1852
Anal. Chem. 1981, 53, 1852-1854
Amperometric Determination of Ammonia Gas with Immobilized Nitrifying Bacteria Isao Karube," Tadashi Okada, and Shuichl Suruki Research Laboratory of Resources Utilization, Tokyo Institute of Technology, Nagatsuta-cho, Midori-ku, Yokohama, 227, Japan
A mlcroblai sensor consisting of immobilized nltrifylng bacteria and an oxygen electrode has been developed for the amperometrlc determination of ammonia gas. The response time for the determhatlon of ammonia was withln 4 min. A linear relationship was observed between the current decrease and the ammonia concentration below 42 mg/L. The mlnimum concentration for the determination of ammonia was 0.1 mg/L. The current decrease was reproducible within f4% of the relative error. The seiectlvity of the microbial sensor for ammonia was satisfactory. The current output of the sensor was almost constant for more than 10 days and 200 assays.
The determination of ammonia is important in clinical, environmental, and industrial process analyses. The determination of ammonia is presently performed by spectrophotometric methods. However, these methods require a long reaction time and additional reagents ( I ) . On the other hand, electrochemical monitoring of ammonia may have definite advantages. An ammonia gas electrode consisting of a combined glass electrode and a gas permeable membrane is usually used for this purpose (2). In this case, the determination must be performed under strong alkaline conditions (above pH 11). The ammonia electrode is based on potentiometric detection of ammonia. However, volatile compounds such as amines often interfere with the determination of ammonia. Recently, several new ammonium ion electrodes with better selectivity have been developed (3-5). These new types of ammonia sensors utilizing an ammonium selective membrane [silicone rubber-nonactin membrane ( 3 ) or PVC-nonactin membrane ( 4 , 5 ) ]are based on potentiometric ammonia gas determination. Therefore, an ammonia sensor based on amperometry is desirable for the electrochemical determination of ammonia. As previously reported, many microbial sensors consisting of immobilized whole cells and an oxygen electrode have been developed for the determination of BOD (6,7), antibiotics (8), acetic acid (9),methane gas (IO),mutagen (21)and so on. The concentration of these organic compounds could be determined from the respiration activity of immobilized whole cells with an oxygen electrode. Nitrifying bacteria contains two genera of bacteria. One genus (i.e., Nitrosomonas sp.) of bacteria utilizes ammonia as the sole source of energy and oxygen is consumed by the respiration as follows
NH3 + 1.502
Nitrosornonas sp.
NOz- + H2O $. H+
The other genus (i.e., Nitrobacter sp.) of bacteria oxidizes nitrite to nitrate as follows:
NO2- + 0.502
Nitrobacter
The oxidation of both substrates ("3,
0p.
* N03-
NOz-) proceeds a t a
0003-2700/81/0353-1852$01.25/0
high rate. And oxygen uptake by the bacteria can be directly determined by the oxygen electrode attached to the immobilized bacteria. Therefore, ammonia is determined by the microbial sensor using immobilized nitrifying bacteria and an oxygen electrode. As previously reported, a microbial electrode for ammonium ion has been developed (12). However, the selectivity of the electrode was poor because the electrode responded to glucose and other nutrients. This might be caused by contaminated bacteria in the electrode. Therefore, addition of antibiotics such as chloramphenicol in a sample solution was required for the determination of ammonium ion. In this paper, an improved ammonia gas electrode using a gas permeable membrane, immobilized nitrifying bacteria, and an oxygen electrode is described. EXPERIMENTAL SECTION Materials. Reagents were commercially available analytical reagents or laboratory grade materials. Deionized water was used for all procedures. Microorganisms. The nitrifying bacteria were obtained from Aginomoto Central Research Institute. These bacteria were isolated from activated sludges in a waste water treatment facility (Kawasaki,Japan). The unidentified bacteria were used for the microbial electrode. These nitrifying bacteria were cultured in 1 L of medium containing 6 g of (NH4)+304,0.5 g of K2HP04, 0.05 g of MgS04.7Hz0,4 mg of CaCl2.2H20,and 10 g of CaCOB per 1L of distilled water. The medium (pH 8.0) was placed in a vessel (2 L) and the bacteria were cultured under aerobic conditions (l/s VVM (volume of air/min)/(volume of medium)) for more than 2 weeks at 25 "C. The pH of the culture medium was controlled to 8.0 by adding 1N NazCOs. Cresol red (0.05 ppm) was added as a pH indicator. Calcium and magnesium sal& were sterilized separately to avoid precipitation. Immobilization