Atomic Spectroscopy 24(5)

Issues also available

electronically.

(see inside front cover)

ASPND7 24(5) 159–194 (2003)ISSN 0195-5373

AtomicSpectroscopy

September/October 2003 Volume 24, No. 5

In This Issue:

Characterization of Environmental Samples in an Ophiolitic Area of Northern Italy Using ICP-OES, ICP-MS, and XRFM. Bettinelli, G.M. Beone, C. Baffi, S. Spezia, and A. Nassisi ..................................... 159

Cloud Point Extraction Preconcentration and ICP-OES Determination of Trace Chromium and Copper in Water SamplesJing Li, Pei Liang, Taqing Shi, and Hanbing Lu ......................................................... 169

A Rapid Method for the Determination of Selenium in Blood and Serum by ETAAS With Zeeman Background CorrectionArturo Montel, José Manuel de Pradena, José María Gervas, Guadalupe Álvarez Bustamante, and José Luis López-Colón .................................... 173

Flow Injection AAS Determination of Cd, Cu, and Pb at Trace Levels in Wine Using Solid Phase Extraction Andréa Pires Fernandes, Mercedes de Moraes, and José Anchieta Gomes Neto .... 179

Tantalum Effects on the ICP-OES Determination of Trace Elements in Tantalum Powder G. Anil, M.R.P. Reddy, Arbind Kumar, and T.L. Prakash .......................................... 185

Mixed Matrix Effect of Easily and Non-easily Ionizable Elements in the ICP-OES Analysis of Trace Impurities in Potassium Tantalum Fluoride G. Anil, M.R.P. Reddy, Arbind Kumar, and T.L. Prakash .......................................... 190

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

PerkinElmer Life andAnalytical Sciences710 Bridgeport AvenueShelton, CT 06484-4794 USAwww.perkinelmer.com

Editor

Anneliese LustE-mail: [email protected]

Technical Editors

Glen R. Carnrick, AADennis Yates, ICPKenneth R. Neubauer, ICP-MS

SUBSCRIPTION INFORMATIONAtomic SpectroscopyP.O. Box 3674Barrington, IL 60011 USAFax: +1 (847) 304-6865

2003 Subscription Rates• U.S. $60.00 includes third-class

mail delivery worldwide; $20.00 extra for electronic file.

• U.S. $80.00 includes airmail delivery; $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 File• For electronic file, send 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 © 2003PerkinElmer, Inc.All rights reserved.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)

Guidelines for Authors

AtomicSpectroscopy

Atomic 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 3000subscribers 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 manuscriptshould be submitted to theeditor by mail plus electronic fileon disk or e-mail in the followingmanner:1. Mail original of text, double-

spaced, plus graphics inblack/white.

2. Provide text and tables in .docfile and figures in doc or tiffiles.

3. Number the references in theorder they are cited in the text.

4. Submit original drawingsor glossy photographs and figure captions.

5. Consult a current copy ofAtomic Spectroscopy for format.

6. Or e-mail text and tables indoc file and graphics in docor tif files to the editor:[email protected] [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.

A copyright transfer form issent to the author for signature.

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.

Additional reprints can bepurchased, but the requestmust be made at the time themanuscript is approved forpublication.

Anneliese LustEditor, Atomic SpectroscopyPerkinElmerLife and Analytical Sciences710 Bridgeport AvenueShelton, CT 06484-4794 USA

Vol. 24(1), January/February 2003

PerkinElmer is a registered trademark of PerkinElmer, Inc.AAnalyst, Optima, GemTip, and WinLab are trademarks of PerkinElmer Life and AnalyticalSciences.SCIEX and ELAN are registered trademarks of MDS Inc., a division of MDS Inc.Dell is a registered trademark of Dell Computer Corporation.Hewlett Packard and LaserJet are registered trademarks of Hewlett Packcard Corporation.Milli-Q is a trademark of Millipore Corporation.Ryton is a registered trademark of Phillips Petroleum Company.Suprapur is a registered trademark of Merck & Co.Triton is a registered trademark of Union Carbide Chemicals and Plastics Technology Corporation.Tygon is a trademark of Norton Co.Viton is a registered trademark of DuPont Dow Elastomers.Registered names and trademarks, etc. used in this publication even without specific indication therof are not to be considered unprotected by law.

159

Characterization of Environmental Samples in an Ophiolitic Area of Northern Italy

Using ICP-OES, ICP-MS, and XRF*M. Bettinellia, G.M. Beoneb, C. Baffib, S. Speziaa, and A. Nassisic

a Laboratorio Igiene Ambientale e Tossicologia Industriale, Fondazione "S. Maugeri," Clinica del Lavoro e della Riabilitazione I.R.C.C.S., via Ferrata, 8, 27100 Pavia, Italy

b Istituto di Chimica Agraria e Ambientale, Facoltà di Agraria, Università Cattolica S. Cuore, via Emilia Parmense, 84, 29100 Piacenza, Italy

c ARPA Sezione Provinciale di Piacenza, Area Analitica Specialistica Agropedologia, Loc. Gariga, 29027 Podenzano-Pc, Italy

Atomic SpectroscopyVol. 24(5), September/October 2003

INTRODUCTION

Environmental studies requireaccurate and precise informationon trace element distribution in theecosystems. In this respect, parti-tion and transfer coefficients can be extremely useful to representthe distribution and flow ofelements among the various envi-ronmental compartments which areoften characterized by extremelycomplex matrices.

In environmental studies, thechoice of the most suitable analyti-cal methods and procedures bywhich good quality data can beobtained is of primary importance.

Generally, the use of validatedmethods for which precision valuesare well known and measurementuncertainty has already been evalu-ated in compliance with the latestinternational guidelines is recom-mended (1,2). Validation of thesemethods requires accuracy testswith certified samples (3,4) or,when not available, comparisonbetween different independent analytical techniques on real samples (5,6).

Today, the characterization ofenvironmental matrices is obtainedusing complete dissolution of thesamples, followed by monoelement(FAAS or GFAAS) or multielementdetermination with spectroscopic

techniques (ICP-OES or ICP-MS).Other techniques, such as X-ray fluorescence (XRF), can providedirect analysis of the sample. Thecurrent trend in analytical methodsfavors closed-vessel solubilizationrather than the traditional open-vessel systems, since the closed-vessel method makes it possible toobtain quantitative recoveries forvolatile elements of toxicologicalinterest and improves methoddetection limits (MDL) by minimiz-ing the required amount of reagents(7–10).

Multielement techniques aregradually replacing monoelementtechniques because they provide a complete elemental profile of thesamples (11) and ensure a signifi-cant reduction in analysis time.

In ophiolitic areas, environmen-tal samples such as rocks,sediments, soils, and plants arecharacterized by high concentra-tions of Ni and Cr. However, thedetermination of these elementscan pose serious problems due tothe complexity of the matrices, par-ticularly when siliceous materialsare present which cannot easily bedissolved completely with com-mon strong mineral acids (12).

Ophiolitic environments arecharacterized by the presence ofmafic or ultramafic rocks, partly orcompletely serpentinized. Theserocks usually show high concentra-tions of Ni and Cr (sometimes up toa few thousands mg/kg), whichdecrease progressively as the rockacidity rises.

ABSTRACT

In Ponte Barberino, an ophi-olitic environment in the TrebbiaRiver Valley near the town ofPiacenza, Northern Italy, rocksand soils are characterized byhigh concentrations of Cr and Ni.The determination of macro,micro, and rare earth elements inrock, soil, sediment, and plantsamples was carried out by ICP-OES, ICP-MS, and XRF after alka-line fusion or microwave acidmineralization.

For sediment samples,microwave acid digestion (8 mLaqua regia + 2 mL HF), followedby ICP-OES, provided resultscomparable to those obtained byXRF.

In the case of plants,microwave acid mineralization (7 mL HNO3 + 1 mL H2O2 + 0.2mL HF) followed by ICP-OESanalysis was particularly suitablefor the determination of traceelements in leaves of the mostcommon botanical species foundin the investigated area. In addi-tion, soil/plant transfer factorscharacterized the Alyssum spp.species as a hyperaccumulator ofNi.

Alkaline fusion was found tobe the best dissolution procedurefor rock samples; recoveriesranged between 95 and 103% forCr and Ni, respectively.

Also, XRF and ICP-OES analy-sis proved to be very reliable andprovided comparable results. Inrock samples, ICP-MS allowedthe determination of trace andrare earth elements notdetectable by the othertechniques.

*Corresponding author.Tel: ++39 0382/ 592311 Fax: ++39 0382/ 592072e-mail: [email protected]

160

Since ophiolitic soils wereformed as deposits (colluvium oralluvium) due to the alteration ofmafic rocks, Ni and Cr arecommonly found co-precipitatedwith Fe and Mn oxides. Unlike Fe2+

and Mn2+ ions, which precipitate insitu, Ni2+ and Cr3+ ions are relativelymore stable in water solutions andtherefore can migrate over longerdistances. In these soils, Ni2+ andCr3+ ions are often found in alumi-nosilicate structures or as oxidesmixed with Fe3+; their solubilitybeing inversely proportional to soilacidity (pH). The determination ofNi and Cr concentration in soils andplants from an ophiolitic site is cru-cial for environmental monitoring,since both elements can be consid-ered natural tracers of geologicalorigin (13).

In the present work, analyseswere carried out in two differentlabs in order to evaluate the applic-ability of microwave dissolutionprocedures, in conjunction withmultielement spectroscopic analy-ses (ICP-OES/ICP-MS), for the ele-mental analysis of rocks, sediments,and plants collected in an ophioliticarea. X-ray fluorescence was usedto compare the different analyticalmethods, while alkaline fusion was employed as an alternativetechnique to closed-vessel aciddigestion for the solubilization ofsamples with very complex matri-ces.

EXPERIMENTAL

Area of Study

The analyzed area (Figure 1) islocated upstream of the TrebbiaRiver Valley. This area is part of anextensive formation of serpentineoutcrops, covering several hundredsquare kilometers, with high con-centrations of Cr and Ni rangingfrom 500 to 1220 µg g–1 and 900 to 1500 µg g–1, respectively.

Samples of rocks, sediments, andsome of the most common botani-

cal species (i.e., Alyssum spp.,Helichrysum, and Euphorbia) werecollected in this area in September 2001.

Instrumentation

Two different microwavedigesters were used in each lab:MLS 1200 Milestone (FKV, Sorisole,Bergamo, Italy) and MDS-2000(CEM, Matthews, NC, USA). Theinstrumental operating conditionsreported in Table I are thosedescribed in a previous paper (14).

Two inductively coupled plasmaoptical emission spectrometerswere used: A PerkinElmer Optima™4300 DV simultaneous ICP-OES(PerkinElmer Life and AnalyticalSciences, Shelton, CT, USA),equipped with an AS-90 autosam-pler, a cross-flow nebulizer, and aScott-type Ryton® spray chamber,and a Jobin Yvon 24 sequential ICP-

OES (Jobin Yvon, France) with across-flow nebulizer.

A Perkin Elmer SCIEX ELAN®5000 inductively coupled plasmamass spectrometer was also used(PerkinElmer SCIEX Instruments,Concord, Ontario, Canada). Thiswas equipped with a PerkinElmerGemTip™ nebulizer and a corro-sion-resistant Scott-type spray cham-ber. The operating conditions werereported in previous papers (12,14)and are summarized in Tables II–IV.

XRF analyses were performedusing a Philips PW1400 spectrome-ter with a Rh-target x-ray tube andPhilips x-14 software package(Milan, Italy). The instrument has afour-position internal sample turretwith a 72-position external sampler(PW1500/15 autosampler) whichallows analysis of the ratio of thestandard and several unknowns in abatch.

Fig. 1. Location of serpentinized ophiolitic outcrops in the Apennines of thePiacenza and Parma provinces, Commune of Bobbio, Piacenza, Italy. The investigated area is marked with a circle.

161

Vol. 24(5), Sept./Oct. 2003

TABLE IOperating Conditions for Microwave Systems

Plant Leaves a

CEM 2000 / Milestone 1200

Step 1 2 3Power (W) 250 400 500

Hold time (min) 2 2 15

a 250 mg of plant leaves treated with 7 mL HNO3, 1 mL H2O2, and0.2 mL HF in each of six vessels.

Soils and Sediments a

CEM 2000 / Milestone 1200

Step 1 2 3 4 5b 6Power (W) 250 400 600 0 300 0

Hold time (min) 8 4 6 2 3 2

a 250 mg of soil / sediment treated with 2 mL HF, 8 mL aqua regia,and 2 mL H3BO3 in each of six vessels.b The boric acid-satured solution was added before step #5.

TABLE IIICP-OES Instrumental Operating Parameters

PerkinElmer Jobin Yvon Optima 4300 DV Model 24Simultaneous ICP Sequential ICP

Frequency 40 MHz free-running 40.68 MHzIncident power 1300 W 900 WReflected power <5 W <5 WAr gas flow rate:

Outer gas 15 L min–1 13 L min–1

Auxiliary gas 0.5 L min–1 <1 L min–1

Aerosol gas 0.8 L min–1 0.9 L min–1

NebulizerModel GemTip™ cross-flow Scott-type cross-flow Sample flow rate 1.0 mL min–1 1.5 mL min–1

Sample pump Peristaltic pump Gilson Minipuls™ 2Computer Dell® Optiplex GX150 Siemens Nixdorf PCD-4RsxAPrinter HP LaserJet® 2200D Fujitsu DL1150

Autosampler Model AS90 –

TABLE IIIWavelengths Used for the PerkinElmer

Optima 4300 DV ICP-OES

Element Wavelengths (nm)

Cd 214.438 – 228.802a

Co 230.786 – 228.616 a

Cr 205.552 – 206.149 – 267.716 a

Cu 324.754 – 327.396 – 234.754 a

Mn 257.610 a – 260.569

Ni 231.604 a – 221.647

Pb 220.353 a – 261.418

Zn 206.200 – 213.856 a

a Wavelength used for the analysis.

162

Reagents and StandardSolutions

The following Suprapur®reagents (E. Merck, Darmstadt, Germany) were used: nitric acid(65% m/v), hydrochloric acid (37%m/v), hydrofluoric acid (40% m/v),and reagent-grade boric acid. TheICP mass spectrometer and opticalemission spectrometers were cali-brated using certified multielementsolutions diluted with water thatcontained the same amount of acidsas the samples. The high-puritywater used in this analysis was pro-duced with a Milli-Q™ deionizingsystem (Millipore, Bedford, MA,USA).

Materials and Procedures

SedimentsA four-layer sediment sample

(0–15 cm, 15–25 cm, 25–38 cm and 38–50 cm) was air-dried and sievedto obtain the fine earth [fractionswith a diameter (Ø) of < 2 mm].

An aliquot of each layer was thenground to 0.2 mm in a planetarymill with agate balls. About 0.250 gof this fraction (Ø < 0.2 mm) wassubjected to microwave acid diges-tion with HF/HNO3/HCl, followingthe program reported in a previouspaper (12). The solutions obtainedfrom the sample mineralizationwere diluted to 50 mL and analyzedby ICP-OES and ICP-MS. An aliquotof the 0.2-mm fraction was also ana-lyzed by XRF. Eighteen elementswere determined in the sedimentsample. The solutions derived fromthe HF/HNO3/HCl procedure wereanalyzed by two laboratories usingtwo different instrumental configu-rations. In order to evaluate theuncertainties associated with theinstrumental readings and thepreparation phase, the solutionsprepared by LAB A were analyzedby LAB B and vice versa.

The possibility of using two dif-ferent wavelengths for every ana-lyte in order to rapidly verify thepresence of spectral interferences

TABLE IVInstrumental Parameters for ICP-MS Analysis

RF power 1085 WPlasma argon 14.6 L/minNebulizer flow 1.1 L/minAuxiliary flow 0.83 L(minSample flow rate 1 mL/minNebulizer Cross-flowData Peak-hop transientResolution NormalReading time 150 msDwell time 50 msSweeps / replicates 3Number of replicates 5Sample time 1’06"Sample read delay 50 sAutosampler wash delay 60 sCalibration mode External calibrationCalibration standard 10, 20, 30, 50, 100 µg/L

Curve fit Linear through zero

Isotopes MassAs 75Ba 138Be 9Cd 114Ce 140Co 59Cr 52Cu 63Cs 133Ga 71Gd 158Hf 180Ho 165La 139Mo 98

Interelemental corrections:

75As = 75As – 3.087 77Se + 2.619 82Se

51V = 51V – 3.09 53Cr + 0.353 52Cr

Isotopes MassMn 55Nd 142Ni 60Pb 208Rb 85Sb 121Sm 152Sr 88Ta 181U 238V 51Y 89Yb 174Zn 66Zr 90

163

Vol. 24(5), Sept./Oct. 2003

was evaluated using thePerkinElmer Optima 4300 DV(Table III). The concentration dif-ference measured for all the ana-lytes at the two wavelengths wasnot larger than 4–5 %; therefore, atthese operating conditions, interfer-ences were not a problem. Usingthe wavelengths marked with an“a” in Table III, the short-term pre-cision for 1 mg/L of different ele-ments (%RSD for five replicates)was always less than 1.2%.

Plants and SoilsTrace elements were determined

in plant leaves of the botanicalspecies Alyssum spp.,Helichrysum, and Euphorbia. The collected samples were driedin a thermostatic oven at 40°C andthen ground, without washing, in a planetary mill with agate balls inorder to obtain 0.2-mm particles.Approximately 0.250 g of the dried0.2-mm fraction was submitted tomicrowave acid digestion (using amixture of 7 mL HNO3 + 1 mL H2O2

+ 0.2 mL HF), following the pro-gram reported in a previous paper(14) and also shown in Table I. Themineralized solutions were thendiluted to 50 mL and six elementsin the botanical samples weredetermined by ICP-OES. The same

procedure was applied forsediment samples, and six trace ele-ments were determined in the soilobtained from beneath these plants.

RocksA rock sample (serpentinized

ophiolite) was analyzed as follows:(a) microwave acid digestion fol-lowed by ICP-OES or ICP-MS analy-sis; (b) alkaline fusion with lithiumtetraborate followed by ICP-OESand ICP-MS analysis; (c) directanalysis by XRF on a pellet sample.Eleven elements were determinedby ICP-OES, 22 samples by ICP-MS,and 17 samples by XRF.

