- Email: [email protected]
Heterotrophic nitrification and aerobic denitrification of high-strength ammonium in anaerobically digested sludge by Alcaligenes faecalis strain No. 4
Journal of Bioscience and Bioengineering VOL. 117 No. 6, 737e741, 2014 www.elsevier.com/locate/jbiosc
Heterotrophic nitrification and aerobic denitrification of high-strength ammonium in anaerobically digested sludge by Alcaligenes faecalis strain No. 4 Makoto Shoda* and Yoichi Ishikawa Able Corporation, 7-9 Nishigoken-cho, Shinjuku-ku, Tokyo 216-0812, Japan Received 25 September 2013; accepted 29 November 2013 Available online 7 January 2014
Alcaligenes faecalis strain No. 4 which is capable of heterogeneous nitrification and aerobic denitrification, was used to remove high-strength ammonium (approximately 1 g NHD 4 -N/l) from digested sludge, the product of an anaerobic digestion reactor, in which methane was produced from excess municipal sewage sludge. Repeated batch operations were conducted at 20 C and 30 C for 550 h, using a jar fermentor. The removal ratios of high-strength ammonium reached 90e100% within 24 h, and the average ammonium removal rate was 2.9 kg-N/m3/day, more than 200 times higher than that in conventional nitrificationedenitrification processes. During these operations, the cell density was maintained at 108e109 cells of A. faecalis strain No. 4/ml. At 3% NaCl in the digested sludge, strain No. 4 exhibited an ammonium removal rate of 3 kg-N/m3/day. Ó 2013, The Society for Biotechnology, Japan. All rights reserved. [Key words: Alcaligenes faecalis; Heterotrophic nitrification; Aerobic denitrification; High-strength ammonium; Anaerobic digestion]
Conventional ammonium removal in wastewater treatment plants consists of nitrification by autotrophs under aerobic conditions, followed by denitrification by heterotrophs under anaerobic conditions. The disadvantages of this system are as follows: (i) the rate of nitrification is slow, and (ii) the autotrophs are vulnerable to high loads of ammonium and organic matter. Therefore, treatment of wastewater containing high-strength ammonium by the conventional method requires wastewater dilution and a large-scale treatment plant due to the long hydraulic retention time required for nitrification (1e3). In a previous paper (4), we showed that Alcaligenes faecalis strain No. 4, which has the ability to carry out heterotrophic nitrification and aerobic denitrification, removed significant amounts of high-strength ammonium and COD from crude piggery wastewater. Recently, several bacteria have been reported to carry out heterotrophic nitrification and aerobic denitrification under lowstrength ammonium conditions (5e10). As an alternative to the conventional method, an anammox method has been intensively studied, mainly because this operation does not require the addition of a carbon source (11e15). However, the problems of the extremely slow growth rate of anammox bacteria and the vulnerability of these bacteria to high concentrations of organic carbon and ammonium must be solved to enable the practical development of wastewater treatment system, especially for wastewaters containing high amounts of ammonium and organic matter. Due to recent trends of limiting fossil energy consumption, sustainable methods of energy production including anaerobic digestion or bioethanol production have been attracting increasing
* Corresponding author. Tel.: þ81 3 260 0485; fax: þ81 3 260 0513. E-mail address: [email protected] (M. Shoda).
attention. In anaerobic digestion, livestock waste, municipal garbage, and waste from the food industry are used for the digestion, leading to the production of wastewater containing a high concentration of ammonium. Therefore, the development of an effective method of the wastewater treatment is a crucial factor enabling the production of methane. In this paper, A. faecalis strain No. 4 was applied to remove highstrength ammonium from digested sludge generated in a municipal anaerobic digestion plant to assess the possibility of efficient biological treatment of the wastewater and to observe the performance of strain No. 4 under several operational conditions. MATERIALS AND METHODS Strain used The detailed characteristics of A. faecalis strain No. 4 were described in previous papers (4,16). Cultured cells of strain No. 4 were mixed with a 50% glycerol solution in vials and stored at 84 C. For each pre-culture, one vial was used as the strain No. 4 inoculum. Medium used The pre-culture of strain No. 4 used a synthetic medium containing 14 K2HPO4, 6 KH2PO4, 15 trisodium citrate dihydrate, 2 (NH4)2SO4, 0.2 MgSO4$7H2O (all units in g per liter), and 2 ml of trace mineral solution. The trace mineral solution contained the following components (g per liter): 57.1 EDTA (2,20 ,200 ,200 0 -(Ethane-1,2-diyldinitrilo) tetraacetic acid) 2Na, 3.9 ZnSO4$7H2O, 7 CaCl2$2H2O, 5.1 MnCl2$4H2O, 5.0 FeSO4$7H2O, 1.1 (NH4)6Mo7O24$4H2O, 1.6 CuSO4$5H2O, and 1.6 CoCl2$6H2O. The reactor used The reactions were carried out in a small-scale jar fermentor (total volume 1 L, working volume 300 ml) (BMJ-01PI, Able Corp., Tokyo). Dissolved oxygen (DO) concentrations and pH values were monitored with a DO sensor (SDOC-12F, Able Corp., Tokyo) and a pH sensor (Easyferm Plus 225, Hamilton Bonaduz AG, Bonaduz, Switzerland) inserted into the fermentor. The temperature was controlled at 20 C or 30 C. The oxygen transfer coefficients, kLa were varied by changing the agitation speed from 300 to 700 rpm at a constant air supply rate of 300 ml per min. Experimental material The digested sludge was supplied by Yokohama Municipal Sewage Treatment Center (Yokohama, Japan) where the excess municipal dehydrated activated sludge was digested at 37 C in a 6000 ton-scale anaerobic
1389-1723/$ e see front matter Ó 2013, The Society for Biotechnology, Japan. All rights reserved. http://dx.doi.org/10.1016/j.jbiosc.2013.11.018
738
SHODA AND ISHIKAWA
J. BIOSCI. BIOENG.,
