ANALYST, FEBRUARY 1989, VOL. 114 137 Determination of Iron Species in Wine by Ion-exchange Chromatography - Flame Atomic Absorption Spectrometry Radmila Ajlec and Janez Stupar Joief Stefan institute, E. Kardelj University, Jamova 39, 61000 Ljubljana, Yugoslavia The direct coupling of ion-exchange chromatography to flame atomic absorption spectrometry (AAS) has been achieved by employing a Babington type nebuliser. The system enables all the processes on the column to be followed directly at flow-rates of between 1 and 5 ml min-I. The potential of the system was investigated for the determination of various iron species in synthetic samples containing iron(1l) and iron(ll1) in ionic or chelated form by employing various ion-exchange (Dowex 50-X8, Dowex 1 -X8) and sorptive (Amberlite XAD-2) resins, respectively.In some instances where direct coupling was impossible, owing to the physical properties of the effluent or eluent, conventional analyses of chromatographically separated iron species were performed by flame AAS. The optimum concentration range, limit of detection and reproducibility of measurement were also determined for a particular column capacity. When direct coupling was employed, the detection limit for the separated iron species was 15 ug with a relative standard deviation (RSD) of +3% and, using the conventional method of analysis, 2-5 pg with an RSD of 2 1%. On the basis of these results the system was applied to the determination of the ratio of iron(ll) to iron(ll1) in wines. Keywords: Iron speciation; ion-exchange chromatography - flame atomic absorption spectrometry; Babington nebuliser; wines The determination of metal species is a complex problem that is specific for a particular sample.A possible approach to surmount such difficulties efficiently and rapidly is to couple various chromatographic and optical spectrometry methods. The main obstacle to the efficient direct coupling of ion chromatography and/or liquid chromatography to flame atomic absorption spectrometry (AAS) lies in the difference between the optimum flow-rates through the column and the conventional pneumatic nebuliser. Several possibilities for overcoming this problem have been investigated. Direct coupling by modifying the flow-rates through the chromato- graphic column and the pneumatic nebuliser of a flame AAS system' makes it impossible to obtain the optimum conditions of measurement for either liquid chromatography or flame AAS.By employing optimum flow-rates for liquid chromato- graphy, the noise level becomes troublesome owing to the resulting low flow-rates through the pneumatic nebuliser used in flame AAS.2 The system employed for the atomic absorption injection method has also been used to connect a liquid chromatography column to flame AAS.3.4 In this instance drops of eluent fell into a funnel attached to the nebuliser capillary and were aspirated sequentially into thc flame. The peak height of the absorption signals corresponded to the concentration of the measured metal in a particular drop, and the sum of all the absorption signals to the total amount of a particular metal species.Some workers5.h employed a vented capillary tube to connect a high-perfor- mance liquid chromatography (HPLC) column to the pneu- matic nebuliser of a flame AAS system. Ebdon et al.7 constructed a special interface consisting of platinum wire spirals mounted on a rotating disc to couple HPLC to flame AAS. By rotation of the disc, the drops from the column were captured on the platinum wires, desolvated using a micro- Bunsen burner and introduced sequentially into the flame. The direct coupling of an ion-exchange chromatography column to the nebuliser of a flame AAS system has been described previously.