Nik Meriam Nik Sulaiman* Shaliza Ibrahim Sarah Lim Abdullah · Membrane Bioreactor for the treatment of natural rubber wastewater 95 The objectives are: i to adapt available laboratory

92 Int. J. Environmental Engineering, Vol. 2, Nos. 1/2/3, 2010

Copyright © 2010 Inderscience Enterprises Ltd.

Membrane Bioreactor for the treatment of natural rubber wastewater

Nik Meriam Nik Sulaiman* Faculty of Engineering, Department of Chemical Engineering, University of Malaya, 50603 Lembah Pantai, Kuala Lumpur, Malaysia Fax: +603-7967 5319 E-mail: [email protected] *Corresponding author

Shaliza Ibrahim Faculty of Engineering, Department of Civil Engineering, University of Malaya, 50603 Lembah Pantai, Kuala Lumpur, Malaysia E-mail: [email protected]

Sarah Lim Abdullah Trident Consultants, B-3A-4, Megan Avenue 1, 189, Jalan Tun Razak, 50400 Kuala Lumpur, Malaysia E-mail: [email protected]

Abstract: A membrane bioreactor was set up using two pieces of Kubota flat sheet membranes, with nominal pore size of 0.4 ∝m and a total effective area of 0.2 m2, to treat wastewater from a latex concentrate processing factory. A steady-state MLSS concentration was attained at 8500 mg/L during biomass acclimatisation. A critical flux of 0.01224 m3/m2h was determined using the flux step method, whereas an optimum operating flux for fouling minimisation was determined from the sub-critical flux region, to be 0.009 m3/m2h at a COD loading of 3500 mg/L. The BOD3 and COD removal efficiencies were 96.78% and 96.99%, respectively.

Keywords: MBR; membrane bioreactor; natural rubber wastewater; critical flux; optimal flux; COD loading.

Reference to this paper should be made as follows: Sulaiman, N.M.N., Ibrahim, S. and Abdullah, S.L. (2010) ‘Membrane Bioreactor for the treatment of natural rubber wastewater’, Int. J. Environmental Engineering, Vol. 2, Nos. 1/2/3, pp.92–109.

Membrane Bioreactor for the treatment of natural rubber wastewater 93

Biographical notes: Nik Meriam Nik Sulaiman is a Professor at the Department of Chemical Engineering, University of Malaya (UM). She is also currently the Director of the Institute of Research Management and Consultancy. She has numerous publications in membrane applications in wastewater treatment and separation processes and also has keen interests in air pollution issues and life cycle management. She has been instrumental in initiating Cleaner Production projects in UM, and is focused on pursuing the concept of sustainable industrial development in this country.

Shaliza Ibrahim is a Chemical Engineer by training and is based at the Department of Civil Engineering for the Environmental Engineering Programme. She has numerous publications to her credit, on mixing aspects of bioreactor design and operation. She is also involved in projects on wastewater treatment, water quality, and the implementation of Cleaner Production (CP) in Malaysia.

Sarah Lim Abdullah received a Chemical Engineering Degree for her first degree academic training. Upon graduation she pursued a Master of Engineering Science programme at the University of Malaya and initiated this project in MBR application for the process industries. She successfully completed her masters in 2007. She is currently employed as an environmental consultant at Trident Consultants.

1 Introduction

The production of latex concentrate is one of the two types of processes in raw natural rubber processing; the other is the production of Standard Malaysian Rubber (SMR). Ninety-five percent of latex concentration in Malaysia is achieved by the centrifugation process. Figure 1 shows a flow chart of the latex concentrate production. The by-product, skim latex, is further treated with acid to recover the remaining rubber content. Skim serum combined with wastewater from the rest of the plant form the overall effluent. Large quantities of water are used in the dilution of field latex and the cleaning of materials and machinery. The effluent typically contains materials originating from field latex such as small amounts of uncoagulated latex, serum containing substantial quantities of proteins, sugars, lipids, as well as inorganic and organic salts. Formic, acetic, phosphoric or sulphuric acids, ammonia and TMTD-ZnO are the chemicals used in natural rubber processing plant.

