Anal. Chem. 1989, 6 1 , 2785-2791
2785
Determination of Chloride and Available Chlorine in Aqueous Samples by Flame Infrared Emission S . W. Kubala, D. C. Tilotta, M. A. Busch, and K. W. Busch* Department of Chemistry, Baylor University, Waco, Texas 76 798
A specially designed system, using a flame infrared emission (FIRE) detector, was developed to permit the determination of chloride (Ci-) in water and available chlorine (Ci,, HOCI, and O W ) in liquid bleach. Chloride ion was converted to molecular chlorine (Ci,) by addition of concentrated sulfuric acid and a saturated solution of potassium permanganate. Bleach samples were treated with sulfuric acid to convert ail hypochlorite and hypochlorous acid to Ci,. After treatment, molecular chlorine was purged from solution with He and introduced into a hydrogen/entrained-air flame. A portion of the HCI emission intensity at 3150-2425 cm-' was monitored by using a lead selenide detector in conjunction wlth a 3.8-pm optical band-pass filter. Peak emission intensity measurements from repeated injectlons of aqueous NaCi standards gave a relative standard deviation of 3.34%, and calibration curves were linear up to the maximum concentration of chloride investigated (10 mM NaCi). The average relative standard deviations of the chloride and available chlorine determinations measured with the FIRE system were found to be 4.39% and 1.84%, respectively. The accuracy of the FIRE technique was determined by comparing the available chlorine resuits wlth those obtained by iodometric titration for three commercial bleach samples and was found to be 2.97% . The detection limit for chloride and available chlorine was found to be 1.59 ppm (4.49X lo-' mM Ci-) and 1.61 ppm (2.27X lo-' mM Ci,), respectively. Elevated levels of bromlde produced a negative interference in the determination of chloride with the FIRE system, but iodide and phosphate did not Interfere.
INTRODUCTION The element chlorine is widely distributed in nature and is used extensively in its various oxidation states (1). Aqueous molecular chlorine ((21,) and hypochlorite (OCl-), for example, are employed as bleaching agents (2) and as disinfectants to prevent the spread of waterborne diseases (3). Because of the reactivity of the higher oxidation states of chlorine, the element occurs in nature primarily as the chloride ion (C1-) and is one of the major inorganic constituents of surface waters, groundwaters, and wastewaters. In seawaters, chloride levels, expressed as chlorinity, are approximately related to salinity and can be used to determine the concentrations of all other biounlimited elements present in a sample ( 4 ) . Chloride concentration is also used as an indicator of water condition. For example, elevated chloride concentrations in the sewerage of coastal areas may signal seawater intrusion into the system, while in potable water they are often associated with wastewater contamination ( 5 ) . In process waters, chloride concentrations are monitored regularly since elevated levels are generally associated with increased deterioration of metallic pipes and structures (5),while in cooling water, they are used to indicate the cycles of concentration (6). Because of the widespread use and occurrence of the many forms of chlorine, analytical methods for their determination
are of great importance. A large number of methods exist for the determination of chloride ion and chlorine in aqueous samples. These include chromatographic (7-11), spectrometric (12-21), potentiometric (22-28), and titrimetric (29,30) procedures, with the most widely used methods involving titration of the sample. For the titrimetric determination of aqueous chloride, various argentometric methods exist, which use either indicators (29-31) or potentiometers (27,28) to detect the endpoint. Alternatively, mercuric nitrate can be used to titrate chloride ion with the use of diphenylcarbazone as an indicator (5). For the determination of chlorine in bleach bath liquors and natural and treated waters in concentrations greater than 1 mg/L ( 5 ) ,iodometric titration is the method of choice. For chlorine levels less than this amount, amperometric titrations are preferred ( 5 ) ,but require greater operator skill to avoid loss of chlorine through mechanical stirring. Poor endpoints are also a problem unless the electrodes are properly cleaned and conditioned. Alternatively, N,N-diethyl-p-phenylenediamine (DPD) can be used to determine dissolved chlorine colorimetrically or titrimetrically with ferrous ammonium sulfate ( 5 ) . All of these titrimetric methods suffer severe interference by a variety of species including such anions as bromide, iodide, cyanide, sulfide, and orthophosphate (for chloride) and other oxidizing agents (for chlorine). Because of the problems associated with existing procedures, new analytical methods for the determination of chloride ion and chlorine in aqueous samples could be of great importance in a number of disciplines. The potential of flame infrared emission (FIRE) spectroscopy as a means of quantitative analysis has only recently become apparent (32-35). Previous work in