as 29(1) 1

Issues also available

electronically.

(see inside front cover)

ASPND7 29(1) 1–38 (2008)ISSN 0195-5373

AtomicSpectroscopy

January/February 2008 Volume 29, No. 1

In This Issue:Separation of Trace Impurities from Boric Acid Using Cloud Point Extraction for TheirDetermination by Dynamic Reaction Cell Inductively Coupled Plasma Mass SpectrometryA. C. Sahayam, Shiuh-Jen Jiang, and Feng-yi Chen ............................................................. 1

Synergic Solvent Extraction of Rare Earth Elements Using Mixed Ligand Complexes ofHexafluoroacetylacetone and Tri-n-butylphosphate and Their Determination inEnvironmental Waters by Low Temperature ETV-ICP-MSChaomei Xiong, Bin Hu, and Zucheng Jiang ........................................................................ 6

Line Selection and Evaluation of Radio Frequency Glow Discharge Atomic EmissionSpectrometry in the Determination of Trace Amounts of Niobium in ElectronBeam-melted Tantalum SheetG. Anil, M.R.P. Reddy, S. Rajesh Kumar, N.R. Munirathnam, and T.L. Prakash .............. 16

Trace Molybdenum Determination in Drinking Waters by USN-ICP-OES AfterSolid Phase Extraction on Ethyl Vinyl Acetate Turnings-Packed MinicolumnLuis Escudero, Raúl A. Gil, José A. Gásquez, Roberto A. Olsina, and Luis D. Martínez ... 21

Determination of Hexavalent Chromium in Welding Fumes by Flow Injection FlameAtomic Absorption Spectrometry After Dynamic Alkaline Ultrasound-assistedExtraction/Anion Exchange PreconcentrationM.C. Yebra and R.M. Cespón ............................................................................................... 27

Ion Pair Solid Phase Extraction for the Indirect Determination of Gatifloxacinby Flow Injection Flame Atomic Absorption Spectrometry Han-Ying Zhan, Mei-Zhen Ning, Zhi-Qi Zhang, and Li-Ping Kang ................................... 32

PerkinElmer is a registered trademark of PerkinElmer, Inc.Dynamic Reaction Cell and DRC are trademarks of PerkinElmer, Inc.SCIEX and ELAN are registered trademarks of MDS SCIEX, a division of MDS Inc.Gilson is a registered trademark and Minipuls is a trademark of Gilson, Villiers-Le-Bell, France.Milli-Q is a trademark of Millipore Corporation, Bedford, MA, USA.Rheodyne is a registered trademark of Rheodyne, Inc.Sep-pak is a registered trademark of Waters, Milford, MA, USA.Suprapur is a registered trademark of Merck & Co., Darmstadt, Germany.Triton is a registered trademark of Union Carbide Chemicals & Plastics Technology Corporation.Tygon is a trademark of Norton Co.

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. 29(1) Jan./Feb. 2008

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

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

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

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

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

Atomic Spectroscopy for format.

6. Editor’s e-mail:[email protected]

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

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

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

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

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

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

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

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

EDITORAnneliese LustE-mail: [email protected]

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

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

2008 Subscription Rates• U.S. $70.00 includes first-class

mail delivery worldwide $20.00 extra for electronic file.

• 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 © 2008PerkinElmer, Inc.All rights reserved.http://www.perkinelmer.com

UMI Microform Editions ofAtomic Spectroscopy available from:National Archive Publishing Company300 N. Zeeb RoadAnn Arbor, MI 48106-0998 USAhttp://www.napubco.com

1

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

Separation of Trace Impurities from Boric Acid Using Cloud Point Extraction for Their Determination

by Dynamic Reaction Cell Inductively Coupled Plasma Mass Spectrometry

A. C. Sahayama, *Shiuh-Jen Jiangb, and Feng-yi Chenb

a National Centre for Compositional Characterisation of Materials (CCCM), Hyderabad, India b Department of Chemistry, National Sun Yat-sen University, Lien-hai Road, Kaohsiung 80424, Taiwan

Atomic SpectroscopyVol. 29(1), Jan./Feb. 2008

INTRODUCTION

Boric acid has several industrialapplications, including the nuclearindustry to slow down the rate atwhich fission is occurring (1) bybeing dissolved in primary coolant,since boron has high absorptioncross-section for neutrons (2). Boricacid is a precursor to prepare B2O3

which is used in the optical waveguide industry (3,4). The presenceof impurities in boric acid willaffect the properties of the prod-uct. Boric acid is also used to com-plex with HF where HF is used todissolve samples, especially for ICP-based techniques where the pres-ence of even traces of HF erodesthe nebulizer (5). In such cases,preliminary knowledge of the con-centration of impurities in boricacid is essential. It is always desir-able to have a simple and fast pro-cedure for the analysis of suchmaterials. Though boric acid is solu-ble in water, it is necessary to sepa-rate the boron from the traceanalytes prior to analysis. This isimportant since boron is known toclog injectors and nebulizers. Boronremoval procedures based onboron volatilization as methoxyborate are time consuming (6,7).Hence an alternative procedurethat is fast and suitable for routineanalysis is of interest.

Surfactant-mediated cloud pointextraction has been reported to befast and simple and has beenapplied to the environmental moni-toring of toxic metals and also spe-

ABSTRACT

A method for the separationof Mn, Fe, Ni, Co, Cu, Zn, Ga, Cd,and Pb from boric acid has beendeveloped using cloud pointextraction (CPE). The impuritieswere complexed with 8-hydroxy-quinoline (8-HQ) at pH 6.5 priorto cloud point extraction usingTriton X-114. The percentagerecoveries of trace impuritiesadded to the sample were 80-102% (60% for Fe) whereas theseparation of boron was 99.9%.The pre-concentration factorsobtained after CPE separationwere 120–150. The processblanks were in ng g–1 levels, andthe detection limits were in therange 40 pg g–1 to few ng g–1

using dynamic reaction cellinductively coupled plasma massspectrometer (DRC ICP-MS). Theinterference of 40Ar16O+ on thedetermination of 56Fe+ has beenalleviated in DRC mode usingNH3 as a reaction gas. Themethod has been applied to thedetermination of Mn, Fe, Ni, Co,Cu, Zn, Ga, Cd, and Pb in twoboric acid samples. The methodhas been verified with avolatilization procedure whereboron was volatilized as methoxyborate using methanol.

analysis of boric acid. Inductivelycoupled plasma mass spectrometry(ICP-MS) is a versatile technique forthe determination of elements attrace and ultra-trace levels. How-ever, in materials analysis the majormatrix elements need to be sepa-rated prior to determination to pre-concentrate impurities and alleviatematrix-based interferences (13,14).The aim of the present work is todevelop a method for the separa-tion of trace impurities from boricacid using cloud point extraction,followed by their measurementwith an inductively coupled plasmadynamic reaction cell mass spec-trometer (ICP-DRC-MS). The condi-tions for cloud point extraction ofimpurities, such as pH and the con-centration of complexing agents,are also reported.

EXPERIMENTAL

Instrumentation

An ELAN® 6100 DRC ICP-MS(PerkinElmer SCIEX, Concord,Ontario, Canada) was used for thedeterminations. Samples were intro-duced by a cross-flow pneumaticnebulizer with a Scott-type spraychamber. NH3 reaction gas(99.999% purity) was obtainedfrom Air Liquide (Taiwan). Theoperating parameters are shown inTable I.

Reagents and Samples

Suprapur® grade HNO3 wasobtained from Merck (Darmstadt,Germany). All dilutions were car-ried out using de-ionized water of18.2 MΩ cm–1 resistivity preparedusing a Milli-Q™ system (Millipore,

ciation of inorganic compoundsafter complexation with suitablereagents (8–12). Though widelyused in environmental analysis, theapplication of cloud point extrac-tion to materials analysis is yet tobe explored.

In the present study, an attempthas been made to apply cloud pointextraction (CPE) to the trace metal

2

extraction procedures are reportedto be suitable for the separation ofneutral molecules that arehydrophobic in nature. Trace met-als are converted to hydrophobicmolecules by complexing with suit-able reagents. The organo metalliccompounds thus obtained areextracted into a surfactant- richphase using cloud point extraction.Solutions with a high salt contentappear to have no negative effecton the cloud point extraction oftrace elements, which is a majorproblem for the analysis by instru-mental techniques such as ICP-MS(8,15). For these reasons, anattempt has been made to applycloud point extraction to the sepa-ration and pre-concentration oftrace elements from a concentratedboric acid solution (1 g in 30 mL).

Selection of Conditions forCloud Point Extraction

8-Hydroxyquinoline is a well-known complexing agent thatforms complexes with Mn, Fe, Co,Ni, Cu, Zn, Ga, Cd, and Pb. It hasbeen used for the pre-concentra-tion of transition metals from sea-water (16). However, the matrixelement in boric acid does not havean affinity for 8-HQ. Hence, 8-HQ isused to complex the impurities ofinterest. Triton X-114 has been cho-sen as the surfactant for the cloudpoint extraction due to its lowercloud point, defined as the temper-ature above which an aqueous solu-tion of water soluble surfactantbecomes turbid. As the complexa-tion is pH-dependent, a suitable pHhas been selected by evaluating apH range from 3–7. The results areshown in Table II. As can be seen,although the optimum pH is differ-ent for some elements (i.e., pH 4for Co), the variation in the recov-eries is smallest at pH 6.5. Hence,to extract all the elements in a sin-gle step, pH 6.5 was selected sincethe recovery of most of theelements is highest at this pH. Thevolume of Triton X-114 selected

Bedford, MA, USA). 8-Hydroxy-quinoline (8-HQ) was from Riedel-de Haen (Seelze, Germany). Boricacid samples were procured fromShowa (Tokyo, Japan) and Sigma(St. Louis, MO, USA). Triton® X-114was from Sigma, and methanolfrom Merck. Glycerol used for thevolatilization procedure wasobtained from Showa. Multi-element standard solutions (10 µg mL–1) were obtained fromSPEX CertiPrep (Metuchen, NJ,USA). 10% (m/v) 8-HQ in methanolwas prepared daily. 10% (m/v) Tri-ton X-114 was prepared in Milli-Qwater. A pH 6.5 solution was pre-pared using KOH (Merck) andpotassium hydrogen phthalate(Fisher, Fair Lawn, NJ, USA). Allchemicals were used without fur-ther purification.

Procedure

Cloud point extraction of tracemetals from boric acid has beencarried out using Triton X-114. To aboric acid solution (1g in 30 mL pH6.5 solution), 1 mL 10% 8-HQ wasadded and mixed well. Then, 2 mLof 10% Triton X-114 was added,shaken, and kept at 70 oC for 1 hr.

This solution was then cooled inice for 30 minutes, and the surfac-tant-rich phase separated anddiluted to 5 mL using 2% (v/v)HNO3. Rh (1 ng mL–1) was added as an internal standard, and thesolutions were analyzed by ICP-DRC-MS. A blank solution was alsoprepared by processing 30 mL pH6.5 solutions as above. The percent-age recoveries were calculated bytreating the sample spiked with 50 ng of multi-element standard. Tooptimize the pH of the solution, theabove procedure was repeated atdifferent pH values.

RESULTS AND DISCUSSION

The analysis of boric acid wasattempted using 1% (m/v) boricacid solution by ICP-DRC-MS. How-ever, the determination of most ofthe elements could not be carriedout as the concentrations of impuri-ties are at lower ng g–1 levels beforedilution, which is significant toreduce the boron content. Toimprove the detection limits, a pre-concentration step involving theseparation of the matrix element isnecessary. Surfactant-mediated

TABLE IOperating Parameters of ICP-DRC-MS

ICP-MS PerkinElmer SCIEX ELAN 6100 DRC IIPlasma Conditions

Outer Gas Flow 15 L min–1

Intermediate Gas Flow 1.13 L min–1

Carrier Gas Flow 1.05 L min–1

RF Power 1100 WData Acquisition

Dwell Time 100 msScan Bode Peak HoppingSweeps per Reading 10

Isotopes Monitored 55Mn, 56Fe, 58Ni, 59Co, 63Cu, 64Zn, 71Ga, 114Cd, 208Pb

DRC ParametersCell Gas NH3

Cell Gas Flow 0.4 mL min–1

Rpa 0.0

Rpq 0.7

3

Vol. 29(1), Jan./Feb. 2008

is heated to 70 oC since it is foundto be necessary to achieve a firmsurfactant-rich phase (glue) whichenabled its easy separation. More-over, increases in temperaturealways enhance a chemical reac-tion, which was found to be favor-able in this case. Pre-concentrationfactors, calculated for a 30-mL ini-tial solution and 0.2 mL glue usingthe following equation, are in therange of 120–150:

Concentration of Analyte

in Surfactant-rich Phase

Pre-concentration Factor =______________________

Concentration of Analyte

in the Initial Solution

was 2 mL for a 30-mL sample solu-tion, which is above the criticalmicelle concentration of Triton X-114. Volumes lower than 2 mLwere not attempted due to the diffi-culty in managing smaller surfac-tant-rich volumes from 30-mLsample solutions.

The concentration of the com-plexing agent appears to be critical.A 1% (m/v) 8-HQ in methanol solu-tion was selected initially. Prelimi-nary results indicate that therecoveries increased with volumeof complexing agent. However, therecoveries are rather poor, rangingfrom 40–56% (32% and 16% for Cd

and Pb), for a 6-mL solution -possi-bly due to the solubility ofcomplexes in methanol. Hence, toobtain better recoveries and todecrease the volume of methanol inthe solution, a 10% (m/v) 8-HQ inmethanol solution has been chosen.The volume of the 8-HQ requiredhas been optimized by varying thevolume from 0.25 mL to 1.0 mL. Asshown in Table III, 1.0 mL resultedin better recoveries and, therefore,was selected. All studies were car-ried out by processing a 50-ng mul-tielement standard solution of theanalytes of interest. Though thecloud point temperature of TritonX-114 is around 25 oC, the solution

TABLE IIPercentage Recoveriesa of Impurities With Change in pH (n=3)

Isotope pH of Solution

3 4 5 6 6.5 755Mn 95.5 ± 0.6 96.7 ± 0.8 98.2 ± 0.7 99.5 ± 0.9 99.8 ± 0.9 98.3 ± 0.856Fe 52.8 ± 1.0 53.4 ± 0.8 53.8 ± 0.8 59.5 ± 0.7 60.4 ± 0.7 58.9 ± 1.058Ni 86.5 ± 1.2 89.6 ± 1.1 96.8 ± 1.2 99.3 ± 0.9 105.1 ± 1.3 99.8 ± 1.059Co 84.0 ± 0.8 92.6 ± 0.6 89.5 ± 0.8 86.5 ± 0.7 84.6 ± 0.7 82.6 ± 0.863Cu 85.5 ± 0.9 87.8 ± 0.8 88.3 ± 0.7 87.0 ± 0.8 88.7 ± 0.7 87.2 ± 0.764Zn 82.2 ± 1.0 84.8 ± 0.8 83.0 ± 0.9 87.3 ± 0.6 87.8 ± 0.6 84.1 ± 0.971Ga 95.4 ± 0.8 91.1 ± 0.9 95.9 ± 0.8 96.8 ± 0.7 98.2 ± 0.6 98.3 ± 0.6114Cd 65.0 ± 1.1 76.0 ± 0.8 81.9 ± 0.6 82.0 ± 0.6 82.2 ± 0.6 72.8 ± 0.9208Pb 54.5 ± 0.9 75.0 ± 0.6 74.9 ± 0.6 78.8 ± 0.5 80.5 ± 0.6 61.5 ± 0.8

a 30-mL solution, 2 mL 10% Triton X-114, 1 mL 10% 8-HQ, 50 ng of each analyte.Uncertainties are expressed as 1 σ variation in measurements of three different experiments.

TABLE IIIPercentage Recoveriesa of Impurities With Change in Volume (mL) of 10% 8-HQ (n=3)

Isotope Volume of Solution (mL)

0.25 0.50 0.75 1.055Mn 58.0 ± 1.2 93.1 ± 0.9 96.1 ± 0.8 98.0 ± 0.856Fe 48.1 ± 1.4 52.9 ± 1.1 56.4 ± 0.9 60.6 ± 0.858Ni 83.5 ± 1.0 89.0 ± 0.8 89.6 ± 0.8 103.0 ± 1.059Co 73.5 ± 1.1 83.3 ± 0.9 85.3 ± 0.9 87.5 ± 0.863Cu 59.1 ± 1.0 74.0 ± 0.8 88.4 ± 0.8 94.9 ± 0.564Zn 64.0 ± 1.3 76.6 ± 1.1 92.4 ± 0.7 93.6 ± 0.671Ga 78.6 ± 0.9 88.8 ± 0.6 90.0 ± 0.6 90.0 ± 0.6114Cd 38.5 ± 1.1 57.0 ± 0.9 76.1 ± 0.7 82.8 ± 0.5208Pb 26.5 ± 1.8 41.4 ± 1.2 49.0 ± 0.9 61.0 ± 0.7

a 30-mL solution at pH 6.5, 2 mL 10% Triton X-114, 50 ng of each analyte.Uncertainties are expressed as 1 σ variation in measurements of three different experiments.

4

Separation of Analytes FromBoric Acid

The recoveries of analytes fromboric acid were carried out by pro-cessing varying amounts of boricacid (0.1 to 1.0 g in 30 mL) afterspiking 50 ng of each analyte. Ascan be seen in Table IV, there is notmuch difference in recoveries, evenwith 1 g of boric acid. Hence, theanalysis has been carried out using1 g of boric acid in 30 mL of pH 6.5solution. Higher amounts (>1 g) ofboric acid were not attemptedsince the dissolution of higheramounts of boric acid in 30 mLbuffer solution is difficult. The sepa-ration of boron was found to be99.9%.

Application

Since the recoveries of the ana-lytes are adequate to carry out theanalysis, the procedure was appliedto the analysis of two boric acidsamples. The determinations werecarried out using ICP-MS. However,in order to measure 56Fe, thedynamic reaction cell (DRC‘) modewas used to alleviate the interfer-ence of 40Ar16O+ on 56Fe+. The reac-tion gas used was NH3; the cell gasflow rate and rejection parameter q value were 0.4 mL min–1 and 0.7,respectively (17). The quantifica-tions were carried out using exter-nal calibration (the concentrationsof Fe, Cd, and Pb were computedby applying the correction factorfor their corresponding recoveries).The method was verified by com-paring the results with those of thevolatilization procedure (6). A com-parison of the results obtained bythe present procedure and thevolatilization procedure is shown inTable V. As can be seen, most ofthe elements are in the lower ng g–1

levels. Discrepancy in the concen-tration of Fe in Showa boric acid ispossibly due to the inhomogeneousdistribution of Fe in boric acid.

