- Email: [email protected]
Denitrification of drinking water in a two-stage membrane bioreactor by using immobilized biomass
Bioresource Technology 128 (2013) 804–808
Contents lists available at SciVerse ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Short Communication
Denitrification of drinking water in a two-stage membrane bioreactor by using immobilized biomass Matjazˇ Ravnjak a,b, Janez Vrtovšek b, Albin Pintar b,⇑ a b
HTZ Velenje, I.P., d.o.o., Partizanska cesta 78, SI-3320 Velenje, Slovenia Laboratory for Environmental Sciences and Engineering, National Institute of Chemistry, Hajdrihova 19, SI-1001 Ljubljana, Slovenia
h i g h l i g h t s " Nitrate removal in a two-stage anoxic/oxic biofilm membrane bioreactor. " Specific biocarriers enable immobilization of efficient and long-lasting microbiota. " High nitrate conversions were obtained, without the formation of by-products. " Shear forces induced by collisions minimize cake formation on membrane surface.
a r t i c l e
i n f o
Article history: Received 7 March 2012 Received in revised form 11 October 2012 Accepted 12 October 2012 Available online 23 October 2012 Keywords: Biocarrier Biofilm Denitrification Drinking water Membrane bioreactor
a b s t r a c t Nitrate removal from polluted groundwater was investigated in a two-stage anoxic/oxic biofilm membrane bioreactor. The process was carried out with ethanol as a carbon source (corresponding C/ N ratio of 1.4–2.5) and commercially available Biocontact-N biocarriers (Nisshinbo, Japan) to enable immobilization of highly efficient and long-lasting microbiota. At a residence time of the liquid phase equal to 2.5 h, nitrate conversions higher than 99% were obtained without the formation of nitrite and ammonium ions. The concentration of total organic carbon in the reactor discharge was very similar to the content of organic matter in tap water. The biocarriers minimized the occurrence of suspended filamentous bacteria, and the utilization of increased shear force facilitated collisions of floating biocarrier particles with the outer membrane surface, preventing membrane fouling and resulting in stable operation of the system for 40 days. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Groundwater can become contaminated with nitrates due to intensive agricultural activities and uncontrolled use of fertilizers. Due to the fact that high nitrate concentrations in drinking water sources represent a direct threat to human health, strict water standards for drinking water have been established worldwide. For example, European directive 98/83/EC sets maximum concentrations for nitrate and nitrite ions at 50 and 0.5 mg/L, respectively, for drinking water intended for human consumption. Ion exchange, reverse osmosis and electro-dialysis fail to completely eliminate nitrate ions and yield concentrated waste brines that require further treatment or disposal (McAdam and Judd, 2006). Heterotrophic biological denitrification of water containing nitrates is feasible, but may produce effluents with high microbial counts unless a final stage for the removal of microbes is included ⇑ Corresponding author. Tel.: +386 1 47 60 237; fax: +386 1 47 60 460. E-mail address: [email protected] (A. Pintar). 0960-8524/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2012.10.055
followed by chlorination or UV treatment (McAdam and Judd, 2007). The other disadvantage of biological denitrification process is the requirement for additional carbon source in order to achieve appropriate nitrate reduction, a fact that can lead to a need to eliminate residual organic carbon content in the effluent. As an alternative, membrane processes have been employed (McAdam and Judd, 2007, 2008). Introduction of a two-stage (anoxic/oxic) process with the membrane module inserted in the oxic (aerated) part could solve the problems connected with the occurrence of nitrite ions as well as bioparticles and organic carbon residuals in treated water. This type of membrane bioreactor (MBR) with suspended biomass has already been used for removal of nitrate from industrial wastewater (Shen et al., 2009), for simultaneous carbon, nitrogen and phosphorus removal from synthetic wastewater (Fu et al., 2009) as well as for nitrate removal from polluted surface water (Buttiglieri et al., 2005). The introduction of biocarriers into the MBR system can have beneficial effects on denitrification kinetics (Vrtovšek and Roš, 2006; Yang et al., 2009; Ivanovic and Leiknes, 2011) and membrane operation cycles, due to a reduction
M. Ravnjak et al. / Bioresource Technology 128 (2013) 804–808
in membrane fouling (Arabi and Nakhla, 2009; Ivanovic and Leiknes, 2012). The objective of the present study was to investigate the overall performance of an anoxic/oxic biofilm MBR (BMBR) system, employing commercially available biomass carriers for the removal of nitrate from polluted groundwater. Efficiency of biological processes as well as the membrane module performance was monitored at two nitrate loadings and C/N ratios.
