ANALYST, JANUARY 1991. VOL. 116 35 Two-step Coprecipitation Method for Differentiating Chromium Species in Water Followed by Determination of Chromium by Neutron Activation Ana I ysis Chi-Ren Lan, Chia-Liang Tseng and Mo-Hsiung Yang* Institute of Nuclear Science, National Tsing Hua University, 30043 Hsinchu, Taiwan Zeer B. Alfassi Department of Nuclear Engineering, Ben Gurion University, 84 102 Beer Sheva, Israel A two-step method for the determination of Crvl and Crlll in natural waters was developed. The method is based on the variation of the coprecipitation yields with Pb(PDC)2 (PDC = pyrrolidine dithiocarbamate) as a function of pH. By using two different pH values both species can be determined separately. Firstly, Crvl was coprecipitated at pH 4.0 and then Cr"1 was separated at pH 9.Total chromium was determined by reduction of Crvl followed by coprecipitation at pH 9. The validity of the procedure was checked with the National Institute of Standards and Technology Standard Reference Material 1643b Trace Elements in Water and the result was found to be in good agreement with the certified value. Keywords: Water; chromium species analysis; ammonium pyrrolidine dithiocarbamate; coprecipitation; neutron activation analysis Chromium is present in natural waters in two different oxidation states, Cr111 and CrVI. The former is considered an essential element in mammals, whereas the latter is con- sidered to be a toxic material.14 Thermodynamic calculations indicated that in natural waters Cr should exist almost exclusively as CrV1.7 However, it was found experimentally that the actual ratio of Cr"' to CrV1 in natural waters varied from 0.02 to 0.99.8 Arrhenius and Bonattig pointed out that this variation and contradiction with theory might be due to the in situ coprecipitation of chromate only, with strontium or barium sulphate.This selective coprecipitation can also be used as an analytical tool for the separation of the two species followed by their individual determination. Chuecas and Riley10 studied the coprecipitation of Cr from water using W r as a radiotracer. They found that both aluminium and iron(m) hydroxides (hydrated oxides) will coprecipitate CrlI1 effi- ciently. The pH range for 99% coprecipitation is considerably larger for iron(1ri) hydroxide (pH 7.0-9.0) than for aluminium hydroxide (pH 7.5-8.0).When coprecipitating CrV1 spiked with slCrv1, about 1.2% of the Cr was precipitated. This might be due to partial coprecipitation caused by a small amount of slCr111 in the radiotracer. Fukai11 measured both Cr111 and CrV1 in sea-water by coprecipitation with iron(r1r) hydroxide first from the untreated water and then after reduction of the sample with sodium sulphite in acidic medium. Several studies have been carried out on the extraction of Cr species by means of ammonium pyrrolidine dithiocarbamate (APDC)-ethyl methyl ketone (EMK) or diethyl dithiocarbamate(DDTC) with EMK. However, the percentage extraction varied considerably (5&100%).12,13 De Jong and Brinkman14 selectively determined CrVI and Cr111 in sea-water using solvent extraction,. They found that CrVI was extracted with high efficiency (>99%) from various acidic solutions with tertiary amines.These workers used a pH of 2 (0.01 mol dm-3 HCI) and Aliquat 336 as the extractant; Aliquat 336 is a mixture of methyl trialkylammonium chlorides with alkyl groups that are mainly C8-Cl0. The extracting organic solvent was toluene. Chromium(1ii) was not extracted at all in this medium. However, by using the same extractant at pH 6-8 and in the presence of at least 1 mol dm-3 thiocyanate (in CC14 rather than toluene, in order to dissolve * To whom correspondence should be addressed. the KSCN), CrI11 was quantitatively extracted while none of the CrV1 was extracted in this pH range. One of the methods of Cr speciation recommended by the US Environmental Protec- tion Agency (EPA)15 involves the extraction of CrVI