Application of iminodiacetate chelating resin muromac A-1 in on-line

Spectrochimica Acta Part B 58(2003) 1541–1552

0584-8547/03/$ - see front matter� 2003 Elsevier B.V. All rights reserved.doi:10.1016/S0584-8547Ž03.00101-0

Application of iminodiacetate chelating resin muromac A-1 inon-line preconcentration and inductively coupled plasma optical

emission spectroscopy determination of trace elements innatural waters�

E. Vassileva* , N. Furuta,1

Faculty of Science and Engineering, Department of Applied Chemistry, Chuo University, 1-13-27, Kasuga, Bunkyo–ku,Tokyo 112-8551, Japan

Received 4 October 2002; accepted 7 May 2003

Abstract

On-line system incorporating a microcolumn of Muromac A-1 resin was used for the developing of method forpreconcentration of trace elements followed by inductively coupled plasma(ICP) atomic emission spectrometrydetermination. A chelating type ion exchange resin has been characterized regarding the sorption and subsequentelution of 24 elements, aiming to their preconcentration from water samples of different origins. The effect of columnconditioning, pH and flow rate during the preconcentration step, and the nature of the acid medium employed fordesorption of the retained elements were investigated. A sample(pH 5) is pumped through the column at 3 mlmin and sequentially eluted directly to the ICP with 3 M HNOyHCl mixtures. In order to remove residual matrixy1

3

elements from the column after sample loading a short buffer wash was found to be necessary. The effectiveness ofthe matrix separation process was illustrated. The procedure was validated by analyzing several simple matrices,Standard River water sample as well as artificial seawater. Proposed method can be applied for simultaneousdetermination of In, Tl, Ti, Y, Cd, Co, Cu and Ni in seawater and for multielement trace analysis of river water.Recovery at 1mg l level for the determination of investigated 24 elements in pure water ranged from 93.1 to 96%y1

except for Pd(82.2%) and Pb(88.1%). For the same concentration level for seawater analysis recovery was between81.9 and 95.6% except for Hg(38.2%).� 2003 Elsevier B.V. All rights reserved.

Keywords: Inductively coupled plasma optical emission spectrometry; On-line preconcentration; Imminodiacetic acid resin; Traceelements; Natural waters

� This paper was presented at the 5th European Furnace Symposium and 10th International Solid Sampling Colloquium withAtomic Spectroscopy, held in Blagoevgrad, Bulgaria, September 2002 and is published in the Special Section ofSpectrochimicaActa Part B, dedicated to that conference.

*Corresponding author.E-mail address: [email protected](E. Vassileva).On leave from: University of Sofia, Faculty of Chemistry, 1, James Bourchier av., 1126 Sofia, Bulgaria.1

1542 E. Vassileva, N. Furuta / Spectrochimica Acta Part B 58 (2003) 1541–1552

1. Introduction

Direct determination of trace elements inextremely low concentrations by modern atomicspectroscopy techniques is in many cases difficult.This is not only due to the insufficient sensitivityof the methods, but also to the matrix effect. Manypapers published recommend the use of off-lineand on-line preconcentration and separation sys-tems as a means of reducing or eliminating spec-troscopic and non-spectroscopic interferences insample with high saline content. Nowadays thereis considerable interest in the development of newmethods or in adaptations of existing methods, foruse ‘on-line’ with the measurement instrumentw1–4x. The major advantage of the on-line procedureis in general the possibility of working in a closedsystem with a significant reduction of airbornecontamination and a fairly high sample throughout,which justifies current trends towards an increasinguse of on-line preconcentration techniques. Morerecently, the use of various types of adsorbent forthe preconcentration and separation of trace ele-ments has gained popularity because of:(i) Thehigh concentration factor achieved;(ii) The simpleoperation;(iii ) The possibility of using a closedcirculation system; and(iv) The combination withdifferent modern analytical techniques permittingin the case of inductively coupled plasma-massspectrometry(ICP-MS) and inductively coupledplasma-atomic emission spectrometry(ICP-OES)simultaneous multi-element determination.Although systems for separation and preconcenta-tion of analytes from the matrix by extractionw5xand co-precipitation methodsw6x have also beeninvestigated, the on-line column preconcentrationsystems using chelating ion-exchange resins havearoused most interest. Two chelating functionalgroups namely iminodiacetate(IDA) and 8-hydro-xyquinoline (8-HQ) have mainly been used forthe analysis of seawater. Different successfulflow–injection methods using a column of silica-immobilized 8-HQ for the determination of traceelements in natural waters have been reportedw7,8xThe disadvantage of the 8-HQ procedures are thatthere is no commercial source of 8-HQ at presentand the material must be synthesized. Severalgroups have automated the preconcentration of

