tomics pectroscopy - a and b spectroscopy 26(3).pdf · pectroscopy vol. 26(3) may/june 2005...

Issues also available

electronically.

(see inside front cover)

ASPND7 26(3) 89–130 (2005)ISSN 0195-5373

AtomicSpectroscopy

May/June 2005 Volume 26, No. 3

In This Issue:Speciation of Chromium With Cloud Point Extraction Separation and Determination by ICP-OESPei Liang and Jing Li........................................................................................... 89

Direct Generation of Stibine From Slurries and Its Determination by ETAAS Using Multivariate OptimizationM.J. Cal-Prieto, M. Felipe-Sotelo, A. Carlosena, and J.M. Andrade .................. 94

Application of Biological Substrates to the Speciation Analysis of Cr(III) and Cr(VI)Belén Parodi, Griselda Polla, Liliana Valiente, and Patricia Smichowski..... 102

Flow Injection Determination of Cd in Meat Samples Using a ContinuousLixiviation/Preconcentration System Coupled to a Flame AASS. Cancela-Pérez and M.C. Yebra-Biurrun....................................................... 110

Medical Application of Fast Furnace Program Used in the ETA-AAS Determination of Cu and Zn in Blood Plasma of Children With Down SyndromeDenny R. Fernández, Aracelis del C. Vásquez, Maigualida Hernández, Ana M. Ocando, José G. Manzanilla, Marisol Soto, Francisco Alvarez, and Víctor A. Granadillo .................................................................................. 117

Determination of Tellurium by ETAAS With Preconcentration by Coprecipitation With Lanthanum HydroxideS. Cerutti, M. Kaplan, S. Moyano, J.A. Gásquez, and L.D. Martinez .............. 125

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

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

AtomicSpectroscopy

Vol. 26(3) May/June 2005

Guidelines for AuthorsAtomic Spectroscopy serves as a medium for the disseminationof general information togetherwith new applications and analytical data in atomicabsorption spectrometry.

The pages of Atomic Spectroscopy are open to allworkers in the field of atomicspectroscopy. There is no chargefor publication of a manuscript.

The journal has around 1500subscribers on a worldwidebasis, and its success can beattributed to the excellentcontributions of its authors aswell as the technical guidanceof its reviewers and theTechnical Editors.

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

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

Atomic Spectroscopy for format.

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

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

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

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

In the interest of speed ofpublication, a pdf file of the type-set text is e-mailed to the corre-sponding author beforepublication for final approval.

Once the issue has been printed,each author receives a final pdf fileof the article including 50 compli-mentary copies of the article andseveral copies of the completeissue.

Additional reprints can bepurchased, but the requestmust be made before printing.

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

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

EDITORAnneliese LustE-mail: [email protected]

TECHNICAL EDITORSGlen R. Carnrick, AADennis Yates, ICPKenneth R. Neubauer, ICP-MS

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

2005 Subscription Rates• U.S. $60.00 includes third-class

mail delivery worldwide; $20.00 extra for electronic file.

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

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

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

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

Back Issues/Claims• Single back issues are available

at $15.00 each.• Subscriber claims for missing

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

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

Copyright © 2005PerkinElmer, Inc.All rights reserved.htpp//:www.perkinelmer.com

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

PerkinElmer and HGA are registered trademarks and AAnalyst and Optima are trademarks ofPerkinElmer, Inc.Excel is a registered trademark of Microsoft Corporation.Milli-Q is a trademark of Millipore Corporation.Minipuls is a trademark of Gilson, Villiers-Le-Bell, France.Suprapur is a registered trademark of Merck, Germany.Triton is a registered trademark of Union Carbide Chemicals & Plastics Technology Corporation.Tygon is a trademark of Norton Co.Registered names and trademarks used in this publication, even without specific indicationthereof, are not to be considered unprotected by law.

89Atomic SpectroscopyVol. 26(3), May/June 2005

Speciation of Chromium With Cloud Point ExtractionSeparation and Determination by ICP-OES

*Pei Liang and Jing LiCollege of Chemistry, Central China Normal University

Wuhan 430079, P.R. China

INTRODUCTION

In recent years, there has beenan increasing demand for informa-tion about speciation because thetoxicity of some elements dependson their chemical form (1).Dissolved chromium is usuallyfound in natural waters in two dif-ferent oxidation states, such asCr(III) and Cr(VI), which have con-trasting physiological effects. Cr(III)is considered an essential trace ele-ment for the maintenance of aneffective glucose, lipid, and proteinmetabolism in mammals (2). On theother hand, Cr(VI) can be toxic forbiological systems (3,4). It is water-soluble and extremely irritating andtoxic to human body tissue owingto its oxidizing potential and per-meability of biological membranes(5,6). Therefore, it is extremelyimportance to accurately define thequantity of both valence forms inenvironmental samples. A greatnumber of speciation studies ofchromium have been carried out insolid and liquid samples (7,8).

The Cr content in natural watersis normally at the low µg L–1 levels.Few analytical techniques are capa-ble of the direct differentiation ofthe chemical forms of Cr in water.The direct determination ofchromium by inductively coupledplasma mass spectrometry (ICP-MS)is possible, but these instrumentsare expensive and not commonlyavailable in an analytical laboratory.As a result, preliminary species sep-aration and preconcentration isrequired before their detection bysensitive analytical techniques (9).For the speciation of chromium,the separation methods reported in

aqueous phase, where the surfac-tant concentration is close to thecritical micelle concentration(CMC). The small volume of thesurfactant-rich phase obtained withthis methodology permits thedesign of extraction schemes thatare simple, low cost, highlyefficient, fast, and of lower toxicityto the environment than thoseextractions that use organicsolvents. Any species that interactswith the micelle system, eitherdirectly (generally hydrophobicorganic compounds) or after a pre-requisite derivatization reaction(e.g., metal ions after reaction witha suitable hydrophobic ligand), maybe extracted from the initial solu-tion and preconcentrated.

Cloud point methodology hasbeen used to separate and precon-centrate organic compounds as astep prior to their analysis in hydro-dynamic analytical systems such asliquid chromatography (17) andcapillary electrophoresis (18). Thephase separation phenomenon hasalso been used for the extractionand preconcentration of metal ionsafter the formation of sparinglywater-soluble complexes (19,20).Recently, CPE as a preconcentra-tion step in conjunction with detec-tion by FIA spectrofluorimetry,FAAS, ETAAS, and HPLC for thespeciation analysis of Cr, Cu, and Fespecies has also been reported(21–28).

The aim of the present work wasto apply CPE as a separation andpreconcentration step for induc-tively coupled plasma optical emis-sion spectrometry (ICP-OES)speciation of chromium. In thedeveloped system, 1-phenyl-3-methyl-4-benzoylpyrazol-5-one(PMBP) was used as the chelatingagent and Triton® X-100 as theextractant. The experimental para-

ABSTRACT

A sensitive and selectivemethod has been developed forthe speciation of chromium innatural water, based on cloudpoint extraction (CPE) andinductively coupled plasma opti-cal emission spectrometry (ICP-OES). Cr(III) reacts with1-phenyl-3-methyl-4-benzoylpyra-zol-5-one (PMBP), yielding ahydrophobic complex which isentrapped in the surfactant-richphase, whereas Cr(VI) remains inthe aqueous phase. Thus, separa-tion of Cr(III) and Cr(VI) can berealized. Total chromium wasdetermined after reduction ofCr(VI) to Cr(III) with ascorbicacid, the reducing reagent. Thedetection limit for Cr(III) was0.81 µg L–1, with an enrichmentfactor of 20 and a relative stan-dard deviation (RSD) of 3.2%(n=11, C=100 µg L–1). The pro-posed method was applied to thespeciation of chromium in nat-ural water samples with satisfac-tory results.

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

the literature are usually based oncoprecipitation (10), solvent extrac-tion (11), solid phase extraction(12,13), and ion chromatography(14).

Separation and preconcentrationbased on cloud point extraction(CPE) is becoming an importantand practical application of surfac-tants in analytical chemistry(15,16). The technique is based onthe property of most non-ionic sur-factants in aqueous solutions toform micelles which become turbidwhen heated to a temperatureknown as the cloud point tempera-ture. Above the cloud point tem-perature, the micelle solutionseparates into a surfactant-richphase of small volume and a diluted

90

meters affecting the CPEseparation/preconcentration andthe subsequent ICP-OES determina-tion of the chromium species wereinvestigated in detail. The proposedmethod has been applied to thedetermination and speciation ofchromium in tap and lake watersamples with satisfactory results.

EXPERIMENTAL

Instrumentation

A PerkinElmer Optima™ 2000DV ICP spectrometer (PerkinElmerAnalytical and Life Sciences, Shel-ton, CT, USA), equipped with thestandard torch assembly and con-ventional cross-flow nebulizer, wasused for the determination ofchromium. The instrumental oper-ating conditions and the analyticalwavelength used are listed in TableI. The pH values were measuredwith a Mettler Toledo 320-S pHmeter [Mettler Toledo Instruments(Shanghai) Co. Ltd., Shanghai, P.R.China], supplied with a combinedelectrode. A thermostated bathmaintained at the desired tempera-tures was used for the cloud pointexperiments. An 80-2 centrifuge(Changzhou Guohua Electric Appli-ance Co. Ltd., P.R. China) was usedto accelerate the phase separation.

Standard Solutions andReagents

Stock solutions (1.000 g L–1) ofCr(III) were prepared by dissolvingCrCl3·6H2O (The First Reagent Fac-tory, Shanghai, P.R. China) in 0.1mol L–1 hydrochloric acid. Stocksolutions (1.000 g L–1) of Cr(VI)were prepared by dissolvingK2Cr2O7 (The First Reagent Factory,Shanghai, P.R. China) in 0.1 mol L–1

nitric acid. The non-ionic surfactantTriton X-100 was obtained fromAmresco and was used without fur-ther purification. A 1.0×10–2 mol L–1

solution of PMBP was prepared bydissolving appropriate amounts ofthis reagent in absolute ethanolfrom the commercially available

product. A buffer solution of 0.1mol L–1 NaAc-HAc was used to con-trol the pH of the solutions. A 10%aqueous ascorbic acid solution wasprepared fresh daily. All otherreagents were of analytical reagentgrade or better. Doubly distilledwater was used throughout thestudy. The pipettes and vesselsused were kept in 10% nitric acidfor at least 24 h and subsequentlywashed four times with doubly dis-tilled water.

Cloud Point Extraction Procedure

For CPE, 10-mL aliquots of asolution containing the analytes,Triton X-100, and PMBP, bufferedat a suitable pH, were kept in thethermostatic bath and maintainedat 80oC for 25 min. Since the surfac-tant density is 1.07 g mL–1, the sur-factant-rich phase can settlethrough the aqueous phase. Phaseseparation was accelerated by cen-trifuging the solution at 3000 rpmfor 5 min. After cooling in an ice-bath, the surfactant-rich phasebecame viscous and was retained atthe bottom of the tube. The aque-ous phases can readily be discardedsimply by inverting the tube. Inorder to decrease the viscosity andfacilitate sample handing prior tothe ICP-OES assay, 0.5 mL 0.1 molL–1 HNO3 was added to the surfac-tant-rich phase. The final solution

was introduced into the nebulizerof the spectrometer by con-ventional aspiration.

Sample Analysis

Tap water was collected in ourlaboratory. Lake water wascollected from East Lake, Wuhan,P.R. China. All water samples werefiltered through a 0.45-mm mem-brane filter (Tianjin Jinteng Instru-ment Factory, Tianjin, P.R. China)and stored at 4oC. The water sam-ple must not be acidified beforestorage because this would changethe chemical species (29).

The water sample was dividedinto two parts:

• Cr(III) determination: To a 10-mL water sample, 1.0 mL of a solu-tion (containing 36 g L–1 TritonX-100 and 1.0×10–2 mol L–1 PMBP)and 1.0 mL of 0.1 mol L–1 NaAc-HAc buffer solution (pH 5.0) wereadded for CPE separation. Afterphase separation, 0.5 mL 0.1 molL–1 HNO3 was added to the surfac-tant-rich phase. The final solutionwas analyzed by ICP-OES.

• Total chromium determination:To a 10-mL water sample, 0.1 mL10% aqueous ascorbic acid wasadded. Then it was assayed asdescribed above.

TABLE IICP-OES Instrumental Operating Conditions

and Analytical Wavelength

Incident power 1100 WPlasma gas (Ar) flow rate 15 L min–1

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

Nebulizer gas (Ar) flow rate 0.8 L min–1

Observation height 15 mmIntegration time 3 sSolution pump rate 0.8 L min–1

Wavelength Cr 283.5 nm

91

Vol. 26(3), May/June 2005

RESULTS AND DISCUSSION

Effect of pH on CPE of Cr(III) and Cr(VI)

The separation of metal ions using the CPE methodinvolves the prior formation of a complex with suffi-cient hydrophobicity in order to be extracted into thesmall volume of the surfactant-rich phase, thus obtain-ing the desired preconcentration. The formation of ametal complex and its chemical stability are the twoimportant factors influencing CPE. The pH plays aunique role on the metal complex formation and sub-sequent extraction, and it proved to be a critical para-meter for the determination of the chromium species.

Figure 1 shows the effect of pH on the extraction ofCr(III) and Cr(VI). It can be seen that extraction wasquantitative for Cr(III) in the pH range of 3.0~7.0, andCr(VI) could not be extracted in the study pH range. Inorder to separate Cr(III) and Cr(VI), a pH of 5.0 wasselected for subsequent work.

Effect of PMBP Concentration

A 10-mL solution containing 1.0 µg of Cr(III) in 3.0g L–1 Triton X-100, at a medium buffer of pH 5.0 con-taining various amounts of PMBP, was subjected to thecloud point preconcentration process. The extractionyield for Cr(III) as a function of the concentration ofthe chelating agent is shown in Figure 2. The yieldincreases up to a PMBP concentration of 5×10–4 molL–1 and reaches near quantitative extraction efficiency.A concentration of 8×10–4 mol L–1 was chosen toaccount for other extractable species that might poten-tially interfere with the assaying of Cr(III).

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 the extraction efficiency ofCr(III) within the Triton X-100 range of 0.1~6.0 g L–1

was examined and the results are shown in Figure 3.Quantitative extraction was observed when the TritonX-100 concentration was above 3.0 g L–1. A concentra-tion of 3.0 g L–1 was chosen as the optimum surfactantconcentration in order to achieve the highest possibleextraction efficiency.

Fig. 1. Effect of pH on the extraction recoveries of Cr(III) andCr(VI)100 ng mL–1 Cr(III) and Cr(VI); 8×10–4 mol L–1 PMBP;3.0 g L–1 Triton X-100.

Fig. 2. Effect of PMBP concentration on the extraction recoveryof Cr(III) 100 ng mL–1 Cr(III); 3.0 g L–1 Triton X-100; pH 5.0.

Fig. 3. Effect of Triton X-100 concentration on the extractionrecovery of Cr(III)100 ng mL–1 Cr(III); 8×10–4 mol L–1 PMBP;pH 5.0.

92

Effects of Equilibration Temperature and Time

It was desirable to employ theshortest equilibration time and thelowest possible equilibration tem-perature as a compromise betweencompletion of extraction and effi-cient separation of phases. Thedependence of extractionefficiency upon equilibration tem-perature and time was studied rang-ing from 60–120oC and 5~30 min,respectively. The results showedthat an equilibration temperature of80oC and an equilibration time of25 min were adequate to achievequantitative extraction.

Effect of Viscosity on ICP-OESDetermination

In order to decrease the viscosityof the surfactant-rich phase and tofacilitate its handling and introduc-tion into the nebulizer of the spec-trometer, some solutions, such as0.1 moL L–1 HNO3, methanol, andethanol alone or containing 0.1moL L–1 HNO3, have been used asthe diluent after separation of thetwo phases. Due to the difficultiesof introducing organic solvents intothe plasma, 0.1 moL L–1 HNO3 waschosen as the diluent for this study.For added volumes of 0.1 moL L–1

HNO3 larger than 0.5 mL, the finalsolution has no influence on theplasma by conventional aspiration.Thus, 0.5 mL 0.1 moL L–1 HNO3

was added to the surfactant-richphase after CPE.

Interferences

In view of the high selectivityprovided by ICP-OES, the onlyinterferences studied were thoserelated to the preconcentrationstep. Cations that may react withPMBP and are extracted to themicelle phase were studied. Thetolerance limits of the coexistingions, defined as the largest amountresulting in the recovery of Cr(III)less than 90%, are given in Table II.It can be seen that the majorcations in the water samples haveno obvious influence on CPE ofCr(III) under the selectedconditions.

Detection Limits and Precision

According to IUPAC definition,the detection limit (3σ) of thismethod for Cr(III) with an enrich-ment factor of 20 is 0.81 µg L–1; andthe relative standard deviation(RSD) is 3.2% (n=11, C=100 µg L–1).

Sample Analysis

The accuracy of the proposedmethod was examined by determin-ing total chromium in environmen-tal water reference materials (ERMs,GSBZ 5009-88, P.R. China). Theanalytical value (637±40 µg L–1)was in good agreement with thecertified value (640±36 µg L–1).

The proposed method wasapplied to the speciation of Cr(III)and Cr(VI) in tap and lake water

samples collected in Wuhan, P.R.China. For calibration purposes, theworking standard solutions weresubjected to the same preconcen-tration procedure as used for theanalyte solutions. In addition, therecovery experiments of differentamounts of Cr(III) and Cr(VI) werecarried out. The results in Table IIIshow that the recoveries were rea-sonable for trace analysis, rangingfrom 96–104%.

CONCLUSION

The feasibility of chromium spe-ciation in water has been demon-strated based on cloud pointextraction of Cr(III) with PMBP inthe presence of the surfactant Tri-ton X-100 and sequential determina-tion by ICP-OES. The developedmethod is definitely simple, repro-ducible, and highly sensitive,because of the distinct and advanta-geous features of CPE (in situ andsingle-step extraction). The methodwas successfully applied to the spe-ciation of chromium in tap and lakewater samples, and the precisionand accuracy of the method are sat-isfactory. The method may also beused for the speciation ofchromium in matrices other thanwater.

Received August 27, 2004.

TABLE IIEffect of Coexisting Ions

on the Determination of Cr(III) (100 ng mL–1)

Coexisting Tolerance Ions Limit of Ions

K+, Na+, Ca2+ 1000 mg L–1

Mg2+, Cd2+, Co2+, Ni2+, Pb2+ 100 mg L–1

Mn2+, Zn2+ 10 mg L–1

TABLE IIIDetermination of Cr(III) and Cr(VI) in Natural Water Samples

Samples Added (µg L–1) Founda (µg L–1) Recovery (%)Cr(III) Cr(VI) Cr(III) Cr(VI)b Total Cr(III) Cr(VI)

Tap water 0 0 4.53 1.85 6.38 – –5.0 5.0 9.56 6.63 16.19 101 96

10.0 10.0 14.20 12.05 26.25 97 102

Lake water 0 0 3.17 5.52 8.69 – –5.0 5.0 8.06 10.47 18.53 98 99

10.0 10.0 13.57 15.32 28.89 104 98

a Mean of five determinations. b Calculated value.