with APDC into isobutyl methyl ketone (IBMK); however, several problems are associated with this method.16 Isozoki et a1.17 studied the different ionic species of Cr in natural water using ion-exchange chromatography; the work was similar to that of Naranjit et a1.18 and Leyden et a1.19 Isozoki et al.used a chelating column of Chelex 100 for the quantitative adsorption of Cr"1 while CrVI was not absorbed and passed through the column. Chromium(v1) was deter- mined after reduction to Cr"1. Miyazaki and Barnes20 used a poly(dithi0carbamate) chelating resin. Only CrV* was retained on the column, the Cr"' passed through. Total Cr was obtained after oxidizing CrIII to CrV' with KMn04 in acidic media.Wai et a1.21 differentiated between Cr"' and CrVI by extracting CrVI from natural waters into chloroform with DDTC followed by back-extraction into aqueous HgII solu- tion for determination by graphite furnace atomic absorption spectrometry. The Cr"1 remaining in the extracted solution can be oxidized to CrVI with KMn04 and then extracted with DDTC. Subramanian22 developed procedures using the APDC- IBMK extraction system for the selective determination of Cr111, and the simultaneous determination of Cr"* plus CrV1, without the need to convert Cr"1 into CrV1. He used phthalate buffer and found that at about 0.02-0.1% of phthalate, both species were extracted efficiently, whereas above 0.8% phthalate, only CrVI was quantitatively extracted (all in the pH range 2.5-4.0).For both atomic absorption spectrometry and inductively coupled plasma atomic emission spectrometry and also for spectrophotometric determination it is preferable to obtain the concentrated sample in a liquid phase; however, for neutron activation analysis (NAA) and also for X-ray fluor- escence spectrometry it is preferable to have the sample in the form of a solid. This is particularly true for Cr as thermal neutron activation leads to two radionuclides of which the short-lived radio- nuclide "Cr has a low abundance of the parent isotope (2.36%), a relatively small cross-section for formation (0.36 barn) and, most significantly, it is almost a pure (3-emitter and emits very few y-rays (0.043%). The lower limit of detection using this radionuclide is very high and measurement of Cr by36 ANALYST, JANUARY 1991, VOL.116 NAA is carried out using the long-lived radionuclide 51Cr (half-life, 27.71 d). In order to obtain high sensitivity the sample should be irradiated at high fluxes for long periods of time (several days or at least 10-20 h). With these long irradiation times liquid samples will suffer considerable radiolysis, leading to the formation of large amounts of gases, which are likely to lead to explosion of the irradiation ampoule. Hence a solid sample should be used and, rather than drying the liquid sample, it is preferable to pre-concen- trate the trace elements by coprecipitation. The best copreci- pitants for NAA will be compounds that have small thermal neotron absorption cross-sections and which do not form radioisotopes on neutron absorption or where the radioiso- topes formed are very short-lived and/or are only (3-emitters.Materials that fulfil these criteria are organic compounds of lead and bismuth. In earlier w0rks~23-27 the precipitation of several trace elements from natural waters and biological fluids with Pb(PDC)2 and Bi(PDC)3 (PDC = pyrrolidine dithiocarbamate) have been described; the present work is concerned with the two species of Cr, viz. , Cr"' and CrvI. Nakayama et ~ ~ 1 . ~ 8 found that in the pH range 4-10 both CrIII and CrVI were coprecipitated with bismuth oxide. Hence they first coprecipitated Cr"1 with iron(m) hydroxide at about pH 8, and then CrV1 was collected at the same pH with bismuth oxide. Pik et a1.29 first coprecipitated CrIII with iron(rI1) hydroxide at pH 8.5 and then coprecipitated CrVI from the remaining solution with Co(PDC)2 at pH 4.Experimental All chemicals used were