2. Methods 2.1. Experimental set-up and operating conditions The experimental set-up used in the biological denitrification tests is schematically illustrated in Fig. 1. The main component of the system is a reactor made of Plexiglas divided into two compartments (anoxic and oxic) with a perforated baffle. The volumes of the anoxic and oxic parts were 5 and 10 L, respectively. Peristaltic pumps (Masterflex) were used for feeding the influent (2.0 L/h) and permeation of the effluent through the membrane module. An ultrasonic lever sensor (Flowline, model Echoswitch II) connected to a peristaltic pump controlled the liquid level above the membrane module in the oxic part of the reactor. Aerobic conditions in the fully mixed oxic stage were provided by continuous aeration (500 L/h) using a porous flexible plastic tube, connected to an air compressor and air flow meter. Intensive aeration also managed to efficiently mix the reactor content (liquid phase and biocarriers). A mechanical stirrer (200 rpm) equipped with a Visco JetÒ impeller (Heidolph) was used for efficient mixing of the anoxic stage content. A flat-sheet membrane (cartridge type 203, Kubota Ltd., Japan), made of chlorinated polyethylene with a nominal pore size of 0.4 lm and an effective surface area of 0.3 m2, was submerged in the oxic part of the reactor unit. The anoxic stage of the reactor unit was filled with 2 L of polyurethane-based Biocontact-N carriers (Nisshinbo Chemical, Japan) with a nominal pore size of 5–10 lm. The oxic stage was filled with 3 L of polyethylene-based Mutag Biochip carriers (Umwelttechnologie AG, Germany) with 3000 m2/m3 effective specific surface area. The BMBR system was inoculated with microbial suspension extracted from a local municipal sewage water treatment plant. The recorded adaptation period was approximately 2 weeks. Nitrate-containing water (influent) was prepared from tap water and potassium nitrate (p.a. grade, Merck). The feed concentrations of nitrate ions were either 70 or 150 mg/L. Potassium phosphate (K2HPO4, p.a. grade, Merck) served as source of P (1.0 mg/L P). The P/N ratio was in the range of 0.03–0.06. The
805
influent solution was kept in a separate 100-L tank, from which it was fed into the anoxic stage of the BMBR system. Ethanol (p.a. grade, Aldrich) was used as a carbon source. The solution (0.5%) was prepared daily and kept in a separate 10-L tank, from which it was precisely fed into the anoxic stage of the BMBR system by means of a multi-channel peristaltic pump (Ismatec, model IPC). As alkalinity is produced in the biological denitrification process, i.e. during the transformation of liquid-dissolved nitrate ions to gaseous nitrogen, its level was maintained in the range of 220– 270 mg/L CaCO3 with the addition of HCl. An aqueous solution of approximately 0.012 M HCl (p.a. grade, Aldrich), prepared daily, was kept in a separate 10-L tank, from which it was fed into the anoxic stage of the BMBR system. All experiments were performed at a constant temperature of 20 ± 2 °C. 2.2. Analytical methods On-line measurements of pH value, dissolved oxygen (DO) and oxidation–reduction potential (ORP) in the anoxic and oxic stages of the BMBR system were performed during the experiments. Samples withdrawn daily from the anoxic bioreactor effluent and the membrane module permeate were analyzed using standardized ion chromatography methods (Dionex, model DX-120). Gas-phase samples from the headspace above the anoxic stage were analyzed for N2O by means of a gas chromatograph (Agilent, model 7890A) equipped with a PoraPlot Q capillary column and TCD. Alkalinity measurements were made by employing potentiometric titrations of aqueous-phase samples using a standardized solution of 0.1 M HCl and a T50 titrator (Mettler Toledo). The total amount of organic substances in samples from the anoxic reactor effluent and membrane module permeate was determined by measuring the total organic carbon (TOC) content. The samples were filtered through a 0.2-lm syringe-type filter before measurements in order to eliminate the contribution of biomass. TOC content was determined applying a high-temperature catalytic oxidation (HTCO) method carried out at 750 °C by using an advanced TOC analyzer (Teledyne Tekmar, model Torch) equipped with a high-pressure NDIR detector, by subtracting the measured inorganic carbon (IC) from the measured total carbon (TC) content. Microbiological quality of water from the anoxic reactor effluent, aerobic reactor effluent and membrane module permeate was determined according to the ISO 6222:1999 standard for enumeration of culturable microorganism on nutrient agar. Turbidity of the permeate was measured according to the EN ISO 7027 standard using a Spectroquant photometer (Merck, model NOVA 60A). 3. Results and discussion 3.1. Nitrate removal
Fig. 1. Schematic display of the experimental set-up.
Maximal treatment capability of the BMBR system with the influent nitrate concentration of 150 mg/L and target effluent nitrate concentration <2 mg/L was determined first. A constant influent C/N ratio of 2.5 with the adequate flow rate of aqueous ethanol solution was maintained in this set of experiments. In the start-up period that lasted approximately 30 days, the BMBR system was adapted to the operating conditions. An appropriate flow rate of diluted aqueous HCl solution fed to the system in order to keep alkalinity in the anoxic stage at the influent level was also determined in this period.
806
M. Ravnjak et al. / Bioresource Technology 128 (2013) 804–808