93


REFERENCES

1. A. Kot and J. Namiesnik, TrendsAnal. Chem. 19, 69 (2000).

2. C. Barnowski, N. Jakubowski, D.Stuewer, and J.A.C. Broekaert, J.Anal. At. Spectrom. 12, 1155(1997).

3. A.S. Prasad and D. Oberleas, TraceElements in Human Health andDisease, Academic Press, 1976, pg.79.

4. A.B. Gwizala III, S.K. Johnson, S. Mol-lah, and R.S. Houk, J. Anal. At.Spectrom. 12, 503 (1997).

5. C.M. Andrles, N. Jakubowski, andJ.A.C. Broekaert, Spectrochim.Acta 52B, 189 (1997).

6. R.K. Mugo and K.J. Orians, Anal.Chim. Acta 271, 1 (1993).

7. M.J. Marques, A. Salvador, A.E.Morales-Rubio, and M. de laGuardia, Fresenius’ J. Anal. Chem.362, 239 (1998).

8. M.J. Marques, A. Salvador, A.E.Morales-Rubio, and M. de laGuardia, Fresenius’ J. Anal. Chem.367, 601 (2000).

9. M. Sperling, X.F. Yin, and B. Welz,Analyst 117, 629 (1992).

10. L. Wang, B. Hu, Z.C. Jiang, and Z.Q.Li, Internatl. J. Environ. Anal.Chem. 82, 387 (2002).

11. Z.H. Wang, M. Song, H. Ma, and S.Liang, Microchim. Acta 134, 95(2000).

12. M. Sperling, S.K. Xu, and B. Welz,Anal. Chem. 64, 3101 (1992).

13. P. Liang, T.Q. Shi, H.B. Lu, Z.C.Jiang, and B. Hu, Spectrochim.Acta 58B, 1709 (2003).

14. P. M. Paquet, J. F. Gravel, P. Nobert,and D. Boudreau, Spectrochim.Acta 53B, 1907 (1998).

15. H. Tani, T. Kamidate, and H. Watan-abe, J. Chromatogr. A 780, 229(1997).

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

17. F. Merino, S. Rubio, and D. Perez-Bendito, J. Chromatogr. A 962, 1(2002).

18. R. Carabias-Martinez, E. Rodriguez-Gonzalo, J. Dominguez-Alvarez,and J. Hernandez-Mendez, Anal.Chem. 71, 2468 (1999).

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

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

21. E.K. Paleologos, C.D. Stalikas,S.M.T. Karayanni, G.A. Pilidis, andI. Miltiades, J. Anal. At. Spectrom.15, 287 (2000).

22. E.K. Paleologos, C.D. Stalikas, andM.I. Karayannis, Analyst 126, 389(2001).

23. E.K. Paleologos, C.D. Stalikas,S.M.T. Karayannis, and M.I.Karayannis, Anal. Chim. Acta 436,49 (2001).

24. X.S. Zhu, Z.C. Jiang, B. Hu, and M.F.Li, Chinese J. Anal. Chem. 31, 1312(2003).

25. A.N. Tang, D.Q. Jiang, Y. Jiang, S.W.Wang, and X.P. Yan, J.Chromatogr. A 1036, 183 (2004).

26. C.D. Stalikas, E.K. Paleologos, andM.I. Karayanni, Anal. Bioanal.Chem. 373, 237 (2002).

27. E.K. Paleologos, C.D. Stalikas,S.M.T. Karayannis, and M.I.Karayannis, Anal. Chim. Acta 458,241 (2002).

28. D.L. Giokas, J. Antelo, E.K. Paleolo-gos, F. Arce, and M.I. Karayannis,J. Environ. Monit. 4, 505 (2002).

29. T.M. Florence, Talanta 29, 345(1982).


*Corresponding author.E-mail: [email protected]: +34-981-167000Fax: +34-981-167065

Direct Generation of Stibine From Slurries and Its Determination by ETAAS Using Multivariate Optimization

M.J. Cal-Prieto, M. Felipe-Sotelo, *A. Carlosena, and J.M. AndradeDepartment of Analytical Chemistry, University of A Coruña,

Campus A Zapateira s/n, E-15071, A Coruña, Spain

ABSTRACT

A simple and fast analyticalmethod combining slurry prepa-ration, hydride generation, andtrapping on iridium-treatedgraphite tubes is presented todetermine Sb in soil samples byelectrothermal atomic absorp-tion spectrometry (HG-ETAAS).Chemometric optimization ofthe slurry preparation and gener-ation of stibine was carried out(Plackett-Burman designs andSimplex optimizations). Slurrieswere prepared in 9 mol L–1 HCl,and stibine was generated using0.4 mol L–1 HCl and 2.5% (m/v)NaBH4 (150 mL min–1 Ar flow).Further, a heated quartz tubeatomization system was used(HG-AAS).

The slurry HG-ETAAS methodwas validated with nine certifiedreference materials: three soils,two sediments, two coal flyashes, and two coals. The lowestLOD obtained was 0.08 µg g–1

and good overall precision wasachieved (RSD <7%). The coalsamples required a previous ash-ing step (450°C, up to constantweight). The analyte extractionto the liquid phase was alsostudied.

INTRODUCTION

Up to the present, hydride gen-eration coupled to atomic absorp-tion spectrometry (HG-AAS) hasbeen the most convenienttechnique used to determinehydride-forming elements in aque-ous samples and digested solid sam-ples (1). The chemical vaporgenerated from slurried sampleshas been an attractive methodologysince the early 1980’s because (a)the combination of using slurriesand hydride generation offered asimple and fast sample pretreatmentstep (better than conventionalmethods), (b) selectivity and sensi-tivity were good enough to deter-mine low concentrations ofanalytes in complex matrices, (c)there is low risk of sample contami-nation, and (d) there is low analyti-cal cost (2). It was Haswell et al. (3)who first determined arsenic insolid environmental samples as slur-ries by HG-AAS. This approach hasalso been applied to the determina-tion of lead (4–7) and antimony (8).

Hydride generation and graphitefurnace, employing tubes coatedwith permanent modifiers, lead toin situ trapping of the vapors gener-ated directly from the samples. Inthis case, the graphite tube actseither as a pre-concentration or asan atomization medium. Thismethodology was reviewed byMatusiewicz and Sturgeon (9) whoemphasized its advantages, namelyachieving higher sensitivity and lessprone to interferences. They

reported that many works focusedon the use of permanent modifiersas trapping agents, with the Pt-group of elements (Ir, Zr, Nb, andW) as the most employed. Degrada-tion of the coating of the atomizer(and its control) is of great concernbecause it is possible to enlarge theuseful analytical lifetime of the tubeby recoating the atomizer. Addition-ally, other operational conditionsmust be considered, including theautomatic control of the capillarytransfer line.

Only recently has an extensionof this methodology been developedwhich allows the analysis of solidsamples in the form of slurries. Flo-res et al. (10) determined Hg in coalslurries (1 mol L–1 HNO3) by HG-ETAAS (Au as the permanent modi-fier), where particle size andleaching time are critical. The com-bination of batch hydride genera-tion and in situ preconcentration ina commercial device demonstratedits advantages to quantify As, Sb, Se,Sn, and Hg in beer and wort (11).Ir-treated graphite tubes were usedwith satisfactory results to deter-mine As, Bi, Ge, and Se (12,13) inenvironmental and biological sam-ples as well as As in sediments andcoal slurries (14).

Over the last several years, someauthors coupled slurry-hydride gen-eration and atomic emission (15,16)or used atomic fluorescence spec-trometry (17) for multielementalanalysis. Hydride generation fromsolid samples also revealed itself asa suitable procedure for metal spe-ciation (2). The sequential and on-line extraction of Sb(III) and Sb(V)from blood and liver tissues wasreported for HG-AAS determination(18). A method to determine As(III)and total inorganic As in inorganic(marine sediment, soil, rock salt)and biological samples withoutsample digestion required a contin-uous flow of hydride into thegraphite tube (19).

To the best of our knowledge,the slurry-HG-ETAAS methodologyhas not yet been applied to thedetermination of Sb in environmen-tal matrices. Such a procedurewould be useful to analyze environ-mental matrices and monitoranthropogenic input of heavy met-

95


als in the environment because, asalready reported (20), Sb is a goodtracer of pollution in soils and sedi-ments. Therefore, a new analyticalmethod combining slurry prepara-tion, hydride generation (batch sys-tem), and trapping of the vapors onan Ir-coated graphite furnace tubewas implemented. Robustness wassought out by using chemometrictools: 8-trial Plackett-Burman frac-tional factorial experimentaldesigns to extract the significantvariables for slurry preparation andSimplex optimization. Two atomiza-tion modes were studied as well: Ir-coated graphite tubes (HG-ETAAS)and heated quartz tubes (HG-AAS).In contrast to previous works (12),a true optimization algorithm wasused (Simplex) and all of the mainanalytical stages were reviewed to(a) assess the influential variablesand (b) to optimize them by takingtheir intrinsic interaction intoaccount.

EXPERIMENTAL

Instrumentation

A PerkinElmer AAnalyst™ 800atomic absorption spectrometer(PerkinElmer Life and AnalyticalSciences, Shelton, CT, USA) wasemployed for sample analysis,equipped with Zeeman backgroundcorrection, AS-800 autosampler, Sbelectrodeless discharge lamp (415mA, 217.6 nm and 0.7 nm spectralbandwidth), and pyrolyticallycoated graphite tubes with pre-inserted platforms, treated with Ir.A PerkinElmer MHS-10 hydride gen-erator was coupled to the autosam-pler quartz capillary using a PTFEtransference line.

A vibrating ball mill, Pulverisette7 Fritsch Planetary Micro Mill,(Fritsch Industries, Idar-Oberstein,Germany), equipped with zirconcups and zirconia balls, was used toreduce the particle size. A J2-MCBeckman centrifuge (Fullerton, CA,USA) was employed when necessary.

Reagents and StandardSolutions

All reagents were of analyticalgrade. High-purity water (Milli-Q™Water System, Millipore, Madrid,Spain) was employed throughout.The acids (HNO3 and HCl) wereBaker Instra-analyzed grade (J.T.Baker, Phillipsburg, PA, USA).

The Sb(III) stock solution (1000g L–1) was from Panreac (Barcelona,Spain) and working Sb standardswere prepared daily by dilutingappropriate aliquots of the stocksolution.

Sodium tetrahydroborate(Merck, Darmstadt, Germany), dis-solved in 2% (m/v) sodium hydrox-ide (Panreac), was used as thereducing agent, prepared daily andfiltered before use. A 5.5-g L–1 irid-ium chloride solution was preparedfrom solid reagent (99.9%, Aldrich,Milwaukee, WI, USA) and injectedonto the graphite tubes to providea trapping surface.

All glassware, plastic ware,pipette tips, and storage bottleswere soaked in 10% (v/v) HNO3 for24 h and then rinsed with high-purity water at least three timesprior to use.

Nine certified reference materi-als (CRMs) were used to assessaccuracy (trueness): two NCR CRMsoils (GBW 07401 and GBW 07409;National Research Council of Certi-fied Reference Materials, P.R.China), one NCR CNRC marine sed-iment (BCSS-1, National ResearchCouncil of Canada), two BCR CRMgeological materials (calcareousloam soil, CRM 141, and estuarinesediment, CRM 277, EU CommunityBureau of Reference), four NISTcoal-related materials (coal fly ash1633a and 1633b, coals 1632c and1435, National Institute ofStandards and Technology,Gaithersburg, MD, USA).

Analytical Procedure

The slurries were prepared bysuspending 0.5 g of a sample in 9mol L–1 HCl, adding 1 mL of ethanol(to wet the particles), and homoge-nizing magnetically. The final vol-ume was 10 mL or 20 mLdepending on the Sb concentrationin the sample.

Stibine was generated using abatch system in order to avoideventual blockage in the manifoldand the connectors (frequentlywhen continuous systems areused). A slurry aliquot (0.1–0.2 mL)was placed in the reaction vesseland acidified to 0.4 mol L–1 withHCl. Next, a 2.5% (m/v) NaBH4

solution was added (0.38 mL min–1)and an Ar flow rate of 150 mL min–1

applied to improve contactbetween slurry and reductant. ForHG-ETAAS, Ar sweeps the stibinethrough the transfer line and to theautosampler quartz capillary, andthen introduces it into the graphitetube (1000ºC, 30 s). The hydride isretained on the Ir-coated graphitetube and then atomized at 2300ºCfor 5 s. Table I summarizes theexperimental conditions. For HG-AAS analysis, the hydride vapor wasdirectly introduced into the heatedquartz cell.

TABLE IExperimental Conditions for the

Direct Generation of StibineFrom Slurries and

Its Determination by HG-ETAAS

Slurry Aliquot 0.1–0.2 mLReaction Volume 5 mLNaBH4 2.5 % (m/v)HCl 0.40 mol L–1

Ar Flow 150 mL min–1

Atomizer Ir-coated graphite tubeWavelength/Slit 217.6 (0.7) nm

Lamp Current 415 mA

96


HG-AAS

All of the variables influencingthe slurry preparation and stibinegeneration, as well as theinstrumental conditions, were care-fully examined in order to measureSb by HG-AAS. An 8-trial Plackett-Burman design revealed their influ-ence on the slurry preparation(preliminary studies not shownhere) as follows: (a) volume ofHNO3 and HCl to prepare the liquidmedia, (b) particle size (grinding,as a binary variable yes/no), and(c) use of the ultrasonic bath (tohomogenize the slurries andincrease analyte extraction). Onlythe HCl volume was a statisticallysignificant factor. Accordingly, aunivariate scheme was adequate foroptimization and 3-mL of HCl (c)was found to be the optimum vol-ume.

In order to evaluate the analyti-cal figures of merit, both an aque-ous and standard additionscalibration were performed. Thelatter mode showed an undesirabledouble peak; the explanation could

be that the Sb added by the spikesleads to a very easy-to-release stib-ine, which is atomized first andthus yields a frontal narrow peak.However, stibine generated fromthe Sb in association with the solidparticles will be released a littlelater and will lead, presumably, tothe broadest profile. This is due tothe different kinetic mechanismsinvolved in stibine generation anddepends on whether the analyte isin the solid or liquid phase (21).No samples were quantified by thisprocedure.

HG-ETAAS

The unsatisfactory resultsobtained using the methodologydescribed above, even for the cali-bration standards, made it neces-sary to look for an alternative forthe determination of Sb. Althoughvapor trapping in pre-coatedgraphite tubes has been reported asa suitable way for trace metal deter-mination in slurries (10,13), onlyone reference was found for mea-suring Sb in solid environmentalsamples (11). In the present work,stibine was generated directly fromsoil/sediment slurries, trapped in an

Ir pre-coated atomizer, and Sb wasdetermined by electrothermal atom-ization. Two main stages needed tobe studied, such as slurry prepara-tion and hydride generation.

Slurry Preparation

A saturated Plackett-Burmanexperimental design was employedto study the slurry preparation inthe same way as HG-AAS (Table II).From the experimental t-values,only the amount of HNO3 (c) andHCl (c) used was statistically signifi-cant (ttab = 1.96, 95% confidence).The negative sign of HNO3 (c) canbe due to the oxidative character ofSb, which might form the insolubleSb(V) silicate complexes in thepresence of the oxidizing acids(yielding low extraction recoveries)(22). Accordingly, HNO3 (c) wasnot added. The most remarkabledrawback of the Plackett-Burmandesigns is that confounding effectshave to be evaluated, i.e., to eluci-date whether the effects can beattributed only to a particular vari-able or to the variable itself plussome other two-variable combina-tions (23). In our study, noconfounding effects were observed.

TABLE IIExperimental Conditions for Each Trial of the Experimental Design and Statistical Effect

Calculated for Each Variable During Slurry Preparation, HG-ETAAS Analysis

Variable

Trial HNO3 HCL ‘Grinding’ Ultrasonic ‘Dummy’ ‘Dummy’ ‘Dummy’ Absorbance(mL) (mL) Bath (min)

1 2.5 7.5 Yes 0 - - - 0.2982 0 7.5 Yes 30 - - - 0.4023 0 0 Yes 30 - - - 0.2464 2.5 0 No 30 - - - 0.1785 0 7.5 No 0 - - - 0.3076 2.5 0 Yes 0 - - - 0.1877 2.5 7.5 No 30 - - - 0.278

8 0 0 No 0 - - - 0.291

Statistical Effect

Variable HNO3 HCl ‘Grinding’ Ultrasonic ‘Dummy’ ‘Dummy’ ‘Dummy’(mL) (mL) Bath (min)

texperimental –2.6 3.3 0.6 0.2 –1.1 –0.3 –1.3

97


Following, a univariate approachwas taken to optimize the volumeof HCl (c). The acid increased theanalyte response, although thehighest volumes (>7.5 mL) causedan irregular trend. Hence, 7.5 mL ofHCl (c) was set as the optimumvalue (9 mol L–1). In addition, 4 mLof ethanol was added to improveslurry withdrawal [final concentra-tion 20% (v/v)], but this induced aviolent reaction resulting in samplesputtering. After assessing differentamounts of ethanol (0.5 to 3 mL), itwas observed that 1 mL was goodenough to ensure good pipetting,precise results, and low blank val-ues. However, a slight reduction inthe atomic signal was observed.

Two additional factors have tobe taken into account since accu-racy and precision greatly dependon (a) volume of the slurry with-drawn to generate stibine and (b)sample-mass/ liquid phase-volumeratio which is a very important fac-tor in slurry analysis (24). Thisratio, in turn, depends on theexpected analyte content of thesample.

Two slurry volumes (10 and 20mL), several sample masses (from0.05 g to 1.00 g), and different vol-umes of slurry aliquots withdrawnto generate stibine (10 to 200 µL)were investigated. Poor precisionfigures (RSD >15%) were obtainedfor both slurry volumes when lowsample mass and low aliquot vol-umes were employed (even thoughthey yielded slightly higherresponses). Therefore, high levelsof sample mass (0.25–0.50 g for10 mL; 0.5–1.0 g for 20 mL) and100 to 200 µL slurry aliquots hadto be employed.

Optimization of the Generationof Stibine

Once the slurry is prepared,hydride generation can be consid-ered. Currently, a classical trial anderror approach is employed eventhough it cannot assure that a real

optimum is found when allvariables are considered simultane-ously (mainly, if they interact).Therefore, a multivariate optimiza-tion procedure was employed. Wecarried out a Simplex optimizationwhich is an iterative experimentalprocess whereby the algorithm pro-poses new experimental trials andevaluates the responses. The Sim-plex starts from an initial design ofn+1 experiences (n = number ofvariables) and follows a sequentialprocess where the worst conditionsare eliminated based on theresponse variable. This results inthe Simplex evolving towards morefavorable conditions. Additionally,experiences retained during then+1 steps must be re-evaluated,which prevents the Simplex to bestuck around a false favorableresponse. Finally, experiments lyingoutside the practical limits of thevariables are given very unfavorableresponses, forcing the Simplex toleave those impractical values;more details are given elsewhere(25). For this study, a modifiedNelder and Mead Simplex wasemployed.