digester. The main characteristics of the digested sludge are as follows: pH 7.3, 24 mg/l volatile fatty acids, 2700 mg/l total nitrogen, 1200 mg/l ammonium-nitrogen, 150 mg/l soluble BOD, 1000 mg/l total BOD, 900 mg/l soluble COD and 20,000 mg/l total COD. Experimental procedure The strain No. 4 cells were pre-cultivated in 100 ml synthetic medium in a 500 ml shaking flask at 30 C at a shaking speed of 100 strokes per min (spm) for 2 days. As control experiments, the ammonium removal was measured under the following three experimental conditions using a jar fermentor: (i) 250 ml of digested sludge plus 50 ml of water, (ii) 250 ml of digested sludge plus 50 ml of water and 20 g of trisodium citrate dihydrate, and (iii) 250 ml of digested sludge plus 50 ml of the strain No. 4 culture. The ammonium removal of the three controls was compared with the removal of the sample containing 250 ml of digested sludge, 50 ml of strain No. 4 culture and 20 g of trisodium citrate dihydrate. In each operation, the jar fermentor was operated at 30 C at an airflow rate of 300 ml per min and an agitation speed of 700 rpm. In repeated batch experiments, 50 ml of the pre-culture of strain No. 4, 250 ml of the digested sludge and 20 g of trisodium citrate dihydrate were mixed in the fermentor, and the treatment of the ammonium was conducted at an air flow rate of 300 ml per min and an agitation speed of 700 rpm. One ml of the culture was sampled periodically, and the concentration of ammonium was determined. After the ammonium concentration was confirmed to be reduced by more than 90% of the initial concentration, 50 ml of the culture was used for the subsequent treatment by adding a fresh 250 ml of digested sludge and 20 g of trisodium citrate dihydrate. The cell numbers of strain No. 4 were determined at the start and at the end of each cycle of batch cultivation. In the previous paper (16), we demonstrated the optimal C/N ratio for strain No. 4 was 10, indicating that at this ratio, nitrogen and carbon sources were simultaneously consumed. Based on the ratio, 1 g-N and 10 g-C was balanced and thus 10 g-C corresponded to 38 g of trisodium citrate dihydrate. If no other carbon source existed in the sludge, 38 g of trisodium citrate dihydrate should be added. In this experiment, 20 g of trisodium citrate dihydrate was arbitrarily chosen by expecting existence of some carbon sources in the sludge. Although the anaerobic digestion was conducted at 37 C, the sludge was stored at room temperature in the treatment center. Thus, the temperatures 20 C and 30 C were selected. As the temperature 30 C was the optimal temperature for No. 4 in the removal of ammonium, which was shown in a previous paper (16), the value of removal rate at 30 C will be optimal. The temperature 20 C can be the most probable temperature in practical operation. Analytical method The ammonium concentration was determined using an ammonium sensor (SNH-10, Able Corp., Tokyo). To determine the number of cells of strain No. 4, the sampled culture was diluted and plated on synthetic agar plates containing the synthetic medium and 1.5% agar, and then the plates were incubated at 30 C for 2 days. As it was previously confirmed that strain No. 4 grew on the plates significantly faster than other cells indigenous to the digested sludge and that strain No. 4 exhibited characteristic morphological features, the colonies that appeared on the plates after 2 days were counted as strain No. 4 cells and the cell concentration was expressed as cells/ml.
RESULTS Control experiments Fig. 1 shows the change in ammonium concentration under four experimental conditions using the digested sludge. The ammonium removal by the digested sludge containing strain No. 4 cells and citrate as a carbon source was significantly higher than the removal under the other three
FIG. 1. Ammonium removal from municipal wastewater by A. faecalis strain No. 4 under four different experimental conditions using anaerobically digested sludge. Each batch treatment was conducted at 30 C at an agitation speed of 700 rpm in a jar fermentor. The samples consisted of the digested sludge only (diamonds), the digested sludge with trisodium citrate dihydrate (squares), the digested sludge with strain No. 4 cells (triangles) and the digested sludge with both strain No. 4 cells and trisodium citrate dihydrate (circles).
FIG. 2. The ammonium concentration in the digested sludge during repeated batch treatment with A. faecalis strain No. 4. These experiments were conducted at 30 C at an agitation speed of 700 rpm in a jar fermentor.
conditions. The control containing digested sludge and strain No. 4 cells achieved 20e30% ammonium removal in 10 h, presumably because some carbon sources in the digested sludge were consumed by strain No. 4. Therefore, the additional carbon sources were needed for strain No. 4 to remove more than 90% of the ammonium from the sludge. Ammonium removal in the repeated batch experiment Fig. 2 shows the change in ammonium concentration over times in a repeated batch experiment at 30 C, and Fig. 3 shows the change in the number of strain No. 4 cells during the same experiment. Agitation speed was controlled at 700 rpm, resulting in an oxygen transfer coefficient, kLa of 450 h1, which guaranteed aerobic conditions by maintaining more than 2 mg/l of DO concentration in the reactor. More than 90% of ammonium was removed within 10e20 h, and the number of strain No. 4 cells varied between 108 and 109 cells/ml. The average ammonium removal rate during the experimental period was 2.9 kg-N/m3/day. This value is significantly higher than that in conventional nitrificationedenitrification processes, and similar to that in an efficient anammox process (11,14). Between 169 and 221 h, the operation was stopped and the jar fermentor was left statically at room temperature. When the operation resumed, ammonium removal was observed without any delay, indicating that interrupted operation causes no adverse effect on the activity of strain No. 4. Figs. 4 and 5 present the change in ammonium concentration and the change in strain No. 4 cell number, respectively, during operation at 20 C. Although the average ammonium removal rate decreased to 1.5 k-g/m3/day, half of the rate at 30 C, a relatively high rate of ammonium removal was maintained. The suspension of the operation between 214 and 337 h in Fig. 4 resulted in some delay in the ammonium removal once the operation was re-started. However, the rate of ammonium removal from the second batch
FIG. 3. Change in the number of A. faecalis strain No. 4 cells in the same experiment shown in Fig. 2.