analytes using chelating resins incorporated intothe flow manifold with subsequent atomic absorp-tion spectrometry(AAS) or ICP-OES w2,9–11xdetection. This method gives simultaneously mul-tielement detection limits, which are over 20 timesbetter than for conventional continuously aspiratedsystems for Ba, Be, Cd, Co, Cu, Mn, Ni and Pb.These studies show that CHELEX 100 has thedisadvantage of swelling and shrinking at differentpH value. Kingston et al.w12x demonstrated theapplicability of a CHELEX 100 column procedureto the isolation and 100-fold concentration of Cd,Co, Cu, Fe, Mn, Ni, Pb and Zn from seawater. Itwas claimed that prior to the elution of the traceelements, alkali and alkaline earth metals are elutedalmost completely from the resin with 1 M ammo-nium acetate solution. The simplicity of the pro-cedure as proposed by Kingston et al.w12x togetherwith its attractive analytical performance and widerfield of application, induced many analysts toapply the IDA resins to the trace analysis ofdifferent types of water and biological samplesw13x.Muromac A-1 is more highly cross-linked IDA

resin based on a macroporous substrate with notendency for volume changed. This is an importantadvantage over CHELEX 100. Unfortunately, theselectivity of all IDA resins is less favorable forseawater than the selectivity of 8-HQ and theremoval of matrix component is essential for thesuccess of the methodw13–15x. From the experi-ences with IDA resin to date, it becomes apparentthat the optimum experimental condition forenrichment and elution procedure is still elusive.Additional work is required to establish the bestcondition for the extraction and determination ofa large number of elements simultaneously from avariety of aqueous samples.In the present study, on-line preconcentration

and matrix separation method was investigated.The manifold incorporates a microcolumn ofMuromac A-1 resin. Attention was focused onoptimization of the conditions for the multi-ele-ment preconcentration of trace elements in differ-ent type waters, their complete elution andsubsequent simultaneous ICP-OES determination.

1543E. Vassileva, N. Furuta / Spectrochimica Acta Part B 58 (2003) 1541–1552

Table 1ICP-OES operating conditions

Plasma gas flow 16.0 l miny1

Intermediate gas flow 1.26 l miny1

Aerosol carrier gas flow 1.10 l miny1

Plasma generator frequency 27.12 MHzIncident power 1.30 kW

2. Experimental

2.1. Reagent

Standard solutions(1000 mg l for investigat-y1)

ed elements were purchased from Kanto ChemicalCo. (Tokyo, Japan). Working standard solutionswere prepared daily with distilled and deionizedwater and stored in precleaned polypropylene con-tainers. The HNO , NH OH and CH COOH3 4 3

(HOAc) used in this work were high purity grade(Tama Chemical Co., Tokyo, Japan). The waterused was Milli-Q water, prepared by further puri-fication of de-ionized water by a Milli-Q system(Milli-Q SP ICP-MS, Millipore, Tokyo, Japan).Suprapur grade salts(Suprapur, Merck, Darmstadt,Germany) were used for the preparation of testsolutions. Artificial seawater solution, which was10 times more concentrated than normal sea-water was prepared by dissolving 22.8 g ofCaCl Ø4H O (Suprapur, Merck, Darmstadt, Ger-2 2

many), 7.4 g of KCl(Suprapur, Merck, Darmstadt,Germany), 106.8 g of Mg(NO ) Ø6H O (Pro Anal-3 2 2

ysis, Merck, Darmstadt, Germany) and 274 g ofNaCl (Suprapur, Merck, Darmstadt, Germany) in1000 ml Milli-Q water (the concentration of Ca,Mg, Na and K in this solution were 5 g l , 10 gy1

l , 107.8 g l and 3.9 g l respectively)y1 y1 y1

A solution of 1 M CH COONH was prepared3 4

by diluting 14.7 ml of 96% wyv HOAc and 17 ml25% wyv NH OH solution in 250 ml Milli-Q4

water. The pH was adjusted with 2 M HNO or3

NH OH solutions. Muromac A-1 50-100 mesh, Na4

form resin was obtained from Muromachi KagakuKogyo Kaisha, Ltd(Tokyo, Japan). River waterstandard solution(natural river water in whichsome elements were added purposely) wasobtained from Analytical Chemical Society ofJapan.

2.2. Instrumentation

The ICP-OES measurements were performedwith an ICP spectrometer(Spectroflame CompactSyE with two polychromators setting grating of3600 grooves mm and 1800 grooves mm ,y1 y1

respectively, and one monochromator setting agrating of 2400 grooves mm). The polychro-y1

mators has 43 simultaneous analytical channels.The operating conditions are summarized in Table1. The analytical lines for the monochromatorwere selected on the basis of their net and back-ground intensities as well as their freedom fromspectral interference overlaps. All measurementswere made using the transient scan mode withintegration time for each element 50 ms, intervaltime 50 ms and 1200 point per peak. Peak areawas used for all quantification work.The pH adjustment was performed by means of

Basis pH meter, from Denver Instrument Company(USA).

