ANALYST, FEBRUARY 1991, VOL. 116 199 Determination of Humic Acid and Iron(iii) by Solid-state Spectrophotometry to Study Their Interactions Kunio Ohzeki, Miyoko Tatehana, lshoshi Nukatsuka and Ryoei lshida Department of Chemistry, Faculty of Science, Hirosaki University, 036 Hirosaki, Japan A simple and sensitive method for the determination of humic acid has been developed. The method is based on the adsorptive enrichment of humic acid using a finely divided anion-exchange resin, collection of the resin on a membrane filter by filtration as a circular thin-layer, and direct measurement of the absorbance of the resultant thin-layer of resin by densitometry at 470 nm. Up to 80 pg of humic acid in 100 ml of sample solution can be determined, the limit of detection being 1.3 pg.The effect of iron(ll1) is masked with ethylenediaminetetraacetic acid (EDTA). Iron(ll1) is also determined by densitometry at 600 nm after enrichment on the thin-layer of resin as a complex with ammonium pyrrolidine dithiocarbamate (ammonium pyrrolidin-I-yldithioformate) (APDC), the limit of detection being 0.06 pg of iron(1ii). In the presence of humic acid, the blank value is obtained by masking the iron(iii) as the EDTA complex, and calculating the net absorbance due to the APDC complex. The methods have been used to investigate the effect of humic acid on the formation of filterable iron(il1) species, which can pass through a 0.45 pm membrane filter. The possibility of characterizing humic acid based on the formation of a complex with iron(i1i) has been shown. Keywords: Humic acid determination; iron(iii) determination; densitometry; interaction between iron(///) and humic acid; filterable iron species The importance of speciation in a water sample for under- standing the toxicity, bioavailability, bioaccumulation and transport of a particular element has been previously evalu- ated.' Filtration with a 0.45 pm membrane filter is a universally applied first step to separate the particular element into the fractions termed dissolved and particulate.1.2 Humic acid has been recognized to play an important role in the mobilization, transportation and immobilization of metals in water.3 As iron is the most abundant heavy metal in water, the interactions between iron and humic acid have been widely investigated.3-9 While the association of colloidal particles of iron hydroxide with humic substances and the existence of iron hydroxide coated with humic substances have been widely accepted, 1-7 the formation of dissolved complexes or chelates between the functional groups of the humic substances and iron has also been accepted as significant in order to keep the iron in solution.3-5 The solubility of humic acid-iron com- plexes has been found to be dependent on the humic acid : iron ratio.6 To understand the speciation of iron in water, it is important to study the effect of humic acid on the formation of filterable iron species which pass through a 0.45 pm membrane filter .The aims of this work were to develop simple methods for the determination of humic acid and iron(ii1) at the hundred ppb level and to carry out fundamental studies of the interactions between humic acid and iron(rr1).Experimental Apparatus A Shimadzu CS-920 chromatoscanner was used for the densitometric measurement of the resin-phase absorbance. The apparatus has the capacity to transform convex calibra- tion graphs into linear graphs. A Hitachi-Horiba Type M-711 pH meter was also used. Toyo KG-25 and KG-47 filter holders with Toyo 0.45 pm membrane filters of cellulose nitrate and 0.40 pm Nucleopore filters were used for the filtration of the sample solution. A Toyo KG-25 filter holder with Toyo 0.65 pm membrane filters was used to collect the anion-exchange resin particles for the preparation of the thin-layer of resin. Reagents Unless stated otherwise, all reagents used were of analytical- reagent grade.Water, purified using an osmotic membrane, was distilled twice and used. 10 Iron(1rr) standard solution, 1000 ppm, pH 1 . Prepared from ammonium iron(ii1) sulphate dodecahydrate. A working solution, containing 10 pg ml-1 of iron, was prepared by dilution of the standard solution with dilute sulphuric acid, maintaining a pH of about 1. Ammonium pyrrolidine dithiocarbamate (ammonium pyrrolidin-1 -yldithioformate) (APDC) solution, 0.4% mlv. Sodium acetate solution, 30% mlv. Sodium perchlorate solution, 4 mol dm-3. Humic acid solutions. Humic acids obtained from Wako Pure Chemical and Nacalai-tesque were used as received. The sodium salt of humic acid obtained from Aldrich was converted to the acid form by precipitation from 0.1 mol dm-3 hydrochloric