Continuous Flow Determination of Nitrite with Membrane Separation/Chemiluminescence Detection Toyoaki Aoki Laboratory of Environmental Chemistry, College of Engineering, University of Osaka Prefecture, Mozu-umemachi, Sakai 591, Japan
A method for the determination of nitrite in water utilizing a membrane separation process and a chemiluminescence detector, with the addition of air-stripping and air-carrier, is proposed. The microporous poly(tetrafluoroethylene) tube was used as separator to transfer nitric oxide (reduced by iodine in acidic media) into a gas phase. Air-stripping was used to enhance the separation. Chemiluminescence signals produced from the reaction of nitric oxide with ozone were linearly proportional to the concentration of nitrite from 10 ppb (pg as NIL) to 5 ppm (mg N/L).The relative standard deviation (n =5) was 0.7% at 0.1 ppm. The time elapse from starting the sample flow until the signal reached a stable level was 1.5 min.
INTRODUCTION Nitrite is an anion which occurs in a wide range of natural systems. In the environment, nitrite is produced in the nitrification process, in which ammonia is oxidized by certain bacteria. Also, nitrite results from the reduction of nitrate by denitrification of bacteria. The toxicology of nitrite in humans can be both acute and chronic. Acute effects result from the oxidation of ferrous hemoglobin to ferric methemoglobin by nitrite. Chronic effects may possibly result through the relationship of nitrite to the formation of N-nitroso compounds, which are suspected to be human carcinogens (Cox and Frank, 1982). Nitrite is generally determined by its reaction with an amine to form a diazonium salt which in turn can be coupled with a second amine to form an azo dye. Sulfanilamide and N-(1 -naphthyl)ethylenediamine(NED) have been used as the reagents because of the high molar extinction of the azo dye formed (Riley and Skirrow, 1965; Strickland and Parsons, 1972). However, this method is interfered with by suspended and colored substances. Previous papers have reported continuous flow methods for the determination of chlorine (Aoki and Munemori, 1983), ammonia (Aoki, Uemura and Munemori, 1983), nitrate (Aoki et al., 1986), aqueous ozone (Aoki, 1988), carbon dioxide (Aoki et al., 1988) and chloramines (Aoki, 1989) in water by utilizing membrane-separation with tubular microporous poly(tetrafluoroethylene) (PTFE). The microporous PTFE permits the diffusion of volatile species, but not of ions and suspended substances in liquids. This paper presents the application of this separation technique to the chemiluminescence determination of nitrite in water with an air-carrier. Chemiluminescent reactions are becoming increasingly popular for trace analysis, because of their high sensitivity, rapidity, and the relatively simple instrumentation. Cox (1980) developed methods for the
determination of nitrate and nitrite which measures the chemiluminescence intensity from the reaction of nitric oxide, the species reduced by. various reductants, with ozone at atmospheric pressure. In this paper, we applied this chemiluminescent reaction to the continuously selective determination of nitrite in water.
EXPERIMENTAL Reagents. All reagents were of analytical reagent grade. Doubly distilled water prepared from an all-Pyrex still was used in the preparation of all solutions. Apparatus and procedure. A schematic diagram of the apparatus used in the determination of nitrite is shown in Fig. 1. The sample was first mixed with 1.0 mol/L potassium iodide ( R l ) and then with 0.18 mol/L sulfuric acid (R2) at a flow rate of 0.8 mL/min by water pressure resulting from a 400 mm height difference. The flow rate was 8 mL/min as the sample was then driven through separation unit S1 using peristaltic pump PI. The separation unit comprised an outer tube (5.0 mm OD, 4.0 mm ID) made of PTFE and an inner tube (2.8 mm OD, 2.0 mm ID, maximum pore size 3.5 pm) made of microporous
w
Recorder
A #t
Figure 1. Schematic diagram of the apparatus used In the determlnation of nitrite in water
CCC-0269-3879/90/0128-0130$1.50 128 BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO 3, 1990
@ Heyden & Son Limited, 1990
CF-DETERMINATION O F NITRITE WITH MEMBRANE-SEPARATION-CL
FTFE (Model TB002, Japan Gore-Tex Co., Okayama, Japan). The length of the separation unit was 600 mm. Clean air was bubbled at the mixing point ( M ) through a cylindrical PTFE tube (0.5 mm ID) at a rate of 8 mL/min by a peristaltic pump (PZ) to enhance the permeation of produced nitric oxide through the microporous PTFE membrane into the inner tube where clean air flows at a rate of 600 mL/min. The clean air was obtained by using activated charcoal (C2) and silica gel (CI) columns. Since the water vapor permeating through the microporous PTFE membrane interferes with the detection of nitric oxide by the chemiluminescencemethod, it was removed by using a Nafion permeation tube system (S2) positioned before the chemiluminescence analyzer (A) (Kimoto Co. NO analyzer). The water vapor permeated through the walls of the Nafion tubing (Ho et al., 1983) and was removed by a countercurrent of dry air swept over the outersurface of the tubing by an air pump (P3). The length of the Nafion tubing system (S2) was 600 mm. Concentration of nitric oxide in the dried air was continuously monitored at the chemiluminescence analyzer (A) where it was mixed with ozone generated in that analyzer.