The concentrations of HCl andNaBH4, as well as the argon flowrate, were optimized simulta-neously to generate stibine from aGBW 07401 certified soil (0.5 g/20mL). An irregular Simplex wasdeveloped using the Matlab® pro-gram (Mathworks Inc., Newton,MA, USA), where the maximumabsorbance of the atomic peak isthe dependent variable. Two Sim-plex were developed starting fromdifferent experimental conditionsto guarantee that an absolute opti-mum is reached. Since the multidi-mensional space generated by theSimplex is hard to visualize, theresponse surface obtained after aprincipal component analysis of theexperimental conditions is shownin Figures 1A and 2A. The ‘principalcomponent scores vs. absorbance’plot resembles the ‘true’ responsesurface quite well, because the PC1

and PC2 subspace explains up to99% (Simplex 1) and 98% (Simplex2) of the variance. It is quite easy tofollow the Simplex movement inthe PC1-PC2 subspace (Figures 1and 2); this approach was also pro-posed by Koch et al.(26). Both Sim-plex lead to similar optimumconditions: 0.4 mol L–1 HCl, 150 mLmin–1 Ar flow and 2.5% NaBH4 (Sim-plex 1) and 0.5 mol L–1 HCl, 180mL min–1 Ar flow and 3% NaBH4

(Simplex 2). For this study, the firstoptimum was selected to preservetube lifetime and for economic rea-sons.

Furnace Temperature Program

Electrothermal atomization of Sbfrom stibine needs different furnacetemperature programs than for con-ventional Sb atomization. The dry-ing and pyrolysis steps aresubstituted with a hydride-trappingstep, followed by Sb atomizationitself. Then, a cleaning step is rec-ommended to avoid memoryeffects (although this is unlikelybecause the sample matrix is notintroduced into the graphite tube).For trapping, several coatings wereproposed in the literature althoughPd is generally recommended. Nev-ertheless, Ir was chosen due to itslower volatility and good properties(27,28). Three replicates (85 µLeach) of a 5.5-g L–1 Ir solution wereused for tube coating following thefurnace program presented in TableIV.

A GBW 07401 soil slurry wasprepared to evaluate the best fur-nace temperature for trapping thehydride. Stibine was generated anda trapping curve built(temperatures ranged from200–1000ºC, atomization tempera-ture were fixed at 2000ºC). Tem-peratures lower than 1000ºC didnot retained the analyte, while1000ºC was enough toquantitatively trap the hydride. Theminimum collection time requiredto obtain maximum absorbancewas determined by testing the dif-

98

Fig. 1. Response surface obtained for Simplex 1. Absorbances vs. PC1 and PC2 scores of the experimentalconditions. Optimum values: HCl = 0.4 mol L–1, NaBH4 = 2.5% (m/v),Ar flow = 150 mL min–1.

Fig. 2. Response surface obtained for Simplex 2. Absorbances vs. PC1 and PC2 scores of the experimentalconditions. Optimum values: HC] = 0.5 mol L–1, NaBH4 = 3% (m/v),Ar flow = 180 mL min–1.

99


ferent lengths of time required forvapor generation plus stibinesweeping; and 30 s was selected forthis study. Finally, the atomizationtemperature was evaluated from2000–2500ºC. Based on the profileof the atomic peak, 2300ºC waschosen, and the final program ispresented in Table III.

The lifetime of the Ir-coatedtubes was monitored with controlcharts and 200 firing cycles werefound to be the average useful life-time. After recoating, the graphitetube can be used for an additional100 cycles.

Figures of Merit

The linear response extended upto 2 ng mL–1 of Sb. The standardadditions calibration method wasselected because its slope differedstatistically from that of the aque-ous calibrations (0.1782 mL ng–1

and 0.1142 mL ng–1, respectively,95% confidence level).

The instrumental limits of detec-tion (LOD) and limits of quantita-tion (LOQ) were 0.16 and 0.52 ngmL–1, respectively. The methodLOD and method LOQ depend onsample mass, volume of slurry, andvolume of the aliquot of slurry usedto generate the stibine. The lowestLOD was 0.08 µg g–1 (LOQ of 0.25µg g–1), which agrees with thatreported for the analysis of wortslurries by HG-ETAAS (0.07 µg g–1)(11). As expected, both the methodLOD and the method LOQ arehigher than the instrumentalLOD/LOQ because of the dilutionsthe samples suffer during the ana-lytical procedure (slurry prepara-tion and stibine generation),although they seem to be adequateto determine Sb in environmentalsamples. The characteristic mass (±standard deviation) was 11 ± 2 ngfor aqueous standards.

The instrumental precision wasstudied using two aqueousstandards (0.5 and 1.0 ng mL–1 Sb)and two soil samples with quite dif-

ferent Sb concentrations. The RSDswere satisfactory, <1.6% and <3.3 %for standards and samples, respec-tively. Furthermore, the overall pre-cision of the method was quitegood, the RSD values <7%, which iswithin the range usually reportedfor this methodology (2).

Accuracy (trueness) of themethod was studied in two ways. Itis important to note that since theuse of an oxidizing medium wasavoided to prepare the slurry and astrong reductant was used to gener-ate the hydride, it follows that stib-ine is generated only from Sb(III).First, the instrumental accuracy wastested using recovery assays. Differ-ent amounts of an aqueous Sb stan-dard were added to several slurriesof the GBW 07401 soil sample andthe final slurries were analyzed sixtimes. All recoveries were good,from 103–109%. The overall accu-racy of the method was assessed byanalyzing several CRMs: three soils,two sediments, two coal fly ashes,and two coals. The latter threematrices were selected to ascertainthat the method might be extendedto such complex matrices. Goodagreement was found between theexperimental and reference (guide)values for all CRMs (Table IV); theexperimental confidence intervals

overlapped with the certified val-ues (Student’s t-test, 95%confidence level). Accordingly, itcan be concluded that the combina-tion of slurry preparation, hydridegeneration, and HG-ETAAS is a suit-able method for the direct determi-nation of Sb in soils, sediments, andcoal fly ash. Unfortunately, lowrecoveries were achieved for thecoal samples (30%).

Four alternatives were studiedto improve the results for coals.First, a surfactant [sodium pyro-phosphate, 0.05% (m/v)] was addedto the slurries to improve manualpipetting. Second, an extractionstep using an ultrasonic bath wasperformed (30 min) before genera-tion of the stibine. This wouldincrease the extraction of the ana-lyte to the liquid phase and, accord-ingly, the efficiency during thegeneration of the stibine. Neverthe-less, the recoveries were lowerthan 35%. Next, the composition ofthe liquid media was changed. Theuse of HNO3 and some mixtureswith other reagents such as H2O2 orH2SO4 were tested to obtain an oxi-dant medium and, hence, to elimi-nate the organic matrix present incarbon. Unfortunately, recoveriesdid not improve, maybe becausethe organic matrix hindered the

TABLE IIIExperimental Conditions Used to Coat Ggraphite Tube With Ir and

Furnace Program to Determine Sb in Soil Slurries by HG-ETAAS(including trapping)

Ir Coating Program Temp. Tramp Thold Ar Flow

(ºC) (s) (s) (mL min–1)

Dry 1 150 30 40 250Dry 2 200 20 30 250

Atomization 2000 0 5 0

Furnace ProgramTrapping 1 1000 1 30 250Trapping 2 1000 1 2 250Atomization 2300 0 5 0

Cleaning 2300 1 3 250

100

generation of stibine as reportedpreviously (29). Finally, it wasdecided to ash the samples toremove the organic matrix beforecarrying out any other analyticalsteps (450ºC up to constantweight). This option yielded satis-factory results (Table IV).

Extraction of Antimony into theLiquid Phase of the Slurry

The percentage of analyteextracted into the liquid phase ofthe slurry was determined to evalu-ate whether stibine is generatedfrom the solubilized analyte or fromthe solid particles (only the super-natant was analyzed after centrifug-ing the slurry at 9000 rpm). Thefigures were 68 and 91% for soilsGBW 07401 and GBW 07409,respectively. Since their total recov-eries were 92 and 95%, respectively(Table IV), it has to be assumed thatstibine is generated from both theliquid and the solid phases.

This hypothesis seems to be con-firmed by the fact that although alow extraction was obtained forsediment BCSS-1 (~30%), the over-all recovery was good (88%). Even

though the extraction for coal flyash was much higher (~85%), thetotal recovery was similar (88%),probably because stibine is noteasily developed from such refrac-tory particles. Coal samples showedclose to 100% extraction when theywere ashed previously.

CONCLUSION

In this study it was demonstratedthat the direct determination of Sbin slurries through its previous con-version to stibine, which becomestrapped on an Ir-coated graphitetube, constitutes a reliable analyti-cal procedure for the analysis ofsolid samples. The procedureavoids typical problems associatedwith the sample matrix when thedirect analysis of slurries is carriedout.

Chemometric tools (experimen-tal design and Simplex optimization)allowed the successful optimizationof both the generation of stibineand its quantitation. The HCl con-centration was critical for the slurrypreparation.

The HG-AAS technique was notvalid for the direct analysis of slur-ries, whereas the HG-ETAAS tech-nique yielded accurate and preciseresults (LOD 0.08 µg g–1, precisionRSD <7%). The direct determina-tion of Sb in several complex matri-ces (soils, sediments, coal fly ash,and coals) was suitable asevidenced by the analysis of certi-fied reference materials. For coals, aprevious sample ashing is manda-tory.

ACKNOWLEDGMENTS

This work was partiallysupported by the Vicerrectoradode Investigación of the Universityof A Coruña.

M.J.C.P. and M.F.S. acknowledgethe Galician government (Xunta deGalicia) and the Spanish Ministry ofEducation, respectively, for theirPh.D. grants.

Received January 18, 2005.

REFERENCES

1. R. Kalähne, G. Henrion, A. Hulan-icki, S. Garbos, and M. Walcerz,Spectrochim. Acta Part B, 52, 1509(1997).

2. H. Matusiewicz, AppliedSpectroscopy Reviews 38, 263(2003).

3. S.J. Haswell, J. Mendham, M.J. Blut-ler, and D.C. Smith, J. Anal. At.Spectrom. 3, 731 (1988).

4. Y. Madrid, J. Meseguer, M. Bonilla,and C. Cámara, Anal. Chim. Acta237, 181 (1990).

5. Y. Madrid, M. Bonilla, and C.Cámara, Analyst 115, 563 (1990).

6. I. López-García, I. Nuño-Peñalver,and M. Hernández -Córdoba,Mikrochim. Acta 109, 211 (1992).

7. C. Cabrera, Y. Madrid, andC.Camara, J. Anal. At. Spectrom. 9,1423 (1994).

8. M.B. de la Calle-Guntiñas, Y. Madrid,and C. Camara, Analyst 116, 1029(1991).

TABLE IV Sb Concentrations of CRMs Analyzed by HG-ETAAS

(± confidence interval; n=6; 95% confidence level)

Sb Concentration (± IC, µg g–1)

Material Certified Obtained RecoveryValue Value (%)

Soil GBW 07401 0.87 ± 0.03 0.80 ± 0.10 92Soil GBW 07409 0.21 ± 0.03 0.20 ± 0.04 95Calcareous Loam Soil BCR 141 (0.7 ± 0.3)a 0.5 ± 0.2 (71)Marine Sediment BCSS-1 0.59 ± 0.06 0.5 ± 0.2 88Estuarine Sediment BCR 277 (3.9 ± 1)a 4.2 ± 1.2 (108)Coal Fly Ash 1633a 6.8 ± 0.4 6.0 ± 2.3 88Coal Fly Ash 1633b (6)a 5.1 ± 0.9 (85)Coal 1632c (Bituminous) 0.461 ± 0.029 0.14 ± 0.15 30Coal 1632c (Bituminous)b 0.461 ± 0.029 0.54 ± 0.03 117

Coal 1635 (Subbituminous)b (0.14)a 0.16 ± 0.02 (118)

( )a Not certified.b Previous ashing step (450°C up to constant weight).

101


9. H. Matusiewicz and R.E. Sturgeon,Spectrochim. Acta Part B 51, 377(1996).

10. E.M.M. Flores, B. Welz, and A.J.Curtius, Spectrochim. Acta Part B56, 1605 (2001).

11. H. Matusiewicz and M. Mikolajczak,J. Anal. At. Spectrom. 16, 652(2001).

12. J. Moreda-Piñeiro, P. López-Mahía,S. Muniategui-Lorenzo, E. Fernán-dez-Fernández, and D. Prada-Rodríguez, Spectrochim. Acta PartB 57, 883 (2002).

13. J. Moreda-Piñeiro, P. López-Mahía,S. Muniategui-Lorenzo, E. Fernán-dez-Fernández, and D. Prada-Rodríguez, J. Anal. At. Spectrom.17, 721 (2002).

14. M.A. Vieira, B. Welz, and A.J. Cur-tius, Spectrochim. Acta Part B 57,2057 (2002)

15. A. Schwingel Ribeiro, M. AntunesVieira, and A. J. Curtius,Spectrochim. Acta Part B 59, 243(2004).

16. M.A. Vieira, T. Dillenburg, B. Welz,and A.J. Curtius, J. Anal. At. Spec-trom. 19, 297 (2004).

17. P. Cava-Montesinos, M.L. Cervera,A. Pastor, and M. de la Guardia,Talanta 62, 175 (2004).

18. Y. Petit de Peña, O. Vielma, J.L.Burguera, M. Burguera, C. Rondon,and P. Carrero, Talanta 55, 743(2001).

19. H. Matusiewicz and M.Mroczkowska, J. Anal. At. Spec-trom. 18, 751 (2003).

20. M.J. Cal-Prieto, A. Carlosena, J.M.Andrade, M.L. Martínez, S. Muni-ategui, P. López-Mahía, and D.Prada, Water, Air, and Soil Pollut.129, 333 (2001).

21. J. Moreda-Piñeiro, P. López-Mahía,S. Muniategui-Lorenzo, E. Fernán-dez-Fernández, and D. Prada-Rodriguez, Anal. Chim. Acta 461,261 (2002).

22. M.J. Nash, J.E. Maskall, and S.J. Hill,J. Environ. Monit. 2, 97 (2000).

23. K. Jones, Intern. Lab. 32 (1986).

24. M.J. Cal-Prieto, M. Felipe-Sotelo, A.Carlosena, J. M. Andrade, P. López-Mahía, S. Muniategui, and D. Prada,Talanta 56, 1 (2002).

25. B.G.M. Vaadeginste, D.L. Massart,L.M.C. Buydens, S. De Jong, P.J.Lewi, and J. Smeyers-Verbeke,Chemometric: A Text Book, Else-vier (1998).

26. I. Koch, C.F. Harrington, K.J.Reimer, and W. R. Cullen, Talanta44, 771 (1997).

27. D.L. Tsalev, A.D´Ulivo, L. Lampug-nani, M. di-Marco, and R. Zamboni,J. Anal. At. Spectrom. 10, 1003(1995).

28. J. Murphy, G. Schlemmer, I.L.Shuter, P. Jones, and S.J. Hill, J.Anal. At. Spectrom. 14, 1593(1999).

29. C. Cámara, Cámara de grafito y gen-eración de hidruros en absorciónatómica, in: M. Blanco, V. Cerdá, A.Sanz-Medel, (Eds.), EspectroscopíaAtómica Analítica, Associación deCiencias y Técnicas Medioambien-tales (AEST), Palma de Mallorca,Spain (1990)


Corresponding author. E-mail: [email protected]

Application of Biological Substrates to the SpeciationAnalysis of Cr(III) and Cr(VI)

Belén Parodia, Griselda Pollab, Liliana Valientec, and *Patricia Smichowskid

a Instituto Nacional de Tecnología Industrial, Mecánica, Laboratorio de Microscopía Electrónica, Casilla de Correo 157, B1650KNA-San Martín, Argentina

b Comisión Nacional de Energía Atómica, Unidad de Actividad Física, Centro Atómico Constituyentes, Av. Gral. Paz 1499, B1650KNA-San Martín, Argentina

c Instituto Nacional de Tecnología Industrial, Química, Laboratorio de Análisis de Trazas, Casilla de Correo 157, B1650KNA-San Martín, Argentina

d Comisión Nacional de Energía Atómica, Centro Atómico Constituyentes, Unidad de Actividad Química, Av. Gral. Paz 1499, B1650KNA-San Martín, Argentina

ABSTRACT

Baker’s yeast (Saccharomycescerevisiae) was used as substratefor the selective retention ofCr(III) in the presence of Cr(VI)from aqueous solutions. Theexperiments were performed inbatch as a suitable and simplemethod to obtain information ofchromium uptake. The effect ofchemical and physical variablesaffecting the degree of uptake ofchromium was tested in order toselect the optimum analyticalconditions for the selective Crretention by Saccharomycescerevisiae. The parameters stud-ied were reaction media and pH,amount of biomass, temperature,and contact time. The influenceof some concomitant ions up to50 mg L–1 was also tested.

During all the steps of theoptimization process, Cr(VI)remained in solution, whileCr(III) was accumulated by theyeast cells. Quantitative determi-

nations of the Cr species in thebiomass and supernatant solutionwere carried out by means offlame atomic absorption spec-trometry (FAAS). A preconcentra-tion factor of 15 was achieved forCr(III) when 75 mL of water wasprocessed. The detection limitsfor Cr(III) and Cr(VI) were 0.2and 3.0 ng mL–1, respectively.

Compared with other meth-ods, this approach allows thedetermination of Cr(III) andCr(VI) in different phases. A mor-phological study and microscopi-cal characterization of the yeastcell wall were carried out byScanning Electron Microscopy(SEM) complemented withenergy dispersive X-ray analysis(EDAX). The speciation analysisof inorganic Cr in different kindsof spiked natural waters was per-formed following the proposedmethod. The recoveries in eachphase and in all cases werebetween 89 and 108%.

INTRODUCTION

The persistence, fate, bioavail-ability, and toxicological effects oftrace elements in the environmentare directly related to the physico-chemical forms in which theyoccur. Chromium represents one ofthe most widely studied elementsdue to the different toxicologicaland biological behavior of itsspecies in living organisms. Triva-lent Cr compounds are much lesstoxic than those of hexavalent

chromium (1,2). Owing to the highoxidation potential of Cr(VI) andthe ease to penetrate biologicalmembranes, it is approximately 100times more toxic than Cr(III) (3).Hexavalent chromium is also apotent carcinogenic agent, andacute and chronic toxicityproblems associated with Cr(VI)exposure include ulceration of theskin, inflammation of the larynx, aswell as damage to the kidneys andlungs (4).

A variety of separation anddetection techniques have beenproposed for the determination

of Cr species in water samples.More recently, advances have beenmade with more sophisticated tech-niques. Hyphenated techniquesbased on the combination of a pow-erful separation technique with asensitive element-specific detectorare among the most promisingapproaches to determineselectively Cr species in a varietyof matrices. Reported methods nor-mally require relatively complexseparation schemes, such asreversed phase ion-pair high perfor-mance liquid chromatography(HPLC) (5) or capillary electro-phoresis (CE) (6), coupled columnHPLC-inductively coupled plasmamass spectrometry (ICP-MS) (7),and isotope dilution (ID-MS) tech-niques (8) that are not always avail-able in all laboratories. It istherefore essential to investigateother alternatives for the reliableand low-cost determination ofspecies at trace levels. One alterna-tive to inorganic sorbents, such asbentonite (9), charcoal (10), andalumina (11), is to use biologicalsubstrates for metal speciationsince microorganisms are capableof binding dissolved metals. Cellwalls have a common compositionof proteins and carbohydrates withwhich the metallic species canreact.