%"' The flow of eluent through the column was controlled by the aspiration rate of the pneumatic nebuliser. A method of coupling gel chromatography directly to the pneumatic nebuliser of a flame AAS system has also been investigated." An additional water reservoir was con- nected to the tube used to couple the column and nebuliser in order to balance the difference in flow-rates. Although the noise level decreased the sample was diluted.The aim of this work was to develop a novel approach for the direct coupling of ion chromatography to flame AAS. A Babington nebuliser was employed and various ion-exchange (Dowex 50-X8, Dowex 1-XS) and/or sorptive (Amberlite XAD-2) resins were investigated for use in iron speciation. The potential of the proposed coupling technique was evaluated and the system was applied to iron speciation in wines. Experimental Apparatus Detection system For direct coupling of ion-exchange chromatography to flame AAS, an atomic absorption spectrometer was constructed, consisting of a Babington type nebuliser.12 a glass spray chamber, Varian hollow-cathode lamps, a 6-cm slot burner, a Carl Zeiss Jena SPM-2 grating monochromator, a Hamamatsu K-213 photomultiplier tube, a Varian AA-6 indicating module (Type IM-6), a Hewlett-Packard 17501 A recorder and an electronic integrator. 13 A Varian AA-5 atomic absorption spectrometer was employed for iron measurement when direct coupling was not possible.The instrument parameters of the two spectrometers are given in Table 1. Table 1. Instrument parameters used for measurement of iron absorbances Parameter Wavelengthlnm . . Spectral band passinm LampcurrentlmA . . Burnerslot/cm .. . . Flame . . . . . . Height above the burner topimm Nebuliser . . . . Flow-rate . . . . Background correction . . . . Spectrometer constructed for direct coupling 248.5 0.2 7 6 Air - acetylene 10 Babington type Variable (1-4.5mlmin-') Deuterium hollow- cathode lamp Varian AA-5 spectrometer 248.5 0.2 7 1 0 Air - acetylene 1 0 Pneumatic type Fixed ( 5 ml min 1 ) Deuterium hollow- cathode lamp138 ANALYST, FEBRUARY 1989, VOL. 114 Separution system The system was developed for the separation of iron species, employing ion-exchange [Dowex 50-X8 (5&100 mesh) and Dowex 1-X8 (50-100 mesh)] and/or sorptive [Amberlite XAD-2 (20-50 mesh)] resins. Glass columns (6 ml) (75 X 10 mm i.d.) were used for all measurements. This volume ensured that the capacity of the columns was not exceeded, with respect to the concentration of cations and anions in the wines.A peristaltic pump (Ismatec MS 4 Reglo), which allowed variable flow-rates of between 1 and 4.5 ml min-1 to be used, was connected to one end of the column. The other end was coupled directly to the Babington nebuliser. The system enabled the peristaltic pump to be joined either to the bottom or the top of the column. The direct coupling of the ion-exchange column to the Babington nebuliser is shown schematically in Fig. 1. Reagents Suprapur (Merck) acids and doubly distilled water were used for the preparation of sample and standard solutions. All other chemicals were of analytical-reagent grade. A standard iron(II1) stock solution (1000 p.p.m.) was prepared by dissolving 7.22 g of Fe(N0&.9H20 in 10 ml of concentrated nitric acid and diluting to 1000 ml with water.The solution was standardised by titration with EDTA. A standard iron(I1) stock solution (1000 p.p.m.) was prepared by dissolving 4.96 g of FeS04.7Hz0 in 5 ml of concentrated sulphuric acid and diluting to 1000 ml with water. The solution was standardised by titration with EDTA. A standard iron(I1) solution (100 p.p.m.) of iron - phenan- throline was prepared by dissolving 2.5 g of 1,lO-phenanthro- line monohydrochloride in water, adding 10 ml of iron(I1) solution (1000 p.p.m.) and diluting to 100 ml with water. Dowex 50-X8 (50-100 mesh) (hydrogen form) and Dowex 1-X8 (5(%100 mesh) (chloride form) ion-exchange resins and Amberlite XAD-2 (20-50 mesh) sorptive resin (Fluka) were used.To \ Th ree-way Drain \ stopcock Babington nebuliser Peristaltic Pump Fig. 1. nebuliser Direct coupling of an ion-exchange column to a Babington 2.25 M HCI 2.0 M HCI n 1.75 M HCI 0.3 1 . . O L I -Time Fig. 2. Elution curves for Fell': resin. Dowex 50-X8.5&10O mesh, in 0 . 