acid. About 0.1 g of each humic acid was accurately weighed and dissolved in 50 ml of a 0.1 rnol dm-3 sodium hydroxide solution using sonication for 30 s.The pH was adjusted to between 6 and 7 by the addition of hydrochloric acid and the solution was then filtered using a 0.45 pm membrane filter of 47 mm diameter. As the filters clogged rapidly, four or five were required to filter 50 ml of the solution. The resultant filtrate was diluted to 500 ml with water. The used membrane filters were air-dried and the amount of residue on each weighed. The concentration of humic acid in the filtrate solution was determined by subtracting the residue on the filters." The concentrations of humic acid solutions from Wako, Nacalai-tesque and Aldrich were 142, 120 and 158 ppm, respectively. Anion-exchange resin suspension (ARS), 7.8 pequiv ml- I.Prepared from a macro-reticular type Amberlyst A-27 resin in the chloride form according to the method reported.lO General Procedure for the Determination of Humic Acid in the Presence of Iron( 111) A 45 ml aliquot of an acidified sample solution (pH about 1.2) containing less than 80 pg of humic acid and less than 5.0 pg of iron(ri1) is placed in a 100 ml beaker. A 0.5 ml portion of 10 mmol dm-3 ethylenediaminetetraacetic acid (EDTA), 1.25 ml of 4 mol dm-3 sodium perchlorate and 2 ml of 30% sodium200 ANALYST, FEBRUARY 1991, VOL. 116 acetate solutions are added successively. The final volume is adjusted to 50 ml with water. The pH of the resultant solution is 4.7.Then 1.3 ml of ARS are added to the solution and the mixture is stirred for 5 min. The resin particles are collected on a 0.65 ym membrane filter by filtration under suction and a circular thin-layer of approximately 17 mm diameter and 0.03 mm thick is prepared. The wet membrane filter holding the thin-layer of the anion-exchange resin is placed on a white plastic plate in the densitometer. The integrated absorbance of the humic acid concentrated in the resin phase is measured, with the use of the linearizer, at 470 and 600 nm by scanning the thin-layer over an area of 24 x 30 mm2. The absorbance measured at 470 nm is used for the determination of humic acid; the absorbance measured at 600 nm is used as the blank value for the determination of iron(m), as described below.General Procedure for the Determination of Iron(m) With APDC in the Presence of Humic Acid The procedure for the determination of iron(Ir1) as the APDC complex is the same as for the determination of humic acid, except that a 0.5 ml portion of 0.4% APDC is added to the sample solution in place of the EDTA solution, and the absorbance of the resultant coloured thin layer of the anion-exchange resin is measured at 600 nm. The net absorbance due to the iron(n1)-APDC complex is determined by subtracting the blank value obtained in the presence of EDTA, as described above. General Procedure to Study the Effect of Humic Acid on the Formation of Filterable Iron Species A 45 ml aliquot of the sample solution containing 500 yl of the 10 ppm iron(Ir1) solution and 1.25 ml of a 4 mol dm-3 sodium perchlorate solution is placed in a 100 ml beaker and various amounts of humic acid are added up to about 80 yg.The solution is then stirred for 1 min using a magnetic stirrer. A 0.3 ml aliquot of a 30% sodium acetate solution is added and the final volume adjusted to 50 ml with water, the pH value of the resultant solution being 5.6. After stirring for 1 min the solution is filtered through a 0.45 pm membrane filter of cellulose nitrate under suktion from a water pump. The filter is then removed from the filter support and the support washed with 0.5 ml of 6 mol dm-3 hydrochloric acid and 5 ml of water. The washings are combined with the filtrate and the whole is allowed to stand for 30 min. Duplicate experiments are carried out to obtain a set of filtrates and then each filtrate is subjected to the determination of humic acid and iron(II1).Results and Discussion Determination of Humic Acid Humic acid has been determined by various methods, including ultraviolet/visible spectrophotometry,3,6.8,9,12-14 fluorimetry,l5J6 chemiluminescence, 17 and thermal-lens spec- trometry.18 Of these, spectrophotometry has been the most widely used. As there are similarities in the absorption spectra of humic acid and iron(rI1) hydroxide, masking the effect of the iron12 or removal of the iron altogether14 is usually required before the determination of humic acid is possible. The simultaneous determination of iron and humic acid is carried out by measuring the absorbances at different wavelengths.19 In this paper, a method of determination based on the enrichment