ANALYST, FEBRUARY 1992, VOL. 117 125 Critical Evaluation of Three Analytical Techniques for the Determination of Chromium(vi) in Soil Extracts Radmila MilaW, Janez Stupar, Nevenka Kofuh and Janez KoroSin Joief Stefan Institute, University of Ljubljana, 61000 Ljubljana, Jamova 39, Slovenia, Yugoslavia Three different analytical techniques [Ir5-diphenylcarbazide spectrophotometry, chelating ion-exchange chromatography (Chelex-IOO), and ion-pairing reversed-phase high-performance liquid chromatography (RP-HPLC) combined with electrothermal atomic absorption spectrometry (ETAAS)] were critically evaluated for the determination of CrV1 in soil extracts. Spectrophotometry was not applicable t o the analysis of most soil extract samples owing t o its high limit of detection (LOD = 30 ng cm-3), and the possibility of the instantaneous reduction of Crvl under the acidic conditions employed.A Chelex 100 column, although adequately sensitive (LOD = 1.5 ng cm-3), is inclined t o give higher results as inert and moderately labile Cr"l complexes partially passed through the resin together with CrVt. In addition, very small particles (€0.45 pm) carrying chromium can produce severe positive systematic errors. In order t o avoid this, filtration employing a 0.1 pm filter is recommended. Ion-pairing RP-HPLC was found t o be the most sensitive technique (LOD = 0.3 ng cm-3). It might also give high chromate results if negatively charged Cr1I1 complexes form ion pairs with tetrabutylammonium phosphate and their elution partially coincided with that of CrV1. Fulvate ligands showed this type o f interference.Reversed-phase HPLC is not suitable for analysis of extracts obtained from soils with freshly added tannery waste owing t o the effects of the undestroyed tannery waste matrix. This study showed that each method investigated was vulnerable t o some type of interference. Keywords: Chromate; spectrophotometry; chromatography; electrothermal atomic absorption spectrometry; soil extracts Chromium appears frequently as a pollutant in terrestrial systems. The hexavalent form is of prime concern because of its high toxicity. Studies by several workers of the formation and fate of Crvl in soil have produced controversial results. This may be owing to the lack of reliable analytical data on Crvl in soil extracts.A number of different analytical techniques are available for the determination of CrV1;1-26 however, most of them cannot be applied efficiently to complex samples such as soil extracts. Samples containing Crvl and Cr1I1 at nanogram levels have been concentrated on anion- and/or cation-exchange resins. 1-4 Investigations based on various extraction techniques5.6 combined with atomic absorption spectrometry (AAS) or inductively coupled plasma atomic emission spectrometry (ICP-AES) have also been carried out. Coprecipitation of CrV1 with lead sulfate7 is satisfactory for water and waste water samples which are low in chloride and sulfate ions (found to produce severe negative interferences). A flow injection method for the determina- tion8 of Cr042- and Cr"' was also reported.The 1,5-diphenyl- carbazide spectrophotometric method was used for the determination of Crvl in soil extract samples,9-14 but high positive interferences were also observed. 14 Several papers have been published on the determination of trace amounts of Crvl and Crlll in water samples using reversed-phase high- performance liquid chromatography (RP-HPLC) combined with AAS.ls-I9 A method has been developed20 for the determination of Crvl in aqueous samples or sample extracts using ion chroma- tography coupled with ICP mass spectrometry or colorimetry. Trace amounts of Crv' and Crrr1 in water samples have also been determined using chelating ion-exchange resins.21-2s The main advantage of these resins is their high purity. A comparative study26 of the lead sulfate coprecipitation, the 1,5-diphenyIcarbazide spectrophotometric and the chelation- extraction methods was carried out for the determination of Crvl in the presence of a large excess of Cr1I1.The aim of this work was to develop a reliable analytical technique for the determination of Crvl in soil extract samples. The 1,5-diphenylcarbazide spectrophotometric method and chelating ion-exchange chromatography (Chelex 100) and ion-pairing RP-HPLC separation, both combined off-line with electrothermal AAS (ETAAS) as a Cr specific detector, were critically evaluated. In addition, various interferences asso- ciated with the sample matrix were carefully investigated for each method examined. Experimental Apparatus and Procedures Detection systems A Varian Cary Model 16 spectrophotometer adjusted to a wavelength of 540 nm was used for the determination of Crvl by the 1,5-diphenylcarbazide spectrophotometric method .9 Separated chromium species (chelating ion exchange, RP-HPLC) were determined on a Varian AA 575 atomic absorption spectrometer with a Perkin-Elmer HGA 76B graphite furnace.The