Anal. Chem. 1989, 61 733-735 I
733
Simultaneous Determination of Bromide and Iodide as Acetone Derivatives by Gas Chromatography and Electron Capture Detection in Natural Waters and Biological Fluids Ldszl6 Maros,* Mdria Kdldy, a n d S a r o l t a Igaz Institute of Inorganic and Analytical Chemistry, L. Eotvos University, Muzeum krt. 4/B,1088 Budapest, Hungary
Oxidation of bromide and iodide ions in acidic solutions In the presence of acetone forms the corresponding acetone derivatives. Iodate was reduced with thiosulfate prior to the determination. After extraction with benzene the bromo- and lodoacetone were measured by gas chromatography using electron capture detectlon. The bromide and lodlde contents of rainwater, drinking water, river water, seawater, oil brine, common salt, cow milk, and human blood serum were determined. The relative standard deviations for bromide at lo-' M and for iodide at lo-' M concentration were 1.9% and 3.0%, respectively, using a 10-mL sample for the determination without preconcentration.
INTRODUCTION Bromine and iodine are trace elements in natural waters and biological fluids in the concentration range of 10"L-10-8 M. Bromide (1)and iodide (2)can be determined separately by spectrophotometric methods based on catalytic reactions in this concentration range. These methods, however, are time-consuming, and iodide determination is subject to interference from certain divalent cations (3). X-ray fluorescence spectrometric methods (4-7) and neutron activation analysis (7) were used for seawaters and biological fluids. Ion chromatographic separations followed by UV detection (8,9), conductometry (lo),ion-selective electrode analysis (11),and catalytic photometric detection (12)are reported in the literature for bromide and iodide determination in different matrices (13). Highly sensitive determinations of iodide were obtained by using gas chromatography and electron capture detection of iodoketone derivatives (14-17). Methods for determinations of bromide concentration in serum have been reported, through oxidation of bromide to bromine followed by bounding the bromine to cyclohexene (18)or to 2,4-dimethylphenol (19)and assaying the bromo derivative by gas chromatography. The derivatization with epoxides, forming and the thermal decomposition of tethaloalkanols (20,21), raalkylammonium halides to form alkyl halides (22,23)can also be applied for gas chromatographic analysis of halides, but these methods are not sensitive enough to determine bromide and iodide in natural waters and biological fluids without preconcentrations. For the selective and highly sensitive simultaneous determination of bromide and iodide we suggest a gas chromatographic separation combined with electron capture detection of the acetone derivatives. The method for iodide determination based on oxidation with iodate in thenpresence of acetone followed by an extraction of iodoacetone with nhexane (16)gives inaccurate results, especially if other reductive substances are present that reduce iodate to iodine or iodide. The quantitative oxidation of iodide and bromide in the presence of acetone to iodo- and bromoacetone can be performed by chromate and permanganate, without the interference of other reducing substances. Instead of n-hexane, benzene was found to be more convenient for the extraction
of bromo- and iodoacetone from aqueous solutions, due to the significantly higher solubilities of these compounds in benzene. The sample preparation for the gas chromatographic analysis consists of the oxidation of bromide and iodide in the presence of acetone and of a single-step extraction of the derivatives with benzene. EXPERIMENTAL SECTION Reagents. Chemicals used were of analytical reagent grade. Acetone and benzene were redistilled before use. A 3 x lo4 M p-dichlorobenzene (internal standard) solution was prepared in benzene. Distilled water (prepared from chlorinated tap water which contains bromoacetone and iodoacetone) was purified as follows: 2 L of distilled water was boiled vigorously for 10 min; then after the addition of 2 g of sodium hydroxide, it was redistilled. All solutions were prepared with redistilled water. Solutions of 5 M perchloric acid, 5 M phosphoric acid, 0.001 M sodium thiosulfate, 0.01 M chromic acid, 0.02 M potassium permanganate, 0.1 M oxalic acid, and 0.6 M sodium tungstate were used. The 5 M perchloric acid (which contains bromate and iodate impurities) was purified as follows: 0.1 g of sodium sulfite was added to 100 mL of solution and boiled for 30 min; then the sulfite was oxidized with 2 mL 0.02 M potassium permanganate solution and boiled for a further 5 min. After cooling, it was made up to the original volume. The 0.02 M potassium permanganate solution (100 mL) was acidified with 1mL of 5 M perchloric acid and boiled for about 5 min before use. Standard bromide solutions of 1 X and 1 x lo4 M, and standard iodide solutions of 1 X lo4 and 1 X lod M concentrations were prepared from 1 X 10-1M potassium bromide and potassium iodide solutions. Apparatus. A Chromatron Model GCHF-5 (GDR) gas chromatograph equipped with tritium electron capture detector was used for sample analysis. The column was a 2 m x 3 mm stainless steel column packed with 5 % neopentylglycol sebacate on Chromosorb G HP (80-100 mesh). Nitrogen was used as a carrier gas with a flow rate of 40 mL/min. The temperature of the column oven and detector was 140 "C; the glass-lined injector block temperature was set at 210 "C. A Chinoin Model Digint 80 L (Hungary) electronic integrator was used for sample level calculations. The column was preconditioned at 200 "C for 24 h by using 40 mL/min nitrogen as a carrier gas. For daily work a 30-60 min precondition was necessary at 190 "C. Calibration. Standards for calibration were prepared with a concentration ratio of bromide to iodide of 1O:l. In glassstoppered test tubes of 20-mL volume, the corresponding aliquots of standard bromide and iodide solutions measured with a 1O-wL microsyringe were added to 10 mL of redistilled water. The concentrations of the calibrating standards were 1,2,4,6, ..., 16 X lo-' M for bromide and 1, 2, 4, 6, ..., 16 X M for iodide, respectively. The solutions (and a blank) were mixed with 0.5 mL of 5 M perchloric acid, 0.5 mL of acetone, and 0.1 mL of 0.02 M potassium permanganate solution. The mixture was allowed to stand for 60 min; then the permanganate was reduced with 0.1 mL of 0.1 M oxalic acid. The solutions were mixed with 0.1 mL of 0.001 M sodium thiosulfate and, after 5 min, with 0.1 mL of 0.01 M chromic acid. After 30 min 2.0 mL of benzene solution of p-dichlorobenzene was given to the mixtures. The mixtures were shaken vigorously and after a few minutes standing 1-wL aliquots of the benzene layers were taken for gas chromatographic analysis.