The accuracy of these analyticalmethods (expressed as % Recovery)was estimated by using the follow-ing certified samples: CRM 320River Sediment for sediments, NIST1570a Spinach Leaves for plants,NIM Dunite and NBS-688 Basalt forrocks. These certified materialswere treated according to the man-ufacturer’s instructions. Intralabora-tory (repeatability or %r) andinterlaboratory (reproducibility or%R) precisions were estimatedaccording to ISO standard 5725(15). The statistical methods usedin this study are consistent withthose reported in the literature (16).

RESULTS AND DISCUSSION

Sediments

Table V shows the concentra-tions of Cd, Co, Cr, Cu, Mn, Ni, Pband Zn observed in the CRM 320River Sediment. Acceptable accu-racy (recoveries between 89 and106%) and precision values (%r <10and %R <26) were obtained for allof the elements determined, exceptCd (uncertified), which was present at a concentration of about 0.5 µg g–1.

Particularly good were the recov-eries for Cr and Ni, elements abun-dant in ophiolitic areas and difficultto solubilize with common acidmixtures, such as HNO3, HNO3/HCl, and HNO3/HClO4. The ICP-OES/ICP-MS and XRF techniquesprovided comparable results asshown in Table VI. This proves thatXRF is particularly suitable for ageneral characterization ofelements since it is a quick, non-destructive technique, while ICP-OES and ICP-MS provide a widerelemental profile, including fortrace elements and rare earths. XRFinstrument detection limits (IDL)were several µg g–1, slightly higherthan the values found for ICP-OES.

The major limits of the XRF technique are related to accuracy,which depends on the types (matri-ces and interelement ratios) of thesamples used to calibrate the instru-ment. On the other hand, the limitsfor ICP-OES are mostly due to prob-lems arising during the dissolutionsteps which do not always permit a complete dissolution of the samples.

The information given in TableVI shows that Al, Ba, and K concen-trations increase as the depth of thesoil layers increases, while the Ca,Mg, Ni, and Cr concentrationsdecrease. The Cr/Ni ratio remainsconstant in all layers (R Cr/Ni 1.45),except the 0–15 cm depth (R Cr/Ni1.15). This confirms the commonorigin of both analytes from

TABLE VCRM 320 River Sediment

ICP-OES Determination of Cd, Co, Cr, Cu, Mn, Ni, Pb, and Zn (µg g–1) After Digestion With HF / aqua regia

Element Certified Values Found Values Recovery r R(µg g–1) (µg g–1) (%) (%) (%)

Cd NP 0.533 ± 0.026 ND 21 45

Co 17.3 – 22.1a 20.5 ± 0.80 ND 7 26

Cr 138 ± 7.0 122 ± 6.0 88 9 12

Cu 44.1 ± 0.1 42.0 ± 2.1 95 8 9

Mn 619 – 786a 697 ± 44 ND 10 12

Ni 75.2 ± 1.4 72.4 ± 2.5 96 6 12

Pb 42.3 ± 1.6 40.6 ± 10.3 96 10 22

Zn 142 ± 3 138 ± 7 97 10 14

a Indicative values. NP = Not Present. ND = Not Determined.

164

ophiolitic rocks upstream of theinvestigated area. The Al/Ti ratio isconstant at various depths (averagevalue: 18.2 ± 1.7), which meansthat both analytes can be used asnormalization elements in order tocalculate enrichment factors fortrace elements in environmentalbiomonitoring studies (17).

Plants and Soils

ICP-OES determination of Al, Cr,Cu, Fe, Mn, Ni, and Zn in botanicaland soil samples was performedafter closed-vessel microwave aciddissolution, a procedure similar tothe one used for the sediment sam-ples. Analytical accuracy and preci-sion can be enhanced by adding HFto the solubilization mixture, asshown in a previous paper (14).This ensures the complete dissolu-tion of materials particularly resis-

tant to acids, e.g., compounds con-taining Cr and Ni.

The accuracy values found inboth labs for plant reference mater-ial NIST 1570a Spinach Leaveswere between 89 and 100% (TableVII), with good precision values forCr and Ni.

Table VIII shows the concentra-tion of some analytes in the botani-cal species collected in the areaunder investigation. High Ni valueswere observed in the Alyssum spp.samples (approx.1 %), whileEuphorbia shows a significant con-centration of Mn (approx. 0.16%).Additional analyses were carriedout in the soil obtained frombeneath these plants using thesame procedures described for sed-iments in order to determine theconcentration of Ca, Cr, Cu, Fe,Mg, Mn, Ni, and Zn.

Table IX shows all the valuesfound, along with the soil/planttransfer factors. This parameter(TF), important and often used inenvironmental studies, is the ratiobetween the concentration of anelement in a plant and its total con-centration in the soil. This providesuseful information on the cyclingof elements between two differentcompartments. As a rule, the TFvalues found in uncontaminatedagricultural and forestry soils areless than one unit. The values givenin Table IX confirm this rule,except for Mn in the Euphorbiaspecies and Ni in Alyssum spp.While a preferential accumulationof Mn is quite evident (TF 1.26),the high Ni value (TF 8.3) in theAlyssum spp. species suggestshyperaccumulation, as alreadyreported in the literature (18,19).The Ca/Mg ratio observed in

TABLE VIDetermination of Major (%) and Minor (µg g–1) Elements in a Sediment Sample

Collected in Ponte Barberino(Analysis by XRF or ICP-OES/ICP-MS After Sample Dissolution With HF/aqua regia in a Microwave Oven)

Depth 0–15 cm 15–25 cm 25–38 cm 38–50 cmElement XRF ICP-OES/ XRF ICP-OES/ XRF ICP-OES/ XRF ICP-OES/

ICP-MS ICP-MS ICP-MS ICP-MS

Al (%) a 5.59 5.70 8.29 8.20 9.33 9.55 10.1 10.5Ca (%) a 9.18 9.04 2.55 2.36 1.81 1.59 1.59 1.50Fe (%) a 9.27 9.40 7.70 8.02 8.26 8.49 9.21 9.55K (%) a 1.65 1.39 2.17 2.01 2.35 2.44 2.51 2.75Mg (%) a 2.09 1.89 1.52 1.50 1.46 1.40 1.38 1.33Na (%) a 0.31 0.28 0.22 0.20 0.20 0.19 0.14 0.15Si (%) a 23.8 24.20 25.7 23.88 26.5 26.87 25.9 26.59Ti (%) a 0.29 0.30 0.49 0.55 0.53 0.52 0.54 0.51Ba (µg g–1) a 231 250 299 320 319 330 420 411Co (µg g–1) b 11.7 28.0 24.2 34.5 26.0 33.6 26.4 32.0Cr (µg g–1) b ND 197 ND 169 ND 149 ND 125Cu (µg g–1) b 31.0 51.0 43.3 44.9 44.5 43.8 42.9 38.0Mn(µg g–1) a 797 845 936 902 995 972 1024 942Ni (µg g–1) b 123 171 131 116 98 103 90.9 86Pb (µg g–1) b 16.4 9.12 30.3 25.2 28.9 16.0 42.8 13.4Sr (µg g–1) a 417 437 242 224 229 220 251 181V (µg g–1) b 62.6 74.0 90.8 ND 97.0 ND 11.0 ND

Zn (µg g–1) a 90.1 101 113 123 115 149 129 123

a Determined by ICP-OES. b Determined by ICP-MS. ND = Not Determined.

165

Vol. 24(5), Sept./Oct. 2003

TABLE VII NIST 1570a.Spinach Leaves, ICP-OES Determination of Cr, Cu, Fe, Mn, Ni and Zn (µg g–1)

After Sample Digestion With HF/HNO3/H2O2 in a Microwave Oven

Element Certified Values Found Values Recovery r R(µg g–1) (µg g–1) (%) (%) (%)

Cr NP 1.63 ± 0.45 ND 35 82Cu 12.2 ± 0.6 10.9 ± 1.27 89 19 35Fe NP 260 ± 30 ND 8 36Mn 75.9 ± 1.9 70.9 ± 8.8 93 10 38Ni 2.14 ± 0.10 2.15 ± 0.36 100 46 47

Zn 82 ± 3 79.2 ± 8.73 97 6 34

NP = Not Present. ND = Not Determined.

TABLE VIIIAlyssum spp., Helichrysum, and Euphorbia Leaves

ICP-OES Determination of Ca, Cr, Cu, Fe, Mg, Mn, Ni and Zn (µg g–1) After Microwave Digestion With HF/HNO3/H2O2

Determination of Ca/Mg Ratio

Element Alyssum spp. Helichrysum Euphorbia(µg g–1) (µg g–1) (µg g–1)

Cr 23 ± 2.83 1.89 ± 0.23 7.14 ± 0.13Cu 4.0 ± 1.41 11.4 ± 0.20 4.52 ± 0.05Fe 1221 ± 210 197 ± 4.15 344 ± 3.21Mn 102 ± 24.7 134 ± 1.72 1626 ± 14.8Ni 10375 ± 1711 13.3 ± 0.31 22.9 ± 0.22Zn 37 ± 16.3 79.2 ± 8.73 36 ± 0.83

Ca/Mg 0.050 0.82 0.81

TABLE IX Concentrations of Cr, Cu, Fe, Mn, Ni and Zn in Ophiolitic Soil Beneath Botanical Species

Soil/Plant Transfer Factors (TF) for Alyssum spp., Helichrysum, and Euphorbia

Element Soil TF TF TF(µg g–1) Alyssum spp. Helichrysum Euphorbia

Cr 1160 ± 203 0.020 0.002 0.006Cu 46 ± 1.99 0.087 0.248 0.098Fe 44252 ± 9575 0.028 0.004 0.008Mn 1289 ± 268 0.079 0.104 1.261Ni 1260 ± 306 8.324 0.010 0.018

Zn 114 ± 11 0.319 0.415 0.315

166

botanical species, particularly inAlyssum spp., was less than oneunit, which is further evidence ofthe ophiolitic nature of theanalyzed site, and in accordancewith the results obtained by Pelosi(20) and Vergnano Gambi (21).

Rocks

Elemental determination in rockswas carried out by comparing theresults obtained by three differentanalytical methods: XRF, ICP-OES,and ICP-MS.

Table X shows the concentrationvalues of both major and trace ele-ments found in the NIM Dunite cer-tified sample and in the ophioliticrocks collected in Ponte Barberino.These samples were analyzed usingalkaline fusion for dissolution, fol-lowed by ICP-OES and XRF analysis.

to the XRF calibration which isstrongly affected by the availabilityof certified reference materials withmatrices similar to those of the sam-ples used.

In the real ophiolite sample, theagreement between ICP-OES andXRF results was on the whole satis-factory, except for Ca and Cuwhich were clearly underestimatedby XRF. The ICP-MS technique, dueto lower instrumental detection lim-its (IDL), allowed the determinationof approximately 20 trace and rareearth elements which cannot bedetermined by ICP-OES at theseconcentration levels. In this study,good accuracy was obtained foralmost all analytes using ICP-MS,except for Ho, Sb, Zr, Sm, and Yb;for the last two analytes, onlyindicative values are reported in theliterature.

CONCLUSION

This study demonstrates thataccurate analytical data can beobtained in very complex environ-ments, such as an ophiolitic site, by using solubilization proceduresbased on the type and nature of thematrices. The use of multielementindependent techniques with ana-lytical sensitivities suitable for thedetermination of trace metal con-centrations, the use of certified reference materials with matricessimilar to the samples investigatedand, most of all, the use of intra-and interlaboratory comparisonsare essential to verify the reliabilityof an analytical method in terms ofaccuracy and precision.

For sediment samples, the use of microwave digestion in conjunc-tion with ICP-OES and ICP-MS pro-vided very good analytical results,comparable to those obtained byXRF for most of the elements. Theuse of HF in the acid mixtureemployed for sample dissolutionprovided quantitative recoveries forthe determination of the total con-tent of analytes, particularly for Cr

Table X does not show the resultsof some preliminary testsperformed by microwave aciddigestion, since the recovery of alarge number of the analytes wasincomplete.

Alkaline fusion at about 1000°Censured a complete solubilizationof silicates, with recoveriesbetween 80 and 120% for NIMDunite (Table X) and between 76and 114% for NBS-688 Basalt (TableXI). Cr and Ni percentage recover-ies ranged from 89 to 100%.

Following the comparisonbetween the employed techniques,the recovery of most of the analytesin the NIM Dunite material (TableX) was better with ICP-OES thanXRF; Ba, Sr, Pb and Cu concentra-tions were not detected by X-rayfluorescence. This is probably due

TABLE X ICP-OES and XRF Determination of

Major (%) and Minor (µg g-1) Elements in NIM Dunite and in an Ophiolitic Rock Sample

Elements NIM Dunite Ophiolitic RockCertified Found Values Found ValuesValues ICP-OES XRF ICP-OES XRF

Al (%) 0.20 0.21 0.18 0.66 0.75Ca (%) 0.20 0.22 0.28 1.26 0.67Fe (%) 11.87 12.15 18.1 8.82 9.08K (%) 0.01 0.01 0.01 0.01 0.01Mg (%) 26.1 24.7 20.1 14.7 13.6Na (%) 0.04 0.03 0.02 0.02 < DLSi (%) 18.18 19.0 20.3 19.6 20.1Ti (%) 0.012 0.011 0.016 0.05 0.07Ba (µg g–1) 10.0 10.8 < DL 55 <DLCo (µg g–1) 210 198 122 103 61.9Cr (µg g–1) 2900 2760 ND 1170 NDCu (µg g–1) 10.0 12.3 < DL 90.1 28.9Mn(µg g–1)) 1704 1654 1032 799 839Ni (µg g–1) 2050 2110 1971 1840 1941Pb (µg g–1) 7.00 8.43 < DL 0.70 <DLSr (µg g–1) 3.00 4.11 < DL 5.90 4.28V (µg g–1) 40.0 43.1 20.9 83.0 90.4

Zn (µg g–1) 90.0 79.7 65.5 47.2 51.7

DL = Detection Limit. ND = Not Determined.

167

Vol. 24(5), Sept./Oct. 2003

TABLE XIICP-OES/ICP-MS Determination of Macroelements (%), Trace Elements and REEs (µg g–1)

in NBS-688 Basalt and in an Ophiolitic Rock After Alkaline Fusion

NBS-688 Basalt Ophiolitic Rock

Element Certified Values Found Values Recovery Found ValuesICP-OES/ICP-MS (%) ICP-OES/ICP-MS

Al (%) 9.18 ± 0.05 9.41 ± 0.02 102 0.66Ca (%) 8.70a (8.47 ± 0.36)b 8.88 ± 0.09 102 1.26Fe (%) 7.23 ± 0.03 7.43 ± 0.03 103 8.82K (%) 0.16 ± 0.07 0.12 ± 0.05 75 0.01Mg (%) 5.10a 5.41 ± 0.04 106 14.7Na (%) 1.60 ± 0.02 0.90 ± 0.01 56 0.02Si (%) 22.6 ± 0.05 22.30 ± 0.03 99 19.6Ti (%) 0.70 ± 0.06 0.68 ± 0.04 97 0.05Ba (µg g–1) 200a (197 ± 12)b 189 ± 6 94 55As (µg g–1) 2.4a ND ND 0.59Be (µg g–1) NA 1.58 ± 0.86 ND 0.95Cd (µg g–1) NA 1.10 ± 0.01 ND 1.19Ce (µg g–1) 13.3a (13 ± 2)b 10.75 ± 0.2 81 0.47Co (µg g–1) 49.7a (49 ± 3)b 50 ± 4.6 101 72.9Cr (µg g–1) 332 ± 9 367 ± 1 110 1170Cs (µg g–1) (0.24 ± 0.15)b ND ND 1.19Ga (µg g–1) 17.4a 16.35 ± 0.07 91 3.56Gd (µg g–1) (3.2 ± 0.4)b 2.14 ± 0.08 67 < 0.2Hf (µg g–1) 1.6a (1.55 ± 0.08)b 1.22 ± 0.01 76 < 0.2Ho (µg g–1) (0.81 ± 0.01)b 0.30 ± 0.09 40 < 0.2La (µg g–1) (5.3 ± 0.4)b 4.27 ± 0.01 80 0.47Mo (µg g–1) NA 2.44 ± 0.86 ND 8.07Nd (µg g–1) (9.6 ± 1.1)b 7.02 ± 0.28 79 0.47Ni (µg g–1) (158 ± 30)b 162 ± 3 102 1840Rb (µg g–1) 1.91 ± 0.01 1.58 ± 0.86 83 0.95Sb (µg g–1) (0.30 ± 0.20)b 1.16 ± 0.78 45 0.36Sm (µg g–1) 2.79a (2.5 ± 0.2)b 1.89 ± 0.09 68 < 0.2Ta (µg g–1) (0.31 ± 0.07)b 0.24 ± 0.01 77 < 0.2U (µg g–1) 0.37a (0.31 ± 0.02)b 0.42 ± 0.26 114 < 0.2V (µg g–1) 250a (242 ± 8)b 236 ± 4.9 94 154Y (µg g–1) (19 ± 8)b 16.0 ± 0.35 84 2.14Yb (µg g–1) 2.09a (2.05 ± 0.20)b 1.58 ± 0.01 76 < 0.2Zn (µg g–1) 58a (84 ± 10)b 62.0 ± 1.41 107 48

Zr (µg g–1) (60.0 ± 1.9)b 46.0 ± 1.27 77 1.90

NA = Not Available. ND = Not Determined. a Indicative value.b Consensus value reported by Gladney (22).

168

which is often difficult to dissolvecompletely.

For rock samples, alkaline fusionseems to be the most suitable pro-cedure to obtain complete recover-ies for elements such as Cr and Ni.

XRF and ICP-OES provided com-parable results for a large numberof elements, while ICP-MS made itpossible to obtain a wider range ofchemical information, including forsome trace and rare earth elements.This is particularly important froma geological and environmentalpoint of view.

Microwave acid digestionfollowed by ICP-OES analysis wasparticularly suitable for the deter-mination of Ca, Cr, Cu, Fe, Mg, Mn,Ni, and Zn in botanical samples.The analysis of Alyssum spp. leavesand the calculation of the soil/planttransfer factor showed that thisspecies is a hyperaccumulator forNi.

Generally, the tests performedfor this study suggest that for a firstscreening of samples, XRF is moresuitable than a multielement spec-troscopic technique by virtue of itsreduced analytical times. On theother hand, ICP-OES and ICP-MScan perform better characterizationof a wider profile of elemental con-centrations, which includes bothtrace elements and rare earths.