In 1977, 97 million litres of rubber processing effluent was discharged daily in Malaysia. This volume of effluent contained about 200 tonnes BOD per day (Chin et al., 1978). A year later, about 100 million litres of effluent were reported to be discharged daily into streams and rivers from the rubber processing factories in Malaysia (John and Ong, 1979). The Malaysian Government enacted the Environmental Quality Regulations (Raw Natural Rubber Industry) in 1978 to control polluted effluent from rubber processing factories. Effluent from the factories has to be treated to meet regulatory standards before being discharged into the waterways (Isa et al., 1988). The standard for effluent from latex concentrate factories falls under the Third Schedule, included in Tables 3 and 4.

94 N.M.N. Sulaiman et al.

Latex concentrate factory effluent is the most difficult type of wastewater from rubber processing factories, but with the establishment of EQA 1978, improvement of the compliance status increased significantly from 75% in 1993 to 90% in 1997.

Since the contents of natural rubber factory effluents are biodegradable, a number of biological treatment processes, such as anaerobic-cum-facultative lagoons, anaerobic-cum-aerated lagoons, aerated lagoons and oxidation ditches have been developed specifically for raw natural rubber factory effluents. The main shortcomings of these processes include large land area requirement, high energy consumption for the aerators, longer effluent treatment period, odour problems, and high operating and maintenance costs. These circumstances in turn lead to frequent non-compliance to the legal discharge limits.

Figure 1 Effluent generation from latex concentrate production (see online version for colours)

Research into membrane technology applications in the natural rubber industry has been initiated for concentrating natural rubber serum by reverse osmosis (Lau, 1994), treatment of skim latex serum using ultrafiltration with air sparging (Harunsyah and Sulaiman, 2002) and concentration of natural rubber field latex using tubular cross flow ultrafiltration system (Veerasamy et al., 2003). Presently, there is no reported research on the use of MBR system for the treatment of effluent from natural rubber processing factories.

An MBR system substitutes the clarifier in a conventional aerobic system with microfiltration or ultrafiltration membrane modules. Unlike the clarifier, the membrane is a microbial barrier that can capture the biomass for recirculation inside the bioreactor. This technology has many benefits such as smaller space requirement, disinfection, high loading rate capability, better solids removal and retention of high-molecular weight organic compounds that can enhance the biodegradation process (Le-Clech et al., 2003).

In this study, the hydraulic performance of an MBR process for natural rubber latex concentrate processing effluent treatment is examined.


The objectives are:

i to adapt available laboratory scale MBR design parameters for application in the treatment of latex concentrate processing effluent

ii to determine the critical flux, Jc, of the membrane for fouling minimisation

iii to obtain the optimum operating flux for latex concentrate processing effluent treatment to enhance permeate quality and minimise fouling

iv to determine the effect of COD loading concentration on permeate quality at the optimum flux established in (iii).

2 Materials and methods

2.1 Wastewater

Skim serum was used as the wastewater in MBR experiments instead of the overall factory effluent, owing to the former’s higher pollution load. Otherwise, a large quantity of effluent would need to be collected. But since the COD and BOD3 of skim serum are about five times those of the effluent, dilution of skim serum was done to get the desired loading. The overall effluent (skim serum together with the washings from the centrifuge bowls, floor and milling of skim rubber) was, nevertheless, also collected to determine the typical range of the pollutant parameter values for latex concentrate processing factories. The samples were kept at pH 4 in the cold room. On the basis of the Standard Methods approved by DOE, Malaysia, 1995 (Department of Environment, 1995), the pH, Biochemical Oxygen Demand (BOD), Chemical Oxygen Demand (COD), Suspended Solids (SS), AN and TN were measured for the feed wastewater as well as for the permeate streams from the experiments.

2.2 Membrane Bioreactor

A laboratory scale MBR model, designed by Lim and Ujang (2004a) for the treatment of Palm Oil Mill Effluent (POME), was adapted with minor adjustments for latex concentrate processing effluent treatment. The anoxic section of MBR system was eliminated due to the lower strength of the latex concentrate processing effluent compared with POME.