0003-2700/89/0361-0733$01.50/00 1989 American Chemical Society
734
ANALYTICAL CHEMISTRY, VOL. 61, NO. 7, APRIL 1, 1989
Preparation of Samples. Ten milliliters of rainwater, river water, or tap water, 1mL of mineral water, and 10 pL of seawater or oil brine were diluted with redistilled water to 10 mL, and the procedure was followed as described under Calibration. From common salt (sodium chloride) a 2% solution was prepared. A 1-mL aliquot was diluted to 10 mL and acidified with 0.5 mL of 5 M phosphoric acid (instead of perchloric acid). The procedure was the same as described above, except that permanganate was reduced with oxalic acid after 30 min. From hydrochloric acid 10 mL of 0.1 M solution was used for sample preparation without further acidification. The time of permanganate oxidation was 5 min. Cow milk and human blood serum were deproteinized before derivatization as follows. In a centrifuge tube of 10-mL volume, 1 mL of sample was mixed with 0.1 mL of 0.001 M sodium thiosulfate, 0.5 mL of sodium tungstate solution, and 3.4 mL of redistilled water, and the mixture was acidified with 1 mL of 0.5 M perchloric acid. Simultaneously a blank was prepared. After being allowed to stand for 10 min, the solution was centrifuged for 5 min at 5OOO rpm. A 1-mLaliquot of the clear solution diluted to 10 mL with redistilled water was used for sample preparation. In the case of cow milk 0.5 mL of 0.02 M potassium permanganate was used for the oxidation. The total bromine and iodine content of cow milk and human blood serum was determined after alkaline ashing. A 1-mL aliquot of the sample was mixed with 1 mL of 5 M sodium hydroxide solution in a Pyrex tube (20 X 150 mm) and dried at 105 "C for 20 h. Simultaneously a blank was prepared. The tubes were heated in a furnace for 30 min at 150 "C and for 60 min at 600 O C . The ash was dissolved in 5 mL of redistilled water and filtered on a paper. A 1-mL aliquot of the filtrate diluted to 10 mL with redistilled water was used for sample preparation described under Calibration. The inorganic iodine concentration of seawater and common salt was individually determined by using 10 mL of seawater or 10 mL of 2% common salt solutions. The solutions were mixed with 0.5 mL of 5 M perchloric acid and 0.1 mL of 0.001 M sodium thiosulfate. After 5 min 0.5 mL of acetone and 0.1 mL of 0.01 M chromic acid solution were added. After 30 min, 2.0 mL of a benzene solution of p-dichlorobenzenewas given to the solutions.
