-to-follow style, the reader is guided from method development to calibration and validation of theinstrument to the use of accessories and software in modern graphite furnace AA.

3. Concepts, Instrumentation and Techniques in InductivelyCoupled Plasma Optical Emission SpectrometryAuthors: Charles B. Boss and Kenneth J. FredeenOrder No. 005446B (free of charge)Ordering information: http://www.las.perkinelmer.com or contact your localPerkinElmer representative.

This book presents the general characteristics of ICP-OES and ICP-OES instrumentation. Itdiscusses ICP-OES methodologies including their application for the analysis of samples inthe various industries such as agriculture and foods, biological and clinical, geological,environmental and water, metals, and organics.

4. Practical Guide to ICP-MS Author: Robert Thomas, Scientific Solutions (www. scientificsolutions1.com)Published in 2004 by Marcel Dekker Ordering and price information:www.dekker.com/servlet/product/productid/5319-4

The brand new reference book presents this powerful trace-element technique as a practicalsolution to real-world problems. The basic principles of ion formation/transportation/detection,common interferences, peak quantitation, sample preparation, contamination issues, routinemaintenance and application strengths of ICP-MS are described in a way that is easy tounderstand for both experienced users and novices of the technique. In addition ICP-MS iscompared with AA and ICP-OES in the areas of detection capability, dynamic range, samplethroughput, ease of use and cost of ownership. The book concludes with an excellent chapteron the most important testing criteria when evaluating commercial instrumentation.

Books on the AAS, ICP-OES, ICP-MS Techniques

AS 2 8(1) 1-29 Spectroscopy 28(1).pdf · intestinal digestion and absorption. These analyses were made by ICP-MS and were compared to those by hydride generation atomic fluores-cence

Documents