Concentrations of nitrate and nitrite ions measured in withdrawn samples from the anoxic stage of the BMBR system in the 30-day period of continuous operation are presented in Fig. 2a. The corresponding properties and composition (average values) of the effluent from the anoxic stage are listed in Table 1. Similarly, nitrate-spiked tap water with an adjusted inlet nitrate concentration of 20 mg/L above the maximum allowed value was treated in the same BMBR system in the second set of experiments. The results obtained after determining an optimal influent C/N ratio in the 30-day start-up period for the target effluent nitrate concentration of 10 mg/L as well as an adequate flow rate of aqueous HCl solution are presented in Fig. 2b and Table 1. In the given range of operating and reaction conditions, the BMBR system operated in the integral regime (i.e. at high nitrate conversions) and no decrease in activity was observed regarding the removal of nitrate as a function of time. A further increase in the stirring rate resulted in no measurable increase in nitrate disappearance rates in both sets of experiments, which implies that a stirrer speed of 200 rpm was sufficient to avoid an influence of external solid–liquid mass-transfer on reactor performance. In both sets of experiments, the concentration of N2O in the headspace above the anoxic stage was below 0.1 vol.%, as determined by GC analysis. The results obtained during the 1st set of experiments (influent nitrate concentration equals to 150 mg/L) show that when the organic carbon was present in excess, nitrate conversions higher than 99% could be achieved in the fully mixed anoxic stage at a hydraulic retention time (HRT) of 2.5 h. The nitrite concentrations were below the level of detection of 0.3 mg/L. Similar results to the ones illustrated in Fig. 2 were achieved at much longer HRTs of 9 h in an anoxic reactor with suspended biomass (Buttiglieri et al., 2005), or at HRT of 8 h in an anoxic biofilm reactor (Wang et al., 2009). The performance of the biofilm anoxic stage of the BMBR system in the present study was comparable to that of a fluidized-bed biofilm reactor (Rabah and Dahab, 2004) which, however, required much higher specific energy input to obtain the same level of nitrate conversion. The results of a detailed study carried out by Nohuglu et al. (2002) in a suspended biomass system indicated that nitrite concentration is the limiting factor in determining the duration of a denitrification process in order to obtain a water stream of suitable quality. Ethanol is a suitable carbon source since a wide spectrum of microorganisms can utilize it (Nyberg et al., 1996; Aspegren et al., 1998; McAdam and Judd, 2007). Furthermore, mechanical
a
3.2. Organic carbon removal High concentrations of organic matter were typically found in groundwater treated by highly loaded biofilm denitrification systems (Rabah and Dahab, 2004; Vrtovšek and Roš, 2006; Wang et al., 2009). Although TOC content is usually not a standard limiting factor, the remaining ethanol or methanol content in treated water of more than 5 mg/L (expressed as TOC) is not acceptable. In this respect, an optimal (usually much lower) influent C/N ratio must be defined in a single-stage denitrification system in order to obtain a suitable organic carbon concentration in treated water (Wang et al., 2009). TOC conversions obtained during the treatment of water in the anoxic stage of the BMBR system were in the range of 65–75%. After the effluent from the anoxic stage was sequentially treated in the intensively aerated and well-mixed oxic stage, TOC concentrations were further reduced to the influent level in both sets of experiments (Fig. 3). At the same time, treated water was simultaneously saturated with oxygen. A further increase in air flow rate resulted in no decrease of TOC concentration in the reactor discharge, which confirms that the extent of TOC removal was determined by processes facilitated by microorganisms in the
b
-
NO3 feed concentration = 150 mg/L
10
and textural characteristics of the biocarrier material (BiocontactN) used in the anoxic stage of the BMBR system provided suitable environment for the development of highly efficient and long-lasting biofilm microbiota. Optimal pH conditions for biological denitrification in the range of 7.0–7.5 (Shen et al., 2009; Wang et al., 2009) were obtained with the continuous feed of aqueous HCl. At the same time, alkalinity was kept at the influent level with a minimal increase in the chloride concentration in the treated water of about 4 mg/L. When polluted groundwater with an influent nitrate concentration of 70 mg/L and a C/N ratio of 1.4 was treated in the BMBR system, an optimum was found at which the effluent nitrate and nitrite concentrations were well below the limits of 50 and 0.5 mg/L, respectively (Fig. 2b). It is evident that ethanol consumption was significantly minimized without altering the nitrate concentration in the reactor discharge. For comparison, in a suspended biomass MBR system investigated by McAdam and Judd (2007), an optimal C/N ratio of 1.52 was determined when ethanol was used as a carbon source. At an acceptable HRT of 2.5 h, the quality of the anoxic stage effluent was very similar to the quality of tap water, except for DO and ORP values as well as TOC concentrations.
-
NO3 feed concentration = 70 mg/L
10
10
8
8
8
6
6
4
4
10
4
2
2
-
-
4
-
NO3
2
2
-
NO2 0 30
35
40
45
Time [days]
50
55
0 60
NO2 [mg/L]
6
-
6
-
NO3 [mg/L]
8
NO3 [mg/L]
NO3 NO2
NO2 [mg/L]
-
0 30
35
40
45
50
55
0 60
Time [days]
Fig. 2. Nitrate and nitrite concentrations measured in the effluent from the anoxic stage at the influent nitrate concentration of (a) 150 mg/L and (b) 70 mg/L.
807
M. Ravnjak et al. / Bioresource Technology 128 (2013) 804–808 Table 1 Average properties and composition of anoxic and oxic stage effluents obtained during the experiments with different influent nitrate concentrations. Parameter
Influent nitrate conc. (mg/L) Anoxic stage
Temperature (°C) pH value Dissolved oxygen (DO) (mg/L) Oxidation–reduction potential (ORP) (mV) Hardness (mg/L) (as CaCO3) Alkalinity (mg/L) (as CaCO3) Conductivity (lS/cm) Total organic carbon (TOC) (mg/L) NO3 (mg/L) NO2 (mg/L) Cl (mg/L) SO42 (mg/L) Na+ (mg/L) K+ (mg/L) NH4+ (mg/L) Ca2+ (mg/L) Mg2+ (mg/L)
a
70
150
70
19.6 7.2 0.1 115 171 233 515 21 1.0 <0.3 14.3 18.2 3.4 102.5 <0.3 51.5 13.2
19.8 7.1 0.1 103 143 207 524 10 9.0 <0.3 12.6 17.9 5.1 73.8 <0.3 49.8 11.7
19.7 7.2 7.3 230 116 225 512 5.0 1.0 <0.3 13.7 17.2 4.2 98.6 <0.3 50.7 12.3
19.9 7.1 7.8 255 143 212 543 5.0 9.0 <0.3 12.6 17.9 5.3 70.7 <0.3 47.9 10.6
b
-
NO3 feed concentration = 150 mg/L
40
Oxic stage
150
-
NO3 feed concentration = 70 mg/L
40
Effluentanoxic
Effluentanoxic
Effluentoxic
Influent
30
TOC [mg/L]
TOC [mg/L]
Effluentoxic
Influent
30
20
20
10
10
0
0 30
35
40
45
50
55
60
Time [days]
30
35
40
45
50
55
60
Time [days]
Fig. 3. Total organic carbon (TOC) concentrations measured in effluents from the anoxic and oxic stages at the influent nitrate concentration of (a) 150 mg/L and (b) 70 mg/L.