Received June 10, 2003.

REFERENCES

1. UNI CEI ENV 13005, Guidaall’incertezza di misura (2002).

2. EURACHEM/CITAC GUIDE: 2000,Quantifying Uncertainty in Analyti-cal Measurement, Laboratory ofGovernment Chemist, Second Edi-tion (2000).

3. Ph. Quevauviller, R. Herzig, and H. Muntau, Sci Tot. Environ. 187,143 (1996).

4. M. Bettinelli, C. Baffi, G.M. Beone,and S. Spezia, At. Spectrosc.21(2), 60 (2000).

5. Ph. Quevauviller, D. vanRenterghem, H. Muntau, and B.Griepink, Int. J. Environ. Anal.Chem. 53, 233 (1993).

6. M. Bettinelli, S. Spezia, and G. Bizzarri, At. Spectrosc. 17, 133(1996).

7. H.M. Kingston and L.B. Jassie, Intro-duction to microwave samplepreparation Theory and practice,American Chemical Society, Wash-ington, D.C., 772 pp (1998).

8. H. Matusiewicz, R.E. Sturgeon, andS.S. Berman, J. Anal. At. Spectrom.4, 323 (1989).

9. C. Minoia, M. Bettinelli, A. Ronchi,and S. Spezia, (a cura di), Appli-cazioni dell’ ICP-MS nel laboratoriochimico e tossicologico, Morganed. Tecniche, 450 pp. (2000).

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

11. R. Djingova and J. Ivanova, Talanta57, 821 (2002).

12. M. Bettinelli, C. Baffi, G.M. Beone,and S. Spezia, At. Spectrosc. 21(2),50 (2000).

13. A. Nassisi, Suoli tipici in: Guidaall’escursione pedologica, Con-vegno Nazionale Società Italianadella Scienza del Suolo, Parva Natu-ralia 20 (2002).

14. C. Baffi, M. Bettinelli, G.M. Beone,and S. Spezia, Chemosphere 48,299 (2002).

15. ISO (International Organization forStandardization) 5725. Precision oftest methods – Determination ofrepeatability and reproducibilityfor a standard test method by inter-laboratory tests. In: ISO StandardHandbook 3, Statistical Methods,3rd ed. Genève, Switzerland(1986).

16. G.T. Vernimont, Interlaboratoryevaluation of an analytical process.In: Spendley W, editor. Use of sta-tistics to develop and evaluate ana-lytical methods 3rd ed. AOAC.Arligton, VA, USA, 183 pp (1990).

17. M. Bettinelli, M. Perotti, S. Spezia,C. Baffi, G.M. Beone, F. Alberici, S. Bergonzi, C. Bettinelli, P. Can-tarini, and L. Mascetti, Microchem-ical Journal 73, 131 (2002).

18. R.R. Brooks and C.C. Radford,Nickel accumulation by Europeanspecies of the genus Alyssum,Proc. R. Soc. London, B 200, 217(1978).

19. O. Vergnano Gambi, R.R. Brooks,and C.C. Radford, L’accumulo di nichel nelle specie italiane delgenere Alyssum, Webbia 33(2),269 (1979).

20. P. Pelosi, R. Fiorentini, and C.Galoppini, On the nature of nickelcompounds in alyssum bertolonii,Desv. II, Agr. Biol. Chem. 40(8),1641 (1976).

21. O. Vergnano Gambi, L. Pancaro,and C. Formica, Investigations on a nickel accumulating plant:Alyssum bertolonii, Desv. I. Nickel,calcium and magnesium contentand distribution during growth,Webbia 32(1), 175 (1977).

22. E.S. Gladney, B.T. O’Malley, I. Roelandts, and T.E. Gills, Compilation of elemental concen-tration data for NBS clinical, biological, geological, and environ-mental standard reference materi-als, NBS Spec. Publ. 260–11(1987).

Atomic SpectroscopyVol. 24(5), September/October 2003

*Corresponding author.e-mail: [email protected] FAX: +86-27-67867955

169

Cloud Point Extraction Preconcentration and ICP-OES Determination of

Trace Chromium and Copper in Water SamplesJing Li, *Pei Liang, Taqing Shi, and Hanbing Lu

College of Chemistry, Central China Normal University, Wuhan 430079, P. R. China

INTRODUCTION

Several analytical techniquessuch as atomic absorptionspectrometry (AAS), inductivelycoupled plasma optical emissionspectrometry (ICP-OES), and induc-tively coupled plasma mass spec-trometry (ICP-MS) are available forthe determination of trace metalswith sufficient sensitivity for mostapplications. However, the determi-nation of trace metal ions in naturalwaters is difficult due to variousfactors, particularly their low con-centrations and matrix effects. Pre-concentration and separation cansolve these problems and lead to a higher confidence level and easydetermination of the traceelements. There are many methodsof preconcentration and separationsuch as liquid-liquid extraction(LLE) (1), ion-exchange techniques(2), co-precipitation (3), sorptionon the various adsorbents such aschelate resin, silica gel, andactivated carbon, etc. (4–7).

Separation and preconcentrationbased on cloud point extraction(CPE) is becoming an importantand practical application in the useof surfactants in analytical chem-istry (8,9). The technique is basedon the property of most non-ionicsurfactants in aqueous solutions toform micelles and become turbidwhen heated to a temperatureknown as the cloud point tempera-ture. Above the cloud point, themicellar solution separates into asurfactant-rich phase of a small vol-ume and a diluted aqueous phase,in which the surfactant concentra-

tion is close to the critical micellarconcentration (CMC). Any analytesolubilized in the hydrophobic coreof the micelles will separate andbecome concentrated in the smallvolume of the surfactant-rich phase.

The use of preconcentrationsteps based on phase separation bycloud point extraction (10,11)offers a convenient alternative tomore conventional extraction sys-tems. The small volume of the sur-factant-rich phase obtained withthis methodology permits thedesign of extraction schemes thatare simple, economical, highly effi-cient, speedy, and of lower toxicityto the environment than thoseextractions that use organicsolvents. CPE also provides results

comparable to those obtained withother separation techniques.Accordingly, any species that inter-acts with the micellar system,either directly (generally hydropho-bic organic compounds) or after aprerequisite derivatization reaction(e.g., metal ions after reaction witha suitable hydrophobic ligand) maybe extracted from the initial solu-tion and may also be preconcen-trated.

Cloud point methodology hasbeen used to separate and precon-centrate organic compounds as a step prior to their determinationin hydrodynamic analytical systemssuch as liquid chromatography(12,13) and capillary electrophore-sis (14). The phase separation phe-nomenon has also been used for the extraction and preconcentra-tion of metal ions after the forma-tion of sparingly water-solublecomplexes (15-18). U (19), Er (20),and Gd (21) were determined byspectrophotometry, Pd (22) byroom temperature phosphores-cence, Cu (23), Cd (24), Ni and Zn(25), Cr(III) and Cr(VI) (26) byFAAS after CPE using complexingagents.

In the present work, we reportthe results obtained in a study ofcloud point preconcentration ofchromium and copper after the formation of a complex with 8-hydroxyquinoline (8-Ox), andlater analysis by inductively cou-pled plasma optical emission spec-trometry using Triton® X-100 asthe surfactant. The proposedmethod was applied to the determi-nation of chromium and copper inwater samples.

ABSTRACT

A new method is proposed forthe determination of tracechromium and copper in watersamples by inductively coupledplasma optical emissionspectrometry after cloud pointextraction (CPE). The method isbased on the complexation ofmetal ions with 8-hydroxyquino-line in the presence of non-ionicmicelles of Triton X-100. Theeffect of experimental conditionssuch as pH, concentration ofchelating agent and surfactant,equilibration temperature andtime on cloud point extractionwas studied. Under the optimumconditions, the detection limitswere 1.29 and 1.31 ng mL-1 forchromium and copper, respec-tively, with relative standard devi-ations of 4.3% and 4.8%,respectively (n=11). Theproposed method was applied tothe determination of tracechromium and copper in watersamples with satisfactory results.

170

EXPERIMENTAL

Instrumentation

An Optima™ 2000 DVinductively coupled plasma optical emission spectrometer(PerkinElmer Life and AnalyticalSciences, Shelton, CT, USA) wasused. The operating conditions andthe analytical wavelengths are sum-marized in Table I. The pH valueswere controlled with a Model 320-SpH meter (Mettler Toledo Instru-ments Co. Ltd., Shanghai, P.R.China) supplied with a combinedelectrode. A Model 80-2 centrifuge(Changzhou Guohua Electric Appli-ance Co. Ltd., Changzhou, Jiangsu,P.R. China) was used to acceleratethe phase separation.


The non-ionic surfactant TritonX-100 was obtained from Amrescoand was used without further purifi-cation. Stock standard solutions ofchromium and copper at a concen-tration of 1000 µg mL–1 wereobtained from the National Instituteof Standards (Beijing, P. R. China).Working standard solutions wereobtained by appropriate dilution of the stock standard solutions. A 1.2x10–2 mol L–1 solution of 8-Oxwas prepared by dissolution inethanol from the commerciallyavailable product. A stock buffersolution (0.1 mol L–1) was preparedby dissolving appropriate amountsof Na2B4O7 .10H2O in doubly dis-tilled water. All other reagents wereof analytical reagent grade and allsolutions were prepared in doublydistilled water. The pipettes andvessels used for trace analysis werekept in 10% nitric acid for at least24 h and subsequently washed fourtimes with doubly distilled water.

Procedures

For the CPE, aliquots of 10 mL ofa solution containing the analytes,Triton X-100 and 8-Ox buffered at a suitable pH were kept in the ther-

mostatic bath maintained at 100oCfor 25 min. Since the surfactantdensity is 1.37 g mL–1, the surfac-tant-rich phase can settle throughthe aqueous phase. The phase sepa-ration was accelerated by centrifug-ing at 3000 rpm for 5 min. Aftercooling in an ice-bath, the surfac-tant-rich phase became viscous andwas retained at the bottom of thetube. The aqueous phases can read-ily be discarded simply by invertingthe tubes. In order to decrease theviscosity and facilitate sample hand-ing prior to the ICP-OES assay, 1.0 mL 0.1 mol L–1 HNO3 was addedto the surfactant-rich phase. Thefinal solution was introduced intothe nebulizer of the spectrometerby conventional aspiration.

Analysis of Water Samples

Prior to the above preconcentra-tion procedure, all water sampleswere filtered through a 0.45-µmpore size membrane filter toremove suspended particulate mat-ter and were then stored at 4oC inthe dark. To a 10-mL water sample, 1.0 mL of a solution containing 3.0 g L–1 Triton X-100 and 1.2x10–2

mol L–1 8-Ox plus 1.0 mL of 0.1 mol L–1 Na2B4O7 . 10H2O buffersolution (pH 9.0) was added. Afterphase separation, 1 mL 0.1 mol L–1

HNO3 was added to the surfactant-rich phase. The samples wereassayed as described in the previ-ous section.


Effect of pH

Cloud point extraction ofchromium and copper wasperformed in different pH buffersolutions. The separation of metalions by the CPE method involvesprior formation of a complex withsufficient hydrophobicity to beextracted into the small volume ofsurfactant-rich phase, thus obtain-ing the desired preconcentration.The extraction yield depends onthe pH at which complex formationoccurs.

Figure 1 shows the effect of pHon the extraction of chromium andcopper complexes. It was foundthat extraction was quantitative forchromium and copper in the pHrange of 7.0–10.0. Hence, a middlerange of pH at 9.0 was chosen forthese analytes.

Effect of Buffer Concentration

The influence of buffer amountwas carried out in which the otherexperimental variables remainedconstant. The results showed thatabove 1.0 mL of buffer solution,added to 10 mL of solution, no vari-ation took place in the extractionyield. Thus, 1.0 mL buffer solutionwas added in all subsequent experi-ments.

TABLE IICP-OES Instrumental Operating Conditions and

Wavelengths of Emission Lines Examined

Parameters

RF power 1.3 kWPlasma gas (Ar) flow rate 15 L min–1

Auxiliary gas (Ar) flow rate 1.0 L min–1

Nebulizer gas (Ar) flow rate 0.5 L min–1

Pump flow rate 0.8 mL min–1

Observation height 15 mm

Wavelength Cr 283.6 nm; Cu 324.8 nm

171

Vol. 24(5), Sept./Oct. 2003

Effect of 8-Ox Concentration

For the two cations studied, 10 mL of a solutioncontaining 0.5 µg of Cr and Cu in 3.0 g L–1 Triton X-100 and at a medium buffer pH of 9.0, containing vari-ous amounts of 8-Ox, were subjected to the cloudpoint preconcentration process. The extraction yieldas a function of the concentration of the complexingagent is shown in Figure 2. The yield increases up toan 8-Ox concentration of 8x10–4 mol L–1 for Cr and Cuand reaches near quantitative extraction efficiency. A concentration of 1x10–3 mol L–1 was chosen toaccount for other extractable species.

Effect of Triton X-100 Concentration

A successful CPE would be that which maximizesthe extraction efficiency through minimizing thephase volume ratio; thus, maximizing its concentratingfactor. The variation in extraction efficiency of Cr andCu within the Triton X-100 range of 0.1–6.0 g L–1 wasexamined and the results are shown in Figure 3. Quan-titative extraction was observed when the Triton X-100 concentration was above 3.0 g L–1. Thus, aconcentration of 3.0 g L–1 was chosen as the optimumsurfactant concentration in order to achieve the high-est possible extraction efficiency.

Effects of Equilibration Temperature and Time

It was desirable to employ the shortest equilibrationtime and the lowest possible equilibration temperatureas a compromise between completion of extractionand efficient separation of phases. The dependence ofextraction efficiency upon equilibration temperatureand time was studied with a range of 60–120oC and5–30 min, respectively. The results showed that anequilibration temperature of 100oC and an equilibra-tion time of 25 min was adequate to achieve quantita-tive extraction.

Interferences

In view of the high selectivity provided by ICP-OES,the only interferences studied were those related to thepreconcentration step. Cations that may react with 8-Oxand extracted to the micelle phase were studied. Theresults are shown in Table II. It can be seen that the pres-ence of major cations and anions has no obviousinfluence on CPE under the selected conditions.

Fig. 1. Effect of pH on the extraction recovery:50 ng mL–1 Cr and Cu, 1x10–3 mol L–1 8-Ox, 3.0 g L–1 Triton X-100.

Fig. 2. Effect of 8-Ox concentration on the extraction recovery:50 ng mL–1 Cr and Cu, 3.0 g L–1 Triton X-100, pH 9.0.

Fig. 3. Effect of Triton X-100 concentration on the extractionrecovery:50 ng mL–1 Cr and Cu, 1x10–3 mol L–1 8-Ox, pH 9.0.

TABLE IITolerance Limits of Coexisting Ions for the

Studied Element Determinations

Coexisting Tolerance Ions Limit of Ions

K+, Na+, Ca2+, Ba2+ 1000Mg2+, Cd2+,Co2+, Ni2+, Pb2+ 100

Mn2+, Zn2+ 10

172

Detection Limits and Precision

According to the definition ofIUPAC, the detection limits (3σ) of this method were 1.29 and 1.31ng mL–1 for chromium and copper,respectively, with relative standarddeviations (RSD) of 4.3% and 4.8%,respectively (n=11).

Determination of Chromiumand Copper in Water Samples

The proposed method wasapplied to the determination of Crand Cu in tap water and lake water.For this purpose, 10 mL of each ofthe samples was preconcentratedwith 3.0 g L–1 Triton X-100 and1x10–3 mol L–1 8-Ox following theproposed method. The results areshown in Table III. For calibrationpurposes, the working standardsolutions were subjected to thesame preconcentration procedureas used for the analyte solutions.

In addition, recovery experi-ments for different amounts of Crand Cu were carried out. Theresults given in Table III indicatethat the proposed method can bereliably used for the determinationof these metal ions in variouswaters.

CONCLUSION

Cloud point extraction offers asimple, rapid, inexpensive, andnon-pollution methodology for pre-concentration and separation oftrace metals in aqueous solutions.Triton X-100 was chosen for theformation of the surfactant-richphase due to its excellent physico-chemical characteristics: low CPtemperature; high density of thesurfactant-rich phase, which facili-tates phase separation easily bycentrifugation; commercial avail-ability and relatively low price; andlow toxicity. 8-Ox is a very stableand fairly selective complexingreagent. The surfactant-rich phasecan be introduced into the nebu-lizer of an inductively coupledplasma optical emission spectrome-

ter by conventional aspiration afterdissolving with 0.1 mol L–1 HNO3.The proposed method gives verylow LODs and good RSDs, and canbe applied to the determination oftrace metals in various water sam-ples.

Received June 12, 2003.

REFERENCES

1. P.L. Malvankar and V.M. Shinde,Analyst 116, 1081 (1991).

2. S.Y. Bae, X. Zeng, and G.M. Murray,J. Anal. At. Spectrom. 13, 1177(1998).

3. L. Elçi, U. Sahin, and S. Öztas,Talanta 44, 1017 (1997).

4. S. Güçer and M. Yaman, J. Anal. At.Spectrom. 7, 179 (1992).

5. J.M. Gladis, V.M. Biju, and T.P. Rao,At. Spectrom. 23, 143 (2002).

6. M. Soylak, U. Sahin, and L. Elçi,Anal. Chim. Acta 322, 111 (1996).

7. B. Mohammad, A.M. Ure, and D.Littlejohn, J. Anal. At. Spectrom. 7,695 (1992).

8. W.L. Hinze and E. Pramauro, Crit.Rev. Anal. Chem. 24, 133 (1993).

9. A. Sanz-Medel, M.R. Fernandez de laCampa, E.B. Gonzalez, and M.L.Fernandez-Sanchez, Spectrochim.Acta 54B, 251 (1999).

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

11. F.H. Quina and W.L. Hinze, Ind.Eng. Chem. Res. 38, 4150 (1999).

12. A. Eiguren Fernández, Z. Sosa Fer-rera, and J.J. Santana Rodriguez,Analyst 124, 487 (1999).

13. R.L. Revia and G.A. Makharadze,Talanta 48, 409 (1999).

14. R. Carabias-Martínez, E. Rodríguez-Gonzalo, J. Dominguez-Alvarez,and J. Hernández-Méndez, Anal.Chem. 71, 2468 (1999).

