methods for the photometric determination of reactive bromine and ...
PII: S0043-1354(00)00216-5
Wat. Res. Vol. 34, No. 18, pp. 4343±4350, 2000 7 2000 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0043-1354/00/$ - see front matter
www.elsevier.com/locate/watres
METHODS FOR THE PHOTOMETRIC DETERMINATION OF REACTIVE BROMINE AND CHLORINE SPECIES WITH ABTS ULRICH PINKERNELL, BERND NOWACK*, HERVEÂ GALLARD and URS VON GUNTEN$ Swiss Federal Institute for Environmental Science and Technology (EAWAG), Ueberlandstr. 133, CH-8600, DuÈbendorf, Switzerland (First received 15 October 1999; accepted 10 March 2000) AbstractÐNew methods for the determination of reactive bromine and chlorine species are presented. Hypobromous acid (HOBr) and all three bromamines species (NH2Br, NHBr2, NBr3) are analyzed as a sum parameter and hypochlorous acid (HOCl), monochloramine (NH2Cl) and chlorine dioxide (ClO2) can be determined selectively. However, no distinction is possible between HOCl and the active bromine species. The bromine and chlorine species react with ABTS (2,2-azino-bis(3ethylbenzothiazoline)-6-sulfonic acid-diammonium salt) to a green colored product that is measured at 405 or 728 nm. Free chlorine and NH2Cl can be measured in the presence of ozone. The method is therefore suitable if combinations of disinfectants are used, such as chlorine/chlorine dioxide or chlorine/ozone. In natural waters, the method provides a detection limit for all chlorine/bromine species of less than 0.1 mM. The colored reaction product is very stable and allows a ®xation of the chlorine/ bromine species in the ®eld and subsequent determination of the absorption in the laboratory. 7 2000 Elsevier Science Ltd. All rights reserved Key wordsÐhypobromous acid, bromamines, chlorine, chlorine dioxide, monochloramine, drinking water, analysis
INTRODUCTION
tion by-products when ozone or chlorine is applied to treat bromide-containing natural waters. It has been shown in studies on bromate formation during ozonation of bromide-containing natural waters that hypobromous acid is a key intermediate (von Gunten and HoigneÂ, 1994). In the presence of ammonia HOBr is scavenged in a fast reaction forming bromamine (NH2Br) and dibromamine (NHBr2) (Inman and Johnson, 1984), which are not oxidized to bromate directly. Therefore, ammonia addition may be one possible control option for bromate minimization (Pinkernell and von Gunten, 1999). To get an overall bromine mass balance during water treatment, it is therefore necessary to assess the fate of these reactive bromamines in addition to HOBr.
Bromine species: Reactive bromine species, such as hypobromous acid (HOBr) and bromamines (NH2Br, NHBr2, NBr3) may occur due to two main reasons during water disinfection: (1) they are used directly as disinfectants; and (2) they are disinfection by-products during ozonation and chlorination of bromide-containing waters: 1. HOBr is used for disinfection of swimming pool water and for control of biofouling (Fisher et al., 1999). Due to the presence of ammonia in these applications, bromamines are formed immediately, which have comparable disinfection properties as HOBr (Floyd et al., 1978). All these active bromine species show a higher toxicity to biofouling organisms as their corresponding chlorine analogues (Fisher et al., 1999). 2. HOBr and bromamines are formed as disinfecPresent address: Institute of Terrestrial Ecology (ITOÈ), Swiss Federal Institute of Technology (ETH), Grabenstr. 11a, CH-8952 Schlieren, Switzerland. $Author to whom all correspondence should be addressed. Tel.: +41-1-823-5270; fax: +41-1-823-5028; e-mail: [email protected] *
HOBr and bromamines can be determined by a DPD (diphenylendiamine)-based method. However, the stability of the colored reaction product can be expected to be low as described for the chlorine analysis (Jandik and EichelsdoÈrfer, 1980). In addition, this method requires internal calibration using standard solutions of HOBr and bromamines.
4343
4344
Ulrich Pinkernell et al.
This is dicult to carry out due to the low stability of these reactive compounds. Chlorine species: Chlorine, consisting of hypochlorous acid (HOCl) and of hypochlorite ion (ClOÿ) at neutral pH, and chlorine dioxide are used worldwide in large quantities for disinfection. Therefore, numerous analytical techniques are available for their determination (APHA, 1989). The need for another analytical method to determine one of the disinfectants alone is therefore not urgent. However, chlorine and chlorine dioxide are also increasingly used as a mixture in water treatment to minimize the formation of by-products (Katz et al., 1994). The exact analysis of these mixtures is troublesome. The standard procedure for the distinction of the two oxidants consists in the addition of glycine to the sample, which masks chlorine by formation of chloroaminoacetic acid (APHA, 1989). Chlorine dioxide can then be measured selectively by a colorimetric method and the chlorine concentration is calculated by dierence of the total oxidant concentration and chlorine dioxide. Chlorite, which is always present when chlorine dioxide is added to water, is also measured and further complicates the analysis (APHA, 1989). A method to distinguish between bromine or chlorine and chlorine dioxide has also been described (Emerson, 1994). However, this method has a detection limit of about 0.5 mg lÿ1 Cl2, which is not suf®ciently sensitive for the analysis of drinking water samples. The method also uses the addition of glycine to suppress free chlorine. It has been reported, however, that glycine interferes with the ClO2 measurement in the DPD method (Jandik and EichelsdoÈrfer, 1980; Palin, 1975). Alternatively, chlorine dioxide can be purged with nitrogen from the water (Aieta et al., 1984). This method is very time-consuming because every sample has to be purged for more than 30 min. Photometric methods for the selective determination of chlorine dioxide with no interference from chlorine have been described (Hofmann et al., 1998; Sweetin et al., 1996). It is, however, not possible to also analyze chlorine using one of these methods. The need for a simple method to distinguish between chlorine and chlorine dioxide at low concentrations is therefore given. A method that uses only one reagent produces a stable color and allows the distinction between chlorine, chloramine and chlorine dioxide without interference of chlorite, should be found. The same need is given for reactive bromine species. As there are no applications where signi®cant concentrations of bromine and chlorine species are present in the water treatment ®eld, a distinction between chlorine and bromine is not necessary. We have found that ABTS (2,2-azino-bis(3-ethylbenzothiazoline)-6-sulfonic acid-diammonium salt), a color reagent, ful®lls these requirements. ABTS is a well-known substrate for enzymatic peroxide tests