spectral bandpass of the monochroma- tor was 0.2 nm. A Varian Techtron hollow cathode lamp was operated at a current of 5 mA. The integrated absorbance of chromium was measured at the 357.9 nm line by injecting 15 mm3 of the sample. The atomization temperature was 2773 K. Background absorption was largely eliminated by careful control of the ashing conditions (ashing temperature between 1673 and 1873K, time 10-30s) and/or by a deuterium background corrector.Platform atomization was found to be superior to wall atomization. High chromate concentrations, employed in some RP-HPLC separations (interference studies), were measured by flame AAS (Varian AA-5, N20-acetylene flame). Separation system All soil extract samples were centrifuged with a Heraeus Model 17s Sepatech Biofuge at 10 000 rev min-1 for 20 min and filtered through 0.1 pm membrane filters, unless stated otherwise. Chelating ion exchange. A system was developed for the separation of chromium species employing Chelex 100 (50- 100 mesh) resin. The batch procedure of Isozaki et al.23 was modified into a column technique to obtain a better separation and lower limits of detection (LODs). Chelex 100 (Na form) resin was first transformed to the NH4 form23 and approxi- mately 0.3 g of the resin was slurried with water and transferred into the column (using a 1 cm3 plastic pipette tip) without prior purification.Quartz wool was placed at each end of the column, and the column was connected to a peristaltic126 0 ANALYST, FEBRUARY 1992, VOL. 117 II A B C 1 pump (Ismatec MS4 Reglo) through which the flow rates could be varied between 0.2 and 4.5 cm3 min-1. The column resin was first treated with 10 cm3 of buffer solution,27 followed by passage of a buffered sample (5 cm3 of sample and 5 cm3 of buffer) at a flow rate of 1 cm3 min-1 and then washed with 15 cm3 of buffer at a flow rate of 4.5 cm3 min-1. The sample eluent was collected in a beaker, acidified with 0.2 cm3 of nitric acid (1 + l ) , evaporated to approximately 2 cm3, transferred into a 5 cm3 calibrated flask and diluted with water to the calibration mark.Chromium not retained by the column was then measured by ETAAS. The same procedure was applied to the blank solution. Although the column capacity was not exceeded by performing 20 consecutive separations (solutions containing 80 ng cm-3 of Cr3+ and 40 ng cm-3 of C++), the multiple use of the same column for the separation of chromium in soil extracts resulted in an increased blank value. Therefore, the column was refilled with fresh resin for each sample analysis. Ion-pairing RP-HPLC. Separation of chromium species was accomplished on a column (150 x 4.6 mm i.d.) packed in-house with LiChrosorb RP CIS particles (5 pm). The eluent was pumped through the column at a rate of 1.0 cm3 min-1 by a gradient Merck-Hitachi 6200 intelligent pump.The sample was introduced onto the column by a Rheodyne injector equipped with a laboratory-built 5 cm3 loop. The minimum volume for reproducible injection of samples on the loop was found to be 12 cm3. Chromium(v1) was separated from other chromium species by the formation of an ion pair with the reagent, tetrabutylammonium phosphate (PIC A). The soil extract (2-10cm3) was mixed with the reagent and diluted with water to 15 cm3 so that the final reagent concentration was 1 x 10-2 mol dm-3. Standard and blank solutions were prepared in the same way and injected into the loop. The pump (operated at a flow rate of 1 cm3 min-1) was programmed in the following steps, which were determined experimentally to ensure quantitative and reproducible sepa- ration.( i ) The prepared sample (15 cm3) was injected into the loop and washed with water onto the column for 15 min (inject position). (ii) The injector was then turned to the load position to ensure the direct elution of the retained chromate by 50% v/v methanol-water. This phase was completed in 12 min. (iii) The column was rinsed with water for 33 min and was then ready for the next separation. The Crvt peak appeared 19 min after injection in a volume of 0.2-0.4 cm3. The. retention time depended on the pH and sample matrix and could vary; therefore, ten fractions of 0.2 cm3 were collected from the eighteenth minute after injection. Chro- mium(v1) was then determined in collected fractions ‘off-line’ by ETAAS.A typical chromatogram for Crvl is presented in Fig. 1. Reagents Merck Suprapur acids and doubly distilled water were used for the preparation of samples and standard solutions. All other chemicals were of analytical-reagent grade. 