immobilized biomass. The properties and composition of water discharged from the BMBR system were very similar to those of tap water. The results also showed that mechanical and surface characteristics of the applied biocarrier material (Mutag Biochip) used in the oxic stage, provided suitable conditions for the development of stable and efficient biofilm microbiota even at very low organic loadings fed from the anoxic stage to the aerobic compartment. 3.3. Membrane module operation Reducing suspended solids concentration with the introduction of biocarriers in the MBR system should help to reduce the effect of extracellular polymeric substances (EPS) on membrane fouling. However, as reported by Yang et al. (2009) and confirmed by Arabi and Nakhla (2009), the use of biocarriers unexpectedly increased the fouling rate of a membrane surface compared to MBRs operated without biomass supports, most probably due to the occurrence of filamentous bacteria, which produced more EPS compared to floc-forming bacteria. Due to specific mechanical properties of the biocarrier material (Biocontact-N) used in the anoxic stage of the BMBR system, the concentration of suspended filamentous bacteria measured in the
oxic stage was so low that it did not affect membrane module operation due to fouling. Correspondingly, a minimum transmembrane pressure (TMP) increase of only 2.6 kPa (20% of initial value) within the time period of 3 months was observed. At the same time, collisions of floating biocarrier particles (Mutag Biochip) with the outer membrane surface in the aerobic compartment considerably increased the shear force, which was otherwise generated by aeration. In the present study, the number of collisions was esti1 mated to be equal to 22; 000 min m2membrane . In this way, deposition of bacteria on the membrane surface (cake formation) was prevented. As a result, the membrane operational period (i.e. a period between subsequent membrane hydrocleaning steps) of 40 days was achieved. In a 3-month experimental period, the membrane module was operated without applying any chemical cleaning procedures. During the 2nd set of experiments (influent nitrate concentration equals to 70 mg/L), the average concentration of suspended microorganisms in the oxic stage was 480,000 CFU/mL, which was reduced by filtration through the membrane module to an average of 5500 CFU/mL in the permeate (final effluent). Rather poor separation efficiency of suspended solids was achieved in the experiments, because a membrane with relatively large average pore size of 0.4 lm was used. It is believed that advanced
808
M. Ravnjak et al. / Bioresource Technology 128 (2013) 804–808
membrane modules with an average pore size <0.1 lm, which could be operated with an automatic back-washing process, might provide much better microbiological quality of water treated in large-scale BMBR plants. However, this issue remains to be investigated. 4. Conclusions The obtained results demonstrate that the two-stage suspended biofilm MBR system exhibits a great potential for the treatment of groundwater polluted with nitrate ions, without any occurrence of nitrite and ammonium ions in treated water. In comparison with suspended biomass MBR systems, much higher denitrification rates were achieved in the anoxic stage of the BMBR system. At the same time, the concentration of residual organic carbon was efficiently controlled in the oxic stage, in which treated water was resaturated with oxygen. It is believed that the achieved 40day operation period of the BMBR system could be further improved by optimizing biosolids/water separation processes. References Arabi, S., Nakhla, G., 2009. Characterization of foulants in conventional and simultaneous nitrification and denitrification membrane bioreactors. Separation and Purification Technology 69 (2), 153–160. Aspegren, H., Nyberg, U., Andersson, B., Gotthardsson, S., Jansen la Cour, J., 1998. Post denitrification in a moving bed biofilm reactor process. Water Science and Technology 38 (1), 31–38.
Buttiglieri, G., Malp, F., Daveri, E., Melchiori, M., Nieman, H., Ligthart, J., 2005. Denitrification of drinking water sources by advanced biological treatment using a membrane bioreactor. Desalination 178, 211–218. Fu, Z., Yang, F., An, Y., Xue, Y., 2009. Simultaneous nitrification and denitrification coupled with phosphorus removal in an modified anoxic/oxic-membrane bioreactor (A/O-MBR). Biochemical Engineering Journal 43 (2), 191–196. Ivanovic, I., Leiknes, T.O., 2011. Impact of denitrification on the performance of a biofilm-MBR (BF-MBR). Desalination 283, 100–105. Ivanovic, I., Leiknes, T.O., 2012. The biofilm membrane reactor (BF-MBR) – a review. Desalination and Water Treatment 37, 288–295. McAdam, E.J., Judd, S.J., 2006. A review of membrane bioreactor potential for nitrate removal from drinking water. Desalination 196, 135–148. McAdam, E.J., Judd, S.J., 2007. Denitrification from drinking water using a membrane bioreactor: chemical and biochemical feasibility. Water Research 41 (18), 4242–4250. McAdam, E.J., Judd, S.J., 2008. Immersed membrane bioreactors for nitrate removal from drinking water: cost and feasibility. Desalination 231, 52–60. Nohuglu, A., Turgay, P., Ergun, Y., 2002. Drinking water denitrification by a membrane bio-reactor. Water Research 36, 1155–1166. Nyberg, U., Andersson, B., Aspegren, H., 1996. Long-term experiences with external carbon sources for nitrogen removal. Water Science and Technology 33 (12), 109–116. Rabah, F.K.J., Dahab, M.F., 2004. Nitrate removal characteristics of high performance fluidized-bed biofilm reactors. Water Research 38 (17), 3719–3728. Shen, J., He, R., Han, W., Sun, X., Li, J., Wang, L., 2009. Biological denitrification of high-nitrate wastewater in a modified anoxic/oxic-membrane bioreactor (A/OMBR). Journal of Hazardous Materials 172 (2–3), 595–600. Vrtovšek, J., Roš, M., 2006. Denitrification of groundwater in the biofilm reactor with a specific biomass support material. Acta Chimica Slovenica 53 (3), 396–400. Wang, Q., Feng, C., Zhao, Y., Hao, C., 2009. Denitrification of nitrate contaminated groundwater with a fiber-based biofilm reactor. Bioresource Technology 100 (7), 2223–2227. Yang, S., Yang, F., Fu, Z., Lei, R., 2009. Comparison between a moving bed membrane bioreactor and a conventional membrane bioreactor on organic carbon and nitrogen removal. Bioresource Technology 100 (1), 2369–2374.