VOL. 117, 2014
AEROBIC DENITRIFICATION OF HIGH-STRENGTH AMMONIUM
739
A
B
FIG. 4. The ammonium concentration under the same experimental conditions as in Fig. 2, except that the experimental temperature was 20 C.
operation after the operation was re-started was similar to the rate of the previous repeated batch operation. In these experiments, pH values were maintained above 8 to facilitate optimal activity of strain No. 4. Effect of DO concentration on ammonium removal In practical operation, DO concentration is related to energy consumption, agitation, air supply, and the activity of strain No. 4. The effects of changes in the oxygen supply rate on ammonium removal were studied. Fig. 6 shows DO concentrations and ammonium concentrations at different oxygen supply rates controlled by changing the agitation speed from 700 to 300 rpm. At 700 rpm, the DO concentration was maintained at more than 2 mg/l during the operation, and the ammonium was completely removed within 10 h. However, when the agitation speed was reduced to 400 or 300 rpm, the DO concentration decreased below 1 mg/l, reducing the ammonium removal rate, indicating that the oxygen supply is an important factor for efficient ammonium removal. Ammonium removal under high salt conditions Strain No. 4 exhibited the unique feature of removing ammonium under high salt conditions. When strain No. 4 cells were cultivated in synthetic medium containing 0%, 3% and 6% NaCl in a shaking flask, the ammonium concentration changed as shown in Fig. 7. The ammonium removal began after induction periods of 1 day at 3% NaCl and 5 days at 6% NaCl, and the ammonium removal rates were similar to that at 0% NaCl. Then, NaCl was added to the digested sludge to 3%, and repeated batch treatment was conducted at 30 C in a jar fermentor. The results of this experiment are shown in Fig. 8. The ammonium removal rate
FIG. 5. Change in the number of A. faecalis strain No. 4 cells under the same experimental conditions as in Fig. 3, except that the experimental temperature was 20 C.
C
FIG. 6. Change in DO (triangles) and ammonium (circles) concentrations at different oxygen transfer coefficients, kLa, 45 h1(A), 150 h1 (B) and 450 h1 (C) when the agitation speeds were 300 rpm (A), 400 rpm (B) and 700 rpm (C), respectively.
reached 3 kg-N/m3/day at the 6th repeated batch operation after the gradual acclimation of strain No. 4 to the saline medium.
DISCUSSION Strain No. 4 has demonstrated the ability to reduce COD and ammonium concentration in high-strength non-diluted piggery wastewater (4). However, additional carbon sources must be supplied to wastewater containing a relatively low amount of carbon sources and a high concentration of ammonium. For the digested sludge used in this research, strain No. 4 achieved a significantly
FIG. 7. Ammonium removal by A. faecalis strain No. 4 at 0% NaCl (triangles), 3% NaCl (squares) and 6% NaCl (circles) in shaking flasks containing 100 ml of synthetic medium at 30 C.
740
SHODA AND ISHIKAWA
J. BIOSCI. BIOENG.,
FIG. 8. Ammonium removal by A. faecalis strain No. 4 in the digested sludge containing 3% NaCl in repeated batch experiment. Symbols: open squares, 1st cycle; closed diamonds, 2nd cycle; closed triangles, 3rd cycle; closed squares, 4th cycle; open triangles, 5th cycle; closed circles, 6th cycle.
high removal rate of ammonium. The digested sludge contained almost no fatty acids, but 20e30% of ammonium removal in the control experiments (Fig. 1) suggested the existence of carbon compounds in the digested sludge that can be utilized by strain No. 4 presumably derived from high total COD. Available carbon sources for strain No. 4 include low- and highmolecular fatty acids, phenol, methanol, some solvents and amino acids (Akuzawa, E., Master’s thesis, Tokyo Institute of Technology, Tokyo, 2009). The experiment performed here included 20 g of trisodium citrate dihydrate. Generally, the C/N ratio of the intracellular components in microorganisms is 10, indicating that 10 units of carbon are used when 1 unit of N is consumed with the C mainly to synthesize cellular materials. In previous experiments, 30e40% of ammonium was reduced to nitrogen gas by strain No. 4 (4). Assuming a similar level of denitrification in these experiments, 0.6e0.7 g-N/l were used for cell synthesis, indicating that 6e7 g-C/l was required. When 20e30% of carbon is available from the digested sludge, approximately 5 g-C/l should be supplied from outside. As 20 g of trisodium citrate dihydrate contained 16 g-C/l, 6 g of trisodium citrate dihydrate is sufficient to enable the complete removal of 1 g-N/l. Actually, when only 7 g of trisodium citrate dihydrate (C/N ¼ 5) was added, almost 100% removal of ammonium was attained at similar removal rates (data not shown). Concerning carbon requirement in strain No. 4, the conventional denitrification process using methanol is compared by using the following reaction. þ NO 3 þ 1:08CH3 OH þ H /0:065C5 H7 NO2 þ 0:47N2 þ 0:76CO2
þ 2:44H2 O
ð1Þ
The C/N ratio in this reaction is 2. The strain No. 4 process demanded C/N ratio 5. In this point, strain No. 4 process is disadvantageous. However, as a total system, strain No. 4 process will be advantageous over the conventional process in that no dilution of high strength of wastewater is required, only single reactor with compact size is needed and significantly high removal rate is possible when less expensive carbon sources from waste or unused resources are available. Under these conditions, this system can achieve efficient ammonium removal. The nitrification rates in the conventional system are mostly reported to be less than 0.01 kg-N/m3/day (1,2), and the immobilization of the cell mass increased this rate by a factor of 10 (3). However, this system remained vulnerable to high-strength carbon or nitrogen sources. Higher ammonium removal rates have been reported using the anammox method, including 0.96 kg-N/m3/day in the SHARONanammox process (15), 2.3 kg-N/m3/day in a fluidized bed using synthetic medium (14) and more than 4 kg-N/m3/day for gel pellets of anammox biomass (11). The problems with this method include that it is time-consuming to produce sufficient amounts of biomass,