Figure 2 shows a diagram of the MBR system. Two flat sheet Kubota membranes, each with an effective area of 0.1 m2 were submerged in a Perspex tank having a working volume of 0.018 m3. Air was introduced through diffusers at a flow rate of 0.015 m3/h. The sparging created coarse bubble aeration, which provided oxygen for the aerobic environment (DO ≥ 2 mg/L) in the reactor in which the activated sludge is kept in suspension and turbulence is created to reduce membrane fouling. Baffles of flat Perspex sheets were located on both sides of the membrane module to concentrate the bubbles on the membrane surface to enhance the scouring action. The Transmembrane Pressure (TMP) was read by a PICO pressure transmitter.


Figure 2 Membrane Bioreactor system

2.3 Biomass acclimatisation

The acclimatisation of the microorganisms in the bioreactor is a necessary start-up step of any biological process. It involves rather slow seeding of wastewater with a cocktail of microorganisms usually obtained from the factory’s Wastewater Treatment Plant (WWTP). For this study, the process was aided by the use of microorganisms from the clarifier of the latex factory’s WWTP.

Under aerobic condition, the pH was maintained at a range between 6 and 8. The method used to determine the Mixed Liquor Suspended Solids (MLSS) and Mixed Liquor Volatile Suspended Solids (MLVSS) were analyses of suspended solids (APHA Standard Method, 2540 D, 1995) and volatile suspended solids (APHA Standard Method, 2540 E, 1995). The remaining sample was returned to the MBR tank to minimise the loss of biomass. The procedure for biomass acclimatisation is as follows:

i 3 L of activated sludge from the clarifier (about 1000 mg/L of MLSS) was fed into the MBR tank.

ii At the beginning of biomass acclimatisation, 100 mL of undiluted serum was fed into the MBR tank daily.

iii 50 mL of sample was collected daily to measure MLSS and MLVSS concentrations.

iv Steps (i)–(iii) were repeated with 200 mL, 300 mL, 400 mL and 500 mL of undiluted serum when MLSS concentration in the MBR tank reached 2000 mg/L, 3000 mg/L, 4000 mg/L and 5000 mg/L, respectively. This is because more food is needed to increase biomass in the reactor.

v The stepwise procedure was continued with the undiluted serum addition fixed at a volume of 500 mL even as the MLSS concentrations keep increasing.

vi When the MLSS and MLVSS remained constant for few days, steady state was assumed, and the batch feeding process was terminated (Lim and Ujang, 2004a).


This acclimatisation stage was used to maximise the microorganism concentration, by operating the MBR without sludge removal except for the samples to measure MLSS and MLVSS. Zero wastage has been reported in studies involving MBR. Muller et al. (1995) treated domestic wastewater for 300 days without wastage, and Benitez et al. (1995) treated pharmaceutical wastewater for 62 days without sludge removal. Pollice et al. (2004) also operated a bench scale MBR for more than 100 days without sludge withdrawal.

2.4 Membrane cleaning

Membrane cleaning must be conducted after every set of experiment. The following procedure was followed:

• The membranes were taken out and soaked in a 0.5% w/w NaOCl solution (manufacturer’s specification) for 1 h. While the membranes were out, of the tank the sludge was kept aerated.

• The membranes were then soaked in distilled water for 30 min.

• The cleaned membranes were tested for integrity using the permeability test.

• The membranes were then put back into the MBR tank ready for the next experiment.

2.5 Permeability test

Permeability test with distilled water was conducted each time after the membranes were cleaned before they were returned to the reactor for the next experimental run, to ensure the membranes’ integrity. The duration of this test was about 200 min. The procedure was:

• 25 L of distilled water was fed into a tank

• The cleaned membranes were submerged in the tank and the suction rate of the peristaltic pump was set at 60 rpm for 10 min

• Permeate was collected in a 100 mL measuring cylinder and the volume was recorded every minute

• The suction rate was then increased by 60 rpm every 10 min and the permeate was collected at each suction rate

• The permeate flow rates were compared with the value obtained at the start of the project before the membranes were used for the first time

2.6 MBR experiments

Experimental runs began when the MLSS and MLVSS reached steady state upon biomass acclimatisation in the MBR. The feed was diluted skim serum whereby the working volume was maintained at 0.018 m3 by a ball valve. A peristaltic pump was used to create suction for permeate to flow out through the membranes to be collected in


another tank, leaving the solids in the bioreactor. Table 1 summarises the values of the operating parameters in this study.