Biological substrates have beenused with preconcentration and, toa lesser extent, for speciationpurposes. Cámara and her group(12–14) employed baker´s yeastbiomass to determine selectivelyCH3Hg+ and Hg2+, Sb(III) and Sb(V),Se(IV) and Se(VI) based on their

103


Microscopy characterization ofthe yeast cells was carried out byscanning microscopy with a Philips515 microscope (Philips ExportB.V., Eindhoven, The Netherlands),equipped with an EDAX PV9100probe (EDAX International Inc.,Prairie, View, IL, USA). An AltronixpH meter (Buenos Aires,Argentina), equipped with a glasselectrode and an AgCl/Ag referenceelectrode, was used for measuringthe pH of the solutions. For separa-tion of the aqueous phase in con-tact with the yeast cells, alaboratory centrifuge was used.

Reagents

All reagents were of analyticalreagent grade unless otherwisestated. Deionized water fromNanopure (Barnstead, Dubuque, IA,USA) was used throughout. Stan-dard solutions of Cr(III) and Cr(VI)were prepared by stepwise dilutionof a 1000-mg L–1 standard solution(Merck, Darmstadt, Germany) justbefore use. Diluted working solu-tions were prepared daily by serialdilutions of this stock solution.

All solutions containing the acidsstudied were prepared at therequired concentrations by dissolv-ing appropriate amounts of eachcompound in deionized water or bydilution. A range of different pHvalues was prepared by adjustingwith HCl or NaOH.

The cation and anion solutions[(As(III), Ca2+, Cd2+, Co2+, Cu2+,Fe3+, Hg(II), Mg2+, Mn2+, Ni2+, Pb2+,Sb(III), V(V), Zn2+] used in interfer-ence study were prepared by

appropriate dilution of the stocksolutions.

Biosorbent Material

The organisms used in this studyas biosorbent material were fromcommercial dry baker’s yeast. Theyeast cells were microbiologicallyidentified by the CNEA Radiobiol-ogy Laboratory as a pure culture ofSaccharomyces cerevisiae. Identifi-cation of the biomass wasperformed by their morphologicalcharacteristics and other proper-ties. These included the ability forfermentation of sucrose, maltose,and galactose and the inability toassimilate potassium nitrate, todegrade urea, and to grow in creati-nine.

The chemical composition andmorphological modification on theyeast surfaces was analyzed in indi-vidual particles and in the bulkusing Scanning Electron Microscopy(SEM) and Energy Dispersive X-RayAnalysis (EDAX). The yeastsamples, before and after Cr accu-mulation, were mounted onaluminum stubs, coated with goldbefore the analysis.

Experimental Procedure

All experiments were performedin batch (by triplicate) as a suitableand simple method to obtain infor-mation of Cr uptake bySaccharomyces cerevisiae. A deion-ized water blank was alsoprocessed and analyzed. Baker’syeast (0.75 g) was placed in cen-trifuge tubes and 0.25 mL of a 2.0-mg mL–1 solution of Cr(III), 0.25 mL

different toxicities. Smichowski etal. (15) demonstrated the effective-ness of yeast cells for the speciationanalysis of inorganic As.

In the specific case ofchromium, different biological sub-strates were evaluated for specia-tion purposes. Neidhart et al. (16)used human red erythrocytes underphysiological conditions for theselective determination ofchromate in the presence of Cr(III).The removal of trivalent and hexa-valent Cr by seaweed biosorbentwas reported by Kratochvil andcoworkers (17). The use of intactand dehydrated cells of Candidautilis (18) was useful for the selec-tive sorption of Cr(VI) in the pres-ence of other metals. Other authorsused immobilized yeast in off-linecolumn procedures. Bag andcoworkers (19) employed sepioliteand Pérez-Corona et al. (20) usedalginate as support.

The aim of this work was toexamine the potential of yeast cellsas a selective substrate for specia-tion analysis of Cr using a simpleand inexpensive batch procedurethat allowed the characterization ofthe metal-binding process. Theeffect of chemical and physicalparameters affecting the degree ofbioaccumulation of Cr(III) andCr(VI) in yeast cells was examined.

EXPERIMENTAL

Instrumentation

A PerkinElmer Model AAnalyst™300 flame atomic absorption spec-trometer (PerkinElmer Life and Ana-lytical Sciences, Shelton, CT, USA),equipped with a deuterium back-ground corrector, was used. APerkinElmer Cr hollow cathodelamp was employed as the lightsource for Cr determination. All sig-nals were monitored at 357.9 nmwith a spectral slit width of 0.7 nm.The instrumental parameters usedfor the Cr determinations are sum-marized in Table I.

TABLE IOperating Conditions for Cr Determination by FAAS

Instrument PerkinElmer, Model AAnalyst 300 AASFlame type Air-C2H2 (reducing, fuel-rich)Burner height 12 mmWavelength 357.9 mmSlit width 0.7 nmLamp current 15 mAMeasurement mode Time average, Read time: 2.0 s, Read delay: 2.0 s

˘

104

of a 2.0-mg mL–1 solution of Cr(VI)and 11.5 mL of acetic acid solution,adjusted to pH 4.0, were added tothe tubes. The samples were manu-ally agitated for 5 min and then cen-trifuged 10 min at 5000 rpm; bothphases were separated by filtration.Cr(VI) was determined in the liquidphase by FAAS. The solid phase wasre-dissolved in 10 mL of a 1.0-molL–1 HCl solution.

To evaluate the effect of differ-ent variables on the Cr uptake, theseparation process was carried outon each oxidation state separately.Cr(III) showed high affinity for theyeast cells and was almostcompletely retained by thebiomass, while Cr(VI) remained inthe supernatant. After optimization,mixtures of Cr(III) and Cr(VI) inwater were prepared to check theability of the yeast cells toselectively retain Cr(III) species.


Morphological Study of YeastCells

The morphological study of thecommercial yeast revealed the uni-form and small size of the particles.The average particle size of the Sac-charomyces cerevisiae was 2.5–3.0mm, and when dried, aggregates oflarger sizes are formed. The cellwall is the initial site of action fortoxic element species. A diversityof active binding sites is containedin the cell wall. An important mech-anism for the accumulation ofmetal ions is their adsorption onthe surface of the substrate throughinteractions with chemicalfunctional groups such as the car-boxylate, hydroxylsulphate, phos-phate, and the amino groups (21).As a result of metal binding, mem-brane damage may occur. Figure 1-ashows a close micrograph of thebiomass before Cr uptake, wherespherical particles can be observed.After Cr uptake, no morphologicalchanges were observed (Figure 1-b),

and the particles remained intact.This study evidenced that noeffects of a chemical process on thecell walls were detected under theexperimental procedure adopted.

Chemical and Physical Parameters Affecting Cr Uptake

The effect of chemical and physi-cal parameters on the Cr uptake byyeast cells was evaluated.Chromium was quantified in thesolid and in the liquid phase. Thenature of the medium and the pH,temperature, contact time, amountof analyte, amount of biomass, andsample volume were the variablesconsidered.

Effect of the Medium and pH on Cr Uptake

The retention of Cr(III) andCr(VI) was evaluated in five media:acetic, phosphoric, citric, oxalic,and tartaric acids. The acidsolutions (0.1 mol L–1) were pre-pared with the pH adjusted bydropwise addition of NaOH or HClsolutions in order to cover therange from 2–10.

In all cases, Cr(III) wassequestered by the yeast cells andCr(VI) remained in the supernatant.The results shown in Figure 2 illus-trate that the uptake mechanism

depends on the medium studiedand to a lesser extent on the pH.It is evident from Figure 2 (a, b, c,d, e) that the acetic and oxalic acidsenable the best separation of bothspecies. For further experiments,acetic acid at pH 4 was selected asthe more convenient acid for Cr(III)and Cr(VI) separation because theseparation of the species is maxi-mum. When this acid was tested,both species were completely sepa-rated at pH 4 because Cr(VI) exhib-ited the lower recovery, whileCr(III) was quantitativelysequestered by the biomass. Thiseffect could be attributed to thedifferent binding sites (functionalgroups) involved in Cr(III) accumu-lation which requires different con-ditions.

Hexavalent Cr retention exhibitslow dependence on pH but is moresensitive to pH changes thanCr(III). It is related to the mecha-nism of metal uptake on the surfaceof the yeast cells and reflects thenature of the physico-chemicalinteraction of both species. Despitethe expected complexity of thebiosorption process, the slightdependence of Cr(III) binding onpH with the different acids evalu-ated suggests that yeast cells mayinteract with the analyte through

Fig. 1. Scanning electron micrograph of yeast cells: (a) before Cr uptake; (b) after Cr uptake.

105


covalent binding. We observedsimilar results for As when Saccha-romyces cerevisiae was used forthe same purpose (15). Even whenthe major mechanisms responsiblefor metal biosorption may includeionic interaction and complex for-mation between metal ions and bio-mass (22), other studies regardingthe nature of bonding betweenmetallic ions and algae cell wallsexplained the greater adsorption ofCu, in comparison with other ele-ments, as an indication of covalentbindings (23).

Effect of Temperature on CrUptake

The influence of temperature onCr sequestration by yeast cells wastested from 0 to 80oC. Figure 3shows that Cr(III) was almost quan-titatively sequestrated by the yeastcells in all of the temperatureranges evaluated. For Cr(VI), a max-imum biosorption was obtained at atemperature of 40ºC. This increasein Cr(VI) retention at an increasedtemperature may be explained by ahigher affinity of the sites for themetal or to an increase in binding

sites on the surface of the biomass.Our results are in good agreementwith those reported by Goyal andcoworkers (24). Room temperaturewas adopted for furtherexperiments since a good separa-tion of Cr(III) and Cr(VI) specieswas obtained.

Fig. 2 (a–e). Effect of pH on Cr(III) uptake. Yeast, 500 mg; Cr concentration, 80 ng mL–1; Contact time, 15 min; Temperature, room temperature; Reaction media: (a) oxalic acid; (b) tartaric acid; (c) acetic acid; (d) phosphoric acid; (e) citric acid.

a)

b)

c)

d)

e)

106

no retention of Cr(VI) was observed. Beyond this value,the retention of the analyte is constant. These are theoptimal conditions to achieve the speciation analysis ofthe inorganic Cr species.

Specific uptake has been demonstrated to depend onthe analyte species and on the nature of the microorgan-isms involved (25). Itoh et al. (26) suggested that electro-static interactions between cells might be a significantfactor in biomass concentration dependency.

In order to estimate the achievable preconcentrationfactor, it is necessary to evaluate the maximum volumeof a sample that can be contacted with the biomass.Increased sample volumes (5–100 mL) of solutions con-

Effect of Contact Time on Cr Uptake

The retention of metals and metalloids by differentmicroorganisms in batch systems occurs in two steps:(a) A passive uptake takes place in the first stagewhich is very rapid; (b) the second step is active andslow and the metal is accumulated by the biomass.Under the conditions previously optimized, a series ofexperiments was carried out at different contact times.Figure 4 demonstrates that the sorption rate of Cr(III)appeared to go through two stages. In the first stage,retention was fast and about 88% of the total Cr(III)uptake by the biomass took place within 10 min. Inthe second stage, between 10 and 120 min, thebiosorption was quantitative and unaffected by con-tact time. We can consider that in a batch system,Cr(III) was completely retained by the biomass in thefirst step. A contact time of 10 min was chosen for fur-ther work.

On the other hand, Cr(VI) exhibited a little reten-tion with a maximum (retention: 20%) at 20 min. Neid-hart et al. (16) reported that yeast cells are able toaccumulate chromates, but the carrier mechanism forpenetration of the cell membrane is too slow. Thesefindings should explain why it is possible to separateboth species under the conditions of our work.

Capacity of Yeast Cells for Cr Retention

The biomass amount is another variable that mayaffect biosorption. The effect of biomass on Cr(III) andCr(VI) uptake was evaluated by using the same con-centration of the analyte (80 ng mL–1) and varying thebiomass between 50 and 750 mg. Figure 5 illustratesthat 750 mg is the minimum amount of biomass neces-sary to remove 98% of the Cr(III) from solution, while

Fig. 3. Effect of temperature on Cr uptake. Yeast, 500 mg; Acetic acid, 0.1 mol L–1; pH, 4; Cr concentration, 80 ng mL–1; Contact time, 15 min.

Fig. 4. Effect of contact time on Cr uptake. Yeast, 500 mg;Acetic acid, 0.1 mol L–1; pH, 4; Cr concentration, 80 ng mL–1;Temperature, room temperature.

Fig. 5. Effect of amount of biomass on Cr uptake.Acetic acid, 0.1 mol L–1; pH, 4; Cr concentration, 80 ng mL–1;Contact time, 10 min; Temperature, room temperature.

107


taining 5.0 mg of Cr(III) wereplaced in contact with 750 mg ofyeast. After following theprocedure described, the solidphase was dissolved in 5 mL of 1.0mol L–1 HCl for Cr(III) determina-tion. This experiment showed thatrecovery of Cr(III) wasindependent of the sample volumeup to 75 mL. The recovery wasevaluated by comparing the signalobtained from a suspension [con-taining 5.0 mg of Cr(III) in 5 mL]with that obtained after preconcen-tration of more diluted solutionscontaining the same amount of ana-lyte. A preconcentration factor of15, calculated as the ratio of initialto final volume, was achieved. It isimportant to note that the recoveryobtained is dependent on the com-plexity of the matrix. We observedlower preconcentration factorswhen tap water was processed.

Study of Potentially InterferingElements

The interference of differentconcomitants is, in general, attrib-uted to the strength of the interac-tion between the metal and the cellcomponents. The effect of severalions on Cr(III) retention was evalu-ated to assess the selectivity of themethod. All tests were carried outunder optimum operating condi-tions and following the general pro-cedure. The test samples analyzedcontained 80 ng mL–1 of Cr(III) and5 µg mL–1 of the potential interfer-ent (50 µg mL–1 for Ca and Mg), andthe results are the average of threereplicate measurements.

Figure 6 depicts the percentageof Cr(III) retention in the presenceof the potential interferent in theacid selected (acetic acid) and inoxalic acid. This study confirmsthat acetic acid is the best alterna-tive not only for the purpose of spe-

ciation but also from the point ofview of selectivity. Since no signifi-cant inhibitory effects wereobserved, it is plausible to concludethat there is no competition forbinding sites between the speciesevaluated and the analyte in themedium selected.

Analytical Performance

The detection limits were calcu-lated following the IUPAC rules onthe basis of the 3σ criterion for 10replicate measurements of theblank signal. For Cr(VI), it was cal-culated in the liquid phase. Thedetection limit for Cr(III) was calcu-lated in the solid phase after disso-lution of the biomass in 10 mL of1.0 mol L–1 HCl. The limit of detec-tion of Cr(III) was then correctedfor the preconcentration factor.The detection limits in each phasewere 3.0 ng mL–1 for Cr(VI) and 0.2ng mL–1 for Cr(III).

Fig. 6. Effect of foreign ions on Cr(III) retention. Yeast, 750 mg; Acetic acid, 0.1 mol L–1; pH, 4; Cr concentration, 80 ng mL–1;Contact time, 10 min; Temperature, room temperature.

108

The relative standard deviation(RSD) for 10 successive measure-ments of a sample containing a finalconcentration of 80 ng mL–1 ofboth, Cr(III) and Cr(VI), were 2.2%and 2.0%, respectively. Table IIsummarizes the analytical perfor-mance of the method.

Application to the SelectiveDetermination of Cr(III) andCr(VI) in Natural Waters

To demonstrate the capability ofSaccharomyces cerevisiae to selec-tively retain Cr(III), different kindsof waters containing both specieswere analyzed. The samples werefiltered through glass microfiberfilters (1.2 mm), acetic acid wasadded to a final concentration of0.1 mol L–1, and the pH wasadjusted to 4. An analytical curveplotted with mixed standard solu-tions containing 0–4.0 µg mL–1 ofCr(III) and Cr(VI) was used to quan-

tify Cr(III) and Cr(VI) in the sam-ples.

The results of the analysis andthe spike recovery data for Cr(III)and Cr(VI) in spiked water samplesare presented in Table III. Recover-ies between 89 and 108% wereobtained, which confirms the relia-bility of the proposed method forCr speciation analysis.

CONCLUSION

This study demonstrated the util-ity of a biological substrate for pre-concentration and speciation ofchromium. Cr(III) wasquantitatively retained and precon-centrated in the yeast cells, whileCr(VI) remained in the supernatant.This fact made possible the deter-mination of each Cr species in sepa-rate phases in contrast to otherstudies (19) where Cr(VI) was cal-culated by substraction. Acetic aciddemonstrated to be an appropriatereaction medium because in theexperimental conditions the fixedCr(III) uptake by Saccharomycescerevisiae was quantitative, selec-

tive, and unaffected by the pH. Nostrict control of the pH is necessarywhich is another advantage of themethod. A preconcentration factorof 15 was reached for Cr(III). Theyeast cells present no chemical haz-ards, and are inexpensive and easyto obtain. The method proposed issimple, sensitive, highly selective,and suitable for routine analysis.

ACKNOWLEDGMENTS

Financial support was providedby the Agencia Nacional de Promo-ción Científica y Tecnológicathrough Project PICT-2000 N° 13-08772. One of the authors (PS)wishes to acknowledge the Interna-tional Atomic Energy Agency(Research contract RC: 11549/R1)for partial financial support.

Received February 10, 2005.

REFERENCES

1. J.O. Nriagu, in Chromium in theNature and Human Environment,Ed. J.O. Nriagu and E. Nieboer,Wiley, New York, Vol. 20, pp 92(1988).

2. S.A. Katz and H. Salem, The biologi-cal and environmental chemistryof chromium, VCH, New York, pp2-4 (1994).

3. E.A. García and D.B. Gomis, Analyst122, 899 (1997).

4. J.M. Christensen, Sci. Total Environ.166, 89 (1995).

5. J. Lintschinger, K. Kalcher, W.Gössler, G. Kölbl, and M. Novic,Fresenius’ J. Anal. Chem. 351, 604(1995).

6. A.R. Timerbaev, O.P. Semenova, W.Buchberger, and G.K. Bonn,Fresenius’ J. Anal. Chem. 354, 414(1996).

7. M. Pantsar-Kallio and P.K.G. Manni-nen, Fresenius’ J. Anal. Chem. 355,716 (1996).

8. R. Nusko and K.G. Heumann, Anal.Chim. Acta 28, 283 (1994).

TABLE IIQuality Parameters for Cr(III) and Cr(VI) Speciation Analysis

Detection Limit (3σ) Precision (%)b

(ng mL–1) for 100 ng mL–1 Cr

Cr(III)a 0.2a 2.2

Cr(VI) 3.0 2.0

a Preconcentration factor: 15, b (n=10).