RESULTS AND DISCUSSION The system was optimized by selecting optimal reagent concentration, air stripping level and the length for microporous PTFE tube. A nitrite concentration of 0.1 mg N / L (0.1 ppm) in water was used in the optimization experiments. Optimization of reagent concentration Figure 2 shows that the relative chemiluminescence intensity (RCI) increased with increasing concentration of potassium iodide and became constant at concentrations above 0.5 mol/L. When potassium bromide was used at the same levels, the RCI decreased to 1/19 compared to that obtained with iodide. In the case of sulfuric acid, the RCI reached a maximum at a concentration of 0.18 mol/L. Phosphoric acid can be used as a substitute since the RCI was very similar to that of sulphuric acid. However, the RCI was reduced to about one-half when acetic acid was used. The optimal concentrations of potassium iodide and sulphuric acid were thus selected to be 1.0 and 0.18 mol/L, respectively.
I
0
10
5 Air ( m l l m i n )
Figure 3. Effect of flow rate of air on relative chemiluminescent intensity (RCI). The flow rate of the sample (0.1 ppm nitrite) was 8 mL/min.
Stripping of nitric oxide with air The technique of air-stripping can be performed to enable more of the nitric oxide as a reduced product to permeate through the membrane. Figure 3 shows the effect of air-stripping on the sensitivity of this detecting system. The RCI increased as the flow rate of air increased. A steady-state condition was reached around a flow rate of 8mL/min, indicating that most of the nitric oxide presented in the aqueous phase diffused into the gas phase in the inner microporous PTFE tube. The sensitivity at this flow rate was about three times more than that when air-stripping was not used. Thus, in the following experiments, 8 mL/min was used as the optimum flow rate of air. Effect of length of microporous PTFE tube The sensitivity of the present method depends on the length of the separation unit. When tfie length of the unit was reduced to 200 mm, the intensity was reduced to 73% of that with a length of 600 mm as shown in Fig. 4. The degree of the reduction was smaller than that in the liquid/liquid membrane separation systems which have been used for determining free chlorine (Aoki and Munemori, 1983) and ammonia (Aoki et al., 1983) in water. This may be due to the faster rate of diffusion of gaseous compounds into air than into aqueous solutions.
20
40
60
L e n g t h ( cm
Figure 2. Effect of concentration of potassium iodide on relative chernilurninescent intensity (RCI). The concentrations of nitrite and sulfuric acid were 0.1 ppm and 0.18rnol/L, respectively.
@ Heyden & Son Limited, 1990
1
Figure 4. Effect of length of separation tube ( S l ) on relative chemiluminescent intensity (RCI). The concentration of nitrite was 0.1 ppm.
BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 3, 1990 129
T. AOKl Table 1. Interference of nitrogenous species (1.0 PPm) Species
0
1
2 3 L o g (concn, ppb )
Nitrite Nitrate Ammonia Ethyl nitrite Albumin Alanine Aspartic-acid Lysine Humic acid
4
Figure 5. Calibration curve for nitrite in water
Relative intensity (%)"
100.00 0.10 0.00 0.03 0.00 0.00 0.00
0.00 0.00
Relative intensity of nitrite at 1 .O ppm is defined as 100%.
a
ferences are summarized in Table 1. Apparent nitrite responses for only nitrate and ethyl nitrite were slightly detected among each species present in aqueous solution at a concentration of 1.0 ppm as shown in Table 1. The other species listed in Table 1 showed no response in the present method. Figure 6. Correlation diagram between present method and NED method.
Application to river water
The time it took from starting the sample flow until the signal level of the chemiluminescent detector reached a stable level was 1.5 min regardless of the PTFE tube length. Consequently, the length of the separation unit was selected to be 600 mm.