Table 1 Typical operating parameters for this study

Parameters Value

Working volume of reactor (V) 0.018 m3 Temperature (T) 25–32°C Aeration rate 0.015 m3/h Dissolved Oxygen (DO) ≥2 mg/L Transmembrane Pressure (TMP) ≤85 kPa Sludge wasted 0 m3/d

Operation was maintained at room temperature and dissolved oxygen concentration of more than 2 mg/L to ensure aerobic conditions. Other operating parameters in this study, i.e., flux, Hydraulic Retention Time (HRT), Solids Retention Time (SRT) and removal efficiency were calculated using equations (1)–(4). The HRT in this study ranged between 7.5 h and 10 h, depending on the operating flux.

( )3 2 Permeate flowFlux, m m hTotal membrane area

J = (1)

Reactor working volumeHRT(h)Permeate flow

= (2)

Reactor working volumeSRT(d)Sludge wasted

= (3)

Influent EffluentRemoval efficiency (%) 100.Influent

−= × (4)

2.7 Critical flux test

The critical flux was determined to enable the selection of a flux value that would minimise fouling and prolong the duration of operation before chemical cleaning was warranted. Below critical value, the flux is directly proportional to the TMP, and steady-state operation is maintained. The flux step method was used to determine the critical flux since it is the most common method compared with other techniques such as hydraulic test and stepwise increase of TMP (Defrance and Jaffrin, 1999). Since the maximum COD of the WWTP in the latex concentrate processing factory is about 3500 mg/L, the skim serum was diluted to a COD of about 3500 mg/L for use in this experiment.

The pump suction rate was increased by 15 rpm every 10 min to increase the flow. Permeate volume collected in a measuring cylinder was recorded every minute. The steps with increasing suction were repeated until TMP started to increase significantly during the 10 min interval, indicating that the critical flux was near. When TMP nearly reached


80 kPa, the suction rate was reduced by 15 rpm every 10 min, and permeate collected was recorded.

2.8 Determination of optimum flux

Defrance and Jaffrin (1999) suggested that optimum flux lies at the upper boundary of the sub-critical flux region, at a value just below the critical flux. It is important to determine the optimum flux so that the MBR can be operated for long periods without chemical cleaning. Flux values of 0.012, 0.011, 0.010 and 0.009 m3/m2h were used with skim serum diluted to COD 3500 mg/L as the feed.

At each flux run, permeate collection in a 100 mL measuring cylinder was recorded every minute. The peristaltic pump was switched off when the TMP reached the maximum operating pressure, 85 kPa, and the time required to reach 85 kPa was recorded. The membranes were cleaned and the collected permeate was analysed for its characteristics. The run was repeated for another flux.

2.9 Effect of COD loading concentration

The COD loading concentrations in the oxidation ditch at the latex concentrate processing factory ranged from 1259 mg/L to 3750 mg/L. COD loadings of 3500, 4500 and 5500 mg/L were selected in this work to determine the effect of loading on permeate quality. The purpose of this experiment is to check the ability of MBR to treat the effluent at higher COD loadings. Skim serum was diluted to the three COD loading concentrations to be used as feed. The optimum flux determined above was used in all the runs. Permeate was collected and analysed every day for a week.

3 Results and discussions

The wastewater characteristics of the skim serum and overall effluent from the factory are given in Table 2.

Table 2 Wastewater characteristics of raw wastewater for microorganism acclimatisation

Quantities No. Parameters Skim serum Effluent 1 PH 3.80 3.88 2 BOD3 (30°C days) (mg/L) 9405 1398

3 COD (mg/L) 15267 3488 4 Suspended solids (mg/L) 605 258 5 Total Nitrogen (mg/L) 2163 468 6 Ammoniacal Nitrogen (mg/L) 1498 326


3.1 Microorganism acclimatisation

Sludge age was almost infinity since no sludge was wasted except for the samples taken for MLSS and MLVSS measurements. Operating an MBR without any sludge wastage should allow the system to reach a steady state at a given sludge concentration. Under this condition, defined as ‘complete sludge retention’ by van-Houten and Eikelboom (1997), the microorganism growth rate balances the decay rate.