1 M HNO,; column, 6 ml; and flow-rate, 4.5 ml min- l . Standard solution. Fe(NO,),.9H,O. 300 pg of Fe Determination of Iron Species In general iron can exist in two oxidation states, i.e., iron(I1) and iron(II1). The possible iron species in aqueous solution14 are hexaaquairon(I1) and hexaaquairon(II1) ions, various negatively and positively charged iron(I1) and iron(II1) complexes and uncharged organic compounds of iron(I1). The predominant iron species depends on the oxidation - reduction conditions, and on the concentration and stability constants of the anions present in the solution.Several methods have been employed for the determination of iron(I1) and iron(II1) in their mixtures, based on spectrophotometry,l5 ion-exchange colorimetry with spectrophotometric detection16 or flow injection analysis," but there are few reports describing the separation of iron(I1) and iron(II1) by ion-exchange chromat- ography. The cation-exchange18 and anion-exchange19 distri- bution coefficients of several cations including iron(T1) in different acidic media and selected data for ion-exchange separations20 have been reported. Iron adsorbed on organic matter in sea water has been concentrated on the sorptive resin Amberlite XAD-2.21 In addition, iron has been deter- mined by forming a l,l0-phenanthroline complex, which was then extracted from aqueous solutions on adsorbent Amber- lite XAD-2.22 This paper describes a system for the determination of iron species employing cation-exchange, anion-exchange and sorp- tive resins.Separation of Ionic Iron(I1) and Ionic Iron(II1) on Dowex 50-X8 Cation-exchange Resin The directly coupled system was used. The column was conditioned and purified by washing it with various concentra- 0.3 I I I -Time 0 ' Fig. 3. flow-rate 3.5 ml min 1 . Standard solution, Fe(N01)q.9H20. 300 vg of Fe Reproducibility of measurement for Fc"': resin. Dowcx 50-X8, 50-100 mesh, in 0.1 M HNO,; column. 6 ml; eluent. 2.25 M HCl; andANALYST. FEBRUARY 1989, VOL. 114 a, 0.2 e z 0 C m a o l - 139 D - 200 pg Fe 100 pg Fe 50 Fe 25 pg Fc 0.3 1 min - 300 pg Fe -Time 0 ' Fig.4. 4.5 ml min-I. Standard solution, Fe(N03),.9H20 Calibration graph for Fe11': resin. Dowex SO-X8, SO-100 mesh, in 0.1 M HNO,; column, 6 ml; eluent, 2.25 M HCl; and flow-rate, 0.4 I I 0 ' +Time Fig. 5 . Elution curves for: A, 300 pg of Ferr (as FeSO4.7H2O); B, 300 ug of Fe"' [as Fe(N0,),.9H20]; and C, 300 pg of Fell f Fell' (1 : 1). Resin, Dowex S0-X8, 5&100 mesh, in 0.1 M HN03; column, 6 ml: eluent, 2.25 M HCI: and flow-rate 4.5 ml min 1 tions (0-3 M) of HCl and doubly distilled water. On the basis of distribution coefficient data18 the resin was washed with 25 ml of 0.1 M HN03 prior to the analysis. Standard solutions of iron were also prepared in 0.1 M HN03. A &300-pg amount of iron was bound on the resin at a flow-rate of 1 ml min-1.Then, 5 ml of solvent were passed through the column at the same flow-rate followed by a further 15 ml of solvent at a flow-rate of 3.5 ml min-1 to separate the non-sorbed species. The optimum conditions for measurement were obtained when iron was eluted at a counter flow-rate of 4.5 ml min-I. The peak area was integrated with an electronic integrator until the signal returned to the base line. A reagent blank was integrated similarly. Sorption and elution of iron(IZ) and iron(III)from the resin Iron(II1) nitrate and iron(I1) sulphate standard solutions were used. The direct