2.3. General procedure for on-line preconcentra-tion and ICP-OES measurement

A schematic diagram of the instrumental set-upis shown in Fig. 1. The microcolumn was condi-tioned by flowing 10 ml of 1 M and 0.05 Macetate buffer solutions through it at a rate of 4ml min . During the preconcentration step(they1

volume of sample preconcentrated was varied inorder to optimize the analyte signal) the flow rateof sample was 3 ml min . After washing proce-y1

dure with 20 ml 0.1 M CH COONH at a flow3 4

rate 5 ml min , the adsorbed trace elements werey1

eluted by the injection of 1 ml of 3 M HNOy33 M HCl in ratio 3:1 and flow rate 0.5 ml min .y1

The eluted sample was introduced into the ICP,where investigated elements were measured simul-taneously. Total time for conditioning of the col-umn, preconcentration, washing procedure, elutionand determination was 25–40 min(according tothe sample volume).When seawater was analyzed to overcome the

problem with salts deposition on the tip of thecentral tube of the ICP torch, the sample passedthrough the column during the preconcentrationstep was discarded into a drain.


Fig. 1. Schematic diagram of the on-line preconcentration set-up.

2.4. Column preparation

The Muromac A-1 microcolumn was made bypacking 0.1 g of the resin into a Teflon tube(i.d.:0.4 mm, length: 40 mm). The aim in the construc-tion of the microcolumn was to prepare a columnwith high capacity, capable of handling large flowrate and allowing elution of the sample in a smallvolume. In order to decrease blank values thecolumn was washed two times with 20 ml of 3 MHNO y3 M HCl (3:1) and water. Washing proce-3

dure allows elution of potentially trace elementcontamination.In order to prevent the precipitation of trace

elements from the start of the preconcentrationprocedure and to avoid any change in the pH ofthe column content during the passage ofthe sample, the resin was treated with 1 MCH COONH solution and the column was rinsed3 4

with 0.05 M CH COONH solution. The pH of3 4

the CH COONH solution was adjusted with nitric3 4

acid to the ‘working’ pH 5. Whether the initiallyprecipitated hydroxide comes in contact with theconditioned resin they dissolve partly or complete-ly during the preconcentration step. This is thereason for the poor agreement of recovery resultsreported in the literature for Mn, Fe and Alw16,17x.

2.5. Test solutions

Pure water test sample(100 ml) was preparedby dissolving trace elements in 95 ml Milli-Q

water. After addition of 5 ml of 1 M CH -3

COONH , the pH was controlled and adjusted to4

pH 5. Artificial seawater test sample(100 ml) wasprepared by mixing 85 ml water(pH adjustedpreviously to 5), 10 ml of a solution of seawatersalts(pH 5) with 10 times concentrated seawaterand finally 5 ml of 1 M CH COONH was added.3 4

Two types of test solutions have been preparedwith concentrations of trace element in the testsamples 0.001mg l and 0.01mg l , respec-y1 y1

tively. During the preconcentration step in the caseof seawater analysis, the pH of the initial effluentshowed a sharp decreasing. Considering the acidityconstant of iminodiacetic acid, it follows that atpH 5 considerable part of the available acid groupsare protonated when seawater sample is passedthrough the column and as a consequence, theH O concentration in the liquid phase of theq

3

column increases. The changes in pH during thepreconcentration step of seawater could be pre-vented if 5 ml of 1 M CH COONH is added3 4

during dilution of 10 times concentrated seawaterand the pH of obtained artificial seawater solutionis adjusted to 5. The influence of the sample flowrate was studied using a constant flow rate of theeluent acid. Changes in flow rate of the samplebetween 1 and 6 ml min resulted in smally1

variations in the signal, if the total volume isconstant. Flow rate of 3 ml min was chosen fory1

all further investigations. Preliminary tests showedthat the sample volume was not an important


factor when the mass of the analyte arriving at thecolumn was kept constant.

2.6. Determination of exchange capacity

2.6.1. Batch determinationA separate solution for each of investigated

element(500 mg ml was prepared in ammoniumy1

acetate buffer(0.1 M at pH 5; pH 2 was used inthe case of Bi) and added to 0.05 g Muromac A-1 resin. The mixture was equilibrated with stirringovernight. The solution was decanted and theconcentration of element in the supernatant liquidwas measured vs. the original concentration. Thedecrease in the concentration of elements was usedto evaluate the capacity using the followingequation:

cyc =VŽ .i fCs

m=1000=Mr

whereC is the capacity(mmol g ), c andc arey1i f

the concentration of the elements before and afterequilibrium, respectively(mg l ), M is the atom-y1

r

ic mass of the element andV is the volume of asolution (ml) equilibrated with a massm (g) ofthe resin.

2.6.2. Dynamic determinationDry resin Muromac 1 was packed into the

column. Two molesylitre buffer solution was usedbecause of high acidity of the standard solution(100 mg l concentration). The capacity wasy1

evaluated in terms of the concentration of elementsin the solution after column. The concentrationbeyond with no further decrease after passing thecolumn was accepted to represent the dynamiccapacity limit. The dynamic capacity was evalu-ated using the equation:

cyc =VŽ .i fC sd m=1000=Mr

wereC is the dynamic capacity mmol g ,c andy1d i

c are the concentration of standard solutionsf

before and after column,V is the volume ofstandard solution(ml), m is the mass of sorbent

in the column andM the atomic mass of ther

element.