TABLE IIIRecovery of Mixtures With Different Concentration Ratios

of Cr(III) and Cr(VI) [Mean value ± standard deviation (n=3),

concentrations are expressed in ng mL–1]

Cr Added Cr Found Recovery (%)

Cr(III) Cr(VI) Cr(III) Cr(VI) Cr(III) Cr(VI)

Tap 50 50 48.9±2.5 53.3±2.8 97 107

Well 10 100 8.9±0.5 98.1±3.9 89 98

River 25 25 26.3±1.5 27.0±1.7 105 108

109


9. A. Chakir, J. Bessiere, K.E. Kacemi,and B. Marouf, J. Hazard. MaterialsB95, 29 (2002).

10. S. Dahbi, M. Azzi, and M. de laGuardia, Fresenius’ J. Anal. Chem.363, 404 (1999).

11. M. Sperling, S. Xu and B. Welz,Anal. Chem. 64, 3101 (1992).

12. Y. Madrid, C. Cabrera, T. Pérez-Corona, and C. Cámara, Anal.Chem. 67, 750 (1995).

13. T. Pérez-Corona, Y. Madrid, and C.Cámara, Anal. Chim. Acta 345,249 (1997).

14. T. Pérez-Corona, Y. Madrid-Albar-rán, C. Cámara, and E. Becerro,Spectrochim. Acta 53B, 321(1998).

15. P. Smichowski, J. Marrero,A. Ledesma, G. Polla, and D.A.Batistoni, J. Anal. At. Spectrom. 15,1493 (2000).

16. S. Neidhart, S. Herwald, Ch. Lipp-man, and B. Straka-Emdem, Frese-nius’ J. Anal. Chem. 337, 853(1990).

17. D. Kratochvil, P. Pimentel, and B.Volesky, Environ. Sci. Technol. 32,2693 (1998).

18. O. Muter, I. Lubinya, D. Millers,L. Grigorjeva, E. Ventinya, and A.Rapoport, Process Biochem. 38,123 (2002).

19. H. Bag, A. Rehber Türker, M. Lale,and A. Tunçeli, Talanta 51, 895(2000).

20. T. Pérez-Corona, Y. Madrid-Albarrán, and C. Cámara, Quim.Anal. 20, 29 (2001).

21. K.J. Blackwell, I. Singleton, and J.M.Tobin, Appl. Microbiol. Biotech-nol. 43, 579 (1995).

22. R. H. Crist, J. R. Martín, D. Carr, J.R. Watson, H. J. Clarke, and D. R.Crist, Environ. Sci. Technol. 28,1859 (1994).

23. R. H. Crist, K. Oberhoser, N. Shank,and M. Nguyen, Environ. Sci. Tech-nol. 15, 1212 (1981).

24. N. Goyal, S.C. Jain, and U.C.Banerjee, Adv. Environ. Research7, 311 (2003).

25. G.M. Gadd, C. White, and L. deRome, Biohydrometall. Sci. Tech-nol. Lett. 421 (1988).

26. M. Itoh, M. Yuasa, and T.Kobayashi, Plant Cell Physiol. 16,1167 (1975).

˘


Corresponding author.E-mail: [email protected]

Flow Injection Determination of Cd in Meat SamplesUsing a Continuous Lixiviation/Preconcentration System

Coupled to a Flame AASS. Cancela-Pérez and *M.C. Yebra-Biurrun

Department of Analytical Chemistry, Nutrition and Bromatology, Faculty of Chemistry,University of Santiago de Compostela, 15782-Santiago de Compostela, Spain

INTRODUCTION

Contamination of the environ-ment by metals is a problem thathas become of regional and evenglobal concern. High metal concen-trations occur within sites contami-nated by atmospheric emissionsfrom point source mining andsmelting operations, and from com-bustion of fossil fuels.

The determination of inorganicelements in foods and biologicalmaterials has become of consider-able interest because of the impor-tance or toxicity of many of theseelements to human health.Cadmium is an important environ-mental toxicant, having adversehealth effects following long-termchronic exposure. This metal canbe easily absorbed by plants andaccumulates in the tissues of herbi-vores, mostly in the liver and kid-ney. Since meat and meat productsform an important part of thehuman diet, cadmium can find itsway into the human food chain (1).Therefore, the monitoring ofendogenous and exogenous toxictrace elements, such as cadmium,in foods is an important aspect offood analysis.

Inductively coupled plasma(ICP) (2), inductively coupledplasma mass spectrometry (ICP-MS)(3-4), and differential pulse anodicstripping voltammetry (DPASV) (5)have been used for the determina-tion of cadmium in meat samples.Atomic absorption spectrometry[graphite furnace (GFAAS) or flameatomic absorption spectrometry

steps are usually not required, butthe method is subject to morechemical interferences and someform of matrix modification isrequired. Flame AAS is simple andinexpensive and therefore widelyused for the determination of metal-lic elements. Flame AAS, using air-acetylene, provides accurate andsensitive results for cadmium deter-mination. Nevertheless, the concen-tration of metallic ions in differenttypes of matrices (such as cadmiumin meat) is often below the detec-tion limits attainable even with themost sensitive detectors. Therefore,in order to determine trace levels oftrace metals using FAAS, it is neces-sary to enrich the sample. Evans etal. (2) used a chelating reagent solu-tion [1.0% (m/v) solution of APDC-DDDC] to form a metal chelatewith cadmium; this chelate wasthen extracted with chloroform.Falandysz (6) preconcentrated cad-mium by forming a complex withAPDC that was extracted withmethyl isobutyl ketone. GomesNeto et al. (7) proposed a minicol-umn packed with a strongly basicanion exchanger (AG-X8 resin) forthe spectrometric determination ofcadmium in foodstuffs, plants, andsimilar products based on the for-mation of a cadmium iodide-Mala-chite Green (MG) associate.

The analysis of solid foodstuffsusually requires a rather complicatedand time-consuming sample prepa-ration. Loss of test elements andsample contamination with theseelements may occur during thisstep, which deteriorates the relia-bility of the analysis. Moreover,sample preparation substantiallyadds to the duration and cost ofanalysis. Wet digestion is the pre-ferred sample preparation proce-

ABSTRACT

A simple flow injectionmethod is proposed, which com-bines continuous acid extraction,preconcentration, and flameatomic absorption spectrometryfor the determination ofcadmium in meat samples. Thedynamic acid extraction step wascarried out by using a continuousultrasound-assisted extractionsystem. The acid extract was pre-concentrated on-line on a mini-column packed with a chelatingresin (Chelite P, with amino-methylphosphoric acid groups),and elution was carried out withhydrochloric acid. Once eluted,cadmium was continuously moni-tored by flame atomic absorptionspectrometry. These steps wereoptimized applying an experi-mental design.

The method allowed a totalsampling frequency of 16 sam-ples per hour. Good precision ofthe procedure was obtained[2.9%, expressed as relative stan-dard deviation (RSD)], includinga high enrichment factor (20.5),and detection and quantificationlimits of 0.014 and 0.067 µg/g for60 mg of sample, respectively.The accuracy of the proposedmethod was verified using CRM183 Pig Kidney standard refer-ence material; the results were ingood agreement with the certi-fied values. The analytical proce-dure was applied to thedetermination of trace amountsof cadmium in real meat samples.

(FAAS)] has shown to meet most ofthe requirements for a reliable andfast determination of this element inmeat. GFAAS has greater sensitivitythan flame atomic absorption spec-trometry and preconcentration

111


dure for the determination of cad-mium in foods. The AnalyticalMethods Subcommitteerecommends use of a sulphuricacid-hydrogen peroxide wet diges-tion with the usual precautions (8).A number of methods for cadmiumdetermination in meat samplesbased on wet digestion have alsobeen proposed. However, the prin-cipal disadvantages of this methodare a very time-consuming process[i.e., overnight at room tempera-ture and 1 hour at 120ºC (9), 6.5 hat different temperatures (10)] andthe use of concentrated acids[nitric acid and hydrogen peroxide(11), a nitric, perchloric andsulphuric acid mixture (10)].Microwave-assisted digestion wascarried out by Crete et al. (11),Cubadda et al. (4), Zhou et al. (3),and Jorhem et al. (12). Despite theuse of microwave digestion, whichis fast and effective, the digestionoperation requires concentratedacids, which generate carcinogenicnitrous vapors as a consequence ofthe organic matrix destruction, sosafety control in the laboratory hasto be taken into account. If dry ash-ing is used, the temperature mustbe kept below 500ºC to preventvolatilization losses. This procedureis also very time-consuming [2 h at100ºC, 1 h at 150ºC, 1 h at 300ºC,and 15 h at 450ºC (13)].

The use of ultrasonic energy is ofgreat help in the pre-treatment ofsolid samples because it facilitatesand accelerates operations such asthe extraction of organic and inor-ganic compounds, slurry dispersion,homogenization, and other analyti-cal stages. It also eliminates totalsample matrix introduction into thenebulizer of the spectrometer,which occurs with solid sampling,slurry sampling, and the digestionprocedures (14–15).

Since its introduction, flow injec-tion analysis (FIA) has establisheditself as a very useful and versatiletechnique (16-17). Some of the

advantages are flexibility, reliability,reproducibility, ease of automation,high sample throughput, small sam-ple volume, and low reagent con-sumption and interferences.

In this paper, we describe a sim-ple and continuous ultrasound-assisted extraction system (CUES)connected to an on-line flow injec-tion manifold for acid leaching ofcadmium, on-line concentration ofthis trace metal from the acidextract using a chelating resin(Chelite P, with aminomethyl-phos-phoric acid groups) , and its deter-mination by flow injection-flameatomic absorption spectrometry(FI-FAAS). This procedurecombines the benefits of the ultra-sound-assisted extraction methods,the high sensitivity of the precon-centration techniques, and theadvantages of the FI system. Themethod was applied to the analysisof meat samples for the determina-tion of cadmium.

EXPERIMENTAL

Instrumentation

A PerkinElmer Model 5000atomic absorption spectrometerwas used (PerkinElmer Life andAnalytical Sciences, Shelton, CT,USA), equipped with a cadmiumhollow cathode lamp. The instru-ment was set at 228.8 nm. Thespectrometer output wasconnected to a PerkinElmer Model50 Servograph Recorder with arange of 5 mV. The FI system wasequipped with two Gilson®Minipuls™ 3 peristaltic pumps(Gilson, France) fitted with Viton®tubes, an ultrasonic bath (Selecta,Barcelona, Spain), six Rheodyne®injection or switching valves (USA),Models 5041 and 5301, a glass mini-column (100 mm x 3 mm i.d., bedvolume 700 µL) (Omnifit, UK). Theends of the minicolumn wereplugged with filter paper (Whatman541) and the Viton minicolumn(120 mm x 1 mm i.d.) was packedwith 50 mg chelating resin for the

preconcentration step. The ends ofthe minicolumn were plugged withglass wool.

Reagents

Ultrapure water of 18.2 MΩ.cmresistivity, obtained from a Milli-Q™water purification system(Millipore, Bedford, MA, USA), wasused for preparation of the reagentsand standards. Hydrochloric acid,nitric acid (Scharlau Chemie,Barcelona, Spain), and 1000 µg/mLcadmium standard solution (Merck,Darmstadt, Germany) were ofreagent grade. The chelating resinused was Chelite P, with amino-methylphosphoric acid groups(Serva Electrophoresis GmbH, Hei-delberg, Germany). The sample par-ticle size was 0.3–0.8 mm andammonium acetate (Merck) of 16mol/L in aqueous solution. The cer-tified reference material was CRM183 Pig Kidney (Community Bureauof Reference, Brussels, Belgium).

Sample Preparation and Procedure

The meat, liver, and kidney sam-ples were purchased at a localsupermarket. Visible fat, connectivetissue, and major blood vesselswere excised. The samples werecut into small pieces, dried at 50ºC,triturated and pulverized in a porce-lain mortar and, after sieving, frac-tions with a particle size less than30 µm were taken for analysis.

The continuous cadmium deter-mination system is shown in Fig-ure1. Meat samples of 50–60 mgwere weighed directly into theglass minicolumn. Then, the mini-column was connected to theCUES. First, the CUES circuit (2 mL)was loaded with the acid leachingsolution (3 M nitric acid). Once theCUES circuit was closed, the leach-ing solution circulated through theminicolumn under ultrasonicenergy action at a flow rate of 3.5mL/min for 2 min. The direction ofthe flow was changed each 30 s inorder to avoid sample accumulation

112

at the end of the minicolumn.Then, the switching valve (SV2)was switched to its opposite posi-tion and the acid extract washomogenized in the mixing coil.After this, the acid extract channelconverged with a buffer solutionstream and 16 M ammoniumacetate was added in order toobtain a pH value of >3. The mix-ture was homogenized in a secondmixing coil and then passedthrough the preconcentration mini-column at a flow rate of 2 mL/min.The cadmium was retained quanti-tatively by formation of a metalchelate, eluted by injection of 92 µLof 3 M hydrochloric acid into awater carrier stream, and swept tothe detector where cadmium wascontinuously monitored.

Standard solutions containingbetween 0.000 and 0.036 µg/mL ofCd in the same acid medium as theleaching solution were introducedinto the flow system.


In the proposed approach, theorder used for optimizing the stepsinvolved was as follows: (a) thevariables affecting the cadmiumpreconcentration step and (b) theleaching step.

Optimization of the Cd Precon-centration Step

The number of variables thataffect the cadmium preconcentra-tion step is very large: HCl concen-tration (eluent solution), elutionflow rate, sample pH to obtain aquantitative retention on the chelat-ing resin, flow rate of the sample,eluent volume, and minicolumndiameter. Plackett-Burman designsare principally used when the num-ber of variables that influence theanalytical system is very large,because such designs permit theuse of a full factorial design and thenumber of factor combinations canbe obtained in multiples of four (18).

In this paper, a Plackett-Burman26 (3/16) factorial type III resolu-

tion design with one center pointwas selected, allowing 6 degrees offreedom and involving 13 non-ran-domized runs. This factorial designwas applied to a standard solutioncontaining 0.04 µg/mL of cadmium.One mL of this solution was passedthrough the preconcentration mini-column. The lower and upper lev-els for each studied variable arelisted in Table I. These values werechosen from available data reportedin previous experiments.

From the results of these analyti-cal data, it can be concluded thatthe cadmium preconcentration-elution process appeared to beaffected positively by one statisti-cally significant factor, namely thesample pH. This means that thecadmium preconcentration-elution

Fig. 1. Flow injection manifold for the whole procedure (CUES and preconcentration device) for cadmium determination inmeat samples. P1 and P2, peristaltic pumps; LS, leaching solution; W, waste; UB, ultrasonic bath; MS, minicolumn containingthe sample; IV, injection valve; SV1-SV5, switching valves; MC1 and MC2, mixing coils; MN, minicolumn containing the chelat-ing resin (Chelite P) and FAAS, flame atomic absorption spectrometer.

TABLE IFactor Levels Studied in the Plackett-Burman Design (26 3/16)

for the Cd Preconcentration-Elution Process

Variable Key Low High Optimum

Sample pH A 1 7 3-8Sample flow rate (mL/min) B 0.5 4.0 2.0Eluent concentration (HCl, M) C 0.1 3.0 3.0Eluent volume (µL) D 70.4 190 92Elution flow rate (mL/min) E 3.0 5.0 3.0

Minicolumn diameter (mm) F 1.0 2.0 1.0

113


efficiency is directly proportionalto the pH and best responses (cad-mium recovery, %) are obtained atthe highest tested level of this vari-able.

All other parameters were notstatistically influential. The diame-ter of the minicolumn, eluent con-centration, and eluent volume wereaffected by a positive estimatedeffect, while the sample flow rateand eluent flow rate were affectedby a negative estimated effect.

In order to improve the analyti-cal characteristics of the method,other experiments outside theframework of the design were car-ried out. Table II shows the resultsof these experiments. The minicol-umn diameter is affected by thelower positive estimated effect. Inorder to avoid a system connectionfit, a minicolumn with a diameterof 1 mm was chosen as optimum.

The eluent volume is the vari-able with an upper non-significantestimated effect and was affectedby a positive sign. Eluent volumesof 70.4, 92, and 190 µL were stud-ied to obtain a greater concentra-tion factor. An eluent volume of70.4 µL was not enough to obtain aquantitative recovery (65%), but 92µL provided the complete elutionof cadmium from the minicolumn;and 92 µL was selected as the opti-mum eluent volume.

Eluent and sample flow rateswere affected by a negativeestimated effect. In order toachieve the greatest sampling fre-quency, the upper values of theflow rates were studied. The eluentflow rate was studied between 3–4mL/min. A flow rate of 4 mL/minprovided a 44.3% recovery, whilethe lower value studied for this vari-able (3 mL/min) was chosen as theoptimum value because a quantita-tive elution was obtained. On theother hand, a sample flow ratebetween 2–3 mL/min was also stud-ied. It was found that a flow rate of

3 mL/min provides a 56% recovery,while a flow rate of 2 mL/min gavea quantitative cadmium recovery of>95%; and this value was selectedas the optimum for cadmium reten-tion.

In order to reduce the hydro-chloric acid concentration (eluent),one experiment was performedusing a lower concentration (2 M).The recovery obtained was notquantitative, so 3 M hydrochloricacid concentration was selected.

Finally, a few experiments werecarried out with the aim of estab-lishing the pH range for quantita-tive retention of cadmium. Theresults of these experiments arelisted in Table III. Cadmium wasretained quantitatively in the pHrange of 3–8. This is a very impor-tant factor because the sample inthe acid extract has a very low pHvalue (3 M nitric acid medium). Anammonium acetate buffer solutionwas proposed in order to achieve

the optimum pH for cadmiumretention. Since quantitative extrac-tion of Cd in meat samples needsan acid medium and the pH opti-mum for Cd retention on thechelating resin ranges from 3–8, anammonium acetate buffer solutionwas proposed. In order to carry outthis pH increase on-line, the con-centration, volume, and flow rate ofthe ammonium acetate buffer solu-tion were studied. A 0.61-g amountof ammonium acetate was requiredto increase the pH from 3–5 for 2–5mL of acid extract.

The buffer solution concentra-tion was studied between 2.65–16M. These concentrations implied abuffer solution volume rangingfrom 0.5–3.0 mL and a buffer solu-tion flow rate from 0.4–1.0mL/min. The principal aim was toachieve a greater samplingfrequency, and the sum of bothflow rates (sample flow rate andbuffer solution flow rate) could not

TABLE IIExperiments Outside the Framework of the

Design to Fine-tune the Preconcentration-Elution Process

Cd Recovery

A B C D E F (%)

7 0.5 3.0 190 3 1 99.77 0.5 3.0 70.4 3 1 65.07 0.5 3.0 92 3 1 103.47 0.5 3.0 92 4 1 44.37 2.0 3.0 92 3 1 96.17 3.0 3.0 92 3 1 56

7 2.0 2.0 92 3 1 74.9

TABLE IIIStudy of pH Range for Quantitative Retention of Cd

CdRecovery

A B C D E F (%)

8 2.0 3.0 92 3 1 106.19 2.0 3.0 92 3 1 46.23 2.0 3.0 92 3 1 98.6

2 2.0 3.0 92 3 1 72

114

exceed a total flow rate of 2mL/min (maximum flow rate toobtain a quantitative cadmiumretention). A 16 M concentration,0.5 mL buffer solution, and 0.4mL/min flow rate were selected asa compromise for the ammoniumacetate channel. The optimum con-ditions for the cadmium preconcen-tration-elution are listed in Table I.