The present method was applied to the determination of nitrite in several rivers. The results were in good agreement with those obtained by the NED method as shown in Fig. 6. The samples analysed by the latter method were filtered with a 0.45 pm membrane filter. An r2 of correlation coefficient ( r ) between the present method and the NED method was 0.974 ( n = 8).
Analytical performance and interference studies ~~~~~
The chemiluminescence signals were linearly proportional to the concentration of nitrite in water from 10.0 ppb to 5.0 ppm as shown in Fig. 5. The detection limit, defined as the concentration at which the signal-tonoise ratio is 3, was 0.9ppb for nitrite in water. The level of noise was mainly affected by the dark current of photomultiplier used in the chemiluminescence analyzer. The relative standard deviation ( n = 5) was 0.7% at 0.1 ppm. Other nitrogenous species were tested to check their levels of interference with the present method. The inter-
CONCLUSIONS As a result of a combination of the membrane-separation process with chemiluminescence detection, both selectivity and sensitivity for determination of nitrite in water were improved. Furthermore, the sensitivity was increased by a factor of three with the introduction of the air stripping technique into this system. The present method takes advantage of elimination of interference from various compounds and suspended substances present in water.
REFERENCES Aoki, T. and Munemori, M . (1983). Anal. Chem. 55, 209. Aoki, T., Uemura, S. and Munemori, M. (1983). Anal. Chem. 55, 1620. Aoki,T., Uemura, S. and Munemori, M.(1986). Environ. Sci. Technol. 20, 515. Aoki, T. (1988). Anal. Lett. 21, 835. Aoki, T., Ito, K. and Munemori, M. (1988). Anal. Lett. 21, 1881. Aoki, T. (1969). Environ. Sci. Technol. 23, 46. Cox, R. D. (1980). Anal. Chem. 52, 332.
130 BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 3, 1990
Cox, R. D. and Frank, C. W. (1982). J. Anal. Toxicol. 6, 148. Ho. H. M., Guilbault, G. G. and Rietz, B. (1983). Anal. Chem. 55,1830. Riley, J. P. and Skirrow, G. (1965). Chemical Oceanography, Vol. 11, pp 370-375, Academic Press, New York, USA. Strickland, J. D. H. and Parsons, T. R. (1972). But/. Fish. Res. Bd. Can., 167, 71. Received 21 April 1989; accepted 30 April 1989.
@ Heyden & Son Limited, 1990
A method for the determination of nitrite in water utilizing a membrane separation process and a chemiluminescence detector, with the addition of air-stripping and air-carrier, is proposed. The microporous poly(tetrafluoroethylene) tube was used as separator to transfer nitric oxide (reduced by iodine in acidic media) into a gas phase. Air-stripping was used to enhance the separation. Chemiluminescence signals produced from the reaction of nitric oxide with ozone were linearly proportional to the concentration of nitrite from 10 ppb (pg as NIL) to 5 ppm (mg N/L).The relative standard deviation (n =5) was 0.7% at 0.1 ppm. The time elapse from starting the sample flow until the signal reached a stable level was 1.5 min.
INTRODUCTION Nitrite is an anion which occurs in a wide range of natural systems. In the environment, nitrite is produced in the nitrification process, in which ammonia is oxidized by certain bacteria. Also, nitrite results from the reduction of nitrate by denitrification of bacteria. The toxicology of nitrite in humans can be both acute and chronic. Acute effects result from the oxidation of ferrous hemoglobin to ferric methemoglobin by nitrite. Chronic effects may possibly result through the relationship of nitrite to the formation of N-nitroso compounds, which are suspected to be human carcinogens (Cox and Frank, 1982). Nitrite is generally determined by its reaction with an amine to form a diazonium salt which in turn can be coupled with a second amine to form an azo dye. Sulfanilamide and N-(1 -naphthyl)ethylenediamine(NED) have been used as the reagents because of the high molar extinction of the azo dye formed (Riley and Skirrow, 1965; Strickland and Parsons, 1972). However, this method is interfered with by suspended and colored substances. Previous papers have reported continuous flow methods for the determination of chlorine (Aoki and Munemori, 1983), ammonia (Aoki, Uemura and Munemori, 1983), nitrate (Aoki et al., 1986), aqueous ozone (Aoki, 1988), carbon dioxide (Aoki et al., 1988) and chloramines (Aoki, 1989) in water by utilizing membrane-separation with tubular microporous poly(tetrafluoroethylene) (PTFE). The microporous PTFE permits the diffusion of volatile species, but not of ions and suspended substances in liquids. This paper presents the application of this separation technique to the chemiluminescence determination of nitrite in water with an air-carrier. Chemiluminescent reactions are becoming increasingly popular for trace analysis, because of their high sensitivity, rapidity, and the relatively simple instrumentation. Cox (1980) developed methods for the
determination of nitrate and nitrite which measures the chemiluminescence intensity from the reaction of nitric oxide, the species reduced by. various reductants, with ozone at atmospheric pressure. In this paper, we applied this chemiluminescent reaction to the continuously selective determination of nitrite in water.