3. Results and discussions

3.1. Influence of pH

pH is a critical parameter for achieving quanti-tative preconcentration of trace elements. A pH inthe range 4.8–5.0 appeared to be the most appro-priate for preconcentration with Muromac A resinof the following investigated elements: Al, Bi, Co,Cd, Cu, Cr, Fe, Hg, Ga, In, Mo, Mn, Ni, Pb, Sc,Tl, Ti, Sn, U, V, W Y, Zn and Zr. For someelements quantitative recovery is already attainedat pH 2: Ti, Zr, U, In, Ga, Th. These observationsare in compliance with the stability constant ofthese elements with similar chelating agent suchas ethylenediaminetetraacetic acid and nitrilotri-acetic acid w18,19x. pH 5 was selected for thefuture investigation. Below this range, elutionlosses from the resin occurred, whereas at pH)5.5 precipitation of Al, Fe and Mn hydrous oxidesmay become significant. In the case of seawaterat pH 6–7 not only trace elements can be chelated,but also Ca and Mg as well. Indeed, comparedwith cations such as Ca, Mg, Na and K, theselectivity for transition elements is higher, buttheir concentrations are 5–7 orders of magnitudelower than those of the matrix elements in sea-water. As a result, trace elements are not or areonly partly sequestered on the resin. The aboveresults show the possibility of concentrating simul-taneously a number of trace elements working ata constant pH, which make Muromac A-1 resinpromising resin as a solid-phase extractant inmultielement trace analysis.

3.2. Dynamic and batch column capacityevaluation

Evaluation of the capacity of Muromac A-1 wasperformed in bath and dynamic conditions. Goodagreement between the dynamic and batch capacityfor all investigated elements was obtained(Table2). This illustrates that under the condition ofdynamic operation, the column retains analytesrapidly. Capacity values are in many cases consis-


Table 2Capacity, mmol g of Muromac A-1 resin in static andy1

dynamic conditions

Element Static Dynamicconditions conditions

Al 0.87 0.98Bi 0.42 0.94Co 1.09 1.11Cr 0.89 0.95Cu 1.32 1.27Hg 0.87 0.99In 0.87 0.83Mo 0.51 0.63Mn 0.98 1.02Ni 1.01 1.08Pb 0.87 0.98Pd 0.71 0.98V 1.20 1.11Y 1.01 0.96Zn 1.12 1.14Zr 0.38 0.34

tent with values reported from the manufacture ofthe resin. The capacity observed for Bi and Zr waslower than 1 mmol g . The possible explanationy1

is the phenomena described as ‘polymer effect’w19x. The dynamic capacity of column wasbetween 40 and 60% lower when the investigationwas performed from artificial seawater. Preconcen-tration of trace elements using ion-exchange-che-lating resin is based on its chelation underconditions where the matrix elements are onlyionogenically bound or not bound at all. Suchconditions can be obtained by changing the amountof resin. Its total capacity should be large in respectof the amounts of trace elements retained on thecolumn and in the same time sufficiently small toensure that only small part of matrix componentsis bound to the resin. In our investigation the totalamount of test elements is under resin’s capacity,whereas the amount of the matrix elements exceedsthe column capacity more than 40 times.

3.3. Elution

Since almost all of investigated elements formstable complexes(formation constants are in therange from 1=10 for Mn to 1=10 for Fe),14 25

one can expect that the acid concentration required

for elution increases with increasing the formationconstant of the element-IDA complex. When 3 MHNO was used quantitative recovery was obtained3

for the majority of investigated elements. However,complementary addition of HCl improves therecovery of Hg, Sn, W and Pd. To minimize thetime needed for quantitative elution and eluatedispersion during elution step the volume of eluentshould be reduced. It is possible only if a highlyefficient eluent is used. The injection of smallervolume eluent contributes to obtain higher precon-centration factor, to reduce blank value and ismore convenient for work in flow conditions. Ascan be noticed from Fig. 2, 1 ml 3 M HNOyHCl3