of humic acid with the use of a finely divided anion-exchange resin followed by the densitometric measure- ment of the resin-phase absorbance is proposed.The details of the method are studied according to the general procedure described under Experimental, both with and without the addition of iron(1n). Absorption spectra Absorption spectra of humic acid collected in the resin phase showed an absorption maximum at 470 nm, as shown in Fig. 1, whereas, the absorbance of a humic acid solution obtained by the usual spectrophotometric method simply decreased with increasing wavelength in the visible range, no maximum being found in the spectrum.3.19 Effect of the amount of ARS The effect of the amount of ARS was investigated according to the general procedure, with the addition of different amounts of ARS of up to 2.0 ml.With increasing amounts of added ARS, the absorbance of the resultant thin-layer of resin increased sharply and reached a maximum at 0.5 ml of ARS. The absorbance then gradually decreased with increasing amounts of added ARS; this was probably because the resultant thin-layer of resin became thicker than the effective optical path length. Effect of p H The absorbance of humic acid collected in the resin phase from solutions with various pH levels was found to increase slightly with increasing pH in the range studied (1.5-7.5) as reported.13 The absorbance, however, was almost constant in the pH range 3.4-5.8. The pH of the solution for the determination of humic acid was, therefore, fixed at pH 4.7 and the determination of iron(n1) with APDC was carried out at pH 4.7, as described later.0.4 1 400 500 600 Wavelengthhm Fig. 1 Absorption spectra of humic acid and iron(ii1)-APDC complex in the thin layer of the anion-exchange resin: A, humic acid from Wako, 71.0 pg; B, iron(m)-APDC complex, iron(m) 5.00 pg; C, resin blank 100 s Y a, C ([I e 8 50 a (D 0 0.1 0.2 0.3 0.4 0.5 Salt concentration/mol dm-3 Fig. 2 Effect of salt concentration on the fixation of humic acid to the thin-layer of the anion-exchange resin. Humic acid (Wako), 71.0 pg; anion-exchange resin, Amberlyst A-27,lO. 1 pequiv (1.3 ml of ARS of 7.8 pequiv ml-1); sample volume, 50 ml. A , Sodium chloride: B, sodium perchlorate. The absorbance of the thin-layer of resin prepared without the addition of the salt was taken as 100%ANALYST, FEBRUARY 1991, VOL.116 201 Effect of sodium perchlorate The effect of salt concentration on the resin-phase absorbance was investigated using sodium chloride and sodium perchlor- ate. While a constant absorbance was observed in the presence of sodium chloride in the range from 0.05 to 0.5 rnol dm-3, as shown in Fig. 2, it decreased markedly with increasing amounts of added sodium perchlorate. Consequently, it was clear that the fixing of humic acid was suppressed in the presence of the perchlorate ion. Although the adsorption of humic acid on the anion-exchange resin of styrene-divinyl- benzene copolymer was reported to be irreversible ,12,21 the above results indicated that there was competition between the perchlorate ion and the anion of humic acid for the ion-exchange sites.The recovery could be improved by increasing the amount of added ARS to more than 1.3 ml, however, the use of large amounts of resin could result in the following disadvantages: firstly, a longer time period for the collection of the resin particles by filtration; and secondly, the decrease of the absorbance of the resultant thin-layer of resin with increasing thickness, as already described. As the interactions between iron(iii) and humic acid were investi- gated in the presence of 0.1 rnol dm-3 sodium perchlorate, the calibration graph for the determination of humic acid was prepared in the presence of 0.1 rnol dm-3 sodium perchlorate in order to compensate for the recovery loss of humic acid due to the perchlorate ion.Effect of sample volume The recovery of humic acid was examined for various volumes of sample solution ranging from 30 to 100 ml, all containing 71.0 pg of humic acid from Wako in the presence of 0.1 rnol dm-3 sodium perchlorate. A constant recovery of the humic acid was obtained from each of the different sample volumes. Effect of iron(ii1) The absorbance of the thin-layer of rcsin prepared from sample solutions containing a known amount of humic acid and various amounts of iron(1ii) was found to increase with increasing amounts of added iron(ii1); this is probably owing to the formation of iron(ii1) hydroxide and the iron(ri1) complex with humic acid. 19 The effect of iron(iii) was successfully masked by the addition of EDTA.The resultant iron(m)- EDTA complex was not retained by the