EXPERIMENTAL Reagents. All reagents were of analytical reagent grade. Doubly distilled water prepared from an all-Pyrex still was used in the preparation of all solutions. Apparatus and procedure. A schematic diagram of the apparatus used in the determination of nitrite is shown in Fig. 1. The sample was first mixed with 1.0 mol/L potassium iodide ( R l ) and then with 0.18 mol/L sulfuric acid (R2) at a flow rate of 0.8 mL/min by water pressure resulting from a 400 mm height difference. The flow rate was 8 mL/min as the sample was then driven through separation unit S1 using peristaltic pump PI. The separation unit comprised an outer tube (5.0 mm OD, 4.0 mm ID) made of PTFE and an inner tube (2.8 mm OD, 2.0 mm ID, maximum pore size 3.5 pm) made of microporous
w
Recorder
A #t
Figure 1. Schematic diagram of the apparatus used In the determlnation of nitrite in water
CCC-0269-3879/90/0128-0130$1.50 128 BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO 3, 1990
@ Heyden & Son Limited, 1990
CF-DETERMINATION O F NITRITE WITH MEMBRANE-SEPARATION-CL
FTFE (Model TB002, Japan Gore-Tex Co., Okayama, Japan). The length of the separation unit was 600 mm. Clean air was bubbled at the mixing point ( M ) through a cylindrical PTFE tube (0.5 mm ID) at a rate of 8 mL/min by a peristaltic pump (PZ) to enhance the permeation of produced nitric oxide through the microporous PTFE membrane into the inner tube where clean air flows at a rate of 600 mL/min. The clean air was obtained by using activated charcoal (C2) and silica gel (CI) columns. Since the water vapor permeating through the microporous PTFE membrane interferes with the detection of nitric oxide by the chemiluminescencemethod, it was removed by using a Nafion permeation tube system (S2) positioned before the chemiluminescence analyzer (A) (Kimoto Co. NO analyzer). The water vapor permeated through the walls of the Nafion tubing (Ho et al., 1983) and was removed by a countercurrent of dry air swept over the outersurface of the tubing by an air pump (P3). The length of the Nafion tubing system (S2) was 600 mm. Concentration of nitric oxide in the dried air was continuously monitored at the chemiluminescence analyzer (A) where it was mixed with ozone generated in that analyzer.
RESULTS AND DISCUSSION The system was optimized by selecting optimal reagent concentration, air stripping level and the length for microporous PTFE tube. A nitrite concentration of 0.1 mg N / L (0.1 ppm) in water was used in the optimization experiments. Optimization of reagent concentration Figure 2 shows that the relative chemiluminescence intensity (RCI) increased with increasing concentration of potassium iodide and became constant at concentrations above 0.5 mol/L. When potassium bromide was used at the same levels, the RCI decreased to 1/19 compared to that obtained with iodide. In the case of sulfuric acid, the RCI reached a maximum at a concentration of 0.18 mol/L. Phosphoric acid can be used as a substitute since the RCI was very similar to that of sulphuric acid. However, the RCI was reduced to about one-half when acetic acid was used. The optimal concentrations of potassium iodide and sulphuric acid were thus selected to be 1.0 and 0.18 mol/L, respectively.
I
0
10
5 Air ( m l l m i n )
Figure 3. Effect of flow rate of air on relative chemiluminescent intensity (RCI). The flow rate of the sample (0.1 ppm nitrite) was 8 mL/min.
Stripping of nitric oxide with air The technique of air-stripping can be performed to enable more of the nitric oxide as a reduced product to permeate through the membrane. Figure 3 shows the effect of air-stripping on the sensitivity of this detecting system. The RCI increased as the flow rate of air increased. A steady-state condition was reached around a flow rate of 8mL/min, indicating that most of the nitric oxide presented in the aqueous phase diffused into the gas phase in the inner microporous PTFE tube. The sensitivity at this flow rate was about three times more than that when air-stripping was not used. Thus, in the following experiments, 8 mL/min was used as the optimum flow rate of air. Effect of length of microporous PTFE tube The sensitivity of the present method depends on the length of the separation unit. When tfie length of the unit was reduced to 200 mm, the intensity was reduced to 73% of that with a length of 600 mm as shown in Fig. 4. The degree of the reduction was smaller than that in the liquid/liquid membrane separation systems which have been used for determining free chlorine (Aoki and Munemori, 1983) and ammonia (Aoki et al., 1983) in water. This may be due to the faster rate of diffusion of gaseous compounds into air than into aqueous solutions.