Contents lists available at SciVerse ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Short Communication
Denitrification of drinking water in a two-stage membrane bioreactor by using immobilized biomass Matjazˇ Ravnjak a,b, Janez Vrtovšek b, Albin Pintar b,⇑ a b
HTZ Velenje, I.P., d.o.o., Partizanska cesta 78, SI-3320 Velenje, Slovenia Laboratory for Environmental Sciences and Engineering, National Institute of Chemistry, Hajdrihova 19, SI-1001 Ljubljana, Slovenia
h i g h l i g h t s " Nitrate removal in a two-stage anoxic/oxic biofilm membrane bioreactor. " Specific biocarriers enable immobilization of efficient and long-lasting microbiota. " High nitrate conversions were obtained, without the formation of by-products. " Shear forces induced by collisions minimize cake formation on membrane surface.
a r t i c l e
i n f o
Article history: Received 7 March 2012 Received in revised form 11 October 2012 Accepted 12 October 2012 Available online 23 October 2012 Keywords: Biocarrier Biofilm Denitrification Drinking water Membrane bioreactor
a b s t r a c t Nitrate removal from polluted groundwater was investigated in a two-stage anoxic/oxic biofilm membrane bioreactor. The process was carried out with ethanol as a carbon source (corresponding C/ N ratio of 1.4–2.5) and commercially available Biocontact-N biocarriers (Nisshinbo, Japan) to enable immobilization of highly efficient and long-lasting microbiota. At a residence time of the liquid phase equal to 2.5 h, nitrate conversions higher than 99% were obtained without the formation of nitrite and ammonium ions. The concentration of total organic carbon in the reactor discharge was very similar to the content of organic matter in tap water. The biocarriers minimized the occurrence of suspended filamentous bacteria, and the utilization of increased shear force facilitated collisions of floating biocarrier particles with the outer membrane surface, preventing membrane fouling and resulting in stable operation of the system for 40 days. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Groundwater can become contaminated with nitrates due to intensive agricultural activities and uncontrolled use of fertilizers. Due to the fact that high nitrate concentrations in drinking water sources represent a direct threat to human health, strict water standards for drinking water have been established worldwide. For example, European directive 98/83/EC sets maximum concentrations for nitrate and nitrite ions at 50 and 0.5 mg/L, respectively, for drinking water intended for human consumption. Ion exchange, reverse osmosis and electro-dialysis fail to completely eliminate nitrate ions and yield concentrated waste brines that require further treatment or disposal (McAdam and Judd, 2006). Heterotrophic biological denitrification of water containing nitrates is feasible, but may produce effluents with high microbial counts unless a final stage for the removal of microbes is included ⇑ Corresponding author. Tel.: +386 1 47 60 237; fax: +386 1 47 60 460. E-mail address: [email protected] (A. Pintar). 0960-8524/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2012.10.055
followed by chlorination or UV treatment (McAdam and Judd, 2007). The other disadvantage of biological denitrification process is the requirement for additional carbon source in order to achieve appropriate nitrate reduction, a fact that can lead to a need to eliminate residual organic carbon content in the effluent. As an alternative, membrane processes have been employed (McAdam and Judd, 2007, 2008). Introduction of a two-stage (anoxic/oxic) process with the membrane module inserted in the oxic (aerated) part could solve the problems connected with the occurrence of nitrite ions as well as bioparticles and organic carbon residuals in treated water. This type of membrane bioreactor (MBR) with suspended biomass has already been used for removal of nitrate from industrial wastewater (Shen et al., 2009), for simultaneous carbon, nitrogen and phosphorus removal from synthetic wastewater (Fu et al., 2009) as well as for nitrate removal from polluted surface water (Buttiglieri et al., 2005). The introduction of biocarriers into the MBR system can have beneficial effects on denitrification kinetics (Vrtovšek and Roš, 2006; Yang et al., 2009; Ivanovic and Leiknes, 2011) and membrane operation cycles, due to a reduction
M. Ravnjak et al. / Bioresource Technology 128 (2013) 804–808
in membrane fouling (Arabi and Nakhla, 2009; Ivanovic and Leiknes, 2012). The objective of the present study was to investigate the overall performance of an anoxic/oxic biofilm MBR (BMBR) system, employing commercially available biomass carriers for the removal of nitrate from polluted groundwater. Efficiency of biological processes as well as the membrane module performance was monitored at two nitrate loadings and C/N ratios.