anion-exchange resin in the presence of 0.1 mol dm-3 perchlorate ion. Calibration graph A calibration graph was prepared for the determination of humic acid using the general procedure without the addition of iron(iii). A graph showing good linearity was obtained for levels of up to about 80 pg of each humic acid. The regression lines of the calibration graph for the determination of humic acid from Nacalai-tesque, Wako and Aldrich were: y = 1293x + 7558; y = 1125x + 7646; and y = 1056x + 7394, respectively, where y is the integrated absorbance in arbitrary units including the blank value due to the anion-exchange resin, and x is the amount of humic acid in pg.The relative standard deviation (RSD) for 42.6 pg of humic acid from Wako was 3.2% (n = 5) and the limit of detection (LOD) was 1.3 pg, based on three times the standard deviation of the blank value. As stated above, the recovery of humic acid was independent of the sample volume for values of up to 100 ml, accordingly, the regression equation was valid for sample volumes of up to 100 ml. Determination of Iron(u1) With APDC Ammonium pyrrolidine dithiocarbamate has been widely used as a solvent extraction reagent and a spectrophotometric reagent for copper and other metals.22,23 Iron(1Ir) was reported to form a black precipitate of Fe(PDC)3 on reaction with APDC.22 The determination of iron with APDC by thin-layer spectrophotometry has already been reported,24 but the details of the method were not revealed.The details of the method are according to the general procedure described above, both with and without addition of humic acid. Absorption spectrum The absorption spectrum of the iron(ii1)-APDC complex in a thin-layer of the anion-exchange resin showed a small maximum at about 600 nm, while the absorbance spectrum of the blank thin-layer decreased with increasing wavelength, as shown already in Fig. 1. Accordingly, the subsequent measurements were carried out at 600 nm. Effect of amount of ARS The effect of the amount of ARS on the absorbance of the iron(ii1)-APDC complex in the resultant thin-layer of the anion-exchange resin was examined with the addition of various amounts of ARS up to 2.0 ml. Although the iron(Ir1)-APDC complex could be collected on the membrane filter directly, without the addition of the anion-exchange resin, reproducible absorbance was only obtained by fixing the complex in the thin-layer of the anion-exchange resin.The resin-phase absorbance increased with increasing amounts of added ARS, up to 0.5 ml and then gradually decreased. The increase in the absorbance with increasing amounts of added ARS, up to 0.5 ml, was probably because the optical path-length would be increased with the increasing thickness of the resultant thin-layer. The decrease in the absorbance with increasing amounts of added ARS, above 0.5 ml, was probably because the resultant thin-layer became thicker than the optical path-length. These tendencies are characteristic of absorption measurements by densitometry.When the absorb- ance is measured with the use of a conventional spectro- photometer it increases with increasing thickness of the thin-layer. Effect of p H The effect of pH on the absorbance of the iron(ir1)-APDC complex was studied according to the general procedure described above, in the pH range 1.3-6.9. The absorbance of the complex in the thin-layer of the anion-exchange resin was found to be constant and maximum in the pH range 3.0-5.5, as was observed in the extraction of the iron(ii1)-APDC complex with 2,6-dimethylheptan-4-one.25 Effect of sodium perchlorate The absorbance of the iron(ii1)-APDC complex in the thin- layer of the anion-exchange resin showed a constant and maximum value in the presence of sodium chloride or sodium perchlorate in the range 0.05-1 .O rnol dm-3.Effect of sample volume The effect of sample volume on the fixing of iron(iii), as the APDC complex, to the anion-exchange resin was examined with various sample volumes ranging from 20 to 100 ml. The recovery of iron was independent of the sample volume studied. Hence, the calibration graph was valid for sample volumes of less than 100 ml. The sensitivity of the densito- metric determination of iron with bathophenanthrolinedi- sulphonate (BPS)26 is higher than with the proposed method, but the reduction of iron(iii) to iron(1i) is essential before the colour producing reaction with BPS can occur. Effect of foreign ions The effect of iron(ii), copper(ii), nickel(ii), cobalt(ii) and zinc(i1) on the determination of 3.0 pg of iron(ii1) was examined.The results are summarized in Table 1. Iron(rr) was202 ANALYST, FEBRUARY 1991, VOL. 116 Table 1 Effect of