of the Bacteria. The bacteria (0.3 g wet weight) were suspended in 5 mL of sterilized water and the suspension was dripped onto a porous acetylcellulosemembrane (Millipore Co., Type HA, 0.45 pm pore size, 4.7 mm diameter, 150 pm thickness) with slight suction. The bacteria were retained on the acetylcellulose membrane. Assembly of the Microbial Electrode. The scheme of the microbial electrode is shown in Figure 1. The oxygen electrode (Model U-1, Ishikawa Seisakujo, Tokyo) consisted of a Teflon membrane (50 pm thick), a platinum cathode, a lead anode, and a sodium hydroxide (30%) electrolyte. The porous membrane retaining immobilized bacteria was cut into a circle (0.8 cm diameter) and soaked in the buffer (0.1 M KHZPO4-NaOHbuffer pH 8.0). Because nitrifying bacteria are activated in an alkaline condition (pH 7-9) (13) and fixed on the surface of a Teflon membrane of the oxygen electrode, the bacterial membrane was covered with a gas-permeableTeflon membrane (Millipore Co., Type FH, 0.5 pm pore size) and fastened with rubber rings. Procedures. The system consisted of a cell (100 mL) with the microbial electrode, magnetic stirred (1000 rpm), and a recorder (TOA Electronics, Model CDR-11A). The microbial sensor was inserted into a sample solution (glycine-NaOH buffer, pH 10,ionic strength 0.1) (50 mL) and the sample solution was saturated with dissolved oxygen and stirred magnetically while measurements were taken. Nitrifying bacteria require dissolved oxygen above 0.7 mg/L for nitrification, below that concentration,nitrification does not occur. The temperature of the cell was maintained at 30 A 0.1 "C by a thermostated bath. The current obtained from 0 1981 American Chemical Society
ANALYTICAL CHEMISTRY,, VOL. 53, NO. 12, OCTOBER 1981
4
1853
3
2
1
Flgure 1. The microbial sensor system for ammonia: (1) gas permeable Teflon membrane; (2) immobilized cells; (3) Teflon membrane; (4) cathode (Ft); (5) anode (Pb); (6) electrolyte (NaOH); (7) magnetic stirrer; (8) amplifier; (9) recorder.
1
0
1
Tire
1
8
1
(rnin)
Flgure 2. Response curves of the microbial sensor for ammonlin. The sensor was inserted into sample solution containing (1) 35,(2) 2!1, and (3)10.5 mg/L ammonium hydroxide. The experiments were petformed at pH 10 and 30 O C .
the electrode was directly connected with a recorder through a 2 kQ resistance.
RESULTS AND DISCUSSION Response of t h e Electrode. Figure 2 shows typical response curves of the sensor based on immobilized nitrifying bacteria. The current at time zero was obtained with the buffer (pH 10.0, glycine-NaC1-NaOH) saturated with oxygen. The pH of a sample solution had to be kepi, sufficiently above the pK value for ammonia1 (9.1 at 30 "C) because ammonium ions cannot pass through the gas permeable membrane. These currents of the response curves correspond to the endogenous respiration level of the immobilized bacteria. When the ammonia solution was injected into the buffer solution, ammonium ion clhanged to ammonia gas. Ammonia gas permeated through the gas permeable membrane and was assimilated by the immobilized bacteria. Oxygen was then consumed by the bacteria so that the concentration of dissolved oxygen around the membrane decreased. The current decreased until it reached a steady state which indicated that the consumption of oxygen by the bacteria and the diffusion of oxygen from the sample solution to the bacteria membrane were in equilibrium. The steady-state current depended on the concentration of ammonia. In this system, ammonia gas escapes to the air for a long time. However, the response time for the determination of ammonia was within 4 min. Therefore, only a small amount of ammonia gas escapes to the air. This steady-state phase continued for over 20 min with this microbial sensor and a commercial potentiometric ammonia gas sensor (Horiba Co., Model 5002-05T) under identical purging conditions. When the sensor was inserted in tap water, the current of the sensor returned to its initial level within 5 min. A microbial sensor consisting of immobilized Nitrosomonas europaea and an oxygen electrode has been prepared for the determination of ammonia1 (12). However, the reusability of the sensor was very poor. This result was caused by the inactivation of bacterium with nitrite formed (14). Therefore, stable nitrifying bacteria are required for the microbial sensor. The nitrifying bacteria vvhich contained Nitrosomonas sp. and Nitrobacter sp., isolated from waste water treatment
6
7
8
9
10 11
12
OH
Flgure 3. Effect on current decrease of the microbial sensor. Thse experiments were performed at the same conditions described in Figure 2 except for buffer solutlon [pH 6-7, phosphate buffer; pH 10-12, glycine buffer (ionic strength 0. l)] employed.