(Bergmeyer, 1986) and percarboxylic acid analysis (Pinkernell et al., 1997). The colorless ABTS is oxidized undergoing a one-electron transfer as depicted in Scheme 1. The reaction product ABTS + is a stable green colored radical. It has a broad absorbance spectrum containing several maxima with high molar absorptivities at 415, 650, 728, and 815 nm (Scott et al., 1993). It has been shown before that because the reaction with ABTS is fast and has a known stoichiometry, a direct calculation of the analyte using the molar absorptivity of the colored oxidation product is possible (Pinkernell et al., 1997). EXPERIMENTAL SECTION
Materials All water used was from a Barnstead B-Pure system. All chemicals have been of p.a. quality (Fluka) if not stated otherwise. ABTS (2,2-azino-bis(3-ethylbenzothiazoline)-6sulfonic acid-diammonium salt, Aldrich) stock solutions (1 g/l) were used for all experiments and stored at 48C. After a few weeks the ABTS solution had a slightly higher blank absorption and was replaced. A 1 mM KI solution (puriss p.a., Fluka, Buchs, Switzerland) was prepared weekly. A phosphate buer (0.5 M) with a pH of 6.1 was prepared from NaH2PO4 and NaOH. Dilution of this buffer with the sample and the other reagents results in a pH of 6.5 in the ®nal solution. Stock solutions of chlorine free hypobromous acid (HOBr) were prepared from a 0.8 mM solution of potassium bromide (KBr) by addition of 1 mM ozone at pH 4 (10 mM phosphate buer). After 24 h, the residual ozone was removed by purging the solution with nitrogen gas for 15 min. The HOBr solution was standardized by direct photometric determination of the hypobromite (OBrÿ) at 329 nm (E=332 Mÿ1 cmÿ1) (Troy and Margerum, 1991) after adjusting the pH to 11 by sodium hydroxide. The yield was typically 95% and the bromine solution was stable for several days when stored at 48C. All three bromamines, monobromamine (NH2Br), dibromamine (NHBr2) and tribromamine (NBr3), were prepared from HOBr and ammonia according to Inman and Johnson (1984) and Galal-Gorchev and Morris (1965). They were checked for their purity by direct measurements of their characteristic UV-spectra. No pure solutions of NHBr2 could be prepared. They always con-
Scheme 1. Chemical structure of 2,2-azino-bis(3-ethylbenzothiazoline)-6-sulfonate (ABTS) and its oxidation product (ABTS +).
Analysis of bromine and chlorine species tained small quantities of NH2Br and NBr3 (approx. 10%). A sodium hypochlorite solution (5% in water, Fluka) was used to prepare stock solutions of aqueous chlorine, which was standardized by UV-spectrometry via the formation of triiodide (Bichsel and von Gunten, 1999). Chlorine dioxide (ClO2) was prepared from K2S2O8 and NaClO2 as described by Gates (1997). Solutions of ClO2 were standardized by direct UV-measurement at 359 nm using a molar absorptivity of 1200 Mÿ1 cmÿ1 (Hoigne and Bader, 1994) and also by formation of triiodide (Bichsel and von Gunten, 1999). The two methods for ClO2 gave similar results (e.g. direct UV: 1.83 mM ClO2; via triiodide: 1.84 mM ClO2). Monochloramine (NH2Cl) was prepared by adding HOCl to NH4NO3 (pH 8) at a molar ratio of 1:5. The reaction is complete within seconds. Such solutions were prepared daily. Apparatus All photometric measurements were made on a spectrophotometer (UVIKON 940, Kontron Instruments) in 1, 5, or 10 cm quartz cells. Kinetic experiments were done in the spectrophotometric cell and absorbance data were automatically recorded every second. Analytical conditions A detection wavelength of 405 nm instead of the near absorption maximum at 415 nm was chosen because the absorbance spectrum of the ABTS + provides a broad shoulder at 405 nm and optical ®lters for portable photometers are obtainable for this wavelength. Table 1 gives an overview of the conditions for the analysis of the bromine and chlorine species as described herein. The following procedures can be simply adapted to dierent analyte concentrations by changing the sample volume, ABTS concentration and the pathlength of the measuring cell. Determination of HOBr with ABTS. The procedure is described for a low concentration range of 1±20 mM HOBr. In a 25 ml volumetric ¯ask, 10 ml of the HOBr sample is added to 1 ml of the ABTS solution (1 g lÿ1) and 1 ml 0.05 M H2SO4 and diluted with H2O. The absorption at 405 nm is measured in a 1 cm cell after 1 min. The HOBr concentration is calculated from the molar absorptivity of the ABTS + E(405 nm)=31,600 Mÿ1 cmÿ1 (Pinkernell et al., 1997). Determination of bromamines with ABTS. Same as for HOBr. The concentration ranges decrease for di- and tribromamine due to their higher degree of bromination. Determination of HOBr as Brÿ 3 . For stoichiometry check the concentration range of 0.07±0.7 mM HOBr, 0.5 ml of the sample is added to 8 ml of 1 M NaBr and 1 ml of 0.05 M H2SO4. After dilution with water to 10 ml, the absorbance of the Brÿ 3 is measured at 260 nm in a 1 cm cell (E=37,200 Mÿ1 cmÿ1; Beckwith and Margerum, 1997). All chlorine analyses. For low chlorine concentrations, 20 ml of sample are added to a 25 ml volumetric ¯ask. The resulting total available chlorine should be less than 20 mM. The absorbance is recorded at 405 or 728 nm,
4345