of analytical-reagent grade and the solutions were prepared using doubly distilled, de-ionized water. Two methods were used to determine the coprecipitation yields of CrIII and CrVI with Pb(PDC)2. In the first, the radiotracer 51Cr was used. The coprecipitation yield was measured from the radioactivity counts [with an NaI(T1) detector] of the original radiotracer solution and the precipi- tate. Solutions of 5lCrvI and 51C1-111 were prepared by the method of Collins et a1.30 About 50 mg of K2Cr04 or K2Cr207 were irradiated in the reactor for 2-3 d. The irradiated salt was dissolved in 5 ml of a solution containing 15 mg of Zn(N03)2 and 50 mg of Cr03.After complete dissolution, 3 ml of 1 mol dm-3 KOH were added and the solution was heated at 90-95 "C for about 30 min. While the reaction mixture was still hot, 1 ml of a solution containing 3 mg of Zn(N03)2 was added, with stirring. The resulting suspension was filtered through a 10 x 6 mm i.d. column of alumina or celite washed with 5 ml of 1 mol dm-3 KOH and 5 ml of 1 X 10-4 mol dm-3 KOH, and kept wet with a KOH solution of pH 10. The filtered solution was used as a radiotracer for CrVI. Some of the 51Cr changed its valency due to the Szilard-Chalmers process and was retained on the column as Cr"1. The 51CrIII was eluted from the column with 5 ml of 1 mol dm-3 HC1 and was used as a radiotracer for CrIII.In the second method, standard solutions of unlabelled CrIII and CrVI were used and the standard solution and the precipitate were analysed simultaneously by NAA. The coprecipitation studies were carried out by the addition of 1 ml of a standard solution containing 1 mg ml-1 of either CrI" or CrVI to 250-500 ml of tap water (for the radiotracer experiments) or distilled water (for the NAA experiments) followed by the addition of 2 mg of Pb(N03)2, 5 ml of 1 mol dm-3 acetate buffer and adjustment of the pH with HN03. Then, 100 mg of APDC were added to the solution which was stirred for 30 min. The precipitates were filtered through a 0.45 pm Millipore filter (Gelman) and dried in a desiccator containing silica gel until completely dry. In the NAA experiments the dried samples were introduced into polyethylene vials which were heat-sealed. A standard was prepared by introducing 1 ml of the standard solution into a polyethylene vial and heating to dryness under an infrared lamp.The precipitates and the standard were placed in an irradiation capsule and irradiated for 30 h at a flux of 5 x 1012 n cm-2 s-1. The radioactivity induced in the samples was measured by a Ge(Li) detector connected to a multi-channel analyser, after 1 week of cooling. The coprecipitation yields were calculated as the radioactivity ratio of the samples to the standard. Results and Discussion The results for the recovery of Cr"1 and CrVI by coprecipita- tion with Pb(PDC)2 as a function of pH for 250 ml of distilled water, tap water and sea-water are given in Table 1.As can be seen from these results, it is possible to precipitate CrVI almost exclusively in the pH range 2.5-4.5. Chromium(Ir1) cannot be coprecipitated alone, as at pH 9 about 1644% of CrVI is also precipitated. However, a two-step coprecipitation on the same sample can give information about the concentrations of both CrVI and Cr"'. The sample is adjusted to pH 4 and Pb(N03)2 (2 mg) and APDC (100 mg) are added to the solution. The solution is stirred for 30 min and then filtered on a 47 mm Millipore filter (0.45 pm). The precipitate on the filter-paper is used to determine CrVI. The filtrate is adjusted to pH 9 with 25% ammonia solutions (about 3 ml). Then, 100 mg of APDC and 2 mg of Pb(N03)2 are added, the solution is stirred (for 30 min) and filtered on a 47 mm diameter 0.45 pm Millipore filter and the precipitate is used for the determina- tion of CrIII.Table 2 gives the results for the recovery of CrIII and CrVI from 2 1 of sea-water or tap water spiked with CrIII or CrVI at a total concentration