registration of atomic absorption signals showed that quantitative sorption of up to 300 pg of iron(TT1) and/or iron(I1) ions was obtained on the resin column in 0.1 M HNO? and in the presence of 0.1 M chloride ion, 0.1 M citrate ion or 0.01 M oxalate ion.The optimum conditions for elution were investigated by employing various concentrations of HCl. The results obtained for iron(II1) are shown in Fig. 2. Iron(I1) was eluted similarly and it was found that the optimum eluent for desorption of iron(I1) and iron(II1) from the resin was 2.25 M HCl. Reproducibility of measurement, calibration graph and detec- tion limit The reproducibility of measurement for an iron(II1) nitrate standard solution (300 pg) is shown in Fig. 3. It is evident that the proposed method has good reproducibility of measure- ment with a relative standard deviation (RSD) of k3%. The calibration graph for iron(II1) is shown in Fig. 4.A linear relationship between the peak area and the amount of iron(II1) in the standard solution was obtained in the range 0-300 pg of iron. The detection limit, calculated on the 30 basis, was 15 pg of iron(II1) ion. Sensitivity of measurement of iron(ZZI), iron(II) and their mixtures The sensitivity of measurement was compared for 300 pg of iron(II1) ion, 300 pg of iron(I1) ion and 300 pg of both iron(II1) and iron(I1) ions in a mixture (1 + 1). all in 0.1 M HN03. The recorder traces obtained are shown in Fig. 5 and it can be seen that both iron(TI1) and iron(I1) ions are eluted simultaneously under the described experimental conditions. In mixtures containing various concentrations of iron(I1) and iron(TI1) ions in 0.1 M HN03 only the total ionic iron can be determined, whereas the uncharged organic iron passes through the cation-exchange resin column.It was found that ionic iron could be calibrated with a standard solution containing iron(II1) or iron(I1) ions. Separation of Ionic Iron(II1) on Dowex 1-X8 Anion-exchange Resin The same system and conditions of measurement were employed as described for the determination of ionic iron on Dowex 50-X8 resin. The resin was purified by washing it with 0-3 M HC1 and doubly distilled water. Sorption and elution of negatively charged iron complexes from the resin Various negatively charged complexes were studied. It was found that the sorption of iron(II1) ion as FeCL- in strongly acidic (3 M HCl) media was not quantitative. Although iron(II1) ion was quantitatively sorbed on the resin in 0.1 M oxalic acid, clogging of the nebuliser occurred during elution with 0.1 M HC1 owing to crystallisation of the oxalic acid.The use of citric acid as a complexing agent for iron(TT1) was then investigated. It was found that quantitative sorption of iron(II1) ion occurred when both the standard solutions and the resin were prepared in 0.5 M citric acid. Various concentrations of HC1 (0.5-2.25 M) were also investigated for the elution of iron (Fig. 6) and 2.0 M HC1 was found to be optimum. Reproducibility of measurement and detection limit The reproducibility of measurement was tested by performing six successive determinations of 300 pg of iron(II1) [as140 ANALYST, FEBRUARY 1989, VOL. 114 2.25 M HCI 2.0 M HCI 1.5 M HCI 0 ' I - Time Fig.6. Standard solution. Fe(N0,),.9H20. 30C pg of Fc Elutlon curyes for FeIII: resin, Dowex 1-XX. S(k100 mesh, in 0.5 M citric acid; column, 6 ml; and flow-rate, 4.5 ml min-1. 1 min H A 1 A B ) \ - Time Fig. 7. Sorption of 300 ug of Fell (as FeSC>,.7H,O) on Dowex 1-XX resin. 