2. Methods 2.1. Experimental set-up and operating conditions The experimental set-up used in the biological denitrification tests is schematically illustrated in Fig. 1. The main component of the system is a reactor made of Plexiglas divided into two compartments (anoxic and oxic) with a perforated baffle. The volumes of the anoxic and oxic parts were 5 and 10 L, respectively. Peristaltic pumps (Masterflex) were used for feeding the influent (2.0 L/h) and permeation of the effluent through the membrane module. An ultrasonic lever sensor (Flowline, model Echoswitch II) connected to a peristaltic pump controlled the liquid level above the membrane module in the oxic part of the reactor. Aerobic conditions in the fully mixed oxic stage were provided by continuous aeration (500 L/h) using a porous flexible plastic tube, connected to an air compressor and air flow meter. Intensive aeration also managed to efficiently mix the reactor content (liquid phase and biocarriers). A mechanical stirrer (200 rpm) equipped with a Visco JetÒ impeller (Heidolph) was used for efficient mixing of the anoxic stage content. A flat-sheet membrane (cartridge type 203, Kubota Ltd., Japan), made of chlorinated polyethylene with a nominal pore size of 0.4 lm and an effective surface area of 0.3 m2, was submerged in the oxic part of the reactor unit. The anoxic stage of the reactor unit was filled with 2 L of polyurethane-based Biocontact-N carriers (Nisshinbo Chemical, Japan) with a nominal pore size of 5–10 lm. The oxic stage was filled with 3 L of polyethylene-based Mutag Biochip carriers (Umwelttechnologie AG, Germany) with 3000 m2/m3 effective specific surface area. The BMBR system was inoculated with microbial suspension extracted from a local municipal sewage water treatment plant. The recorded adaptation period was approximately 2 weeks. Nitrate-containing water (influent) was prepared from tap water and potassium nitrate (p.a. grade, Merck). The feed concentrations of nitrate ions were either 70 or 150 mg/L. Potassium phosphate (K2HPO4, p.a. grade, Merck) served as source of P (1.0 mg/L P). The P/N ratio was in the range of 0.03–0.06. The
805
influent solution was kept in a separate 100-L tank, from which it was fed into the anoxic stage of the BMBR system. Ethanol (p.a. grade, Aldrich) was used as a carbon source. The solution (0.5%) was prepared daily and kept in a separate 10-L tank, from which it was precisely fed into the anoxic stage of the BMBR system by means of a multi-channel peristaltic pump (Ismatec, model IPC). As alkalinity is produced in the biological denitrification process, i.e. during the transformation of liquid-dissolved nitrate ions to gaseous nitrogen, its level was maintained in the range of 220– 270 mg/L CaCO3 with the addition of HCl. An aqueous solution of approximately 0.012 M HCl (p.a. grade, Aldrich), prepared daily, was kept in a separate 10-L tank, from which it was fed into the anoxic stage of the BMBR system. All experiments were performed at a constant temperature of 20 ± 2 °C. 2.2. Analytical methods On-line measurements of pH value, dissolved oxygen (DO) and oxidation–reduction potential (ORP) in the anoxic and oxic stages of the BMBR system were performed during the experiments. Samples withdrawn daily from the anoxic bioreactor effluent and the membrane module permeate were analyzed using standardized ion chromatography methods (Dionex, model DX-120). Gas-phase samples from the headspace above the anoxic stage were analyzed for N2O by means of a gas chromatograph (Agilent, model 7890A) equipped with a PoraPlot Q capillary column and TCD. Alkalinity measurements were made by employing potentiometric titrations of aqueous-phase samples using a standardized solution of 0.1 M HCl and a T50 titrator (Mettler Toledo). The total amount of organic substances in samples from the anoxic reactor effluent and membrane module permeate was determined by measuring the total organic carbon (TOC) content. The samples were filtered through a 0.2-lm syringe-type filter before measurements in order to eliminate the contribution of biomass. TOC content was determined applying a high-temperature catalytic oxidation (HTCO) method carried out at 750 °C by using an advanced TOC analyzer (Teledyne Tekmar, model Torch) equipped with a high-pressure NDIR detector, by subtracting the measured inorganic carbon (IC) from the measured total carbon (TC) content. Microbiological quality of water from the anoxic reactor effluent, aerobic reactor effluent and membrane module permeate was determined according to the ISO 6222:1999 standard for enumeration of culturable microorganism on nutrient agar. Turbidity of the permeate was measured according to the EN ISO 7027 standard using a Spectroquant photometer (Merck, model NOVA 60A). 3. Results and discussion 3.1. Nitrate removal
Fig. 1. Schematic display of the experimental set-up.
Maximal treatment capability of the BMBR system with the influent nitrate concentration of 150 mg/L and target effluent nitrate concentration <2 mg/L was determined first. A constant influent C/N ratio of 2.5 with the adequate flow rate of aqueous ethanol solution was maintained in this set of experiments. In the start-up period that lasted approximately 30 days, the BMBR system was adapted to the operating conditions. An appropriate flow rate of diluted aqueous HCl solution fed to the system in order to keep alkalinity in the anoxic stage at the influent level was also determined in this period.
806
M. Ravnjak et al. / Bioresource Technology 128 (2013) 804–808