0.300 A standard Crlll stock solution (1000 pg cm-1) was prepared by dissolving 1.0 g of Cr metal powder (99.99%) in 20 cm3 of HCI (6 mol dm-3) followed by dilution to 1OOOcm3 with water. Chromium(1ir) citrate, maleate, oxalate and fulvate com- plexes (1000 pg cm-3) were prepared by mixing CrCI3 stock solution with an appropriate amount (3 : 1 ligand to Cr ratio) of citric, maleic and oxalic acid, respectively. Solid fulvic acid isolated from peat soil (relative molecular mass about 1000) was used to form chromium(rI1) fulvate.The concentration of fulvic acid in the solution (145 pg (3111-3) was more than adequate for complexation of added Crttl. A standard Crvl stock solution (1000 pg cm-3) was prepared by dissolving 2.828 g of potassium dichromate in 1000 cm3 of water. Chelating resin Chelex 100, Na form, 50-100 mesh was obtained from Sigma. Tetrabutylammonium phosphate (PIC A) (0.5 mol dm-3, Supelco) stock solution was used as an ion-pairing reagent. Methanol (Merck) for chromatography was employed in HPLC measurements. Formic acid-potassium hydroxide buffer solutions were applied27 for the pH range 2.5-5.0 and cellulose nitrate membrane filters, 0.45-0.05 pm, and 25 mm diameter (Sar- torius), were used in the filtration procedure. Sample Preparation Soil samples were prepared by shaking 2.00 g of moist soil for 2 h with 20cm3 of KH2P04 (1.5 x 10-2moldm-3), then centrifuging and decanting.The KH2P04 solution (1.5 X 10-2 mol dm-3) was used to extract efficiently water-soluble chromate and chromate sorbed on various oxides and clay particles.9 Samples were then filtered through 0.1 pm mem- brane filters and the concentration of total soluble chromium was determined by ETAAS. Aliquots of these solutions were used for the determination of different chromium species. It was found experimentally that filtering through 0.45 pm membrane filters did not remove particles from the solution efficiently. These fine particles containing chromium might produce large positive systematic errors if separation is performed on ion-exchange columns.This problem is virtually eliminated by filtering through 0.1 pm filters. Results and Discussion Parameters Influencing Chromium Speciation and the Study of Interferences Spectrophotometry Two similar procedures were tested: addition of acidified 1,5-diphenylcarbazide reagent to the sample9 (Procedure I), and addition of 1,5-diphenylcarbazide reagent to the sample before acidification14 (Procedure 11). Procedure I: a lcm3 aliquot of azide reagent was added to 10 cm3 of soil extract and the magenta colour was compared with standard Crv’ solutions at 540 nm after 20 min. For preparation of the azide reagent, 120cm3 of 85% v/v H3P04 were diluted with 280 cm3 of distilled water and added to 0.4 g of 1,5-diphenylcarbazide dissolved in 100 cm3 of 95% v/v ethanol.Procedure 11: a 0.4 cm3 aliquot of 1,5-diphenylcarbazide solution (0.1 g of 1,5- diphenylcarbazide dissolved in 10 cm3 of acetone) was added to 10 cm3 of soil extract. Then, 0.2 cm3 of H2S04 (20 g of 95% H2S04 dissolved in 80 cm3 of distilled water) was added to the sample. The magenta colour was compared with standard Crvl solutions at 540nm after 20 min. Analysis of soil extract matrices to which Crvl had been added indicated that both procedures produced similar results. The reproducibility of measurement for six parallel deter- minations of Crv’ (50 ng cm-3) was found to be k2%. The LOD (30) for Crv’ in aqueous standard solutions was 10ngcm-3. In soil extracts, the matrix influenced the sensitivity of the measurements. The LOD in soil extracts was found to be 30 ng cm-3, and therefore analysis of most soil extracts was not possible employing this technique. TheANALYST, FEBRUARY 1992, VOL.117 method could only be applied in particular situations where soils heavily polluted with Cr were investigated. Parameters influencing the determination of Cr"' in soil extracts. The influence of organically complexed Cr1I1 present in soil extract samples on the determination of Crvl was studied. To 1 pg cm-3 of Crvl, 2.5 pg cm-3 of Cr"' complexes were added and Crv' was determined by Procedure I and Procedure 11. The results are presented in Table 1. It can be seen from Table 1 that interferences from organically complexed Cr"' species were almost negligible in Procedure 11. The slightly lower results obtained by Procedure I can be explained by partial reduction of chromate in the presence of reductants under low pH conditions. Influence of particles and soil matrix on the determination of Cr"l in soil extracts.Clay and peat soil extracts were prepared as described previously. After centrifugation, the supernatant was separated from the solid residue either by decanting, decanting and filtering through a 0.45 pm filter or filtering through a 0.1 pm filter. The total concentration of soluble chromium in these soil extracts was below 1 ng cm-3. A 500 ngcm-3 concentration of Crvl was added to each soil extract and Crvl was determined by spectrophotometry (Procedure I). The influence of particles in the soil extracts and the influence of the soil matrix were studied. The results are presented in Table 2.Two effects influencing the results in Table 2 could be observed. The first is light scattering caused by particles in solution, which contributes to severe positive interferences, particularly in soils with a high clay