10 min after addition of ABTS because the reaction is slower than the corresponding with HOBr. Determination of total chlorine. One milliliter ABTS solution (1 g lÿ1), 0.15 ml KI solution (1 mM) and 3 ml phosphate buer (pH 6.1) are added to a 25 ml volumetric ¯ask. The water sample is added and the ¯ask is ®lled with chlorine-free water to 25 ml. 0.05 ml NH4NO3 (0.1 M) can additionally be added to the ¯ask together with ABTS to avoid consumption of free chlorine by natural organic matter (see below). Determination of monochloramine. 0.05 ml NaNO2 (0.1 M) is added to a 25 ml volumetric ¯ask to quench HOCl or ClO2 (Table 1). If the sample has a pH of less than 7, buer has to be added to raise the pH to 8 for nitrite addition. Then, the water sample is added. After 4 min, 1 ml ABTS (1 g lÿ1), 3 ml phosphate buer (pH 6.1) and 0.15 ml KI (1 mM) are added and the ¯ask is ®lled with chlorine-free water to 25 ml. Determination of chlorine dioxide (for explanation see below). 0.2 ml glycine (50 g lÿ1), 0.2 ml HgIICl2 (3 g lÿ1), 3 ml phosphate buer (pH 6.1) and 1 ml ABTS (1 g lÿ1) are added to a 25 ml volumetric ¯ask. Then, the water sample is added and the ¯ask is ®lled with chlorine-free water to 25 ml. The absorbance is read after 2 min. Determination of total chlorine in the presence of ozone. 0.05 ml NH4NO3 (0.1 M) is added to a 25 ml ¯ask before addition of the water sample. After 1 min, 0.05 ml NaNO2 (0.1 M) is added to quench ozone. Again after 1 min, 1 ml ABTS (1 g lÿ1), 3 ml buer (pH 6.1) and 0.15 ml KI (1 mM) are added and the ¯ask is ®lled with water to 25 ml. The blanks (not chlorinated natural water) prepared as described above have absorptions of 0.999 each) was determined. The detection limits for all bromine species were determined as three times the variability of the blank signal using a 1 cm cell and were found to be: HOBr and NH2Br 0.05 mM, NHBr2 0.03 mM and NBr3 0.02 mM. The precision for the analysis of all bromine species was determined by the analysis of 10 samples each and resulted in a standard deviation of 1% for all bromine species. Calibration graphs for HOCl, NH2Cl, and ClO2 are linear in the range 1±10 mM (0.07±0.7 mg lÿ1 chlorine) with correlation coecients of 0.998 or better, using a 1 cm cell and a detection wavelength of 405 nm (Fig. 3). Calibration curve for HOCl in the low concentration range from 0.1 to 1.5 mM (0.007±0.1 mg lÿ1 chlorine) measured at 728 nm in a 5 cm cell also gives a good correlation coecient (r 2 > 0.999). The detection limits for the dierent chlorine species were determined in river water as three times the variability of the blank signal using a 5 cm cell. At 728 nm it was found to be 0.04 mM for HOCl and NH2Cl and 0.1 mM for ClO2. For all natural water analyses, the absorption was measured at 728 nm because the background absorption of the water matrix is much lower than at 405 nm. This results in a lower detection limit despite the smaller value of E. Standard deviation was found to be less than 4% for all chlorine species in pure water (n = 33 samples). For mixtures of HClO, ClO2 and NH2Cl,
Stability of ABTS + The colored ABTS + decays with a ®rst-order kinetic. At room temperature and in pure water, its half-life is approximately 47 h. Therefore, the decay in the absorption is less than 1 and 5% after 30 min and for 5 h of storage, respectively. At 48C, the half-life increases to a value of 357 h. Under these conditions, it can be stored for 24 h with less than 3% of decrease in the absorption. Linear range, detection limits and precision The linear ranges for these analytical methods using ABTS depend on the analytical procedure. Concentrations in the lower micromolar range have been chosen for all analytes to describe their linear ranges. These ranges can be shifted to higher concentrations by decreasing the sample volume and increasing the pathlength of the measuring cell. The linear range for the determination of HOBr in a 1 cm cell was found to be 1±20 mM HOBr with
Fig. 4. Behavior of reactive chlorine species (HOCl, NH2Cl, ClO2) during Limmat river water treatment (Limmatwasserwerk ZuÈrich, Switzerland). A mixture of chlorine/chlorine dioxide is added to the river water. For speci®c conditions see text.
Analysis of bromine and chlorine species
the detection limit and the precision for the determination of HOCl calculated as the dierence between the total active chlorine and the sum of ClO2 and NH2Cl will depend on the concentrations of each compound. In this case, precision can be signi®cantly lower than for pure solution of HOCl. Application to drinking water samples The fate of reactive chlorine species at a drinking water treatment plant in ZuÈrich using chlorine/ chlorine dioxide as initial disinfectant has been determined using the new method. A mixture of chlorine/chlorine dioxide (4.3 mM (0.31 mg lÿ1)/ 1.3 mM (0.088 mg lÿ1)) is added to the river water (pH 8, 2 mM bicarbonate, 1.5 mg/l DOC, 15 mg lÿ1 ammonia (01 mM), 0.2 mM Brÿ), followed by sedimentation, ozonation, and sand- and activated carbon ®ltration. Finally, a mixture of chlorine/ chlorine dioxide is added for disinfection again (2.1 mM (0.149 mg lÿ1)/0.7 mM (0.047 mg lÿ1)). Figure 4 shows a typical development of the concentrations of HOCl, ClO2 and NH2Cl through the whole process. A rapid consumption of chlorine and chlorine dioxide occurs between the addition and the ®rst sampling point (t = 8 min) with the concurrent formation of NH2Cl. After sedimentation (t = 3 h) HOCl and ClO2 have almost entirely reacted. At this point of the treatment, NH2Cl is still present at a concentration of about 1 mM, which corresponds to the concentration of ammonia in the river water. During ozonation (HRT=30 min) and the subsequent activated carbon ®ltration, NH2Cl disappears completely. The ®nal disinfection results again in the formation of a low concentration of NH2Cl. Despite the small amount of added disinfectants (0.4 mg as Cl2), their concentration can be easily followed by the new method. Compared to the added HOCl the Brÿ concentration in natural waters is low. Once HOBr is formed from HOCl and Brÿ, it is consumed much more rapidly by natural organic matter compared to chlorine. Therefore, the overestimation of free chlorine by the cross-reactivity towards HOBr is negligible.