TABLE IVPercentage Recoveriesa of Impurities

With Change in Weight (g) of Showa Boric Acid (n=3)

Isotope Weight of Boric Acid in Solution (g)

0.1 0.5 1.055Mn 98.4 ± 0.8 98.0 ± 1.0 98.2 ± 1.056Fe 64.0 ± 0.9 60.5 ± 1.1 60.0 ± 1.258Ni 105.3 ± 2.3 104.2 ± 1.9 102.2 ± 1.359Co 88.1 ± 0.6 87.9 ± 0.6 87.4 ± 0.863Cu 96.5 ± 0.9 95.0 ± 0.9 94.8 ± 0.664Zn 91.0 ± 0.8 93.6 ± 0.7 94.3 ± 0.871Ga 91.0 ± 0.6 91.1 ± 0.6 90.8 ± 0.7114Cd 83.4 ± 0.5 82.3 ± 0.5 80.8 ± 0.5208Pb 79.7 ± 0.8 80.5 ± 0.6 80.0 + 0.6

a 30-mL solution, 2 mL 10% Triton X-114, 1 mL 10% 8-HQ, 50 ng of each analyte.Uncertainties are given as 1σ variation in measurements of three different experi-ments.

TABLE VComparison of Concentrations Obtained

by the Present Procedure and Volatilization Procedurea

Isotope Showa Boric Acid (ng g–1) Sigma Boric Acid (ng g–1)

CPE (n=3) Vol (n=2) CPE (n=3) Vol (n=2)55Mn 1.62 ± 0.10 2.05 ± 0.10 4.63 ± 0.93 4.80 ± 0.4856Fe 87.6 ± 26.0 14.8 ± 2.1 447 ± 53 427 ± 2858Ni 11.3 ± 0.6 10.8 ± 0.3 23.9 ± 2.9 19.5 ± 2.059Co 0.11 ± 0.05 0.04 ± 0.01 1.11 ± 0.23 0.75 ± 0.2263Cu 2.90 ± 0.32 3.45 ± 0.82 18.0 ± 1.2 22.3 ± 1.864Zn 2.30 ± 0.80 2.86 ± 0.55 760 ± 30 630 ± 3671Ga <LOD <LOD <LOD <LOD114Cd 0.10 ± 0.02 0.05 ± 0.01 0.14 ± 0.01 0.21 ± 0.06208Pb 3.56 ± 0.13 4.93 ± 0.09 1.72 ± 0.41 0.88 ± 0.20

a CPE = Cloud Point Extraction, Vol = Volatilization.Uncertainties are given as 1σ variation in measurements of three different experi-ments.

5

The detection limits were calcu-lated as the signal corresponding tothree times the standard deviationof repeated measurements (n = 5)of a process blank solution; theseare shown in Table VI. As can beseen, the limits of detection are inng g–1 levels. The only short-comingof the procedure is slightly elevatedprocess blank levels due to thehigher quantities of the complexingagent used. However, the back-ground levels can be decreased byusing ultra-pure reagents in accor-dance with clean lab protocols.

CONCLUSION

The utility of cloud point extrac-tion for the analysis of boric acidhas been demonstrated. The recov-eries are near quantitative for mostof the elements, even in the pres-ence of higher concentrations ofboric acid. The procedure reportedis simple, fast, easy to adopt, andcan be used as an alternative to theexisting procedures.

ACKNOWLEDGMENT

This research was supported bya grant from the National ScienceCouncil of the Republic of Chinaunder Contract NSC 95-2113-M-110-007.

Received November 14, 2007.

REFERENCES

1. http://en.wikipedia.org/wiki/Boric_acid

2. J. I. Kroschwitz, Encyclopedia ofChemical Technology, 4th ed.,Wiley, Vol. 17 (1996).

3. W. Gerhartz, Ulman’s Encyclopediaof Industrial Chemistry, 5th ed.,VCH, Vol. A-4 (1985).

4. H. Bernhard Pogge, Electronics Mate-rials Chemistry, Marcel Dekker(1986).

5. J. Sucharova and I. Suchara, Anal.Chim. Acta 576, 163 (2006).

6. K. Dash, D. Karunasagar, S.Thangavel, and S. C. Chaurasia, J.Chromatogr. A, 1002, 137 (2003).

7. J.-C. Landry, M.-F. Landry, and D.Monnier, Anal. Chim. Acta 62, 177(1972).

8. X. Zhu, X. Zhu, and B. Wang, J. Anal.At. Spectrom. 21, 69 (2006).

9. B. B. Chen, B. Hu, and M. He, RapidCommun. Mass Spectrom. 20,2894 (2006).

10. P. Wu, Y. Zhang, Y. Lv, and X. Hou,Spectrochim. Acta Part B, 61, 1310(2006).

11. Y. Li, B. Hu, and Z. Jiang, Anal.Chim. Acta 576, 207 (2006).

12. J. Nan, Y. Jiang, and X.-P.Yan, J.Anal. At. Spectrom. 18, 946 (2003).

13. A. C. Sahayam, S.-J. Jiang, and C.-C.Wan, J. Anal. At. Spectrom. 19, 407(2004).

14. R.-L. Ueng, A. C. Sahayam, S.-J.Jiang, and C.-C. Wan, J. Anal. At.Spectrom. 19, 681 (2004).

15. R. Carabias-Martinez, E. Rodriguez-Gonzalo, B. Moreno-Cordero, J. L.Perez-Pavon, C. Garcia-Pinto, andE. Fernandez Laespada, J. Chro-matogr. A, 902, 251 (2000).

16. S. Nakashima, R. E. Sturgeon, S. N.Willie, and S. S. Berman, Fresenius’J. Anal. Chem. 330, 592 (1988).

17. C.-F. Yeh, and S.-J. Jiang, J. Chro-matogr. A, 1029, 255 (2004).

Vol. 29(1), Jan./Feb. 2008

TABLE VIExperimental Blanks and Detection Limits (ng g–1)

Isotope Blank D.L. (n=5) (3 σ)

55Mn 0.40 0.356Fe 140 9.958Ni 27 2.959Co 0.26 0.1763Cu 5.9 6.064Zn 5.0 6.371Ga 0.075 0.04114Cd 0.84 0.66208Pb 0.3 0.79

D.L. = Detection Limit.

6Atomic SpectroscopyVol. 29(1), Jan./Feb. 2008

Synergic Solvent Extraction of Rare Earth ElementsUsing Mixed Ligand Complexes of

Hexafluoroacetylacetone and Tri-n-butylphosphate and Their Determination in Environmental Waters

by Low Temperature ETV-ICP-MSChaomei Xiong, *Bin Hu, and Zucheng Jiang

Department of Chemistry, Wuhan University, Wuhan 430072, P. R. China

INTRODUCTION

Rare earth elements (REEs) arewidely used in functional materials,catalysts, and fertilizer (1). Previousresearch has revealed that the useof fertilizers containing REEs havecontributed immensely to the effi-cient development and yields ofcrops, vegetables, and fruits.Hence, in view of these industrialand agricultural activities, more and

more REEs are continuouslyreleased into the environment andmay eventually affect living organ-isms (2,3). Therefore, monitoringtechniques for REEs are required.Of the various methods used forthe determination of REEs, induc-tively coupled plasma optical emission spectrometry/mass spec-trometry (ICP-OES/MS) are usuallythe most favorable choices, due totheir excellent analytical character-istics, including high sensitivity,accuracy, and simplicity of opera-tion (4,5).

In ICP-OES/MS, sample introduc-tion techniques play a key role notonly in improving the sensitivityand selectivity, but also in extend-ing the application of analyticalmethods. Therefore, great attentionhas been focused on sample intro-duction techniques in the field ofatomic spectrometry and massspectrometry. Conventional pneu-matic nebulization (PN) has the dis-advantages of low intake efficiency(less than 2% of aspirated solutionenters the plasma), high sampleconsumption, and difficulty in ana-lyzing samples with high salt con-tent and viscosity. Electrothermalvaporization (ETV), as a means ofsample introduction for ICP-OES/MS, has been demonstrated topossess the following distinct mer-its: low sample consumption, hightransport efficiency, low absolutedetection limit, and in situ removalof the organic matrix before sampleintroduction into the plasma,which leads to extensive applica-tion in analytical ICP-OES/MS (6).However, with direct vaporizationof refractory elements [REEs, andhigh field strength elements (HFSE)-Nb, Ta, Zr, Hf] by ETV methods,low sensitivity and serious memoryeffects are usually encountered dueto high boiling points or the forma-tion of refractory carbides. In orderto improve the introduction effi-ciency, conversion of the analytesto more volatile halogenatedspecies with halogenating agents,such as poly(tetrafluoroethylene)(PTFE), has been suggested (7,8).

In recent years, considerablyimportant developments have beenachieved in the exploration and

ABSTRACT

A synergic solvent extractionsystem has been developed forthe rapid extraction of rare earthelements (REEs) from aqueousmedia with cyclohexane contain-ing hexafluoroacetylacetone(HHFA) and tri-n-butylphosphate(TBP). The resulting mixed ligandcomplexes of REEs-HHFA-TBPwere introduced into a graphitefurnace for the determination ofREEs by low temperature elec-trothermal vaporization induc-tively coupled plasma massspectrometry (ETV-ICP-MS). Inthe absence of TBP or HHFA, theREEs were not efficientlyextracted with cyclohexane-HHFA or TBP, and also could notbe vaporized at low vaporizationtemperature, possibly becausethese metal chelates containcoordinated water. Various fac-tors influencing the extraction ofREEs, including the pH values ofthe solution, the type of extrac-tion solvent, and the phase ratio,have been systematically investi

*Corresponding author.E-mail: [email protected]:: 0086-27-68754067

gated. It was found that by usingHHFA and TBP as chelatingreagents, REE-HHFA-TBP com-plexes could be quantitativelyextracted into cyclohexane at pH6.0 with a phase ratio (aqueousphase to organic phase) of nomore than 10. The factors affect-ing the vaporization and trans-portation of analytes, such as theconcentration of HHFA and TBP,the drying temperature and time,and the vaporization tempera-ture, were also studied; the opti-mal experimental conditionswere established for ETV-ICP-MS.Under the optimized conditions,the detection limits of themethod for 15 REEs were in therange of 0.017 ng L–1 [Eu(III)] -0.277 ng L–1 [Nd(III)], and therelative standard deviations(RSDs) for the determination ofREEs at 0.05 ng mL–1 werebetween 4.5% [Ce(III)] and 9.9%[La(III)]. The proposed methodwas applied to the analysis ofenvironmental samples, and therecoveries for the spiked sampleswere in the range of 84–119%.

7

Vol. 29(1), Jan./Feb. 2008

application of organic chelatingreagents as chemical modifiers inETV-ICP-OES/MS. In contrast to thehalogenating agents, organic chelat-ing reagents used as chemical modi-fiers have given prominence tolower vaporization temperatures,higher introduction efficiencies,and lower memory effects becauseof the complexes’ thermal stabilityand volatility. Some organic chelat-ing reagents, including 8-hydrox-yquinoline (8-HQ) (9,10),1-phenyl-3-methyl-4-bonzoyl-5-pyra-zone (PMBP) (11), acetylacetonate(AcAc) (12), ammonium pyrrolidinedithiocarbamate (APDC) (13),sodium diethyldithiocarbamate(DDTC) (14), benzoylacetone(BZA) (15), thenoyltrifluoroacetone(TTA) (16), ethylene diaminetetraacetic acid (EDTA) (17), andpolyhydroxy compounds (8), havebeen used to improve the sensitiv-ity of determination and decreasethe vaporization temperature ofanalytes in graphite furnace.

Recently, analysts have focusedmore attention on the study of syn-ergic extraction of REEs and theextraction mechanism. Synergicsolvent extraction, as a liquid-liquidextraction mode, involves the appli-cation of a mixture of two or morechelating reagents to extract theanalyte of interest. The distributioncoefficient is prominently largerthan the summation of distributioncoefficients when each of theextraction reagents is usedseparately to extract the analyte ofinterest. At present, many binary,ternary, and even multinary syner-gic solvent extraction systems havebeen studied. Synergic solventextraction has some merits, includ-ing high extraction efficiency, shortextraction time, and good selectiv-ity (18,19).

The chelates of REEs and β-dike-tones were widely studied in thefield of gas chromatography andelectrothermal vaporizationbecause of their good thermal sta-bility and volatility. By using

thenoyltrifluoroacetone (TTA) as achelating reagent, Jacquelot et al.(20) developed a method for theseparation of Sc and Al at 270 oC bygas chromatography, Fan et al. (21)established a method to determineSc and Y in biological samples bylow temperature ETV-ICP-OES.Xiong et al. (22) developed a sensi-tive and rapid method for low tem-perature ETV-ICP-OES deter-mination of La and Eu after separa-tion/preconcentration by a micro-column packed with immobilized1-phenyl-3-methyl-4-bonzoyl-5-pyra-zone (PMBP) on microcrystallinenaphthalene. It should be notedthat the examples mentioned aboveare typically unitary extraction sys-tems. However, metal ions, with acoordination number greater thantwice the charge, will satisfy thecoordination requirement viahydrate formation, solvate forma-tion, or adduct formation whenreacting with bidentate chelatingagents. Hydrate formation generallyis accompanied by poor solventextraction efficiency, particularlywith halohydrocarbon solvents,while solvent formation or adductformation will permit or enhanceextractions. As a typical example,REEs can easily form hydrated com-plexes with bidentate agents,which are extracted with poor effi-ciency by organic extraction sol-vents because of hydrate formation.At the same time, the complexeshave less thermal stability and mayeasily decompose to oxides. Inorder to solve this problem, theaddition of a neutral donor, such astri-n-butylphosphate (TBP) (23), di-n-butylsulphoxide (DBSO) (24), tri-n-butylphosphine oxide (TBPO)(25), tri-n-octylphosphine oxide(TOPO) (26), dimethylsulfoxide(DMF) (27), benzoic acid (28),isobutylamine (IBA) (29), 2,2’-bipyridine (bpy), or 1,10-phenan-throline (phen) (30), to the organicextraction solvent system not onlygreatly increases the percentageextracted, but also forms mixed lig-and complexes that are thermally

stable. Some of the mixed ligandcomplexes have been analyzed bygas chromatography. Butts et al.(23) stated that the addition of TBPto the HHFA-extraction solvent sys-tem greatly increased the percent-age extraction. Burgett et al. (24)reported that DBSO provides a simi-lar synergic enhancement of theextraction of REEs with1,1,1,2,2,6,6,7,7,7-decafluoro-3,5-heptanedione [H(FHD)] and themixed ligand complexes of thecerium group lanthanides weresuccessfully analyzed by gas chro-matography.

Considering the above facts, itmay be stated that a synergic sol-vent extraction system can greatlyincrease the extraction efficiencyand strengthen the thermal stabilityof complexes. The aim of thispaper is to study the binary syner-gic solvent extraction system ofREEs with hexafluoroacetylacetone(HHFA) and tri-n-butylphosphate(TBP), and develop a new methodby combining synergic extractionwith low temperature ETV-ICP-MSfor the determination of REEs inenvironmental waters.

EXPERIMENTAL

Instrumental and Operating Conditions

A modified commercially WF-4Cgraphite furnace (Beijing SecondOptics, Beijing, P.R. China) wasused as an electrothermal vaporizerand was connected to an Agilent7500a ICP-MS (Japan). Details onthe modification of the graphitefurnace and its connection withICP-MS have been described previ-ously (15). The ICP-MS operatingconditions were optimized with aconventional pneumatic nebuliza-tion method prior to connectionwith the ETV device. Pyrolyticallycoated graphite tubes were usedthroughout this work. The operat-ing conditions for ETV-ICP-MS andthe temperature program are sum-marized in Table I.

8

An Intrepid XP Radial ICP-OES(Thermo, Waltham, MA, USA) witha concentric nebulizer and aCinnabar cyclonic spray chamberwas used for studying the synergicextraction; the operating conditionsfor ICP-OES are listed in Table II. AUV-2000 spectrophotometer (Shi-madzu, Japan) was used to studythe mechanism of vaporization. ThepH values were monitored with a320-S pH meter (Mettler ToledoInstruments Co. Ltd., Shanghai, P.R.China) supplied with a combinedelectrode. A SY1200 ultrasonic sys-tem (Shengyuan Ultrasonic Appara-

tus Equipment Co. Ltd., Shanghai,P.R. China) was used to acceleratethe extraction. An AD-72 centrifuge(The Tenth Factory of OperationAppliance, Shanghai, P.R. China)was used to for phase separation.

Chemicals and Reagents

The lanthanides and yttriumstock solutions (1.0 g L–1) were pre-pared by dissolving the appropriateamount of their oxides or nitrates(Shanghai Reagent Co. Ltd., Shang-hai, P.R. China) in 1:1 (v/v) HNO3

using the recommended method(31). A mixture of stock solution

containing 50 µg mL–1 of target ana-lytes in 2% (v/v) HNO3 wasprepared by diluting the formerstock solutions. All the oxides ornitrates used were of Specpure“grade. Hexafluoroacetylacetone(HHFA, 99%) was from AcrosOrganics, New Jersey, USA. Allother chemical reagents were ofanalytical grade. 0.1 mol L–1 sodiumhydroxide and 0.1 mol L–1 HNO3

were used to adjust pH. High-puritydeionized water was obtained froma Labconco system (18.2 MΩ·cm)and used throughout this work.

All laboratory ware wasthoroughly cleaned by soaking innitric acid 20% (v/v) for at least 24 h before use.

General Procedure

In order to study synergic extraction, 2-mL aliquots of the pH-adjusted rare element earth sam-ple solutions were placed in 5-mLcentrifuge tubes, and 0.2 mL cyclo-hexane, containing only 0.002 mol L–1

TBP, or 0.001 mol L–1 HHFA, orboth, was added. The tubes werethen stoppered and shaken. Aftercentrifugation, the residual aqueousmedia (lower layer phase) wasdirectly collected and determinedby ICP-OES. The quantity ofanalytes extracted by the extractionsolvent could be calculated by sub-tracting the determined results byICP-OES from the original concen-trations.

