Operational performance of a submerged membrane bioreactor for reclamation of bath wastewater

Process Biochemistry 40 (2005) 125–130

Operational performance of a submerged membrane bioreactor for reclamation of bath wastewater Rui Liu, Xia Huang∗ , Lvjun Chen, Xianghua Wen, Yi Qian Environment Simulation and Pollution Control State Key Joint Laboratory, Department of Environmental Science and Engineering, Tsinghua University, Beijing 100084, PR China Received 13 August 2003; received in revised form 6 November 2003; accepted 25 November 2003

Abstract Bath wastewater is an ideal wastewater-reclamation source for its large discharge amount, simple pollutant composition and low pollutant content. The feasibility of reclaiming bath wastewater with a membrane bioreactor was investigated in a pilot plant of 10 m3 per day at an organic load of 0.50–1.85 kg-COD/(m3 per day). The operation was continued for 216 days without sludge discharge and chemical cleaning of membrane modules. The quality of the effluent obtained met the wastewater reclamation standard of China, with COD < 40 mg/l, NH4 + −N <0.5 mg/l and anionic surfactant (AS) <0.2 mg/l. Biological treatment removed most pollutants in the influent, degrading 34–85% of COD and 98% of anionic surfactant. The membrane separation balanced the unstable biological treatment of COD but demonstrated no contribution to anionic surfactant removal. Inorganic substances were found to accumulate in the bioreactor. The main reason for membrane fouling was considered to be sludge adhesion and gel layer formation over the outer membrane surface and microbial growth over the inner membrane surface. © 2003 Elsevier Ltd. All rights reserved. Keywords: Membrane bioreactor; Wastewater reuse; Wastewater reclamation; Filtration; Membrane fouling; Membrane cleaning

1. Introduction Wastewater reclamation has been studied mostly on municipal wastewater so far. Few literature have concentrated on bath wastewater although this is even a better source for wastewater reclamation than municipal wastewater because of its large discharge amount, simple pollutant composition and low pollutant content. Reclaiming bath wastewater separately from the sewage is considered helpful to ensure the quality of the recycled water and consequently to reduce the exposure risk to various unknown hazard substances. However, operational problems may take place when bath wastewater is treated with conventional biological wastewater treatment processes. A high content of surfactants in the bath wastewater exposures the treatment system to a risk of sludge foaming. Moreover, a low content of organics may result in sludge concentrations not high enough to remove surfactants efficiently. ∗ Corresponding author. Tel.: +86-10-62772324; fax: +86-10-62771472. E-mail address: [email protected] (X. Huang).

0032-9592/$ – see front matter © 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.procbio.2003.11.038

The membrane bioreactor is a promising technology for wastewater treatment and recycling due to its high running performance such as the excellent and stable effluent quality, high organic loading rate, compact structure as well as low excess sludge production [1,2]. By substituting the settling tank in a conventional activated sludge process with a membrane filtration device, all micro-organisms are retained in the bioreactor and the hydraulic retention time (HRT) becomes completely independent on the sludge retention time (SRT) [3]. High sludge concentration can therefore be achieved even in a short HRT. Some macromolecules are also retained in the bioreactor so that the contact time of activated sludge and pollutants is elongated, facilitating efficient removal of slowly biodegradable pollutants [4,5]. The effluent of the membrane bioreactor is normally free of bacteria and has a potential in municipal and industrial reuse [6,7]. The advantages of the membrane bioreactor as mentioned above benefit reclamation of bath wastewater. In this study, long-term performance was investigated in a membrane bioreactor to treat bath wastewater. The effluent quality was monitored and subsequently compared with the current water reclamation standard of China. Contributions

126

R. Liu et al. / Process Biochemistry 40 (2005) 125–130

of biodegradation and membrane separation to COD and anionic surfactant (AS) removals were examined. The growth of activated sludge and the behaviour of inorganic substances were investigated. Membrane fouling development was monitored and analysed.

intermittently extracted through membrane modules by suction of a pump at a fixed filtration flux. The corresponding filtration flux and HRT were respectively 13 l/(m2 h) and 3.6 h. A suction mode of 13 min on and 4 min off was adopted. No sludge was discharged. No chemical cleaning of membrane modules was carried out.