CONCLUSIONS
The reaction of ABTS with BrI and ClI is strictly stoichiometric, so the molar absorptivity E can be used for the calculation of the concentration and a calibration is not necessary. The advantage of using ABTS over DPD consists in the much higher stability of the color that is formed. With ABTS, the dierent oxidant species can be ®xed in the ®eld and measured after transfer to the laboratory. The determination of hypobromous acid and all three bromamines using ABTS was found to be re-
4349
liable and accurate for concentrations down to 0.1 mM. The selectivity of the reaction of ABTS towards ClO2 and the other chlorine species allows a very simple determination of ClO2. The detection limits for the chlorine species are 0.07 mM for NH2Cl and HOCl (0.005 mg lÿ1 Cl2) and 0.1 mM for ClO2 (0.007 mg lÿ1). The main disadvantage of the method is that free chlorine cannot be measured independently. No conditions could be found where HOCl reacts with ABTS but NH2Cl did not. Therefore, free chlorine can only be determined by dierence. Moreover, no distinction is possible between HOCl and HOBr. AcknowledgementsÐU.P. thanks the Compagnie GeÂneÂrale des Eaux, Paris, and B.N. and H.G. thank the Waterworks ZuÈrich for ®nancial support. REFERENCES
Aieta E. M., Roberts P. V. and Hernandez M. J. (1984) Determination of chlorine dioxide, chlorine, chlorite, and chlorate in water. Am. Water Works Ass. 76, 64±70. APHA American Water Works Association Water Pollution Control Federation (1989) Standard Methods for the Examination of Water and Wastewater, 17th ed. American Public Health Association, Washington, DC. Beckwith R. C. and Margerum D. W. (1997) Kinetics of hypobromous acid disproportionation. Inorg. Chem. 36, 3754±3760. Bergmeyer H. U. (1986) Methods of Enzymatic Analysis I, 3rd Edition, pp. 210±217, Verlag Chemie, Weinheim. Bichsel Y. and von Gunten U. (1999) Determination of iodide and iodate by ion chromatography with postcolumn reaction and UV/visible detection. Anal. Chem. 71, 34±38. Emerson D. W. (1994) Microdetermination of bromine, chlorine, and chlorine dioxide in water in any combination. Microchem. J. 50, 116±124. Fisher D. J., Burton D. T., Yonkos L. T., Turley S. D. and Ziegler G. P. (1999) The relative acute toxicity of continuous and intermittent exposures of chlorine and bromine to aquatic organisms in the presence and absence of ammonia. Water Res. 33, 760±768. Floyd R., Sharp D. G. and Johnson J. D. (1978) Inactivation of single poliovirus particles in water by hypobromite ion, molecular bromine, dibromine, and tribromine. Environ. Sci. Technol. 12, 1031±1035. Galal-Gorchev H. and Morris J. C. (1965) Formation and stability of bromamide, bromimide, and nitrogen tribromide in aqueous solution. Inorg. Chem. 4, 899±905. Gates D. J. (1997) The Chlorine Dioxide Handbook. American Water Works Association, Denver, CO. Hofmann R., Andrews R. C. and Ye Q. (1998) Comparison of spectrophotometric methods for measuring chlorine dioxide in drinking water. Environ. Technol. 19, 761± 773. Hoigne J. and Bader H. (1994) Kinetics of reactions of chlorine dioxide (OClO) in water. 1. Rate constants for inorganic and organic compounds. Water Res. 28, 45± 55. Inman G. W. and Johnson J. D. (1984) Kinetics of monobromamine disproportionationÐdibromamine formation in aqueous ammonia solutions. Environ. Sci. Technol. 18, 219±224. Jandik J. and EichelsdoÈrfer D. (1980) Anmerkungen zur gemeinsamen Bestimmung von Chlor und Ozon in Schwimmbeckenwasser nach der DPD-Methode von
4350
Ulrich Pinkernell et al.
Palin; comments on the determination of chlorine and ozone in swimming pool water using the DPD method of Palin. Archiv des Badwesens 80, 90±91. Jensen J. N. and Johnson J. D. (1990) Interferences by monochloramine and organic chloramines in free available chlorine methods. 1. Amperometric-titration. Environ. Sci. Technol. 24, 981±985. Katz A., Narkis N., Orshansky F., Friedland E. and Kott Y. (1994) Disinfection of euent by combinations of equal doses of chlorine dioxide and chlorine added simultaneously over varying contact times. Water Res. 28, 2133±2138. MacCrehan W. A., Jensen J. S. and Helz G. R. (1998) Detection of sewage organic chlorination products that are resistant to dechlorination with sul®te. Environ. Sci. Technol. 32, 3640±3645. Palin A. T. (1975) Current DPD methods for residual halogen compounds in water. J. Am. Water Works Ass. 67, 32±33. Pinkernell U., LuÈke H.-J. and Karst U. (1997) Selective photometric determination of peroxycarboxylic acids in the presence of hydrogen peroxide. Analyst 122, 567± 571. Pinkernell U. and von Gunten U. (1999) Control options
for bromate minimization during ozonation processes: kinetically based approach. In Proceedings of the 14th Ozone World Congress, Dearborn, MI. International Ozone Association, Stamford, CT, pp. 441±450. Qualls R. G. and Johnson J. D. (1983) Kinetics of the short-term consumption of chlorine by fulvic-acid. Environ. Sci. Technol. 17, 692±698. Scott S. L., Chen W. J., Bakac A. and Espenson J. H. (1993) Spectroscopic parameters, electrode-potentials, acid ionization-constants, and electron-exchange rates of the 2,2'-azinobis(3-ethylbenzothiazoline-6-sulfonate) radicals and ions. J. Phys. Chem. 97, 6710±6714. Sweetin D. L., Sullivan E. and Gordon G. (1996) The use of chlorophenol red for the selective determination of chlorine dioxide in drinking water. Talanta 43, 103±108. Troy R. C. and Margerum D. W. (1991) Nonmetal redox kineticsÐhypobromite and hypobromous acid reactions with iodide and with sul®te and the hydrolysis of bromosulfate. Inorg. Chem. 30, 3538±3543. von Gunten U. and Hoigne J. (1994) Bromate formation during ozonation of bromide-containing waters: interaction of ozone and hydroxyl radical reactions. Environ. Sci. Technol. 28, 1234±1242.