RESULTS AND DISCUSSION According to our studies, the principle for iodide determination by converting it to iodoacetone can also be used for bromide determination. Oxidation of bromide with permanganate in the presence of a high excess of acetone leads to the formation of monobromoacetone, which can be extracted from aqueous solution and measured quantitatively by gas chromatography with electron capture detection. Inorganic reductants do not interfere if permanganate is present in sufficient excess. In most cases the organics, such as olefmes and aldehydes, which can cause loss of bromine, react fast with permanganate, so the interference of these compounds, especially in the presence of a high excess of acetone, can be avoided. In order to study the effect of organics, measurements were performed with 10 mL of lo4 M bromide-lo-' M iodide standard solutions containing M glucose, allyl alcohol, and formic acid, respectively. Two milliliters of 0.02 M sodium permanganate was applied for the oxidation. Within the experimental error no difference was found between the organic-containing solution and the calibrating standard of the same bromide and iodide concentration. In a solution of a strong acid (0.25 M perchloric acid was used) a small portion of the chloride also reacts with permanganate, and chloroacetone is formed. Bromoacetone can be separated from chloroacetone on the column used here, but at a 3-4-fold excess of chloroacetone, accurate values for bromoacetone cannot be obtained. Chloride has a catalytic effect on bromide oxidation and bromoacetone formation. In a solution of a weak acid (0.25 M phosphoric acid was used) the oxidation of chloride is negligible. However, the higher the concentration of chloride present, the faster the oxidation of bromide. Therefore the oxidation of bromide in sodium
chloride ([Br-]/[Cl-] = 1.2 X was performed in phosphoric acid solution. The methods described for determination of iodide content in cow milk by gas chromatography and electron capture detection were based either on the formation of iodoacetone by oxidation with iodate (16)or on that of iodobutanone by oxidation with nitrite (17)or chromate (24). Iodoketones were extracted with n-hexane in each case. The low partition coefficient of iodoacetone between n-hexane and water, 2, does not allow the extractive concentration of iodoacetone. The partition coefficient of iodobutanone, 8, makes possible a favorable extractive concentration, but the two isomers, 1iodo-2-butanone and 3-iodo-2-butanone, result in two peaks on the chromatogram (14, 15). The partition coefficients of bromoacetone and iodoacetone were found to be 0.6 and 1.7 for n-hexane and water and 10.0 and 16.1 for benzene and water, respectively. The use of benzene instead of n-hexane for the extraction gives the possibility of favorable extractive concentration of the acetone derivatives. A special problem may arise in the determination of iodine in natural waters, because inorganic iodine can be present as iodide, molecular iodine, and iodate (11). The simultaneous oxidation of iodide and bromide by permanganate in the presence of acetone leads to less than a stoichiometric amount of iodoacetone formation, since the oxidation of iodide to iodate and the iodoacetone formation run parallel. The iodate originally present and formed in a certain amount by the oxidation of iodide with Permanganate was converted to iodoacetone as follows: the excess of permanganate was removed with oxalate, and after the reduction of iodate by thiosulfate iodide was oxidized by chromate to form iodoacetone, with acetone. For the GC separation of the acetone derivatives, the use of the neopentylglycol sebacate stationary phase was found to be the most convenient. The less polar OV 17 gave no sharp peaks, and by the use of the more polar neopentylglycol adipate phase the internal standard p-dichlorobenzene overlapped with the iodoacetone peak. The column was stable over more than one thousand injections without notable change of the separation characteristics. Chromatograms are shown in Figure 1. The following relative standard deviations (RSD,%) in the simultaneous determinations were found: for bromide 1.0% at the concentration of 1 X lo* M, for iodide 1.7% at the concentration of 1 x lo-' M, for bromide 1.9% at the concentration of 1 x lo-' M for iodide 3.5% at the concentration of 1 X lo-* M (n = 7). The values for the blank sample were M with 3.9% RSD found as follows: for bromide 1.0 X M with 10.4% RSD (n = 7). The and for iodide 3.0 X detection limit (3~7)of bromide was 1.2 X lo4 M and that of iodide was 0.9 X M. The detection limit of our method for bromide and iodide is 1-2 orders of magnitude lower than that of the other methods used previously and makes possible the determination of these components in freshwaters-involving rainwaters-without preconcentrations of the samples. Linear detector response was found up to concentration of 3 X lo-' M. It can be seen in Figure 2 that the calibration curves have different characteristics for the two ions. In the concentration range of practical interest the calibration curve for iodide is linear but for bromide the detector response is not linear and no extrapolation can be made. The relative response ratios are fairly constant, though the sensitivity of the detector can show a change up to 120-30% from day to day, depending on the column or septa bleeding. The stability of bromoacetone is higher than that of the iodoacetone in the benzene solutions. After a week long
ANALYTICAL CHEMISTRY, VOL. 61, NO. 7, APRIL 1, 1989
735
Table I. Characteristic Bromide and Iodide Concentrations of Natural Waters, Cow Milk, Human Blood Serum, and Chloride Matrices
range of the concentrations found, M
sample
b
a
bromide
rainwater river and lake water
mineral water oil brine seawater cow milk
human blood serum
"t, 6
? 4
2
'
0
6
"
'-t 4'
2
-
-
t
0
time (min) time (min) Flgure 1. Chromatograms for bromide and iodide determinations: (a) reagent blank, (b) drinking water; (S) internal standard. Sensitivity was changed at 160 s from 3 to 30.
(0.5-1.9) (2.0-8.9) (4.6-5.0) (0.2-1.0) (3.7-7.7) (2.4-3.2) (2.1-3.8)
X
x 10-7 X X X X X
10"
lo-'
iodide (0.5-4.7) X (0.2-1.1) x 10-7 (4.2-8.0) X (0.4-1.1) X lo-' (1.2-4.4) X (1.5-2.2) X 10" (0.12-2.1) X 10"
mol/mol of chloride iodide
sample
bromide
common salt hvdrochloric acid
1.2 x 10-5 4.2 X lo4
6.7 X lo-*