In this study, 3 L of activated sludge from the WWTP clarifier (MLSS about 1000 mg/L) was used at the beginning of microorganism acclimatisation process. The process took about 50 days to achieve steady state, at which point the MLSS concentration variation was less than 10% (Ong et al., 2004). The ratio of MLVSS to MLSS was 0.8 at steady state. Rosenburger et al. (1999) had demonstrated continuous MBR operation for three years with no reduction in the volatile fraction, and remaining constant at approximately 75%. Hasar et al. (2002) used an MBR to treat domestic wastewater, which recorded a ratio of 0.7, with the maximum MLSS and MLVSS concentrations at 14650 mg/L and 10245 mg/L, respectively. Pollice et al. (2004) also documented that the MLVSS/MLSS ratio was stable at 80% over the whole experiment in their MBR process for municipal wastewater treatment.

Figure 3 shows the MLSS and MLVSS variation in the MBR system during the acclimatisation period. Microorganism acclimatisation can be understood using the microorganism batch growth curve (Davis and Cornwell, 1998). At the beginning of acclimatisation (the first 10 days), the MLSS and MLVSS concentration increased slightly. This is the lag phase of microorganism growth, where digestion of food occurs and the necessary enzymes are developed.

Figure 3 Variation of MLSS and MLVSS in MBR during the acclimatisation period

Between days 10 and 15, both MLSS and MLVSS concentration kept increasing, but with occasional sporadic decrease. A possible explanation could be that some microorganisms in the reactor were adapting themselves to degrade non-active microorganisms. This sporadic decrease was also reported by Choo and Stensel (2000) using their sequencing batch MBR treatment for nitrogen removal. Nah et al. (2000) also recorded this behaviour when MBR was used for household wastewater treatment.

Both MLSS and MLVSS concentrations increased rapidly between days 25 and 35, and then declined between days 35 and 45, where reproduction slowed down because food supply remained constant. During days 45 and 60, the MLSS and MLVSS


concentrations achieved steady state at about 8500 mg/L and 6800 mg/L, respectively. Lim and Ujang (2004a) reported that the biomass acclimatisation in their submerged MBR system for POME treatment was about 45 days, with MLSS and MLVSS concentrations at 7400 mg/L and 4700 mg/L, respectively. This gives their study a MLVSS to MLSS ratio of about 0.64.

3.2 Critical flux test

The flux step method used here is preferred over the TMP-step method since it provides better control of the flow material depositions on the membrane surface (Le-Clech et al., 2003).

Figure 4 shows the results obtained during the critical flux test. The critical flux, Jc, in this study was 0.01224 m3/m2h at MLSS and MLVSS concentrations of 8500 mg/L and 6800 mg/L, respectively.

Figure 4 Relationship between flux and TMP in the critical flux test

During the flux increase phase, TMP remained constant at 0 kPa at a suction rate of 15 rpm, corresponding to a flux of 0.00423 m3/m2h. At 600 s the suction rate was increased to 30 rpm, to increase the flux to 0.01389 m3/m2h. This flux was maintained for the time interval between 600 s and 1200 s, and the TMP was observed to increase slightly from 0.76 kPa to 1.37 kPa. This regime showed that membrane fouling had started to occur.

The suction rate was then increased and maintained at 45 rpm for the time interval 1200–1800 s. This resulted in the flux increasing to 0.02646 m3/m2h. The increasing rate of TMP became significant during this time interval due to fouling. At 1800 s, the suction rate raised to 60 rpm. Over the 10 min interval the TMP increased rapidly from 18.15 kPa to 66.97 kPa, but the flux only increased slightly to 0.03024 m3/m2h as the pores of the membranes were probably blocked. Since the maximum operating pressure of the membranes was 85 kPa (manufacturer’s specification), the step increase in flux was discontinued.