Optimization of ContinuousUltrasonic Acid Extraction of Cd

In this case, the same experimen-tal design was used: a Plackett-Bur-man 26 (3/16) factorial type IIIresolution design. This factorialdesign was applied to 60 mg of rab-bit liver with a cadmium concentra-tion of 0.165 µg/g. The sampleconcentration was obtained usingan off-line acid digestion procedure.Six experimental variables wereoptimized: nitric acid andhydrochloric acid concentrations(leaching solution), sonicationtime, leaching temperature, flowrate of the CUES, and leaching vol-ume. To optimize the CUES, cad-mium was measured on-line byFAAS with a flow injection systemthat involves the preconcentrationstep previously described. Table IVsummarizes the lower and upperlevels for each studied variable.

The analytical data show thatcadmium extraction appeared to beaffected positively by two statisti-cally significant factors: nitric acidconcentration and sonication time.The flow rate of the CUES,hydrochloric acid concentration,and leaching volume were affectedby a negative estimated effect,which suggests that they inhibit theextraction at the highest levelstested. Thus, a 3.5 mL/min flowrate, 0 M hydrochloric acid concen-tration, and 2 mL volume wereselected as the optimum values.

The sonication temperature wasaffected by a positive estimatedeffect but was not a statistically sig-nificant factor. In order to simplify

the determination, a room tempera-ture of 20ºC was selected as theoptimum temperature for cadmiumextraction. In this case, only threeexperiments outside of the frame-work of this study were carried outand the results are listed in Table V.The aim of this study was toincrease the sample frequency andto reduce the nitric acid concentra-tion. The sonication time was stud-ied between 2.0–1.5 min. Since2 min was enough to extract cad-mium quantitatively, but 1.5 minprovided an 82.3% recovery, 2 minwas selected as the optimum soni-cation time.

Finally, the possible reduction ofnitric acid concentration was stud-ied. A 1.5 M nitric acid concentra-tion was tested but was found notto be enough to obtain a quantita-tive cadmium extraction. Bestresults were obtained with 3 MHNO3 and this value was chosen asthe leaching solution. The optimumconditions for continuous ultra-sonic acid extraction are listed inTable IV.

Two sample variables that canaffect the acid extraction processwere studied, namely, meat parti-cle size and amount of sample.Using the optimum CUES and pre-concentration conditions, particlesizes between 30–100 µm weretested. The results obtained indi-cate that this variable does notaffect the extraction process. Sinceoptimum results were obtainedwhen a sample amount of 60 mgwas used to optimize the CUES sys-tem, we supposed that smallerquantities would also be suitable.We also studied the maximumamount of sample that is quantita-tively leached in order to increasethe sensitivity of the method andselected 80 mg. The results demon-strated that sample amounts greaterthan 60 mg produce high pressurein the CUES and cause loss of sam-ple. Therefore, the maximum sam-ple amount that can be used is 60mg.

TABLE IVFactor Levels Studied in the Plackett-Burman Design (26 3/16) for the

Continuous Ultrasonic Acid Extraction of Cd From Meat Samples

Variable Key Low High Optimum

HNO3 concentration (M) G 0 3 3HCl concentration (M) H 0 3 0Sonication time (min) I 0.5 5.0 2Leaching volume (mL) J 2.0 5.0 2Leaching temperature (ºC) K 20 70 20

Flow rate of the CUES (mL/min) L 3.5 6.0 3.5

TABLE VExperiments Outside the Framework of the Design to Fine-tune the

Continuous Ultrasonic Acid Extraction Process

CdG H I J K L Recovery (%)

3 0 2 2 20 3.5 99.7

3 0 1.5 2 20 3.5 82.3

1.5 0 2 2 20 3.5 70.8

115


Features of the Method

The calibration graph was run(n=7) under optimum chemical andflow conditions for the globalprocess, the equation wasabsorbance = 7.8 x10–4 + 3.12 X(r = 0.999), where X is the Cd con-centration expressed as µg/mL. Thelinear ranged from 0.067–3.243µg/g.

The precision of the preconcen-tration step was verified using 1 mLCd standard solution containing0.024 µg/mL Cd (n = 11). The pre-cision for the global processobtained for the real samples waschecked on a rabbit liver samplecontaining 0.165 µg/g Cd (n = 11).The results obtained and expressedas the relative standard deviationwere 1.9 and 2.9%, respectively.The limit of detection (LOD), basedon three times the standard devia-tion (n = 30) of the blank, wasfound to be 0.014 µg/g. The quan-tification limit (LOQ) based on 10times the standard deviation (n =30) of the blank was found to be0.067 µg/g. The validation of themethod was performed using a cer-tified reference material CRM183Pig Kidney, with a cadmium con-tent of 2.71 ± 0.15 µg/g. The cad-mium content obtained for thecertified reference material by theproposed method (mean ± SD,n = 3) was 2.76 ± 0.05 µg/g, whichagrees with the certified value.

The concentration factor of themethod based on the rate betweenthe direct calibration graph and FIAcalibration graph was found to be20.5. The resin capacity based onthe maximum amount of cadmiumthat was able to retain the chelatingresin was 7.07 µmol Cd/g resin.The sample throughput using theproposed method was about 16samples/h.

Analysis of the Samples

The method was applied to thedetermination of cadmium in vari-ous meat samples. The concentra-

tion of cadmium in these samplesranged from 0.096–0.417 µg/g. Theresults obtained with the proposedmethod were compared with thoseachieved by a conventional off-linesample digestion method with con-centrated nitric acid, a preconcen-tration step previously optimizedby using the chelating resin CheliteP, and determination by FAAS. Tocompare the results obtained byboth methods, the paired t-test wasapplied (19). As shown in Table VI,both methods do not give signifi-cantly different values, thus theagreement between the two meth-ods is satisfactory.

CONCLUSION

The continuous ultrasound-assisted extraction system (CUES)combined with a preconcentrationstep and coupled with a FI mani-fold has demonstrated to be a rapid,precise, and accurate sample pre-treatment procedure for the acidleaching of cadmium, preconcen-tration of this trace metal, and itsdetermination by FAAS in solid sam-ples. The main goals obtained withthe method proposed are a reduc-tion of sample contamination aswell as analyte loss. This is becauseless manipulation of the sample isrequired, reduced sample andreagent amounts are needed, and areduction in the sample prepara-

tion time is achieved which signifi-cantly increases sample throughput.

This procedure combines thebenefits of ultrasound-assistedextraction with the high sensitivityprovided by the preconcentrationstep and the inherent advantages ofthe FI systems.

ACKNOWLEDGMENTS

The authors are grateful for thefinancial support provided by theInstituto Nacional de Investigacióny Tecnología Agraria y Alimentaria(INIA). Ministerio de Ciencia yTecnología (Project No. CAL01–043).S.C. acknowledges a grant from theXunta de Galicia.

Received July 14, 2004.

TABLE VIDetermination of Cd in Meat Samples and Paired t-Test

Sample [Cd]A (µg/g) [Cd]B (µg/g) Recovery (%)

Chicken 0.134 0.135 99.3Turkey 0.157 0.157 100Pig 0.133 0.130 102.3Calf 0.101 0.096 105.2Mutton 0.107 0.106 100.9Rabbit liver 0.165 0.168 98.2

Mutton kidney 0.408 0.417 97.8

A: Off-line acid digestion, preconcentration with a chelating resin (Chelite P) and FAAS Cd determination.

B: Cd concentration obtained by the present method.Experimental value of t = 0.33; Critical value of (tn–1=6, P=0.05) = 2.45

116

REFERENCES

1. J.M. Llobet, G. Falco, C. Casas, A.Teixido, and J.L. Domingo, J.Agric. Food Chem. 5, 838 (2003).

2. W.H. Evans and D. Dellar, Analyst197, 977 (1982).

3. H. Zhou and J. Liu, At. Spectrosc. 18,115 (1997).

4. F. Cubadda, A. Raggi, A. Testoni, andF. Zanasi, J. AOAC Int. 85, 113(2002).

5. J. Kumpulainen, M. Paakki, and R.Tahvonen, Fresenius’ J. Anal.Chem. 338, 423 (1990).

6. J. Falandysz, Food Addit. Contam. 8,71 (1991).

7. J.A. Gomes Neto, H. Bergamin Filho,R.P. Sartini, and E.A.G. Zagatto,Anal. Chim. Acta. 306, 343 (1995).

8. Analytical Methods Subcommittee,Analyst 100, 761 (1975).

9. M. Lopez Alonso, J.L. Benedito, M.Miranda, C. Castillo, J. Hernández,and R.F. Shore, Food Addit. Con-tam. 17, 447(2000).

10. C.D.C. Salisbury and W. Chan, J.AOAC Int. 74, 587 (1991).

11. M. Crete, F. Potvin, P. Walsh, J.L.Benedetti, M.A. Lefebvre, J.P.Weber, G.Paillard, and J. Gagnon,Sci. Total Environ. 66, 45 (1987).

12. L. Jorhem and J. Engman, J. AOACInt. 83, 1189 (2000).

13. N. Ybañez, R. Montoro, and A.Bueso, Rev. Agroquim. Tecnol.Aliment. 23, 510 (1983).

14. A. Moreno-Cid, M.C. Yebra, S. Cancela, and R.M. Cespón, Anal.Bioanal. Chem. 377, 730 (2003).

15. M.C. Yebra, A. Moreno-Cid, R.M.Cespón, and S. Cancela, At. Spec-trosc. 24, 31(2003).

16. K.K. Stewart, G.R. Beecher, and P.E.Hare, Fed. Proc. 33, 1439 (1974).

17. J. Ruzicka and E.H. Hansen, Anal.Chim. Acta. 78, 145 (1975).

18. S.N. Deming and S.L. Morgan,Experimental Design: A Chemo-metric Approach, Elsevier, Amster-dam, The Netherlands (1994).

19. J.C. Miller and J.N. Miller, Statisticsfor Analytical Chemistry, Ellis Hor-wood, Chichester, UK (1994).


*Corresponding author.E-mail: [email protected]: + 58-261-759 81 54Fax: + 58-261-759 81 25

Medical Application of Fast Furnace Program Used in the ETA-AAS Determination of Cu and Zn

in Blood Plasma of Children With Down SyndromeDenny R. Fernándeza, Aracelis del C. Vásqueza, Maigualida Hernándeza, Ana M. Ocandoa,

José G. Manzanillaa, Marisol Sotob, Francisco Alvarezb, and *Víctor A. Granadilloa

a Laboratorio de Instrumentación Analítica, Departamento de Química, Facultad Experimental de Ciencias, Maracaibo 4011, Venezuela

b Unidad de Genética, Facultad de Medicina, Universidad del Zulia, Maracaibo 4011, Venezuela

INTRODUCTION

Copper and zinc are essentialelements for the human organismand are indispensable componentsof a great number of enzymes andproteins of the body’s systems.These metals are obtained throughthe daily intake of foods (1) and aninadequate consumption affects thebody’s immune system (2).

Copper is found in various tis-sues and organs of the humanbody, where the liver, brain, andbones contain the highest Cu con-centrations. High levels ofplasmatic Cu are responsible forgastrointestinal malfunctions,anorexia, infantile malnutrition,nausea, frequent vomiting, dysfunc-tion of the central nervous system,and the well-known copper fever(3). Low levels of Cu can cause cer-tain types of anemia since cerulo-plasmine (an enzyme associatedwith Cu) diminishes the metal’scapacity to promote hematopoiesis(3) and affects mental development(4).

Zinc has catalytic and regulatoryfunctions in many biological sys-tems (4), and is absorbed from theduodenum and the jejunum like aprotein complex in the mucous tis-sues. Zinc is a component of theproteins that participate in the for-mation of collagen and in the pro-duction of insulin, is required forthe duplication and development ofcells, responsible for sexual matura-

tion (4,5), and involved in therepair and transcription of deoxyri-bonucleic acid (6). A deficiency inZn manifests itself in poor growthdevelopment, hypogonadysm, lackof appetite, and alteration in thescarring of wounds. A lack of Znintake is also related to illnessesand vesicular lesions of the skin andcauses breathing infections (6,7).

Down syndrome is a veryfrequent chromosomal anomaly,occurring in about one in every 700births (3). It is widespread aroundthe world, and its congenital dys-functions are characterized by dif-ferent degrees of mental retardationand multiple physical defects (7). Inchildren with Down syndrome, theimmune system is depressed(2,6,7), and causes breathing infec-tions, where high levels of the thy-rotropin regulatory hormone of thethyroid (TRH) are observed (6,7).

Since Cu and Zn are present attrace levels in different types ofclinical samples, it is important touse an analytical instrumental tech-nique reliable and modern to deter-mine these elements with adequateaccuracy and precision. Numerousanalytical methods for Cu and Zndetermination have been developeddue to the clinical relevance ofthese metals (4,8–12). Thetechniques most widely used forthe determination of Cu and Zn areinductively coupled plasma massspectrometry (ICP-MS) (8), anodicstripping voltametry (ASV) (9),flame atomic absorption spectrome-try (FAAS) (4), and electrothermalatomization atomic absorptionspectrometry (ETA-AAS) (10–12).

ABSTRACT

The medical application of afast furnace program is presentedusing ETA-AAS-based methods forthe determination of copper andzinc in blood plasma of childrenwith Down syndrome. The ana-lytical parameters were evaluatedto establish the reliability andreproducibility of the ETA-AASmethod for clinical purposes.The accuracy was verified byanalyzing two certified materials,comparing the target and experi-mental values for each metal, andobtaining mean relative errors of0.63% for Cu and 1.50% for Zn.Additionally, the recovery studieswere carried out, with recoveriesof 100.6% and 100.2% for Cu andZn, respectively.

The average precision(expressed as RSD) was 0.97% forCu and 1.21% for Zn. The detec-tion limit (3 σ) was 0.1 µg/L forboth analytes. The characteristicmass was 5.3 pg for Cu and 0.9pg for Zn using a 20-µL sampleinjection volume. The metalsconcentrations were establishedfor 70 blood plasma samplesfrom children with Down syn-drome and from controls (35samples each), with concentra-tion ranges from 366–1616 µg/LCu and 1183–2660 µg/L Zn forthe experimental group, and500–1500 µg/L Cu and 683–3817µg/L Zn for the control group.

The ETA-AAS-based methodsdescribed for Cu and Zn repre-sent an analytical and clinical toolfor the evaluation of this kind ofgenetic anomaly, and permit theaccurate, precise, and interfer-ence-free determination of cop-per and zinc in clinical samples.

118

The ETA-AAS technique is mostwidely used for the determinationof Cu and Zn in biological, environ-mental, clinical, and food samples.It allows the trace determination ofthese two metals and provides lowdetection limits (i.e., in µg/L). Thistechnique favors the determinationof these metals in blood plasma attrace levels due to its high sensitiv-ity, excellent selectivity, and mini-mum requirements in themanipulation and treatment of thesamples, thus decreasing contami-nation problems (12). ETA-AAS alsorequires small sample volumes,which is a very important aspectfor this type of clinical analysis. Inaddition, the graphite furnace isable to reach high and controlledtemperatures, allowing the efficientelimination of the concomitantwithout significant loss of the ana-lytes under study. Another advan-tage of this technique is the use ofparameters that maximize instru-mental performance such as theability to use an analyticalisoformer, the use of air or anothergas (as internal alternate gas) duringthe pyrolysis step for the removalof carbon residues, the addition ofsurfactants, and the use of normalor pyrolytically coated graphitetubes with or without L’vov plat-form. The main difficulties with thistechnique are volatilization prob-lems of the analyte and spectralinterferences. However, these typesof interferences are corrected byusing the background corrector,which allows the correction of highbackground absorbance values gen-erated by the matrix. The volatiliza-tion problems are frequentlycorrected by the addition of an ana-lytical isoformer (13). Despite theabove-mentioned analytical advan-tages, sequential multielementaldeterminations using ETA-AAS arevery time-consuming. Therefore, itis necessary to develop analyticalmethods that use fast temperatureprograms of ≤45 s in comparison tothe conventional temperature pro-grams of ≥90 s required for the

complete dry-pyrolyze-atomizecycles in the graphite furnace.

The aim of this work is to pre-sent two fast furnace programs forthe ETA-AAS determination of cop-per and zinc in blood plasma ofchildren with Down syndrome andof a control group (healthy infants).These analytical methods areemployed like reliable and repro-ducible tools for the clinical evalua-tion of this kind of geneticanomaly.

EXPERIMENTAL

Instrumentation

For the analytical determinationsof Cu and Zn, a PerkinElmer Model2380 atomic absorption spectrome-ter was used, equipped with a deu-terium arc continuum backgroundcorrection system and coupled to aPerkinElmer Model HGA®-500graphite furnace (PerkinElmer Lifeand Analytical Sciences, Shelton,CT, USA). A PerkinElmer Model AS-40 autosampler was also employed.Wall atomization on the pyrolyti-cally coated graphite tubes wasused for all experiments. Hollowcathode lamps were individually

used as the primary radiationsources for Cu and Zn. Table Ishows the instrumental operatingparameters and temperature pro-grams used for the determination ofCu and Zn in the blood plasma sam-ples.

Reagents and StandardSolutions

All chemical reagents used forthis study were of analytical grade.The concentrated solutions of cop-per and zinc (~1000 mg/L) wereprepared from copper metal pow-der (Fisher Scientific, Pittsburgh,PA, USA) and using a commercialconcentrated solution of Zn (Riedelof Haën, Hannover, Germany).

The standard solutions for thecalibration curves were performeddaily by direct dilution of the con-centrated solution of each analytein 0.01 M nitric acid (Merck, Darm-stadt, Germany). For both metals,the concentrations of the aqueousstandard solutions for the workingcurves were 5, 10, 15, 20, 25, and30 µg/L. These standards were pre-pared from intermediate solutionsof 100 mg/L Cu and 100 mg/L Zn.

TABLE IOperational Parameters and Fast Temperature Programs Used in the

Graphite Furnace for the Determination of Cu and Zn by ETA-AAS

Wavelength 324.8a/213.9b nm Inert gas ArgonSlit width 0.7 nm Injection temp. 20°CLamp current 10.0 mAc Graphite tube Pyrolytically coated

Sample volume 20 µL Atomization type Wall

Temperature ProgramStep Dry 1 Dry 2 Pyrolysis Atomization Clean-out

Temp. (ºC) 80 200 900a/1000b 1800a/2200b 2700Ramp (s) 1 10 5 0 1Hold (s) 5 1 5 5 2Read 0Rec. –5Internal flow rate

Ar (mL/min) 300 300 300 0 300

a For Cu. b For Zn. c For both metals using hollow monocathode lamps for each analyte.