51)-I00 mesh. in 0.5 M citric acid; column, 6 ml. A, Passage of the Fe" solution through the ion-exchange resin at a flow-rate of 1.0 ml min-1; and B. elution from the column with 2.0 M PIC1 at a flow-rate o f 4.5 ml min-' Fe(N03)3.9H20] in 0.5 M citric acid, eluting with 2.0 M HCl. The RSD was found to be +8%. It was also observed that the sensitivity of measurement decreased with increasing number of measurements owing to crystallisation of citric acid and subsequent clogging of the burner slot.Better results and reproduciblity of measurement (i 1%) were obtained when the eluent was collected in a 25-ml calibrated flask and iron was measured by flame AAS employing a conventional method of analysis. Under these conditions the detection limit was 5 pg of iron(II1) ion. Sorption of iron(l1) ion on the resin A standard solution of iron(I1) ion and a solution of the resin were prepared in 0.5 M citric acid, and 300 pg of iron(T1) ion were passed through the ion-exchange resin using 2.0 M HCl as eluent. The eluate was analysed immediately by AAS. The recorder traces are shown in Fig. 7 from which it can be seen that iron(I1) was not sorbed on the resin column. This allows the separation of iron(I1) and iron(II1) ions in their mixture.Determination of iron(l1) und iron(ll1) ions in synthetic mixtures Synthetic mixtures of iron(I1) and iron(II1) ions were pre- pared in 0.5 M citric acid in various concentration ratios. Iron(I1) ion passed through the column and was collected in a 25-ml calibrated flask, and iron(TI1) ion was then eluted with 2.0 M HCl and also collected in a 25-ml calibrated flask. Equivalent standard solutions of iron(I1) and iron(II1) were prepared separately in 25-ml calibrated flasks in the same concentrations as used in the synthetic mixtures and in the same solvent. This enabled the concentration of the separa- ted iron species to be determined in the solutions by flame AAS employing a conventional procedure of analysis. The results are presented in Table 2, from which it can be seen that the proposed technique is satisfactory for the separation of Table 2.Separation of iron(I1) and iron(II1) ions in synthetic mixtures on Dowex 1-X8 (5Crl00 mesh) anion-exchange resin in 0.5 M citric acid. Resin volume, 6 ml; eluent, 2.0 M HCI; flow-rate, 4.5 ml min-l Iron( 11) ion Iron(I1I) ion Added1 Yg 200 100 40 20 200 200 200 Found/ Yg 200 101 42.5 24.5 199 199 199 Recovery, Added/ Foundi Recovery, 100 200 201 100.5 101 200 200 100 106 200 198.5 99.3 122 200 199 99.5 99.5 100 102 102 99.5 40 42.5 106 99.5 20 24.5 122 Yo CCg ug O/O iron(I1) and iron(I11) ions in ratios ranging from 1 : 5 to 5 : 1 (re cover y 99- 1 06 "/" ) . It is also evident that iron can be speciated by measuring either iron(I1) or iron(TII), depending on the total amount of iron and the ratio of iron(I1) to iron(II1) ions.Separation of Uncharged Organic Iron on Amberlite XAD-2 Sorptive Resin The resin was purified by several washings with methanol and distilled water. Iron - phenanthroline standard solutions were used and various organic solvents (methanol, acetone and ethanol) were used as eluents. When direct coupling of the column to the flame AAS system was employed, poor reproducibility of measurement (RSD = i 15-20%) was obtained owing to variation of the nebulisation efficiency and the stoicheiometry of the flame during elution. For this reason sorbed species were eluted from the column into 25-ml calibrated flasks and iron was determined by flame AAS in the conventional way. Influence of p H on iron - phenunthroline sorption and elution from the resin The resin and a standard solution of iron - phenanthroline (100 pg of Fe) were prepared at pH 1-6, employing various concentrations of HN03 (0.005-0.1 M), and sorbed on the resin column, The solvent passing through the resin was collected in a 25-ml calibrated flask and the non-sorbed iron was measured by flame AAS.The efficiency of the sorption is shown in Fig. 8 as a function of pH from which it can be seen that quantitative sorption was obtained at pIi <3.5. In view of these results 100 pg of iron (as iron - phenanthroline) were