Wat. Res. Vol. 34, No. 18, pp. 4343±4350, 2000 7 2000 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0043-1354/00/$ - see front matter
www.elsevier.com/locate/watres
METHODS FOR THE PHOTOMETRIC DETERMINATION OF REACTIVE BROMINE AND CHLORINE SPECIES WITH ABTS ULRICH PINKERNELL, BERND NOWACK*, HERVEÂ GALLARD and URS VON GUNTEN$ Swiss Federal Institute for Environmental Science and Technology (EAWAG), Ueberlandstr. 133, CH-8600, DuÈbendorf, Switzerland (First received 15 October 1999; accepted 10 March 2000) AbstractÐNew methods for the determination of reactive bromine and chlorine species are presented. Hypobromous acid (HOBr) and all three bromamines species (NH2Br, NHBr2, NBr3) are analyzed as a sum parameter and hypochlorous acid (HOCl), monochloramine (NH2Cl) and chlorine dioxide (ClO2) can be determined selectively. However, no distinction is possible between HOCl and the active bromine species. The bromine and chlorine species react with ABTS (2,2-azino-bis(3ethylbenzothiazoline)-6-sulfonic acid-diammonium salt) to a green colored product that is measured at 405 or 728 nm. Free chlorine and NH2Cl can be measured in the presence of ozone. The method is therefore suitable if combinations of disinfectants are used, such as chlorine/chlorine dioxide or chlorine/ozone. In natural waters, the method provides a detection limit for all chlorine/bromine species of less than 0.1 mM. The colored reaction product is very stable and allows a ®xation of the chlorine/ bromine species in the ®eld and subsequent determination of the absorption in the laboratory. 7 2000 Elsevier Science Ltd. All rights reserved Key wordsÐhypobromous acid, bromamines, chlorine, chlorine dioxide, monochloramine, drinking water, analysis
INTRODUCTION
tion by-products when ozone or chlorine is applied to treat bromide-containing natural waters. It has been shown in studies on bromate formation during ozonation of bromide-containing natural waters that hypobromous acid is a key intermediate (von Gunten and HoigneÂ, 1994). In the presence of ammonia HOBr is scavenged in a fast reaction forming bromamine (NH2Br) and dibromamine (NHBr2) (Inman and Johnson, 1984), which are not oxidized to bromate directly. Therefore, ammonia addition may be one possible control option for bromate minimization (Pinkernell and von Gunten, 1999). To get an overall bromine mass balance during water treatment, it is therefore necessary to assess the fate of these reactive bromamines in addition to HOBr.
Bromine species: Reactive bromine species, such as hypobromous acid (HOBr) and bromamines (NH2Br, NHBr2, NBr3) may occur due to two main reasons during water disinfection: (1) they are used directly as disinfectants; and (2) they are disinfection by-products during ozonation and chlorination of bromide-containing waters: 1. HOBr is used for disinfection of swimming pool water and for control of biofouling (Fisher et al., 1999). Due to the presence of ammonia in these applications, bromamines are formed immediately, which have comparable disinfection properties as HOBr (Floyd et al., 1978). All these active bromine species show a higher toxicity to biofouling organisms as their corresponding chlorine analogues (Fisher et al., 1999). 2. HOBr and bromamines are formed as disinfecPresent address: Institute of Terrestrial Ecology (ITOÈ), Swiss Federal Institute of Technology (ETH), Grabenstr. 11a, CH-8952 Schlieren, Switzerland. $Author to whom all correspondence should be addressed. Tel.: +41-1-823-5270; fax: +41-1-823-5028; e-mail: [email protected] *
HOBr and bromamines can be determined by a DPD (diphenylendiamine)-based method. However, the stability of the colored reaction product can be expected to be low as described for the chlorine analysis (Jandik and EichelsdoÈrfer, 1980). In addition, this method requires internal calibration using standard solutions of HOBr and bromamines.
4343
4344
Ulrich Pinkernell et al.
This is dicult to carry out due to the low stability of these reactive compounds. Chlorine species: Chlorine, consisting of hypochlorous acid (HOCl) and of hypochlorite ion (ClOÿ) at neutral pH, and chlorine dioxide are used worldwide in large quantities for disinfection. Therefore, numerous analytical techniques are available for their determination (APHA, 1989). The need for another analytical method to determine one of the disinfectants alone is therefore not urgent. However, chlorine and chlorine dioxide are also increasingly used as a mixture in water treatment to minimize the formation of by-products (Katz et al., 1994). The exact analysis of these mixtures is troublesome. The standard procedure for the distinction of the two oxidants consists in the addition of glycine to the sample, which masks chlorine by formation of chloroaminoacetic acid (APHA, 1989). Chlorine dioxide can then be measured selectively by a colorimetric method and the chlorine concentration is calculated by dierence of the total oxidant concentration and chlorine dioxide. Chlorite, which is always present when chlorine dioxide is added to water, is also measured and further complicates the analysis (APHA, 1989). A method to distinguish between bromine or chlorine and chlorine dioxide has also been described (Emerson, 1994). However, this method has a detection limit of about 0.5 mg lÿ1 Cl2, which is not suf®ciently sensitive for the analysis of drinking water samples. The method also uses the addition of glycine to suppress free chlorine. It has been reported, however, that glycine interferes with the ClO2 measurement in the DPD method (Jandik and EichelsdoÈrfer, 1980; Palin, 1975). Alternatively, chlorine dioxide can be purged with nitrogen from the water (Aieta et al., 1984). This method is very time-consuming because every sample has to be purged for more than 30 min. Photometric methods for the selective determination of chlorine dioxide with no interference from chlorine have been described (Hofmann et al., 1998; Sweetin et al., 1996). It is, however, not possible to also analyze chlorine using one of these methods. The need for a simple method to distinguish between chlorine and chlorine dioxide at low concentrations is therefore given. A method that uses only one reagent produces a stable color and allows the distinction between chlorine, chloramine and chlorine dioxide without interference of chlorite, should be found. The same need is given for reactive bromine species. As there are no applications where signi®cant concentrations of bromine and chlorine species are present in the water treatment ®eld, a distinction between chlorine and bromine is not necessary. We have found that ABTS (2,2-azino-bis(3-ethylbenzothiazoline)-6-sulfonic acid-diammonium salt), a color reagent, ful®lls these requirements. ABTS is a well-known substrate for enzymatic peroxide tests