Concentrations of nitrate and nitrite ions measured in withdrawn samples from the anoxic stage of the BMBR system in the 30-day period of continuous operation are presented in Fig. 2a. The corresponding properties and composition (average values) of the effluent from the anoxic stage are listed in Table 1. Similarly, nitrate-spiked tap water with an adjusted inlet nitrate concentration of 20 mg/L above the maximum allowed value was treated in the same BMBR system in the second set of experiments. The results obtained after determining an optimal influent C/N ratio in the 30-day start-up period for the target effluent nitrate concentration of 10 mg/L as well as an adequate flow rate of aqueous HCl solution are presented in Fig. 2b and Table 1. In the given range of operating and reaction conditions, the BMBR system operated in the integral regime (i.e. at high nitrate conversions) and no decrease in activity was observed regarding the removal of nitrate as a function of time. A further increase in the stirring rate resulted in no measurable increase in nitrate disappearance rates in both sets of experiments, which implies that a stirrer speed of 200 rpm was sufficient to avoid an influence of external solid–liquid mass-transfer on reactor performance. In both sets of experiments, the concentration of N2O in the headspace above the anoxic stage was below 0.1 vol.%, as determined by GC analysis. The results obtained during the 1st set of experiments (influent nitrate concentration equals to 150 mg/L) show that when the organic carbon was present in excess, nitrate conversions higher than 99% could be achieved in the fully mixed anoxic stage at a hydraulic retention time (HRT) of 2.5 h. The nitrite concentrations were below the level of detection of 0.3 mg/L. Similar results to the ones illustrated in Fig. 2 were achieved at much longer HRTs of 9 h in an anoxic reactor with suspended biomass (Buttiglieri et al., 2005), or at HRT of 8 h in an anoxic biofilm reactor (Wang et al., 2009). The performance of the biofilm anoxic stage of the BMBR system in the present study was comparable to that of a fluidized-bed biofilm reactor (Rabah and Dahab, 2004) which, however, required much higher specific energy input to obtain the same level of nitrate conversion. The results of a detailed study carried out by Nohuglu et al. (2002) in a suspended biomass system indicated that nitrite concentration is the limiting factor in determining the duration of a denitrification process in order to obtain a water stream of suitable quality. Ethanol is a suitable carbon source since a wide spectrum of microorganisms can utilize it (Nyberg et al., 1996; Aspegren et al., 1998; McAdam and Judd, 2007). Furthermore, mechanical
a
3.2. Organic carbon removal High concentrations of organic matter were typically found in groundwater treated by highly loaded biofilm denitrification systems (Rabah and Dahab, 2004; Vrtovšek and Roš, 2006; Wang et al., 2009). Although TOC content is usually not a standard limiting factor, the remaining ethanol or methanol content in treated water of more than 5 mg/L (expressed as TOC) is not acceptable. In this respect, an optimal (usually much lower) influent C/N ratio must be defined in a single-stage denitrification system in order to obtain a suitable organic carbon concentration in treated water (Wang et al., 2009). TOC conversions obtained during the treatment of water in the anoxic stage of the BMBR system were in the range of 65–75%. After the effluent from the anoxic stage was sequentially treated in the intensively aerated and well-mixed oxic stage, TOC concentrations were further reduced to the influent level in both sets of experiments (Fig. 3). At the same time, treated water was simultaneously saturated with oxygen. A further increase in air flow rate resulted in no decrease of TOC concentration in the reactor discharge, which confirms that the extent of TOC removal was determined by processes facilitated by microorganisms in the
b
-
NO3 feed concentration = 150 mg/L
10
and textural characteristics of the biocarrier material (BiocontactN) used in the anoxic stage of the BMBR system provided suitable environment for the development of highly efficient and long-lasting biofilm microbiota. Optimal pH conditions for biological denitrification in the range of 7.0–7.5 (Shen et al., 2009; Wang et al., 2009) were obtained with the continuous feed of aqueous HCl. At the same time, alkalinity was kept at the influent level with a minimal increase in the chloride concentration in the treated water of about 4 mg/L. When polluted groundwater with an influent nitrate concentration of 70 mg/L and a C/N ratio of 1.4 was treated in the BMBR system, an optimum was found at which the effluent nitrate and nitrite concentrations were well below the limits of 50 and 0.5 mg/L, respectively (Fig. 2b). It is evident that ethanol consumption was significantly minimized without altering the nitrate concentration in the reactor discharge. For comparison, in a suspended biomass MBR system investigated by McAdam and Judd (2007), an optimal C/N ratio of 1.52 was determined when ethanol was used as a carbon source. At an acceptable HRT of 2.5 h, the quality of the anoxic stage effluent was very similar to the quality of tap water, except for DO and ORP values as well as TOC concentrations.
-
NO3 feed concentration = 70 mg/L
10
10
8
8
8
6
6
4
4
10
4
2
2
-
-
4
-
NO3
2
2
-
NO2 0 30
35
40
45
Time [days]
50
55
0 60
NO2 [mg/L]
6
-
6
-
NO3 [mg/L]
8
NO3 [mg/L]
NO3 NO2
NO2 [mg/L]
-
0 30
35
40
45
50
55
0 60
Time [days]
Fig. 2. Nitrate and nitrite concentrations measured in the effluent from the anoxic stage at the influent nitrate concentration of (a) 150 mg/L and (b) 70 mg/L.
807
M. Ravnjak et al. / Bioresource Technology 128 (2013) 804–808 Table 1 Average properties and composition of anoxic and oxic stage effluents obtained during the experiments with different influent nitrate concentrations. Parameter
Influent nitrate conc. (mg/L) Anoxic stage
Temperature (°C) pH value Dissolved oxygen (DO) (mg/L) Oxidation–reduction potential (ORP) (mV) Hardness (mg/L) (as CaCO3) Alkalinity (mg/L) (as CaCO3) Conductivity (lS/cm) Total organic carbon (TOC) (mg/L) NO3 (mg/L) NO2 (mg/L) Cl (mg/L) SO42 (mg/L) Na+ (mg/L) K+ (mg/L) NH4+ (mg/L) Ca2+ (mg/L) Mg2+ (mg/L)
a
70
150
70
19.6 7.2 0.1 115 171 233 515 21 1.0 <0.3 14.3 18.2 3.4 102.5 <0.3 51.5 13.2
19.8 7.1 0.1 103 143 207 524 10 9.0 <0.3 12.6 17.9 5.1 73.8 <0.3 49.8 11.7
19.7 7.2 7.3 230 116 225 512 5.0 1.0 <0.3 13.7 17.2 4.2 98.6 <0.3 50.7 12.3
19.9 7.1 7.8 255 143 212 543 5.0 9.0 <0.3 12.6 17.9 5.3 70.7 <0.3 47.9 10.6
b
-
NO3 feed concentration = 150 mg/L
40
Oxic stage
150
-
NO3 feed concentration = 70 mg/L
40
Effluentanoxic
Effluentanoxic
Effluentoxic
Influent
30
TOC [mg/L]
TOC [mg/L]
Effluentoxic
Influent
30
20
20
10
10
0
0 30
35
40
45
50
55
60
Time [days]
30
35
40
45
50
55
60
Time [days]
Fig. 3. Total organic carbon (TOC) concentrations measured in effluents from the anoxic and oxic stages at the influent nitrate concentration of (a) 150 mg/L and (b) 70 mg/L.