Reducing the suction rate to 45 rpm at 2400 s showed a super-critical condition where the TMP increased rapidly from 66.29 kPa to 83.42 kPa, although the flux was reduced to 0.01913 m3/m2h. At 3000 s, the suction rate was reduced to 30 rpm with a flux of 0.01224 m3/m2h. The critical flux region occurred when the TMP dropped from


80.23 kPa to 42.56 kPa and only varied between 39.36 kPa and 41.47 kPa, which is small enough to be considered constant. When flux was changed to 0.00423 m3/m2h between the time 3600 s and 4200 s, TMP decreased dramatically from 37.63 kPa to 8.49 kPa. This indicated the sub-critical condition. Hence, the critical flux was determined to be 0.01224 m3/m2h in this study, with COD loading of 3500 mg/L.

A similar method was used by Le-Clech et al. (2003). Critical flux rate was recorded at 10 L/m2h in their vertical-mounted submerged tubular MBR for synthetic wastewater with MLSS concentration at 3 g/L. Lim and Ujang (2004b) reported a critical flux rate of 0.0112 m3/m2h in their submerged MBR for POME treatment using the flux step method. In their study, the TMP obtained during the flux decrease phase was higher than during the flux increase phase, and vice versa for the flux, which is similar to the results of this study.

The critical flux indicates the point at which fouling starts to become severe, but does not yield predictive absolute data for extended operation (Le-Clech et al., 2003). There were observations that fouling occurs after a certain period even though the MBR process was operated under critical flux or at low flux rates (Defrance and Jaffrin, 1999; Cho and Fane, 2002). However, the critical flux of a process still depends on the cross flow velocity and feed concentration as well as the MLSS concentration. The higher cross flow velocity or the lower MLSS and feed concentrations, the higher the value of critical flux obtained (Field et al., 1995). A similar conclusion was made by Madaeni (1997) using Millipore hydrophilic 0.22 µm microfiltration membranes.

3.3 Determination of optimum flux

As discussed in critical flux concept, it is important to ascertain a flux in the sub-critical flux region, where the tendency for membrane fouling is low, which will make it easier to clean. Slight increases in flux after this region can lead to rapid TMP changes with time. Selected flux values from sub-critical region were used to determine the optimum flux for this system. Critical flux was determined to be 0.01224 m3/m2h, thus 0.012 m3/m2h, 0.011 m3/m2h, 0.010 m3/m2h and 0.009 m3/m2h were selected for these experiments.

It was observed that a decrease of flux occurred with time although sub-critical fluxes were used. The flux had to be controlled throughout the experiment by increasing the suction rate. Constant flux needed to be maintained to study the TMP variations with time. The TMP variations with time at different flux rates are illustrated in Figure 5. After each run of flux value, the membranes were cleaned for the next run.

When the MBR was operated at flux rate of 0.012 m3/m2h, the TMP increase was recorded at an average rate of 3.51 kPa/hr, and reached 84.25 kPa after one day of operation. The experiment was terminated since its maximum operating pressure is 85 kPa. This means that membrane cleaning would be required daily if this system operated at 0.012 m3/m2h.

At a flux of 0.011 m3/m2h, the time for TMP to reach its maximum operating pressure was about two days. The TMP was observed to increase at an average rate of 1.76 kPa/h, to reach 84.59 kPa after the second day of operation.


Figure 5 TMP variations using different flux

The process was conducted continuously without the need for membrane cleaning for about five days under the operating flux of 0.010 m3/m2h. TMP increased at an average of 0.71 kPa/hr when the value reached 84.79 kPa.

When the operating flux was fixed at 0.009 m3/m2h, the TMP was constant at 0 kPa for the first 18 h of operation. After this time, TMP increased gradually and reached 52.18 kPa after 30 days (average rate = 0.0725 kPa/h). This indicates that the MBR process can be operated at this flux without membrane cleaning for more than a month.

Ahn and Song (2000) reported an MBR operation for municipal wastewater treatment without membrane cleaning, for 77 days, at a sub-critical flux of 0.04 m3/m2h. According to Howell (1995), no fouling would occur and the system would operate in a total clean regime if the operation is conducted below the critical flux. Kubota (2004) reported maintaining an operation at a low flux for a submerged MBR system using flat sheet microfiltration membrane for over a year without chemical cleaning. The only cleaning activity was the constant flushing of membrane surface by cross flow velocity generated by the air bubbles from the aeration system (Ujang and Le-Clech, 2003).