a long period of time is required to stabilize the system, and rapid recovery of the system once inefficient removal occurs is difficult. On the other hand, it is easy to cultivate strain No. 4 in a synthetic medium with a doubling time of 2e3 h and it is easy to immobilize the cells on different immobilization carriers (Joo, H.S., Ph.D. thesis, Tokyo Institute of Technology, Tokyo, 2005). When cultured strain No. 4 cells were stored at 4 C, high activity was maintained for several months, and the cells remained tolerant to high osmotic pressure (data not shown). Recently, many bacteria have been found to exhibit heterotrophic nitrification and aerobic denitrification, such as Rhodococcus sp. (5), Diaphorobacter sp. (6), Bacillus sp. (7), Bacillus methylotrophicus (8), A. faecalis (9) and Pseudomonas sp. (10). Although these bacteria were cultivated in the initial ammonium concentration range from 30 to 300 mg/l and the removal rates were in the range of 0.08e0.6 kg-N/m3/day, these bacteria hold significant potential for ammonium treatment. Although strain No. 4 is not osmophilic, the cells were able to achieve ammonium removal under high saline conditions. In our basic experiment, strain No. 4 was found to synthesize the osmoprotectant, hydroxyectoine during the lag time when the cells were exposed to high salt concentration (Tsunoda, H., Master’s thesis, Tokyo Institute of Technology, Tokyo, 2007) as shown in Fig. 7. As most microorganisms are vulnerable to wastewater with high saline concentrations or highstrength solvents due to the resulting high osmotic pressure, strain No. 4 is able to effectively remove ammonium under such conditions after a certain acclimation period. Thus, this new system can remove highstrength ammonium from marine aquaculture wastewater. Another advantage of strain No. 4 is that the cells of strain No. 4 can suppress the growth of plant pathogens, as we previously reported (17,18). This indicates that the excess sludge of strain No. 4 generated in the treatment facility can be reutilized in farmland to protect against plant diseases. In relatively small-scale reactors like this jar fermentor, oxygen supply capacity is lower than that in practical large-scale reactors. The power requirement in wastewater treatment is one of important factors considered to be in operation. Thus, DO level in large scale reactors can be maintained at lower agitation speeds and the power requirement for strain No. 4 will be almost equivalent to that in conventional aerobic nitrification process. The difference in delay of ammonium removal was observed between 20 C and 30 C after shutdown experiments (Figs. 2 and 4). The exact reason of difference of delay between 20 C and 30 C is not clear. However, the atmospheric temperature during 30 C operation was higher than that at 20 C and the shutdown period at 30 C was 2 days, but that at 20 C was 5 days. These factors may be reflected in the delay of the recovery at 20 C. The next step in this research is to perform continuous treatment using high-strength ammonium (2000e3000 mg-N/l) without dilution of the original samples using suspended or immobilized strain No. 4 cells. ACKNOWLEDGMENTS The authors thank Yokohama Municipal Sewage Treatment Center (Yokohama, Japan) for supplying the digested sludge and valuable information on the sludge. References 1. Kuai, L. and Verstraete, W.: Ammonium removal by the oxygen-limited autotrophic nitrification-denitrification system, Appl. Environ. Microbiol., 64, 4500e4506 (1998). 2. Franco-Rivera, A., Paniagua-Michel, J., and Zamora-Castro, J.: Characterization and performance of constructed nitrifying biofilms during nitrogen bioremediation of a wastewater effluent, J. Ind. Microbiol. Biotechnol., 34, 279e287 (2007).
VOL. 117, 2014 3. Rostron, W. M., Stuckey, D. C., and Young, A. A.: Nitrification of high-strength ammonia wastewater: comparative study of immobilized media, Water Res., 35, 1169e1178 (2001). 4. Joo, H. S., Hirai, M., and Shoda, M.: Piggery wastewater treatment using Alcaligenes faecalis strain No. 4 with heterotrophic nitrification and aerobic denitrification, Water Res., 40, 3029e3036 (2006). 5. Chen, P. Z., Li, J., Li, Q. X., Wang, Y. C., Li, S. P., Ren, T. Z., and Wang, L. G.: Simultaneous heterotrophic nitrification and aerobic denitrification by bacterium Rhodococcus sp. CPZ24, Bioresour. Technol., 116, 266e270 (2012). 6. Khardenavis, A. A., Kapley, A., and Purohit, H. J.: Simultaneous nitrification and denitrification by diverse Diaphorobacter sp., Appl. Microbiol. Biotechnol., 77, 403e409 (2007). 7. Yang, X. P., Wang, S. M., Zhang, D. W., and Zhou, L. X.: Isolation and nitrogen removal characteristics of an aerobic heterotrophic nitrifying-denitrifying bacterium, Bacillus subtilis Al, Bioresour. Technol., 102, 854e862 (2011). 8. Zhang, Q. L., Liu, Y., Ai, G. M., Miao, L. L., Zheng, H. Y., and Liu, Z. P.: The characteristics of a novel heterotrophic nitrification-aerobic denitrification bacterium, Bacillus methylotrophicus strain L7, Bioresour. Technol., 108, 35e44 (2012). 9. Zhao, B., An, Q., He, Y. L., and Guo, J. S.: N2O and N2 production during heterotrophic nitrification by Alcaligenes faecalis strain NR, Bioresour. Technol., 116, 379e385 (2012). 10. Zhu, L., Ding, W., Feng, L. J., Kong, Y., Xu, J., and Xu, X. Y.: Isolation of aerobic denitrifiers and characterization for their potential application in the bioremediation of oligotrophic ecosystem, Bioresour. Technol., 108, 1e7 (2012).
AEROBIC DENITRIFICATION OF HIGH-STRENGTH AMMONIUM
741
11. Furukawa, K., Inatomi, Y., Qiao, S., Quan, L., Yamamoto, T., Isaka, K., and Sumio, T.: Innovative treatment system for digester liquor using anammox process, Bioresour. Technol., 100, 5437e5443 (2009). 12. Molinuevo, B., Garcia, M. C., Karakashev, D., and Angelidaki, I.: Anammox for ammonia removal from pig manure effluents: effect of organic matter content on process performance, Bioresour. Technol., 100, 2171e2175 (2009). 13. Strous, M., van Gerven, E., Zheng, P., Kuenen, J. G., and Jetten, M. S. M.: Ammonium removal from concentrated waste streams with the anaerobic ammonium oxidation (Anammox) process in different reactor configurations, Water Res., 31, 1955e1962 (1997). 14. Van de Graaf, A. A., deBruijin, P., Robertson, L. A., Jetten, M. S. M., and Kuenen, J. G.: Autotrophic growth of anaerobic ammonium-oxidizing microorganisms in a fluidized bed reactor, Microbiology, 142, 2187e2196 (1996). 15. Van Dongen, U., Jetten, M. S. M., and van Loosdrecht, M. C. M.: The SHARONAnammox process for treatment of ammonium rich wastewater, Water Sci. Technol., 44, 153e160 (2001). 16. Joo, H. S., Hirai, M., and Shoda, M.: Nitrification and denitrification in highstrength ammonium by Alcaligenes faecalis, Biotechnol. Lett., 27, 773e778 (2005). 17. Honda, N., Hirai, M., Ano, T., and Shoda, M.: Antifungal effect of a heterotrophic nitrifier Alcaligenes faecalis, Biotechnol. Lett., 20, 703e706 (1998). 18. Honda, N., Hirai, M., Ano, T., and Shoda, M.: Control of tomato damping-off caused by Rhizoctonia solani by the heterotrophic nitrifier Alcaligenes faecalis and its product, hydroxylamine, Ann. Phytopathol. Soc. Japan, 65, 153e162 (1999).