fraction. It is therefore reasonable to expect that the results previously reported by Bartlett and James9 overestimate the oxidation of Cr"' to chromate in soils with a high content of MnOz. The formation of chromate from soluble Crrl' species in soils due to the presence of Mn02 was actually confirmed in our laboratory but the levels of chromate found in the soils were significantly lower. In addition, Heringer Donmez and Kalenberger14 by not performing filtration actually reported anomalously high Crvl levels in soil (28.7% clay) leachates, which is consistent with our observation.The second effect observed from Table 2 is a slightly lower recovery in filtered extracts. This is probably due to the partial reduction of chromate in the acidic medium of the 1,5-diphenyIcarbazide reagent by reducing substances in the soil extract. Chelating ion exchange-ETAAS Influence of pH on the sorption of chromium. The resin and standard solutions of Crl" (CrCI3) (200 ng cm-3) and Crvl (40 ng cm-3) were prepared at pH 3-5 in formic acid-potas- Table 1 Influence of organically complexed Cr"' on the determination of CrV' by the 1,5-diphenylcarbazide spectrophotometric method Procedure I Procedure 11 Addedpg ~ m - ~ CrV' CrV1 Cr"' CrV' pgcm-3 (%) pgcm-3 (%) 2.5 (Citrate) 1 .o 0.94 94 0.94 94 2.5 (Oxalate) 1.0 0.85 85 0.94 94 2.5 (Maleate) 1.0 0.98 98 1.00 100 2.5 (Fulvate) 1.0 0.89 89 0.95 95 found/ Recovery found/ Recovery Table 2 Influence of particles in soil extracts and soil matrix on the spectrophotometric determination of CrV' (Procedure I) Particle Sample CrV' added/ CrV1 found/ Recovery size/pm characteristic ng cm-3 ng cm-3 (Yo) >0.45 Clay soil 500 907 181 <0.45 Clay soil 500 47 1 94 (0.1 Clay soil 500 469 94 Peat soil 500 503 101 Peat soil 500 472 94 Peat soil 500 463 93 127 sium hydroxide buffer solutions.27 The efficiency of sorption as a function of pH is shown in Fig.2. A quantitative sorption of Crrrr was obtained in the pH range examined, while most of the Crvl passed through the column resin. Optimum con- ditions for the separation of Cr"' and Crvl were found at pH 3.5-4.5.A pH of 4.0 was chosen for further work, owing to the increased possibility of Crvl reduction in the lower pH range. Even at this pH, about 15% of added Crvl (40 ng ~ m - ~ ) was retained on the resin column. Separation of Cr"' and Crvl ions in synthetic mixtures. Synthetic mixtures of Cr"' (CrCI3) and Crvl (K2Cr207) solutions were prepared at pH 4.0 in various concentration ratios. Chromium(v1) was measured in the eluate. The results are presented in Table 3. The proposed separation of Cr"' from Crvl is satisfactory for CrV1 concentrations not exceeding 20 ng cm-3, probably because of the efficient elution from the resin column. Most of the soil extracts analysed were in this concentration range; at higher concentrations of Crvl, samples should be diluted prior to separation.The reproducibility of measurement, tested for six parallel determinations of Crvl (10 ng cm-3, was found to be +_5.5%. The LOD (30) for CrV1 in aqueous standard solutions and soil extracts was 1.5 ng cm-3. Influence of Cr"' complexes on the separation and determi- nation of Crv' in soil extracts. The existence of negatively charged low relative molecular mass organic complexes of chromium has been demonstrated in soil pore waters.28 Similarly, water-soluble Crlll in the soil solution is expected to be bound to some of the soil borne organic ligands. The presence of these Cr"' complexes in soil extracts might affect the determination of chromate. For this reason, the influence of negatively charged low and high relative molecular mass organic complexes of Crl'l on the determination of Crv' was studied in the concentration range expected in soil extracts. The results are presented in Table 4.It is evident that moderately labile and inert Crlll organic complexes passed partially through the resin column and apparently yielded higher Crvl values. Chromium(II1) citrate and fulvate produced the most severe positive interferences. As these ligands can be found in most soils and waste materials .- ; I s I I 3 4 5 PH Fig. 2 Efficiency of sorption of: A, chromium(m) (200 ng cm-3 of Cr3+) and B, chromium(v1) (40 ng cm-3 of @+), on Chelex 100 resin (50-100 mesh) as a function of pH Table 3 Separation of Cr"' (CrCI3) and CrV' (K2Cr207) ions in synthetic mixtures on Chelex 100 chelating resin (50-100 mesh) at pH 4.0 and determination of CrV1 by ETAAS Addedng cm-3 Cr"' CrV1 200 200 200 100 200 40 200 20 200 4 200 2 CrV' found/ ng cm-3 172 86 35.8 19.9 4.1 1.9 Recovery of CrV1 (% ) 86.0 86.0 89.5 99.6 103.0 95 .O128 ANALYST, FEBRUARY 1992, VOL.117 Table 4 Influence of organically complexed Cr"' on the determination of CrV1 by chelating ion exchange-ETAAS Addedng~rn-~ CrV' found Recovery Cr"' CrV1 ng cm-3 (%) 70 (Citrate) 30 75 .O 250 70 (Oxalate) 30 42.5 142 70 (Maleate) 30 32.5 108 70 (Fulvate) 30 58.2 194 in measurable amounts, analysis of soil extracts