2. Materials and methods

2.2. Analytical items and methods

2.1. Experimental system

All items on the quality of the influent, supernatant and effluent, together with the mixed liquor suspended solids (MLSS) and mixed liquor volatile suspended solids (MLVSS) were measured according to the standard methods [8]. AS was determined as methylene blue active substances (MBAS) with a colorimetric method and the linear alkylbenzene sulphonate (LAS) was taken as the reference substance [8]. Supernatant was obtained by centrifuging the mixed liquor for 15 min at 4000 rpm (LG10-2.4A, Beijing Medical Centrifuge Corporation) then filtering through a membrane of 0.45 ␮m. Scanning electron microphotographs (SEMs) of the fouled membrane fiber were taken by a scanning electron microscope (HITACHI S-570) after preparation following the standard procedure.

The membrane bioreactor applied in this study was a completely mixed aeration tank (effective capacity of 1.5 m3 ) in which were submerged eight hollow fibre membrane modules (total filtration area of 32 m2 ) (Fig. 1). The aeration tank was divided into one riser zone (cross-flow area of 0.32 m2 ) and two down-comer zones (cross-flow area of 0.23 m2 each) by two baffle plates. Membrane modules (polyethylene, pore size 0.4 ␮m, filtration area 4 m2 each, Mitsubishi Rayon Co. Ltd.) were allocated into two layers by four parallel rows in the riser zone. Air was supplied right below membrane modules for supplying oxygen, mixing and inducing cross flow over the membrane surface. The water level was maintained constant by a floatswitch connected to an electromagnetic valve in the influent pipe. The effluent flow rate and the transmembrane pressure were respectively monitored with sensors. Raw wastewater was collected from a public bathroom in Tsinghua University, Beijing, China. The bathroom served more than 3000 students daily. Wastewater was mainly produced from a shower bath. Discharge from a small toilet in the bathroom was also combined, but the amount was very small. Fresh wastewater was collected once per day into an elevated tank after filtration with a stainless steel screen of 1.2 mm to remove garbage and hairs. Then after filtration through another stainless steel screen of 0.9 mm, the stored wastewater was supplied into the bioreactor through the electromagnetic valve. The membrane bioreactor was designed for treating 10 m3 of wastewater daily. The mixed liquor in the bioreactor was

Electromagnetic valve Influent

Control system Flow rate sensor

Screen

Pressure sensor

Floatswitch

Effluent Pump

Membrane modules Flow meter Baffle plate Air diffuser Blower Bioreactor

Fig. 1. Schematic flow diagram of experimental apparatus.

3. Results and discussion 3.1. Effluent quality The reactor was continuously operated for 216 days while the quality of the feed water and process effluent was periodically monitored. A stable and excellent effluent quality in terms of COD and AS was obtained all the time despite the fact that the correspondent values of the influent respectively fluctuated in a range of 126–322 and 3.46–8.90 mg/l (Fig. 2). As a whole, effluent COD concentrations were 2–37 mg/l with a mean value of 18 mg/l and AS concentrations were 0.03–0.21 mg/l with a mean value of 0.08 mg/l. Other items on quality of the influent and effluent were also determined when the operation went to pseudo-steady state (Table 1). All suspended solids (SS) and most of the biological oxygen demand (BOD5 ) in the influent were removed after being treated with the membrane bioreactor. Compared with the French grey, turbid, shampoo-odoured influent, the effluent appeared clear, transparent and odourless. The effluent quality met the water reuse standard of China (CJ25.1–89) (Table 1), thus the effluent could be used for car washing, land watering and toilet flushing. Sludge foaming often occurred in the reactor owing to the high content of anionic surfactant in the feed. Brown foam sometimes was as thick as 15 cm over the activated sludge surface and some of the foam flowed out of reactor from the overflow pipe. The effluent quality, however, was not negatively influenced, showing advantage over the conventional activated sludge system in which the foam would

R. Liu et al. / Process Biochemistry 40 (2005) 125–130

127

Table 1 Quality of the influent, treated effluent and water reclamation standard of China (CJ25.1–89) Items