15. H. Watanabe and H. Tanaka, Talanta 25, 585 (1978).

16. H. Watanabe, T. Kamidate, S.Kawamorita, K. Haraguchi, and M.Miyajima, Anal. Sci. 3, 433 (1987).

17. T. Saitoh, Y. Kimura, T. Kamidate,H. Watanabe, and K.Haraguchi,Anal. Sci. 5, 577 (1989).

18. H. Watanabe, T. Saitoh, T. Kamidate, and K. Haraguchi,Mikrochim. Acta 106, 83 (1992).

19. M.E.F. Laespada, J.L. Perez Pavón,and B. Moreno Cordero, Analyst118, 209 (1993).

20. M.F. Silva, L. Fernández, R. Olsina,and D. Stacchiola, Anal.Chim. Acta 342, 229 (1997).

21. M.F. Silva, L. Fernández, and R.Olsina, Analyst 123, 1803 (1998).

22. S. Igarashi and K. Endo, Anal. Chim.Acta 320, 133 (1996).

23. C.H. Wang, D.F. Martin, and B.B.Martin, J. Environ. Sci. Health A 31,1101 (1996).

24. C. García Pinto, J.L. Pérez Pavón, B. Moreno Cordero, E.RomeroBeato, and S. García Sánchez,J. Anal. At. Spectrom. 11, 37(1996).

25. M.C. Cerrato Oliveros, O. Jimenesde Blas, J.L. PérezPavón, and B.Moreno Cordero, J. Anal. At. Spectrom. 13, 547 (1998).

26. E.K. Paleologos, C.D. Stalikas, S.M. Tzouwara-Karayanni, G.A.Pilidis, and M.I. Karayannis, J. Anal.At. Spectrom. 15, 287 (2000).

TABLE IIIDetermination of Cr and Cu in Natural Water Samples (n = 3)

Samples Added (ng mL–1) Measured (ng mL–1) Recovery (%)Cr Cu Cr Cu Cr Cu

Tap water 0 0 6.39 10.3 – –10 10 16.6 20.1 102 9820 20 26.5 30.2 101 99.5

Lake water 0 0 8.69 3.71 – –10 10 18.5 13.6 98.1 98.9

20 20 28.8 23.8 101 100

173Atomic SpectroscopyVol. 24(5), September/October 2003

and latent. In children from five tothirteen years of age, selenium defi-ciciency causes Kashin-Beck dis-ease, whose main symptoms arearticulation deformation and hyper-trophy, and hyaline cartilage degen-

INTRODUCTION

Selenium is an element thatbelongs to group VIA of the Peri-odic Table of the Elements. With anatomic mass of 78.96 Daltons, itshows intermediate propertiesbetween metals and non-metals. Itcan be found in four states of oxida-tion: 0, 2–, 4+ and 6+. The valenceof 6+ is the most common in inor-ganic selenium salts, while Se2– isthe most frequent for organic sele-nium compounds.

Selenium was discovered byBerzelius in 1817. It is not a verycommon element in nature; itsconcentration in the earth’s crustscarcely reaches 0.09 mg/kg, eventhough in some areas it can exist inconcentrations high enough to pro-duce livestock poisoning. The mainsources of exposure are industrialemission, principally from electri-cal, metal treatment, glass, paintand varnish production, and thechemical industries (1–3).

Selenium is an essential elementto human life. A deficiency in sele-nium causes an endemic cardiomy-opathy, the so-called Keshansyndrome, affecting mainlychildren and women during theirchild-bearing years, with symptomssuch as muscular pain, striatedmusculature degeneracy, conges-tive cardiomyopathy and, finally,multifocal myocardium necrosis.Once the illness has beendiagnosed, the ingestion ofselenium is of no therapeutic value.There are four different types ofthis illness, classified by their seri-ousness: acute, subacute, chronic,

eration. Other problems related toselenium deficiency have beenreported such as neutrofile bacteri-cide capacity depletion, and reduc-tion of 5-diiodine-ase enzymeactivity (responsible for triiodothy-ronine production from thyroxine)(4,5).

Selenium deficiency canproduce malfunction in differentorgans such as the brain, cardiovas-cular system, liver, and muscles.Epidemically, chronic deficiencyhas been associated to certain typesof cancer, while carcinogenesisinhibitory action of the upper sele-nium levels has been found in someexperimental models (6). Also, sele-nium depletion in blood has beenrelated to some neurological degen-erative illnesses similar toAlzheimer’s (7).

In blood, selenium is distributedin the following way: between 40 to 45% in plasma and 55 to 60%in red blood cells, with serum and plasma concentrations approxi-mately the same. Erythrocyte sele-nium regeneration is a relativelyslow process that lasts more or less120 days. In plasma it is muchfaster, which can be a keycontributing factor for the controlof selenium ingestion. In erythro-cytes, most of the selenium isbound to hemoglobin andglutathione peroxidase. The activ-ity of this enzyme represents amajor part of total enzymatic bloodactivity, but only 3% of plasmaenzymatic activity (8–10).

Selenium determination in bio-logical samples could be of clinicalimportance in cases of excessive orinappropriate amounts of ingestionof this element. Most published

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

ABSTRACT

Selenium is an essential ele-ment to human life. Deficiency inselenium causes an endemic car-diomyopathy, also called Keshansyndrome and, in children, theKashin-Beck disease. Therefore,selenium determination in blood,serum, and biological samples isof clinical importance.

A rapid method of analysis,based on Electrothermal AtomicAbsorption Spectroscopy withZeeman Background Correction(ETAAS), has been developed.During the method developmentprocess, the modifier concentra-tion (palladium nitrate) has beenoptimized, as well as the mineral-ization and atomization tempera-tures.

A sensitive method has beenobtained, with a detection limitin the samples of 0.4 µmol/L, acharacteristic mass of 0.49 µmol,and a linear range of 5.07 µmol/L.Accuracy and precision determi-nations for the blood sampleswere carried out using SeronormTrace Elements Whole Blood ref-erence samples. An average preci-sion and a recovery of 4.5% and102.3%, respectively, wereobtained. For serum samples, thestandard addition technique wasused to evaluate precision, result-ing in a maximum value of 1.7%.Accuracy was calculated by par-ticipating in an interlaboratoryproficiency study, conducted bySurrey University in Guilford, Sur-rey, United Kingdom, ExternalQuality Assessment Scheme.

A Rapid Method for the Determination of Selenium in Blood and Serum by ETAAS

With Zeeman Background Correction*Arturo Montel, José Manuel de Pradena, José María Gervas, Guadalupe Álvarez Bustamante,

and José Luis López-ColónInstituto de Medicina Preventiva del Ejército, Hospital Central de la Defensa,

Glorieta del Ejército s/n, 28047 Madrid, Spain

174

studies use erythrocytes, wholeblood or plasma as short-term indi-cators of selenium ingesta. Mainlyplasma or serum samples are usedin today’s clinical monitoring.Acute poisoning produces nervous-ness, dizziness and convulsions,while chronic poisoning causesanemia, mucosae irritation, pale-ness, gastrointestinal problems,hepatosplenomegaly, blindness,weakness of limbs, and respiratoryproblems, also causing a character-istic ‘garlic’ odor in breath, perspi-ration, and urine (11).

Currently, the method of choicefor selenium determination is Elec-trothermal Atomic AbsorptionSpectrometry with Zeeman Back-ground Correction (12-14). Thehigh volatility of selenium and itscompounds in complex matricesmakes the addition of a matrixmodifier indispensable to fix sele-nium and to reach pyrolysis tem-peratures as high as possible,enabling the removal of the matrixwithout selenium loss (15).

In the past, most methods usedfor the determination of seleniumused nickel as the matrix modifier(11). However, current methodsuse Pd (13,16), Pd/Mg, ascorbicacid, or hydroxylamine hydrochlo-ride (J. Fischer, 1998) (15,17–19).

EXPERIMENTAL

Instrumentation

PerkinElmer AAnalyst™ 600Electrothermal Atomic Absorption

Spectrometer, with LongitudinalZeeman Background Corrector andAS-800 Autosampler (PerkinElmerLife and Analytical Sciences, Shel-ton, CT USA).

Pyrolytically coated graphitetubes with transversely heated integrated platforms and end caps.

PerkinElmer AA WinLab™ software, Version 4.2.

PerkinElmer selenium lamp, EDL type.

Reagents

Ultrapure water (resistivity 18MΩ x cm) purified with a Milli-Q™system (Millipore, Bedford, MA,USA), for solution preparation.

Triton® X-100 (Merck, Darm-stadt, Germany).

Nitric acid, Suprapur® grade(Merck, Darmstadt, Germany).

Selenium aqueous standard solu-tion, 12.6 µmol/L (1 g/L) in 2%HNO3, for atomic absorption spec-troscopy (PerkinElmer).

Palladium modifier for graphitefurnace, 10g/L of Pd(NO3)2 in 15%HNO3 (Merck, Darmstadt,Germany).

Internal quality control samples:Nycomed (Seronorm TraceElements Whole Blood) for bloodand Nycomed (Seronorm TraceElements Serum) for serum. Inboth samples, the theoretical val-ues and acceptability ranges wereindicated.

To assure measurement trace-ability and to assess method accu-racy, the interlaboratoryproficiency study of Surrey Univer-sity, "External Quality AssessmentScheme," was employed.

Analytical Procedure

Experimental ConditionsTable I shows the instrumental

settings used for the seleniumdetermination in blood and serumsamples.

Sample Pretreatment Serum samples should be cen-

trifuged for 10 minutes at 2500rpm, while blood samples requirean homogenization time of at least20 minutes. Serum samples, bloodsamples, and standard solutionswere diluted 1/10 (v/v) with thematrix modifier solution, and thenhomogenized, avoiding bubblegeneration during this process.

Graphite Furnace TemperatureProgram

Table II shows the temperatureprogram used for selenium ingraphite furnace analysis. This pro-gram has been studied in order toobtain a quick method that ensuresthe best accuracy and precision.Each one of the temperature rampswas evaluated with the purpose ofdeveloping a sensitive and linearmethod for clinical ranges andapplications (20).

In ramp temperature optimiza-tion studies, elimination of thematrix effect caused by the pres-ence of iron was taken into

TABLE IInstrumental Operating

Conditions

Wavelength 196 nmSlit width 0.7 nmEDL lamp current intensity 290 mAReading Peak areaIntegration time 3 s

Sample volume 20 µL

TABLE II Graphite Furnace Temperature Program

Temp. Ramp Time Flow Gas Reading(ºC) (s) (s) (mL/min)

1 110 1 20 250 Argon No2 130 1 30 250 Argon No3 1250 5 30 250 Argon No4 2000 0 3 0 Argon Yes

5 2500 1 3 250 Argon No

175

Vol. 24(5), Sept./Oct. 2003

account. This element causes lamp energy absorption ata wavelength close to that of selenium, producing a highmatrix interference that even Zeeman correction cannottotally compensate for. Optimization of the analyticalconditions, selection of the optimum concentration ofthe matrix modifier, and the temperatures of the mineral-ization and atomization stages are described below.


Optimization of the Matrix Modifier Solution The matrix modifier solution was prepared by adding

the following amounts to a 100-mL volumetric flask: 1 mL of palladium modifier for graphite furnace analysis[10 g/L Pd(NO3)2 in 15% HNO3 solution] plus 0.1 mL ofTriton® X-100, then bringing the solution to volumewith ultrapure water.

Figure 1 shows the dependence of absorbance withthe concentration of palladium in the matrix modifiersolution. The greater the palladium concentration, thehigher the absorbance until reaching a concentration of72 mg/L corresponding to 1.44 µg of palladium in eachinjection of sample (20 µL), where a stabilization can beobserved. As a conservative approach, an optimum valueof 90 mg/L (1.8 µg / 20 µL sample) was selected.

Optimization of Mineralization Stage

The pyrolysis temperature selected was based on theone that permitted the best organic matter destructionwithout loss of analyte, and at the same time obtainedminimum matrix effect and higher signal-to-noise ratio(see Figure 2).

Optimization of Atomization Stage

The temperature allowing both a higher response anda greater relation between analyte and background sig-nals (see Figure 3) was selected.

Linearity

Calibration standards were prepared from a concen-trated standard solution of 12.6 µmol/L (1g/L), initiallydiluted with 1% HNO3 (v/v) in a volumetric flask until a concentration of 0.13 µmol/L (10 mg/L) was achieved.This solution is stable for at least three months in refrig-erated conditions. Working standards of 0.63, 1.27, 2.53,5.07, and 7.60 µmol/L (50, 100, 200, 400 and 600 µg/L,

Fig. 1. Dependence of absorbance vs. modifier concentration.An aqueous standard with a final concentration of 0.125µmol/L was used.

Fig. 2. Dependence of analyte absorbance (continuous line)and background absorbance (dotted line) vs. pyrolysis tem-perature with a fixed temperature of atomization at 2000ºC.A Seronorm Trace Elements Whole Blood "level 1" sample wasused.

Fig. 3. Dependence of analyte absorbance vs. atomizationtemperature with a fixed temperature of pyrolysis at 1250ºCin a Seronorm Trace Elements Whole Blood "level 1" sample.

176

respectively) were prepared bydiluting this solution with ultrapurewater. These working standardswere diluted 1/10 (v/v) with thematrix modifier solution in order to handle the standards in the sameway as the samples. As reagentblank solution, ultrapure waterdiluted 1/10 (v/v) with the matrixmodifier solution was used.

The linearity of the calibrationcurve was evaluated analyzing theformer working standard solutionsdiluted as samples in order toobtain a final range within 0 and0.76 µmol/L (0 to 60 µg/L). A linearresponse was observed up to a con-centration of 0.507 µmol/L (see Figure 4). As a working curve, theconcentration range within 0 to0.507 µmol/L (0 to 40 µg/L) wasadopted. At each point, this curveshowed a maximum relative stan-dard deviation of 3.7%.

A slope of 1.51·10–4 ± 0.01·10–4

(p=0.05) was obtained. Slope vari-ance, expressed in terms ofpercentage of the mean value, was0.8%, for a probability of 95%. Thecorrelation coefficient was 0.9992.To carry out a proportionality test,confidence limits of Y-axisintercept of the line obtained withstandard solutions were calculated(95% probability). The valueobtained was 7·10–4± 6.7·10–4

(n=15, p=0.05), including zero.

The Matrix Effect

To prove that the measuredabsorbance is free of interferencesfrom the matrix, a blood sample,spiked with increasing amounts ofselenium, was analyzed. The slopeof the regression line obtained afterrepresenting the measuredabsorbances against the addedamounts was calculated.

The slope of the regression lineobtained in the measurement ofaqueous selenium standards(1.51·10–4 ± 0.01·10–4, p=0.05) fallsinside the slope confidence interval

of the line obtained using the stan-dard addition method and spikedblood samples (1.63·10–4 ±0.14·10–4, p=0.05) (see Figure 5).Thus, the sensitivity of the methoddoes not seem to be affected by thepresence of the remaining bloodconstituents.

In the addition tests, the recover-ies obtained were 95.2, 103.2, and106% for concentrations of 0.48,0.76, and 1.26 µmol/L (40, 60, and100 µg/L), respectively.

Quantification and DetectionLimits

The instrumental detection limitwas calculated from the analysis of30 reagent blank samples. A valueof 0.04 µmol/L (0.34 µg/L) wasobtained for standard solutions cor-responding to 0.4 µmol/L (3.4 µg/L)for blood and serum samples whenthe dilution proposed in this proce-dure was used. The experimentalcharacteristic mass was 495 µmol(39.1 pg), which is lower than thecharacteristic mass granted in theinstrument manufacturer’s(PerkinElmer) specifications, i.e.,570 µmol or 45 pg (21).

Fig. 4. Confidence and prediction intervals of the calibration curve obtained withaqueous standards for a probability of 95%.

Fig. 5. Comparison between the slope of the regression line (obtained with thestandard addition method, spiking selenium amounts to a blood sample) and thecalibration line. The dotted lines represent the confidence interval of the calibra-tion line (95% probability). The parallel line to that obtained with the standardaddition method falls inside this interval.

177

Vol. 24(5), Sept./Oct. 2003

Precision and Accuracy

Reference samples (SeronormTrace Elements Whole Blood "level1," Lot 404107) were employed forprecision and accuracy tests. Forthis sample, the target value was1.01 µmol/L (80 µg/L) of Se. Thetests were carried out on three different days (10 times each day);an average recovery of 102.3% wasobtained, with a precision value of4.5%. The mean values of each dayfulfil the requirements of the Stu-dent’s statistical t-test for meancomparison, with a 95%significance level (Table III).

The tests for assessing methodaccuracy for the serum sampleswere carried out on 12 serum sam-ples within the interlaboratory pro-ficiency study, conducted by SurreyUniversity, "External Quality Assess-ment Scheme." For the samplesreceived, all values obtained fellwithin the range determined by the mean plus/minus one standarddeviation of the participating labo-ratories, named Standard DeviationIndex (SDI) (see Figure 6).

Method precision for the serumsamples was studied from the analy-sis of serum samples to whichincreasing quantities of seleniumwere added in order to obtain finalconcentrations in the samples of0.88,1.39,1.89 and 2.90 µmol/L.(70, 110, 150 and 230 µg/L). Theobtained standard deviation valuesvaried between 0.94% and 1.7%.

CONCLUSION

No statistically significant differ-ence between the slopes of thelines obtained by measuring aque-ous standards and by the standardaddition method was observed.Thus, no matrix effect wasobserved which makes it possibleto calibrate the method with aque-ous standards. The calibration linecorresponding to aqueousstandards was shown to be linear inthe interval between 0 and 5.07µmol/L (400 µg/L) which fulfils the

requirements of the proportionalitytest, as the regression line passesthrough zero.

The experimental detection limitof the method for serum and bloodsamples was 0.4 µmol/L (3.4 µg/L).It turned out to be useful in detect-ing selenium deficiencies, both inserum and blood, as a result of theconsiderable difference among thedetection limit and the referenceselenium values in the above-men-tioned fluids.

The tests carried out for the ref-erence blood sample, Seronorm,and in the interlaboratoryproficiency study for serum sam-ples indicate that the methodshows suitable precision and accu-racy values.

Received August 19, 2002.Revised text received August 12, 2003.