this laboratory has shown that the discrete infrared emission band a t 4.42 wm (2262 cm-') produced by vibrationally excited carbon dioxide is useful analytically as a means of detecting organic materials introduced into the flame (33,34)and also for determining total inorganic carbon (TIC) in aqueous samples (35). Spectroscopic studies in this laboratory using Fourier transform techniques have also demonstrated that other discrete infrared emission bands can be observed in a hydrogen-air flame due to the presence of vibrationally excited molecules (32). For example, combustion of chlorine-containing compounds in the flame produces HCl, and under high-resolution conditions, the P and R branches of the HC1 infrared emission can be easily detected above the flame background in the region 3200-2400 cm-l. Since the HC1 emission band lies between the water emission band a t 3800-3200 cm-' and the carbon dioxide emission band centered a t 2262 cm-' (32), the strong, well-resolved infrared emission from HCl should also be useful analytically for the determination of C1 in a variety of chlorine-containing samples. Since chlorine gas reacts rapidly with hydrogen under flame conditions to form HCl Hz C1p 2HC1 (1) any reaction that generates molecular chlorine in a quanti-
+
0003-2700/89/0361-2785$01.50/0 0 1989 American Chemical Society
+
2786
ANALYTICAL CHEMISTRY. VOL. 61. NO. 24. DECEMBER 15. 1989
tative manner could serve as the basis of an analytical procedure emdovine FIRE as a hiehlv snecific means of detection. As one example, samples c o n k k n g dissolved chloride ion could he oxidized to molecular chlorine according to the following half-cell (Eo = -1.36 V)
2Cl-
-
Cl,(aq)
+ 2e-
(2)
using such strong oxidants as permanganate ion (Eo= +1.51 V in acid solution), peroxide ion (E? = +1.77 V in acid solution), or peroxydisulfate ion (Eo = +2.01 V). The resulting chlorine gas could then be purged from solution by using an inert gas and introduced into a hydrogenlair flame to form excited HCI, which could be detected by means of its infrared emission. A second example of an analysis that could be carried out in this manner is the determination of available chlorine in bleaches prepared from molecular chlorine and hypochlorite. I n solution these species produce hypochlorous acid (pK. = 7.60 at 25 "C) according to eq 3 and 4, respectively, and the
CI,
-
+ 2Hz0 HOC1 + H,O+ + C1OCI- + HzO HOC1 + OH-
-
B
(3) (4)
term available chlorine refers to the total oxidizing power of the solution due to chlorine, hypochlorous acid, and hypochlorite ion, expressed in terms of an equivalent quantity of CI, (36). (The distribution of CI between these three species is temperature- and pH-dependent.) Since eq 3 is readily reversible (K4-I = 2.2 X lo3at 24 "C), addition of acid leads to the rapid generation of dissolved molecular chlorine, which can be purged from solution, converted to HCI in the flame, and detected by infrared emission as described previously. This paper reports the development of a new chlorinespecific method for the direct determination of chloride and available chlorine in aqueous samples based on the principle of flame infared emission (FIRE). The FIRE-chlorine analyzer described in this paper consists of two commercially available purge devices coupled to a FIRE detector. Aqueous chlorine-containing samples are treated chemically to convert the chloride or hypochlorous acid present into molecular chlorine (C12),which is then purged from the sample cell by using an inert carrier gas and introduced into a hydrogen/ entrained-air h e to produce vibrationally excited HCI. The FIRE detector consists of a miniature capillary-head burner that supports the hydrogenlentrained-air flame, an optical collection lens, a hand-pass filter to isolate the HCI emission band, and a PbSe photoconductive detector. EXPERIMENTAL SECTION Apparatus. The FIRE-ehlorine instrument shown in Figure 1consists of two major sections: (A) the chlorine generation and purging apparatus and (B) the FIRE detection system. The purging apparatus has been described previously (35)and consists of two demountable purge tubes connected in parallel with polyethylene tubing and two, three-way switching valves. One of the purge tubes serves as the sample chamber while the w o n d serves as the reference chamber. Each purge device consists of a 5 m L demountable purge tuhe that can be disconnected for sample introduction and cleaning. A septum located at the top of eqch puge tube allows the samples and reagents to be introduced by syringe (sulfuric acid syringe: Model No. 2300, Bectan-Dickson & Co., Rutherford, NJ; KMn04 syringe: Model No. 1001, Hamilton Co., Reno, Nv; water sample syringe: Model No. 1002, Hamilton Co.). Helium gas, maintained at a flow rate of 130 mL/min, was used to purge evolved CI2from the aqueous solution into the flame of the FIRE detector. The optimum flow rate of hydrogen in the capillary-head burner was determined to be 324 mL/min with combustion supported only by entrained air. The supply pressures of the helium and hydrogen were regulated at 0.15 atm by using triple-stage regulation (35).