Influent

Effluent

Reclamation standard

Items

Influent

Effluent

Reclamation standard

SS (mg/l) NH4 + −N (mg/l) COD (mg/l) BOD5 (mg/l) AS (mg/l)

15–50 0.6–1.0 130–322 99–212 3.5–8.9

NDa <0.5 <40 <5 <0.2

≤10 ≤20 ≤50 ≤10 ≤1.0

pH Colour (TCU) Turbidity (NTU) Odour Coliforms (cell/l)

5.8–6.3 – 146–185 Shampooz-odor –

6.5–7.1 <3 <1 NDa NDa

6.5–9.0 ≤30 ≤10 NDa ≤3

a

Values below determination thresholds.

overflow with the effluent and consequently deteriorate the effluent quality. 3.2. Contributions of biodegradation and membrane separation to pollutant removals The membrane bioreactor is a technological combination of biological treatment with membrane filtration. In such a process, the membrane separated from the effluent not only suspended solids and micro-organisms but also a variety of soluble organic substances [9,10]. The contact time of the activated sludge and organic pollutants are thereby elongated, facilitating efficient removal of slowly biodegradable pollutants and thus upgrading the effluent quality. Contributions of biodegradation and membrane interception to pollutant removal were respectively evaluated in terms of removal efficiencies. Removal efficiency from biodegradation was calculated as the reduced pollutant content in the supernatant from the influent divided by the influent content. Removal efficiency from membrane interception was calculated as the reduced content in the filtrate

from the supernatant divided by the influent content. The integrated contributions of biodegradation and membrane interception to pollutant removal are the removal efficiency of the whole process. Respective contributions of biodegradation and membrane interception to COD removal are shown in Fig. 3(a). 78.5–99.9% of the influent COD was removed in the whole process, where the biological treatment took the main responsibility, reducing 34–85% of COD. The membrane separation further intercepted some pollutants from the biologically treated effluent (the supernatant of the mixed liquor), counter-balancing the instability of the bioreactor and consequently ensuring stable quality of process effluent. The pollutant removal performance and the effluent quality did not change significantly with organic volumetric loads ranged in 0.50–1.85 kg-COD/(m3 per day). Occasionally low COD removal efficiencies compared with those in municipal wastewater treatment [11,12] was attributed to low

COD (mg/L)

400 300 200 100 (a)

0

AS (mg/L)

10 8 6 4 2 0 0 (b)

50

100 150 200 Operational time (d)

250

Fig. 2. Profiles of COD and anionic surfactant (AS) concentrations in the influent (䊏) and effluent (䊊). (a) COD concentrations. (b) AS concentrations.

Fig. 3. Contributions of biological treatment and membrane separation to (a) COD removal and (b) anionic surfactant (AS) removal.

R. Liu et al. / Process Biochemistry 40 (2005) 125–130

3.3. Sludge concentration In this study, dewatered activated sludge from a municipal wastewater treatment plant was seeded into the bioreactor after filtration with a stainless steel screen of 0.9 mm. MLSS in the activated sludge was initially around 4 g/l, but gradually decreased and levelled off at about 1.3 g/l. It is difficult to explain why the MLSS stabilised at so low a value when no sludge was discharged. Loss of sludge with sludge foaming may be a reason but this should not be the main contributor. The largest possibility might reside in the composition of the bath wastewater, which contains a variety of bactericidal substances from shampoo, body soap or other cleaning agents. Moreover, some surfactants also inhibit microbial growth [13]. It is necessary to carry out further study on the microbial community composition and microbial growth kinetics in the reactor in order to find the reasons. Inorganic substances accumulated in the bioreactor. The ratio of MLVSS to MLSS decreased with the operational time, especially during the first 89 days when MLVSS/MLSS was reduced from 83.4% to as low as 64.6% (Fig. 4). This inorganic accumulation was mostly attributed to the open structure of the bioreactor as well as its location. In the room of about 20 m2 where the membrane bioreactor was located, a boiler was used for heating bath waters in the public bathroom. Much dirt was produced when coal was supplied into the boiler and when the boiler was heated. The dirt in the air entered the bioreactor, mixed into flocs and then was retained by membrane separation. The decreasing tendency of MLVSS/MLSS was much slowed down 89 days later when the boiler stopped running. There was no indi100