foreign ions on the determination of iron(i1i). Amount of iron(iir), 3.00 pg; sample volume, 50 ml Amount Ion added/ yg Fe"t 3.00 Cu" 3.00 1.00 Ni" 3.00 1.00 CO" 3.00 1.00 Zn" 6.00 3.00 Iron( 111) found*& 2.99 2.86 2.98 2.81 3.08 3.03 2.98 3.08 2.95 Error (% ) -0.3 -4.6 -0.7 -6.3 2.7 1 .o -0.7 2.7 -1.7 * Results of duplicate determinations. t In the presence of hydroxylamine hydrochloride. Table 2 Effect of humic acid on the determination of iron(1ii) Amount of Humic acid Number of iron(rrr)/yg added*/yg Iron found/yg experiments 5.00 2.00 14.2 2.06 42.6 2.00 71.0 2.08 Mean 2.05 SDt 0.21 14.2 5.10 42.6 5.11 71 .O 5.02 Mean 5.08 SD 0.09 * Humic acid from Wako.t SD = standard deviation. found to react with APDC in a similar manner to iron(iii). Consequently, in the presence of iron(i1) the total amount of iron can be determined. A small negative error was observed in the presence of the same amount of copper(i1) and nickel(ii), however, the concentrations of these ions in common water samples are much lower than that of iron(111).27 Effect of humic acid In the presence of humic acid, the sum of the absorbances from the iron(ii1)-APDC complex and humic acid was measured at 600 nm. The absorbance of humic acid was measured with the use of another aliquot of the sample solution after masking the iron(ii1) with the EDTA complex and the difference calculated. The effect of humic acid was thus effectively eliminated, as shown in Table 2.Calibration graph The calibration graph for the determination of iron(rri) was prepared according to the general procedure without the addition of humic acid. The calibration graph showed good linearity, the correlation coefficient ( r ) being 0.9998 in the concentration range 1.0-5.0 pg of iron(il1). The regression equation of the calibration graph was y = 21878~ - 4774, where y is the net value of the integrated absorbance in arbitrary units and x the amount of iron(Ir1) in pg. For the determination of less than 1.0 pg of iron(1ri) the regression equation of y = 17969~ was used. The RSD for 4.0 pg of iron(iii) was 2.4% ( n = 5 ) and the LOD was 0.06 pg, based on three times the standard deviation of the blank value. Absorption of Iron(m) on the Membrane Filter Effect of p H on the formation of iron(iii) hydroxide The effect of pH on the formation of the filterable iron(ii1) species was investigated by the use of a 0.45 pm membrane fitter of cellulose nitrate.The solution pH was adjusted with 100 C .- C 2 - 0 2 4 6 8 10 12 14 PH Fig. 3 Effect of pH on the adsorption of iron on to a 0.45 pm membrane filter of cellulose nitrate from 50 ml of solution containing 5.00 yg of iron(m) and 0.1 mol dm-3 sodium perchlorate. Solid line indicates the theoretical curve for the formation of iron(m) hydroxide. The following equilibria and the stability constants were used for the calculation, the ionic strength being 0.1 mol dm-320 Fe(OH)3(s) = Fe(OH)2+ + 20H- log K S I = - 26.16 Fe(OH)3(s) = Fc(OH)~+ + OH- log Ks2 = - 16.74 Fe(OH)3(s) + OH- = Fe(OH)?- 2Fe(OH)3(s) = Fe2(OH)24+ + 40H- log K = - 50.8 log Ks4 = -5 100 0 2 4 6 8 10 12 PH Fig.4 Effect of pH on the fixing of the humic acid and iron on to the 0.45 ym membrane filter of cellulose nitrate. Amount of iron(iii), 5.00 pg; amount of humic acid (Wako), 14.2 yg; sample volume, 50 ml. A, Adsorption of iron(m) in the presence of humic acid; B, adsorption of humic acid in the absence of iron(iI1); C, adsorption of humic acid in the presence of iron(m) the use of sodium hydroxide and dilute hydrochloric acid solutions. The amount of iron(1n) in the filtrate was deter- mined by the proposed method and plotted against the pH of the filtrate, as shown in Fig. 3. Iron(iri) was almost quantita- tively collected on the membrane filter as a pale yellow precipitate from solutions in the pH range of about 6-10.The results of the calculation to estimate the effect of pH on the formation of iron(r1r) hydroxide are also illustrated in Fig. 3. There were good similarities between the pH dependence of the formation of iron(m) hydroxide and the adsorption of iron(iii) on to the membrane filter. Effect of pH on the adsorption of humic acid on the membrane filter The effect of pH on the adsorption of humic acid on to the membrane filter of cellulose nitrate was examined both in the absence and presence of iron(ii1). The solution pH was adjusted with the use of hydrochloric acid, sodium acetate, ammonium acetate and sodium hydroxide solutions. Humic acid was not retained by the membrane filter from.solutions with a pH of above about 6, as shown in Fig.4. In the presence of iron(iiI), however, some of the humic acid became fixed on to the