Different operating fluxes correspond to different HRT values, as calculated from equation (2). Lower flux means higher HRT, so for rates of 0.009, 0.010, 0.011 and 0.012 m3/m2h, equation (2) gave HRTs of 10, 9, 8.18 and 7.5 h, respectively. Table 3 shows the permeate qualities obtained at various flux values. In general, the longer HRT allowed for more degradation of the wastewater, thus improving the parameter values. All parameters except BOD3 met the discharge limits. BOD3 only complied at 0.09 m3/m2h.

pH increased progressively from 6.92 to 7.05 as flux reduced from 0.012 m3/m2h to 0.09 m3/m2h. It can be deduced that the longer HRT allows the biomass to degrade more fatty acids in the MBR tank and therefore, the acidity of permeate was lower. BOD3, COD, AN and TN increased with higher flux.

Ahn and Song (2000) made similar observations for BOD3 and COD in the municipal wastewater treatment using MBR. Total nitrogen (TN) increased from 163 mg/L to 189 mg/L, whereas AN increased from 126 mg/L to 147 mg/L, respectively, with increase in flux from 0.009 m3/m2h to 0.012 m3/m2h.


Table 3 Permeate characteristics at different flux rates

Permeate quality (average)

Parameters Units Raw feed

DOE discharge standard limits (third schedule)

Flux at 0.009

m3/m2h

Flux at 0.010

m3/m2h

Flux at 0.011

m3/m2h

Flux at 0.012

m3/m2h

pH – 3.88 ± 0.5 6–9 7.05 6.98 6.95 6.92 BOD3 mg/L 1398 ± 50 50 (100)* 45 56 90 117 COD mg/L 3488 ± 50 400 105 112 129 158 Total Nitrogen

mg/L 468 ± 50 300 163 168 174 189

Ammoniacal Nitrogen

mg/L 326 ± 50 250 126 134 139 147

Suspended solids

mg/L 258 ± 50 100 (150)* 57 53 49 38

*No single value to exceed the limit in the bracket.

The only parameter that showed an inverse relationship with flux rate was the suspended solid in permeates. This might be due to the build up of thin biofilms and foulants on the membrane surface, which occurred more rapidly at higher fluxes, thus resulting in better solids removal.

3.4 Effect of COD loading concentrations

COD loading concentrations of 3500, 4500 and 5500 mg/L were selected to check the ability of the MBR system to react to variable shock loadings of COD. Figures 6 and 7 show the permeate qualities in terms of COD and BOD concentrations at different COD loadings at a flux rate of 0.009 m3/m2h. The permeate quality analyses are summarised in Table 4.

Figure 6 Permeate quality for COD effluent


Figure 7 Permeate quality for BOD effluent

Table 4 Permeate characteristics at different COD loading concentrations

Permeate quality (Average)

Parameters Units

DOE discharge standard limits (third schedule)

COD at 3500 mg/L

COD at 4500 mg/L

COD at 5500 mg/L

pH – 6–9 7.05 6.89 6.46 BOD3 mg/L 50 (100)* 45 120 247 Total Nitrogen mg/L 300 163 198 258 Ammoniacal Nitrogen mg/L 250 126 187 235 Suspended solids mg/L 100 (150)* 57 59 58

*No single value to exceed the limit in the bracket.

In this study, the critical parameters of interest are the COD and BOD3 levels. However, only COD values at specified shock loadings are tested due to the non-linearity of the dilution factor of the skim serum to sample BOD3. Hence difference in the BOD3 readings can be attributed to the fact that the sample homogeneity was not the same for each sampling event. Also, BOD3 readings naturally fluctuate and are more case-sensitive than the COD readings. This is because the dissolved oxygen of samples is significantly affected by the condition of the samples and this directly affects the BOD3 readings.