sorbed on the resin column at pH 2.5 and then eluted into a 25-ml calibrated flask. The iron content was measured by flame AAS. Methanol was found to be an efficient eluent.ANALYST, FEBRUARY 1989. VOL.114 141 Reproducibility of measurement and detection limit The reproducibility of measurement was tested on a standard solution containing 50 pg of iron (as iron - phenanthroline). The RSD for six parallel determinations was found to be k 1% and the detection limit was 2 pg of iron. Sorption of ionic iron(Il) and ionic iron(III) A 100-pg amount of ionic iron(I1) (as sulphate) and ionic iron(II1) (as nitrate and iron citrate complex) was passed through the column resin under conditions that ensured the quantitative retention of uncharged organic iron on the column. The ionic forms of iron(I1) or iron(II1) were found to pass quantitatively through the resin column and this allowed quantitative separation of uncharged organic iron from ionic iron species.Determination of Iron Species in Wine Employing Ion- exchange Sorptive Chromatography and Flame AAS In wines, iron(I1) exists as the positively charged hexaaqua- iron(I1) ion and as uncharged organic iron(I1) (colouring matter), whereas iron(II1) exists as the positively charged hexaaquairon(II1) ion and is also partially complexed with organic acids (such as citric and oxalic acids) to give negatively charged ions. Positively charged iron(TI1) ions tend to form undesirable precipitates with phosphate ions and tannin, which can affect the appearance of the wine (i.e.. its clarity). The ratio of iron(I1) to iron(II1) in wine depends on the oxidation - reduction conditions. During the fermentation process a reducing atmosphere is present and so iron(T1) is the predominant species. However, during decanting and bottling, the wine is in contact with the air and the iron(1I) is then oxidised to iron(II1).To obtain clear wines (i. e. , without a precipitate), commer- cial producers reduce the iron concentration to below 5 mg 1-1 using K4Fe(CN),.23 Owing to the different stoicheiometries of the reactions of iron(I1) and iron(II1) with K,Fe(CN), it is CC ; 75 .- 0 0 1 2 3 4 5 6 PH Fig. 8. Amberlite XAD-2 (2040 mesh) resin as a function of pH Efficiency of iron - phenanthroline sorption (100 pg of Fe) on necessary to know not only the total concentration of iron, but also the ratio of iron(I1) to iron(II1). Commercially, the exact amount of K,Fe(CN), that has to be added is determined experimentally. In this work eight types of red and white wine from different geographical areas were analysed.Some of the wines were bottled by commercial producers, whereas others were local wines produced by small growers. Determination of Total Iron Concentration For the determination of the total concentration of iron in wine 5 ml of the sample were acidified with 0.5 ml of 6 M HCl and the solution was diluted to 25 ml. The iron content was measured (using aqueous standard solutions) by flame AAS. No measurable background was detected. The reproducibility of measurement was tested by performing six parallel analyses on a sample of Malvazija Vipava white wine. The RSD was found to be +0.5%. All other samples were analysed in duplicate. Determination of Iron Species in Wine The proposed method was applied to the determination of iron species in wine, employing various ion-exchange and/or sorptive resins.The results are summarised in Table 3. Determination of ionic iron(IZ) and ionic iron(III) in wine on Dowex 50-X8 cation-exchange resin Direct coupling of the column to the flame AAS system was employed using the optimum experimental conditions deter- mined previously. Samples of wine, standard solutions and column resin were prepared in 0.1 M HN03. An amount equivalent to 10CL150 pg of ionic iron was sorbed on the