(3:1) is appropriate eluent volume for elution andthe subsequent ICP-OES determination. The influ-ence of the flow rate on the recovery of studiedelements was checked in the range of 0.2–2 mlmin . Maximum recovery was obtained in rangey1

from 0.2 to 1 ml min . Thus, a flow rate of 0.5y1

ml min was used for elution.y1

3.4. Effect of the matrix

The problematic nature of seawater(0.47 moll sodium chloride and 0.01 mol l potassiumy1 y1

chloride) w16,20,21x made it incompatible withdirect ICP-OES analysis. The cationic part of thismatrix is easily removed by chelating ion exchangebecause of the low complex formation constant ofalkali ions with IDA. In the other side theexchange sites not occupied by trace metals remainas magnesium or calcium complexes and theiramount is dominated by the exchange capacity ofthe separation column. The higher exchange capac-ity of the separation column causes a higherremaining amount of alkaline earth metals. Thisproblem is quite important in on-line application.As one can expect the concentration of the analyteand the remaining matrix elements increases duringthe elution step. This may lead to serious interfer-ence problems by molecular ions or to changes inthe plasma conditionsw16,22x.Using the fact that alkaline earth metals are

partially removed by buffer solutionw14,20x oneinvestigation of residual matrix concentration afterseawater preconcentration and optimization of suit-able washing procedure was undertaken. Fig. 3


Fig. 2. Elution histogram for 0.1mg of investigated elements, bound to 0.1 g of Muromac A-1 resin. Eluent-3 M HNOyHCl acid3

mixture in ratio 3:1.

Fig. 3. Recovery of trace elements as a function of the CH COONH concentration, used for column washing(pH 5.2). (d) – Bi,3 4

(h) – Mo, (�) – Pb,(j) – Tl, (m) – V.


Fig. 4. Matrix removal as a function of the CH COONH concentration, used for column washing after preconcentration on the3 4

microcolumn of 100 ml artificial seawater.(�) – Ca,(j) – Mg, (m) – Na.

indicates that the recovery of trace elements in ourinvestigation tends to decrease, when molarity ofwashing CH COONH solution increases more3 4

than 0.1 M. The recovery of Pb, Tl and Bi appearsto be more susceptible to concentration ofCH COONH solution than the recovery of V and3 4

Mo. Washing solution with concentration above0.1 M does not contribute significantly for furthermatrix removal—Fig. 4. In this study 0.1 MCH COONH was selected as the appropriate con-3 4

centration of the washing solution.It is of interest to establish the final proportion

of Na, Ca and Mg on the column after 100 mlseawater has been passed through the column. Itmight be assumed that Ca is dominant in the finalsolution as the affinity for Ca ions is the highestamong these three metal ions. Pai et al.w23x foundfor CHELEX 100 resin that after a large amountof seawater had passed through, the resin wereequilibrated with all ions and the ratio of majorions on the exchange sites gradually reached con-stant values no matter what the original concentra-tion was.In the present investigation Ca, Mg and Na was

removed by 95, 98, 99.9%, respectively, and thetotal matrix concentration in the acidic eluate isapproximately 0.6%. Such a concentration is suf-

ficiently low to exclude the occurrence of interfer-ences in on-line ICP-OES analysis of the studiedelement during the measurement step. Ca and Mgions are completely retained on the resin, whentheir total concentration in the sample is undercapacity of the column. Following the washingprocedure in this case does not contribute for thematrix removal at all.Fig. 5 shows matrix removal as a function

of volume of washing solution (0.1 MCH COONH , pH 5.2) used for 4-mm diameter3 4

column, containing 0.1 g resin. It can be concludedthat 20 ml CH COONH assures considerable3 4

reduction of residual matrix concentration, whereasa larger volume does not contribute significantlyto further matrix removal.As regards the amount of resin we should stress

again that in the enrichment step the amount ofresin in the column must be kept to a minimumfor several reasons:(i) amount of acidic andwashing solutions required for the elution of traceelements and washing of the column, respectively,increases with increasing the amount of resin;(ii)the amount of matrix elements not removable fromthe resin also increases;(iii ) total time of precon-centration and ICP determination of trace elementsbecome longer; and(iv) the loss of retained trace


Fig. 5. Matrix removal as a function of the volume of 0.1 M CH COONH , used for column washing after preconcentration on the3 4

microcolumn of 100 ml artificial seawater.(�) – Ca,(d) – K, (j) – Mg, (m) – Na.

elements is connected with increasing the volumeof a washing solution.

3.5. Preconcentration efficiency

The efficiency of preconcentration system wasdetermined by passing through the column twosamples with volume 50 ml and concentration of1 mg l and 10mg l , respectively, at pH 5.y1 y1

The integrated signals for 0.05mg ml and 0. 5y1

mg ml for each of the analytes in 1% HNOy13

were compared with integrated signals after pre-concentration. All analyses were performed induplicate and the integrated intensities were cor-rected for the background. Since the integratedintensities were compared, two values should bethe same, irrespective of the preconcentration fac-tor, if the preconcentration is quantitative andsensitivity remains constant. Table 3 gives theresult of the comparison for pure water solutionand artificial seawater. In the case of pure water,the quantitative recovery(above 95%) wasobtained for all investigated elements, except forPd, Pb and Mn for both concentration levels. Theprecision, measured as %R.S.D., was between 1.5and 5%(except for Pd 7.1% and for W 5.9% at

concentration level 1mg l ). In the case ofy1

artificial seawater, the investigated analytical pro-cedure allows quantitative recovery for some ele-ments – Co, Cu, Ga, In, Sc, Ti and Zr. Theprecision for preconcentration of 50 ml seawatersample with a concentration of 10mg l wasy1

between 2.9 and 14.9%, where as the precision forthe preconcentration of the same sample volumeat concentration level 1mg l was in the rangey1

of 4.9–18.9%.