4

80 MLVSS/MLSS

3

60

2

40 MLVSS

1

MLSS

20

0

MLVSS/MLSS (%)

MLSS or M LVSS (g/L)

5

0 0

50

100 150 200 Operational time (d)

250

Fig. 4. Profiles of sludge concentrations and MLVSS/MLSS.

cation demonstrated that effluent quality was influenced by the accumulation of inert substances (Fig. 2(a)), although decreased microbial activity with inorganic accumulation was once reported [14]. 3.4. Membrane fouling The transmembrane pressure was adopted as an indicator for membrane fouling since the filtration flux was fixed in most cases during the experiment. Evolutions of the transmembrane pressure and the filtration flux were monitored, as shown in Fig. 5. The filtration flux gradually increased and levelled off at 13.75 l/(m2 h) before Day 90. Transmembrane pressure normally remained below 13 kPa, but occasionally demonstrated an abrupt increase. The sporadic elevation of transmembrane pressure was mainly related to several operational accidents when no wastewater was supplied into the reactor owing to problems of the floatswitch. The liquid was continuously sucked out without influent until almost all membranes were bared on Days 57 and 90 and one-fourth of membranes were bared on Day 65. The filtration flux was greatly increased owing to the reduction of the available membrane filtration surface. MLSS was increased due to concentration. Transmembrane pressure was therefore abruptly elevated as a consequence of the increase in both filtration flux and MLSS. In order to retard membrane fouling, tap water was added to the normal liquid level and filtration was stopped for 24 h. Membrane permeability was almost completely restored. Membrane fouling became irreversible after Day 90 and the transmembrane pressure rapidly increased. On Day 100, all membranes were bared but the transmembrane pressure demonstrated no decrease after filtration was stopped for 24 h. In addition, the transmembrane pressure rapidly increased from Days 92–93 and during Days 124–126 when the filtration flux was slightly increased to 15 l/(m2 h). The increased pressure did not decrease even when the filtration flux was reduced to 11 l/(m2 h). The filtration performance became unstable from Day 183 when the rotating speed of the suction pump had to be increased often to maintain the filtration flux. The transmembrane pressure continuously increased and the experiment was stopped on Day 216. 80 70

16 14

60 50

12 10

40 30 20

8 6 4

10 0 0

2 0 50

100 150 200 Operational time (d)

Flux (L/(m2·h))

COD concentrations in the influent but not the decreased process performance. AS removal efficiency remained as high as 98% above in the whole process (Fig. 3(b)). In contrast with COD removal, nearly all AS was removed with biological treatment while membrane separation displayed almost no contribution. The idleness of membrane separation might be attributed to the small size of anionic surfactant molecules, as observed in ammonia removal during treatment of municipal wastewater [12].

Transmembrane pressure (kPa)

128

250

Fig. 5. Profiles of the transmembrane pressure and the filtration flux.

R. Liu et al. / Process Biochemistry 40 (2005) 125–130

129

Fig. 6. SEM photos of a fouled membrane fiber on Day 216. (a) Fouling status of the membrane outer surface; (b) a magnified photo for deposition of micro-organisms and suspended solids on the membrane outer surface; (c) a magnified photo for formation of a gel layer over the membrane outer surface; (d) a magnified photo for the outer surface of chemically cleaned membrane; (e) fouling status of the membrane inner surface; and (f) a magnified photo for the fouled membrane inner surface.

SEMs of several fouled membrane fibers were taken at the end of the operation. Foulants were visible on both of the outer surface and inner surface of the membrane fibre (Fig. 6). Fouling to the membrane outer surface was mostly attributed to the formation of a cake layer as well as a gel layer. The cake layer was formed by slight deposition of some suspended solids (Fig. 6(a)), which were mainly composed of a diverse society of micro-organisms in a matrix of viscous extracelluar substance (Fig. 6(b)). Under the cake layer a heavily developed gel layer was recorded (Fig. 6(c)), which made membrane pores vague even invisible compared with the chemically cleaned membrane (Fig. 6(d)). As for fouling of the membrane inner surface, micro-organisms seemed to contribute principally (Fig. 6(e) and (f)). Filamentous bacteria acted as the skeleton with cocci or bacilli as the matrix in the cake layer. The pore appeared clear with almost no gel layer formed over the surface.