membrane filter together with iron(1ii) above about pHANALYST, FEBRUARY 1991, VOL. 116 203 4.5. As the solution pH increased above 8, most of the humic acid added appeared in the filtrate together with iron(iii). When each sample solution of different pH was allowed to stand for 20 min before the filtration, no significant difference was found in the pH dependence of the formation of the filterable iron species in the presence of humic acid. It was safely concluded that there are strong interactions between iron(ii1) and humic acid above about pH 4.5. Effect of humic acid concentration The effect of the concentration of humic acid on the formation of the filterable iron(ri~) species was examined at pH 5.6 and 8.4, respectively.The results obtained with the use of humic acid from Nacalai-tesque are shown in Fig. 5. Until the added amount of humic acid reached a certain level, both iron and humic acid were not found in the filtrate, however, after that, the amount of filterable iron(iii) species was found to increase with increasing added amount of humic acid and then reached a constant and maximum value. Almost the same results were obtained at pH 8.4 except that the plot was shifted to the left of that obtained at pH 5.6. As it was postulated that the iron found in the filtrate was accompanied by a fixed amount of humic acid (as shown in Table 3) it was expected that the complexation between iron(111) and humic acid occurred in a fixed molar ratio.Assuming the formation of a 1 : 1 complex, the relative molecular mass of humic acid could be calculated, as shown in Table 3, in which the results obtained with humic acids from Wako and Aldrich are also included. Among the 5.0 4.0 0-l 5 . w 2" 3.0 - .- Lc C -u .- : 2.0 Lc S 2 - 1 .o 0 60 0-l 3. . 4- 40 Y- C TI S 3 0 .- Lc z 20 .- s I 0 20 40 60 Humic acid added/yg Fig. 5 Effect of humic acid concentration on the formation of the filterable iron(iI1) species: A, the maximum amount of iron(m) in the filtrate; B, humic acid in the filtrate which is expected to be complexed with iron(1Ii) of A. 1 and 3, pH 8.4; and 2 and 4, pH5.6. Initial amount of iron(In), 5.00 pg; sample volume, 50 ml; ionic strength, 0.1 mol dm-3; and membrane filter.0.45 pm cellulose nitrate humic acids examined, that of Aldrich was found to be most effective for the formation of the filterable iron(1ri) species; in the presence of about twice the amount of humic acid over 5 pg of iron(iii), more than 90% of the iron was converted to the filterable species. As the average relative molecular mass of humic acid was estimated to be in the range 600-1000,28 the formation of iron-humic acid complexes with molar ratios of 2 : 1 and 3 : 1 could also be expected.3 Effect of other complexing agents The effect of complexing agents, other than humic acid, on the formation of the filterable iron(rI1) species was also examined at pH 5.6. As expected, nitrilotriacetic acid (NTA) was found to react with iron(ni) to form a filterable 1 : 1 complex, as shown in Fig.6. On the other hand, citrate, tartrate and pyrophosphate ions were much more effective than NTA, even when they were present at less than the stoichiometric amounts required for the formation of a 1 : 1 complex with iron(m), as considerable amounts of iron(m) were found in the filtrate. The effect of standing time on the formation of a filterable iron(i1i) species was examined with the use of 50 ml portions of a sample solution containing 5.0 pg (89.5 nmol) of iron(ii1) and 20 nmol of pyrophosphate. The amount of iron in the resultant filtrate was independent of standing time for up to 30 min. Accordingly, it was assumed that citrate, tartrate and pyrophosphate ions react with iron(ii1) to form filterable complexes in which the molar ratio of iron to each ligand 0 20 40 60 80 100 Ligand addedhmol Fig.6 Effect of complexing agents on the formation of the filterable iron(1ir) species: A, sodium pyrophosphate; B, sodium citrate; C, sodium tartrate; D, NTA; and E, sodium oxalate. pH 5.6, other conditions as in Fig. 5 Table 3 Results for the determination of iron and humic acid in 50 ml portions of sample solution containing 5.00 vg of iron(m) and various amounts of humic acid after filtration (see Fig. 5) Relative Amount of Amount of molecular Source of iron in humic acid Ratio of mass of humic acid PH A*/% in B*/pg B to A humic acid? Aldrich 5.6 4.53 12.0 2.6 142 8.4 4.54 11.4 2.5 140 Nacalai-tesque 5.6 4.00 19.0 4.8 268 8.4 4.29 20.3 4.7 262 Wako 5.6 3.25 18.4 5.7 31 8 8.4 3.96 22.1 5.6 313 * A and B areas found in Fig.5 . j- The formation of a 1 : 1 iron-humic acid complex was assumed.204 ANALYST, FEBRUARY 1991, VOL. 116 exceeded 