119


Triton® X-100 0.01% (v/v)(Riedel of Haën) was used to dilutethe blood plasma. All of the solu-tions were prepared with watergrade I ASTM (14) in polypropylenecalibrated flasks (Nalgene Labware,Nalge, Rochester, NY, USA). Theconcentrated nitric acid (Merck)contained concentrations of Cu andZn which are not detectable byETA-AAS.

Evaluation of the accuracy of themethods was carried out in theanalysis of certified reference mate-rial HPS 590321 Trace Metals inDrinking Water (from High-PurityStandards, HPS, Charleston, SC,USA) and a standard referencematerial NIST 1566a Oyster Tissue(from the National Institute of Stan-dards and Technology, NIST,Gaithersburg, MD, USA).

Population Under Study

This study was carried out for agroup of 70 children: 35 childrenwith Down syndrome and 35healthy children (the controlgroup), ranging in age from 6 monthsto 6 years. Informed consent wasobtained from the parents of bothgroups studied. The clinical proto-col was approved by an ad hoccommittee. These childrenbelonged to poor families locatedin the IV and V socio-economic stra-tum according to the modified Graf-far’s method for Venezuela (15).

Collection of Blood Plasma Samples

The blood plasma samples werecollected in the Unit of Genetics,University Hospital of MaracaiboCity, Venezuela. Approximately 5mL of whole blood was collectedfrom each individual and placedinto polypropylene tubes, whichcontained sodium heparin as theanticoagulant (i.e., 50 µL for eachmL of blood sample). The bloodsamples were centrifuged at 15,000rpm for 10 minutes. Supernatantswere used as the plasma samples.The obtained plasma was placed in

another clean polypropylene tube,and stored at 4ºC to avoid decom-position and bacterial growth. Priorto the spectrometric determinationof Cu and Zn, the plasma sampleswere diluted 100-fold with TritonX-100 0.01% (v/v); the certifiedmaterials were diluted 10-fold with0.01 M nitric acid. Each test portionwas prepared in triplicate and ana-lyzed five times in order to obtainthe mean, standard deviation, andrelative standard deviation. Statisti-cal analyses were carried out byconventional methods using com-mercial statistical programs (i.e.,Origin 6.0, Excel® 2000, etc.). Dif-ferences were considered statisti-cally significant when p was <0.05.


The methodologies used for theETA-AAS determination of Cu andZn in blood plasma of children withDown syndrome were based on thephysicochemical characteristics ofthese elements [i.e., melting (mp)and boiling (bp) points of theatomic precursor of the elementconsidered]. These elements areamong the intermediate metals andare neither volatile nor refractory,which suggests that atomization ofboth occurs via an oxide (13).Thus, Cu and Zn have metal oxideswith a relatively high mp and bp[Cu: mp of 1235°C and bp of1800°C; and Zn: mp of 1975°C andbp of >2000°C (16)], which makesthem very stable at the pyrolysistemperatures considered. Based onthese physicochemical characteris-tics, it can be concluded that (a)the use of STPF conditions for thedetermination of Cu and Zn byETA-AAS is unnecessary, and (b)moderate pyrolysis temperatures(~1000°C) can be applied in thegraphite furnace for the efficientremoval of the concomitants with-out significant volatilization loss ofCu and Zn.

Consequently, it is possible todevelop fast temperature programs(FTP) for the determination of Cu

and Zn by ETA-AAS without usingSTPF conditions. Thus, the interfer-ence effects of the background onthe analytical signals (spectral inter-ferences), were minimized using aFTP for the graphite furnace.

The fast temperature programswere optimized for the ETA-AASdetermination of Cu and Zn usingmoderate temperatures and shortramp and hold times in the pyroly-sis step.

Selection of Graphite FurnacePrograms

The temperature programs forthe graphite furnace wereoptimized using aqueous standardsolutions of Cu and Zn (~5.00 µg/Lof each, equivalent to 100 pg of Cuor Zn); a 10-fold diluted Trace Met-als in Drinking Water sample,which has Cu and Zn masses of 41and 1394 pg, respectively; and a100-fold diluted real sample ofblood plasma (164 pg of Cu and192 pg of Zn). The temperatures ofthe pyrolysis and atomization stepswere optimized by increasing thetemperature and taking theintegrated absorbance from thediluted real blood plasma sampleswith a known mass of Cu and Zn[as previously reported for ETA-AASmethodologies (17)], and observingthe maximum values for the inte-grated absorbance (Figure 1). Fromthe analysis of the pyrolysis andatomization curves, two interestinginferences were observed: (a) theoptimum pyrolysis temperaturesthat guarantied the efficientremoval of the concomitant of inor-ganic and organic origin in the eval-uated matrix were 900°C and1000°C for Cu and Zn, respectively;and (b) the optimum atomizationtemperatures for Cu and Zn were1800°C and 2200°C, respectively(Figure 1). These temperatureswere necessary to ensure goodelimination of the matrix and alsoto achieve high sensitivity, reliabil-ity, and reproducibility. Moreover,the results demonstrated that with

120

Fig. 1. Pyrolysis (A1 and B1) and atomization (A2 and B2) temperature optimiza-tion curves for Cu (A) and Zn (B) determinations by ETA-AAS, where: - - aqueousstandard (100 pg of Cu or Zn), -- diluted Trace Metals in Drinking Water HPSCRM 590321 (41 pg of Cu and 1394 pg of Zn), and -- real blood plasma sample(164 pg of Cu and 192 pg of Zn).

these temperatures, the determina-tion of these metals can be carriedout without loss due to volatiliza-tion, and thus obtain excellentaccuracy and precision for the realsamples and the certified materials.

There is controversy regardingthe pyrolysis and atomization tem-peratures used in the graphite fur-nace for the determination of Cuand Zn in different types of matri-ces. Some authors report highpyrolysis temperatures rangingfrom 1200–1300°C for Zn(10,11,18); and for copper, lowtemperatures ranging from 700–1000°C (11) or between 300 and800°C (19). Other authors reportatomization temperatures rangingfrom 2100–2300°C for Cu and1900–2300°C for Zn (10,11,18).These authors used short ramp andhold times (~10–25 s) in the drystep, ~40 s in the pyrolysis step,and a few seconds (~3–5 s) for theatomization step. These differencesin the pyrolysis and atomizationtemperatures are related principallyto the type of sample matrix used(i.e., organic or inorganic).

Evaluation of Analytical Parameters

The accuracy of the spectromet-ric methods proposed for the deter-mination of Cu and Zn was verifiedanalyzing two certified materials:HPS 590321 Trace Metals in Drink-ing Water and NIST 1566a OysterTissue. These materials were previ-ously mineralized by using areported method (17) to provideaqueous solutions whose final Cuand Zn concentrations were withinthe range of the metal concentra-tions expected in the clinical sam-ples evaluated. The results obtainedin this study are presented in TableII, which shows a similaritybetween the target and the experi-mental values. No significant statis-tical differences (p <0.05) wereobserved. This parameter was alsoevaluated by performing recoverystudies of each analyte added to the

121


TABLE IIAccuracy Studies for the ETA-AAS Determination

of Cu and Zn in Certified Materials

Concentration of Cu and Zna

(mean ± SD, µg/L)

CertifiedMaterials Target Experimental Relative Error (%)

CRM 590321b 20.00 ± 0.10 20.01 ± 0.60 0.05

(70.00 ± 0.50) (72.01 ± 1.70) (2.85)

SRM 1566ac 66.30 ± 4.30 67.10 ± 3.60 1.20

(830.30 ± 57.00) (831.50 ± 59.00) (0.14)

a Values in parenthesis are for zinc.b Trace Metals in Drinking Water from the High-Purity Standards (HPS, USA).c Oyster Tissue from the National Institute of Standards and Technology (NIST, USA).

TABLE IIIWithin- and Between-run Precision Studies for the Determination

of Cu and Zn by ETA-AASa

Meanb Within-runc Between-runsd

Sample (µg/L) SD (µg/L) RSD (%) SD (µg/L) RSD (%)

Plasma1 1.60 0.01 0.60 0.02 1.20

(1.40) (0.01) (0.70) (0.02) (0.05)2 11.20 0.05 0.44 0.16 1.40

(1.02) (0.03) (3.00) (0.01) (0.98)3 12.68 0.08 0.59 0.17 0.01

(1.62) (0.04) (2.50) (0.02) (0.94)4 32.26 0.39 1.22 0.01 0.01

(2.03) (0.02) (0.75) (0.18) (1.90)Drinking water 2.00 0.01 0.50 0.01 0.05

(7.10) (0.02) (0.28) (0.14) (1.97)

a Values in parenthesis are for zinc.b Concentrations of Cu and Zn in the diluted test portion of real sample.c Aliquot samples prepared by triplicate, five runs each.d Triplicate samples per analysis; five runs each and analyzed over a five-day period.

blood plasma samples, resulting inaverage recoveries of 100.6% and100.2% for Cu and Zn, respectively,and verifies appropriate accuracyfor the described methods.

The within- and between-runprecisions of the developed meth-ods were evaluated for four realblood plasma samples diluted 100-fold (Table III). Three aliquots ofeach real sample were analyzed(five runs each) using the instru-mental and operating conditionspreviously listed. The obtainedresults show average RSDs of 0.60%(Cu) and 1.08% (Zn) for the within-run and 1.34% (Cu) and 1.33% (Zn)for the between-run precisions.These results can be consideredadequate for these types of analy-ses.

A spectral interference studywas performed by carrying out abackground study in which the per-formance of the continuum sourcebackground corrector was evalu-ated to ensure its efficacy in theappropriate compensation of thespectral interferences. This proce-dure was described in detail else-where (4,13). Briefly, comparisonsbetween background-correctedintegrated absorbances and thoseobtained by subtracting the back-ground signals from theuncorrected absorbances wereestablished. Mean relative errors of1.87% and 1.16% were found for Cuand Zn, respectively. These factsindicate that the deuterium arc waseffective in compensating for thespectral interferences.

Non-spectral interference studieswere carried out by comparing theslopes of the working curves withthose obtained by the method ofstandard additions. For ETA-AASdeterminations of Cu in bloodplasma, the equations obtained forthe standard addition and calibra-tion curves were: A = 0.0151c + 0.0216, r = 0.9998(p <0.001); and A = 0.0146c – 0.0013, r = 0.9999

122

(p <0.001), respectively (where A =integrated absorbance, c = concen-tration, p = statistical error, and r =correlation coefficient). The meanrelative error between slopes was3.3%. These results implied theabsence of non-spectral interferencesin the ETA-AAS analyses for Cu andpermitted the use of either the cali-bration curves or the standard addi-tions methods for metalsquantification. However, bloodplasma Cu levels were evaluatedagainst aqueous copper standardcalibration curves.

The linearity of the calibrationcurve for Cu and Zn is very smalland only appears up to 15 µg/L Cuand 10 µg/L Zn, which indicatesthat its dynamic interval is quitelimited. As a consequence, quantifi-cation of these metals in the bloodplasma samples was carried out forextrapolation of the calibrationcurve, which has the form of aparabola arc (Figure 2). Similar con-centrations of Cu and Zn were usedto elaborate the calibration curves(~ 5, 10, 15, 20, 25, and 30 µg/L).These curves allowed the analyticaldetermination of Cu and Zn in theblood plasma samples.

The quantification of Zn in theblood plasma samples wasperformed by extrapolation of theworking curve in the form of aparabola arc because the spectro-metric dynamic range of Zn is verysmall (i.e., up to 10 µg/L). Figure 3shows the parallelism between theslopes of the calibration and stan-dard addition curves obtained forZn, indicating that non-spectralinterferences did not exist in theFTP-ETA-AAS determinations of thismetal. These results reinforced theanalytical use of the calibrationcurves for quantification purposesof Cu and Zn in blood plasma ofchildren with Down syndrome byETA-AAS.

The characteristic masses were5.3 and 0.9 pg/0.0044 s–1 for Cuand Zn, respectively, using a 20-µL

Fig. 2. Calibration curves for the ETA-AAS determination of Cu (A) and Zn (B) inblood plasma of children with Down syndrome.

123


with respective mean values of1839 ± 261 µg/L and 1374 ± 867µg/L. These results could be due tosimilar food intake of the infantilepopulation under study and thealeatory sampling carried out.

In accordance with the estab-lished Cu and Zn norm for healthyindividuals, the blood plasma Cumean concentrations listed in TableIV show a deficit of the metal inboth infantile populations[~900–1300 µg/L Cu (4)]. The lowCu levels found could becomeharmful for the children studiedsince levels below those considerednormal can cause illnesses such asanemia, high cholesterol, and coro-nary dysfunction (21). The copperlevels for Turkish Down children asreported by Cenquiz et al. (1)(~750 ± 170 µg/L Cu) are not statis-tically different from those found inthis work, indicating that the nutri-tional state of children living inpoor socio-economic areas is verysimilar to those in other parts of theworld.

Table IV also shows that the Znconcentrations found in the bloodplasma samples indicate a signifi-cant increase of this metal in com-parison to the established Zn normfor good health [~900–1100 µg/LZn (4)]. According to studies car-ried out by Artacho et al. (6) andMadaric et al. (22) no clinicalpathologies are associated withhigh levels of zinc in blood plasma.It can therefore be concluded thatthe high Zn level of the studiedpopulations does not pose anyhealth risks. Statistical differences(p <0.01) were found between theblood plasma Zn concentrations ofDown infants and the controls(Table IV). As is true for Cu, thefindings for Zn are related to thelow socio-economic stratum of thechildren studied (15). Cengiz et al.(2) reported blood plasma Zn con-centrations of 660 ± 180 µg/L forTurkish children with Down syn-drome, which are below the normas previously indicated. The blood

Fig. 3. Study of non-spectral interferences for the determination of Zn in bloodplasma samples of children with Down syndrome by ETA-AAS, showing theadequated parallelism between calibration ( ) and standard addition () curves.

injection volume was used. Experi-mental characteristic masses for Cuwere lower than the reported value(~6.4 pg) of the instrument used(20). The characteristic mass for Znwas higher than the reported value(~0.1 pg) by the spectrophotome-ter used (20).

The limits of detection (definedand calculated as three times thestandard deviation of the blanks,expressed in µg/L as a concentra-tion unit) in the blood plasma sam-ples analyzed were 0.1 µg/L forboth copper and zinc. In the caseof Cu, this value was lower than thereported value (~0.3 µg/L) by theinstrument employed (20). In solidsamples, Vale et al. (10) reported adetection limit of 0.014 µg/g Cu.For Zn, the detection limit was sim-ilar to the reported value by theinstrument. Several authorsreported detection limits rangingfrom 0.2 to 4.0 µg/L (7,11,18).

The methods developed wereused in our laboratory and allowedthe analysis of 70 samples within an8-hour working day. In addition toreducing the total analysis time, the

number of samples analyzed wasincreased, excellent reproducibilityand reliability was achieved, show-ing that the methods are suitablefor use in clinical laboratories ofhospitals. From an analytical pointof view, it can be stated that thereis no difference between samplescoming from children with or with-out Down syndrome, implying thatthe matrices of the evaluated bloodplasma samples are similar innature.

Levels of Cu and Zn in DownSyndrome

The concentrations of copperand zinc in blood plasma samplesof children with Down syndromeand the controls are listed in TableIV. For Cu, no statistically signifi-cant differences (p > 0.01) betweenDown and the control infantilepopulations were observed. Asshown in Table IV, the Cu meanvalues (± SD, µg/L) were 805 ± 261and 767 ± 288, for Down and thecontrols, respectively. For Zn,a significant statistical difference(p <0.01) was found betweenDown children and the controls,

124

plasma Zn mean values reported byCengiz et al. were very differentfrom those found in this investiga-tion (~1839 ± 361 µg/L Zn). Thiselevated Zn mean concentrationcan be ascribed to the nutritionalconditions of the children selectedfor this work.

CONCLUSION

A medical application of the fastfurnace program for the ETA-AASdetermination of copper and zincin blood plasma of children withDown syndrome and healthy con-trols was successfully carried out.This method provides analyticaland clinical tools for the evaluationof this kind of genetic anomaly.

These fast electrothermal atom-ization atomic absorption spectrom-etry methodologies allowed for theaccurate, precise, and interference-free determination of copper andzinc in clinical samples without theuse of two basic paradigms of theSTPF conditions: isoformation andplatform atomization.

ACKNOWLEDGMENTS

The authors would like toexpress their gratitude to La Univer-sidad del Zulia (LUZ, Maracaibo,Venezuela). The authors alsoacknowledge the previous instru-mental endowment obtainedthrough of the finished researchesand financed by Fondo Nacional deCiencia y Tecnología (FONACIT,Caracas, Venezuela) and Consejo de

Desarrollo Científico y Humanístico(CONDES, LUZ). This research waspartially supported by theresearchers from Laboratorio deInstrumentación Analítica (LIA-LUZ).

Received June 7, 2004.Revision received March 17, 2005.

REFERENCES

1. S.N. Willie, Y. Lida, and J.W.McLaren, At. Spectrosc. 19, 67(1998).

2. M. Cengiz, M. Seven, S. Cengiz, A.Yuksel, and M.Y. Iscan, TraceElem. Electr. 3, 156 (2000).

3. A.A. Ganeev, V.S. Vergizova, A.I.Drobyshev, S.E. Pogarev, V.V.Ryzhov, and S.E. Shelupov, J. Anal.Chem. 54, 61 (1999).

4. D.R. Fernández, Determination ofcooper and zinc in blood plasma ofchildren with Down syndrome byelectrothermal atomization atomicabsorption spectrometry. L.U.Z.,Maracaibo, Venezuela (2003).Bachelor Thesis.

5. H.M. Tvinnereim, R. Eide, T. Riise, G.Fosse, and G.R. Wesenberg, Sci.Total Environ. 226, 201 (1999).

6. R. Artacho, M.D. Ruíz López, C.Gámez, A. Puerta, and M.C. López,Sci. Total Environ. 205, 159(1997).

7. M.A. Moreno, C. Marín, F. Vinagre,and P. Ostapczuk, Sci. Total Envi-ron. 229, 209 (2000).

8. S. Sakao, and H. Vchida, Anal. Chim.Acta 382, 215 (1999).

9. Y. Bonfil, E. Brand, and Kirowa-Eisner, Anal. Chim. Acta 387, 85(1999).

10. M.G.R. Vale, M.M. Silva, B. Welz,and E.C. Lima, Spectrochim. Acta56, 1859 (2001).

11. P.R.M. Correia, E. Oliveira, and P.V.Oliveira, Anal. Chim. Acta 458, 321(2002).

12. B. Semprún de Villasmil, M.Hernández, M.C. Rodríguez, andV.A. Granadillo, Ciencia 10, 68(2002).

13. V.A. Granadillo, L. Parra deMachado, and R.A. Romero, Anal.Chem. 66, 3624 (1994).

14. Standard Specification for Water;ASTM D 1193-77; American Societyfor Testing and Materials, Philadel-phia, PA, USA (1977).

15. H. Méndez Castellano,"Estratificación social. Método deGraffar modificado paraVenezuela". Arch. Ven. Puericult.Pediatr. 49, 93 (1986).

16. CRC Handbook of Chemistry andPhysics, Weast, R.C., Ed., 65th ed.;CRC Press, Inc.: Boca Raton, FL(USA), 1984/1985, p B-95 y B-159.