For low temperature ETV-ICP-MSdetermination of REEs, 0.1 mL ofcyclohexane, containing 0.1 molL–1 HHFA and 0.2 mol L–1 TBP, wasadded into 1 mL mixed REE solu-tions by microsyringe, and thenextracted by an ultrasonic systemfor 15 minutes. After centrifugation,10 µL of the resulting organic phasewas injected into the graphite fur-nace with a microsyringe for fur-ther ETV-ICP-MS analysis.

The blank consisted of 1 mLhigh-purity deionized water,adjusted to pH 6.0 by 0.1 mol L–1

TABLE IOperating Parameters of ETV-ICP-MS

ICP-MS PlasmaRF Power 1250 WOuter Gas (Ar) Flow Rate 15 L·min–1

Intermediate Gas (Ar) Flow Rate 0.9 L·min–1

Nebulizer Gas (Ar) Flow Rate 0.7 L·min–1

Sampling Depth 7.0 mmSampler/Skimmer Diameter Orifice Nickel 1.0 mm/0.4 mm

Time-resolved Data AcquisitionScanning Mode Peak-hoppingDwell Time 20 msIntegration Mode Peak AreaPoints Per Spectral Peak 1

Electrothermal VaporizerCarrier Gas Flow Rate 0.4 L·min–1

Drying 100 oC, Ramp 10 s, Hold 20 sVaporization 1400 oC, Hold 5 sCooling 50 oC, Hold 5 sCleaning 2500 oC, Hold 4 s

Sampling Volume 10 µL

TABLE IIOperating Parameters of Intrepid XP Radial ICP-OES

RF Generator Power 1150 WFrequency of RF Generator 27.12 MHzCoolant Gas (Ar) Flow Rate 14 L·min–1

Carrier Gas (Ar) Flow Rate 0.6 L·min–1

Auxiliary Gas (Ar) Flow Rate 0.5 L·min–1

Max Integration Times 10 s

Emission Line Er, 349.910 nm

9

Vol. 29(1), Jan./Feb. 2008

and Tm(III) in the post-extractionorganic phase was also studied overa pH range of 1.0 – 9.0 by ETV-ICP-MS; the experimental results areshown in Figure 2. It can be seenthat the pH had a similar effect onREEs because of their similar chemi-cal properties. As described above,the highest percentage extractionwas obtained and remained con-stant when the pH approached 5.5.Therefore, pH 6.0 was selected forthe synergic extraction of REEs.

Choice of Extraction SolventThe complexes have different

distribution coefficients in differentextraction solvents: the higher thedistribution coefficient, the moreanalytes in extraction solvent.Therefore, selecting a suitableextraction solvent is very impor-tant. The effects of some extractionsolvents, including chloroform,benzene, tetrachloride, toluene,and cyclohexane, on the extractionof target analytes have been stud-ied, and the results are shown inFigure 3. While keeping other con-ditions constant and changing theextraction solvent from chloroformto cyclohexane, the intensity of theREEs increased accordingly, whichmeans that the complex had ahigher distribution coefficient in

pH range tested, the percentageextraction was less than 50% if onlyHHFA or TBP was present in cyclo-hexane. But, when both HHFA andTBP were coexisting in cyclohexane,the percentage extraction increasedwith pH, and more than 95%extraction was achieved at pH 5.4and higher. It can be concludedthat TBP provided a synergicenhancement of extraction of REEswith HHFA. The possible reasonwas that HHFA formed hydratedchelates with REEs [as in reaction(1)] and the coordinated water inhydrated chelates resulted in poorextraction efficiency for REEs.When the neutral donor (TBP)coexisted, the coordinated water inthe hydrated chelates were substi-tuted by TBP [as in reaction (2)],and the hydrated chelates changedto dehydrated complexes thatcould be extracted with an organicsolvent. Hence, all the above-men-tioned factors have contributed tothe quantitative extraction of REEs.

REEs(III) + 3 HFA- + xH2O →REEs(HFA)3·x(H2O) (1)

REEs(HFA)3·xH2O + 2 TBP →REEs(HFA)3·2 TBP + xH2O (2)

The effect of pH on the synergicextraction of La(III), Eu(III), Er(III),

NaOH and 0.1 mol L–1 HNO3; wasthen subjected to the same extrac-tion procedure as described above.The determined values for REEswere obtained after subtracting theblank values. Calibration standardswere also subjected to the sameextraction procedure.

Sample Preparation

Environmental water samples,such as lake water, river water andtap water, were collected, filteredthrough 0.45-µm membrane filters(Tianjin Jinteng Instrument Factory,Tianjin, P.R. China) and acidified.The pH values of the samples wereadjusted to 6.0 before extractionand then subjected to the extrac-tion procedure described above.


Optimization of SynergicSolvent Extraction System

Effect of pH on Synergic SolventExtraction

The pH of the solution is one ofthe most important factors thataffects the formation and extractionof the chelate in all parameters.Figure 1 shows the percentageextraction of Er(III) in the pH rangeof 0.5–7.0. As can be seen, over the

Fig. 1. Effect of pH on the percentage extraction of Er(III).Conditions: aqueous phase, Er(III), 0.5 µg mL–1, 2 mL;organic phase, HHFA, 0.001 mol L-1, TBP, 0.002 mol L–1, 0.2 mL; the operating conditions for ICP-OES are listed inTable II.

Fig. 2. Effect of pH on the percentage extraction of La(III),Eu(III), Er(III), and Tm(III). Conditions: aqueous phase,concentration of analytes, 0.1 ng mL–1, 1 mL; organic phase,HHFA, 0.1 mol L–1, TBP, 0.2 mol L–1, 0.1 mL; the operatingparameters for ETV-ICP-MS are listed in Table I.

10

cyclohexane. A comparison of thecharacteristics of cyclohexane withother extraction solvents is shownin Table III. As can be seen, thecharacteristics of cyclohexane,including the dielectric constant,the dipole moment, and solubilityin water, are more amenable toextract the mixed ligand complexesinto organic phase than the othersolvents. Hence, cyclohexane waschosen as the extraction solventbecause of its greater synergiceffect.

Study of Phase RatioThe effect of phase ratio on the

extraction of target analytes hasbeen studied. A 0.1-mL aliquot of

cyclohexane containing 0.1 mol L–1

HHFA and 0.2 mol L–1 TBP wasused to extract 0.1, 0.25, 0.5, 1, 2,3, 4, and 5 mL of solution contain-ing 0.05 ng mL–1 REEs. The resultsindicate that when the volume ofaqueous phase is not more than 10times that of the organic phase, thesignal intensity of the analytesremained constant and thendecreased with an increase inphase ratio.

Optimization of ETV-ICP-MS

Vaporization Behaviors of MixedLigand Complexes

Butts et al. (23) reported that themixed ligand complexes of REEswith HHFA and TBP can be

analyzed by gas chromatography,which implied that the mixed lig-and complexes of REEs with HHFAand TBP could be vaporized by lowtemperature electrothermal vapor-ization. Figure 4 shows the signalprofiles of Nd(III) at 1400 oC in dif-ferent extraction systems. A strongand sharp signal profile of Nd(III) isfound when the organic extractionphase consisted of both HHFA andTBP. However, no signal wasobtained in the absence of HHFAand TBP in cyclohexane under thesame conditions. Figure 5 alsoshows that the analytes extractedby both HHFA and TBP could bequantitatively vaporized at a lowtemperature (1400 oC) and with nomemory effects at a high vaporiza-tion temperature [2500 oC (Figure5 c)]. No signal was observed at thelow temperature of 1400 oC whenonly HHFA was used as the chelat-ing reagent in cyclohexane (Figure5 b). At the same time, no signalwas obtained when the blank solu-tion was extracted by both HHFAand TBP (Figure 5 a). In view of theabove facts, it may be stated thatthe analytes were vaporized in the

Fig. 3. Effect of the extraction solvents on the signal inten-sity of Er(III). Conditions: aqueous phase, Er(III), 0.1 ng mL–1, 1 mL; organic phase, HHFA, 0.1 mol L–1, TBP,0.2 mol L–1, 0.1 mL; extraction solvent, a, HCCl3; b, benzene;c, CCl4; d, toluene; e, cyclohexane; the operating parametersfor ETV-ICP-MS are listed in Table I.

Fig. 4. Signal profile of Nd(III) in ETV-ICP-MS with HHFAand TBP as chemical modifier by ETV-ICP-MS. Conditions:aqueous phase, pH 6.0; extraction solvent, cyclohexane; a,10 pg Nd(III) extracted only by cyclohexane; b, 10 pgNd(III) extracted only by 0.2 mol L–1 TBP; c, 10 pg Nd(III)extracted only by 0.1 mol L–1 HHFA; d, 10 pg Nd(III)extracted by both of 0.1 mol L–1 HHFA and 0.2 mol L–1 TBP;e, blank solution extracted by both of 0.1 mol L-1 HHFA and0.2 mol L–1 TBP; f, memory effect of ‘d’ at cleaning tempera-ture 2500 oC; the operating parameters for ETV-ICP-MS arelisted in Table I.

TABLE IIIComparison of Dielectric Constant, Dipole Moment,

and Solubility of Extraction Solvents

Extraction Solvent CHCl3 Benzene CCl4 Toluene Cyclo- hexane

Dielectric Constant 4.81 2.28 2.24 2.38 2.02

Dipole Moment 1.03 0 0 0.37 0

Solubility in Water (%) 0.822 0.180 0.077 0.0627 0.006

11

Vol. 29(1), Jan./Feb. 2008

Fig. 5. Typical signal profiles of 10 pg Y(III), La(III), Gd(III) and Lu(III) by ETV-ICP-MS. Conditions: aqueous phase, pH 6.0;extraction solvent, cyclohexane; a, blank solution extracted by 0.1 mol L–1 HHFA and 0.2 mol L–1 TBP; b, extracted only by 0.1 mol L–1 HHFA; c, extracted by both of 0.1 mol L–1 HHFA and 0.2 mol L–1 TBP; the operating parameters for ETV-ICP-MS arelisted in Table I.

12

form of mixed ligand complexes.The phenomenon was differentfrom the other ETV systems usingβ-diketone as the chemical modifier(12,15,16). The possible reason forthis is that, in the absence of TBP,hydrated chelates decompose toform oxides by losing coordinatedwater in the process of heating; theresulting oxides cannot be vapor-ized at low temperatures. WhenTBP coexists with HHFA in cyclo-hexane (as a strong neutral donor),TBP substituted the coordinatedwater in chelates and formed dehy-drated complexes which preventedthe decomposition of thecomplexes.

Effect of Drying Temperature andTime on Signal Intensity

The effect of drying temperaturesfrom 40 oC to 100 oC was studied.The experimental results indicatethat there is no notable effect ofdrying temperature on the signalintensity. Therefore, 100 oC waschosen as the drying temperature.

Using 100 oC as the drying tem-perature, the effect of drying timewas studied. The results showedthat a drying time longer than 40 swould decrease the signal intensity,so in this work, the drying time of20 s was selected.

Effect of Vaporization Tempera-ture on Signal Intensity

The effect of vaporization tem-perature on signal intensity wasstudied. The results are shown inFigure 6. As can be seen, when thevaporization temperature changedfrom 200 oC to 2000 oC, the inten-sity of analytes increased andreached maximum at 1300 oC,remaining constant as the tempera-ture increased further. The resultsindicate that the REEs were vapor-ized into the plasma in the form ofsome volatile complexes. In thisstudy, 1400 oC was chosen as thevaporization temperature.

Optimization of Concentration ofHHFA

Figure 1 indicates that whenthe pH values of the solution arehigher than 5.5 and concentrationsof HHFA and TBP are 0.001 and0.002 mol L–1 in cyclohexane,respectively, the analytes aredetected in the residual liquidphase, which means that the ana-lytes in the liquid phase were com-pletely extracted by the mixedorganic phase. However, whenETV-ICP-MS was used to determinethe analytes in the post-extractionorganic phase, no signal wasobserved at the vaporization tem-perature of 1400 oC. It was alsofound that with an increase in con-centration of HHFA and TBP, aweak intensity was observed atthe vaporization temperature of1400 oC by ETV-ICP-MS. This phe-nomenon showed that excessivereagents should be used to preventthe decomposition of mixed ligandcomplexes during vaporization inETV and to transport the analytesinto the plasma (16). Therefore,optimization of the concentrationof the chelating reagent is neces-sary.

Keeping other conditions con-stant, the effect of HHFA concentra-tion from 0.002 to 0.2 mol L–1 onthe intensity of the analyte wasstudied by ETV-ICP-MS. It wasfound that the intensity of the tar-get analyte increases with the con-centration of HHFA, and reachesmaximum at 0.04 mol L–1 HHFA incyclohexane containing 0.2 mol L–1

TBP. The signal remained constantat higher concentrations. Theexperimental results indicate thatusing excessive chelating reagent isvery important for the sample intro-duction technique of low tempera-ture electrothermal vaporization.Therefore, 0.1 mol L–1 HHFA wasused in this work.

Optimization of Concentration ofTBP

Keeping the concentration ofHHFA constant at 0.1 mol L–1 ,along with other parameters, theeffect of TBP concentration from0.017 to 0.3 mol L–1 on the inten-sity of the analyte was studied. Itwas found that the intensity of theanalyte increases along with anincrease in TBP concentration, andattaines maximum at 0.15 mol L–1,remaining constant at higher con-

Fig. 6. Effect of the vaporization temperature on the signal intensity in ETV-ICP-MS. Conditions: aqueous phase, La(III), Eu(III), Er(III) and Tm(III), 0.1 ng mL–1, pH 6.0; organic phase, 0.1 mol L–1 HHFA and 0.2 mol L–1 TBP, 0.1 mL; the operating parameters except for vaporization temperature for ETV-ICP-MS arelisted in Table I.

13

Vol. 29(1), Jan./Feb. 2008

centrations. Excessive TBP substi-tutes for the H2O in the chelate andforms thermally stable mixed ligandcomplexes. To prevent the decom-position of the complex., 0.2 mol L–1

TBP was used in future work.

Proof of Formation and Vapor-ization of Mixed Ligand Com-plexes of REEs-HHFA-TBP

In order to verify the formationand vaporization of the mixed lig-and complexes of REEs-HHFA-TBP,the following experiments weredesigned: A 100-µg mL–1 La(III),Eu(III), Er(III), or Tm(III) solution atpH 6.0 was added into an equal vol-ume of cyclohexane solution con-taining HHFA and TBP in a ratio of0.1:0.2 mol L–1, respectively. Afterextraction and centrifugation, theETV heating program was run 10times with 10-µL injections of theresulting organic solution. Thesample vapor produced duringvaporization was collected in

cyclohexane and analyzed with aUV-VIS spectrophotometer. Theexperimental results are shown inFigure 7. It can be seen that theobtained absorption spectrum ofthe collected solution was very sim-ilar to that of standard metal-HHFA-TBP complexes in cyclohexane thatwere prepared according to refer-ence 23. The above experimentalresults indicate that the REEs wereindeed vaporized and transportedas mixed ligand complexes fromthe graphite furnace to the plasma.In accordance with reference 23,the mixed ligand complexes ofREEs-HHFA-TPB might beREEs(HFA)3·2TBP.

Interferences

Various metal ions were addedindividually into a tube containing0.05 ng mL–1 REEs, and thenadjusted to the desired pH of 6.0with 0.1 mol L–1 HNO3 and NaOH.The tube was then subjected to the

procedure described above, andthe effects of coexisting ions on theintensity of the REEs were investi-gated. Table IV shows the concen-trations of interfering ions whenthe error of determination was lessthan ±10%.

Fig. 7. UV spectra of La-, Eu-, Er-, and Tm-HHFA-TBP chelates in cyclohexane. (1) HHFA and TBP; (2) La-HHFA-TBP; (3) Eu-HHFA-TBP; (4) Er-HHFA-TBP; (5) Tm-HHFA-TBP; a-e, standard solution; a’-e’, sample vapor collected in cyclohexane.

TABLE IVTolerance Concentrations

of Coexisting Ionsa

Ion Tolerance Limit(µg mL–1)

K(I), Na(I) 4000Ca(II), Mg(II) 400Fe(III) 50Al(III), Cu(II), Zn(II) 10Mn(II), Ti(IV), Zr(IV) 5Cl- 7000NO3

– 6000

SO42– 1600

a REEs: 0.05 ng mL–1

14

Analytical Performance

In accordance with IUPAC rec-ommendations, the limits of detec-tion (LODs) of the methods werecalculated as three times the stan-dard deviation of the reagent blanksignal. The detection limits (3σ),absolute detection limit and preci-sion (0.05 ng mL–1, n = 9) are listedin Table V.

Application

The proposed method wasapplied to the determination ofultratrace REEs in environmentalwater, and the analytical resultsalong with the recoveries for thespiked samples are listed in TableVI. As can be seen, the REE concen-trations were in the range of 0.52(Lu) – 412.9 (Y) ng L–1, 0.61 (Lu) –305.9 (Y) ng L–1, and 0.48 (Lu) –270.1 (Y) ng L–1 for lake water,river water and tap water, respec-tively. Because there is no certifiedreference material of environmentalwater available in our laboratory,the spiking test was carried in thiswork to validate the method. Ascan be seen in Table VI, the recov-eries for the spiked samples werein the range of 84 – 119%.

CONCLUSION

A simple and sensitive methodhas been developed for the deter-mination of trace amounts of REEsin environmental water samples bycombining synergic solvent extrac-tion with low temperature ETV-ICP-MS. The experimental results haveshown that the addition of a neutraldonor-TBP to the HHFA-cyclohexanesystem not only increased the extrac-tion efficiency, but also improvedthe vaporization behavior of REEsat the low temperature of 1400 oC.The study of vaporization mecha-nism has shown that the REEs werevaporized and transported into theICP as their mixed ligand chelates.The developed method has beenapplied for the determination ofREEs in environmental waters withsatisfactory results.

ACKNOWLEDGMENT

Financial supports from ScienceFund for Creative Research Groupof NSFC (No. 20621502) and NCET-04-0658, MOE of China is gratefullyacknowledged.

Received November 19, 2007.