20
40
60
L e n g t h ( cm
Figure 2. Effect of concentration of potassium iodide on relative chernilurninescent intensity (RCI). The concentrations of nitrite and sulfuric acid were 0.1 ppm and 0.18rnol/L, respectively.
@ Heyden & Son Limited, 1990
1
Figure 4. Effect of length of separation tube ( S l ) on relative chemiluminescent intensity (RCI). The concentration of nitrite was 0.1 ppm.
BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 3, 1990 129
T. AOKl Table 1. Interference of nitrogenous species (1.0 PPm) Species
0
1
2 3 L o g (concn, ppb )
Nitrite Nitrate Ammonia Ethyl nitrite Albumin Alanine Aspartic-acid Lysine Humic acid
4
Figure 5. Calibration curve for nitrite in water
Relative intensity (%)"
100.00 0.10 0.00 0.03 0.00 0.00 0.00
0.00 0.00
Relative intensity of nitrite at 1 .O ppm is defined as 100%.
a
ferences are summarized in Table 1. Apparent nitrite responses for only nitrate and ethyl nitrite were slightly detected among each species present in aqueous solution at a concentration of 1.0 ppm as shown in Table 1. The other species listed in Table 1 showed no response in the present method. Figure 6. Correlation diagram between present method and NED method.
Application to river water
The time it took from starting the sample flow until the signal level of the chemiluminescent detector reached a stable level was 1.5 min regardless of the PTFE tube length. Consequently, the length of the separation unit was selected to be 600 mm.
The present method was applied to the determination of nitrite in several rivers. The results were in good agreement with those obtained by the NED method as shown in Fig. 6. The samples analysed by the latter method were filtered with a 0.45 pm membrane filter. An r2 of correlation coefficient ( r ) between the present method and the NED method was 0.974 ( n = 8).
Analytical performance and interference studies ~~~~~
The chemiluminescence signals were linearly proportional to the concentration of nitrite in water from 10.0 ppb to 5.0 ppm as shown in Fig. 5. The detection limit, defined as the concentration at which the signal-tonoise ratio is 3, was 0.9ppb for nitrite in water. The level of noise was mainly affected by the dark current of photomultiplier used in the chemiluminescence analyzer. The relative standard deviation ( n = 5) was 0.7% at 0.1 ppm. Other nitrogenous species were tested to check their levels of interference with the present method. The inter-
CONCLUSIONS As a result of a combination of the membrane-separation process with chemiluminescence detection, both selectivity and sensitivity for determination of nitrite in water were improved. Furthermore, the sensitivity was increased by a factor of three with the introduction of the air stripping technique into this system. The present method takes advantage of elimination of interference from various compounds and suspended substances present in water.
REFERENCES Aoki, T. and Munemori, M . (1983). Anal. Chem. 55, 209. Aoki, T., Uemura, S. and Munemori, M. (1983). Anal. Chem. 55, 1620. Aoki,T., Uemura, S. and Munemori, M.(1986). Environ. Sci. Technol. 20, 515. Aoki, T. (1988). Anal. Lett. 21, 835. Aoki, T., Ito, K. and Munemori, M. (1988). Anal. Lett. 21, 1881. Aoki, T. (1969). Environ. Sci. Technol. 23, 46. Cox, R. D. (1980). Anal. Chem. 52, 332.
130 BIOMEDICAL CHROMATOGRAPHY, VOL. 4, NO. 3, 1990
Cox, R. D. and Frank, C. W. (1982). J. Anal. Toxicol. 6, 148. Ho. H. M., Guilbault, G. G. and Rietz, B. (1983). Anal. Chem. 55,1830. Riley, J. P. and Skirrow, G. (1965). Chemical Oceanography, Vol. 11, pp 370-375, Academic Press, New York, USA. Strickland, J. D. H. and Parsons, T. R. (1972). But/. Fish. Res. Bd. Can., 167, 71. Received 21 April 1989; accepted 30 April 1989.
@ Heyden & Son Limited, 1990