Flgure 1. Schematic diagram of FIRE instrument for determination of chloride and available chlorine in aqueous samples, showing (A) purge sectwn and (E) FIRE detector. Key: He. helium cylinder; H,. hydrogen cylinder; PbSe, lead selenide detector; V. three-way valve. The optical system of the FIRE detector was modified from that described previously by the addition of a 5-cm-focal-length, f / 1 concave mirror (Model No. 44340, Oriel Corp., Stratford, CT) installed behind the hydrogen/air flame. This back collection mirror, in conjunction with the CaF, lens, directed the infrared emission from the flame onto the PhSe detector. In addition to the installation of the back collection mirror, the FIRE detector was also modified by replacing the 4.4-pm CO, optical band-pass filter with a 3.8-pm optical hand-pass filter (Model No. S-902-079,Spectrogon, Secaucus, NJ) to spectrally isolate a portion of the HCI emission consisting of most of the more intense part of the P branch. This optical band-paas filter possessed a full widt.h at half-maximum (fwhm) height of 0.18 pm and was placed immediately in front of the detector. The flame infrared emission was detected by using a PbSe photoconductive cell operated at room temperature with a bias potential of 45 V. The detector preamplifier circuit, lock-in amplifer, and recurderlintegrator have been previously discussed, and except for a slight change in the preamplifier circuit to allow for a larger bias voltage, these components were used without modification (33, 34). Reagents. All chemicalswere of ACS reagent grade and were used without further purification. A stock solutiop of 100 mM NaCl (Mallinckrodt, Inc., St. Louis, MO) was prepared by dissolving NaCI, dried a t 120 "C for 24 h, in deionized water. Standard NaCl solutions, having concentrations of 0.1,0.5,1.0, 2.0, 5.0, 8.0, and 10 mM, were prepared before use by diluting aliquots of the stock solution to the appropriate volumes. A saturated solution of KMnO, (Mallinckrodt) used in conjunction with concentrated H2S04(Mallinckrodt) served as the oxidizing agent for the aqueous chloride determination. The saturated KMnO, solution was boiled and filtered to remove any MnO, that might be present. Aqueous solutions of &NO, (Tbom Smith, Inc., Beulah, MI) and Na2S203(Sargent-WelchScientific Co., Skokie, IL) were used in the titrimetric analyses of aqueous chloride and available chlorine, respectively. Procedure. Prior to use, the flame in the FIRE system was ignited, and the instrument was allowed to w m up until a stable base line was obtained on the chart recorder. As part of the warmup procedure, He purge gas was directed through the dry reference purge tuhe and into the flame. When the instrument had stabilized, the analysis for aqueous chloride or available chlorine was carried out according to the appropriate procedure indicated below. Aqueous Chloride. To construct a calibration curve for chloride by using the FIRE system, the sample purge tube was disconnected, a 1.0" volume of an aqueous chloride standard was volume of concentrated placed on the glass frit, and a 0.5"
ANALYTICAL CHEMISTRY, VOL. 61, NO. 24, DECEMBER 15, 1989
HzSO4 was added with a syringe. The purge assembly was then reconnected, and the He flow was switched from the reference purge tube to the sample purge tube by using the dual, three-way valve system. The acidified standard solution was purged for approximately 70 s. A 0.1-mL aliquot of the saturated KMnO, solution was injected through the septum and into the sample chamber with a syringe. The chlorine gas produced from the resulting oxidation of the aqueous chloride in the sample was purged from the solution and introduced into the flame where it formed vibrationally excited HCl. After the resulting HC1 infrared emission peak had been recorded, the He flow was switched back through the reference purge