facilities was cultured and employed for this microbial sensor. In this newly microbial electrode system, nitrite was used b y Nitrobacter sp., but the concentration of nitrate was low in the immobilized bacterial layer, and nitrate did not inhibit the nitrifying metabolism (25). Inhibition of nitrification process was not observed in the experiments of continuous usage below 42 mg/L for 200 assays. However, the response time of the microbial sensor is longeir than that of potentiornetric electrode. Since the electrode is covered with triple membranes (a Teflon membrane, a bacterial membrane, and a gas permeable membrane), the diffusion of substrates such as ammonia and oxygen is a rats determining factor of the response. Employment of thin membranes may improve the sensitivity and the response of the sensor. Furthermore, the determination of ammonia can be done within 1 min by the rate assay. Calibration. A liniear relationship was observed between the current decrease (tlhe current difference between the initial and the steady state) imd the ammonia concentration below 42 mg/L (current decrease 4.7 PA). The minimum concentration for the determination of ammonia was 0.1 mg/L (signal to noise, 20; reproducibility, f 5%). The reproducibility of the current decrease was examined by using the same sample. The current decrease was reproducible within f 4 % of the relative error when a sample solution containing 21 mg/L of ammonium hydrooxide was employed. The standard devia tion was 0.7 mg/L in 20 experiments. Thus the amperometric determination of ammonia became possible by the microbial sensor. The sensitivity of the microbial sensor was almost at the same level as that of a glasei electrode, and its miniinum measurable concentration was 0.1 mg/L. Selectivity of the Microbial Sensor. The selectivity of the microbial electrode for ammonia was examined. The sensor did not respond to volatile compounds such as acetic acid, ethyl alcohol, and amines (diethylamine, propylamine, and butylamine) or to involatile nutrients such as glucose, amino acids, and metal ions (potassium ion, calcium ion, and zinc ion). On the other hand, preliminary experiments showed that the microbial sensor gradually responded to nutrients such as glucose and acetic acid (22). These results were caused from growth of the contamiinated bacteria in the membrane. The response to organic compounds depends on the assirnilability of the immobilized bacteria. Therefore, in this paper, the immobilized nitrifying bacteria on the electrode was covered with a gas permeable membrane and only volatile compounds could penetrate through the membrane . Nitrifying bacteria, which utilize ammonia, did not assimilate acetic acid, ethyl alcohol, volatile amines, g l u m , amino acids, and metal ions. Therefore, these substances did not
1854
ANALYTICAL CHEMISTRY, VOL. 53, NO. 12, OCTOBER 1981
n
1
Cell content
0
2
Effect of cell content on current decrease of the microbial sensor. The experiments were performed at the same conditions described in Figure 2 except for the cell content in the electrode employed. Figure 4.