immobilized biomass. The properties and composition of water discharged from the BMBR system were very similar to those of tap water. The results also showed that mechanical and surface characteristics of the applied biocarrier material (Mutag Biochip) used in the oxic stage, provided suitable conditions for the development of stable and efficient biofilm microbiota even at very low organic loadings fed from the anoxic stage to the aerobic compartment. 3.3. Membrane module operation Reducing suspended solids concentration with the introduction of biocarriers in the MBR system should help to reduce the effect of extracellular polymeric substances (EPS) on membrane fouling. However, as reported by Yang et al. (2009) and confirmed by Arabi and Nakhla (2009), the use of biocarriers unexpectedly increased the fouling rate of a membrane surface compared to MBRs operated without biomass supports, most probably due to the occurrence of filamentous bacteria, which produced more EPS compared to floc-forming bacteria. Due to specific mechanical properties of the biocarrier material (Biocontact-N) used in the anoxic stage of the BMBR system, the concentration of suspended filamentous bacteria measured in the
oxic stage was so low that it did not affect membrane module operation due to fouling. Correspondingly, a minimum transmembrane pressure (TMP) increase of only 2.6 kPa (20% of initial value) within the time period of 3 months was observed. At the same time, collisions of floating biocarrier particles (Mutag Biochip) with the outer membrane surface in the aerobic compartment considerably increased the shear force, which was otherwise generated by aeration. In the present study, the number of collisions was esti1 mated to be equal to 22; 000 min m2membrane . In this way, deposition of bacteria on the membrane surface (cake formation) was prevented. As a result, the membrane operational period (i.e. a period between subsequent membrane hydrocleaning steps) of 40 days was achieved. In a 3-month experimental period, the membrane module was operated without applying any chemical cleaning procedures. During the 2nd set of experiments (influent nitrate concentration equals to 70 mg/L), the average concentration of suspended microorganisms in the oxic stage was 480,000 CFU/mL, which was reduced by filtration through the membrane module to an average of 5500 CFU/mL in the permeate (final effluent). Rather poor separation efficiency of suspended solids was achieved in the experiments, because a membrane with relatively large average pore size of 0.4 lm was used. It is believed that advanced
808
M. Ravnjak et al. / Bioresource Technology 128 (2013) 804–808
membrane modules with an average pore size <0.1 lm, which could be operated with an automatic back-washing process, might provide much better microbiological quality of water treated in large-scale BMBR plants. However, this issue remains to be investigated. 4. Conclusions The obtained results demonstrate that the two-stage suspended biofilm MBR system exhibits a great potential for the treatment of groundwater polluted with nitrate ions, without any occurrence of nitrite and ammonium ions in treated water. In comparison with suspended biomass MBR systems, much higher denitrification rates were achieved in the anoxic stage of the BMBR system. At the same time, the concentration of residual organic carbon was efficiently controlled in the oxic stage, in which treated water was resaturated with oxygen. It is believed that the achieved 40day operation period of the BMBR system could be further improved by optimizing biosolids/water separation processes. References Arabi, S., Nakhla, G., 2009. Characterization of foulants in conventional and simultaneous nitrification and denitrification membrane bioreactors. Separation and Purification Technology 69 (2), 153–160. Aspegren, H., Nyberg, U., Andersson, B., Gotthardsson, S., Jansen la Cour, J., 1998. Post denitrification in a moving bed biofilm reactor process. Water Science and Technology 38 (1), 31–38.
Buttiglieri, G., Malp, F., Daveri, E., Melchiori, M., Nieman, H., Ligthart, J., 2005. Denitrification of drinking water sources by advanced biological treatment using a membrane bioreactor. Desalination 178, 211–218. Fu, Z., Yang, F., An, Y., Xue, Y., 2009. Simultaneous nitrification and denitrification coupled with phosphorus removal in an modified anoxic/oxic-membrane bioreactor (A/O-MBR). Biochemical Engineering Journal 43 (2), 191–196. Ivanovic, I., Leiknes, T.O., 2011. Impact of denitrification on the performance of a biofilm-MBR (BF-MBR). Desalination 283, 100–105. Ivanovic, I., Leiknes, T.O., 2012. The biofilm membrane reactor (BF-MBR) – a review. Desalination and Water Treatment 37, 288–295. McAdam, E.J., Judd, S.J., 2006. A review of membrane bioreactor potential for nitrate removal from drinking water. Desalination 196, 135–148. McAdam, E.J., Judd, S.J., 2007. Denitrification from drinking water using a membrane bioreactor: chemical and biochemical feasibility. Water Research 41 (18), 4242–4250. McAdam, E.J., Judd, S.J., 2008. Immersed membrane bioreactors for nitrate removal from drinking water: cost and feasibility. Desalination 231, 52–60. Nohuglu, A., Turgay, P., Ergun, Y., 2002. Drinking water denitrification by a membrane bio-reactor. Water Research 36, 1155–1166. Nyberg, U., Andersson, B., Aspegren, H., 1996. Long-term experiences with external carbon sources for nitrogen removal. Water Science and Technology 33 (12), 109–116. Rabah, F.K.J., Dahab, M.F., 2004. Nitrate removal characteristics of high performance fluidized-bed biofilm reactors. Water Research 38 (17), 3719–3728. Shen, J., He, R., Han, W., Sun, X., Li, J., Wang, L., 2009. Biological denitrification of high-nitrate wastewater in a modified anoxic/oxic-membrane bioreactor (A/OMBR). Journal of Hazardous Materials 172 (2–3), 595–600. Vrtovšek, J., Roš, M., 2006. Denitrification of groundwater in the biofilm reactor with a specific biomass support material. Acta Chimica Slovenica 53 (3), 396–400. Wang, Q., Feng, C., Zhao, Y., Hao, C., 2009. Denitrification of nitrate contaminated groundwater with a fiber-based biofilm reactor. Bioresource Technology 100 (7), 2223–2227. Yang, S., Yang, F., Fu, Z., Lei, R., 2009. Comparison between a moving bed membrane bioreactor and a conventional membrane bioreactor on organic carbon and nitrogen removal. Bioresource Technology 100 (1), 2369–2374.