1. Stable and excellent effluent was obtained with COD<40 mg/l, NH4 + −N <0.5 mg/l, anionic surfactant <0.2 mg/l, no colour, no odour and free of SS, meeting the wastewater reuse standard of China. Sludge foaming often occurred but gave no visible impact on effluent quality. 2. Most COD and almost all anionic surfactant in the influent were biologically removed by activated sludge. Membrane separation acted as an equilibrator to the unstable biological treatment of COD while demonstrated no contribution to anionic surfactant removal. 3. MLSS concentration in the bioreactor stabilised at around 1.3 g/l. Accumulation of inorganic substances was observed. 4. A slight deposited cake layer and a heavy developed gel layer contributed to the fouling of the membrane outer surface. Attachment of micro-organisms contributed to the fouling of the membrane inner surface.

4. Conclusions Reclamation of bath wastewater with conventional processes encounters sludge foaming problems. In this study, the operating performance in a pilot plant of a submerged membrane bioreactor for treating bath wastewater was investigated. The operation was continued for 216 days. The long-term pollutant removal performance and membrane permeability were studied and the following results obtained:

References [1] Brindle K, Stephenson T. Application of membrane biological reactors for the treatment of wastewaters. Biotech Bioeng 1996;49:601– 10. [2] Xing CH, Tardieu E, Qian Y, Wen XH. Ultrafiltration membrane bioreactor for urban wastewater reclamation. J Membr Sci 2000;177:73–82.

130

R. Liu et al. / Process Biochemistry 40 (2005) 125–130

[3] Muller EB, Stouthamber AH. Aerobic domestic wastewater treatment in a pilot plant with complete sludge retention by cross-flow filtration. Water Res 1995;29(4):1179–89. [4] Scott JA, Smith KL. A bioreactor coupled to a membrane to provide aeration and filtration in ice-cream factory wastewater remediation. Water Res 1997;31(1):69–74. [5] Huang, X, Gui, P, Wen, X-H, Qian, Y. Treatment of refractory organic compound using a membrane bioreactor. Paper presented at the 21th Century-International Symposium on Membrane Technology and Environment Protection, September 18–21 2000; Beijing, China. [6] Dijk L van, Roncken GCG. Membrane bioreactors for wastewater treatment: the state of the art and new developments. Water Sci Technol 1997;35(10):35–41. [7] Jeffrey P, Seaton RAF, Stephenson T, Parsons S. Infrastructure configurations for wastewater treatment and reuse: a simulation based study of membrane bioreactors. Water Sci Technol 1998;38(11):105– 11. [8] APHA/AWWA/WEF. Standard methods for the examination of water and wastewater. 19th ed. Washington, DC: USA; 1995.

[9] Nagaoka H, Ueda S, Miya A. Influence of bacterial extracellular polymers on the membrane separation activated sludge process. Water Sci Technol 1996;34(9):165–72. [10] Huang X, Liu R, Qian Y. Behaviour of soluble microbial products in a membrane bioreactor. Process Biochem 2000;36(5): 401–6. [11] Chiemchaisri C, Yamamoto K. Performance of membrane separation bioreactor at various temperatures for domestic wastewater treatment. J Membr Sci 1994;87:119–29. [12] Liu R, Huang X, Chen L-J, Wang C-W, Qian Y. A pilot study on domestic wastewater treatment using a submerged membrane bioreactor. J Environ Sci Health 2000;A35(10):14–33. [13] Wu H, Kuzmenko A, Wan S, Schaffer L, Weiss A, Fisher JH, et al. Surfactant proteins A and D inhibit the growth of Gram-negative bacteria by increasing membrane permeability. J Clin Invest 2003;111(10):1589–602. [14] Yamamoto K, Hiasa M, Mahmood T, Matsuo T. Direct solid–liquid separation using hollow fiber membrane in an activated sludge aeration tank. Water Sci Technol 1989;21:43–54.