TABLE IIIDetermination of Se in a Sample With a Known Se Concentration:

Seronorm Whole Blood "level 1"(Selenium concentration: µmol/L)

1st day 2nd day 3rd day

1.06 1.03 1.001.01 0.95 1.031.00 0.95 1.061.10 1.10 1.001.10 1.01 1.071.03 1.10 1.051.02 1.07 1.011.06 1.01 1.021.05 1.07 0.98

0.92 1.07 1.03 Total

Mean 1.03 1.03 1.02 1.03RSD (%) 5% 5.5% 2.8% 4.5%Recovery 102.5% 102.6% 101.6% 102.30%Practical "t" 1.567 1.512 1.807

Theoretical "t" (p=0.05) 2.262 2.262 2.262

Fig. 6. Standard Deviation Index (SDI) for the serum samples of the External Qual-ity Assessment Scheme of the University of Surrey.

178

REFERENCES

1. R.R. Lauwerys, "Selenio," In: Toxicologia industrial e intoxica-ciones profesionales. Ed. Masson(1994).

2. C. Gonzalez, "Elementos traza esen-ciales en la practica clinica," Rev.Quimica Clinica (April 9–11,1994).

3. J. Ladron de Guevara and V. Moya,"Selenio y sus compuestos," In:Toxicología Médica Clínica y Laboral, 1st Ed., McGraw-Hill-Interamericana (1995).

4. OMS, "Selenium," Aspects sanitaireset nutritionnels des oligoelementset des elements en traces, pp. 105– 222 (1997).

5. M. Kantola and T. Vartiainent, J. Trace Elements Med Biol 15,11–17 (2001).

6. J.M. Queralto, "Toxicologia," In: Bioquímica Clínica, Ed. Barcanova(1994).

7. J.F. Emard, J.P. Thouez, and D. Gau-vreau, Soc. Sci. Med. 40(6), 847–58(1995).

8. S.G. Patching and P.H.E. Gardiner, J. Trace Elements Med. Biol. 13,193–214 (1999).

9. S.H. Bugel, B. Sandtrom, and E.H.Larsen, J. Trace Elements Med.Biol. 14, 198–204 (2001).

10. A. Zachara, U. Trafikowska, A.Adamowicz, E. Nartowicz, and J. Manitius, J. Trace Elements Med.Biol. 15, 161-166 (2001).

11. T.M.T. Sheehanl and D. J. Halls,Ann. Clin. Biochem. 36, 301–315(1999).

12. M. Sanz and C. Diaz, Chem. Rev.227–257 (1995).

13. P. Bermejo, M.J. Lorenzo, A.Bermejo, J.A. Cocho, and J.M.Fraga, Forensic Science Interna-tional 107, 149–156 (1999).

14. M. Feuerstein and G. Schlemmer,At. Spectrosc. 20(5), 180–1855(1999).

15. E. Sabe, R. Rubio, and L. Garcia-Beltran, Anal. Chim. Acta 398, 279-287 (1999).

16. P.H.E. Gardiner, D. Littlejohn, D.J. Halls, and G.S. Fell, J. TraceElements Med. Biol. 9, 74–81(1995).

17. J.L. Ficher and C.J. Rademeyer,Spectrochim. Acta 53, 537-548(1998).

18. J. Ficher and C.J. Rademeyer, Spectrochim. Acta 53, 549–567(1998).

19. R. Sabe, R. Rubio, and L. Garcia-Beltran, Anal. Chim. Acta 419,121–135 (2000).

20. A. Montel, J.L. Lopez Colon, J.M. De Pradena, G. Alvarez, and T. Hoya, Quím. Clín. 17, 88 (1998).

21. The THGA graphite furnace: Tech-niques and recommended condi-tions, PerkinElmer Life andAnalytical Sciences, Shelton, CTUSA (1992).



Flow Injection AAS Determination of Cd, Cu, and Pb atTrace Levels in Wine Using Solid Phase Extraction

Andréa Pires Fernandes, *Mercedes de Moraes, andJosé Anchieta Gomes Neto

Instituto de Química, Universidade Estadual PaulistaP.O. Box 355, 14801-970 Araraquara-SP, Brazil

INTRODUCTION

Flame atomic absorption spec-trometry (FAAS) has for many yearsbeen used in the analysis of varioussamples (1). However, in somecases of direct determination oftrace metals, there is insufficientsensitivity and limit of detection(LOD) offered by FAAS unless pre-concentration steps are used.

In chemical analysis, the maindifficulties associated with manualsample preparation are time andreagent consumption, the possibil-ity of sample contamination, loss ofanalyte, or risk to the analyst,amongst others. These shortcom-ings can be circumvented with on-line sample pretreatment (2,3). Theflow injection technique has madeon-line sample preparation proce-dures even more attractive becauseit is an excellent tool for solutionmanagement (2).

The coupling of FIA-FAAS forpreconcentration and the determi-nation of Cd, Cu, and Pb using solidreversed phase sorbent C18 (octade-cyl groups bonded silica gel) anddiethylammonium-N,N-diethyldithiocarbamate (DDTC)(4,5–7), or ammoniumdiethyldithiophosphate (ADTP)(7–12) as the complexing agentscan be efficiently applied to theanalysis of water (5), and biological(6) and environmental (7,8) sam-ples.

The ammonium salt of O,O-diethyldithiophosphoric acid(DDTP) was first studied in the1960s by Bode and Arnswald (13).It was demonstrated that this ligandcomplexes several transition metals

ABSTRACT

A comparative study isreported between C18 bondedsilica gel and powdered polyeth-ylene (PE) as sorbent for Cd, Cu,and Pb determination usingammonium diethyldithiophos-phate (ADTP) as the complexingagent in a flow injection system.The complexes were formed in0.14 mol L–1 HNO3 andprocessed in a simple flow sys-tem comprising a peristalticpump, a manual injector-commu-tator, and a sorbent-packed mini-column. Ethanol was selected asthe eluent and analytes in theeluate were determined by flameatomic absorption spectrometry.The optimum concentration ofthe complexing agent was 0.1%(m/v) ADTP for Cu and Pb deter-mination using either C18 or PE,and 0.25% (m/v) ADTP for Cddetermination using PE. The sam-ple loading flow rates were 5.0,3.6, and 3.0 mL min–1 for Cu, Pb,and Cd, respectively. The best

elution flow rate was 6.5 mLmin–1. For a 60-sec preconcentra-tion time, the sampling rate was40 h–1 and the enrichment factorsof 33, 36, and 11 times (C18) or18, 22, and 23 times (PE) wereobtained for Cu, Pb, and Cd,respectively. The limits of detec-tion (LOD) were 1.6 µg L–1 Cu,11 µg L–1 Pb, and 2.0 µg L–1 Cdusing C18 or 2.9 µg L–1 Cu, 19 µg L–1 Pb, and 1.0 µg L–1 Cdusing PE, respectively. The rela-tive standard deviations (n = 12)were typically <2%, <2%, and<6% for Cd, Cu, and Pb, respec-tively. The recoveries of Cd, Cu,and Pb added to wine samplesvaried from 96–99%, 97–102%,and 90–99%, respectively, usingC18 or PE. Accuracy was checkedfor Cd, Cu, and Pb determinationin six wine samples digested byblock digestor and open-vesselmicrowave-assisted digestion sys-tems. The results revealed thatC18 was more efficient for Cu andPb determination, while PE wasthe best sorbent for Cd.

and semi-metals in an acidicmedium, but does not react withalkali and alkaline earth metal ele-ments and others, such as Al, Cr,Fe(II), Ga, Ge(IV), Ir, Mn, Nb, Pt,Rh, Ru, Se(VI), Ta, Th, Ti, V, W, Zn,and Zr, depending on the natureand concentration of the acid. Thisreagent has been used for the sepa-ration and preconcentration of dif-ferent analytes such as Pb, Cu, Cd,Mo, Bi, and Ag in several samplesfor FAAS (7,11,14,15) orelectrothermal atomic absorptionspectrometry (ETAAS) detection(8,14,16–18). Since the ADTP com-plexes only As(III) and not As(V), a method for the on-line determina-tion of these species in waters byETAAS was also proposed (19). On-

line separations by complexationwith this ligand for ultra-trace multi-element determination in watersand biological samples byinductively coupled plasma massspectrometry (ICP-MS) have alsobeen proposed (20).

Although solid sorbents such ascarbon (14,16,18), C18 (5–12,15,19)and knotted polyethylene tubing(17) have been proposed as solidsupport for Cd, Cu, or Pb, more sys-tematic studies on the use of C18

and polyethylene for Cd, Cu and Pbdetermination are lacking in theliterature.

In this study, the performance ofC18 bonded silica and powderedpolyethylene-packed mini-columns

180

for on-line preconcentration of Cd,Cu, and Pb using ADTP as the com-plexing agent was evaluated. A flowinjection system coupled to a flameatomic absorption spectrometerwas used to perform the detectionautomatically. As analytical demon-stration, the determination of ana-lytes in wine samples will bepresented. The performance of theFIA procedures with C18 and PE-packed mini columns was checkedafter analyzing digested wine sam-ples and the accuracy was alsochecked using spike recovery tests.

EXPERIMENTAL

Instrumentation

A Model AAnalyst™ 100 flameatomic absorption spectrometerwas used (PerkinElmer Life andAnalytical Sciences, Shelton, CT,USA), equipped with a deuteriumarc background corrector forabsorbance measurements at thewavelengths of 228.8 nm (Cd),324.8 nm (Cu), and 283.3 nm (Pb).PerkinElmer Lumina™ hollow cath-ode lamps, operating at 4 mA, 15mA, and 10 mA for Cd (P/N N305-0115), Cu (P/N N305-0157), and Pb(P/N N305-0121), respectively,were used. The air-acetylene flamewas adjusted and the operating con-ditions were those recommendedby the manufacturer, unless speci-fied otherwise.

The FIA system comprised ahome-made manual injector-com-mutador, a sorbent-packed minicol-umn, an Ismatec IPC–8 peristalticpump equipped with Tygon®pumping tubes used to manage thecomplexing agent, sample, and ana-lytical solutions. A Viton® organicsolvent-resistant pumping tube wasused to aspirate the ethanol (elu-ent). A poly-tetrafluoroethylene tub-ing (0.8 mm i.d) was used as thereaction coil and transmission lines.Mini columns were prepared byfilling a Tygon tube (50 mm x 2.7mm i.d.) with about 30 mg C18 or60 mg of powdered polyethylene.

The ends of the packed mini-col-umn were filled with glass wool toavoid sorbent losses. This mini-column was initially flushed withethanol for cleaning.

FIA System

Figure 1 shows the FIA systemdesigned for Cd, Cu, and Pb deter-mination. It consists of two steps:preconcentration and elution/detection. In the position of injec-tor-commutator (IC) as specified inFigure 1A, the sample (S) mergeswith the ADTP reagent stream (R)at the confluent point x. Metal com-plexes formed inside the reactioncoil (RC) are retained by the C18 orPE-packed mini-column (PC). Theother metals and species that donot form complexes with ADTPpass through PC and are discharged(W). At the same time, ethanol (E)continuously flows through thenebulizer of the spectrometer.When the injector was changed tothe other position (Figure 1B), themini-column (PC) is inserted intothe eluent channel. Ethanol passesthrough the mini-column in reversedirection and the eluate containingthe analyte flows towards the detec-tor. The passage of the eluate zonethrough the flame of the spectrom-eter produces a transientabsorbance signal with the peakheight proportional to the analytecontent in the sample. After peakmaximum measurement, the injec-tor-commutator is switched back tothe initial position and anothercycle is started.

Reagents, Reference Solutions,and Samples

High-purity deionized waterusing a Milli-Q™ water purificationsystem (Millipore, Bedford, MA,USA) was used throughout. All ref-erence solutions and samples wereacidified to 0.14 mol L–1 with Supra-pur® grade nitric acid (Merck,Darmstadt, Germany). Six Cd or Cuand five Pb calibration solutions(Cd: 5.00, 25.0, 50.0, 75.0, 100, 150

µg L–1; Cu: 10.0, 50.0, 100, 150,200, 250 µg L–1, Pb: 25.0, 50.0,75.0, 100, 150 µg L–1) wereprepared daily in 0.14 mol L–1

HNO3 by diluting the individualstock solutions (1000 mg L–1;Titrisol®, Merck). Ethanol was usedas the eluent without previoustreatment.

A 1.0% (m/v) ADTP solution was prepared weekly by dissolving1.00 g of the ammonium salt(Aldrich, Milwaukee, WI, USA) inabout 5 mL of ethanol and dilutingup to 100-mL volume with water.This solution was purified by pass-ing it through a C18 bonded, silica-packed column, then stored in arefrigerator. The 0.1% and 0.25%(m/v) ADTP solutions wereprepared by dilution of the ADTPstock solution in water.

Octadecyl bonded silica (C18)from commercially available Sep-Pak cartridges (Waters, Milford, CT,USA) was used to prepare the mini-columns.

Brazilian red and white winesamples were purchased at a localsupermarket in Araraquara City,Brazil.

Preparation of Samples

The samples were prepared intriplicate by two different proce-dures involving acid mineralizationin open systems. A 100-mL sample,20 mL of concentrated HNO3, and600 µg V2O5 were sequentiallyadded to a digestion tube (21),which was transferred to theTE05/50 digestion block (Tecnal,Piracicaba, Brazil). The temperaturewas then increased to 120oC andkept steady for 8 h. All digests wereclear solutions and the final volumewas made up to 100 mL with Milli-Q™ water. The wet ashing diges-tion procedure in an open vesselsystem employed the Star System 6microwave oven (CEM, Matthews,NC, USA), equipped with glass ves-sels. A 50-mL wine sample was

181

Vol. 24(5), Sept./Oct. 2003

placed into acid-cleaned vessels and10 mL of concentrated nitric acidadded. The flasks were placed inthe microwave oven cavity and theheating program as listed in Table Iwas run. All digests were clear solu-tions and the final volume wasmade up to 50-mL volume withMilli-Q water.

TABLE IOptimized Heating Program

of Wine Digestion Using Microwave Oven

Step Ramp Temp. Hold Rea- Vol-gent ume

(min) (oC) (min) (mL)

1 5 70 5 – –

2 0.1 70 0.1 HNO3 10

3 5 95 15 – –

Optimization

The influence of variation inflow rates of the sample (S), ADTPsolution (R), and ethanol (E) on the absorbance was carried outwith an additional peristaltic pump.With this strategy, one channel wasvaried while the others were unal-tered. The rotation speed of theperistaltic pump was varied to pro-duce different flow rates of S and Rwithin the range of 1–10 and 0.2–2mL min–1, respectively. These flowrates were simultaneously changedin order to eliminate dilutioneffects.

The influence of variation of theeluent flow rate was studied from 2 to 8 mL min–1.

The influence of variation of concentration of reagent R onabsorbance was evaluated in the0.05–1% (m/v) ADTP range.

As the system was designed on a time-based injection mode, theinfluence of loading time onabsorbance was studied within the1–10 minute range.


FIA System Optimization

The FIA system as depicted inFigure 1 was studied in relation tochemical and flow variables inorder to obtain optimum analyticalconditions.

The effect of sample flow rate onabsorbance was studied in the 1–10mL min–1 range for all elements. Nosignificant measured peak heightsfor 50 µg L–1 Cu were observed for

flow rates < 1 mL min–1 using C18 orPE. The faster the rotation speed ofthe peristaltic pump, the higher thepeak height. The signals increasedsteeply with flow rates up to 10 mLmin–1 (0.12 A) and 7 mL min–1 (0.03A) for C18 and PE, respectively,above which they became constant.In these situations, the height of theanalytical signals for Cu increasedabout 10 times for C18 and 5 timesfor PE when the sample flow ratewas varied from 1 to 10 mL min–1

and from 1 to 7 mL min–1, respec-

Fig. 1. Flow diagram of the system at the sampling (A) and elution (B) positionsfor Cd, Cu, and Pb determinations. IC: injector-commutator; x: confluent point; S: sample or reference solutions (Cd:3.0 mL min–1 ; Cu: 5.0 mL min–1; Pb: 3.6 mL min–1); R: ADTP solution (Cd: 0.6 mLmin–1 ; Cu: 1.4 mL min–1; Pb: 1.0 mL min–1) ; E: ethanol (6.5 mL min–1); W: waste;RC: reaction coil (200 mm x 0.8 mm i.d.); PC: preconcentration column (50 mm x2.7 mm i.d.); FAAS: spectrometer (Cd: 228.8 nm; Cu: 324.8 nm; Pb: 283.3 nm).

182

tively. Flow rates higher than 10 mLmin–1 caused high hydrodynamicimpedance of the flow system,which resulted in solution leakingthrough the connecting points orconfluences. In general, the ratio ofmeasured absorbance from C18 andPE extractions in the same flow ratewere ~ 5. As a compromise withregard to sensitivity, sample con-sumption, and system stability, theselected flow rate of sample loadingfor Cu determination using C18 andPE was 5 mL min–1. It should bepointed out that the flow rate ofsample loading could be adjusted to10 or 7 mL min–1 if a higher sensi-tivity were to be required.

When a 100-µg L–1 Pb samplesolution was processed in the flowsystem with a C18 or PE minicolumnand a sample carrier stream of < 1 mL min–1, no significantabsorbance was observed. It wasfound that the faster the rotationspeed of the peristaltic pump, thehigher the peak height up to 4 mLmin–1 for both C18 (0.03 A) and PE(0.017 A); then the signalsdecreased steeply up to 10 mLmin–1, resulting in absorbance simi-lar to 1 mL min–1 (0.018 A for C18;0.006 A for PE). In general, the ratioof measured absorbance from C18

and PE extractions in the same flowrate was ~ 2. As a compromisebetween sensitivity and system sta-bility, the loading flow rate selectedfor Pb determination using C18 andPE was 3.6 mL min–1.

The flow rate of sample loadingis an important parameter for theCd determination in the proposedflow system with a C18 or PE mini-column. For a 50-µg L–1 Cd solution,it was found that the faster the rota-tion speed of the peristaltic pump,the higher the peak height up to 3 mL min–1 using C18 (0.075 A)and PE (0.16 A); then the signalsdecreased steeply up to 10 mLmin–1, resulting in an absorbancesimilar to that of the blank solution(0.003 A). Measurements in the

1–4 mL min–1 range revealed thatthe absorbance of the Cd solutionsextracted using PE was ~ 2 timesthat obtained with C18. The flowrate of the carrier stream selectedfor the Cd determinations with C18

and PE mini-column was 3 mL min–1.

Ethanol concentration is anotherimportant parameter. The studyshowed that diluted ethanol cannotbe used as the eluent because unsat-isfactory signals were observed dueto poor elution.