of 20 ng ml-1. As can be seen from this table even at this low concentration there is nearly a 100% recovery (about 95-105%). Total Cr Determination Total Cr can be determined either by oxidation of Cr"1 to CrVI and coprecipitation at pH 4 or by reduction of CrVI to Cr"' and coprecipitation at pH 9.0. It is easier to reduce CrVI than to oxidize CrIII and consequently the latter option was preferred. Chromium(v1) was reduced with NaHS03. Table 1 Recovery of CrlI1 and Crw by coprecipitation with Pb(PDC)2 as a function of pH. All values in per cent.Conditions: water sample, 250 ml; Cr"' or Crw, 100 pg; APDC, 100 mg; and Pb(N03)2, 2 mg Distilled water Tap water Sea-water PH 3 4 5 6 7 8 9 Cr"1 CrVI Cr"' Crw Cr"' Crm 0.8 92 0.8 92 0.5 95 1.8 98 1.8 95 0.5 95 3.6 95 3.6 95 0.5 95 13.2 95 13 95 4 34 88 58 66 58 67 14 93 52 93 52 96 14 97 44 97 44 95 16 Table 2 Recovery tests carried out by spiking sea-water and tap water samples with Cr"' or CrVI. Water sample, 2 1; Cr, 20 ng ml-l Recovery (%) Sample CrVI Sea-water 101.6 95.7 105.3 Aver age : 100.9 Tap water 102.9 101 .o 100.7 Average : 101.5 Cr"1 98.5 99.2 102.1 99.9 98.6 95.9 103.1 99.2ANALYST, JANUARY 1991, VOL. 116 37 Table 3 Recovery of CrVI by reduction and precipitation at pH 9 as a function of the amount of NaHS03 added (CrVI, 104 pg) AmountofNaHS03/mg 1.7 6.7 16.7 33.3 66.6 83.5 167 500 Recovery (% ) 25.4 28.7 41.5 56.3 100 99.3 97.6 99.9 Sample (2 I) w pH = 4.0 100 mg of APDC 2 mg of Pb(NO3I2 stirred for 30 rnin 62 mg of NaHS03 boiled for 5 min i o o mg of APDC 2 mg of Pb(NO3j2 stirred for 30 min pH = 9.0 2 mg of Pb(N03j2 stirred for 30 min mg of APDC NAA NAA NAA CrVl only Crlll only Total amount of Cr Fig 1 Scheme for the speciation of Cr"1 and CrVI and for the determination of total Cr Different amounts of NaHS03 were added to a series of solutions containing 104 pg of CrvI.The solutions were stirred and then placed in a microwave oven for heating (2 min at 100% power followed by 3 min at 60% power). After boiling for 5 min the solutions were rapidly cooled in an ice-bath.The pH was adjusted to 9 with 1 ml of 25% ammonia solution and the Cr"1 in the solution was then determined as described above. Table 3 shows the percentage recovery of CrVI obtained with this method as a function of the amount of NaHS03 added. It can be seen that 66.6 mg of NaHS03 are sufficient for complete reduction and recovery. Consequently in later experiments, about 70 mg of NaHS03 were used for the reduction of CrVI in the procedure for the determination of total Cr. Fig. 1 illustrates schematically the determination of the two species of Cr and of total Cr. This method of separate determination of the two species of Cr has advantages over most of the previous methods in that it does not require the oxidation of Cr"1 to Crvr or the reduction of Crvr to Cr"' which is usually carried out with an excess of the reagents.The proposed method does, however, suffer from the disadvan- tage that even if one is interested only in Cr"', one must first coprecipitate CrVI. If only CrIII is required, then coprecipita- tion with hydrated iron(i1i) hydroxide might be a better procedure. Real Samples Three samples of sea-water, well water and tap water were analysed according to the scheme illustrated in Fig. 1. The results are given in Table 4. It can be seen that the agreement between the values for CrVI plus Cr"1 and total Cr is good. A certified reference water, viz., National Institute of Standards and Technology Standard Reference Material 1643b Trace Elements in Water, was also analysed as a quality control material during the measurement of the total amount of Cr.Table 4 Results for the determination of CrW3-W in natural waters. Results given are mean k standard deviation ( n = 3) Foundng ml-I Total Cr Sample CrV1 Cr"' Sea-water 0.10 f 0.01 0.49 f 0.04 0.54 k 0.03 Well water 0.13 2 0.05 0.11 _t 0.02 0.25 _+ 0.02 Tap water 0.14 k 0.04 0.20 k 0.05 0.33 