(Bergmeyer, 1986) and percarboxylic acid analysis (Pinkernell et al., 1997). The colorless ABTS is oxidized undergoing a one-electron transfer as depicted in Scheme 1. The reaction product ABTS + is a stable green colored radical. It has a broad absorbance spectrum containing several maxima with high molar absorptivities at 415, 650, 728, and 815 nm (Scott et al., 1993). It has been shown before that because the reaction with ABTS is fast and has a known stoichiometry, a direct calculation of the analyte using the molar absorptivity of the colored oxidation product is possible (Pinkernell et al., 1997). EXPERIMENTAL SECTION
Materials All water used was from a Barnstead B-Pure system. All chemicals have been of p.a. quality (Fluka) if not stated otherwise. ABTS (2,2-azino-bis(3-ethylbenzothiazoline)-6sulfonic acid-diammonium salt, Aldrich) stock solutions (1 g/l) were used for all experiments and stored at 48C. After a few weeks the ABTS solution had a slightly higher blank absorption and was replaced. A 1 mM KI solution (puriss p.a., Fluka, Buchs, Switzerland) was prepared weekly. A phosphate buer (0.5 M) with a pH of 6.1 was prepared from NaH2PO4 and NaOH. Dilution of this buffer with the sample and the other reagents results in a pH of 6.5 in the ®nal solution. Stock solutions of chlorine free hypobromous acid (HOBr) were prepared from a 0.8 mM solution of potassium bromide (KBr) by addition of 1 mM ozone at pH 4 (10 mM phosphate buer). After 24 h, the residual ozone was removed by purging the solution with nitrogen gas for 15 min. The HOBr solution was standardized by direct photometric determination of the hypobromite (OBrÿ) at 329 nm (E=332 Mÿ1 cmÿ1) (Troy and Margerum, 1991) after adjusting the pH to 11 by sodium hydroxide. The yield was typically 95% and the bromine solution was stable for several days when stored at 48C. All three bromamines, monobromamine (NH2Br), dibromamine (NHBr2) and tribromamine (NBr3), were prepared from HOBr and ammonia according to Inman and Johnson (1984) and Galal-Gorchev and Morris (1965). They were checked for their purity by direct measurements of their characteristic UV-spectra. No pure solutions of NHBr2 could be prepared. They always con-
Scheme 1. Chemical structure of 2,2-azino-bis(3-ethylbenzothiazoline)-6-sulfonate (ABTS) and its oxidation product (ABTS +).
Analysis of bromine and chlorine species tained small quantities of NH2Br and NBr3 (approx. 10%). A sodium hypochlorite solution (5% in water, Fluka) was used to prepare stock solutions of aqueous chlorine, which was standardized by UV-spectrometry via the formation of triiodide (Bichsel and von Gunten, 1999). Chlorine dioxide (ClO2) was prepared from K2S2O8 and NaClO2 as described by Gates (1997). Solutions of ClO2 were standardized by direct UV-measurement at 359 nm using a molar absorptivity of 1200 Mÿ1 cmÿ1 (Hoigne and Bader, 1994) and also by formation of triiodide (Bichsel and von Gunten, 1999). The two methods for ClO2 gave similar results (e.g. direct UV: 1.83 mM ClO2; via triiodide: 1.84 mM ClO2). Monochloramine (NH2Cl) was prepared by adding HOCl to NH4NO3 (pH 8) at a molar ratio of 1:5. The reaction is complete within seconds. Such solutions were prepared daily. Apparatus All photometric measurements were made on a spectrophotometer (UVIKON 940, Kontron Instruments) in 1, 5, or 10 cm quartz cells. Kinetic experiments were done in the spectrophotometric cell and absorbance data were automatically recorded every second. Analytical conditions A detection wavelength of 405 nm instead of the near absorption maximum at 415 nm was chosen because the absorbance spectrum of the ABTS + provides a broad shoulder at 405 nm and optical ®lters for portable photometers are obtainable for this wavelength. Table 1 gives an overview of the conditions for the analysis of the bromine and chlorine species as described herein. The following procedures can be simply adapted to dierent analyte concentrations by changing the sample volume, ABTS concentration and the pathlength of the measuring cell. Determination of HOBr with ABTS. The procedure is described for a low concentration range of 1±20 mM HOBr. In a 25 ml volumetric ¯ask, 10 ml of the HOBr sample is added to 1 ml of the ABTS solution (1 g lÿ1) and 1 ml 0.05 M H2SO4 and diluted with H2O. The absorption at 405 nm is measured in a 1 cm cell after 1 min. The HOBr concentration is calculated from the molar absorptivity of the ABTS + E(405 nm)=31,600 Mÿ1 cmÿ1 (Pinkernell et al., 1997). Determination of bromamines with ABTS. Same as for HOBr. The concentration ranges decrease for di- and tribromamine due to their higher degree of bromination. Determination of HOBr as Brÿ 3 . For stoichiometry check the concentration range of 0.07±0.7 mM HOBr, 0.5 ml of the sample is added to 8 ml of 1 M NaBr and 1 ml of 0.05 M H2SO4. After dilution with water to 10 ml, the absorbance of the Brÿ 3 is measured at 260 nm in a 1 cm cell (E=37,200 Mÿ1 cmÿ1; Beckwith and Margerum, 1997). All chlorine analyses. For low chlorine concentrations, 20 ml of sample are added to a 25 ml volumetric ¯ask. The resulting total available chlorine should be less than 20 mM. The absorbance is recorded at 405 or 728 nm,
4345