column (8-50 ml of wine), depending on the concentration of iron in the sample. Although the colouring matter was sorbed irreversibly on the resin, this did not affect the capacity of the ion exchanger (i.e., the sensitivity of measurement for standard solutions was not changed).The sorbed ionic species were eluted with 2.25 M HCl and the absorbances under the elution curve were integrated. The Malvazija Vipava wine sample was subjected to six parallel analyses; the RSD was &3%. All other samples were analysed in duplicate. Determination of uncharged organic iron in wine on Amberlite XAD-2 sorptive resin As an alternative to using Dowex 50-X8 resin. the uncharged organic iron and ionic iron could also be separated on Amberlite XAD-2 resin. The samples and resin column were prepared in 0.02 M HN03 (pH 2.8). Samples of white wine (10 ml) or red wine (1 ml) were then passed through the column. The colouring matter and uncharged organic iron were retained, whereas ionic iron passed through the column and Table 3.Determination of iron species in wine employing various ion-exchange and sorptive resins Resin Species adsorbed Species passing through the column Dowex 50-XS (50-100 mesh) Dowex 1-X8 (5C100 mesh) in 0.1 M HN03 . . . . . . . . . . Ionic iron(I1) and ionic iron(IT1) Uncharged organic iron in 0.5 M citric acid . . . . . . . . Ionic iron(I1I) as the negatively Ionic iron(I1) and uncharged Amberlite XAD-2 (2C-50 mesh) . . . . Uncharged organic iron Ionic iron(I1) and ionic iron(1TI) charged citrate complex organic iron (1) Total iron = ionic iron + uncharged organic iron. (2) Ionic iron = ionic iron(I1) + ionic iron(II1). (3) Uncharged organic iron = total iron - ionic iron. (4) Ionic iron(I1) = total iron - ionic iron(II1) - uncharged organic iron.142 ANALYST, FEBRUARY 1989, VOL.114 Table 4. Determination of total iron and iron species employing various ion-exchange and sorptive resins Total concentration Uncharged Sample of iron/pg ml-* Iron(II), Yo Iron(III), % organic iron, O h Red wine - Teran Kras“ . . . . . . . . RefoSkKoper* . . . . . . BlatinaMostart . . . . . . Renski rizling Ljutomeri . . ZilavkaMostart . . . . . . MalvazijaKoper” . . . . . . MalvazijaVipava* . . . . Kralj evin a * . . . . . . . . White wine - Local wine (small grower). -t Bottled wine (commercial producer). 10.59 20.0 13.68 4.88 3.97 12.06 6.41 2.41 36.2 46.6 5.8 24.2 17.4 40.6 65.5 6.7 63.8 53.4 94.2 75.8 82.6 59.4 32.0 93.3 0 0 0 0 0 0 2.5 0 was collected in a 2.5-ml calibrated flask. Both the ionic iron and the subsequently eluted uncharged organic iron were measured as described previously. The reproducibility of measurement was tested by performing six parallel analyses on the sample of Malvazija Vipava wine and was found to be il% for ionic iron and i2% for uncharged organic iron.All other samples were analysed in duplicate. The results obtained showed good agreement with those obtained for the determination of ionic iron on Dowex 5O-X8 cation-exchange resin. It was found that any iron present in wine was unlikely to be adsorbed on the colouring matter (excluding any precipitate). Determination of ionic iron(lII) in wine on Dowex 1-X8 an ion -exchange resin The wine samples and the resin column were prepared in 0.5 M citric acid. A 3-20-ml volume of wine was passed through the column, eluting with 2.0 M HC1, and the eluate was collected in a 2.5-ml calibrated flask.The iron absorbances were measured as described previously. The colouring matter was again irreversibly sorbed on the resin, but this did not affect the capacity of the ion exchanger. The sample of Malvazija Vipava wine was subjected to six parallel analyses; the RSD was i2%. All other samples were analysed in duplicate. By combining the data obtained for the total concentration of iron, ionic iron