Heterotrophic nitrification and aerobic denitrification of high-strength ammonium in anaerobically digested sludge by Alcaligenes faecalis strain No. 4 Makoto Shoda* and Yoichi Ishikawa Able Corporation, 7-9 Nishigoken-cho, Shinjuku-ku, Tokyo 216-0812, Japan Received 25 September 2013; accepted 29 November 2013 Available online 7 January 2014
Alcaligenes faecalis strain No. 4 which is capable of heterogeneous nitrification and aerobic denitrification, was used to remove high-strength ammonium (approximately 1 g NHD 4 -N/l) from digested sludge, the product of an anaerobic digestion reactor, in which methane was produced from excess municipal sewage sludge. Repeated batch operations were conducted at 20 C and 30 C for 550 h, using a jar fermentor. The removal ratios of high-strength ammonium reached 90e100% within 24 h, and the average ammonium removal rate was 2.9 kg-N/m3/day, more than 200 times higher than that in conventional nitrificationedenitrification processes. During these operations, the cell density was maintained at 108e109 cells of A. faecalis strain No. 4/ml. At 3% NaCl in the digested sludge, strain No. 4 exhibited an ammonium removal rate of 3 kg-N/m3/day. Ó 2013, The Society for Biotechnology, Japan. All rights reserved. [Key words: Alcaligenes faecalis; Heterotrophic nitrification; Aerobic denitrification; High-strength ammonium; Anaerobic digestion]
Conventional ammonium removal in wastewater treatment plants consists of nitrification by autotrophs under aerobic conditions, followed by denitrification by heterotrophs under anaerobic conditions. The disadvantages of this system are as follows: (i) the rate of nitrification is slow, and (ii) the autotrophs are vulnerable to high loads of ammonium and organic matter. Therefore, treatment of wastewater containing high-strength ammonium by the conventional method requires wastewater dilution and a large-scale treatment plant due to the long hydraulic retention time required for nitrification (1e3). In a previous paper (4), we showed that Alcaligenes faecalis strain No. 4, which has the ability to carry out heterotrophic nitrification and aerobic denitrification, removed significant amounts of high-strength ammonium and COD from crude piggery wastewater. Recently, several bacteria have been reported to carry out heterotrophic nitrification and aerobic denitrification under lowstrength ammonium conditions (5e10). As an alternative to the conventional method, an anammox method has been intensively studied, mainly because this operation does not require the addition of a carbon source (11e15). However, the problems of the extremely slow growth rate of anammox bacteria and the vulnerability of these bacteria to high concentrations of organic carbon and ammonium must be solved to enable the practical development of wastewater treatment system, especially for wastewaters containing high amounts of ammonium and organic matter. Due to recent trends of limiting fossil energy consumption, sustainable methods of energy production including anaerobic digestion or bioethanol production have been attracting increasing
* Corresponding author. Tel.: þ81 3 260 0485; fax: þ81 3 260 0513. E-mail address: [email protected] (M. Shoda).
attention. In anaerobic digestion, livestock waste, municipal garbage, and waste from the food industry are used for the digestion, leading to the production of wastewater containing a high concentration of ammonium. Therefore, the development of an effective method of the wastewater treatment is a crucial factor enabling the production of methane. In this paper, A. faecalis strain No. 4 was applied to remove highstrength ammonium from digested sludge generated in a municipal anaerobic digestion plant to assess the possibility of efficient biological treatment of the wastewater and to observe the performance of strain No. 4 under several operational conditions. MATERIALS AND METHODS Strain used The detailed characteristics of A. faecalis strain No. 4 were described in previous papers (4,16). Cultured cells of strain No. 4 were mixed with a 50% glycerol solution in vials and stored at 84 C. For each pre-culture, one vial was used as the strain No. 4 inoculum. Medium used The pre-culture of strain No. 4 used a synthetic medium containing 14 K2HPO4, 6 KH2PO4, 15 trisodium citrate dihydrate, 2 (NH4)2SO4, 0.2 MgSO4$7H2O (all units in g per liter), and 2 ml of trace mineral solution. The trace mineral solution contained the following components (g per liter): 57.1 EDTA (2,20 ,200 ,200 0 -(Ethane-1,2-diyldinitrilo) tetraacetic acid) 2Na, 3.9 ZnSO4$7H2O, 7 CaCl2$2H2O, 5.1 MnCl2$4H2O, 5.0 FeSO4$7H2O, 1.1 (NH4)6Mo7O24$4H2O, 1.6 CuSO4$5H2O, and 1.6 CoCl2$6H2O. The reactor used The reactions were carried out in a small-scale jar fermentor (total volume 1 L, working volume 300 ml) (BMJ-01PI, Able Corp., Tokyo). Dissolved oxygen (DO) concentrations and pH values were monitored with a DO sensor (SDOC-12F, Able Corp., Tokyo) and a pH sensor (Easyferm Plus 225, Hamilton Bonaduz AG, Bonaduz, Switzerland) inserted into the fermentor. The temperature was controlled at 20 C or 30 C. The oxygen transfer coefficients, kLa were varied by changing the agitation speed from 300 to 700 rpm at a constant air supply rate of 300 ml per min. Experimental material The digested sludge was supplied by Yokohama Municipal Sewage Treatment Center (Yokohama, Japan) where the excess municipal dehydrated activated sludge was digested at 37 C in a 6000 ton-scale anaerobic
1389-1723/$ e see front matter Ó 2013, The Society for Biotechnology, Japan. All rights reserved. http://dx.doi.org/10.1016/j.jbiosc.2013.11.018
738
SHODA AND ISHIKAWA
J. BIOSCI. BIOENG.,