REFERENCES

1. X.K. Xu, Z.J. Wang, Y.S. Wang, and I.Kazuyuki, Aust. J. Soil Res. 41, 741(2003).

2. B. Wen, D.A Yuan, X.Q. Shan, F.L. Li,and S.Z. Zhang, Chem. SpeciationBioavail. 13, 39 (2001).

3. S.L. Tong, W.Z. Zhu, Z.H. Gao, Y.X.Meng, R.L. Peng, and G.C. Lu, J.Environ. Sci. Health Part A-Toxic/Hazard. Subst. Environ. Eng.39, 2517 (2004).

4. X. Cao, G. Zhao, M. Yin, and J. Li,Analyst 123, 1115 (1998).

5. J.S. Suleiman, C.Z. Huang, X.L. Pu, N.Zhang, and B. Hu, At. Spectrosc.27(6), 200 (2006).

6. Z.C. Jiang, L.Q. Tian, X.K. Chen, andB. Hu, "Modern atomic emission

spectrometry analysis", SciencePublishing Company, Beijing,China (1999).

7. G.Q. Xiang, Z.C. Jiang, M. He, and B.Hu, Spectrochim. Acta B 60, 1342(2005).

8. B. Hu, Z.C. Jiang, T.Y. Peng, and Y.C.Qin, Talanta 49, 357 (1999).

9. S. Tao, and T. Kumamaru, J. Anal. At.Spectrom. 11, 111 (1996).

10. X.S. Zhu, H. Bin, Z.C. Jiang, Y.L.Wu, and S. Xiong, Anal. Chim. Acta471, 121 (2002).

11. T.Y. Peng, and Z.C. Jiang, J. Anal.At. Spectrom. 13, 75 (1998).

12. Y.L. Wu, B. Hu, T.Y. Peng, and Z.C.Jiang, Anal. Chim. Acta 439, 153(2001).

13. L.B. Xia, B. Hu, Z.C. Jiang, Y.L. Wu,R. Chen, and L. Li, J. Anal. At. Spec-trom. 21, 362 (2006).

14. Z.F. Fan, B. Hu, Z.C. Jiang, and F.Yang, Anal. Chim. Acta 510, 45(2004).

15. L.B. Xia, B. Hu, Z.C. Jiang, Y.L. Wu,and Y. Liang, Anal. Chem. 76, 2910(2004).

16. Z.F. Fan, B. Hu, and Z.C. Jiang,Spectrochim. Acta B 60, 65 (2005).

TABLE VAnalytical Performances for ETV-ICP-MS

Element Isotope Rel. DLa Abs. DLb RSDc

(ng L–1) (fg) (%)

Y 89 0.089 0.89 7.3La 139 0.096 0.96 9.9Ce 140 0.070 0.70 4.5Pr 141 0.079 0.79 6.1Nd 146 0.277 2.77 6.7Sm 147 0.052 0.52 7.7Eu 151 0.017 0.17 9.0Gd 157 0.113 1.13 6.8Tb 159 0.049 0.49 5.6Dy 163 0.157 1.57 6.6Ho 165 0.047 0.47 7.4Er 166 0.040 0.40 4.9Tm 169 0.081 0.81 6.3Yb 172 0.019 0.19 7.0

Lu 175 0.044 0.44 8.1

a Relative detection limit; b Absolute detection limit; c REEs, 0.05 ng mL–1, n = 9.

15

Vol. 29(1), Jan./Feb. 2008

17. H.H. Lu, and S.J. Jiang, Anal. Chim.Acta 429, 247 (2001).

18. H. Imura, M. Ebisawa, M. Kato, andK. Ohashi, J. Alloy. Compd. 408-412, 952 (2006).

19. J. Yin, B. Hu, M. He, and Z.C. Jiang,Anal. Chim. Acta 594, 61 (2007).

20. P. Jacquelot, and G. Thomas, J.Chromatogr. 66, 121 (1972).

21. Z.F. Fan, B. Hu, Z.C. Jiang, and S.Q.Li, Anal. Bioanal. Chem. 378, 456(2004).

22. H.C. Xiong, B. Hu, T.Y. Peng, S.Z.Chem, and Z.C. Jiang, Anal. Sci. 15,737 (1999).

23. W.C. Butts, and C.V. Banks, Anal.Chem. 42, 133 (1970).

24. C.A. Burgett, and J.S. Fritz, Talanta20, 363 (1973).

25. R.F. Sieck, and C.V. Banks, Anal.Chem. 44, 2307 (1972).

26. K.S.R. Murthy, R.J. Krupadam, andY. Anjaneyulu, J. Chromatogr. Sci.36, 595 (1998).

27. M.F. Richardson, and R.E. Sievers,Inorg. Chem. 10, 498 (1971).

28. I. Matsubayashi, E. Ishiwata, T.Shionoya, and Y. Hasegawa,Talanta 63, 625 (2004).

29. W.G. Scribner, and A.M. Kotecki,Anal. Chem. 37, 1304 (1965).

30. S. Nakamura, S. Takei, and K. Akiba,Anal. Sci. 18, 319 (2002).

31. Handbook of Analytical Chemistry,Chemical Industry Press, Beijing, P.R. China 2, 163 (1982).

TABLE VIAnalytical Results of REEs in Environmental Water Samples (Mean ± SD, n=3)

Element Lake Water Yangtze River Water Tap WaterAdded Found Recovery Found Recovery Found Recovery (ng L–1) (ng L–1) (%) (ng L–1) (%) (ng L–1) (%)

Y 0 412.9±2.2 305.9±12.5 270.1±12.750.0 472.0±27.6 118 359.7±21.2 108 328.7±5.9 117

La 0 101.5±8.0 69.3±8.2 66.4±5.650.0 161.0±11.4 119 111.3±9.2 84 114.5±2.8 96

Ce 0 74.0±3.3 73.6±2.4 86.1±5.650.0 127.9±8.3 108 121.0±5.1 95 130.3±9.0 88

Pr 0 6.6±1.4 9.1±0.6 4.2±0.650.0 53.6±2.6 94 65.0±4.3 112 60.5±6.3 113

Nd 0 206.1±23.0 45.7±6.3 67.6±5.750.0 264.0±38.9 116 98.0±8.6 105 114.1±3.6 93

Sm 0 12.7±1.1 20.8±1.0 19.0±1.850.0 60.6±4.5 96 72.6±2.0 104 69.0±3.5 100

Eu 0 4.6±0.3 10.0±0.4 12.0±0.950.0 58.8±4.1 108 56.9±3.2 94 59.6±3.0 95

Gd 0 14.3±0.9 21.3±3.1 8.3±0.350.0 59.2±0.9 90 67.1±2.7 92 64.3±2.5 112

Tb 0 17.1±0.9 23.2±3.1 11.4±0.650.0 65.3±7.7 96 73.6±8.1 101 60.3±1.6 98

Dy 0 24.6±3.6 18.9±2.5 8.6±0.850.0 74.2±3.7 99 66.3±4.6 95 61.3±1.9 105

Ho 0 10.3±0.9 8.2±0.3 12.3±0.250.0 60.3±2.8 100 55.3±8.7 94 65.7±2.5 107

Er 0 12.3±0.6 26.9±3.6 13.5±0.650.0 58.1±3.8 92 71.7±8.5 90 61.7±2.9 96

Tm 0 5.4±0.5 10.5±1.8 4.9±0.550.0 54.4±4.0 98 55.9±0.6 91 53.9±4.8 98

Yb 0 11.4±0.1 8.3±0.6 3.0±0.150.0 56.5±7.1 90 59.8±0.7 103 49.4±1.7 93

Lu 0 0.52±0.03 0.61±0.05 0.48±0.05

50.0 52.6±4.1 104 50.3±4.4 99 53.6±3.3 106


*Corresponding author.E-mail address:[email protected] Present address:Department of Chemistry, Biosystem Research Complex,Clemson University, Clemson, S.C, 2963-1905, USA

Line Selection and Evaluation of Radio Frequency Glow Discharge Atomic Emission Spectrometry

in the Determination of Trace Amounts of Niobium in Electron Beam-melted Tantalum Sheet*G. Anil, M.R.P. Reddy, S. Rajesh Kumar, N.R. Munirathnam, and T.L. Prakash Analytical Division, Centre for Materials for Electronics Technology (C-MET),

IDA Phase-III, HCL Post, Cherlapalli, Hyderabad - 500 051, India

ABSTRACT

Trace niobium (Nb) wasdetermined in an electron beam-melted (EBM) tantalum (Ta)sheet by a radio frequency glowdischarge optical emission spec-trometry (RF-GD-OES) methodusing argon (Ar) as the dischargegas. The discharge operatingparameters were optimizedbased on both the raw analytesignal intensity (S) and the signal-to-background ratio (S/B). Amethodical approach was used toselect a Ta interference-free lineof Nb for the analysis of trace Nbin a Ta sheet based on signal-to-noise (S/N) measurements with aModel Jobin-Yvon high-resolution1 m JY-10,000 RF monochroma-tor having a 2400/2 grooves/mmholographic grating.

The analytical lines selectedwere used to evaluate the degreeof line interference and limits ofdetection. Nb determination inan EBM Ta sheet was performedto estimate the precision andaccuracy of the analytical result.The relative standard deviation(RSD) of the intensity for the Nbline was around 3% for withinruns and around 4% for betweenruns which represents the overallanalytical performance of thistechnique on the precision; theaccuracy was compared withICP-OES analysis and was satisfac-tory. The detection limits for Nbin the EBM Ta sheet wasestimated to be 2 µg/g. Thedetails of sensitive line selectionand sputtering parameters arepresented and discussed.

INTRODUCTION

C-MET Hyderabad is activelyinvolved in the preparation of highvoltage capacitor-grade tantalumpowder through electron-beammelting (EBM) and hydriding/dehy-driding routes. Tantalum is usedmainly as a corrosion-resistantmetal in the chemical industries, ashigh temperature heating elements,and in electrolytic capacitors aselectrodes, which are covered byanodic oxidation with Ta as thedielectric. The latter application issignificant, since it has better volu-metric efficiency. The trace Nbalters the electrical characteristicsof the Ta capacitors. The amount ofNb in the Ta powder used for theTa capacitors (1) determines thesize, type, and voltage of the Tacapacitors that one desires to makeand the sintering temperature used.Low voltage capacitor (< 10 V)powder contains a maximum of50 ppm Nb; for high voltage capaci-tors (35 V), the Nb amount shouldbe < 10 ppm. The control of com-position within semi-finished andfinished products is an issue ofmajor importance for producersand component fabricators alike,especially in materials destined forhigh performance applications.Considerable time and effort isexpended in ensuring that thematerials fall within strict composi-

tional limits. The Glow DischargeSpectrometer (GDS) lamp providesa low-pressure argon environment(typically 5-10 Torr) over the sam-ple surface. A high negative poten-tial (2-3) (typically –800 to –1200 V)is applied to the sample; thus, thesample becomes the cathode.Spontaneously produced argon ions(Ar) are accelerated across theanode/cathode gap. These Ar ionscollide with argon gas moleculescausing plasma formation and fur-ther production of Ar ions which iscalled glow discharge plasma. Someof these high velocity Ar ions reachthe sample surface where theysputter (or mill out) materials uni-formly from the sample substrate.This sputtered material diffuses intothe glow discharge plasma where itis dissociated into atomic particlesand finally excited. The light emit-ted from these excited state species(as they collapse back to lowerenergy levels) is characteristic ofthe elements composing the sam-ple. The wavelengths and intensityof the light emission are used toidentify and quantify the composi-tion of the sample.

The aim of this paper was toselect a Ta interference-free line ofNb, since to our knowledge therehas been no definitive and quantita-tive evaluation reported in the liter-ature of the analytical characteristicsof radio frequency glow dischargeoptical emission spectrometry (RF-GD-OES). In addition, the key para-meters of power and pressure wereevaluated in order to obtain betterreproducibility both for the sputter-ing rates and for the emission inten-sities of the selected interference-free spectral lines of Nb.

17

Vol. 29(1), Jan./Feb. 2008

EXPERIMENTAL

Instrumentation

All measurements wereperformed using a Model JY 10,000RF 1.0 m monochromator (Jobin-Yvon, France). A standard Marcus-Grim source (stainless steel) wasused with a 2400/2 grooves/mmholographic grating. The power tothe plasma was supplied by a radiofrequency generator at 13.56 MHz.Cathodic sputtering was used foroptimizing the glow dischargeanalysis. The total analysis time wasoptimized for the EBM Ta sheetover a period of 3 minutes. On initi-ating a discharge, the burning volt-age is normally high in order tobreak through the layer of oxidesand the gases adsorbed at the elec-trode surface; hence, the plasmaattained stability in 30 seconds;after 30 seconds, the plasmareached sputtering equilibrium andsteady state for the next 150 sec-onds, for the matrix element Ta(emission line of Ta: 362.662 nm)for a Ta sheet at 60 W power and7.5 Torr pressure. Figure 1 depictsa sputtered Ta sheet. The highpower required for a Ta sheet isdue to its high hardness: RockwellB hardness 35 (4). A 5, 10, and 15%Nb2O5 mixed carbon pellet wasprepared for line identification ofNb, the power was set at 20 W, andthe pressure at 6 Torr since the pel-let was soft in nature. The Jobin

Yvon (France) software (QuantumIQ-2.22) was employed for control-ling the data acquisition parame-ters. High-purity argon (99.999%)was used as the discharge gas. Alimiting anode orifice diameter of4 mm was used for all analyses. ATa sheet was polished on siliconcarbide abrasive paper. The stan-dardized instrumental parametersfor the determination of trace Nb inan EBM Ta sheet are given in Table I.


Implementation of RF-GD-OESanalysis is dependent on the selec-tion of the most optimum set ofelemental transitions. Since thetransitional element Ta containsnumerous atomic and ionic transi-tions, it is often difficult to choose amatrix interference-free line thatwould optimize the figures of meritfor a given technique. There is adefinitive need for the selection ofa line that has no interference fromthe matrix element Ta. A systematicapproach was followed with a 5,10, and 15 % Nb2O5 mixed carbonpellet for the line identification ofNb. The three sensitive linesselected for Nb were 309.418 nm,316.340 nm, and 319.498 nm (seeFigures 2 a–f). These Nb lines fromthe EBM Ta sheet were overlappedwith the Nb lines from the Nb2O5

doped graphite pellet. Out of thethree lines selected, the 309.418-nm

and 319.498-nm Nb lines have bothleft and right overlap of the matrixline, whereas the 316.340-nm linehas only a left overlap. Since the316.340–nm line had least lineinterference (5), it was selected forthe analysis.

RF-GD-OES analysis wasperformed by fixing two key para-meters: operating pressure andapplied generator output power.The delivered power may differconsiderably from the real powercoupled to the discharge owing topower losses. A change in the pres-sure modifies the impedance of theRF glow space between the RFelectrode and the surfacegrounded, and so could producechanges in the percentage of the RFpower coupled to the discharge.Moreover, the same power can beobtained with different values ofvoltage and current intensity. It isnecessary to measure other parame-ters like voltage in order to knowthe electrical conditions in the dis-charge. It is well established that inGD-OES it can be assumed that thesputtering (6–9) and excitationprocesses are independent. Bychanging the power and pressure,the emission intensities of theselected analytical line were investi-gated at the discharge operatingconditions. The results show thatthe EBM Ta sheet sputtered at theRF power of 60 W and gives maxi-mum intensity. At higher RF power,

Fig. 1. RF-GD-OES sputtered Ta sheet.

TABLE IRF-GD-OES Operating Conditions for EBM Ta Sheet

Spectrometer Jobin-Yvon JY-10, 000 RFGrating Holographic 2400/2 grooves/mmSpectral Range 120-800 nmPower 60 WPressure (Ar) 7.5 TorrArgon Flow Rate 0.25 L/minRinse Rime 10 sPre-integration 40 sAnalysis Duration 60 s

Wavelength 316.340 nm

18

Fig 2 (a). Interference spectra. Nb (309-nm line) peak in EBMTa Sheet.

Fig 2 (d). Interference spectra. Nb (316-nm line) peak in EBMTa Sheet overlapped with Nb in Nb2O5 doped graphite pellet..

Fig 2 (c). Interference spectra. Nb (316-nm line) peak in EBMTa Sheet.

Fig 2 (b): Interference spectra. Nb (309-nm line) peak in EBMTa Sheet overlapped with Nb in Nb2O5 doped graphite pellet.

19

the intensity starts decreasing,while small changes in pressurealter the relative emission yields toa large extent. As mentioned above,changes in pressure give rise tonoticeable modifications in theimpedance of the system. The pres-sure was optimized at 7.5 Torr assuggested by Bengston et al. (9).

The determination of trace Nbusing RF-GD-OES was carried outusing the 316.340-nm line. Thestandard Ta sheet procured fromNFC, Hyderabad, contained 60 ppmNb. These results were verified

using ICP-OES analysis (10–11) andfound to be in agreement. Threedifferent EBM Ta sheets were ana-lyzed using GD-OES and the preci-sion was evaluated. Within-runprecision on the method was evalu-ated on three samples testedrepeatedly for 10 times and theRSD was around 3%. Between-runprecision was also carried out sinceit is the best indicator of a method'soverall precision. For day-to-dayanalysis (20 days) the RSD wasaround 4%. For measuring the accu-racy of the RF-GD-OES method, the

results were compared with ICP-OES analysis. The results of boththe solid and liquid OES measure-ments were found to be in agree-ment and the results are given inTable II. Due to the absence of atrue blank in solid analysis, the limitof detection (LOD) was generatedusing the relative standard devia-tion of the background (RSDB)method of calculation proposed byBoumans and co-workers (12). TheLOD calculated for Nb in a Ta sheetwas found to be 2 µg/g.

CONCLUSION

The study of radio frequencyglow discharge optical emissionspectrometry (RF-GD-OES) for thedetermination of trace Nb in arefractive material such as tantalumhas demonstrated that it is wellsuited to meet the analyticalrequirements. The limit of detec-tion obtained was 2 µg/g. The GD-OES method developed seems to bethe most powerful analytical tool toanalyze Nb concentrations in a Ta

Vol. 29(1), Jan./Feb. 2008

Fig 2 (f). Interference spectra. Nb (319-nm line) peak inEBM Ta Sheet overlapped with Nb in Nb2O5 doped graphitepellet.

Fig 2 (e). Interference spectra. Nb (319-nm line) peak inEBM Ta Sheet.