3.6. Blank values and detection limits

The accurate determination of ng ml levelsy1

of transition metals in natural waters required lowand reproducible procedural blanks. In Table 4,the blank associated with on-line preconcentrationand ICP-OES determination of 100-ml bufferedwater sample are presented. Blank values observedfor Pd, Sn, W and Zr are very close or below theirdetection limits.The limits of detection for investigated elements

are also shown in Table 4. The limits of detectionfor studied elements by direct aspiration are com-pared to those obtained with described analytical


Table 3Recovery(%)"R.S.D.(%) of metal ions in the preconcentration procedure, using Muromac A-1 microcolumn, Sample volume 50ml, ns3

Element Wavelength Pure water Pure water Seawater Seawater(nm) (1 mg l )y1 (10 mg l )y1 (1 mg l )y1 (10 mg l )y1

Al 167.079 96.0"4.0 97.0"2.8 90.3"7.3 90.1"3.6Bi 223.061 95.1"3.6 97.3"2.6 89.4"4.6 83.9"5.1Cd 226.502 95.9"4.8 98.0"4.6 89.9"6.1 92.1"4.2Co 228.616 95.6"4.2 97.0"3.9 95.1"6.7 96.9"5.8Cu 654.738 96.1"4.7 96.8"2.4 95.6"7.2 96.2"2.9Fe 259.94 93.8"4.6 95.6"4.4 86.0"8.9 89.6"6.5Ga 417.206 95.0"4.2 96.2"2.1 95.5"4.9 95.8"3.3Hg 184.950 95.2"4.9 96.0"4.4 38.2"18.9 48.2"14.9In 325.609 96.6"5.0 97.6"3.3 95.6"5.9 95.6"4.0Mn 257.610 90.0"2.4 91.3"5.5 89.1"14.2 69.3"12.6Mo 202.030 93.1"5.0 95.2"5.0 82.1"7.5 84.9"6.0Ni 231.604 95.0"4.3 96.0"4.0 95.3"7.9 92.3"7.9Pb 220.351 88.1"3.6 90.1"3.6 83.1"5.6 90.0"4.1Pd 340.458 82.2"7.1 83.9"5.1 81.9"6.6 83.9"5.1Sn 189.989 95.6"4.0 97.0"3.9 89.1"6.7 91.9"5.8Sc 722.768 96.1"3.6 96.1"3.2 95.1"4.9 95.4"3.7Ti 334.941 95.0"5.7 96.1"4.7 95.4"7.2 95.6"6.2V 292.464 94.8"5.9 96.8"3.6 83.0"11.9 86.0"5.9Tl 190.864 92.0"4.8 95.0"4.2 90.5"5.5 91.5"4.5Y 321.669 95.2"4.9 96.0"4.4 94.2"11.9 95.2"14.9W 207.916 93.5"5.9 95.6"4.0 94.0"7.2 95.6"4.0Zn 213.856 96.0"4.4 97.0"2.4 88.1"5.2 90.1"4.2Zr 213.856 95.0"6.2 95.1"5.4 93.1"6.5 87.1"4.5

Table 4Absolute amount of procedural blank and limits of detection(LOD)

Element Blank (ng) LOD (ng ml )y1 Element Blank (ng) LOD (ng ml )y1

Al 3.54 0.005 Pb 1.4 0.01Bi -LOD – Pd -LOD –Cd 0.51 0.001 Ni 0.34 0.006Co 0.40 0.003 Sn -LOD –Cr 0.31 0.003 Ti 0.22 0.002Cu 0.62 0.007 Tl -LOD –Fe 0.66 0.008 V 1.21 0.006Hg 2.0 0.02 W -LOD –In 0.50 0.001 Y 0.36 0.006Mn 0.27 0.001 Zn 0.28 0.006Mo 0.81 0.009 Zr -LOD –

*Blanks are based on the on-line preconcentration of 100 ml buffered water. Limits of detection(ng ml ) are calculated as threey1

times standard deviation of the concentrations of investigated elements in buffered water(ns3), after direct nebulization.

procedures. For all investigated elements, the col-umn blank seems to be a limited factor. Peak areadetection limits were approximately 20–200 timeslower to those obtained by continuous nebulizationof the same elements.

The limit of detection could be still improvedby reducing the blank value. Procedural blankcould be minimized by further purification of theresin or by reducing the volume of the samplesolutions.