digester. The main characteristics of the digested sludge are as follows: pH 7.3, 24 mg/l volatile fatty acids, 2700 mg/l total nitrogen, 1200 mg/l ammonium-nitrogen, 150 mg/l soluble BOD, 1000 mg/l total BOD, 900 mg/l soluble COD and 20,000 mg/l total COD. Experimental procedure The strain No. 4 cells were pre-cultivated in 100 ml synthetic medium in a 500 ml shaking flask at 30 C at a shaking speed of 100 strokes per min (spm) for 2 days. As control experiments, the ammonium removal was measured under the following three experimental conditions using a jar fermentor: (i) 250 ml of digested sludge plus 50 ml of water, (ii) 250 ml of digested sludge plus 50 ml of water and 20 g of trisodium citrate dihydrate, and (iii) 250 ml of digested sludge plus 50 ml of the strain No. 4 culture. The ammonium removal of the three controls was compared with the removal of the sample containing 250 ml of digested sludge, 50 ml of strain No. 4 culture and 20 g of trisodium citrate dihydrate. In each operation, the jar fermentor was operated at 30 C at an airflow rate of 300 ml per min and an agitation speed of 700 rpm. In repeated batch experiments, 50 ml of the pre-culture of strain No. 4, 250 ml of the digested sludge and 20 g of trisodium citrate dihydrate were mixed in the fermentor, and the treatment of the ammonium was conducted at an air flow rate of 300 ml per min and an agitation speed of 700 rpm. One ml of the culture was sampled periodically, and the concentration of ammonium was determined. After the ammonium concentration was confirmed to be reduced by more than 90% of the initial concentration, 50 ml of the culture was used for the subsequent treatment by adding a fresh 250 ml of digested sludge and 20 g of trisodium citrate dihydrate. The cell numbers of strain No. 4 were determined at the start and at the end of each cycle of batch cultivation. In the previous paper (16), we demonstrated the optimal C/N ratio for strain No. 4 was 10, indicating that at this ratio, nitrogen and carbon sources were simultaneously consumed. Based on the ratio, 1 g-N and 10 g-C was balanced and thus 10 g-C corresponded to 38 g of trisodium citrate dihydrate. If no other carbon source existed in the sludge, 38 g of trisodium citrate dihydrate should be added. In this experiment, 20 g of trisodium citrate dihydrate was arbitrarily chosen by expecting existence of some carbon sources in the sludge. Although the anaerobic digestion was conducted at 37 C, the sludge was stored at room temperature in the treatment center. Thus, the temperatures 20 C and 30 C were selected. As the temperature 30 C was the optimal temperature for No. 4 in the removal of ammonium, which was shown in a previous paper (16), the value of removal rate at 30 C will be optimal. The temperature 20 C can be the most probable temperature in practical operation. Analytical method The ammonium concentration was determined using an ammonium sensor (SNH-10, Able Corp., Tokyo). To determine the number of cells of strain No. 4, the sampled culture was diluted and plated on synthetic agar plates containing the synthetic medium and 1.5% agar, and then the plates were incubated at 30 C for 2 days. As it was previously confirmed that strain No. 4 grew on the plates significantly faster than other cells indigenous to the digested sludge and that strain No. 4 exhibited characteristic morphological features, the colonies that appeared on the plates after 2 days were counted as strain No. 4 cells and the cell concentration was expressed as cells/ml.
RESULTS Control experiments Fig. 1 shows the change in ammonium concentration under four experimental conditions using the digested sludge. The ammonium removal by the digested sludge containing strain No. 4 cells and citrate as a carbon source was significantly higher than the removal under the other three
FIG. 1. Ammonium removal from municipal wastewater by A. faecalis strain No. 4 under four different experimental conditions using anaerobically digested sludge. Each batch treatment was conducted at 30 C at an agitation speed of 700 rpm in a jar fermentor. The samples consisted of the digested sludge only (diamonds), the digested sludge with trisodium citrate dihydrate (squares), the digested sludge with strain No. 4 cells (triangles) and the digested sludge with both strain No. 4 cells and trisodium citrate dihydrate (circles).
FIG. 2. The ammonium concentration in the digested sludge during repeated batch treatment with A. faecalis strain No. 4. These experiments were conducted at 30 C at an agitation speed of 700 rpm in a jar fermentor.
conditions. The control containing digested sludge and strain No. 4 cells achieved 20e30% ammonium removal in 10 h, presumably because some carbon sources in the digested sludge were consumed by strain No. 4. Therefore, the additional carbon sources were needed for strain No. 4 to remove more than 90% of the ammonium from the sludge. Ammonium removal in the repeated batch experiment Fig. 2 shows the change in ammonium concentration over times in a repeated batch experiment at 30 C, and Fig. 3 shows the change in the number of strain No. 4 cells during the same experiment. Agitation speed was controlled at 700 rpm, resulting in an oxygen transfer coefficient, kLa of 450 h1, which guaranteed aerobic conditions by maintaining more than 2 mg/l of DO concentration in the reactor. More than 90% of ammonium was removed within 10e20 h, and the number of strain No. 4 cells varied between 108 and 109 cells/ml. The average ammonium removal rate during the experimental period was 2.9 kg-N/m3/day. This value is significantly higher than that in conventional nitrificationedenitrification processes, and similar to that in an efficient anammox process (11,14). Between 169 and 221 h, the operation was stopped and the jar fermentor was left statically at room temperature. When the operation resumed, ammonium removal was observed without any delay, indicating that interrupted operation causes no adverse effect on the activity of strain No. 4. Figs. 4 and 5 present the change in ammonium concentration and the change in strain No. 4 cell number, respectively, during operation at 20 C. Although the average ammonium removal rate decreased to 1.5 k-g/m3/day, half of the rate at 30 C, a relatively high rate of ammonium removal was maintained. The suspension of the operation between 214 and 337 h in Fig. 4 resulted in some delay in the ammonium removal once the operation was re-started. However, the rate of ammonium removal from the second batch
FIG. 3. Change in the number of A. faecalis strain No. 4 cells in the same experiment shown in Fig. 2.