Another parameter studied wasthe flow rate of the ethanol carrierstream. The flow rate of the eluentwas varied from 3 to 8 mL min–1.Flow rates of < 5 mL min–1 resultedin a very unstable system. When thenebulizer uptake rate (5 mL min–1)exceeded the elution rate, vacuumsegments or gas bubbles formedinside the capillary that connectsthe FIA injector with the nebulizerand no repetitive transient signalswere recorded. The signalsincreased by ~ 40% when the flowrates increased from 5 to 7 mL min–1

for C18 and PE. Since higher flowrates did not significantly alter themeasured signals, the eluent (E) car-rier stream of 6.5 mL min–1 wasselected for C18 and PE.

The length of the reaction coil(RC) was investigated in the 10–50cm range. The signals increasedsteeply with coil lengths up to 20cm, above which they became con-stant. A 20-cm coil length wasselected for further experiments.

ADTP Concentration

The influence of ADTP concentra-tion on the absorbance of Cd, Cu,and Pb is shown in Figure 2. For Cddeterminations, the signalsincreased with an ADTP concentra-tion up to 0.25% (m/v) (0.22 A) and 0.05% (m/v) (0.17 A) for PE andC18, respectively, above which theybecame constant. A 0.05% (m/v)ADTP solution was sufficient for Pband Cu determination with C18.Higher ligand concentrations did not

alter the absorbance. For Cu and Pbdetermination with PE, the signalsincreased with ADTP up to 0.1%(m/v), above which they becamepractically constant.

Analytical Features

Table II shows the analyticalcharacteristics of the flow systemfor Cd, Cu, and Pb determinationwith the PE and C18 mini-column.With the main variables of the pro-cedure optimized for each sorbentmaterial, calibration curves for Cd,Cu, and Pb were obtained by sub-mitting the calibration solutions tothe proposed flow system shown in Figure 1. The enrichment factor(EF) was obtained as the ratio of theslopes of the calibration curveswith and without preconcentration,both obtained with reference solu-tions prepared in 0.14 mo1 L–1

HNO3. The influence of sorbent onsensitivity was evaluated by meansof the analytical curves in the5–150 µg L–1 Cd, 25–150 µg L–1 Pb,and 10–250 µg L–1 Cu ranges. Thecorrelations between absorbance(A) and analyte concentrationswere as follows:

Using PE: A = 3.30 10–5 + 3.05 10–-3 µg L–1 (Cd)A = 3.26 10–3 + 1.05 10–3 µg L–1 (Cu) A = -6.45 10–4 + 1.60 10–4 µg L–1 (Pb)

Using C18: A = 4.48 10–3 + 1.47 10–3 µg L–1 (Cd)A = 6.12 10–3 + 1.91 10–3 µg L–1 (Cu)A = 3.97 10–3 + 2.64 10–4 µg L–1 (Pb)

With direct aspiration of the analytical solutions containing10–2000 µg L–1 Cd, 50–5000 µg L–1

Cu, and 50–20,000 µg L–1 Pb, thecorrelations between absorbance(A) and analyte concentrationswere as follows:

For Cd: A = 1.28 10–4 + 1.34 10–4 µg L–1

For Cu: A = 2.09 10–4 + 5.74 10–5 µg L–1

For Pb: A = 1.34 10–4 + 7.26 10–6 µg L–1

183

Vol. 24(5), Sept./Oct. 2003

For a 60-sec preconcentrationtime, the sampling rate was ~ 40h–1. The enrichment factors for Cu,Pb, and Cd calculated fromabsorbance ratios using curves withand without preconcentration were33, 36, and 11 times and 18, 22, and23 times for C18 and PE,respectively. Additionally, the pre-concentration data revealed that PEwas the best sorbent for Cd whileC18 was best for Cu and Pb.

Determination of Cd, Cu, and Pbin Wine

Table III shows the results of theCd, Cu, and Pb determination withthe flow system consisting of the PEor C18 mini-column. The sampleswere mineralized in triplicatetogether with blanks which allowedfor the contribution of analytes pre-sent in the reagents in the digestionprocedures. Six table wine sampleswere analyzed in order to investi-gate the performance of C18 and PEextractors, and the results werecompared with those obtained withmicrowave-assisted digestion.According to a paired t-test, allresults were in agreement at the95% confidence level. In addition,the recoveries for the spiked sam-ples were within 96–99% for Cd,97–102% for Cu, and 90–99% for Pbusing C18 or PE. The relative stan-dard deviations for most samples(n=12) were lower than 2%, 2%,and 6% for Cd, Cu, and Pb, respec-tively.

The limit of detection (LOD)defined as the concentration ofmetal with a response equivalent tothree times the standard deviation(s) of the blank (n=12) was foundto be 1.6 µg L–1 Cu, 11 µg L–1 Pb,and 2.0 µg L–1 Cd using C18; and 2.9 µg L–1 Cu, 19 µg L–1 Pb, and 1.0 µg L–1 Cd using PE.

Fig. 2. Effect of the ADTP concentration on absorbances of 50 µg L–1 Cd, Cu, and100 µg L–1 Pb using C18 and PE. Curves correspond to the following solid extractor–analyte pairs: PE–Cd (A), C18–Cd (B), C18–Pb (C), C18–Cu (D), PE–Cu (E), and PE–Pb (F).

TABLE IIAnalytical Characteristics of the Proposed FIA System Using PE or C18

Cd Cu Pb

ADTP % (m/v) 1.0 0.1 0.1Sample flow rate (mL min–1) 3.0 5.0 3.6Ethanol flow rate (mL min–1) 6.5 6.5 6.5ADTP flow rate (mL min–1) 0.6 1.4 1.0Calibration (µg L–1) 5–150 10–250 25–150Preconcentration time (s) 60 60 60Sampling frequency (h–1) 42 40 40

C0 (µg L–1) 3.0 2.3 17C18 LOD (µg L–1) 2.0 1.6 11

Mini-column EF 11 33 36C0 (µg L–1) 1.4 4.2 27

PE LOD (µg L–1) 1.0 2.9 19

EF 23 18 22

C0 = characteristic concentration, LOD= limit of detection, EF= enrichment factor.

184

CONCLUSION

This work presents a simple,fast, and accurate method for theflow injection determination of Cd,Cu, and Pb in table wines by FAASusing on-line preconcentration withC18 or PE. The recovery data of theanalytes added to the sampledigests showed that there are nosignificant matrix effects caused bythe original matrix components orthe reagents. Analysis of the recov-ery data for the analytes added tothe original wine samples revealedthat there were no significant ana-lyte losses in both digestion proce-dures.

These studies demonstrated thatthe C18 extractor was more efficientto preconcentrate Cu-DTP and Pb-DTP complexes and determineboth elements in addition to provid-ing the highest enrichment factors,while the PE extractor was moreappropriate for Cd-DTP determina-tion. The on-line solid phase extrac-tion using C18 or PE can be used todetermine Cd, Cu, and Pb in winessince the limits of quantification(LOQ) obtained for Cd, Cu, and Pbexceeded the requirements of theBrazilian Food Regulation (DecreeNo. 55871 from the Ministry ofHealth) which established the maxi-

mum permissible levels for Cd, Cu,and Pb at 0.50 mg L–1, 5.00 mg L–1,and 0.50 mg L–1, respectively.

Received November 17, 2002.

ACKNOWLEDGMENTS

The authors thank the FAPESP(Projects 99/10486-9 and 00/14887-7) for financially supporting thiswork, the fellowship to A.P.F., andthe CNPq for fellowships toJ.A.G.N.

REFERENCES

1. B. Welz and M. Sperling, AtomicAbsorption Spectrometry, 3rd edi-tion, Wiley-VCH, Weinheim, Ger-many, pg. 941 (1999).

2. J. Ruzicka and E.H. Hansen, FlowInjection Analysis; Wiley & Sons:New York, pg. 498 (1994).

3. Z-L. Fang, Flow Injection Atomic,John Wiley & Sons, Chichester, UK(1995).

4. J. Ruzicka and A. Arnswal, Anal.Chim. Acta 216, 243 (1989).

5. Z. Fang, T. Guo, and B. Welz, Talanta38, 613 (1991).

6. M. Sperling, X. F. Yin, and B. Welz,Fresenius’ J. of Anal. Chem. 343,754 (1992).

TABLE IIIResults (n=3) for Cd, Cu, and Pb Determination in Wine Samples Analyzed by the

Proposed FIA System Using PE and C18 With Digestion in Open Systems

FIA-FAASSample Microwave Oven Digestion (µg L–1) Digestion Block (µg L–1)

Cd Cu Pb Cd Cu PbPE C18 PE C18 PE C18 PE C18 PE C18 PE C18

1 <1.0 <2.0 245(1)* 244(1)* 19.0(1.1)* 17.9(1.0)* <1.0 <2.0 243(0)* 244(2.7)* 21.3(1.3)* 20.6(0)*

2 <1.0 <2.0 230(1)* 228(0)* <19 <11 <1.0 <2.0 225(1)* 226(1)* <19 <11

3 <1.0 <2.0 216(0)* 216(0)* <19 <11 <1.0 <2.0 216(0)* 217(2.7)* <19 <11

4 <1.0 <2.0 222(1)* 221(0)* <19 <11 <1.0 <2.0 219(1)* 220(0)* <19 <11

5 <1.0 <2.0 242(1)* 242(1)* <19 22(0)* <1.0 <2.0 240(1)* 241(1)* <19 22.8(1.2)*

6 <1.0 <2.0 245(1)* 248(1)* 23(0)* 26(0)* <1.0 <2.0 244(1)* 247(1)* 22.8(1.3)* 26.6(1.5)*

*S.D.

7. R. Ma, W. Van Mol, and F. Adams,Anal. Chim. Acta 285, 33 (1994).



10. R. Ma and F. Adams, Anal. Chim.Acta 317, 215 (1995).

11. R. Ma and F. Adams, Spectrochim.Acta Part B 51, 1917 (1996).

12. R. Ma, W. Van Mol, and F. Adams,At. Spectrosc. 17, 176 (1996).

13. H. Bode, W. Arnswald, Fresenius’ Z.Anal. Chem. 185, 179 (1962).

14. V. L. A. Monte and A. J. Curtius, J.Anal. At. Spectrom. 5, 21 (1990).

15. S. M. Sella, A. K. Ávila, and R. CCampos, Anal. Lett. 32, 2091,(1999).

16. A. K. Avila and A. J. Curtius, J. Anal.At. Spectrom. 9, 543 (1994).

17. X. P. Yan and F. Adams, J. Anal. At.Spectrom. 12, 459 (1997).

18. M. A. M. da Silva, M. B.O.Giacomelli, and A. J. Curtius,Analyst 124, 1249 (1999).

19. D. Pozebon, V. L. Dressler, J. A.Gomes Neto, and A. J. Curtius,Talanta 45, 1167 (1998).

20. D. Pozebon, V. L. Dressler, and A. J.Curtius, J. Anal. At. Spectrom. 13,363 (1998).

21. C. Mena, C. Cabrera, M. L. Lorenzo,and M. C. López, Sci. Total Envi-ron. 181, 201 (1996).



Tantalum Effects on the ICP-OES Determination of Trace Elements in Tantalum Powder

G. Anil, *M.R.P. Reddy, Arbind Kumar, and T.L. Prakash Centre for Materials for Electronics Technology (C-MET), IDA, Phase-II, HCL Post

Cherlapalli, Hyderabad – 500 051, India

INTRODUCTION

C-MET, Hyderabad, is activelyinvolved in the preparation of lowand high voltage capacitor-gradetantalum powder and lead wiresusing solvent extraction, sodiumreduction, and the hydriding anddehydriding process. The presenceof even trace amounts of certainelements in tantalum powder andwires can adversely affect the elec-trical properties of solid tantalumcapacitors. Transitional elementsare known to increase high fieldleakage of anodic tantalum oxidefilms. Calcium and magnesium aresuspected to increase the DC leak-age of solid tantalum capacitors.Most high-charge tantalum powderscontain one or more dopants whichare added to retard sintering. Thedopant usually involves compoundscontaining boron. The presence ofboron up to 100 mg/L is acceptablein overcoming the DC leakage prob-lem; boron amounts above 100mg/L cause tantalum lead wires toharden and become brittle duringthe sintering process (1,2). Molyb-denum is added to tantalum foil tomodify the etch characteristics inthe manufacture of etched foilcapacitors. Addition of molybde-num also has a very pronouncedeffect in decreasing the susceptibil-ity of pure tantalum metal towardshydrogen embitterment in severecorrosion conditions. The presenceof just a few atomic percent of oxy-gen in tantalum increases hardness,tensile strength, and elasticity butdecreases ductility and corrosionresistance. Micro alloying of zirco-nium and tungsten with amaximum of up to 1000 mg/L with

tantalum can minimize the deleteri-ous effects of oxygen added duringprocessing and extend the usefullife of tantalum products. Cu, Mn,Co, and Al are added during therecycling process of solid tantalumcapacitor scrap (3,4). For these rea-sons, ICP-OES analysis of tantalumpowder is routinely carried out andmany matrix interferences arefound during trace element deter-mination.

This analytical work is very vitalin regard to tantalum capacitor performance and applications. In a previous work, we reported thematrix effects of tantalum in thedetermination of the trace impuri-

ties of Cr, Fe, Nb, Ni, and Ti presentin ore as well added during thereduction stage (5). In this study weextended to include 10 more ele-ments (Zr, Mo, Cu, Mn, Mg, Co, W,Ca, B, and Al). These traces areadded to tantalum powder duringthe addition of dopants, duringmicro alloying, and during recyclingof tantalum products. The elementsZr, Mo, W, and B are doped inorder to increase the capacitanceand life of tantalum products. Cu,Mn, Co, and Al are contaminantsadded during recycling of solid tan-talum capacitors, while Ca and Mgare added during chemical process-ing such as deoxidation, etc.

Earlier works by otherresearchers describe the determina-tion of Zr in a tantalum matrix usingtrace matrix separation by anionexchange, extraction chromatogra-phy, and isotope dilution analysiswith a multi-collector coupled withICP-MS (6-8). Knote et al. (9)reported Zr, Mo, and W determina-tion in a tantalum matrix by usingX-ray fluorescence spectrometry,while other studies report thedirect determination of Cu and Mgby ICP-OES in potassium tantalumlithium niobates and by electrother-mal atomic absorbance spectrome-try in tantalum powder (10,11). Al,Mo, and W were determined in thetantalum matrix by anion exchangecoupled with ICP-MS (12).ConteRosa (13) reported the directdetermination of several elementspresent in high concentrations intantalum products using a 1.0-mmonochromator ICP optical emis-sion spectrometer (13). Shuji (14)determined boron using on-lineanion exchange combined withICP-MS. Padma Subashini et al. (15)describe the determination of tung-sten in niobium tantalite present intin slag using derivative spectropho-

ABSTRACT

In this study, the trace ele-ments Zr, Mo, Cu, Mn, Mg, Co,W, Ca, B, and Al are determinedin tantalum powder, using a 0.64-m monochromator inductivelycoupled plasma spectrometer.The effects of tantalum in thedetermination of trace levelimpurities were studied by deter-mining these elements in solu-tions containing tantalum from1000 to 20,000 mg/L. As the con-centration of tantalum isincreased, the Zr, Mo, Co, W,and Al decreases in intensitycounts while Cu and Mn showedan increase in intensity counts.In the case of boron, there wasno change in intensity. For Caand Mg, +1 states showed anincrease in intensity and +2states showed a decrease. Therelative standard intensities oftrace elements in the tantalummatrix were calculated from1000 to 20,000 mg/L tantalumsolutions. Spectral line profiles ofthe trace elements and spectraloverlaps of tantalum at 1000mg/L for interference-free lineselection is presented and dis-cussed.

186

tometry and ICP-OES. In the pre-sent work, the trace elements weredirectly determined using a 0.64-mmonochromator ICP spectrometerwithout trace matrix separation bychoosing a spectral interference-free tantalum line for trace elementdetermination. Further studies(16–25) of tantalum matrix effectson trace elements have been per-formed and appropriate correctionfactors were calculated.

EXPERIMENTAL

Instrumentation

The instrument used for thedetermination of trace elements intantalum powder was a Model JY-24R inductively coupled plasmaoptical emission spectrometer(Jobin Yvon, Longjumeau Cedex,France). A cross-flow type nebulizerwas used for pneumatic nebuliza-tion. The instrumental parametersare given in Table I.


All chemicals used were ofSuprapur® grade (Merck,Darmstadt, Germany). The single-element standard solutions werealso procured from Merck. For allanalytical work, deionized waterwas used, prepared with a Nanop-ure system (Barnstead, Dubuque,IA, USA).

Experimental Procedure

Tantalum powder was dissolvedin a HF + HNO3 mixture and appro-priate concentrations (ranging from1000 to 20,000 mg/L) wereprepared. Trace element solutionsof 1 mg/L were added to the tanta-lum solution. To overcome analyti-cal bias due to possiblecontamination from matrices,which are present at high concen-trations, the matrix solution wasanalyzed as a blank for all trace element studies.


Matrix-free Line Selection

The spectral interferences com-monly found (26,27) in the selec-tion of analyte peaks are coincidentline overlaps, a wing overlap from amuch more intense line nearby.Complex spectra are most trouble-some in the analysis whenproduced by the major constituent,i.e., tantalum. This is due to thespectral lines from tantalum, whichtend to overlap lines of the analytesor trace elements. Line profiling(28,29) was done for each trace ele-ment. Profiles of a 1-mg/L trace ele-ment standard were plotted andoverlapped with 1000 mg/L tanta-lum; the results are listed in TableII. The interference-free lines werechecked for matrix effects of tanta-lum by doping 1 mg/L standard intoconcentrations of tantalum rangingfrom 1000 mg/L to 20,000 mg/L.The detailed results of traceelements added to different tanta-lum concentrations are given inTable III. The best lines for eachelement were selected by takinginto account the matrix effects andthe sensitivity of the line at 20,000mg/L of Ta.