10 min after addition of ABTS because the reaction is slower than the corresponding with HOBr. Determination of total chlorine. One milliliter ABTS solution (1 g lÿ1), 0.15 ml KI solution (1 mM) and 3 ml phosphate buer (pH 6.1) are added to a 25 ml volumetric ¯ask. The water sample is added and the ¯ask is ®lled with chlorine-free water to 25 ml. 0.05 ml NH4NO3 (0.1 M) can additionally be added to the ¯ask together with ABTS to avoid consumption of free chlorine by natural organic matter (see below). Determination of monochloramine. 0.05 ml NaNO2 (0.1 M) is added to a 25 ml volumetric ¯ask to quench HOCl or ClO2 (Table 1). If the sample has a pH of less than 7, buer has to be added to raise the pH to 8 for nitrite addition. Then, the water sample is added. After 4 min, 1 ml ABTS (1 g lÿ1), 3 ml phosphate buer (pH 6.1) and 0.15 ml KI (1 mM) are added and the ¯ask is ®lled with chlorine-free water to 25 ml. Determination of chlorine dioxide (for explanation see below). 0.2 ml glycine (50 g lÿ1), 0.2 ml HgIICl2 (3 g lÿ1), 3 ml phosphate buer (pH 6.1) and 1 ml ABTS (1 g lÿ1) are added to a 25 ml volumetric ¯ask. Then, the water sample is added and the ¯ask is ®lled with chlorine-free water to 25 ml. The absorbance is read after 2 min. Determination of total chlorine in the presence of ozone. 0.05 ml NH4NO3 (0.1 M) is added to a 25 ml ¯ask before addition of the water sample. After 1 min, 0.05 ml NaNO2 (0.1 M) is added to quench ozone. Again after 1 min, 1 ml ABTS (1 g lÿ1), 3 ml buer (pH 6.1) and 0.15 ml KI (1 mM) are added and the ¯ask is ®lled with water to 25 ml. The blanks (not chlorinated natural water) prepared as described above have absorptions of 0.999 each) was determined. The detection limits for all bromine species were determined as three times the variability of the blank signal using a 1 cm cell and were found to be: HOBr and NH2Br 0.05 mM, NHBr2 0.03 mM and NBr3 0.02 mM. The precision for the analysis of all bromine species was determined by the analysis of 10 samples each and resulted in a standard deviation of 1% for all bromine species. Calibration graphs for HOCl, NH2Cl, and ClO2 are linear in the range 1±10 mM (0.07±0.7 mg lÿ1 chlorine) with correlation coecients of 0.998 or better, using a 1 cm cell and a detection wavelength of 405 nm (Fig. 3). Calibration curve for HOCl in the low concentration range from 0.1 to 1.5 mM (0.007±0.1 mg lÿ1 chlorine) measured at 728 nm in a 5 cm cell also gives a good correlation coecient (r 2 > 0.999). The detection limits for the dierent chlorine species were determined in river water as three times the variability of the blank signal using a 5 cm cell. At 728 nm it was found to be 0.04 mM for HOCl and NH2Cl and 0.1 mM for ClO2. For all natural water analyses, the absorption was measured at 728 nm because the background absorption of the water matrix is much lower than at 405 nm. This results in a lower detection limit despite the smaller value of E. Standard deviation was found to be less than 4% for all chlorine species in pure water (n = 33 samples). For mixtures of HClO, ClO2 and NH2Cl,
Stability of ABTS + The colored ABTS + decays with a ®rst-order kinetic. At room temperature and in pure water, its half-life is approximately 47 h. Therefore, the decay in the absorption is less than 1 and 5% after 30 min and for 5 h of storage, respectively. At 48C, the half-life increases to a value of 357 h. Under these conditions, it can be stored for 24 h with less than 3% of decrease in the absorption. Linear range, detection limits and precision The linear ranges for these analytical methods using ABTS depend on the analytical procedure. Concentrations in the lower micromolar range have been chosen for all analytes to describe their linear ranges. These ranges can be shifted to higher concentrations by decreasing the sample volume and increasing the pathlength of the measuring cell. The linear range for the determination of HOBr in a 1 cm cell was found to be 1±20 mM HOBr with
Fig. 4. Behavior of reactive chlorine species (HOCl, NH2Cl, ClO2) during Limmat river water treatment (Limmatwasserwerk ZuÈrich, Switzerland). A mixture of chlorine/chlorine dioxide is added to the river water. For speci®c conditions see text.
Analysis of bromine and chlorine species
the detection limit and the precision for the determination of HOCl calculated as the dierence between the total active chlorine and the sum of ClO2 and NH2Cl will depend on the concentrations of each compound. In this case, precision can be signi®cantly lower than for pure solution of HOCl. Application to drinking water samples The fate of reactive chlorine species at a drinking water treatment plant in ZuÈrich using chlorine/ chlorine dioxide as initial disinfectant has been determined using the new method. A mixture of chlorine/chlorine dioxide (4.3 mM (0.31 mg lÿ1)/ 1.3 mM (0.088 mg lÿ1)) is added to the river water (pH 8, 2 mM bicarbonate, 1.5 mg/l DOC, 15 mg lÿ1 ammonia (01 mM), 0.2 mM Brÿ), followed by sedimentation, ozonation, and sand- and activated carbon ®ltration. Finally, a mixture of chlorine/ chlorine dioxide is added for disinfection again (2.1 mM (0.149 mg lÿ1)/0.7 mM (0.047 mg lÿ1)). Figure 4 shows a typical development of the concentrations of HOCl, ClO2 and NH2Cl through the whole process. A rapid consumption of chlorine and chlorine dioxide occurs between the addition and the ®rst sampling point (t = 8 min) with the concurrent formation of NH2Cl. After sedimentation (t = 3 h) HOCl and ClO2 have almost entirely reacted. At this point of the treatment, NH2Cl is still present at a concentration of about 1 mM, which corresponds to the concentration of ammonia in the river water. During ozonation (HRT=30 min) and the subsequent activated carbon ®ltration, NH2Cl disappears completely. The ®nal disinfection results again in the formation of a low concentration of NH2Cl. Despite the small amount of added disinfectants (0.4 mg as Cl2), their concentration can be easily followed by the new method. Compared to the added HOCl the Brÿ concentration in natural waters is low. Once HOBr is formed from HOCl and Brÿ, it is consumed much more rapidly by natural organic matter compared to chlorine. Therefore, the overestimation of free chlorine by the cross-reactivity towards HOBr is negligible.