VOL. 117, 2014
AEROBIC DENITRIFICATION OF HIGH-STRENGTH AMMONIUM
739
A
B
FIG. 4. The ammonium concentration under the same experimental conditions as in Fig. 2, except that the experimental temperature was 20 C.
operation after the operation was re-started was similar to the rate of the previous repeated batch operation. In these experiments, pH values were maintained above 8 to facilitate optimal activity of strain No. 4. Effect of DO concentration on ammonium removal In practical operation, DO concentration is related to energy consumption, agitation, air supply, and the activity of strain No. 4. The effects of changes in the oxygen supply rate on ammonium removal were studied. Fig. 6 shows DO concentrations and ammonium concentrations at different oxygen supply rates controlled by changing the agitation speed from 700 to 300 rpm. At 700 rpm, the DO concentration was maintained at more than 2 mg/l during the operation, and the ammonium was completely removed within 10 h. However, when the agitation speed was reduced to 400 or 300 rpm, the DO concentration decreased below 1 mg/l, reducing the ammonium removal rate, indicating that the oxygen supply is an important factor for efficient ammonium removal. Ammonium removal under high salt conditions Strain No. 4 exhibited the unique feature of removing ammonium under high salt conditions. When strain No. 4 cells were cultivated in synthetic medium containing 0%, 3% and 6% NaCl in a shaking flask, the ammonium concentration changed as shown in Fig. 7. The ammonium removal began after induction periods of 1 day at 3% NaCl and 5 days at 6% NaCl, and the ammonium removal rates were similar to that at 0% NaCl. Then, NaCl was added to the digested sludge to 3%, and repeated batch treatment was conducted at 30 C in a jar fermentor. The results of this experiment are shown in Fig. 8. The ammonium removal rate
FIG. 5. Change in the number of A. faecalis strain No. 4 cells under the same experimental conditions as in Fig. 3, except that the experimental temperature was 20 C.
C
FIG. 6. Change in DO (triangles) and ammonium (circles) concentrations at different oxygen transfer coefficients, kLa, 45 h1(A), 150 h1 (B) and 450 h1 (C) when the agitation speeds were 300 rpm (A), 400 rpm (B) and 700 rpm (C), respectively.
reached 3 kg-N/m3/day at the 6th repeated batch operation after the gradual acclimation of strain No. 4 to the saline medium.
DISCUSSION Strain No. 4 has demonstrated the ability to reduce COD and ammonium concentration in high-strength non-diluted piggery wastewater (4). However, additional carbon sources must be supplied to wastewater containing a relatively low amount of carbon sources and a high concentration of ammonium. For the digested sludge used in this research, strain No. 4 achieved a significantly
FIG. 7. Ammonium removal by A. faecalis strain No. 4 at 0% NaCl (triangles), 3% NaCl (squares) and 6% NaCl (circles) in shaking flasks containing 100 ml of synthetic medium at 30 C.
740
SHODA AND ISHIKAWA
J. BIOSCI. BIOENG.,
FIG. 8. Ammonium removal by A. faecalis strain No. 4 in the digested sludge containing 3% NaCl in repeated batch experiment. Symbols: open squares, 1st cycle; closed diamonds, 2nd cycle; closed triangles, 3rd cycle; closed squares, 4th cycle; open triangles, 5th cycle; closed circles, 6th cycle.
high removal rate of ammonium. The digested sludge contained almost no fatty acids, but 20e30% of ammonium removal in the control experiments (Fig. 1) suggested the existence of carbon compounds in the digested sludge that can be utilized by strain No. 4 presumably derived from high total COD. Available carbon sources for strain No. 4 include low- and highmolecular fatty acids, phenol, methanol, some solvents and amino acids (Akuzawa, E., Master’s thesis, Tokyo Institute of Technology, Tokyo, 2009). The experiment performed here included 20 g of trisodium citrate dihydrate. Generally, the C/N ratio of the intracellular components in microorganisms is 10, indicating that 10 units of carbon are used when 1 unit of N is consumed with the C mainly to synthesize cellular materials. In previous experiments, 30e40% of ammonium was reduced to nitrogen gas by strain No. 4 (4). Assuming a similar level of denitrification in these experiments, 0.6e0.7 g-N/l were used for cell synthesis, indicating that 6e7 g-C/l was required. When 20e30% of carbon is available from the digested sludge, approximately 5 g-C/l should be supplied from outside. As 20 g of trisodium citrate dihydrate contained 16 g-C/l, 6 g of trisodium citrate dihydrate is sufficient to enable the complete removal of 1 g-N/l. Actually, when only 7 g of trisodium citrate dihydrate (C/N ¼ 5) was added, almost 100% removal of ammonium was attained at similar removal rates (data not shown). Concerning carbon requirement in strain No. 4, the conventional denitrification process using methanol is compared by using the following reaction. þ NO 3 þ 1:08CH3 OH þ H /0:065C5 H7 NO2 þ 0:47N2 þ 0:76CO2
þ 2:44H2 O
ð1Þ
The C/N ratio in this reaction is 2. The strain No. 4 process demanded C/N ratio 5. In this point, strain No. 4 process is disadvantageous. However, as a total system, strain No. 4 process will be advantageous over the conventional process in that no dilution of high strength of wastewater is required, only single reactor with compact size is needed and significantly high removal rate is possible when less expensive carbon sources from waste or unused resources are available. Under these conditions, this system can achieve efficient ammonium removal. The nitrification rates in the conventional system are mostly reported to be less than 0.01 kg-N/m3/day (1,2), and the immobilization of the cell mass increased this rate by a factor of 10 (3). However, this system remained vulnerable to high-strength carbon or nitrogen sources. Higher ammonium removal rates have been reported using the anammox method, including 0.96 kg-N/m3/day in the SHARONanammox process (15), 2.3 kg-N/m3/day in a fluidized bed using synthetic medium (14) and more than 4 kg-N/m3/day for gel pellets of anammox biomass (11). The problems with this method include that it is time-consuming to produce sufficient amounts of biomass,