The zirconium line at 339.198nm was selected as it had the high-

est sensitivity and showed only a7% increase in intensity withrespect to 1 mg/L. In case ofmolybdenum, the 201.511-nm linewas selected as it showed a 6%increase in intensity. The 224.700-nm copper line showed a 12%decrease. The 280.106-nmmanganese line was selected andshowed a 3% decrease, while the344.199-nm line could not beresolved at this concentration oftantalum. The cobalt line at228.616 nm was selected as it hadthe highest sensitivity and showeda 4% increase in intensity. The alu-minum line at 394.401 nm +1 state(29) showed a 17% increase inintensity, tungsten at the 224.875-nm line showed a 4% decrease.Boron at the 208.959-nm lineshowed no change in intensity. Allof the above elements were of +2state (29). However, calcium at theline of 422.673-nm +1 state (29)showed a 4% increase in intensity,while the line at 396.84-nm +2state (29) showed a 6% decrease inintensity. For calcium, the 396.847-nm line was selected as it had veryhigh sensitivity. The magnesiumline at 383.826 nm +1 state (29)showed a 46% increase and thelines at 280.270 nm and 279.553nm +2 state (29) showed an 8%decrease. The line at 280.270 nm

TABLE IICP-OES Operating Conditions

Sequential spectrometer Jobin Yvon JY-24R, 0.64-m Czerny TurnerGrating Holographic, 3600 grooves mm–1

Spectral range 160–500 nmSlit width 20 nm for entrance and exits, adjustableHDD Detector Dynamic range 2.5 x 1010Image Measures complete emission spectrum

in 2 min (230,000 points)

Plasma torch assembly Fused quartz with capillary injectionNebulizer and spray chamber Cross-flow type nebulizer and Teflon®

dual-tube spray chamberArgon coolant gas flow rate 1.5 L min–1

Argon plasma gas flow rate 12 L min–1

187

Vol. 24(5), Sept./Oct. 2003

TABLE IISpectral Interference of Tantalum on Trace Element Wavelengths (nm)

(1 mg/L of trace element line overlap with 1000 mg/L of tantalum)

Wavelengths for: (Zr ) (Mo) (Cu) (Mn) (Mg) (Co) (W) (Ca) (B) (Al)

256.764a 201.511b 213.598a 257.610c 202.582a 228.616d 200.807c 315.887e 208.893c 221.006c

256.887d 202.030c 217.894c 259.373c 279.482c 230.786a 202.998c 317.993b 208.959b 226.346c

257.139c 203.844c 219.958c 260.569c 277.983c 231.160c 203.503c 393.366b 249.678e 226.910c

257.863c 204.598c 221.458c 279.482c 279.553b 231.405c 205.468c 396.847b 249.773c 226.922c

270.013d 263.876c 221.810c 279.827c 280.270b 231.498e 207.911a 422.673b – 236.705c

272.261b 268.414c 222.778c 280.106b 285.213b 234.426e 208.819c -- – 237.312e

273.486b 277.540c 223.008c 293.306c 293.654d 234.739c 209.475c -- – 237.335e

274.256b 281.615b 224.700d 293.930c 383.231c 235.342d 209.860c -- – 257.510c

316.597b 284.823c 324.754c 294.920c 383.826b 236.379a 216.632c -- – 308.215c

327.305c 287.151c 327.396c 344.199b -- 237.862e 218.935c -- – 309.271e

327.926c 292.339c -- 403.076c -- 238.346c 220.448c -- – 309.284e

330.628e -- -- 403.307c -- 238.636c 222.589c -- – 394.401b

339.198b -- -- -- 238.892c 224.875e -- – 396.152c

343.823d -- -- -- 327.305c 232.609c -- -- --

348.115b -- -- -- -- 239.709c -- -- --

349.621c -- -- -- -- 248.923c -- – --

350.567c -- -- -- -- -- -- – --

355.660b -- -- -- -- -- -- – --

357.247c – – – – – – – – –

a = Right and left wing overlap of tantalum line.b = No tantalum line overlap present in the vicinity of ±0.05 nm.c = Direct tantalum line overlap.d = Right wing overlap of tantalum line.e = Left wing overlap of tantalum line.

was selected for magnesium as ithad high sensitivity. The selectedlines of trace elements were usedfor the analysis by calculating thematrix factor as described in a pre-vious work (5).

Comparison of Results With a Certified Powder andSpike Recovery

The method developed was veri-fied by analyzing 2 g of certifiedtantalum powder STA-18KT (H.C. Starck GmbH & Co., K.G.,Germany) dissolved by using 3 mLHF, 1 mL HNO3, and 5 mL of water;the solution was made up to 100-mL volume (i.e., 20,000 mg/L oftantalum). The results are presented

in Table IV. Similarly, a 2-g/100-mLtantalum powder was spiked withpure standard solutions of 0.25mg/L and 0.5 mg/L trace elementsolution; the % recovery was calcu-lated with and without a matrix fac-tor. The results are presented inTable V. From these experiments itcan be seen that the experimentalvalues, when multiplied using thematrix factor, showed very closeproximity to the certified values aswell as to the recovery results ofthe spiked analytes.

CONCLUSION

The ICP optical emission spec-trometer, equipped with a 0.64-mmonochromator, was successfully

applied for the direct determinationof the trace elements Zr, Mo, Cu,Mn, Mg, Co, W, Ca, B, and Al in tan-talum powder without the need fortrace matrix separation by selectingtantalum spectral interference-freelines and calculating the matrixeffects factor of tantalum at differ-ent concentrations of tantalumranging from 1000 mg/L to 20,000mg/L tantalum. This factor wasapplied to the analysis of certifiedmaterials and was found to be verysuitable. This method and thematrix factors can be applied forthe ICP-OES determination of traceelements in tantalum powder. Theinterference-free lines selected canbe used for the determination ofthese traces in other tantalum-

188

TABLE III. Trace Elements (1 mg/L) Added to Different Ta Concentrations

S.No. Element Wavelength 1,000 mg/L (S.D) 20,000 mg/L (S.D) 5,000 mg/L (S.D) 10,000 mg/L (S.D) 20,000 mg/L (S.D)

1. Zr 272.261 0.99± 0.0029 0.98± 0.0088 0.97± 0.0087 0.95± 0.0065 0.94± 0.0079

2. Zr 273.486 1.0± 0.0018 1.0± 0.0092 1.0± 0.0082 1.0± 0.0096 1.0± 0.0093

3. Zr 274.256 1.0± 0.0065 1.0± 0.0098 1.0± 0.0094 1.01± 0.0075 1.01± 0.0088

4. Zr 316.597 1.0± 0.0018 1.0± 0.0076 1.0± 0.0078 1.02± 0.0087 1.04± 0.0095

5. Zr 339.198 1.0± 0.0035 1.0± 0.0077 1.02± 0.0095 1.04± 0.0075 1.07± 0.0096

6. Zr 348.115 1.0± 0.0055 1.0± 0.0095 1.0± 0.0085 1.0± 0.0076 1.0± 0.0085

7. Zr 355.660 1.0± 0.0071 1.0± 0.0099 CR CR CR

8. Mo 201.511 1.0± 0.0073 1.0± 0.0055 1.0± 0.0086 1.04± 0.0072 1.06± 0.0087

9. Mo 281.615 1.07± 0.061 1.12± 0.0063 1.18± 0.078 1.22± 0.0068 1.31± 0.0086

10. Cu 224.700 1.0± 0.0088 0.98± 0.0064 0.92± 0.0065 0.90± 0.0061 0.88± 0.0084

11. Mn 280.106 1.0± 0.0065 1.0± 0.0065 1.0± 0.0084 0.99± 0.0054 0.97± 0.0081

12. Mn 344.199 1.0± 0.0078 0.99± 0.0075 0.98± 0.0077 CR CR

13. Mg 279.553 1.0± 0.0072 0.98± 0.010 0.97± 0.0068 0.95± 0.0059 0.92± 0.0078

14. Mg 280.270 1.0± 0.0082 1.0± 0.091 0.98± 0.0065 0.94± 0.0061 0.92± 0.0066

15. Mg 285.213 1.0± 0.0091 1.0± 0.0084 0.97± 0.0060 CR CR

16. Mg 383.826 1.46± 0.0092 1.46± 0.0075 1.46± 0.0058 1.46± 0.0093 1.46± 0.0075

17. Co 228.616 1.0± 0.0066 1.0± 0.0071 1.01± 0.0075 1.02± 0.0066 1.04± 0.0074

18. Co 231.498 1.0± 0.0081 1.01± 0.0066 CR CR CR

19. Co 234.426 1.0± 0.0072 1.0± 0.0078 1.0± 0.0079 1.0± 0.0075 1.0± 0.0068

20. Co 235.342 1.0± 0.0093 1.0± 0.0067 1.02± 0.0087 1.04± 0.0088 1.06± 0.0072

21. Co 237.862 1.0± 0.0082 1.02± 0.0089 1.06± 0.0053 CR CR

22. W 224.875 1.0± 0.0078 0.99± 0.0075 0.98± 0.0077 0.97± 0.0071 0.96± 0.0065

23. Ca 317.993 1.0± 0.0062 0.98± 0.0044 0.97± 0.0094 CR CR

24. Ca 393.366 1.0± 0.0064 0.98± 0.0081 0.96± 0.0085 CR CR

25. Ca 396.847 1.0± 0.0076 0.98± 0.0095 0.98± 0.0083 0.96± 0.0093 0.94± 0.0084

26. Ca 422.673 1.0 ± 0.0078 1.0 ± 0.0096 1.01 ±0.0081 1.01 ± .0091 1.04 ±0.0071

27. B 208.959 1.0± 0.0082 1.0± 0.0095 1.0± 0.0094 1.0± 0.0098 1.0± 0.0069

28. Al 394.401 1.0± 0.0077 1.03± 0.0095 1.05± 0.0083 1.09± 0.0076 1.17± 0.0074

CR: Cannot be resolved. S.D: Standard deviation of five readings.

TABLE IV. Analysis of Certified Powder (Results in mg/L)

S.No Elements Wavelength Certified Values Experimental Results x Matrix factor(STA-18KT) (STA-18KT)

1. Zr 339.198 <2 2.05 x 0.93 = 1.915

2. Mo 201.511 <2 1.90 x 0.94 = 1.786

3. Cu 224.700 <1 0.68 x 1.13 = 0.772

4. Mn 280.106 <2 1.62 x 1.03 = 1.67

5. Mg 280.270 9 8.36 x 1.087 = 9.08

6. Co 228.616 N.R 0.35 x 0.96 = 0.336

7. W 224.875 <5 5.10 x 0.96 = 4.90

8. Ca 396.847 <3 2.52 x 1.06 = 2.68

9. B 208.959 <1 0.80 x 1 = 0.80

10. Al 394.401 1 1.19 x 0.85 = 1.01

N.R: Values not reported.

189

Vol. 24(5), Sept./Oct. 2003

related compounds using ICP-OES.This method of analysis is very fastand accurate and can be applied forroutine analysis.

Received September 20, 2002.

ACKNOWLEDGMENTS

The authors would like to thankthe Department of InformationTechnology (DIT), Government ofIndia, New Delhi, for sponsoringthe Tantalum Technology MissionProject. The authors are also thank-ful to Dr. B.K. Das, Executive Direc-tor, C-MET, Dr. K.S.K. Sai, SeniorDirector, Dr. V.C. Sethi, Director,DIT, and Dr. K. Chandrasekar, Sci-entist, IICT, for their helpful discus-sions.

REFERENCES

1. B. Tripp Terrance, Proceedings ofInternational Symposium of Tanta-lum and Niobium, 255 (1988).

2. J.H. Chen, R.M. Bergman, and L.E.Huber, Proceedings of InternationalSymposium of Tantalum andNiobium, 271 (1988).

3. Charles Pokross, Proceedings ofInternational Symposium of Tanta-lum and Niobium, 297 (1988).

4. Meinhard Aits, Proceedings of Inter-national Symposium of Tantalum andNiobium, 333 (1988).

5. G. Anil, M.R.P. Reddy, Arbind Kumarand T.L. Prakash, At. Spectrosc. 23,119 (2002).

6. M. Opilz, G.Wuensch, J. Prakt. Chem.339, 44 (1997).

7. X.J. Yang and C. Pinc, Anal. Chim.Acta 458, 375 (2002).

8. Stefan Weyer, Carsten Munker, MarkRehkamper, and Klaus Metzger,Chemical Geology. 187, 295 (2002).

9. Harald Knote and Viliam Krivan,Anal. Chem. 54, 1858 (1982).

10. I. Segal, E. Dorfman, O.Yoffeo, I.Bezrukavnikov, and A.J. Agranat, At.Spectrosc. 21, 46 (2000).

11. Klaus-Christian Friese and ViliamKrivan, Spectrochim. Acta, Part B,53, 1069 (1998).

12. S. Kozono, R. Takashi, and H.Haraguchi, Anal. Sciences 16, 69(2000).

13. A. ConteRosa, Jean-Michel Mermet,Jose De Anchieta Rodrigues, andJean-Louis Dimartino, J. Anal. At.Spectrom. 12, 1215 (1997).

14. Kozono Shuji, Takahashi Shigeto,and Haraguchi Hiroki, Analyst 127,930 (2002).

15. V. Padma Subashini, M.K. Ganguly,K. Satyanarayana, and R.K. Malhotra,Talanta 50, 669 (1999).

16. Cendrine Dubuisson, EmmanuellePoussel, and Jean Michel Mermet, J.Anal. At. Spectrom. 12, 281 (1997).

17. Jean-Michel Mermet, J. Anal. At.Spectrom. 13, 419 (1998).

18. M. Moltelica-Heino, O.F.X.Donard,and J.M. Mermet, J. Anal. At. Spec-trom. 14, 675 (1999).

19. E. Edivaldo Garcia, A. Ana Rita, A.Nougueira, and A. Joaquim Nobrega,J. Anal. At. Spectrom. 16, 825 (2001).

20. Marco Grotti, Emanuele Magi, andRoberto Frache, J. Anal. At.Spectrom. 15, 89 (2000).

21. Lilli Paama, Eed Parnoja, Mare Must,and Paavo Peramaki, J. Anal. At.Spectrom. 16, 1333 (2001).

22. Anna Krejcova, TomasCernohorsky, and Eva Curdova, J.Anal. At. Spectrom. 16, 1002 (2001).

23. Margarita Villaneuva, Miguel Cata-sus, E. Eric Salin, and Mario Pomares,J. Anal. At. Spectrom. 15, 877 (2000).

24. M. Jesus Cano, L. Jose Todli, VicenteHernandis, and Juan Mora, J. Anal.At. Spectrom. 17, 57 (2002).

25. Bojan Budic, Acta Chim. Slov. 47,153 (2000).

26. L. Aleksieva, N. Daskalova, and S.Velichkov, Spectrochim. Acta, PartB, 57, 1339 (2002).

27. N. Daskalova, G. Gentsheva, andS.Velichkov, Spectrochim. Acta, PartB, 57, 1351 (2002).

28. Akbar Montaser and D.W. Golightly,Inductively Coupled Plasmas in Ana-lytical Spectrometry, Second Edition,VCH Publishers, Inc. (1992).

29. Richard Payling and Peter Larkins,Optical Emission Lines of the Ele-ments, John Wiley and Sons Ltd.(2000).

TABLE V. Recovery of 0.25 mg/L and 0.50 mg/L Trace Elements Spiked in 2 g/100 mL Tantalum Powder

0.25 mg/L Spiked trace elements 0.5 mg/L Spiked trace elementsRecovery % Recovery % Recovery % Recovery %without matrix with matrix without matrtix with matrix

No. Element Wavelength Factor calculation (SD) Factor calculation (SD)

1. Zirconium 339.198 107.21± 0.15 99.96±0.05 107.28± 0.17 100.02±0.04

2. Molybdenum 201.511 106.34±0.19 99.88±0.08 106.18±0.17 99.97±0.06

3. Copper 224.700 88.15±0.13 99.97±0.06 88.12±0.10 100.07±0.06

4. Manganese 280.106 97.25±0.18 100.06±0.07 97.19±0.15 100.03±0.11

5. Magnesium 280.270 92.22±0.14 100.08±0.06 92.29±0.15 99.89±0.07

6. Cobalt 228.616 104.33± 0.19 99.97±0.04 104.38± 0.15 100.02±0.02

7. Tungsten 224.875 96.17±0.09 100.19±0.12 96.26±0.18 100.12±0.07

8. Calcium 396.847 94.28±0.16 100.34±0.26 94.20±0.13 99.97±0.08

9. Boron 208.959 100.05±0.07 99.98±0.03 100.02±0.04 99.82±0.06

10. Aluminum 394.401 117.33±0.25 100.18±0.07 117.29±0.19 99.95±0.08

SD = Standard Deviation of five readings.


Mixed Matrix Effect of Easily and Non-easily IonizableElements in the ICP-OES Analysis of Trace Impurities

in Potassium Tantalum Fluoride G. Anil, *M.R.P. Reddy, Arbind Kumar, and T.L. Prakash

Center for Materials for Electronics Technology (C-MET), IDA, Phase-II, HCL Post Cherlapalli, Hyderabad – 500 051, India

INTRODUCTION

Despite the intensive work forthe development of inductivelycoupled plasma optical emissionspectrometry (ICP-OES) techniques,ICP remains an important field ofstudy. The study of the processesoccurring within the plasma (suchas desolvation, vaporization,atom/ion formation and excitation)is important, particularly from thepoint of their control, so that anyshortcomings existing in atomicemission spectrometric analysis canbe lowered or eliminated, and exist-ing methods for atomic spectromet-ric analysis optimized. However,the fundamental processes thatcontrol analyte emission intensityand spatial dependence are notfully understood. Studies on thematrix interference effects havebeen conducted by several researchgroups (1–9), but the matrix ele-ments studied have been mostlyrestricted to alkali metals (EIE).Presently our laboratory is workingon the different process parametersfor the preparation of K2TaF7 byacid washings and the recrystalliza-tion process. The purity of K2TaF7

is extremely critical for the prepara-tion of tantalum powder.

C-MET in Hyderabad is activelyinvolved in the preparation of lowand high voltage capacitor-gradetantalum powder through solventextraction, sodium reduction,hydriding and dehydriding routes.Analysis of K2TaF7 is routinely car-ried out using a 0.64-m monochro-mator ICP-OES, where matrixinterferences are noted during traceelement determinations. There is a

need to establish a method todetermine the effects of NEIE, EIE,and the mixed effects (10) of NEIEand EIE, while at the same timeestimating their trace element con-tent. Irina Segal et al. (11) deter-mined traces of Cu, Fe, Li, Ti, andV in potassium lithium tantalateniobate crystals using ICP-OES, andalso discussed the spectral interfer-ence of Ta and Nb, but not theirmatrix effects. Almost any matrixelement can cause a change in thesensitivity of a susceptible spectralline. The easily ionizable elementpotassium causes larger effectsthan tantalum, which has a higherionization potential, at least at

observation heights of 14 to 18 mmabove the load coil. The effects arerather complicated and far fromfully characterized. Nevertheless, agrasp of the broad aspects is mean-ingful because it bears on the con-duct of practical analysis. The aimof this work was to investigate theeffects of potassium, tantalum, andthe mixed effects of both elementson the analytical intensity of traceelements.