TABLE IIComparison of RF-GD-OES and ICP-OES Results (ppm)

for Trace Nb Determination in Electron Beam-melted Ta Sheet

Sample Precision %RSD Accuracy

Electron Beam- Within Run Between Run Measured Value ICP-OESmelted Ta Sheet (%) (%) (RF-GD-OES) Analyzed

Results

Sheet-1 3.25 4.35 32.0±0.2 33.0Sheet-2 3.40 3.80 37.0±0.3 38.0

Sheet-3 3.10 4.20 40.0±0.3 41.0

20

sheet. In less than a minute, theanalysis can be carried with goodaccuracy and sensitivity. The use ofthis method also enables a quickquality check at every step of theanalytical process. The standardizedparameters of the methoddescribed, such as power and pres-sure, can also be used for the deter-mination of other trace elements ina Ta sheet.

ACKNOWLEDGMENTS

The authors are thankful to theExecutive Director, C-MET, for hisencouragement during this work.

Received July 25, 2007.

REFERENCES

1. G.S. Upadhyaya, Bull. Mater. Sci.28(4), 305 (2005)

2. Michael R. Winchester and RichardPayling, Spectrochim. Acta Part B59, 607 (2004).

3. G. Anil, M.R.P. Reddy, D.S. Prasad,S.T. Ali, N.R. Munirathnam, andT.L. Prakash, Materials Characteri-zation 58, 92 (2007).

4. E Armantrout, U.S. Patent 3,249,429(1966).

5.A. Bengtson, A. Eklund, and F.Präßler, Fresenius J Anal Chem.335, 836 (1996).

6. Covadonga Perez, Rosario Pereiro,Nerea Bordel, and Alfredo Sanz-Medel, J. Anal. At. Spectrom. 15,67 (2000).

7. R. Payling, T. Nelis, M. Aeberhard , J.Michler , and P. Seris, Surface andInterface Analysis 36(10), 1384(2004).

8. Richard Payling, Johann Michler, Max Aeberhard, and Yuriy Popov,Surface and Interface Analysis,35(7), 583 (2003)

9. Arne Bengtson and Thomas Nelis, J. Anal. Bioanal. Chem.385, 568(2006).

10. G. Anil, M.R.P. Reddy, ArbindKumar, and T.L. Prakash , At. Spectrosc. 23(4), 119 (2002).

11. G. Anil, M.R.P. Reddy, and T.L.Prakash, J. Anal. Chem. 61(7),641(2006).

12. W.J.M. Boumans and J.J.A.M.Vrakking, Spectrochim. Acta, 42B, 819 (1987).


Trace Molybdenum Determination in Drinking Waters by USN-ICP-OES After Solid Phase Extraction onEthyl Vinyl Acetate Turnings-Packed Minicolumn Luis Escuderoa, Raúl A. Gila,b, José A. Gásqueza, Roberto A. Olsinaa,b, and *Luis D. Martíneza,b

a Área de Química Analítica, Departamento de Química, Facultad de Química Bioquímica y Farmacia, Universidad Nacional de San Luis, Chacabuco y Pedernera, San Luis, 5700, Argentina

b Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET),Av. Rivadavia 1917, CP C1033 AAJ, Capital Federal, Argentina

ABSTRACT

This study presents the devel-opment of a novel on-line pre-concentration method forinorganic molybdenum by flowinjection solid-phase extractionwith detection by inductivelycoupled plasma optical emissionspectrometry and ultrasonic neb-ulization (SPE-USN-ICP-OES). Thismethod employs a minicolumnfilled with ethyl vinyl acetateturnings as the adsorbent mater-ial. This system was applied tothe on-line preconcentration anddetermination of molybdenum inwater samples with good selec-tivity and reproducibility; molyb-denum was retained on theminicolumn without furthercomplexation.

The time required for the pre-concentration of a 20-mL sample,injection/data acquisition, andconditioning was about 5.0 min-utes, resulting in a samplethroughput of 12 samples perhour. A 300-fold total enhance-ment factor for this sample vol-ume was obtained with respectto the molybdenum determina-tion by ICP-OES without precon-centration.

The precision for six replicatemeasurements of a solution con-taining 20 µg L-1 Mo was 3.5%relative standard deviation, calcu-lated at peak height. The limit ofdetection (3σ) was 0.04 µg L-1.

This method was successfullyapplied to trace molybdenumdetermination in drinking watersamples and in a certified refer-ence material (Metals in NaturalWaters, NIST CRM 1643e).

INTRODUCTION

Molybdenum (Mo) is an essentialtrace element for plants, animals,and humans, and is the only metalof the second transition serieswhich is of major biological signifi-cance (1). In animals, it is an indis-pensable co-factor for severalenzymes, such as xanthine oxidaseand sulphite oxidase (2, 3). Inplants, although present at low lev-els, Mo is an essential micronutrientand is involved in the biochemicalprocesses related to fixing the N2

of the atmosphere by bacteria.Nevertheless, it is harmful toplants, animals, and humans at highconcentrations.

Although Mo is essential for ani-mals, it becomes toxic at high lev-els (4, 5). It must be emphasizedthat studies pertaining to the role ofMo in man or the environmenthave often been hampered by thelack of sufficiently sensitive analyti-cal methods for determining traceMo levels. For this reason, it is oftendifficult to determine whether thesymptoms attributed to Modeficiency or excess are due to bio-logical variations or simply toexperimental error (6). A betterunderstanding of the role of Mo inhumans, plants, and animals as wellas in the environment depends onimproving the sensitivity and accu-racy of the analytical methodsinvolved (7).

One of the major routes of incor-poration of Mo is through waterintake or uptake. For the analysis ofwater samples, inductively coupledplasma mass spectrometry (ICP-MS)(8, 9) has the analytical capabilityto determine Mo at trace and ultra-trace levels because of its high sen-sitivity and selectivity, and it alsooffers high sample throughput.However, the cost of theinstrumentation is not affordablefor many laboratories. Trace Modetermination in water samples hasbeen successfully carried out byatomic absorption spectrometrywith electrothermal atomization(ETAAS) (10) and inductively cou-pled plasma optical emission spec-trometry (ICP-OES) (11). Thedrawback of these two techniquesis that the Mo concentration in wellwater, tap water (12), seawatersamples (13), etc., is too low to bedirectly determined. Preconcentra-tion is an effective means forextending the detection limits ofthe ICP-OES technique. However,when practiced manually in thebatch mode, the operations are usu-ally too laborious and time-consum-ing.

Preconcentration and the deter-mination of trace Mo has been pro-posed by using N-benzoyl-N-phenyl-hydroxylamine (14) and α-benzoinoxime (15) as complexing reagentsin a liquid-liquid extraction system.Ammonium pyrrolidinedithiocarba-mate (APDC) and activated carbonhas also been used for the precon-centration and determination of Moin silicates by flame AAS (FAAS)(16). Giacomelli et al. (17) deter-mined Mo in steels by ETAAS after

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

22

complexation with ammonium saltof dithiophosphoric acid O,O-diethyl ester and adsorption ontoactivated carbon. Molybdenumdetermination in natural water sam-ples by ETAAS after enrichmentusing pyrocatechol violet has alsobeen proposed (18). Molybdenumhas been determined by ICP-MS(19) and ICP-OES (20) after com-plexation using oxine. Thiocyanatewas used for the preconcentrationand determination of Mo in steel(21). Calmagite and activated car-bon has been used for the separa-tion and preconcentration of traceMo in water samples prior to thedetermination by ICP-OES (13). Inaddition, in recent years there havebeen reports (22-29) utilizing thesolid-liquid separation and precon-centration of trace elements usingmicrocrystalline naphthaleneloaded with organic reagents.

Alternatively, microcolumnspacked with ethyl vinyl acetate(EVA) turnings have been success-fully employed for the on-line pre-concentration of trace elements.The beneficial use of this novelform of EVA as a sorbent materialfor As, Se, and Te on-line precon-centration and determination byatomic absorption spectrometrywas actually demonstrated by ourresearch group (30,31).

The aim of this work was todevelop a sensitive and selectiveflow injection (FI) preconcentra-tion method for the determinationof inorganic Mo in natural watersby on-line SPE, assisted by an EVAturnings-packed microcolumn cou-pled to ICP-OES with ultrasonicnebulization (USN). Because of thevery good chemical resistance andthe unlimited lifetime of the EVAmaterial, the column does not needany regeneration or repacking andremains unaltered after the precon-centration or elution step. More-over, Mo can be retained onto themicrocolumn without using a com-plexing reagent. To the best of our

knowledge, it is the first time thatEVA turnings are used as sorbentmaterial for Mo preconcentration.The proposed method wasoptimized and applied to the selec-tive determination of Mo in naturalwater samples and a certified refer-ence material.

EXPERIMENTAL

Instrumentation

All measurements wereperformed with a Model ICP 2070sequential ICP spectrometer (Baird,Bedford, MA, USA). The 1-mCzerny-Turner monochromator had a holographic grating with1800 mm-1 grooves. A U-5000 ATultrasonic nebulizer (CETAC Tech-nologies, Omaha, NE, USA), involv-ing a desolvation system, was used.

The ICP-OES and ultrasonic neb-ulization (USN) conditions arelisted in Table I. Minipuls™ 3 peri-staltic pumps (Gilson, Villiers-Le-Bel, France) were used. Sampleinjection was achieved using aRheodyne® (Cotati, CA, USA),Model 50, four-way rotary valve.The minicolumn (85 mm long, 4.0 mm i.d.) used as the EVA turn-ings holder was prepared usingcross-linked ethyl vinyl acetate

(EVA) Microline tubing. Tygon®pump tubing (Ismatec, Cole-ParmerInstrument Company, Niles, IL,USA) was employed to propel thesample, reagent, and eluent. The FIsystem used is shown in Figure 1.The 202.030-nm Mo spectral linewas used and the FI system mea-surements were expressed as peakheight emission, which wascorrected against the reagent blank.

Reagents

A moblydenum stock solution(1000 mg L-1) was prepared by dis-solving 920.3 mg (NH4)6Mo7O24 ·4H2O to a final volume of 500 mLwith water and a few drops ofconcentrated nitric acid (Fluka).

Ultra-pure water (18 MΩ cm-1)was obtained from an EASY pureRF (Barnstedt, Dubuque, IA, USA).

All other solvents and reagentswere of analytical reagent grade orbetter, and the presence of Mo wasnot detected in the working range.

Column Preparation

Ethyl vinyl acetate (EVA) is aflexible (rubbery), transparent poly-meric material with good low tem-perature flexibility, good chemicalresistance (to acids, alkalis, and

TABLE IICP-OES and Ultrasonic Nebulizer Instrumental Parameters

ICP-OES Conditions

RF Generator Power Plasma 0.8 kW

Frequency of RF Generator 40.68 MHz

Gas Flow Rate 8.5 L min-1

Auxiliary Gas Flow Rate 1.0 L min-1

Observation Height (above load coil) 15 mm

Analytical Line for Mo 202.030 nm

Ultrasonic Nebulizer Conditions

Heating Temperature 140.0 ºC

Condenser Temperature 4.0 ºC

Carrier Gas Flow Rate 1.0 L min-1

23

Vol. 29(1), Jan./Feb. 2008

alcohols) and high friction coeffi-cient. This material was employedto construct the preconcentrationminicolumn as follows:

The minicolumn was preparedby placing 100 mg of EVA turnings(small particles of spiralled shape)into an empty cylindrical EVA tub-ing using the dry packing method.To avoid loss of EVA particles when

change under reproducible condi-tions.

Sample Preparation

Water samples were filteredthrough 0.45-mm pore size mem-brane filters immediately after sam-pling, and the pH was adjusted topH 7.0 with hydrochloric acid orsodium hydroxide solution, thenstored at 4 ºC in Nalgene® polyeth-ylene bottles (Nalge Nunc Interna-tional, Rochester, NY, USA). Allglass materials used were washedwith 10% (v/v) HNO3 and ultrapurewater before use.

Preconcentration Step

A schematic of the preconcen-tration manifold is shown in Figure1. At the beginning of the precon-centration/determination cycles,the minicolumn was conditionedfor preconcentration first with10 mL of HCl solution (1.0 mol L-1)and finally washed with ultrapurewater until a neutral pH wasachieved (valve V1 in position B).

LoadingMolybdenum solutions were

loaded on the EVA minicolumn at aflow rate of 5.0 mL min-1 with valveV1 in position S and valve V2 inload position (a), while the HClsolution was passing throughoutthe manifold directly to the USNand later to the plasma (4 minutes).

InjectionAfter the loading time, the injec-

tion valve V2 was switched on theinjection position (b) and theretained metal was quantitativelyeluted with the HCl solution at 1.8 mL min-1 directly into the USN(0.5 minutes).

ConditioningAfter Mo determination was

completed, injection valve (V2)was switched back to the loadingposition, and ultrapure water waspassed (by the tubing lines)through the minicolumn (V1 inposition B) in order to eliminate the

Fig. 1. Schematic diagram of the instrumental setup. B, conditioning line; S, sam-pling line; W, waste; V1, two-way rotary valve, V2, load injection valve (a, load posi-tion; b, injection position); M, conical minicolumn; USN, ultrasonic nebulizer; Ar,argon gas supply either for plasma or for USN.

the sample solution passes throughthe minicolumn, a small amount ofquartz wool was placed on bothends of the minicolumn. The col-umn was then connected to a peri-staltic pump with PTFE tubing toform the preconcentration system.This method was used for around600 preconcentration cycles, andthe retention capacity did not

24


Effect of pH on MolybdenumRetention

The pH value plays an importantrole in the retention phenomenonof different ions on the EVA turn-ings. In order to evaluate the effectof pH, a series of sample solutionswere adjusted to a pH range of1.0–12.0 with HCl or sodiumhydroxide and processed according

remaining acid (0.5 minutes).

The emission measurements(peak height) were proportional tothe Mo concentration in the sampleand were used for allmeasurements.

The operating conditions wereestablished and the determinationwas carried out.

to the recommended procedure.The effect of pH on molybdenumretention is shown in Figure 2. Itcan be seen that quantitativeadsorption (close to 100%) wasachieved in the pH range between5.0 and 7.0.

Accordingly, a pH of 6.0 wasselected for further experiments.All these studies were carried outwith synthetic Mo solutions with aconcentration of 20 µg L-1.

Effect of Sample Loading FlowRate on Molybdenum Retention

The sample flow rate throughthe microcolumn is one of thesteps that controls the analysistime. In this study it was verifiedthat with flow rates up to 5.0 mL min-1 there was no effecton analyte recovery, which underoptimum conditions was close to100% (Figure 3). At higher flowrates, the recovery showed an obvi-ous decrease with a diminishingcontact time between the analyteand the adsorbent material.

Effect of Eluent

A satisfactory eluent shouldeffectively elute the analyte in a dis-crete volume in order to obtain thebest analyte recovery. Different elu-ents were tested at a flow rate of1.8 mL min-1 through the column inorder to evaluate and compare theanalyte recovery (95% confidenceinterval n = 6). Best results wereobtained with HCl and, consequently,this acid was employed as the elu-ent. In this way, different concen-trations were tested and we verifiedthat 1.0 mol L-1 was the minimumconcentration necessary to obtainthe best response. At lower HClconcentrations, a lower sensitivitywas obtained due to an increase inthe dispersion effect.

The elution flow rate is a vari-able which is often optimized care-

Fig. 3. Analytical signal (expressed as % relative response) as a function of thesample load flow rate. Mo concentration: 5.0 µg L-1; pH: 6.0; HCl concentration:1.5 mol L-1.

Fig. 2. Analytical signal (expressed as % relative response) as a function of thesample pH. Mo concentration: 5.0 µg L-1; Sample flow rate: 5.0 mL min-1, HCl concentration: 1.0 mol L-1.

25

Vol. 29(1), Jan./Feb. 2008

fully. In this case, this variable wassubjected to the optimal intakeflow rate of the USN which was1.8 mL min-1. Accordingly, a flowrate of 1.8 mL min-1 was selected asthe fixed elution flow rate.

Column Reuse

The stability and regeneration ofthe column were investigated. Dueto the high chemical and physicalresistance of the EVA, the columnwas stable up to at least 500 pre-concentration cycles, and neithercolumn deterioration nor adecrease in Mo recovery wasobserved.

Interference Studies

The effects of common coexist-ing ions in various drinking watersand a certified reference materialNIST CRM 1643e (National Instituteof Technology, Gaithersburg, MD,USA) on the adsorption of Mo onthe EVA turnings minicolumn wereinvestigated. In these experiments,solutions of 20 µg L-1 of Mo(VI) con-taining the added interfering ionswere treated according to the rec-ommended procedure. The toler-ance limits of the coexisting ions,defined as the largest amount of ionthat produces a recovery of Mo lessthan 90% were evaluated. Theresults showed that Cu2+, Zn2+,Cd2+, Ni2+, Co2+, Cr3+, Mn2+, andAl3+ could be tolerated up to atleast 2500 µg L-1; Fe2+ and Fe3+

could be tolerated up to at least5000 µg L-1. Commonlyencountered matrix components,such as alkaline and alkaline earthelements, are not retained on theminicolumn.

Figures of Merit

The time required for the pre-concentration of 20.0 mL of sample(5.0 mL min-1), acquiring injection/data (0.5 min at 1.8 mL min-1), andconditioning was about 5.0 minutes,resulting in a sample throughput of12 samples per hour. A 300-foldtotal enhancement factor (10 for

the minicolumn and 30 for theUSN) for a sample volume of 20.0mL was obtained with respect tothe Mo determination by ICP-OESwithout preconcentration.

The relative standard deviation(RSD) for six replicatemeasurements of a solution con-taining 20 µg L-1 Mo was 3.5%. Thelimit of detection (LOD), calculatedas the amount of Mo required toyield a net peak equal to threetimes the standard deviation ofthe background signal (3σ), was0.04 µg L-1. The calibration curvewas linear from levels close to theLOD up to 500 µg L-1, with a corre-lation coefficient of 0.9989.

Recovery Study and Applicationto Real Samples and ReferenceMaterials

In order to evaluate the Morecovery of this method, 200 mL of

a natural drinking water samplewas collected in our laboratory anddivided into 10 portions of 20 mL.The proposed method was appliedto six portions and the averagequantity of the Mo obtained wastaken as the base value. Then,increasing quantities of Mo(VI)were added to the other aliquots ofthe sample and Mo was determinedby the same method. The recover-ies were in the range of98.3–100.0% (see Table II).

Additionally, the accuracy of theproposed method was evaluated bydetermining Mo in the certified ref-erence material NIST CRM 1643ewith a molybdenum content of121.4 ± 1.3 µg L-1 and a densityequal to 1.016 g mL-1 at 22ºC.