a long period of time is required to stabilize the system, and rapid recovery of the system once inefficient removal occurs is difficult. On the other hand, it is easy to cultivate strain No. 4 in a synthetic medium with a doubling time of 2e3 h and it is easy to immobilize the cells on different immobilization carriers (Joo, H.S., Ph.D. thesis, Tokyo Institute of Technology, Tokyo, 2005). When cultured strain No. 4 cells were stored at 4 C, high activity was maintained for several months, and the cells remained tolerant to high osmotic pressure (data not shown). Recently, many bacteria have been found to exhibit heterotrophic nitrification and aerobic denitrification, such as Rhodococcus sp. (5), Diaphorobacter sp. (6), Bacillus sp. (7), Bacillus methylotrophicus (8), A. faecalis (9) and Pseudomonas sp. (10). Although these bacteria were cultivated in the initial ammonium concentration range from 30 to 300 mg/l and the removal rates were in the range of 0.08e0.6 kg-N/m3/day, these bacteria hold significant potential for ammonium treatment. Although strain No. 4 is not osmophilic, the cells were able to achieve ammonium removal under high saline conditions. In our basic experiment, strain No. 4 was found to synthesize the osmoprotectant, hydroxyectoine during the lag time when the cells were exposed to high salt concentration (Tsunoda, H., Master’s thesis, Tokyo Institute of Technology, Tokyo, 2007) as shown in Fig. 7. As most microorganisms are vulnerable to wastewater with high saline concentrations or highstrength solvents due to the resulting high osmotic pressure, strain No. 4 is able to effectively remove ammonium under such conditions after a certain acclimation period. Thus, this new system can remove highstrength ammonium from marine aquaculture wastewater. Another advantage of strain No. 4 is that the cells of strain No. 4 can suppress the growth of plant pathogens, as we previously reported (17,18). This indicates that the excess sludge of strain No. 4 generated in the treatment facility can be reutilized in farmland to protect against plant diseases. In relatively small-scale reactors like this jar fermentor, oxygen supply capacity is lower than that in practical large-scale reactors. The power requirement in wastewater treatment is one of important factors considered to be in operation. Thus, DO level in large scale reactors can be maintained at lower agitation speeds and the power requirement for strain No. 4 will be almost equivalent to that in conventional aerobic nitrification process. The difference in delay of ammonium removal was observed between 20 C and 30 C after shutdown experiments (Figs. 2 and 4). The exact reason of difference of delay between 20 C and 30 C is not clear. However, the atmospheric temperature during 30 C operation was higher than that at 20 C and the shutdown period at 30 C was 2 days, but that at 20 C was 5 days. These factors may be reflected in the delay of the recovery at 20 C. The next step in this research is to perform continuous treatment using high-strength ammonium (2000e3000 mg-N/l) without dilution of the original samples using suspended or immobilized strain No. 4 cells. ACKNOWLEDGMENTS The authors thank Yokohama Municipal Sewage Treatment Center (Yokohama, Japan) for supplying the digested sludge and valuable information on the sludge. References 1. Kuai, L. and Verstraete, W.: Ammonium removal by the oxygen-limited autotrophic nitrification-denitrification system, Appl. Environ. Microbiol., 64, 4500e4506 (1998). 2. Franco-Rivera, A., Paniagua-Michel, J., and Zamora-Castro, J.: Characterization and performance of constructed nitrifying biofilms during nitrogen bioremediation of a wastewater effluent, J. Ind. Microbiol. Biotechnol., 34, 279e287 (2007).
VOL. 117, 2014 3. Rostron, W. M., Stuckey, D. C., and Young, A. A.: Nitrification of high-strength ammonia wastewater: comparative study of immobilized media, Water Res., 35, 1169e1178 (2001). 4. Joo, H. S., Hirai, M., and Shoda, M.: Piggery wastewater treatment using Alcaligenes faecalis strain No. 4 with heterotrophic nitrification and aerobic denitrification, Water Res., 40, 3029e3036 (2006). 5. Chen, P. Z., Li, J., Li, Q. X., Wang, Y. C., Li, S. P., Ren, T. Z., and Wang, L. G.: Simultaneous heterotrophic nitrification and aerobic denitrification by bacterium Rhodococcus sp. CPZ24, Bioresour. Technol., 116, 266e270 (2012). 6. Khardenavis, A. A., Kapley, A., and Purohit, H. J.: Simultaneous nitrification and denitrification by diverse Diaphorobacter sp., Appl. Microbiol. Biotechnol., 77, 403e409 (2007). 7. Yang, X. P., Wang, S. M., Zhang, D. W., and Zhou, L. X.: Isolation and nitrogen removal characteristics of an aerobic heterotrophic nitrifying-denitrifying bacterium, Bacillus subtilis Al, Bioresour. Technol., 102, 854e862 (2011). 8. Zhang, Q. L., Liu, Y., Ai, G. M., Miao, L. L., Zheng, H. Y., and Liu, Z. P.: The characteristics of a novel heterotrophic nitrification-aerobic denitrification bacterium, Bacillus methylotrophicus strain L7, Bioresour. Technol., 108, 35e44 (2012). 9. Zhao, B., An, Q., He, Y. L., and Guo, J. S.: N2O and N2 production during heterotrophic nitrification by Alcaligenes faecalis strain NR, Bioresour. Technol., 116, 379e385 (2012). 10. Zhu, L., Ding, W., Feng, L. J., Kong, Y., Xu, J., and Xu, X. Y.: Isolation of aerobic denitrifiers and characterization for their potential application in the bioremediation of oligotrophic ecosystem, Bioresour. Technol., 108, 1e7 (2012).
AEROBIC DENITRIFICATION OF HIGH-STRENGTH AMMONIUM
741
11. Furukawa, K., Inatomi, Y., Qiao, S., Quan, L., Yamamoto, T., Isaka, K., and Sumio, T.: Innovative treatment system for digester liquor using anammox process, Bioresour. Technol., 100, 5437e5443 (2009). 12. Molinuevo, B., Garcia, M. C., Karakashev, D., and Angelidaki, I.: Anammox for ammonia removal from pig manure effluents: effect of organic matter content on process performance, Bioresour. Technol., 100, 2171e2175 (2009). 13. Strous, M., van Gerven, E., Zheng, P., Kuenen, J. G., and Jetten, M. S. M.: Ammonium removal from concentrated waste streams with the anaerobic ammonium oxidation (Anammox) process in different reactor configurations, Water Res., 31, 1955e1962 (1997). 14. Van de Graaf, A. A., deBruijin, P., Robertson, L. A., Jetten, M. S. M., and Kuenen, J. G.: Autotrophic growth of anaerobic ammonium-oxidizing microorganisms in a fluidized bed reactor, Microbiology, 142, 2187e2196 (1996). 15. Van Dongen, U., Jetten, M. S. M., and van Loosdrecht, M. C. M.: The SHARONAnammox process for treatment of ammonium rich wastewater, Water Sci. Technol., 44, 153e160 (2001). 16. Joo, H. S., Hirai, M., and Shoda, M.: Nitrification and denitrification in highstrength ammonium by Alcaligenes faecalis, Biotechnol. Lett., 27, 773e778 (2005). 17. Honda, N., Hirai, M., Ano, T., and Shoda, M.: Antifungal effect of a heterotrophic nitrifier Alcaligenes faecalis, Biotechnol. Lett., 20, 703e706 (1998). 18. Honda, N., Hirai, M., Ano, T., and Shoda, M.: Control of tomato damping-off caused by Rhizoctonia solani by the heterotrophic nitrifier Alcaligenes faecalis and its product, hydroxylamine, Ann. Phytopathol. Soc. Japan, 65, 153e162 (1999).