1. However, oxalate, acetate and phosphate ions were not effective for the formation of filterable iron(ir1) species. More than 20 pmol of phosphate ion and 4.5 mmol of acetate ion were required for the conversion of 5.0 pg of iron(m) to the filterable species. Effect of Sample Volume on the Determination of Filterable Iron(m) The effect of sample volume on the determination of the filterable iron(1rr) species was examined in the presence of humic acid. The 50 ml portions of sample solution of pH 5.6, containing 5.0 pg of iron(1ii) and 35 pg of humic acid from Nacalai-tesque, were prepared and filtered successively through a 0.45 pm membrane filter of cellulose nitrate, with diameters of 25 and 47 mm, and a 0.4 pm Nucleopore filter with a 25 mm diameter under suction. The amount of iron(m) in each filtrate decreased throughout the number of filtration runs as shown in Fig.7. As already shown in Fig. 4, about 20% of the iron(Ii1) was fixed on to the membrane filter together with the humic acid from the 50 ml of sample solution, and consequently, the filter may be clogged with the particulate iron(m) species during successive filtrations. A marked decrease in iron(rrr) concentration in the filtrate was observed when a sample of water was filtered with the use of a Nucleopore filter.2 A membrane filter of 47 mm diameter was much better than that of 25 mm diameter for preparation for the determination of filterable iron(Ir1) species.The experi- ments were also carried out in the presence of citrate, tartrate, nitrilotriacetate and pyrophosphate ions. Each complexing agent was added to the sample solution such that a part of iron(rr1) was converted to the filterable species and the remainder was present as the particulate hydroxide. With the 0 I I 1 I I I I 1 0 1 2 3 4 5 6 7 8 Number of fraction Fig. 7 Effect of filtration on the determination of the filterable iron(ii1) species and humic acid. Membrane filters used were; A , 0.45 ym cellulose nitrate of 47 mm diameter; B, 0.45 ym cellulose nitrate of 25 mm diameter; and C, 0.40 ym Nucleopore of 25 mm diameter.Initial amount of iron(m), 5.00 yg; and humic acid (Nacalai-tesque), 36.0 yg. Amount of each fraction, 50 ml addition of citrate, tartrate and nitrilotriacetate ions the concentration of iron(n1) in the resultant filtrate was constant regardless of filtered sample volumes for volumes of up to 400 ml (8 filtration runs). In the presence of pyrophosphate, however, the concentration of iron in the filtrate decreased with an increasing number of filtration runs, as was observed in the presence of humic acid. Determination of Iron and Humic Substances in River-water An aliquot of a river-water sample was divided into 50 ml portions, and each portion was filtered successively through a 0.45 pm membrane filter of cellulose nitrate, of 47 mm diameter, within 3 h of sampling.Another series of filtrates was prepared with the use of a fresh membrane filter. A series of sets of filtrates was thus prepared. Each set of filtrates was then subjected to the procedure for the determination of iron and humic acid. Coloured species in each filtrate collected in the resin phase were determined as for humic acid with the use of a calibration graph prepared from the Aldrich humic acid. The results are shown in Table 4 (the results obtained for a tap-water sample are included for comparison). The amount of iron in each 50 ml filtrate of the river-water sample was almost constant for 6 filtration runs. As the total concentration of iron was found to be 81 ppb (Table 5), it was concluded that 85% of the iron in the river-water was from particulate species and 15% from filterable species.However, the classification of iron in the tap-water sample was meaningless because the concentrations of the filterable iron and therefore the particu- Table 4 Effect of filtration on the determination of filterable iron and humic substances in river- and tap-water samples. The 50 ml portions of the water samples were successively filtered through a 0.45 ym membrane filter of 47 mm diameter River water* Tap-water$ Humic Number of Iron substances Iron filtration found found? found runs (PPb) (PPb) (PPb) 1 11.2 75 24.1 2 12.2 87 8.5 3 11.3 82 0.0 4 12.8 81 0.0 5 13.3 82 6 12.3 88 7 9.3 70 8 10.0 77 9 7.6 70 10 5.4 54 - - - - - - * Collected from a tributary of the Iwaki river. t Calibration graph prepared with the humic acid obtained from 3: Collected after running for more than 30 min.The total Aldrich was used. concentration of iron was 102.4 k 4.0 ppb (n = 6). Table 5 Results of the determination of iron and humic substances in river-water Iron Humic substances* Total