employing the separation technique described here will tend to give higher chromate results. Influence of particles on the determination of Crvl in soil extracts. Extracts of clay soils contain considerable numbers of particles below 0.45 pm, which pass through normal liquid chromatography columns.Therefore, the presence of such particles would lead to serious errors in chromate results if such columns were employed for separation. This effect was carefully examined. After centrifugation, soil extracts were filtered through 0.45,0.2 and 0.1 pm membrane filters and the concentration of chromate was determined in these filtrates. Decreasing Crvl values were observed. With KH2P04 (1.5 x 10-2 rnol dm-3) extracts, no further change in the Crvl results was observed by the use of a 0.05 pm filter. Thus, filtering through a 0.1 pm membrane filter was found to be satisfactory for efficient removal of particles from KH2P04 (1.5 x 10-2 rnol dm-3) soil extracts. On the other hand, aqueous soil extracts in the absence of electrolytes contain substantially more fine particles (light yellow colour).A 0.05 pm mem- brane filter should be employed in this instance if accurate Crvl results are to be obtained. Ion-pairing reversed-phase HPLC-E TAAS Influence of p H on the separation of CrVi. The variation of pH between 4.0 and 7.0 in aqueous standard solutions of chromate resulted in a slightly shifted retention time of Crvl during elution, but had no influence on the separation of Crvl. At pH values lower than 3, a substantial reduction of chromate ion concentration in the presence of electron donors was observed. Influence of C P on the separation of Crvl. A standard solution of chromium(n1) chloride (400 ng cm-3) was pre- pared alone or in synthetic mixtures containing 200- 400 ng cm-3 of Crvl.Positively charged Cr"' species did not form ion pairs with the PIC A reagent and showed no influence on the determination of Crvl. In addition, the calibration graph for Crv' in the presence of 400 ng cm-3 of chromium(II1) chloride was found to be linear in the concen- tration range 200-400 ng cm-3. The reproducibility of measurement tested for six parallel determinations of CrV' (1 ng cm-3) was found to be +4.5%. The LOD (30) for aqueous standard solutions was 0.2 ng cm-3 and for soil extracts, 0.3 ng cm-3. Influence of C P on the separation and determination of Crv' in soil extracts. Negatively charged complex species of Cr"' might form ion pairs with the PIC A reagent analogous to chromate.If the elution of these species coincides with that of chromate, an interference in the determination of the latter would result. These effects were therefore investigated carefully by employing several organic complexes of Crrrr typical of the soil environment. The results are presented in Table 5. It is evident that citrate, oxalate and maleate complexes of Cr1I1 had no influence on the determination of Crvl. In contrast, fulvate complexes formed ion pairs with the PIC A reagent that partially coincided with Crvl elution. The severe positive interference of chromium(Ir1) fulvate on the determination of Crvl (Table 5) was similar to that observed in the use of the chelating ion-exchange separation technique (Table 4). The complexation of Cr'Il with fulvic acid ligands in Table 5 Influence of organically cornplexed Cr"' on the determination of CrV' by ion-pairing RP-HPLC-ETAAS Addedng ~ m - ~ CrV1 found Recovery Cr"' CrV1 ng cm-3 (Yo) 70 (Citrate) 30 29.4 98.0 70 (Oxalate) 30 29.5 98.3 70 (Maleate) 30 29.3 97.6 70 (Fulvate) 30 55.7 185 Table 6 Effect of PIC A reagent concentration on the determination of CrV' in soil extracts (KH2P04, 1.5 x 10-2 mol dm-3) by ion-pairing RP-HPLC-ETAAS Peat soil Clay soil Concentration ofPICA CrV1 CrV1 CrV1 reagent/ added/ found Recovery found Recovery mol dm-3 ng cm-3 ng cm-3 (YO) ng cm-3 (%) 5 x 10-4 500 19 3.7 50 6.3 2 x 10-3 500 210 41.1 330 65.4 1 x 10-2 500 580 116.0 490 98.0 2 x 10-2 500 580 116.0 491 98.2 soil solution, therefore, leads to an overestimate of the chromate content of these soils.Influence of particles on the determination of CrV1 in soil extracts. It was found experimentally that small particles present in soil extracts did not influence the results of the determination of Crv' when separation was performed on the HPLC column. Obviously, the dense packing of the HPLC column, in contrast to the Chelex 100 column, retained particles of <0.45 pm. However, deposition of these particles in the column during consecutive determinations results in a continuous increase of the blank values and in column pressure. Consequently, the lifetime of the column was drastically reduced. The lifetime of the HPLC column was prolonged when soil extracts were filtered through 0.1 pm membrane filters. Influence of the concentration of PIC A reagent on the determination of Crvi in soil extracts.The PIC A reagent was used to form an ion pair with the chromate ions for preconcentration from natural pond-water samples. 