17. J.E. Tahán, V.A. Granadillo, and R.A.Romero, Anal. Chim. Acta 295, 187(1994).

18. P. Bermejo, E. Peña, D. Fompedriña,R. Domínguez, A. Bermejo, J.M.Fraga, and J.A. Cocho, Analyst 126,571 (2001).

19. D. Bohrer, P. Cícero doNascimento, M. Trevisan, and E.Seibert, The Analyst 124, 1345(1999).

20. Analytical Methods for FurnaceAtomic Absorption Spectrometry;PerkinElmer, Überlingen, Germany(1984).

21. D.M. Medeiros, J. Davidson and J.E.Jenkins, Proc. Soc. Exp. Biol. Med.203, 262 (1993).

22. A. Madaric, E. Ginter and J.Kadrabova, Physiol. Res. 43, 107(1994).

TABLE IVETA-AAS Determination of Copper and Zinc in Blood Plasma

of Children With Down Syndrome and a Control Group From the City of Maracaibo, Venezuela

Sample Metal Mean ± SDa C.V.(%) Experimental Rangea

Down Cu 805 ± 261 32.37 366 – 1616Zn 1839 ± 361 19.6 1183 – 2660

Control Cu 767 ± 288 37.57 500 – 1500

Zn 1374 ± 867 63.1 683 – 3817

a In µg/L.

125

Corresponding author.E-mail: [email protected]: +54-2652-430224

Determination of Tellurium by ETAAS With Preconcentration by Coprecipitation

With Lanthanum HydroxideS. Ceruttia,b, M. Kaplana,b, S. Moyanoa, J.A. Gásqueza, and *L.D. Martineza,b

a Departamento de Química Analítica, Facultad de Química, Bioquímica y Farmacia, Universidad Nacional de San Luis, Chacabuco y Pedernera, P.O. Box 375, 5700 San Luis, Argentina

b Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Rivadavia 1917, CP C1033 AAJ, Ciudad de Buenos Aires, Argentina

Atomic SpectroscopyVol. 26(3), May/June 2005

INTRODUCTION

Tellurium is a rare trace elementand used as an additive to improvealloys. Powdered tellurium isemployed as a secondary vulcaniz-ing agent in various kinds of rub-bers (natural rubber andstyrene-butadiene rubbers) since itreduces curing time and endowsthe rubbers with increased resis-tance to heat and abrasion. Due toits photoelectric properties,tellurium and its compounds arealso employed in the semiconduc-tor and electronic industries (1).Because of the extremely low levelsof tellurium in various matrices, asensitive method is required for itsdetermination.

In the past years, several meth-ods have been developed for thedetermination of Te at low concen-trations, among them anodic strip-ping voltammetry (2–4), inductivelycoupled plasma optical emissionspectrometry (ICP-OES) (5,6),inductively coupled plasma massspectrometry (ICP-MS) (7,8), andelectrothermal atomic absorptionspectrometry (ETAAS) (8–11).

Electrothermal atomic absorp-tion spectrometry appears to beone of the most attractiveapproaches for trace element deter-mination (12–14). However, thedirect determination of ultratraceamounts of elements by ETAAS isusually difficult owing to an insuffi-cient instrument detection power.

ABSTRACT

A procedure for the determi-nation of traces of total telluriumis described which combineselectrothermal atomic absorptionspectrometry (ETAAS) with pre-concentration of the analyte bycoprecipitation. The samples,each spiked with lanthanumnitrate (20 mg L–1), areintroduced into the AmberliteXAD-4 resin, and mixed withammonium buffer (pH 9.1). Tel-lurium is preconcentrated bycoprecipitation with the gener-ated lanthanum hydroxide pre-cipitate. The precipitate isquantitatively collected in theresin, and subsequently elutedwith 5% (v/v) nitric acid. Thedetermination is developed withETAAS.

Considering a sampleconsumption of 25 mL, anenrichment factor of 10 wasobtained. The detection limit(3σ) was 0.04 µg L–1 and the pre-cision (relative standard devia-tion) was 3.5% (n=10) at the10-µg L–1 level. The calibrationcurve, using the preconcentra-tion system for tellurium, waslinear with a correlation coeffi-cient of 0.9993. Satisfactoryresults were obtained for thedetermination of tellurium in thestandard reference material NIST1643d Trace Elements in Waterand in a tap water sample.

Consequently, preconcentrationprocedures, such as ion exchange,adsorption, solvent extraction andcoprecipitation, are often neededbefore ETAAS determination.

Separation and preconcentrationtechniques using sorption extrac-

tion (2–4), solvent extraction (10),ion exchange (9), and coprecipita-tion (6) have been employed forthe preconcentration of telluriumin batch and flow injection modes.

The use of a knotted reactor(KR) as collector for precipitateshas been found feasible in flowinjection (FI) on-line precipitation-preconcentration systems. Besides,using KR, stainless-steel filters andpacked-bed filters have also beenused to retain precipitatedcompounds (15–17). Although XADadsorption resins have beenemployed in on-line preconcentra-tion systems for retaining solublecomplexes (18–20), in a previouswork we have reported the use ofXAD-7 resin as a filter packing toretain Pb-diethyldithiocarbamateprecipitate (21).

In the present work, a methodfor the preconcentration oftellurium and its determination byETAAS is proposed. Tellurium waspreconcentrated by coprecipitationwith lanthanum hydroxide using apacked-bed filter with AmberliteXAD-4 resin packing. The method-ology was applied for the analysisof tellurium in the standard refer-ence material NIST 1643d TraceElements in Water and in tap water.

EXPERIMENTAL

Instrumentation

The measurements wereperformed with a Shimadzu ModelAA-6800 atomic absorption spec-trometer (Tokyo, Japan), equippedwith a deuterium background cor-rector, a Model 6500 electrother-mal atomizer, and an ASC-6100

126

min–1 loading flow rates,respectively, was loaded by meansof peristaltic pump into thecolumn. The sample and bufferstreams merged at a point about1 cm upstream of the column. Theprecipitate, which is formed instan-taneously after the merging point,was collected on the XAD-4 resin.The effluent emerging from theresin was discarded. The coprecipi-tated analyte was eluted from theresin with nitric acid up to a finalvolume of 2.5 mL.

After the preconcentration step,50 µL of the analyte solution wasautomatically introduced into thegraphite tube by means of the fur-nace autosampler. Then, theautosampler arm was moved backto the wash position and the atom-ization program was started. Theabsorbance measurements (peakheight) were proportional to thetellurium concentration in the sam-ple, and were used for all measure-ments. With respect to thecharacteristic mass for tellurium,the manufacturer does not reportany value for this analyte. The char-acteristic mass obtained in our labo-ratory was 20 pg.

Lanthanum nitrate solution (0.5%m/v) was made by dissolving 0.66 gof lanthanum nitrate hexahydrate in100 mL of ultrapure water.

The buffer solutions used werein all instances 0.2 mol L–1 ammo-nium chloride, adjusted to theappropriate pH (9.0–9.2) by addi-tion of 0.2 mol L–1 ammonia; theoptimum pH was found to be 9.1.

Ultrapure water (18 MΩ cm–1)was obtained from an EASY pureRF (Barnstedt, Dubuque, IA, USA).

All solvents and reagents were ofanalytical reagent grade or better,and the presence of tellurium wasnot detected in the working range.

Sample Pretreatment

Water samples were filteredthrough a 0.45-µm pore size mem-brane filter. An appropriate amountof 0.5% lanthanum nitrate solutionwas added (so that each samplesolution contained a final concen-tration of 20 mg L–1 lanthanumnitrate) and then acidified to pH 3.0with 0.1 mol L–1 hydrochloric acid.

Operational Procedure

A 25-mL sample and buffer solu-tion at 5.0 mL min–1 and 1.0 mL

autosampler. Wall atomization withstandard high-density graphitetubes (Shimadzu-Tokyo-Japan) wasused. Some experiments weredeveloped comparing the perfor-mance of high density graphitetube, pyrolytic graphite tube, andgraphite tube with L’vov platform.Best results were obtained with thehigh density graphite tube. The useof a matrix modifier was not neces-sary because with the furnace pro-gram used there were no analytelosses. The introduction of oxygenwas not necessary due to the sim-plicity of the matrix under study.A tellurium hollow cathode lamp(Hamamatsu Photonics K.K., Japan)was employed as the radiationsource.

The instrument settings and fur-nace program are detailed in TableI. A Minipuls™ 3 peristaltic pump(Gilson, Villiers-Le-Bell, France) wasused. A home-made microbore glasscolumn (50 mm length; 3 mm inter-nal diameter) fitted with porous 25µm glass frits was used as the resinholder. Pump tubes, Tygon® type(Ismatec, Cole-Parmer InstrumentCompany, Niles, IL, USA), wereemployed to propel the sample,reagent, and eluent.

Reagents

Amberlite XAD-4 resin (Rohm &Haas, Philadelphia, PA, USA) wasused. The particle size wasbetween 20 and 50 mesh with asurface area of 450 m2 g–1. Beforeuse, the surface of the resin wasactivated by immersion in a solu-tion of 4 mol L–1 methanol/hydrochloric acid (1:1). Subsequently,the metal impurities were removedby further washing with 2 mol L–1

HCl solution.

Tellurium standard solution wasprepared by appropriate dilutionsof a 1000 mg L–1 stock solution(Merck, Darmstadt, Germany)immediately before use.

TABLE IInstrumental Operating Parameters and

Furnace Temperature Program for Te Determination

Parameters

Wavelength 214.3 nmSpectral Bandpass 0.2 nmLamp Current 14 mABackground Correction Deuterium Lamp

Furnace Program

Stage Temp. (ºC) Time (s) Argon Gas Flow Ramp Hold (L min–1)

Drying 120 30 - 0.10250 10 - 1.0

Pyrolisis 600 - 20 1.0600 - 5 0.0

Atomization 2500 - 3 0.0 (Read)

Cleaning 2600 - 2 1.0

127



Effect of pH and LanthanumConcentration

Preliminary experiments showedthat the pH and the lanthanum con-centration in the sample were criti-cal for the precipitation reactionitself and the subsequent recoveryof tellurium. Measured on standardsall containing 10 µg L–1 of tellurium,the absorbance remained almostconstant in the pH ranges of9.0–9.2. The results obtained areshown in Figure 1. Moreover, thelanthanum nitrate concentrationwas within the range: 15–30 mgL–1. In the present work, 20 mg L–1

lanthanum nitrate and a pH of 9.1were selected for subsequent stud-ies. These results are in agreementwith those obtained by Tao et al.(15) who used knotted reactors forthe determination of Se .

Selection of Resin

A resin Amberlite XAD-4 packingfilter allowed the tellurium presentin the samples to be efficiently pre-concentrated. This might have beendue not only to retention by filtra-tion of the precipitate, but also pos-sibly due to adsorption effects ofthe precipitate on the resin surface.The resin size constitutes an impor-tant parameter, since it must allowappropriate precipitate retentionwith low hydrodynamicimpedance. The particle size usedin this work (20–50 mesh) permit-ted us to obtain optimum retention(95%) at sample flow rates of 5 mLmin–1.

Sample Loading Rate

The sample flow rate throughthe column is one of the steps thatcontrols the preconcentrationtime. In this study it was verifiedthat with flow rates up to 5.0 mLmin–1 there is no effect on analyterecovery, which under optimumconditions is 95%. Figure 2 showsthat at higher flow rates the recov-ery decreases.

Fig. 1. Effect of pH of loading solutions. Sample loading volume 25 mL, loading flow rate 5 mL min–1, elution flow rate 1.5 mL min–1, Te concentration 10 µg L–1, lanthanum nitrate concentration 20 mg L–1.

Fig. 2. Analysis of sample loading rate. Sample loading volume 25 mL, elution flow rate 1.5 mL min–1, Te concentration 10 µg L–1, lanthanum nitrate concentration 20 mg L–1.

128

Effect of Eluent

A satisfactory eluent shouldeffectively dissolve the precipitatewith a discrete volume in order toobtain a better enrichment factor.Nitric acid turned out to be a goodeluent for the tellurium coprecipi-tate with lanthanum hydroxide. Itwas found that 5% (v/v) nitric acidwas the minimum concentrationnecessary to obtain best response.The optimum flow rate of the elu-ent used was 1.5 mL min–1.

Interference Studies

The proposed coprecipitationsystem can tolerate the presence ofions at the concentration levels thatmay be found in natural water sam-ples. Thus, Cu2+, Zn2+, Cd2+, Ni2+,Co2+, Mn2+, and Fe3+ could be toler-ated up to at least 2500 µg L–1.Commonly encountered matrixcomponents, such as alkali andalkaline earth elements, are notretained on the column.

Analytical Performance

The time required for the pre-concentration of 25 mL of sample(5.0 min, flow rate of 5 mL min–1),elution (1.7 min, flow rate of 1.5mL min–1), and conditioning (0.2min) was about 6.9 min. Addition-ally, the time required for theETAAS determination was about 1.0min, resulting in a sample through-put of seven samples per hour.

A 10-fold total enrichment factorfor a sample volume of 25 mL wasobtained with respect to thetellurium determination by ETAASwithout preconcentration.

The relative standard deviation(RSD) for 10 replicates containing10 µg L–1 of Te was 3.5%. The cali-bration curve was linear, with acorrelation coefficient of 0.9993.The detection limit (DL), calculatedas the amount of Te required toyield a net peak equal to threetimes the standard deviation of thebackground signal (3s), was 0.04 µgL–1.

In comparison to our method,the procedure of multipleinjections would be too time-con-suming. Besides, it would require10 injections of 50 µL each toobtain a signal comparable withthat of the preconcentation proce-dure. On the other hand, webelieve that a drying stage wouldbe indispensable in order to inject500 µL due to the limited capacityof the graphite tube (approximately100 µL).

Recovery and Validation Studies

Since tellurium was not detectedin the tap water samples, we spikedthe samples with a known quantityof tellurium and applied the pre-concentration procedure.

In order to evaluate the telluriumrecovery of this method, 250 mL oftap water sample was collected inour laboratory and divided into 10portions of 25 mL each. The pro-posed method was applied to sixportions and the average quantityof tellurium obtained was taken asthe base value. Then, increasingquantities of tellurium were addedto the other sample aliquots andtellurium was determined by thesame method. The results obtainedwere between 95–102 % (Table II).Even though the method of stan-dard addition is not as useful as themethod of validation; it is consid-ered as a method of validation (22).

Additionally, the proposedmethod was applied to a standardreference material, NIST SRM1643d Trace Elements in Water,with a reference tellurium contentof 1.0 µg L–1. The density of theSRM 1643d sample at 22ºC was1.016 g mL–1. Using our method,the tellurium concentration foundin this SRM was 1.1 ± 0.1 µg L–1.

CONCLUSION

Although the proposed method-ology was not completelyautomated, the on-line preconcen-tration system increases the speedof the preconcentration and analy-sis process.

The preconcentration procedurewith lanthanum hydroxide using apacked-bed filter with AmberliteXAD-4 resin packing and coupledto ETAAS allowed better detectionlimits, approximately 10-fold, incomparison to tellurium determina-tion without preconcentration. Therecovery studies performed indi-cate that the method shows goodreproducibility and accuracy.

ACKNOWLEDGMENTS

This work was supported byConsejo Nacional de InvestigacionesCientíficas y Técnicas (CONICET);Agencia Nacional de PromociónCientífica y Tecnológica (FONCYT)(PICT-BID); and UniversidadNacional de San Luis (Argentina).

TABLE IIRecovery Study

Aliquots Base Value Te Added Te Found Recovery(µg L–1) (µg L–1) (µg L–1) (%)a

1–6 – 0.0 0.0 –

7 0.0 1.0 0.95 95.0

8 0.0 2.0 1.98 99.0

9 0.0 3.0 3.06 102.0

10 0.0 5.0 5.0 100.0

a 100 x [(Found-Base)/Added].

129


14. Z-L. Fang, Spectrochim. Acta Part B53, 1371 (1998).

15. G. Tao and E. Hansen, Analyst 119,333 (1994).

16. S. Nielsen, J.J. Sloth, and E. Hansen,Analyst 121, 31 (1996).

17. Z-L. Fang, Flow Injection Separationand Preconcentration, VCH, Wein-heim, Germany (1993).

18. V. Porta, C. Sarzanini, E. Mentasti,and O. Abollino, Anal. Chim. Acta258, 237 (1992).

19. R.G. Wuilloud, H.A. Acevedo, F.A.Vazquez, and L.D. Martinez, Anal.Letters 35, 1649 (2002).

20. S. Moyano, R.G. Wuilloud, R.A.Olsina, J.A. Gásquez, and L.D.Martinez, Talanta 54, 211 (2001).

21. R.G. Wuilloud, J.A. Salonia, J.A.Gásquez, R.A. Olsina, and L.D.Martinez, Anal. Sciences 17, 457(2001).

22. E. Prichard, G.M. MacKay, and J.Points, Trace Analysis: A structuresapproach to obtaining reliableresults, The Royal Society of Chem-istry, UK (1996).

Received August 13, 2004.Revision received March 14, 2005.

REFERENCES

1. H.G. Seiler, A. Sigel, and H. Sigel,Handbook on Metals in Clinicaland Analytical Chemistry, MarcelDekker, New York (1994).

2. J. Wang and J. Lu, Electroanalysis 6,405 (1994).

3. T. Ferri, S. Rossi, and P. Sangiorgio,Anal. Chim. Acta 361, 113 (1998).

4. M. Baldelli, T. Ferri, A. Liberatori,and B.M. Petronio, Analusis 22,147 (1994).

5. P. Fodor and R.M. Barnes,Spectrochim. Acta 388, 229(1983).

6. L. Halicz and G.M. Russell, Analyst111, 15 (1986).

7. A.G. Coedo and M.T. Dorado, J. Anal.At. Spectrom. 9, 1111 (1994).

8. S.K. Aggarwal, M. Kinter, J. Nichol-son, and D. Herold, Anal.l Chem.66, 1316 (1994).

9. M. Grotti and A. Mazzucotelli, J.fAnal. At. Spectrom. 10, 325 (1995).

10. V.G. Torgov, G.A. Vall, M.G. Demi-dova, and V.T. Yatsenko, Chem.Geol. 124, 101 (1995).

11. T. Wickstrom, W. Lund, and R. Bye,Analyst 120, 2695 (1995).

12. R.E. Sturgeon, Fresenius’ J. Anal.Chem. 355, 425 (1996).

13. E. Vereda Alonso, A. García deTorres, and J.M. Cano Pavón,Talanta 55, 219 (2001).

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 own lan-guage. All instrument controls are right there on the screen, available at your fin-gertips. Even troubleshooting and repairs are easier, with quick-change parts yousimply snap out and snap in. No service visit, no down time. As rugged and reli-able as ever, our newest AAnalyst is a better way to do AA. Experience it for your-self. Talk to a PerkinElmer inorganic analysis specialist today.

© 2005 PerkinElmer, Inc.

Atomic Absorption

Justtouch

and go.

007305_03

tomics pectroscopy - a and b spectroscopy 26(3).pdf · pectroscopy vol. 26(3) may/june 2005...

Documents