k 0.05 The resulting value (18.1 _t 1.5 ng ml-1, mean of three sample analyses f standard deviation) was in good agreement with the certified value (18.6 ng ml-1). Limit of Detection Blank values were measured by using tap water as the metal-free solution after coprecipitation of total Cr. Analysis of five samples of the blank gave an average value of 0.261 k 0.006 ng ml-1. Employing the usual convention, that the detection limit is 4.7 times the square root of the background, or 4.7 times the standard deviation, gave a limit of detection of 0.03 ng ml-1.Mechanism The variation of the amount of CrVI with pH might be associated with the equilibrium between Cr042- and Cr2O72-. At acidic pH, Cr042- is the major species and it is this species that is coprecipitated, while Cr2O72- is not precipitated to any great extent. The precipitation of CrI" at pH 9 cannot be in the form of Cr(PDC)3. It is possible that the insoluble hydrated CrVI oxide is coprecipitated with Pb(PDC)2. It is known31 that Cr"I exists at acidic pH as the hexaaqua ion [Cr(H20)#+, which has a pK of 4. At higher pH the hydroxide ion [Cr(H20),(OH)]2+ is formed, which can give soluble dimers and polymers. At even higher pH values, dark green gels are formed which are coprecipitated with Pb(PDC)2.The authors thank the National Science Council of the Republic of China for financial support of this work. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 References Bowen, H. J., Trace Elements in Biochemistry, Academic Press, New York, 1966. Pankon, J. F., and Janauer, G. E . , Anal. Chim. Acta, 1974,69, 97. Valkovic, V., Trace Element Analysis, Taylor and Francis, London, 1975, p. 107. Schwartz, K . , and Mertz, W., Arch. Biochem. Biophys., 1957, 72, 515. Schwartz, K . , and Mertz, W., Arch. Biochem. Biophys., 1959, 85,292. Mertz, W., Phys. Rev., 1969, 49, 163. Elderfield, H., Earth Planet. Sci. Lett., 1970, 9, 10. Florence, T. M., and Batley, G. E., CRC Crit. Rev. Anal. Chem., 1980,9, 219.Arrhenius, G., and Bonatti, E., Prog. Oceanogr., 1965, 3, 7. Chuecas, L., and Riley, J. P., Anal. Chim. Acta, 1966,35,240. Fukai, R., Nature (London), 1967, 213, 901. Filbert, T. R., and Clay, A. M., Anal. Chim. Acta, 1973, 67, 289. Hiro, R., Owa, T., Takaoka, M., Tanaka, T., and Kawahara, A., Bunseki Kagaku, 1976, 25, 122. De Jong, G. J . , and Brinkman, U. A. Th., Anal. Chim. Acta, 1978, 98,243. Methods for Chemical Analysis of Water and Wastes, US Environmental Protection Agency, Cincinnati, OH, 1979. Sturgeon, R. E., Berman, S. S., Desaulniers, A,, and Russel, D. S . , Talanta, 1980, 27, 85.38 ANALYST, JANUARY 1991, VOL. 116 17 18 19 20 21 22 23 24 25 Isozoki, A., Kumazai, K., and Utsume, S., Anal. Chim. Acta, 1983, 153, 15. Naranjit, D., Thomassen, Y., and Van Loon, J. C., Anal. Chim. Acta, 1979, 110, 307. Leyden, D. E., Channell, R. E., and Blount, C. W., Anal. Chem., 1972, 44, 607. Miyazaki, A., and Barnes, R. M., Anal. Chem., 1981,53, 364. Wai, C. H.. Tsay, L. H., and Yu, J. C., Mikrochim. Acta, Part 11, 1987, 73. Subramanian, K. S., Anal. Chem., 1988, 60, 11. Lavi, N., and Alfassi, Z. B., J. Radioanal. Nucl. Chem, 1988, 126, 361. Lavi, N., Mantel, M., and Alfassi, Z. B., Analyst, 1988, 113, 1855. Lavi, N., and Alfassi, Z. B., J. Radioanal. Nucl. Chem., 1989, 130,71. 26 Lavi, N., and Alfassi, Z. B., Analyst, 1990, 115, 817. 27 Lan, C. R., Sun, Y. C., Chao, J. H., Chung, C., Yang, M. H., Lavi, N., and Alfassi, Z. B., Radiochim. Acta, in the press. 28 Nakayama, E., Kuwamoto, T., Tokoru, H., and Fujingawa, T., Anal. Chim. Acta, 1981, 131, 247. 29 Pik, A. J., Eckert, J. M., and Williams, K. L., Anal. Chim. Acta. 1981, 124, 351. 30 Collins, K. E., Collins, C. H., Yang, M. H., Ke, C. N., Lo, J. M., and Yeh, S. J., J. Radioanal. Chem., 1972, 10, 197. 31 Cotton, F. A., and Wilkinson, G., Advanced Inorganic Chem- istry, Wiley, New York, 4th edn., 1980, p. 727. Paper 0102883E Received June 26th, 1990 Accepted September 13th, I990