18 The optimum concentration of the reagent in the sample solution was reported to be 5 x lO-4mol dm-3. Extraction of water-soluble chromate from a soil sample was performed using 1.5 x 10-2 mol dm-3 KH2P04. Owing to the similarity of Cr042- and H2PO4- ions, the latter should also form an ion pair with the PIC A reagent. Additionally, some similar ions that could react with the reagent might be present in the soil solution. Thus, the optimum concentration of PIC A reagent in soil extracts for quantitative formation of an ion pair with the chromate ion in soil should exceed 1 X 10-2 rnol dm-3.In order to prove this assumption, the following experiment was carried out. Chromium(v1) (500 ng cm-3) was added to each of two soil extracts (peat and clay soil extracted with KH2P04, 1.5 x 10-2 rnol dm-3) in which the contents of total soluble chromium were below 1 ng cm-3. Various concentrations of PIC A reagent were added to 10 cm3 of these solutions and the samples were diluted to 15 cm3 with water. The concentra- tions of PIC A in the final solutions for HPLC separation were between 5 x 10-4 and 2 X 10-2 rnol dm-3. Aqueous standard solutions of 500 ng cm-3 of Crvl were prepared with these concentrations of PIC A reagent. Recovery for the aqueous standard solutions of CrV1 was found to be between 98 and 100% for all concentrations of PIC A examined.The effect of the PIC A reagent concentration on the determination of Crvl in soil extracts (extracted with KH2P04, 1.5 X 10-2 mol dm-3) is presented in Table 6. An additional experiment with aqueous extracts of the same soils indicated about 90% recovery for added Crvl at a concentration of PIC A of 5 x 10-4 mol dm-3. From theseANALYST, FEBRUARY 1992, VOL. 117 129 observations and the data in Table 6, it was concluded that KH2P04 formed ion pairs with the PIC A reagent. The optimum concentration of the reagent depends on the KH2P04 concentration in the solution. A concentration of 1 x 10-2 rnol dm-3 of PIC A reagent was found to be optimum when 1.5 X 10-2 rnol dm-3 KH2P04 is used as the extractant solution. Analysis of Soil Extracts In order to evaluate the capability of the methods for the determination of chromium in soil extracts, various types of soil samples were selected and analysed for total soluble and hexavalent soluble chromium.The selection of soils was such as to provide a wide variety of sample matrices characterized either by the physico-chemical nature of the soil or by the soil and waste material together. The first group represented natural soils of different characteristics (clay, sandy, peat and acid soils) with a low and medium concentration of total chromium (40-120 pg g-1) and three natural serpentine soil samples with a high concentration of total chromium (480- 1300 pg g-1). The total chromium was determined by the procedure described previously.29 The second group consti- tuted some natural soils mixed with tannery waste, which were left to settle for 6 months under atmospheric conditions.The total chromium contents of these soils were 2300-3800 pg g-l. Finally, a clay field soil, which had been treated continuously Table 7 Determination of total Cr, total soluble Cr and soluble CrV' in KH2P04 (1.5 x 10-2 rnol dm-3) extracts of various natural soils by RP-HPLC-ETAAS, ~t = 3 Soil sample No. 1 2 3 4 5 6 7 8 9 10 11 Sample characteristic Sandy soil Clay soil Peat soil Clay soil Sandy soil Clay soil Serpentine soil Serpentine soil Serpentine soil Acid soil Acid soil Total soluble TotalCr/ Cr/ 45 2.4 115 4.5 40 2.9 93 17.0 75 5.6 92 16.5 1324 75.4 485 46.1 825 43.4 127 35.3 85 30.5 Mg-' ngg-' Soluble CrV1 ETA AS)/ <5 <5 <8 10 <5 8.8 53.9 27.2 36.9 (RP-HPLC- ngg-' pH of extract 6.0 5.5 6.3 5.7 5.6 5.7 5.4 5.6 5.9 3.8 4.2 for 17 years with tannery waste, was sampled 4 years after the last waste application.The total chromium concentration of this field was between 1000 and 1400 pgg-1, while the concentration on a nearby meadow, which was indirectly contaminated by tannery waste (wind), was 170 pg g-1. Samples were prepared in triplicate as described under Experimental and analysed by the appropriate method. All of the samples were extracted with KH2P04 (1.5 x 10-2 rnol dm-3), but some of the contaminated soils were also extracted with water in order to demonstrate the effects of chromate adsorption and the presence of colloidal particles in the extract solution on the results of the determination of chromium.The results of these measurements are sum- marized in Tables 7-9. The concentration of soluble (KH2P04, 1.5 X 10-2 mol dm-3) hexavalent chromium in natural soils was generally below the detection limit for spectrophotometry and chelating ion exchange-ETAAS. The RP-HPLC-ETA AS technique, being the most sensitive, was therefore used (Table 7). It is evident that the concentrations of soluble (KH2P04, 1.5 x 10-2 mol dm-3) chromate in natural soil samples are very low, with the exception of the