EXPERIMENTAL

Instrumentation

The instrument used for traceelement analysis in K2TaF7 was aModel JY-24R ICP-OES (JobinYvon, Longjumeau Cedex, France ).A cross-flow nebulizer was used fornebulization. The instrument para-meters are given in Table I.

Reagents and Samples

The acids used were of Supra-pur“ reagent grade (Merck, Darm-stadt, Germany). The single-elementstandard solutions were obtainedfrom Glen Spectra Reference Mate-rials, Middlesex, England. Thedeionized water used was preparedusing the Nanopure system (Barn-stead, Dubuque, IA, USA).

Experimental Procedure

2 g of K2TaF7 was dissolved in 10mL HF. 1 g of K2TaF7 contains0.1994 g of potassium and 0.4614 gof tantalum by weight. Toovercome analytical bias due topossible contamination from thematrices (acids) present at highconcentration levels, the matrixsolution was analyzed as a blank forall trace element studies.


ABSTRACT

The determination of traceimpurities (B, Cr, Fe, Nb, Ni, Ti,Cu, Co, Ca, and Mg) in potassiumtantalum fluoride (K2TaF7) byinductively coupled plasma opti-cal emission spectrometry is com-plicated due to the matrixinterference of tantalum on traceimpurities, a non-easily ionizableelement (NEIE), and due to potas-sium, an easily ionizable element(EIE). In this paper, the effect oftrace elements by NEIE, EIE, andthe mixed effect of NEIE and EIE,was studied with doped traceelements at concentrations of0.25, 0.5, 0.75, and 1 mg/L. Itwas found that NEIE showed anincrease and decrease in intensityin the range of 10% only, whereasEIE showed a decrease up to40%; the mixed matrix effectshowed an initial increase for alltrace elements, except for Cu. Asthe concentration of the mixedmatrix increased, the intensity ofall traces started decreasing,except for Ca and Mg, whichshowed an increase.

191

Vol. 24(5), Sept./Oct. 2003


The matrix effect, M, is definedas the percentage difference in thenet line signals between solutionsof the analyte with and without thematrix. Potassium, an EIE, oftenconstitutes the bulk of the solutemass; it will markedly affect the rateor even the mechanism of soluteparticle vaporization. In turn, themore rapid vaporization of potas-sium salt, or the consequentbreakup of solute particles, wouldyield an overall vaporizationprocess that begins lower in theplasma. As the resulting analyte-containing vapor cloud travelsupward through the plasma, its vol-ume increases by diffusion. Thetemporally integrated emission willtherefore appear to have a "V"shape. As the vapor cloud appearslarger and more dilute, the "V" willmove downward. Hobbs and Olesik(12–14) stated that the extent ofEIE matrix effects on ionization andion excitation was more severenear an incompletely desolvateddroplet or vaporizing particle thanaway from a droplet or particle.This is consistent with lower tem-peratures and electron number den-sities as well as higher analyte andmatrix species concentration nearan incompletely desolvated drop ora vaporizing particle. Depressionsin analyte ion number densities dueto the presence of EIE were, by a

factor of two to three, more severenear an incompletely desolvateddrop or a vaporizing particle thanwhen no droplet or particle was inthe observation volume. Trace ele-ments of 1 mg/L were doped intopotassium solutions containingequivalent weight with respect to0.5 g, 1.0 g, 1.5 g, and 2.0 g/100 mLof K2TaF7. The results showed adecrease in intensity for all traceelements; as the concentration ofEIE increased, the intensity of thetraces further decreased. Theresults are given in Table II.

The trace elements B, Cr, andNb, when doped at a concentrationof 1 mg/L into tantalum solutions[tantalum interference-free lineswere selected (15-16)] with respectto K2TaF7 weight, showed nochange in intensity even with anincrease of tantalum concentration.Iron and Co showed a decrease inintensity while Ni, Ti, Cu, Ca, andMg showed an increase in intensity.The results are given in Table III.The addition of 1 mg/L trace ele-ments into mixed solutions ofpotassium and tantalum showed anincrease in intensity for allelements, except Cu, whichshowed a decrease of 14%. As theconcentration of a mixed matrixincreased, the intensity of tracesstarted decreasing, except for Caand Mg, which showed an increase(see Table IV). Using these results, a

matrix factor (15) was calculatedand applied to the determination oftrace elements doped at concentra-tions of 0.25, 0.50, 0.75, and 1.0mg/L into K2TaF7 2 g/100 mL. Theresults obtained (see Table V)showed that the matrix factor cal-culated was very suitable and canbe applied for trace determinationsin K2TaF7.

CONCLUSION

This study demonstrated thattrace element determinations inK2TaF7 by ICP-OES, using interfer-ence-free lines, is not the completeanalysis but the matrix effects ofEIE, NEIE, and their mixed effectsshould be considered thoroughly inorder to get accurate and repeat-able results. Doping traces at differ-ent concentrations in K2TaF7

showed that, by applying thematrix correction factor, the errorcaused by the matrix was rectified.This type of correction should beconsidered when a properreference material for such an inter-mediate compound is not available.

Received September 20, 2002.

ACKNOWLEDGMENTS

The authors would like to thankthe Department of InformationTechnology (DIT), Govt. of India,New Delhi, for sponsoring the Tan-talum Technology Mission Project.The authors are also thankful to Dr.B.K. Das, Executive Director, C-MET and Dr. K.S.K. Sai, SeniorDirector, Dr. V.C. Sethi, Director,DIT.

TABLE I ICP-OES Operating Conditions

Sequential spectrometer Model JY-24R, 0.64-m Czerny-TurnerGrating Holographic, 3600 grooves mm–1

Spectral range 160–500 nmSlit width 20 nm for entrance and exits, adjustableHDD detector Dynamic range 2.5 x 1010

Plasma torch assembly Fused quartz with capillary injectionNebulizer and spray chamber Cross-flow nebulizer and Teflon®

dual-tube spray chamberArgon coolant gas flow rate 1.5 L min–1

Argon plasma gas flow rate 12 L min–1

192

TABLE IIRecovery (%) of 1 mg/L of Trace Elements Doped Into Potassium Solution

(Equivalent to 0.5 g, 1 g, 1.5 g, and 2.0 g K2TaF7)

Elements Wavelength (nm) 0.5 g (SD) 1.0 g (SD) 1.5 g (SD) 2.0 g (SD)

B 208.959 94.64±0.29 85.24±0.42 84.36±0.53 82.31±0.20 Cr 267.716 80.61±0.33 77.45±0.36 76.56±0.64 74.27±0.53 Fe 274.932 82.49±0.45 79.36±0.54 74.47±0.73 72.39±0.75 Nb 295.088 91.53±0.54 87.68±0.58 82.48±0.68 79.42±0.79 Ni 227.021 98.59±0.68 94.39±0.61 92.39±0.47 91.54±0.85Ti 336.121 93.63±0.76 82.93±0.69 77.32±0.52 74.58±0.73Cu 224.700 89.74±0.65 81.31±0.73 78.41±0.35 62.73±0.68 Co 228.616 95.61±0.85 91.26±0.63 85.49±0.90 81.84±0.81 Ca 396.847 90.61±0.79 74.37±0.56 66.37±0.71 58.81±0.57

Mg 280.270 92.58±0.82 79.35±0.37 68.29±0.82 60.87±0.65

± SD of five readings.

TABLE IIIRecovery (%) of 1 mg/L of Trace Elements Doped Into Tantalum Solution



B 208.959 100.15±0.19 99.93±0.20 99.97±0.23 99.58±0.19 Cr 267.716 100.23±0.16 99.95±0.18 99.90±0.19 99.85±0.17 Fe 274.932 93.255±0.29 92.65±0.27 91.89±0.31 89.69±0.35 Nb 295.088 100.05±0.33 100.08±0.25 100.05±0.29 99.97±0.21Ni 227.021 108.89±0.28 109.54±0.21 111.58±0.19 112.45±0.32 Ti 336.121 103.50±0.31 103.79±0.27 104.86±0.32 105.25±0.25 Cu 224.700 112.06±0.42 112.36±0.48 113.65±0.52 113.97±0.41 Co 228.616 96.10±0.27 96.45±0.24 96.98±0.23 97.21±0.27 Ca 396.847 105.48±0.42 106.21±0.4 106.87±0.50 107.24±0.44

Mg 280.270 101.15±0.39 101.45±0.37 101.78±0.34 101.89±0.43


TABLE IVRecovery (%) of 1 mg/L of Trace Element Doped Into Mixture of Potassium and Tantalum Solution



B 208.959 103.45±0.20 99.51±0.37 94.34±0.77 91.28±0.37 Cr 267.716 112.48±0.62 108.23±0.29 103.23±0.43 101.35±0.48 Fe 274.932 121.13±1.15 114.65±0.45 110.49±0.58 108.25±0.39 Nb 295.088 107.28±0.38 105.68±0.63 102.66±0.36 99.35±0.69 Ni 227.021 106.38±0.44 97.48±0.56 92.58±0.49 88.19±0.25 Ti 336.121 105.22±0.18 103.46±0.58 99.58±0.48 96.23±0.20 Cu 224.700 88.71±0.21 84.71±0.66 80.38±0.40 77.16±0.26Co 228.616 113.36±0.39 104.23±0.19 101.41±0.56 98.61±0.82Ca 396.847 106.35±0.24 108.50±0.45 123.49±0.3 127.42±0.44

Mg 280.270 104.27±0.33 101.18±0.15 108.64±0.58 113.46±0.58


193

Vol. 24(5), Sept./Oct. 2003

TABLE VDifferent Concentration of Traces Doped Into K2TaF7 (2 g/100 mL)

Wave- 0.25 mg/L 0.50 mg/L 0.75 mg/L 1.0 mg/Llength Recovery % (SD) Recovery % (SD) Recovery % (SD) Recovery % (SD)

Without Applying Without Applying Without Applying Without ApplyingElements Factor Corr. Factor Factor Corr. Factor Factor Corr. Factor Factor Corr. Factor

B 208.959 91.49±0.45 100.23 91.45±0.39 100.18 91.38±0.41 100.10 91.31±0.30 99.52

Cr 267.716 101.89±0.51 100.53 101.78±0.48 100.42 101.65±0.45 100.29 101.31±0.43 99.28

Fe 274.932 108.65±0.44 100.36 108.55±0.45 100.27 108.50±0.37 100.23 108.32±0.23 99.65

Nb 295.088 99.15±0.34 99.80 99.14±0.35 99.79 99.18±0.31 99.18 99.35±0.60 99.35

Ni 227.021 88.46±0.32 100.30 88.48±0.34 100.32 88.32±0.31 99.80 88.29±0.23 99.76

Ti 336.121 96.65±0.39 100.43 96.61±0.41 100.40 96.43±0.32 100.28 96.35±0.26 100.20

Cu 224.700 77.33±0.25 100.22 77.46±0.31 100.38 77.33±0.36 99.75 77.25±0.24 99.65

Co 228.616 98.89±0.73 100.28 98.92±0.88 100.31 98.87±0.76 99.85 98.77±0.78 99.75

Ca 396.847 127.88±0.78 100.36 128.05±0.88 100.49 127.57±0.75 99.50 127.49±0.39 99.95

Mg 280.270 113.79±0.74 100.29 113.70±0.75 100.21 113.62±0.65 99.98 113.54±0.43 99.91

± SD of five readings.Corr. Factor = Correction Factor.

REFERENCES

1. Cendrine Dubuisson, EmmanuellePoussel, and Jean Michel Mermet,J. Anal. At. Spectrom. 12, 281(1997).

2. Jean-Michel Mermet, J. Anal. At. Spec-trom. 13, 419 (1998).

3. M. Moltelica-Heino, O.F.X. Donard,and J.M. Mermet, J. Anal. At. Spec-trom. 14, 675 (1999).

4. E. Edivaldo Garcia, A. Ana Rita,A.Nougueira, and A. JoaquimNobrega, J. Anal. At. Spectrom. 16,825 (2001).

5. Marco Grotti, Emanuele Magi, andRoberto Frache, J. Anal. At. Spec-trom. 15, 89 (2000).

6. Lilli Paama, Eed Parnoja, Mare Must,and Paavo Peramaki, J. Anal. At.Spectrom. 16, 1333 (2001).

7. Anna Krejcova, Tomas Cernohorsky,and Eva Curdova, J. Anal. At. Spec-trom. 16, 1002 (2001).

8. Margarita Villaneuva, Miguel Catasus,E. Eric Salin, and Mario Pomares, J.Anal. At. Spectrom. 15, 877 (2000).

9. M. Jesus Cano, L. Jose Todli, VicenteHernandis, and Juan Mora, J. Anal.At. Spectrom. 17, 57 (2002).

10. Bojan Budic, Acta Chim. Slov. 47,153 (2000).

11. Irina Segal, Ethel Dorfman, OlgaYoffe, Irina Bezrukavnikov, andA.J. Agranat, At. Spectrosc. 21, 46(2000).

12. P.J .Galley, M.Glick, andG.M.Hieftje, Spectrochim. Acta,48B, 769 (1993).

13 S.E. Hobbs and J.W.Olesik,Spectrochim. Acta, 52B, 353(1997).

14 X. Romeo, E. Poussel, and J.M. Mer-met, Spectrochim. Acta, 52B, 495(1997).

15. G. Anil, M.R.P. Reddy, ArbindKumar, and T.L. Prakash, At. Spec-trosc. 23, 119 (2002).

16. G. Anil, M.R.P. Reddy, ArbindKumar, and T.L. Prakash, At. Spec-trosc. to be published in 24(5)(2003).

194

2004 Winter Conference on Plasma SpectrochemistryFort Lauderdale, Florida, USA, January 5-10, 2004

The 13th biennial international Winter Conference will be held at the Wyndham Bonaventure Resort and Spa(www.wyndham.com/bonaventure) in Fort Lauderdale, Florida (www.sunny.org).

More than 600 scientists are expected and approximately 300 papers on modern plasma spectromchemistry will bepresented.

Symposium Features

• Elemental speciation, speciation sampling, and sample preparation

• Excitation mechanisms and plasma phenomena

• Flow injection and flow processing spectrochemical analysis

• Glow discharge atomic and mass spectrometry

• Inductively coupled plasma atomic and mass spectrometry

• Laser ablation and breakdown spectrometry

• Microwave atomic and mass spectrometry

• Plasma chromatographic detectors

• Plasma instrumentation, microplasmas, automation, and software innovations

• Sample introduction and transport phenomena

• Sample preparation, treatment, and automation; high-purity materials

• Spectrochemical chemometrics, expert systems, and software

• Spectroscopic standards and reference materials, databases

• Stable isotope analyses and applications

January 2-4, Fri-Sun 62 Continuing Education Short Courses

January 2-4, Fri-Sun Manufacturer’s Seminars

January 4, Sunday 5th Annual Golf Tournament

Daily Six Provocative Panel Discussions

January 6-8, Tue-Thu Workshop on New Plasma Instrumentation

January 4, Sunday USP Roundtable Discussion

Information

For program, registration, hotel, and transportation details, visit our website at:

http:>// www.unix.oit.umass.edu/~wc2004/WinterConf2004

Or contact:

Ramon Barnes, ICP Information Newsletter, Inc., P.O. Box 666, Hadley, MA 01003-0666, USA

Tel: (+1) 413-256-8942, Fax: (+1) 413-256-3746, e-mail: [email protected]

U.S. 800-762-4000 (+1) 203-925-4600

www.perkinelmer.com

There, that’s all the training you need.Walk up to the AAnalyst 200 and let the touch screen guide you through everythingfrom setup to analysis. It practically tells you what to do—and in your ownlanguage. All instrument controls are right there on the screen, available at yourfingertips. Even troubleshooting and repairs are easier, with quick-change partsyou simply snap out and snap in. No service visit, no down time. As rugged andreliable as ever, our newest AAnalyst is a better way to do AA. Experience it foryourself. Talk to a PerkinElmer inorganic analysis specialist today.

© 2003 PerkinElmer, Inc.

Atomic Absorption

Justtouch

and go.

AutoLens™

Optimizes voltagefor each element

PlasmaLok®

Easy optimization and extended cone life

Dynamic Bandpass TuningEfficiently screens out interferenceswhile maximizing analytetransmission

Axial Field™TechnologyOptimizes performance andspeed in all matrices

Platinum Quick-ChangeInterface ConesEasy maintenance,maximum uptime

SimulScan™

Simultaneous dual-stagedetector with 9 orders of dynamic range

High-SpeedQuadrupoleFast transient signal analysis

AutoRes™

Custom resolution minimizesspectral interferences andimproves detection limits

All-Quartz SampleIntroductionMinimizes contamination

When your applications extend beyond the capabilitiesof conventional ICP-MS, you need the power of theinnovative ELAN® DRC II. The DRC II combines the power of patented Dynamic Reaction Cell (DRC)technology with performance-enhancing Axial FieldTechnology, providing uncompromised sensitivity andperformance in all matrices for even the toughestapplications. Unlike collision cell, high-resolution, orcold plasma systems, the DRC II completely eliminatespolyatomic interferences providing ultratrace-leveldetection limits.

The DRC II uses chemical resolution to eliminateplasma-based polyatomic species before they reach thequadrupole mass spectrometer. This ion-moleculechemistry uses a gas to “chemically scrub” polyatomicor isobaric species from the ion beam before they enterthe analyzer, resulting in improved detection limits forelements such as Fe, Ca, K, Mg, As, Se, Cr, and V.

Unlike more simplistic collision cells, patented DRCtechnology not only reduces the primary interference;it eliminates sequential side reactions that create newinterferences. Unless kept in check by DRC technology,these uncontrolled reactions increase spectral complexity and create unexpected interferences.

Dynamic Reaction Cell™ (DRC™) with Axial Field Technologyprovides superior interference reduction and performance for all applications.

Eliminates interferences COMPLETELY

006872_03

Atomic Spectroscopy 24(5)

Documents

Transcript of Atomic Spectroscopy 24(5)