affect the determination of ammonia by the sensor. Therefore, selectivity of this microbial electrode became satisfactory. Effect of pH on Current Decrease. Figure 3 shows the effect of pH on current decrease of the microbial sensor. The current decrease increased with raising pH of a sample solution and reached a plateau above pH 10. In these conditions, most of ammonia vaporized and easily permeated through the gas permeable membrane. Therefore, because the maximum current decrease was observed above pH 10, we aim at the development of an ammonia gas sensor for urine or blood where a drastic pH change is not desirable. Therefore, the buffer solution of pH 10 was employed for further works. Effect of Cell Content on t h e Current Decrease. As described above, the nitrifying bacteria were immobilized on the porous acetylcellulose membrane and the respiration activity of the immobilized bacteria was determined by the oxygen electrode. Therefore, the cell contents on the membrane affected the current output of the sensor. The cell content of the membrane was changed from 0.19 to 2.58 mg/mm3 and their effects on the current of the sensor were examined. Figure 4 shows the effect of cell content on the current decrease of the electrode. The current decrease increased with increasing the cell content and reached the maximum a t a cell content of 1 mg/mm3. Further increase of the cell content decreased the current decrease. These results may be caused from lysis of bacteria at a high bacteria content. A cell content of 1mg/mm3 was employed for the electrode hereinafter. Reusability and Application of the Microbial Sensor. The long-term stability of the microbial sensor was examined with a sample solution containing 33 mg/L of ammonia. The results are shown in Figure 5. The current output of the electrode was almost constant for more than 10 days and 200 assays. Therefore, the microbial sensor can be used for a long time for the assay of ammonia.
2
0
Time
(mg/m3)
6
8 1 0
(days)
Figure 5. Reusability of the microbial sensor. The experimental cGditions were the same conditions in Figure 2.
The microbial sensor was applied to the determination of ammonia in human urine. Urine was diluted with the glycine-NaC1-NaOH buffer (50 times) and employed for experiments. The concentration of ammonia was determined by the electrochemical sensor and the conventional method (1).
Good comparative results were obtained between ammonia concentrations determined by both methods (correlation coefficient 0.9). Therefore, the sensor proposed gives an economical and reliable method for assay of ammonia in biological fluids. A new sensor prepared can be used for a long time and applied to the determination of ammonia in urine. In conclusion, the microbial sensor appears to be quite promising and very attractive for the amperometric determination of ammonia. LITERATURE CITED (1) Tabara, M. "Testlng Methods for Ammonium Ion KOlOl", Japanese Industrial Standard Committee: Tokyo, 1979; pp 116-122. (2) Balley, P. L. "Analysls with Ion-Selective Electrodes"; Heyden: Spectrum House, Hlllview Gardens, London, 1978; Chapter 7. (3) Guilbault, G. G.; Nagy G. Anal. Chem. 1973, 45,417-419. (4) Meyerhoff, M. E. Anal. Chem. 1880, 52, 1532-1534. (5) Meyerhoff, M. E.;Robins, R. H. Anal. Chem. 1960, 52,2383-2387. (6) Karube, I.; Mitsuda, S.; Matsunaga, T.; Suzukl, S. J. ferment. Techno/. 1977, 55,243-248. (7) Hlkuma, M.; Suzuki, H.; Yasuda, T.; Karube, I.; Suzukl, S. Eur. J. Appl. Mlcrobiol. Biotechnol. 1979, 8,289-297. (8) Karube, I.; Matsunaga, T.; Suzukl, S. Anal. Chlm. Acta 1979, 109, 39-44. (9) Hlkuma, M.; Kubo, T.; Yasuda, T.; Karube, 1.; Suzukl, S. Anal. Chlm. Acta 1979, 109, 33-38. (10) Okada, T.; Karube, I.; Suzuki, S. €ur. J . Appl. Microblol. Biotechnol. 1981, 12, 102-106. (11) Karube, I.; Matsunaga, T.; Nakahara, T.; Suzukl, S. Anal. Chem. 1981, 53,1024. (12) Hikuma, M.; Kubo, T.; Yasuda, T.; Karube, I.; Suzuki, S. Anal. Chem. 1980, 52,1020-1024. (13) Boon, B.; Laudelout, H. Blochem. J. 1962, 85,440-447. (14) Anthonisen, A. C.; Loehr, R. C.; Prakasam, T. B. S.; Srlnath, E. G. J.- Water Pollut. Control Fed. 1976. 48, 835-852. (15) Aleem, M. I. H.;Alexander, M. Appl. Microblol. 1960, 8 , 80-84.
RECEIVED for review January 21,1981. Resubmitted May 6, 1981. Accepted June 26, 1981.