serpentine soil samples. Despite the sensitive technique used (LOD = 0.3 ng cm-3) some of the concentrations were below the LOD. The inconsistency in the LOD shown in Table 7 reflects the differences in moisture contents of the soil samples considered.The pH of all the soil extracts was between 5.5 and 6.3 with the exception of the acid soil samples No. 10 (pH = 3.8) and No. 11 (pH = 4.2). When these two samples were analysed by RP-HPLC- ETAAS, the Crvl peak did not appear either in the soil extracts or in the soil extracts to which Crvl was added due to a low pH and the presence of reducing substances in these soils. Chromium(v1) added to the sample extract was reduced immediately. Despite the relatively high total soluble chro- mium concentration in acid soils (Table 7) the expected chromate concentration should be extremely low owing to the nature of these soils. Chromium(v1) was also added to the extracts of other soil matrices. Recoveries obtained were between 98 and 115%, which indicated that RP-HPLC-ETAAS was a suitable technique for the determination of soluble (KH2P04, 1.5 x 10-2 rnol dm-3) hexavalent chromium in natural soils with normal pH values.Table 8 Determination of total Cr, total soluble Cr and soluble CrV1 in KH2P04 (1.5 X by tannery waste using spectrophotometry, chelating ion exchange-ETAAS and RP-HPLC-ETAAS, n = 3 rnol dm-3) extracts of various soils contaminated Soluble Crv'/ng g-1 Soil sample No. I I1 111 IV V VI Sample characteristic Sandy soil* Clay soil* Peat soil* Clay soil? Clay soil? Clay soil? Total Cr/ 2360 2470 3730 1400 1050 169 Pg g-' Total soluble Cr/ 776 1621 960 384 233 37 ngg-' Chelating ion Spectro- exchange- photometry ETAAS 482 517 1245 1389 <750 658 <450 290 ~ 4 5 0 169 33 - RP-HPLC- ETAAS 434 909 822 297 184 34 * Tannery waste treated, analysed 6 months after waste application.t. Seventeen years of continuous tannery waste application, analysed in the fourth year after the last application. pH of extract 6.1 5.3 6.4 6.0 5.8 5.4 Table 9 Determination of total Cr, total soluble Cr and soluble CrV1 in aqueous soil extracts of various soils contaminated by tannery waste using chelating ion exchange-ETAAS and RP-HPLC-ETAAS, n = 3 Soluble Crv'/ng g-1 Soil Sample Total Total soluble Chelating ion RP-HPLC- pH of sample No. characteristic Cr/pg g- * Crhg g- exchange-ETAAS ETAAS extract IV Clay soil* 1400 V Clay soil* 1050 VI Clay soil* 169 238 140 48 183 92 38 143 6.7 62 6.8 <5 5.7 * Seventeen years of continuous tannery waste application, analysed in the fourth year after the last application.130 ANALYST, FEBRUARY 1992, VOL.117 Contaminated soil samples were analysed in triplicate by all three techniques (Table 8). Spectrophotometry could not be applied generally to the analysis of some of the contaminated soil extracts because of the poor LODs (30 ng cm-3 for Crvl). Results for the determination of Crvl obtained by RP-HPLC- ETAAS and chelating ion exchange-ETAAS agreed very well for soil samples treated with tannery waste, analysed 4 years after the last application (samples IV, V and VI). It was found experimentally that after consecutive RP-HPLC-ETAAS analyses of these samples, the blank value was constant but a slight broadening of the chromate peak appeared. When extracts of soil samples freshly treated with tannery waste are analysed (samples I, I1 and III), the influence of the waste matrix should be taken into account.Tannery waste is a protein-based matrix with a high content of CrlI1, organic polymers and reducing substances which could produce interferences in the determination of Crvl. In order to suppress these effects, the soil extracts were diluted 1 + 7.5 prior to RP-HPLC determinations and 1 + 2 prior to chelating ion exchange. The samples were not diluted for spectrophoto- metric analyses because of the poor LOD of the technique. It should be emphasized that with these types of soil extracts excessively high and variable blanks appeared when using RP-HPLC-ETAAS, making the results uncertain and in disagreement with those given by the other two techniques.In addition, the lifetime of the RP-HPLC columns was drastically reduced. Results using spectrophotometry and chelating ion exchange-ETAAS correlated well for these samples, pro- vided that the concentration of chromate in the soil extracts was >30ngcm-3. This good agreement indicates that the concentrations of organic ligands forming relatively inert Cr complexes, present in the samples should be low, otherwise much higher results would be obtained by chelating ion exchange-ETAAS. The accuracy of the result for the freshly treated peat soil sample (sample 111) obtained by chelating ion exchange-ETAAS is questionable as no reliable comparison could be made using the other two techniques. According to the expected interference effects of chromium(I1r) fulvate, the reported value might be too high.A comparison of the results from samples IV, V and VI (Tables 8 and 9) obtained by