TABLE II Concentration of Molybdenum in Drinking Water Samples

and NIST CRM 1643e (95% confidence interval; n = 6)

Sample Mo(VI) Mo(VI) Mo(VI) Recoveryc

Conc. Conc. Conc.Base Added Found

(µg L-1) (µg L-1) (µg L-1) (%)

A (Drinking Water-1st week)a 1.23 ± 0.08 0.00 1.23 ± 0.08 ---

A 1.23 0.30 1.53 ± 0.09 100.0

A 1.23 0.60 1.82 ± 0.10 98.3

A 1.23 0.90 2.12 ± 0.10 98.8

A 1.23 1.20 2.42 ± 0.09 99.1

B (Drinking Water - 2nd week)a 0.97 ± 0.08 0.00 0.97 ± 0.07 ---

C (Drinking Water - 3rd week)a 0.88 ± 0.08 0.00 0.88 ± 0.08 ---

D (Drinking Water - 4th week)a 1.11 ± 0.10 0.00 1.11 ± 0.09 ---

E (NIST CRM 1643e)b 120.50 ± 1.20 0.00 120.50 ± 1.20 ---

a Drinking water collected in our laboratory.b Certified Value = 121.4 ± 1.3 µg L-1

c Recovery (%) = [(found-base)/added] x 100

26

CONCLUSION

On the basis of the results of thisstudy, the ethyl vinyl acetate (EVA)turnings have shown to be a suit-able substrate for the preconcentra-tion of inorganic molybdenum.

The on-line coupling of solidphase extraction with ultrasonicnebulization (SPE-USN) and induc-tively coupled plasma optical emis-sion spectrometry (ICP-OES) for Modetermination in water samplesincreases the speed of preconcen-tration and the analysis process,reduces sample consumption andcontamination risks, and improvessensitivity and selectivity. As anadditional advantage for total Modetermination, this metal can easilybe retained without using a com-plexing reagent. The preconcentra-tion system was able to determineMo in drinking waters at sub-µg L-1

levels with recoveries close to100%.

To the best of our knowledge,this is the first report where EVAwas used as a substrate in an on-line preconcentration system formolybdenum determination.

ACKNOWLEDGMENTS

This work was supported byConsejo Nacional deInvestigaciones Científicas yTécnicas (CONICET); AgenciaNacional de Promoción Científicay Tecnológica (FONCyT) (PICT-BID) and Universidad Nacional deSan Luis (Argentina) (UNSL).

Received February 9, 2007.Revision received January 16, 2008.

REFERENCES

1. G.H. Morrison, CRC Cnt. Rev. Anal.Chem. 8, 287 (1979).

2. M.P. Coughlan (Ed.), Molybdenumand Molybdenum ContainingEnzymes, Pergamon, Oxford, UK(1980).

3. S.J.N. Burgmayer and E.I. Stiefel, J.Chem. Educ. 62, 943 (1985).

4. H. Seiler, A. Sigel, and H. Sigel, Hand-book on Metals in Clinical and Ana-lytical Chemistry, Marcel Decker,New York (1994).

5. N.S. Mandel and G.S. Mandel, J. Am.Chem. Soc. 9 (1975) 2319.

6. I.Császma, E.Andrási, A.Lásztity,E.Bertalan, and D.Gawlik, J. Anal.At. Spectrom., 18 (2003) 1082.

7. Z. Marczenko and R. Lobinski, Deter-mination of Molybdenum in Bio-logical Materials, Vol. 63, No. 11,Pure & App. Chem., p. 1627(1991).

8. K.L.Caldwell, J. Hartel, J. Jarrett, andR.L. Jones, At. Spectrosc. 26, 1-7(2005).

9. M. Dutta, K. Chandrasekhar, andA.K.Das, At. Spectrosc., 26, 14(2005).

10. C.L. Chen, K.S.K. Danadurai, and S.Da Huang, J. Anal. At. Spectrom.16, 404 (2001).

11. C.H. Lee, M.Y. Suh, K.S. Choi, J.S.Kim, Y.J. Park, and W.H. Kim,Anal. Chim. Acta 475, 171 (2003).

12. A.R. Ghiasvand, S. Shadabi, E.Mohagheghzadeh, and P. Hashemi,Talanta 66, 912 (2005).

13. H.C. Santos, M.G.A. Korn, andS.L.C. Ferreira, Anal. Chim. Acta426, 79 (2001).

14. Y. Shijo, M. Suzuki, T. Shimizu, S.Aratake, and N. Uehara, Anal. Sci.12, 953 (1996).

15. A.R. Ghiasvand, S. Shadabi, E.Mohagheghzadeh, and P. Hashemi,Talanta 66, 912 (2005).

16. G.R. Boaventura, J.R. Hirson, andR.E. Santelli, Fresenius’ J. Anal.Chem. 350, 651 (1994).

17. M.B.O. Giacomelli, J.B.B. Silva, andA.J. Curtius, Talanta 47, 877(1998).

18. T. Okutani, K. Noshiro, and A.Sakuragawa, Anal. Sci. 14, 621(1998).

19. G.E. Hall, C.J. Park, and J.C. Pelchat,J. Anal. At.. Spectrom. 2, 189(1987).

20. G.E. Hall, J.C. Pelchat, and K.N.Silva, Analyst 112, 631 (1987).

21. J.C. Andrade, C.J. Cuelbas, S.P.Eiras, Talanta 47 (1998) 719.

22. M.A. Taher, Talanta 50, 559 (1999).

23. Z.F. Fan, B. Hu, and Z.C. Jiang,Spectrochim. Acta B 60, 65 (2005).

24. M.A. Taher, A.M. Dehzoei, B.K.Puri, and S. Puri, Anal. Chim. Acta367, 55 (1998).

25. M.A. Taher, Anal. Chim. Acta 382,339 (1999).

26. R.K. Dubey, A. Bhalotra, M.K.Gupta, and B.K. Puri, Microchem.J. 58, 117 (1998).

27. M.A. Taher, Anal. Chim. Acta 408,153 (2000).

28. B. Cai, B. Hu, H.C. Xiong, Z.H. Liao,L.S. Mao, and Z.C. Jiang, Talanta55, 85 (2001).

29. M.A. Taher, Talanta 52, 181 (2000).

30. R.A. Gil, N. Ferrúa, J.A. Salonia, R.A.Olsina, and L.D. Martinez, J. Haz. Mat. 143, 431 (2007).

31. R.A. Gil, M.M. Kaplan, J.A. Salonia,J.A. Gásquez, and L.D. Martinez,At. Spectrosc. 28, 66 (2007).


*Corresponding author.E-mail:[email protected]: 34981595012

Determination of Hexavalent Chromium in WeldingFumes by Flow Injection Flame Atomic Absorption

Spectrometry After Dynamic Alkaline Ultrasound-assistedExtraction/Anion Exchange Preconcentration

*M.C. Yebra and R.M. Cespón Department of Analytical Chemistry, Nutrition and Bromatology,

Faculty of Chemistry, University of Santiago de Compostela, 15782-Santiago de Compostela, Spain

ABSTRACT

A fast and simple method forthe fully automatic determina-tion of Cr(VI) in welding fumesby flame atomic absorption spec-trometry is presented. Thismethod is based on a dynamicalkaline ultrasound-assistedextraction coupled on-line to aflow-injection preconcentrationminicolumn system. The mini-column was filled with ananionic exchange resin (Dowex1-X8), which increased the ana-lytical sensitivity of the proposedmethod 12.7 times. Under theoptimum conditions, using 10%(v/v) Na2CO3–2% (v/v) NaHCO3

as the extractant solution at aflow rate of 6 mL min-1, 40 ºCand 5 minutes as the extractiontemperature and sonication time,respectively, a retention flowrate of 2.5 mL min-1, 150 µL ofeluent (0.5 M (NH4)2SO4–0.1 MNH3), and an elution flow rate of3.5 mL min-1, a precision of 2%(expressed as the relative stan-dard deviation) and a samplingfrequency of 10 samples perhour were obtained. The pro-posed method was successfullyapplied to the analysis of Cr(VI)in welders’ workplace environ-ments.

INTRODUCTION

Interest and demand for hexava-lent chromium determination in airfrom workplace environments hasincreased because it is toxic andcarcinogenic even at very low lev-els. High concentration of Cr(VI) inthe workplace occurs during workactivities such as stainless steelwelding, thermal cutting, chromeplating, painting, and coatingprocesses (1). The OccupationalSafety and Health Administration(OSHA) has proposed a permissibleexposure limit (PEL) of 0.005 mg ofCr(VI) per cubic meter ofworkplace air (2). Therefore, toprotect the health of exposedworkers it is necessary todetermine Cr(VI) in workplace airusing fast and reliable analyticalprocedures in order to verify thatthe concentration of this elementdoes not rise above the permissiblevalue.

Analysis of Cr(VI) in industrialhygiene samples generally requirestwo major steps: extraction from afilter and detection. An optimalextraction procedure should com-pletely extract Cr(VI) species with-out disturbing the speciesdistribution (3–4). That is, thereshould be not reduction of Cr(VI)and no oxidation of Cr(III). TheNational Institute for OccupationalSafety and Health (NIOSH) andOSHA recommended alkalinebuffer extraction of air filters forhexavalent chromium determina-

energy to accelerate the extractionstep. Thus, Wang el al. (5–6) usedan ammonium buffer; Ndung’u etal. (7) an ammonium and a phos-phate buffer; and Sabty-Daily et al.(8), Scancar et al. (9), and Hazel-wood et al. (10) used a carbonatebuffer. In these procedures,between 15–30 minutes and aroom temperature of 70 ºC as theextraction conditions wereemployed. However, theseapproaches are based on an off-lineextraction and full automation ofthe method herewith describedhas to our knowledge not yet beenproposed.

Measurement of extracted Cr(VI)from workplace samples is gener-ally accomplished by spectropho-tometry by using its selectivereaction with 1,5-diphenylcarbazide(DPC) (11–15). Nevertheless, underproper acidic conditions the ligandreacts little with other transitionmetals. Therefore, solid-phaseextraction is usually used to isolateCr(VI) from Cr(III) and other metalcations. Atomic absorption spec-trometry (AAS) was also an analyti-cal technique used to determineCr(VI) because it is simple, rapid,and readily available. In thesemethodologies, separation of Cr(VI)is ordinarily required becauseCr(III) might also be present, andanalytical sensitivity is approved(16–18). In this sense, anion-exchange resins or membraneswere successfully applied forchromium speciation (5,19–21).

The present paper describes afully automated procedure based

tion in workplace atmospheres andin airborne hexavalent chromiumdetermination with the aim of stabi-lize Cr(VI) and prevent air oxida-tion of Cr(III) (2). In this sense,several methodologies have beenproposed involving alkaline buffersfor extracting and dissolving Cr(VI)species from hygienic samplesbased on the use of ultrasonic

28

on coupling dynamic alkaline ultra-sound- assisted extraction, minicol-umn separation, and a flame atomicabsorption spectrometric (FAAS)detection of Cr(VI) in weldingfumes collected on polyvinyl chlo-ride (PVC) filters through a flowinjection (FI) manifold interface.The developed method has beenapplied to the determination ofCr(VI) in welders’ workplace envi-ronments.

EXPERIMENTAL

Instrumentation

A PerkinElmer Model 5000atomic absorption spectrometerfurnished with a chromium hollow-cathode lamp was used(PerkinElmer Life and AnalyticalSciences, Shelton, CT, USA). Theinstrument was set at 357.9 nm. Anacetylene flow rate of 4.5 L min-1

and an air flow rate of 15.0 L min-1

were employed to obtain a clearyellow flame (reducing). The spec-trometer output was connected toa Perkin Elmer 50 servographrecorder with a range of 5 mV. Thesignals measured were the heightsof the absorbance peaks. The FIsystem (Figure 1) comprises twoGilson® Minipuls™-3 peristalticpumps (Gilson, France), an ultra-sonic bath (Selecta, Spain), fourRheodyne® injection or switchingvalves (USA), Models 5041 and5301, and a glass minicolumn(100 mm x 3 mm i.d., bed volume700 µL) (Omnifit, UK) with theends plugged with filter paper(Whatman 541) used as the extrac-tion unit. The laboratory-mademinicolumn for the on-line precon-centration step was prepared byfilling Tygon® tubes (100 x 1.1 mmi.d.) with 50 mg of Dowex 1-X8resin (20 mesh).

All glassware and plasticwarewas decontaminated with 10% (v/v)nitric acid for at least 48 h andwashed three times with Milli-Q™water before use.

Statistical analysis of the experi-mental designs was carried out bymeans of the Statgraphics PlusV.5.1 statistical package (Manugis-tic, Inc. Rockville, MD, USA).

Reagents

Ultrapure water of 18.3 MΩ cmresistivity obtained from a Milli-Qwater purification system (MilliporeCorporation, Bedford, MA, USA)was used for the preparation ofreagents and standards. All chemi-cal products are of analytical-reagent grade. Cr(VI) and Cr(III)working standard solutions wereprepared from 1000 µg mL-1 stockstandard solutions prepared byK2CrO4 and Cr2O3 (Merck, Darm-stadt, Germany). Diluted standardsolutions were freshly preparedbefore each analysis. Sodium car-bonate and sodium bicarbonate(Merck) were used for preparationof 10% (v/v) Na2CO3–2% (v/v)NaHCO3) (pH 11) which we’veused as the alkaline extraction solu-tion. Dowex 1-X8 strongly basicanion-exchange resin Cl– form (20mesh) (Sigma, USA) was used toisolate and preconcentrate Cr(VI).

Ammonium sulfate and ammonia(Merck) were used for preparationof a buffer solution 0.5 mol L-1

(NH4)2SO4–0.1mol L-1 NH3 (pH 8.3)which was used as the eluent.

Evaluation of SampleHomogeneity

Before starting the optimizationstudy of the dynamic alkalineextraction process, the homogene-ity of the samples was tested. Sixsamples were collected at the sameworkplace site using six filters andsix air sampling pumps. The objec-tive was that between sampling,filter variation should not mask theeffects of the experimental factorsto be optimized. With this aim, aset of six filter samples wascollected for Cr(VI) determinationduring an effective welding time of400 min at a sampling flow rate of2.0 L min-1. Cr(VI) was captured onPVC filters and analyzed using theNIOSH 7600 method (22). Themean concentration obtained forCr(VI) was 1.5 µg m–3 with a rela-tive standard deviation of 3.4%,which is an acceptable value forhomogeneous samples (similar in

Fig. 1. Experimental set-up used for the dynamic alkaline extraction process and FIpreconcentration/determination of Cr(VI) in welding fumes. P1 and P2, peristalticpumps; AES, alkaline extraction solution [10% (v/v) Na2CO3–2% (v/v) NaHCO3]; W, waste; UB, ultrasonic bath; M, minicolumn containing the filter; SS, Cr(VI) stan-dard solution; B, blank; IV, injection valve; SV1-SV3, switching valves; IV1 and IV2,injection valves; E, eluent [0.5 mol L-1 (NH4)2SO4 - 0.1 mol L-1 NH3]; C, carrier (ultra-pure water); MN, preconcentration minicolumn packed with the anion-exchangeresin Dowex 1-X8 and FAAS, flame atomic absorption spectrometer.

29

Vol. 29(1), Jan./Feb. 2008

terms of mass and composition)obtained with different pumps atthe same time and from the samesite.

Sampling and Procedure

Workplace air samples werepumped with a Gilian HFS-513 air-sampling pump with a constanthigh flow of 750 to 5000 mL min-1

(Sensidyne, USA). This pumpwas calibrated up to an air flow of2.0 L min-1 (sampling flow rate).Welding fumes were collected onMillipore (Bedford, USA) polyvinylchloride (PVC) filters (5.0 µm poresize). Each filter is in a samplingcassette of 37 mm diameter. Sam-pling took place in a welding shop,and welding fumes were collectedduring the welding of stainlesssteel.

Air filters were removed fromthe sampling cassette and placedinto a glass minicolumn (extractionunit of the dynamic alkaline ultra-sound-assisted extraction device).This extraction unit was assembledfor the dynamic alkaline extractionprocess (Figure 1). Then, theextraction unit was immersed intothe ultrasonic bath at 40 ºC, andfilled with the extraction solution[2 mL of 10% (v/v) NaHCO3–2%(v/v) Na2CO3] impelled by the peri-staltic pump. The alkaline extrac-tion solution was then circulatedthrough the extraction unit for 5minutes under ultrasonic irradiationat 6.0 mL min-1. During the extrac-tion step, the direction of the leach-ing solution stream was changedeach 30 seconds to prevent com-pactness of the filter in the extrac-tion cell, which could causeoverpressure in the FI system. Afterthe extraction step, the SV2 wasswitched to its other position, andthe alkaline extract arrived at thepart of the FI system where the sep-aration step takes place. Thus, thealkaline extract passed at 2.5 mLmin-1 through the minicolumnplaced on the loop of the elutioninjection valve (IV1) and containing

the anion-exchange resin (Dowex1-X8). The sample matrix was sentto waste, while ultrapure waterflowed through the detector.Finally, Cr(VI) was subsequentlyeluted at 3.5 mL min-1 by injectionof a 150-µL volume of 0.5 mol L-1

(NH4)2SO4–0.1mol L-1 NH3 into awater carrier stream and continu-ously monitored by a flame atomicabsorption spectrometer.


Optimization of the Preconcen-tration Step

With the aim of increasing ana-lytical sensitivity, a preconcentra-tion minicolumn was included inthe FI manifold. The anion-exchange resin Dowex 1-X8 wasselected as a suitable material forthe separation of Cr(VI) because ofthe good results reported in the lit-erature (5,21). According to thesereferences, an ammonium buffer0.5 mol L-1 (NH4)2SO4–0.1 mol L-1

NH3 (pH 8.3) was selected as theeluent.