and uncharged organic iron, the ratio of iron(I1) to iron(II1) was calculated using the equations given in Table 3. The results are presented in Table 4, from which it can be seen that the ratio of iron(I1) to iron(II1) in wines changes during the ageing process. The ratio of iron(I1) to iron(II1) also depends on the type of wine and the fermenta- tion procedure employed.Conclusion The determination of iron species by means of an ion- exchange chromatography - flame AAS technique has been investigated. A Babington type nebuliser was used for direct coupling. The potential of the system was evaluated for various ion-exchange (Dowex .5O-X8, Dowex 1-X8) and/or sorptive (Amberlite XAD-2) resins. The system enabled all the processes on the column to be followed directly at various flow-rates. Hence the optimum conditions for sorption on the resin and elution from the column could be determined accurately and rapidly. The use of direct coupling was affected by the large differences in the physical properties of the effluent and eluent, which changed the efficiency of nebulisa- tion and the flame stoicheiometry during measurement. The density of the effluent also affected the accuracy of the direct coupling measurements by causing clogging of the nebuliser and/or the burner slot. On the basis of experimental data obtained for synthetic standard solutions a system was developed for the determination of the ratio of iron(I1) to iron(I1I) in wines, employing a combination of various ion-exchange and/or sorptive resins.Ionic iron(I1) and ionic iron(II1) were determined by ion-exchange chromatography on Dowex 50-X8 resin with direct coupling to a flame AAS system, uncharged organic iron on Amberlite XAD-2 resin and ionic iron(II1) on Dowex 1-X8 resin, the last two both via conventional analysis by flame AAS.The authors thank the Boris Kidrii: Foundation for providing financial support and Dr. A. R. Byrne of the Joief Stefan Institute for valuable suggestions and linguistic correction of the manuscript. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. References Suzuki, K. T., Anal. Biochem., 1980, 102, 31. Botre, C., Cacace, F., and Cozzani, R . , Anal. Lett., 1976, 9, 825. Slavin, W., and Schmidt, G. J . , J . Chromatogr. Sci., 1979, 17, 610. Atwood, J . G., Schmidt, G. J . , and Slavin, W., J . Chromat- ogr., 1979, 171, 109. Ebdon, L., Hill, S . J., and Jones, P., Analyst, 1985, 110, 515. Hill, S., Ebdon. L., and Jones, P., Anal. Proc., 1986, 23, 6. Ebdon, L., Hill, S., and Jones, P., Analyst, 1987, 112, 437. Manahan, S. E . , and Jones, D. R . , Anal. Lett., 1973, 6, 745. van Loon, J. C., Radziuk, B., Kalm, N., Lichwa, J . , Fernandez, F. J., and Kerber, J. D., At. Absorpt. Newsl., 1977, 16, 79. Kahn, N., and van Loon, J. C.. Anal. Lett., 1978, A l l , 991. Yoza, N., and Ohashi, S . , Anal. Lett., 1973, 6, 595. Mohamed, N., and Fry, R. C., Anal. Chem., 1981, 53, 450. Ajlec, R . , and Stupar, J . , Vestn. Slov. Kern. Drus., 1986,3312, 87. Cotton, F. A . , and Wilkinson, G., “Advanced Inorganic Chemistry,” Wiley, New York, 1962, pp. 707-719. Hoshino, H., and Yotsuyangi, T., Tulantu, 1984, 31, 525. Nigo, S . , Yoshimura, K., and Tarutani, T., Talantu, 1981, 28. 699. Lynch, T. P., Kernoghan, N. J., and Wilson, J. N . , Analyst, 1984, 109, 843. Strelow, F. W. E., Rethemeyer, R . , and Bothma, L. J. C., Anal. Chem., 1965. 37, 107. Strelow, F. W. E . , Liebenberg, C. J . , and von Toerien, S . F., Anal. Chim. Acta, 1969, 47, 251. Saito, N., Pure Appl. Chem., 1984, 56, 523. Florence, T. E . , and Batley, G. E . , Crit. Rev. Anal. Chem., 1980, 9, 259. Willis, R . B., and Sangster, D., Anal. Chem., 1976, 48, 59. Troost, G., “Technologie des Weines,” Ulmer. Stuttgart, 1980, pp. 397-403. Paper 8iO3003 K Received July 25th, 1988 Accepted September 27th, 1988