RP-HPLC-ETAAS reflected the effect of chromate adsorption on clay minerals, which was reported by Bartlett and James.9 On the other hand, from the results obtained by chelating ion exchange-ETAAS two effects were superimposed: the effect due to adsorption and the effect produced by the colloidal particles present in the aqueous extract. Conclusions Light scattering on colloidal particles can produce severe positive systematic errors in the spectrophotometric determi- nation of Crvl in soils containing a high clay fraction. Similarly, particles passing through the ion-exchange column contribute to higher results for Crvl. In HPLC separation, particles had no direct influence but reduced the lifetime of the RP columns. A KH2P04 (1.5 X mol dm-3) extraction solution efficiently released adsorbed chromate9 and pre- vented formation of colloidal solutions.Nevertheless, filtering through a 0.1 pm membrane filter was found to be necessary. A low LOD for Crvl in soil extracts (0.3 ng cm-3) using RP-HPLC-ETAAS enabled the determination of soluble hexavalent chromium in most natural soils. For soil samples contaminated by tannery waste, spectrophotometry was found to be suitable only for heavily contaminated soils. This technique may produce lower chromate results probably owing to partial reduction of Crvl during the measuring procedure in acidic media and a reducing environment. Analysis of samples treated with tannery waste, 4 years after the last application, showed very good agreement in the determination of soluble (KH2P04, 1.5 x 10-2 mol dm-3) Crvl between RP-HPLC-ETAAS and chelating ion exchange -ETAAS techniques, taking into account soil characteristics.Analysis of freshly treated tannery waste soil samples employ- ing RP-HPLC-ETAAS gave uncertain results because of the influence of the waste matrix and the variable blank. Spectrophotometry and chelating ion exchange-ETAAS results for these samples agreed well when the chromate concentration in the soil extracts was above 30 ng cm-3. Despite the considerable experimental efforts associated with the preparation of this paper, the reliable determination of chromium in soil extracts still remains a problem at least for some particular situations.Nevertheless, the experimental evidence presented demonstrates the complex nature of the effects of soil and tannery waste. This work was supported by the Research Council of Slovenia and US Environmental Protection Agency (project JF 908). The authors thank Professor S. A. Katz for valuable sugges- tions in preparing this manuscript and Dr. A. R. Byrne for assistance with linguistic correction. The authors also thank Marko Zupan, for providing some soil samples and their physical characterization. 1 2 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 29 References Fajgelj, A., and Kosta, L., Vestn. Slov. Kern. Drus., 1987, 34, 175. Pankow, J. F., and Januer, G. E., Anal. Chim. Acta, 1974,69, 97. Naranjit, D., Thomassen, Y., and Van Loon, J. C., Anal. Chim. Acta, 1979, 110,307. Minoia, C., Mazzucotelli, A., Cavalleri, A., and Minganti, V., Analyst, 1983, 108, 481. Bergmann, H., and Hardt, K., Fresenius 2. Anal. Chem., 1979, 297, 381. Donaldson, E. M., Talanta, 1980, 27, 779. Vos, G., Fresenius Z. Anal. Chem., 1985. 320, 556. de Andrade, J. C., Rocha, J. C.. and Baccan, N., Analyst, 1985, 110, 197. Bartlett, R., and James, B., J. Environ. Qual., 1979, 8, 31. James, B. R., and Bartlett, R. J., J. Environ. Qual., 1983, 12, 169. James, B. R., and Bartlett, R. J., J. Environ. Qual., 1983, 12, 173. James, B. R., and Bartlett, R. J., J. Environ. Qual., 1983, 12, 177. James, B. R., and Bartlett, R. J., J. Environ. Qual., 1984, 13, 67. Heringer Donmez, L. A., and Kalenberger, W. E., J. Am. Leather Chern. Assoc., 1989, 84, 110. Krull, I. S., Bushee, D., Savage, R. N., Schleicher, R. G., and Smith, S. B., Anal. Lett., Part A , 1982, 15, 267. Krull, I. S., Panaro, K. W., and Gershmann, L. L., J. Chromatogr. Sci., 1983, 21, 460. Lawrence, K. E., Rice, G. W., and Fassel, W. A., Anal. Chern., 1984, 56, 292. Syty, A., Christensen, R. G., and Rains, T. C., At. Spectrosc., 1986, 7 , 89. Syty, A., Christensen, R. G., and Rains, T. C., J. Anal. At. Spectrorn., 1988,3, 193. Roehl, R., and Alforque, M. M., At. Spectrosc., 1990, 11,210. Mayazaki, A., and Barnes, R. M., Anal. Chem., 1981,53,364. Colella, M. B., Siggia, S., and Barnes, R. M., Anal. Chem., 1980, 52, 967. Isozaki, A., Kumagai, K., and Utsumi, S., Anal. Chim. Acta, 1983, 153, 15. Florence, T. M., and Batley, G. E., Talanta, 1976. 23, 179. Knudtsen, K., and O’Connor, G. A., J. Environ. Qual., 1987, 16, 85. Nazario, C. L., and Menden, E. E., J. Am. Leather Chem. Assoc., 1990,85, 212. Perrin, D. D., Aust. J. Chem., 1963, 16, 572. Brown, L., Haswell, S. J., Rhead, M. M., O’Neill, P., and Bancroft, K. C. C., A n a l p , 1983, 108, 1511. Ajlec, R., Cop, M.. and Stupar, J., Analyst, 1988, 113, 585. Paper 1 lOI559A Received April 3, 1991 Accepted September 2, 1991