The main factors that affectedthe preconcentration process werepH of the sample solution, reten-tion flow rate, elution flowrate,and the eluent volume. In order toestablish the best chemical and FIconditions for retention and elutionof the analyte, the procedure wasoptimized. The pH effect was evalu-ated for Cr(VI) preconcentration inthe resin in the range of 4–12. Theresults showed that Cr(VI) isretained quantitatively within therange studied. This range includesalkaline pH, which will be themedium of the samples when thepreconcentration step isaccomplished since it takes placeafter alkaline extraction. Therefore,adjustment of the alkaline sampleextract was not necessary forCr(VI) retention. The retentionflow rate was optimized in the 1–5mL min-1 range. It was found thatthere was a decrease in sensitivity

for flow rates higher than 2.5 mLmin-1; thus, this flow rate wasselected for this study. The effect ofthe eluent flow rate (carrier stream:ultrapure water) was studied in the2.0–5.0 mL min-1 range. An increasein peak height was observed up toa flow rate of 3.5 mL min-1 whichis due to peak broadening at lowflow rates. Thus, an elution flowrate of 3.5 mL min-1 was selectedfor further studies. The influence ofthe eluent volume used was alsostudied, and it was observed that150 µL of 0.5 mol L-1 ammoniumbuffer was sufficient for quantita-tive elution of Cr(VI).

Optimization of the DynamicUltrasound-assisted AlkalineExtraction of Cr(VI)

To extract the Cr present in thehexavalent state, an alkaline 10%(v/v) Na2CO3–2% (v/v) NaHCO3

buffer solution (pH 11) that hadbeen applied to industrial hygienesamples and for hexavalentchromium determination in work-place atmospheres was adopted(2,9–10). The ultrasound-assistedextraction of Cr(VI) from the filterwas carried out using an experi-mental design approach. The fac-tors selected and their levels arepresented in Table I. A factorialscreening half-fraction 24–1 design,which studied the effects of thefour factors in 8 randomized experi-ments plus one centerpoint wasrun. The conclusions of this screen-ing study were that in dynamicCr(VI) alkaline extraction there aretwo factors statistically significantlyaffected by a positive sign: extrac-tion temperature and sonicationtime; that is, these factors overtakethe limit of statistical significance(at 95% confidence). The otherparameters were not statisticallysignificant, with estimated effectsof 6.3 and 0.3 for extraction flowrate and alkaline solution volumeused as extractant, respectively.Therefore, as the extraction flowrate is affected by a positive sign, a

30

value for this variable of 6.0 mL min-1

was chosen as optimum. Withregard to the alkaline solution vol-ume, this variable also is affected bya positive sign. But as can be seenabove, it has a small estimatedeffect, and sin order to shorten theanalysis time, a 2-mL value for thisvariable was chosen as optimum forsubsequent work. Since the screen-ing design only provides tendenciestowards the optimum, a new factor-ial design was performed in orderto fine-tune the extraction tempera-ture and sonication time to obtaina quantitative % recovery with theminimum values of these variables.For this, we used an orthogonalcentral composite design, 22 + starwith two center points, resulting in10 non-randomized runs with 4error degrees of freedom. Axial dis-tance (a) was selected having avalue of 1.07809 in order to estab-lish the orthogonality condition.The results obtained confirmed thatthe sonication time was a statisti-cally influential factor at the 95%confidence level in the range stud-ied (3–5 minutes). While extractiontemperature is affected also by apositive sign, it is not statisticallysignificant in the range studied(40–70 ºC). Considering Figure 2,which shows the response surfaceof this design, we decided to con-

sider as optimum the following val-ues for the variables studied: 5 minutes and 40ºC for sonicationtime and extraction temperature,respectively.

Determination of Cr(VI) in thePresence of Cr(III)

To estimate the recovery forCr(VI) in the presence of Cr(III),the alkaline buffer solution used asextractant was spiked beforeextraction with Cr(III) concentra-tions between 0.05–0.2 µg mL-1.The results show that recoveriesfor Cr(VI) in spiked samples werebetween 97.1 and 98.7%.Therefore, the proposed proceduredoes not disturb the chromiumspecies distribution.

Analytical Characteristics

The calibration graph (n=7) wasmeasured with Cr(VI) standard solu-tions between 0 and 0.35 µg mL-1

under optimum chemical and flowconditions for the whole process.Thus, the equation was: absorbance = 1 x 10–4 + 0.661X(r=0.999), where X is Cr(VI) con-centration expressed as µg mL-1.The limit of detection (LOD) andthe limit of quantification (LOQ)were calculated from 30 measure-ments of unexposed air filters andwere 1.8 and 6.1 µg L-1, respectively.The preconcentration factor of the

method, based on the rate betweenthe calibration graph of theproposed method and the directgraph slopes (without preconcen-tration), was found to be 12.7. Therepeatability (2%) was calculatedfrom 11 analyses of air filters sam-pled at the same time in a station-ary sampling site at a workplaceenvironment by means of the rela-tive standard deviation. The recov-ery of the method was checked byadding to the alkaline extractantsolution 0.1 µg of Cr(VI) and usingthis solution to develop the deter-mination of Cr(VI) in an air sampleby the proposed procedure. Therecovery obtained was 97.1%.

A comparison of the slopes ofthe calibration graph and the stan-dard addition graph using a t-test(95% confidence level) was madeand no difference was observed,which indicated that there is not amatrix effect (22).

The sample throughput (takinginto account the global process)was about 10 samples h–1.

Application

The continuous systemdescribed above was applied to thedetermination of Cr(VI) in weldingfumes from a welding shop duringthe welding of stainless steel. Ascan be seen in Table II, the concen-

TABLE IFactors and Levels Considered in the Factorial

Screening Half-fraction 24–1 Design Used to Optimize the Dynamic Ultrasound-assisted

Alkaline Extraction of Cr(VI) From the Air Filters

Factor Low Level High Level

Extraction Temperature 20 ºC 70 ºC(room temperature)

Sonication Time 0.5 min 5 min

Extraction Flow Rate 3 mL min-1 6 mL min-1

Alkaline Solution Volume 2 mL 5 mLUsed as Extractant [10% (v/v) Na2CO3–2% (v/v)NaHCO3]

Fig. 2. Estimated response surface from the central compositedesign 22 + star applied to optimize the dynamic ultrasound-assisted alkaline extraction of Cr(VI).

31

Vol. 29(1), Jan./Feb. 2008

tration of Cr(VI) obtained was up to1.5 µg m–3. The results obtainedwere also compared with thoseachieved by using the NIOSH 7600method (23). To compare theresults obtained by these two meth-ods, a paired t-test was applied (24)and both methods do not give sig-nificantly different values, thus theagreement between the two meth-ods is satisfactory at a 95% confi-dence level.

CONCLUSION

For the first time, the fullautomation in the determination ofCr(VI) in air samples was accom-plished and herewith reported. Theproposed methodology is based ona dynamic alkaline ultrasound-assisted extraction manifold cou-pled to a flow injection (FI)preconcentration-FAAS system. Thismethod was found to be fast, accu-rate, inexpensive, and simple, andwas accomplished with minimumreagent consumption as comparedwith off-line ultrasound-assistedextraction procedures. The study ofmethod performance demonstratedthe benefits of dynamic ultrasoundtreatment with a quantitative recov-ery and good sensitivity, achievinga preconcentration factor of 12.7and a LOD lower than 2.0 µg L-1.The analysis of workplace air sam-

ples demonstrated the validityof the proposed method for thedetermination of Cr(VI) inwelding fumes.

Received October 8, 2007.

REFERENCES

1. J. N. Dennis, M.J. French, P.J. Hewitt,S.B. Mortazavi, and C.A.J. Redding,Anal. Occup. Hyg. 46, 43 (2002).

2. Department of Labor OccupationalSafety and Health Administration:Safety and Health Topics. Hexava-lent Chromium.http://www.osha.gov/SLTC/hexa-valentchromium/index.html

3. M.J. Marqués, A. Salvador, and A.E.Morales Rubio, Fresenius J. Anal.Chem. 362, 239 (1998).

4. K. Ashley, A.M. Howe, M. Demange,and O. Nygren, J. Environ. Monit.5, 707 (2003).

5. J. Wang, K. Ashley, E.R. Kennedy,and C. Neumeister, Analyst 122,1307 (1997).

6. J. Wang, K. Ashley, D. Marlow, E.C.England, and G. Carlton, Anal.Chem. 71, 1027 (1999).

7. K. Ndung'u, N. Djane, F. Malcus, andL. Mathiasson, Analyst 124, 1367(1999).

8. R.A. Sabty-Daily, K.K. Luk, and J.R.Froines, Analyst 127, 852 (2002).

TABLE IIDetermination of Hexavalent Chromium

Cr(VI) Concentration (µg L–1); Mean ± SD (n=3)Air Samples Proposed Method Reference Method

(NIOSH 7600)

Sample 1 1.50 ± 0.03 1.44 ± 0.07Sample 2 1.25 ± 0.02 1.20 ± 0.06Sample 3 0.10 ±0.00 ndSample 4 0.95 ± 0.01 1.00 ± 0.05Sample 5 0.26 ± 0.01 ndSample 6 0.24 ± 0.01 nd

Sample 7 0.85 ± 0.02 0.90 ± 0.04

Critical value of t (n–1 = 3, P=0.05) = 3.18; Experimental value of t=0.08SD = Standard Deviationnd = not detected

9. J. Scancar, and R. Milacic, Analyst127, 629 (2002).

10. K. J. Hazelwood, P. Drake, K. Ash-ley, and D. Marcy, J. Occup. Envi-ron. Hyg. 1,613 (2004).

11. L. Girard, and J. Hubert, Talanta 43,1965 (1996).

12. J. Wang, K. Ashley, D. Marlow, E.C.England, and G. Carlton, Anal.Chem. 71, 1027 (1999).

13. G. Samanta, C.B. Boring, and P.K.Dasgupta, Anal. Chem. 73, 2034(2001).

14. E.H. Borai, E.A. El-Sofany, A.S.Abdel-Halim, and A.A. Soliman,Trends Anal. Chem. 21, 741(2002).

15. M. Goldoni, A. Caglieri, D. Poli,M.V. Vettori, M. Corradi, P. Apos-toli, and A. Mutti, Anal. Chim. Acta562, 229 (2006).

16. D. Naranjit, Y. Thomassen, and J.C.Van Loon, Anal. Chim. Acta 110,307 (1979).

17. C. Brescianini, A. Mazzucotelli, F.Valerio, R. Frache, and G.Scarponi, Fresenius Z. Anal. Chem.332, 34 (1988).

18. R. Milacic, J. Scancar, and J. Tusek,Anal. Bioanal. Chem. 372, 549(2002).

19. G.L. Ou-Yang, and J.F. Jen, Anal.Chim. Acta 279, 329 (1993).

20. E. Castillo, M. Granados, and J.L.Cortina, Anal. Chim. Acta 464, 15(2002).

21. A. Aparna, M. Sumithra, G.Venkateswarlu, A.C. Sahayam, S.C.Chaurasia, and T. Mukherjee, At.Spectrosc. 27, 123 (2006).

22. M. Blanco, R. Boque, R. Cela, J.Coello, S. Paspoch, M.C. Ortiz, J.Riba, F.X. Rius, A. Ruiz, L.A. Sara-bia, and X. Tomás, Avances enQuimiometría Práctica, Universi-dad de Santiago de Compostela,Santiago de Compostela, Spain(1994).

23. NIOSH Method 7600.http://www.cdc.gov/niosh/nmam/pdfs/7600.pdf

24. J.C. Miller and J.N. Miller, Estadís-tica y Quimiometría para QuímicaAnalítica, Prentice Hall, Madrid,Spain (2002).

32

INTRODUCTION

Flame atomic absorptionspectrometry (FAAS), which is fre-quently used for the determinationof metal ions because it requiresrelatively simple instrumentation,is less used for the indirect determi-nation of non-metallic inorganicions and organic species (1-5) dueto the complex operations andinsufficient sensitivity.

Flow injection (FI) on-linepretreatment of samples is anattractive technique in atomicabsorption spectrometry due to itssimplicity, sensitivity, reducedmatrix effects, and high throughputrates. Preconcentration is an earlyand widely used technique formetal ion analysis. Cold vapor andhydride on-line generation andseparation have been developed forthe determination of some metalelements such as mercury (6-9),antimony (10-11), cadmium (12),lead (13), arsenic (14-15), iodide(16), and selenium (17). Otheron-line reaction and translationtechniques have been employedfor the indirect determination ofcyanide (18), COD (19), and nitriteand nitrate (1). Solid phase extrac-tion (SPE) is a method often usedfor FI on-line separation for theindirect determination of organiccompounds (20-21). Ion pair solidphase extraction (IP-SPE), one ofthe preconcentration and separa-tion methods, has some advantagesover normal phase and ionexchange SPE. First, it is compati-ble with aqueous solutions and

does not require previous dissolu-tion or extraction into a non-polarsolvent. Second, IP-SPE oftenresults in lower detection limitscompared to other sample prepara-tion methods (22).

Gatifloxacin (1-cyclopropyl -6-fluoro -1, 4-dihydro -8-methoxy 7-[3-methyl -1-piperazinyl] -4-oxo- 3-quinolinecarboxylic acid), a syn-thetic broad-spectrum antimicrobialfluoroquinolone, is active againstboth gram-negative and gram-posi-tive bacteria. It is used in the treat-ment of a wide range of infections(23), such as acute bacterial exacer-bation of chronic bronchitis, acutesinusitis, community-acquiredpneumonia, cystitis, and compli-cated urinary tract infections. Sev-eral methods for the determinationof gatifloxacin are described in theliterature using capillaryelectrophoresis (24), high-perfor-mance liquid chromatography (25-28), pressurized capillaryelectrochromatography (29), micro-biological assay (30), spectrofluo-rimetry (31), and UVspectrophotometry (32).Chromatographic techniques aretime-consuming, expensive, requireharsh conditions and much exper-tise. In addition, some chromato-graphic methods suffer frominsufficient sensitivity (Table II,see page 36). Although UVspectrophotometry is simpler,poor sensitivity is inherent.

Gatifloxacin can react withFe(III) to form a complex cation ofgatifloxacin-Fe(III), which can reactfurther with ClO4

- to form a stablecomplex ionic pair of gatifloxacin-Fe(III)-ClO4. In this paper, a FI-FAASmethod is established for the indi-rect determination of gatifloxacin

Ion Pair Solid Phase Extraction for theIndirect Determination of Gatifloxacin by Flow Injection

Flame Atomic Absorption SpectrometryHan-Ying Zhan, Mei-Zhen Ning, *Zhi-Qi Zhang, and Li-Ping Kang

Key Lab of Analytical Chemistry for Life Science of Shaanxi Province, School of Chemistry and Materials Science, Shaanxi Normal University, Xi’an 710062, P.R. China

ABSTRACT

A flow injection flame atomicabsorption spectrometry (FAAS)method with on-line ion pairsolid-phase extraction was devel-oped for the indirect determina-tion of gatifloxacin. In a flowinjection system, gatifloxacinreacted first with Fe(III) to forma 2:1 complex cation, then thecation reacted with ClO4

- to forma neutral ion pair complex ofgatifloxacin-Fe(III)-ClO4 and wasadsorbed on a C18 solid phaseextraction column, while exces-sive Fe(III) was washed withH2O to waste. The ion pair com-plex adsorbed was eluted with amixture of ethanol, water, andacetonitrile (48:32:20, v/v/v),and Fe(III) in the complex wasdetermined with FAAS. The opti-mization of various experimentalconditions was investigated.With a reaction and extractiontime of 60 seconds, the calibra-tion curve was linear, rangingfrom 0.05 to 12.50 µg mL-1 witha relative standard deviation of2.0% (n = 11, c = 5.0 µg mL-1),and the detection limit for gati-floxacin was 0.02 µg mL-1. Theanalytical frequency was 24 sam-ples per hour. The method wasvalidated by the determination ofgatifloxacin in capsules andtablets, and the recoveriesobtained were 95-106%. Thisprocedure not only provides asensitive analytical method forgatifloxacin determination, butalso offers a new approach forthe application of FAAS in theindirect determination of organicspecies.

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

33

Vol. 29(1), Jan./Feb. 2008

based on IP-SPE of gatifloxacin-Fe(III)-ClO4. Combining FI on-lineIP-SPE with FAAS can overcome thedisadvantages of FAAS in the indi-rect determination of non-metallicinorganic ions and organiccompounds, simplifying greatly thepretreatment operations, improvingsensitivity, increasing throughput,and enhancing analytical efficiency.

EXPERIMENTAL

Instrumentation

A Model TAS-986 flame atomicabsorption spectrometer (BeijingPurkinje General Instrument Co.,Ltd., P.R. China) was used for theatomic absorption measurements.The operating conditions are givenin Table I.

A Model IFIS-C intellectual flowinjector (Xi’an Ruike ElectronEquipment Corporation, P.R.China) was employed in this sys-tem. Sep-pak® plus C18 (Waters, Mil-ford, MA, USA) was used as thesolid phase extraction column. The

pH was measured with a ModelCHN868 pH meter (Thermo Orion,Waltham, MA, USA).

Chemicals

All chemicals were of analyticalreagent grade unless otherwisespecified. Water, purified with aMilli-Q™ deionization system (Milli-pore Corporation, Bedford, MA,USA), was used to prepare all solu-tions.

The gatifloxacin reference sub-stance (assigned purity of 97.2%)was purchased from the ChineseNational Institute for the Control ofPharmaceuticals (130518-200402).A stock solution of 1.0 mg mL-1 gati-floxacin was prepared by dissolvingappropriate amounts in water andstoring in a refrigerator. Workingsolutions were prepared daily bydiluting the stock solution with0.10 mol L-1 HAc-Ac- buffer solution(pH 5.0).

Ammonium ferric sulfate (Xi’anChemical Reagent Factory, P.R.China) solution (1.0 mg mL-1) was

prepared by dissolving appropriateamounts in 0.10 mol L-1 dilute HCl.Ammonium perchlorate (ShanghaiChemical Reagent Factory, P.R.China) was dissolved with water toprepare a solution of 0.5 mol L-1.The eluant was prepared by mixingethanol and water with acetonitrilein a proportion of 48:32:20 (v/v/v).

Sample and Preparation

The capsules were obtainedfrom Shandong Luoxinye Co. Ltd.(060215140) and Chengdou Hen-grui Pharmacy Co. Ltd. (051102)and the tablets from Xi’an WanlongPharmacy Co. Ltd (060201).

The total content of the tabletswas weighed and ground to a finepowder using a pestle and mortar.The average weight of one capsuleor one tablet was calculated. Anaccurately weighed 100-mgamount of gatifloxacin (from a com-posite of the mixed contents of 10capsules or tablets) was dissolvedin water and filtered through anordinary filter paper. Then the solu-tion was analyzed.