Filterable Total Filterable River-water (ppb) (PPb) (PPb) (PPb) A t 81.0 f 1.6 12.4 f 0.7 126 f 7 84 f 3 n = 6 n = 5 n = 6 n = 5 B 185.4 f 3.1 17.3 f 0.7 126 f 3 13 f 2 n = 6 n = 5 n = 3 n = 4 * Humic acid from Aldrich was taken as the standard. t River-water A is a tributary of the Iwaki river B.ANALYST, FEBRUARY 1991, VOL. 116 205 late iron were dependent on how much of the tap-water was filtered. The results also indicated that the hydroxide might be the predominant iron species in tap-water.The classification of humic substances in the river-water sample into filterable and particulate species was significant, as a constant amount of humic substances was found in the filtrate, see Table 4. For the determination of the total concentration of iron, the water sample was acidified to pH 1.2 and filtered through a 0.45 pm membrane filter of cellulose nitrate of 47 mm diameter. On the other hand, the determination of total humic substances was carried out as follows. A 1 1 portion of the water sample was acidified to pH 1.2 and 2.0 ml of 0.1 mol dm-3 EDTA solution were added. The solution was allowed to stand for 2 h, after which the pH was adjusted to approximately 11 by the addition of sodium hydroxide and the resultant solution was filtered through a 0.45 pm membrane filter.The pH of the filtrate was adjusted to 4.7 and the coloured species in the filtrate determined as described previously. The results are summarized in Table 5. The total concentration of humic acid in the river-waters was found to be almost the same, whereas a significant difference was found between the concentrations of filterable iron(rr1) species in each water. Conclusion Simple and sensitive methods for the determination of iron and humic acid have been developed and the effects of humic acid and other complexing agents, such as citrate, tartrate, pyrophosphate and nitrilotriacetate ions, on the formation of filterable iron(m) species were investigated. These ligands were found to be effective for the formation of filterable iron species.Except for the nitrilotriacetate ion which formed a 1 : 1 complex with iron(m), the formation of iron-rich com- plexes in which the molar ratio of iron to each ligand exceeded 1 was demonstrated. The total amount of iron and humic substances found in the river-waters examined could be classified into their respective filterable and particulate species. References 1 Florence, T. M., Talanta, 1982, 29, 345. 2 Lexen, D. P. H . , and Chandler, I. M., Anal. Chern., 1982, 54, 1350. 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Schnitzer. M., Hurnic Substances in the Environment, Marcel Dekker, New York, 1972, pp. 54 and 230. Rashid, M. A., and Leonard, J. D., Chern. Geol., 1973, 11,89. Perdue, E. M., Beck, K. C., and Reuter, J. H.. Nature (London), 1976,260,418. Picard, G. L., and Felbeck, G. T.. Jr., Geochirn. Cosrnochirn. Acta, 1976, 40, 1347. Moore, R. M., Burton, J . D., Williams, P. J., and Yound, M. L., Geochirn. Cosrnochirn. Acta, 1979, 43, 919. Tipping, E., Geochirn. Cosrnochirn. Acta, 1981, 45, 191. Hiraide, M., Ishii, M., and Mizuike, A., Anal. Sci., 1988, 4, 605. Ohzeki, K., Minorikawa, M., Yokota. F., Nukatsuka, I . , and Ishida, R., Analyst, 1990, 115, 23. Hiraide, M., Tillekeratne, S. P., Otsuka, K., and Mizuike. A . , Anal. Chirn. Acta, 1985, 172, 215. Nagayama, M., Goto, K., and Yotsuyanagi, T., Kogyo-Yosui, 1963, (61), 24. Hiraide, M., Arima, Y., and Mizuike. M.. Anal. Chirn. Acta, 1987, 200, 171. Nomizu, T., Sanji, M., and Mizuike, A., Anal. Chirn. Acta, 1988,211,293. Almgren, T., Josefsson, B., and Nyquist, G., Anal. Chirn. Acta, 1975, 78, 411. McCrum, W. A., Anal. Proc., 1986, 23, 307. Marino, D. F., and Ingle, J . D., Jr., Anal. Chirn. Acta, 1981, 124. 23. Power, J. F., and Langford, C. H., Anal. Chern., 1988,60,842. Carpenter, P. D., and Smith, J . D . , Anal. Chim. Acta, 1984, 159, 299. Snoeyink, V. L., and Jankins, D., Water Chemistry, Wiley. New York, 1980, pp. 264-267. Miles, C. J., Tsuchall, J. R., Jr., and Brezonik. P. L., Anal. Chern., 1983, 55, 410. Shindo, E., and Morito, M., Kagaku no Ryoiki, 1966, 21,206. Cheng, K. L., Ueno, K., and Imamura, T., Handbook of Organic Analytical Reagents, CRC Press, Boca Raton, FL, 1982, p. 397. Ohzeki, K.. Toki, C., Ishida. R., and Saitoh, T., Analyst, 1987, 112, 1689. Bone, K. M., and Hibbert, W. D., Anal. Chirn. Acta, 1979,107, 219. Shriadah, M. M. A., and Ohzeki, K., Analyst, 1986, 111, 555. Martin, J.-M., and Mwybeck, M., Mar. Chern.. 1979, 7, 173. Wilson. M. A., Vasallo, A. M., Perdue, E. M., and Reuter, J. H., Anal. Chem., 1987, 59, 551. Paper 01023461 Received May 25th, 1990 Accepted October 19th, 1990