Table 5Results obtained for the analysis of certified reference materialafter preconcentration with Muromac A-1 microcolumn,ns3

Element Results obtained Certificate values(mg l )y1 (mg l )y1

Al 62.8"2.8 61.0"2.0Cd 0.67"0.10 1.0"0.02Cr 10.27"0.30 10.1"0.2Cu 10.39"0.30 10.5"0.2Fe 58.60"0.32 57.0"2.0Mn 3.89"0.12 5.4"0.1Ni 9.76"0.50 10.2"0.3Pb 9.74"0.38 9.9"0.2Zn 11.59"0.30 11.3"0.4

Table 6Accuracy and precision of the on-line flow injection precon-centration and ICP-OES determination of trace elements inpure and seawater, using Muromac A-1 microcolumn,ns3,Sample volume 100 ml, Amount added 2.00mg

Element Amount found Amount found(mg"S.D.) (mg"S.D.)pure water seawater

Al 1.92"0.13 1.71"0.18Bi 1.93"0.10 1.68"0.12Cd 1.96"0.12 1.89"0.16Co 1.91"0.10 1.83"0.20Cu 1.94"0.12 1.88"0.16Fe 1.94"0.18 1.66"0.10Hg 1.92"0.20 0.98"0.25In 1.93"0.12 1.96"0.12Mn 1.84"0.09 1.32"0.21Mo 1.82"0.26 1.69"0.14Ni 1.91"0.14 1.78"0.23Pb 1.87"0.22 1.56"0.12Pd 1.52"0.21 1.11"0.20Sn 1.92"0.16 1.38"0.12Ti 1.93"0.11 1.86"0.16Tl 2.07"0.13 1.83"0.19V 1.94"0.07 1.72"0.10W 1.79"0.17 1.22"0.22Y 1.86"0.15 1.87"0.14Zn 1.95"0.17 1.79"0.18Zr 1.95"0.14 1.84"0.12

The preconcentration procedure was validatedby analysis of one certified reference material—river water from Analytical Chemical Society ofJapan. The analytes were quantified by externalcalibration against acidified multi-element purewater standard processed through the same mani-

fold. Three repeated measurements were made foreach element. The calibration curb showed goodlinearity in the concentration range 0–20 ngml . Results for the certified reference materialy1

analysis are given in Table 5. Good agreementbetween the found and certified values wasobtained for all elements measured, except for Mnand Cd. A consistently low result for Mn isprobably due to partial hydrolysis of Mn at pH 5.Table 6 shows the results of the added–found

method applied on pure and artificial seawatersamples for investigated elements. A good agree-ment with the added–found value was achievedfor all the investigated elements(except for Pd)in the case of pure water. For seawater, the resultsbecame somewhat erratic for certain elements.Speciation changes in this complex matrix areinvolved – the formation of stable chloride com-plex ions in seawater could be one of the reasonsfor low recovery of Hg, Sn, Pb and Bi(see alsoTable 3). Good agreement with the added–foundvalue was achieved for Co, Cu, In, Ni, Y and Zr-elements with high value of the formation of theMe–IDA complex constant.For other investigated elements in the case of

seawater analysis the method of standard additionis required.

4. Conclusions

The use of a chelating ion exchange with appro-priate exchange capacity is in many aspects favor-able for the analysis of natural waters. Rapidsurface reactivity and wide elemental applicationof Muromac A-1 made this resin a highly effectivereagent for preconcentration procedure undertakenin this study. It has been shown that the numberof elements were preconcentrated quantitativelyand the proposed method can be applied succes-sively for trace analysis of different types ofnatural waters(river water, tap water, rainwater).In the case of seawater, the complex matrix causedproblems in the preconcentration step and for on-line ICP-OES determination of some elements thestandard addition method is required. For deter-mination of In, Tl, Ti, Y, Cd, Co, Cu and Ni inseawater, the described method can be applieddirectly. The preconcentration factor was between


20 and 200 times, and the detection limits are inorder of 0.001–0.02 ng ml .y1

References

w1x M.S. Jimenez, R. Velarte, J.R. Castillo, Performance ofdifferent preconcentration columns used in sequentialinjection analysis and inductively coupled plasma-massspectrometry for multielemental determination in sea-water, Spectrochimica Acta Part B 57(2002) 391–402.

w2x L.C. Azeredo, R.E. Sturgeon, A.J. Curtius, Determina-tion of trace metals in seawater by graphite furnaceatomic absorption following on-line separation and pre-concentration, Spectrochim. Acta Part B 48(1) (1993)91–98.

w3x D.B. Taylor, H.M. Kingston, D.J. Nogay, D. Koller, R.Hutton, On line solid-phase helation for the determina-tion of 8 metals in environmental waters by inductivelycoupled plasma-mass spectrometry, J. Anal. At. Spec-trom. 11(1996) 187–191.

w4x S.N. Willie, H. Tekgul, R.E. Sturgeon, Immobilizationof 8-hydroxyquinoline onto silicone tubing for thedetermination of trace elements in seawater using flow-injection inductively coupled plasma-mass spectrometry,Talanta 47(1998) 439–445.