device. The sample purge tube was then disconnected and rinsed thoroughly with deionized water to remove excess reagents. This process was repeated for all chloride standards to construct a calibration curve of peak intensity versus chloride concentration. Several natural water samples were collected from sources around the Waco, TX, area and stored according to standard procedures (5). One-milliliter volumes of these natural water samples were treated by using the same procedure as outlined for the preparation of the chloride calibration curve, and the resulting infrared emission signal was recorded. Aqueous chloride concentrations in these natural water samples were read from the calibration curve by using measured peak heights. For comparison purposes, aqueous chloride in these samples was also determined by argentometric titration using 0.0136 N AgN03with potassium chromate as an indicator according to procedures outlined in ref 5. Available Chlorine. Calibration curves for available chlorine (ClZ,HOCl, and OCl-) were constructed by using aqueous chloride standards as described previously. Bleach sample concentrations were read as chloride concentrations from the chloride calibration curve and converted into available C12concentration according to the relationship l/z[C1-l = [Cl,]. To determine available chlorine in bleach by using the FIRE system, a 0.5-mL aliquot of concentrated H2S04and a 1.0-mL volume of deionized water were added to the disconnected sample purge tube as described for the aqueous chloride determinations. The sample purge device was then reconnected and the He purge gas flow switched from the reference purge device to the sample purge device. After this acidified deionized water solution was purged for approximately70 s, a 0.1-mL aliquot of a bleach sample, diluted by a factor of 20 with deionized water, was introduced through the septum with a syringe. The C12gas generated from the acidfiation of the bleach sample was purged from the solution into the flame where it formed vibrationally excited HCl. After the HCl infrared emission peak had been recorded, sample cleanup was performed as described previously. Three commercially available bleach products were selected for the determination of available chlorine. For comparison purposes, available chlorine in these samples was also determined by iodometric titration using 0.151 N Na2S203according to the procedure outlined in ref 5.
RESULTS AND DISCUSSION Instrumental Configuration. The FIRE detection system used in this paper for the determination of chloride and available chlorine is similar to the system that has been previously described for use in total inorganic carbon (TIC) determinations (35).However, since the HC1 infrared emission band occurs in the 3.17-4.24-pm spectral region of the hydrogen flame (Figure 2B), the 4.4-pm optical band-pass filter used to isolate the C 0 2 emission (35) was replaced with a 3.8-pm band-pass filter. The spectra shown in Figure 2 were obtained on a Mattson Cygnus 100 Fourier transform infrared spectrometer under resolution conditions sufficient to reveal the rotational fine structure associated with the P and R branches of the HCl emission band (Figure 2B). As shown in Figure 2A, the 3.8-pm band-pass filter optically isolated the 3.64-4.03-pm portion of the P branch of the HCl emission band. The corresponding flame background spectrum (Figure 2C) shows that this region of the spectrum should be ideal for the detection of hydrogen chloride. A comparison of Figure 2B and Figure 2C also
3.0
4000
3500
MICRONS 3.5 4.0 4.5
3000
2500
2787
5.5
2000
WAVENUMBER Figure 2. Fourier transform infrared spectra from 4000 to 1800 cm-’ plotted on the same relative intensRy scale (not corrected for instrument response): (A) transmission spectrum of the band-pass filter used in the FIRE detector (maximum transmission at 3.8 pm 75% T ,