FI System Setup and OperatingProcedure

The flow injection system setupand the analytical operatingprocesses are shown in Figure 1. ASep-pak® plus C18 column was con-nected to the rotary valve. The thinend of the column was connectedto PTFE tubing toward the detec-tor, and the thick end wasconnected to Tygon® pump tubingtoward the pump. At the beginningof each working day, the C18 col-umn was flushed with methanol,water, and ethanol for 5, 5, and 10minutes, respectively; simultane-ously, the AAS instrument waswarmed up for 20 minutes.

TABLE IOptimum Experimental Conditions

Atomic Absorption SpectrometerLamp Iron Hollow Cathode LampLamp Current 3.0 mAWavelength 248.3 nmAcetylene Flow Rate 2.0 L min-1

Air Flow Rate 8.0 L min-1

Ion Pair Solid Extraction

pH of Sample Solution 5.0

Fe(III) Concentration 100 mg L-1

NH4ClO4 Concentration 0.05 mol L-1

Reaction Coil L1 Length 15 cm

Reaction and Extraction Time 60 sa

Sample Flow Rate 7.3 mL min-1

Sample Time 60 s

Eluant Ethanol, Water andAcetonitrile (48:32:20,v/v/v)

Eluant Flow Rate 2.5 mL min-1

Elute Time 60 s

a The sensitivity would be increased by with extending the time.

34

In the reaction and extractionprocess (A), P1 runs and P2 stops,the stream of the sample solution Smerges with the stream of ammo-nium ferric sulfate solution R1 toform a complex cation ofgatifloxacin-Fe(III) in the reactioncoil (L1), then merges again withthe stream of ammonium perchlo-rate solution R2 to form a stablecomplex ion pair of gatifloxacin-Fe(III)-ClO4 in another reaction coil(L2). Finally, the complex ion pair isadsorbed on the C18 solid phaseextraction column.

In the washing process (B), P2

runs and P1 stops, excessive Fe(III)is washed to be waste. In both ofthe processes, (A) and (B), P3 runs

and the carrier stream provides asteady baseline for the AAS signal.

In the elution and measurementprocess (C), the ion pair complexextracted on the C18 column iseluted by the carrier stream C (elu-ant) into the nebulizer, and theFe(III) in the ion pair complex ismeasured with FAAS. At the sametime, P1 runs and prepares for thenext analytical procedure.


Principle of the Ion Pair SolidPhase Extraction of Gatifloxacin

It is well known thatfluoroquinolones are capable offorming six-membered ring com-

plexes with certain multi-chargedmetal ions through binding of themetal ions at 3, 4-β-diketone residue(33). Experimental results showedthat gatifloxacin (Figure 2), one ofthe fluoroquinolones, could form agatifloxacin-Fe(III) complex withFe(III) in aqueous solution. The sto-ichiometric composition of the gati-floxacin-Fe(III) complex was 2:1 ofgatifloxacin:Fe(III), established withthe Job’s method of continuousvariations (34). The formation ofthe 2:1 complex betweengatifloxacin and Fe(III) was alsoconfirmed with the molar ratiomethod (Figure 3). The complexcomposition is consistent with thecomplexes between other fluoro-quinolones and Fe(III) (35).

Experimental results furthershowed that the gatifloxacin-Fe(III)complex could not be adsorbed onthe C18 solid phase extraction col-umn, but it formed a neutral ionpair compound of gatifloxacin-Fe(III)-ClO4 with ClO4

– in thepresence of perchlorate.Gatifloxacin-Fe(III)-ClO4 could beadsorbed on the C18 column andseparate from Fe(III) andFe(III)–ClO4. This suggested thatgatifloxacin could be indirectlydetermined by FAAS with SPE ofgatifloxacin-Fe(III)–ClO4 and detect-ing Fe(III) in the complex.

Atomic SpectroscopyVol. 29(1), Jan./Feb. 2008

Fig. 1. Schematic diagram of the FI system and operationprocesses for the determination of gatifloxacin with FI-IP-SPE-FAAS. (A) Reaction and extraction process, 60 s; (B) washingprocess, 30 s; and (C) elution and measurement process, 60 s. P1, P2, P3, peristaltic pump; S, sample; R1, Fe(III) solution; R2,ClO4

- (ion reagent); C, carrier (eluant); V, injection value; A,C18 column; L1 and L2, reaction coils; W, waste; FAAS, flameatomic absorption spectrometer.

Fig. 2. Structure of gatifloxacin

35

Vol. 29(1), Jan./Feb. 2008

Optimization of Reaction andExtraction Conditions

The forming reaction and solidphase extraction of the gatifloxacin-Fe(III)–ClO4 ion pair compound isof importance in this system for thedetermination of gatifloxacin withAAS. The variables investigatedwere sample solution pH, Fe(III)solution concentration, NH4ClO4

solution concentration, length ofthe reaction coil, and the flow ratesof the sample and Fe(III) solution.The figure of merit was maximumnet absorbance for the determina-tion of 10 µg mL-1 gatifloxacin.

The effect of sample solution pHon absorbance was investigatedwithin the pH range of 1.0 to 6.0.As shown in Figure 4, absorbance Aincreased with a solution pHincrease up to 3, and then variedslightly. It is considered that at thispH, gatifloxacin mainly exists in itsprotonated form and complexationwith a metal ion is accompanied byliberation of a proton from the car-boxyl group on C3 like otherfluoroquinolone analogues (33).The complex formation betweengatifloxacin and Fe(III) would beinhibited at a stronger acidic pHrange because the liberation of aproton is suppressed. The optimumreaction and extraction conditionschosen were expected to maintainthe standard and sample solutionpH value of 5.0 with 0.10 mol L-1

HAc-Ac- buffer solution.

A series of different concentra-tions of Fe(III) solution (from 10 to150 mg L-1) were tested.Absorbance A increased with anincrease in the concentration ofFe(III) solution up to 75 mg L-1, andexhibited a plateau as the concen-tration of Fe(III) solution increasedabove 75 mg L-1. An ammoniumferric sulfate solution of 100 mg L-1

Fe(III) was chosen.

Ammonium perchlorate acts ascounter ion to form ion pair com-plexes with the complex cation ofgatifloxacin-Fe(III). The sensitivity

for the determination ofgatifloxacin obviously depended onthe concentration of ammoniumperchlorate. The results obtainedfrom 1.0×10-3 to 0.10 mol L-1

NH4ClO4 showed that the bestconcentration of NH4ClO4 was0.05 mol L-1.

A longer reaction coil L1 (up to15 cm) and higher flow rates wouldenhance the absorbance for boththe sample and Fe(III) solution, butthe flow rate was limited by themaximum value possible with theinstrument used. A reaction coil L1

of 15 cm length and flow rates of7.3 mL min-1 were chosen for boththe sample and Fe(III) solution.

The absorbance increases pro-portionally when reaction andextraction times are prolonged,which enhances the sensitivity ofthe determination, but the analyti-cal frequency would be slowed. Itis suggested that different reactionand extraction times should be cho-sen based on the different concen-trations of the samples.

Optimization of ElutingCondition

The eluting variables studiedwere type, concentration and flowrate of the eluant, and eluting time.

Fig. 3. Molar ratio plot for the determination of the complex composition.

Fig. 4. Effect of sample solution pH on the determination of 10 µg mL-1

gatifloxacin.

36

The experimental resultsshowed that 60% aqueous ethanolwas the best eluant for the differentconcentrations of ethanol,methanol, acetonitrile, and acetonesolutions. Addition of a few dropsof acetonitrile into 60% aqueousethanol as the eluant should furtherincrease the absorbance, but it wasfound that the absorbancedecreased when the concentrationof acetronitrile was above 20%. Sothe optimal composition of the elu-ant selected was ethanol and watermixed with acetonitrile in the pro-portion of 48:32:20 (v/v/v).

The influence of flow rate of theeluant was tested ranging from 1.7to 3.3 mL min-1. The results showedthat the absorbance signalincreased with an eluant flow rateup to 2.5 mL min-1 and decreasedabove 2.5 mL min-1. An eluant flowrate of 2.5 mL min-1 was selected inorder to obtain maximum sensitiv-ity and throughput rate. The elutingprocess lasted 60 seconds. Areverse elution mode was used forthe FI filling column on-line precon-centration system (19,21) to avoidirregular filling, but eluting in theopposite direction of loading in thissystem would result in a bigger dis-persion of the eluted ion pair com-plex and in lower sensitivity

because the two end sizes of theSep-pak® plus C18 column are notsymmetrical.

Analytical Performance

Under the optimum experimen-tal conditions (Table I, see page33), absorbance A varied linearlywith the concentration ofgatifloxacin ranging from 0.05-12.5µg mL-1 and fitted the followingequation:

A = (0.0376 ± 0.0005) C +(0.0050 ± 0.0018), R= 0.9992

where C is the concentration of gat-ifloxacin expressed in µg mL-1. Thedetection limit (DL) was 0.02 µgmL-1 (5.33×10-8 mol L-1) ofgatifloxacin calculated as the con-centration corresponding to a signalof three times the standard devia-tion of the blank absorbance for 11injections. The precision of themethod, expressed as the relativestandard deviation, was 2.0%obtained for 11 samples eachcontaining 5.0 µg mL-1 gatifloxacin.The analytical throughput rate wasat least 24 sample injections perhour. Table II compares theanalytical performance of themethod developed in this studywith other methods used.

Interference Studies

The interferences of co-existingexcipients in the determination of5.0 µg mL-1 gatifloxacin were stud-ied under the optimum conditions.A relative error of less than 5% wasconsidered to be within the rangeof experimental error. The resultsobtained are summarized in TableIII. At the given concentrations, nointerference was observed in thedetermination of gatifloxacin.

Analytical Applications

The application of the methodwas verified by the determinationof gatifloxacin in its pharmaceuticalformulations. Simultaneously, com-parison and recovery tests were

TABLE IIITolerance Limit of Foreign

Species in the Determination5 µg mL-1 Gatifloxacin

Foreign Species Tolerated Ratio

Magnesium Stearate 100

Glucose 10,000

Talcum 100

Sucrose 10,000

Lactose 1000

Starch 100

Dextrin 5

TABLE IIAnalytical Performance of Developed Methods for the Determination of Gatifloxacin

Method Linearity DL QL Precision Reference

(µg mL-1) (µg mL-1) (µg mL-1) (%RSD)

This Work 0.05-12.50 0.02 2.0CE 20.0-60.0 1.0 <1.78 (24)HPLC 0.10-6.00 0.1 <2.77 (25)HPLC-DAD 2-20 2.3 3.4 (26)HPLC-fluorescence 0.02-2.00 0.03 2.8 HPLC 0.162-5.000 0.12 0.162 1.7-3.4 (27)HPLC 0.1-10.0, 1-150 0.1, 1.0 1.77-6.82 (28)Pressurized CEC 5.0-50.0 1.0 0.59-1.61 (29)Microbiological Assay 4.0-16.0 1.14 (30)Spectrofluorimetry 0.2-2.0 0.06 0.2 1.3 (31)UV Spectrophotometry 2.4-6.4 0.87 (32)

37

Vol. 29(1), Jan./Feb. 2008

performed to check the validity.Comparison tests were performedas described in the literature (32)with UV spectrophotometric meth-ods. Recovery tests were made byadding the standard solution of gati-floxacin to the sample solution.There was good agreementbetween the proposed and the ref-erence method. The t-test assuredthat the results of both methodsshowed no significant differences(P = 0.95). The proposed methodgave satisfactory average recoveriesas well. All results, as shown inTable IV, indicate that the proposedmethod is reliable for routine analy-sis.

CONCLUSION

In this work, a novel flow injec-tion (FI) on-line solid phase extrac-tion (SPE) procedure wasdeveloped for the extraction of ionpair complex. This procedure hasbeen combined successfully withflame atomic absorption spectrome-try (FAAS) for the sensitive determi-nation of gatifloxacin. As a methodof indirect determination of organiccompounds with FAAS, it has manyadvantages which include simpleroperation, lower detection limit,better precision, and higher samplefrequency. Compared to the analyti-cal performance with other devel-oped methods for thedetermination of gatifloxacin, theproposed method is superior in sen-sitivity and linear range. In fact, thesensitivity could be furtherenhanced by extending the reac-tion and adsorption time. This work

demonstrates that the indirectdetermination of gatifloxacin is fea-sible and that the new approach ofusing FAAS analysis has the poten-tial for routine analysis of organicspecies as well.

ACKNOWLEDGMENTS

The authors gratefully acknowl-edge financial support from theNational Natural Science Founda-tion of China (No. 20575039) andthe Doctor Base Foundation of theChinese Ministry of Education (No.20050718011).

Received May 27, 2007.

REFERENCES

1. M. Noroozifar, M. Khorasani-Motlagh,A. Taheri, and M. Homayoonfard,Talanta 71, 359 (2007).

2. N.T. Abdel-Ghani, A.F.A. Youssef,and M.A. Awady, IL Farmaco 60,419 (2005).

3. M.M. Ayad, A.A. Shalaby, H.E. Abdel-latef, and M.M. Hosny, J. Pharm.Biomed. Anal. 28, 311 (2002).

4. W.F. El-Hawary, J. Pharm. Biomed.Anal. 27, 97 (2002).

5. Z.-Q. Zhang and Y.-C. Jiang, At. Spec-trosc. 22, 429(2001).

6. L.O. Leal, O. Elsholz, R.Forteza, andV. Cerdà, Anal. Chim. Acta 573-574, 399(2006).

7. L. E. Kaercher, F. Goldschmidt,J.N.G. Paniz, E.M. Moraes Flores,and V.L. Dressler, Spectrochim.Acta B 60, 705 (2005).

8. S.R. Segade and J.F. Tyson, J. Anal.At. Spectrom. 18, 268 (2003).

9. J.L. Capelo, I. Lavilla, and C. Bendi-cho, Anal. Chem. 72, 4979 (2000).

10. E. Bolea, D. Arroyo, F. Laborda, andJ.R. Castillo, Anal. Chim. Acta 569,227 (2006).

11. Y. Rodriguez and J.F. Tyson, J. Anal.At. Spectrom. 21, 757 (2006).

12. W. Chuachuad and J.F. Tyson, J.Anal. At. Spectrom. 20, 273 (2005).

13. W. Chuachuad and J.F. Tyson, J.Anal. At. Spectrom. 20, 282 (2005).

14. A.N. Anthemidis and E.K. Martavalt-zoglou, Anal. Chim. Acta 573-574,413 (2006).

15. J.L. Capelo, I. Lavilla, and C. Bendi-cho, Anal. Chem. 73, 3732(2001).

16. O. Haase and J.A.C. Broekaert, Spec-trochim. Acta B 57, 157 (2002).

17. M. Garcia, R. Figueroa, I. Lavilla, andC. Bendicho, J. Anal. At. Spectrom.21, 582 (2006).

18. M. Noroozifara, M. Khorasani-Motlaghb, and S.N. Hosseinia, Anal.Chim. Acta 528, 269 (2005).

19. Z.-Q. Zhang, H.-T. Yan, and L. Yue,At. Spectrosc.25, 191 (2004).

20. Z.-Q. Zhang, Y.-C. Jiang, and H.-T.Yan, At. Spectrosc. 24, 27 (2003).

21. Y.-C. Jiang, Z.-Q. Zhang, and J.Zhang, Anal. Chim. Acta 435, 351(2001).

22. M. C. Carson, J. Chromatogr. A 885,343 (2000).

23. J..M. Blondeau, R. Laskowski, J.Bjarnason, and C. Stewart, Int. J.Antimicrob. Agents 14, 45 (2000).

24. R.T. Sane, S. Menon, A.R. Pathak,A.Y. Deshpande, and M. Mahale,Chromatogr. 61, 303 (2005).

25. S. Al-Dgither, S.N. Alvi, and M.M.Hammami, J. Pharm. Biomed. Anal.41, 251 (2006).

TABLE IVResults for the Determination of Gatifloxacin in Different Forms (n=3)

Sample Proposed Method UV Added Obtained Recovery

(g/capsule) (g/capsule) (g/capsule) (g/capsule) (%)

0602151 0.2149±0.0034 0.2128±0.0010 0.2000 0.1901±0.0026 95.1±1.4

051102 0.0955±0.0025 0.0975±0.0005 0.2000 0.1941±0.0066 97.0±3.2

(g/capsule) (g/capsule) (g/capsule) (g/capsule) (%)

060201 0.1021 ±0.0015 0.1079±0.0004 0.2000 0.2105±0.0011 105.3±1.1

38

26. M.L.S. Cavazos, L.Y.C. González, G.Gallegos de Lerma, and N. Waks-man de Torres, Chromatogr. 63,605 (2006).

27. H.A. Nguyen, J. Grellet, B.B. Ba, C.Quentin, and M.C. Saux, J. Chro-matogr. B 810, 77 (2004).

28. B.R. Overholser, M.B. Kays, andK.M. Sowinski, J. Chromatogr. B798, 167 (2003).

29. H.-X. Lü, X.-P. Wu, Z.-H. Xie, X.-C.Lin, L.-Q. Guo, C. Yan, and G.-N.Chen, J. Sep. Sci. 28, 2210 (2005).

30. H.R.N. Salgado, C.C.G.O. Lopes, andM.B.B. Lucchesi, J. Pharm. Biomed.Anal. 40, 443 (2006).

31. L.Y. Colunga-González, Anal. Lett.38, 2355 (2005).

32. X.-L. Wang, D. Li, and Z.-Y. Nie,China Pharmacist 8, 1014 (2005).

33. A.I. Drakopoulos and P.C. Ioannou,Anal. Chim. Acta 354, 197 (1997).

34. C.Y. Huang, R. Zhou, D.C. Yang,and P. Boon Chock, Biophys.Chem. 100, 143 (2003).

35. D.L. Ross and C.M. Riley, J. Pharm.Biomed. Anal. 12, 1325 (1994).

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

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

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

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

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

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

4. Practical Guide to ICP-MS Author: Robert Thomas, Scientific Solutions (www. scientificsolutions1.com)Published in 2004 by Marcel Dekker Ordering and price information: http://www.crcpress.com/shopping_cart/products/product_detail.asp?sku=DK2933&parent_id=1151&pc=

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

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

PerkinElmer Life and Analytical Sciences710 Bridgeport AvenueShelton, CT 06484-4794 USAPhone: (800) 762-4000 or (+1) 203-925-4602www.perkinelmer.com

008108_01

Vol. 29(1), Jan./Feb. 2008

as 29(1) 1

Documents