CONCLUSIONS
The reaction of ABTS with BrI and ClI is strictly stoichiometric, so the molar absorptivity E can be used for the calculation of the concentration and a calibration is not necessary. The advantage of using ABTS over DPD consists in the much higher stability of the color that is formed. With ABTS, the dierent oxidant species can be ®xed in the ®eld and measured after transfer to the laboratory. The determination of hypobromous acid and all three bromamines using ABTS was found to be re-
4349
liable and accurate for concentrations down to 0.1 mM. The selectivity of the reaction of ABTS towards ClO2 and the other chlorine species allows a very simple determination of ClO2. The detection limits for the chlorine species are 0.07 mM for NH2Cl and HOCl (0.005 mg lÿ1 Cl2) and 0.1 mM for ClO2 (0.007 mg lÿ1). The main disadvantage of the method is that free chlorine cannot be measured independently. No conditions could be found where HOCl reacts with ABTS but NH2Cl did not. Therefore, free chlorine can only be determined by dierence. Moreover, no distinction is possible between HOCl and HOBr. AcknowledgementsÐU.P. thanks the Compagnie GeÂneÂrale des Eaux, Paris, and B.N. and H.G. thank the Waterworks ZuÈrich for ®nancial support. REFERENCES
Aieta E. M., Roberts P. V. and Hernandez M. J. (1984) Determination of chlorine dioxide, chlorine, chlorite, and chlorate in water. Am. Water Works Ass. 76, 64±70. APHA American Water Works Association Water Pollution Control Federation (1989) Standard Methods for the Examination of Water and Wastewater, 17th ed. American Public Health Association, Washington, DC. Beckwith R. C. and Margerum D. W. (1997) Kinetics of hypobromous acid disproportionation. Inorg. Chem. 36, 3754±3760. Bergmeyer H. U. (1986) Methods of Enzymatic Analysis I, 3rd Edition, pp. 210±217, Verlag Chemie, Weinheim. Bichsel Y. and von Gunten U. (1999) Determination of iodide and iodate by ion chromatography with postcolumn reaction and UV/visible detection. Anal. Chem. 71, 34±38. Emerson D. W. (1994) Microdetermination of bromine, chlorine, and chlorine dioxide in water in any combination. Microchem. J. 50, 116±124. Fisher D. J., Burton D. T., Yonkos L. T., Turley S. D. and Ziegler G. P. (1999) The relative acute toxicity of continuous and intermittent exposures of chlorine and bromine to aquatic organisms in the presence and absence of ammonia. Water Res. 33, 760±768. Floyd R., Sharp D. G. and Johnson J. D. (1978) Inactivation of single poliovirus particles in water by hypobromite ion, molecular bromine, dibromine, and tribromine. Environ. Sci. Technol. 12, 1031±1035. Galal-Gorchev H. and Morris J. C. (1965) Formation and stability of bromamide, bromimide, and nitrogen tribromide in aqueous solution. Inorg. Chem. 4, 899±905. Gates D. J. (1997) The Chlorine Dioxide Handbook. American Water Works Association, Denver, CO. Hofmann R., Andrews R. C. and Ye Q. (1998) Comparison of spectrophotometric methods for measuring chlorine dioxide in drinking water. Environ. Technol. 19, 761± 773. Hoigne J. and Bader H. (1994) Kinetics of reactions of chlorine dioxide (OClO) in water. 1. Rate constants for inorganic and organic compounds. Water Res. 28, 45± 55. Inman G. W. and Johnson J. D. (1984) Kinetics of monobromamine disproportionationÐdibromamine formation in aqueous ammonia solutions. Environ. Sci. Technol. 18, 219±224. Jandik J. and EichelsdoÈrfer D. (1980) Anmerkungen zur gemeinsamen Bestimmung von Chlor und Ozon in Schwimmbeckenwasser nach der DPD-Methode von
4350
Ulrich Pinkernell et al.
Palin; comments on the determination of chlorine and ozone in swimming pool water using the DPD method of Palin. Archiv des Badwesens 80, 90±91. Jensen J. N. and Johnson J. D. (1990) Interferences by monochloramine and organic chloramines in free available chlorine methods. 1. Amperometric-titration. Environ. Sci. Technol. 24, 981±985. Katz A., Narkis N., Orshansky F., Friedland E. and Kott Y. (1994) Disinfection of euent by combinations of equal doses of chlorine dioxide and chlorine added simultaneously over varying contact times. Water Res. 28, 2133±2138. MacCrehan W. A., Jensen J. S. and Helz G. R. (1998) Detection of sewage organic chlorination products that are resistant to dechlorination with sul®te. Environ. Sci. Technol. 32, 3640±3645. Palin A. T. (1975) Current DPD methods for residual halogen compounds in water. J. Am. Water Works Ass. 67, 32±33. Pinkernell U., LuÈke H.-J. and Karst U. (1997) Selective photometric determination of peroxycarboxylic acids in the presence of hydrogen peroxide. Analyst 122, 567± 571. Pinkernell U. and von Gunten U. (1999) Control options
for bromate minimization during ozonation processes: kinetically based approach. In Proceedings of the 14th Ozone World Congress, Dearborn, MI. International Ozone Association, Stamford, CT, pp. 441±450. Qualls R. G. and Johnson J. D. (1983) Kinetics of the short-term consumption of chlorine by fulvic-acid. Environ. Sci. Technol. 17, 692±698. Scott S. L., Chen W. J., Bakac A. and Espenson J. H. (1993) Spectroscopic parameters, electrode-potentials, acid ionization-constants, and electron-exchange rates of the 2,2'-azinobis(3-ethylbenzothiazoline-6-sulfonate) radicals and ions. J. Phys. Chem. 97, 6710±6714. Sweetin D. L., Sullivan E. and Gordon G. (1996) The use of chlorophenol red for the selective determination of chlorine dioxide in drinking water. Talanta 43, 103±108. Troy R. C. and Margerum D. W. (1991) Nonmetal redox kineticsÐhypobromite and hypobromous acid reactions with iodide and with sul®te and the hydrolysis of bromosulfate. Inorg. Chem. 30, 3538±3543. von Gunten U. and Hoigne J. (1994) Bromate formation during ozonation of bromide-containing waters: interaction of ozone and hydroxyl radical reactions. Environ. Sci. Technol. 28, 1234±1242.