w5x M.N.A. Dasilva, V.L.A. Frescura, A.J. Curtius, Deter-mination of trace elements in water samples by ultra-sonic nebulization inductively coupled plasma-massspectrometry after cloud point extraction, Spectrochim.Acta Part B 55(7) (2000) 800–813.

w6x H. Sawatari, E. Fujimori, H. Haraguchi, Multielementdetermination of trace elements in seawater by galliumcoprecipitation and inductively coupled plasma-massspectrometry, Anal. Sci. 11(1995) 369–374.

w7x B.W. Wen, X. Shan, S. Xu, Preconcentration of ultratra-ce rare earth elements in seawater with 8-hydroxyquin-oline immobilized on polyacrylonitrite hollow-fibermembrane for determination by inductively coupledplasma-mass spectrometry, Analyst 124(1999)621–626.

w8x S.D. Lofthouse, G.M. Greenway, S.C. Stephen, Minia-turization of a matrix separationypreconcentration pro-cedure for inductively coupled plasma-massspectrometry using 8-hydroxyquinoline immobilized ona microporous silica frit, J. Anal. At. Spectrom. 14(1999) 1839–1842.

w9x S. Caroli, A. Alimonti, F. Petrucci, On-line preconcen-tration and determination of trace elements by flowinjection—inductively coupled plasma-atomic emissionspectrometry, Anal. Chim. Acta 248(1991) 241–249.

w10x X. Wang, R. Barns, Chelating resins for on-line flowinjection pre-concentration with inductively coupledplasma-atomic emission spectrometry, J. Anal. At. Spec-trom. 4 (1989) 509–518.

w11x D. Beauchemin, S.S. Berman, Determination of traceelements in reference water standards by inductivelycoupled plasma-mass spectrometry with on-line precon-centration, Anal. Chem. 61(1989) 1857–1862.

w12x H.M. Kingston, I.L. Barnes, T.Y. Brady, T.C. Rains,M.A. Champ, Anal. Chem. 50(1978) 2064.

w13x S.M. Nelms, G.M. Greenway, D. Koller, Evaluation ofcontrolled-pore glass-immobilized iminodiacetate as areagent for automated on-line matrix separation forinductively coupled plasma-mass spectrometry, J. Anal.At. Spectrom. 11(1996) 907–912.

w14x R.A. Nickson, S.J. Hill, P.J. Workfold, Behavior ofmatrix cations(Ca , K , Mg and Na ) during on-2q q 2q q

line preconcentration and atomic spectrometric detectionof trace metals in natural waters, Anal. Chim. Acta 351(1997) 311–317.

w15x P. Hashemi, A. Olin, Equilibrium and kinetic propertiesof a fast iminodiacetate based chelating ion exchangerand its incorporation in a FI-ICP-OES system, Talanta44 (1997) 1037–1053.

w16x J.W. McLaren, J.W.H. Lam, S.S. Berman, K. Akatsuka,M.A. Azereto, On-line method for the analysis ofseawater for trace elements by inductively coupledplasma-mass spectrometry, J. Anal. At. Spectrom. 8(1993) 279–285.

w17x J.W. Jones, S.G. Capar, T.C. O’Haver, Critical evaluationof a multi-element scheme using plasma emission andhydride evolution atomic-absorption spectrometry forthe analysis of plant and animal tissues, Analyst 107(1982) 353–377.

w18x M. Pesavento, R. Biesuz, M. Gallori, A. Profumo,Sorption mechanism of trace amounts of divalent metalions on a chelating resin containing iminodiacetategroups, Anal. Chem. 62(1993) 2522–2527.

w19x A. Yuchi, T. Sato, Y. Morimoto, H. Mizuno, H. Wada,Adsorption mechanism of trivalent metal ions on che-lating resins containing iminodiacetic acid groups withreference to selectivity, Anal. Chem. 69(1997)2941–2944.

w20x A. Seubert, G. Petzold, J.W. McLaren, Synthesis andapplication of an inert type of 8-hydroxyquinoline-basedchelating ion exchanger for sea-water analysis using on-line inductively coupled plasma mass spectrometrydetection, J. Anal. At. Spectrom. 10(1995) 371–379.

w21x K.E. Jarvis, A.L. Gray, R.S. Houk, Handbook of Induc-tively Coupled Plasma-Mass Spectrometry, Blavkie,Glasgow, 1992.

w22x S. Nakashima, R.E. Sturgen, S.N. Willie, S.S. Berman,Determination of trace metals in seawater by graphitefurnace atomic absorption spectrometry with preconcen-tration on silica—immobilized 8-hydoxyquinoline in aflow-system, Fresenius J. Anal. Chem. 330(1988)592–595.

w23x S.C. Pai, P.Y. Whung, L.R. Lai, Preconcentration effi-ciency of Chelex-100 resin for heavy metals in seawater.Part 1. Effects of pH and salts on the distribution ratiosof heavy metals, Anal. Chim. Acta 211(1988) 257–270.

Documents

Application of iminodiacetate chelating resin muromac A-1 in on-line