At a COD loading of 3500 mg/L, BOD3 was about 1400 mg/L and the average treated effluent quality for seven days was 105 mg/L COD and 45 mg/L BOD3. Both parameters complied with the DOE’s discharge limits for latex concentrate processing factories. At a COD loading concentration of 4500 mg/L, the BOD3 was measured to be 2000 mg/L. This loading concentration caused the average treated effluent COD and BOD3 concentrations to be at higher levels. For COD loading concentration of 5500 mg/L, BOD3 concentration was 2600 mg/L, and the treated effluent COD and BOD3 were 348 mg/L and 247 mg/L, respectively.

For COD loading concentrations at 3500, 4500 and 5500 mg/L, removal efficiencies for COD were 96.99%, 94.36% and 93.63%, respectively, whereas for BOD3 removal was 96.77%, 94.19% and 90.46%, respectively. For COD loading concentrations of 4500 and 5500 mg/L, the treated effluents obtained complied with DOE standard in terms


of COD values but failed to comply for BOD3. Longer HRT would be required for better BOD3 removal if COD loading rate is over 3500 mg/L.

3.5 Permeability test

Permeability test was conducted to determine the integrity of the membranes for natural rubber wastewater treatment. Figure 8 is a flux vs. TMP plot for permeability test of the new membrane. Similar linear relationships were obtained for the membranes after every experiment was conducted. The slopes of the graph represent membrane permeability. Table 5 gives the slope values and flux recovery after every experiment.

Figure 8 Permeability test for new membrane (see online version for colours)

Table 5 Permeability test results

Permeability test of membranes, (Lp)

Slope of permeability test graph (m3/m2h.kPa × 10−5)

Flux recovery (Lp of after tests/Lp of new membranes) (%)

New membrane 8.69 100 After critical flux test 7.89 90.79 After optimum flux test 7.51 86.42 After determination of COD loading 7.30 84

The flux recovery of the membranes after the first critical flux experiments was 90.97%. After the optimum flux experiments, the flux recovery dropped to 86.42%. Finally, a flux recovery of 84% was observed after the experiments with different COD loadings.

Reductions of flux recovery to less than 80% indicates that the membrane has been fouled seriously. In such circumstances, further cleaning would be necessary or the membrane should be replaced (Shamel, 1999). The 84% recovery in this study means that the membrane is applicable for natural rubber wastewater treatment.


4 Conclusions

This study has shown that wastewater from latex concentrate processing factory can be treated with an MBR. There was a 20% reduction on the membrane permeability after treating the wastewater. The critical flux of the process was determined as 0.01224 m3/m2h using the flux step method. The optimum flux of the process was obtained at 0.009 m3/m2h. At this flux rate, the process could be operated more than one month continuously without chemical cleaning of the membranes. The optimum COD loading for the system studied here was 3500 mg/L. The effect of COD loading on permeate quality was also determined. If the COD loading was more than 4000 mg/L, a longer HRT was required to get better COD and BOD removal efficiencies that meet the DOE discharge limits.

The MBR system at a flux of 0.009 m3/m2h had better removal efficiencies over the oxidation ditch process in all parameters. This can be seen from Table 6. Without an anoxic process, AN and TN removal efficiencies were 65% and 61%, respectively, in the treated effluent, which complied with the DOE standard. The MBR system has thus great application potential for natural rubber latex concentrate processing wastewater treatment due to its better removal efficiency, shorter HRT, less sludge production and less space requirement. The modular nature of the system makes it amenable for increased throughput of the effluent.

Table 6 Removal efficiency for oxidation ditch system and MBR process

Parameters Oxidation ditch system (%) MBR process (%)

BOD3 96.03 96.78 COD 96.62 96.99 Suspended solids 75.56 77.91 Total Nitrogen 59.14 65.17 Ammoniacal Nitrogen 49.46 61.35

References Ahn, K.H. and Song, K.G. (2000) ‘Application of microfiltration with a novel fouling control

method for reuse of wastewater from a large-scale resort complex’, Desalination, Vol. 129, pp.207–216.

APHA Standard Method, 2540 D (1995) American Public Health Association, American Waste Works Association, Water Pollution Control Federation (APHA-AWWA-WPFC) (1995) Standard Methods for Examination of Water and Wastewater, 19th ed., America Public Health Association, Washington DC, USA.

APHA Standard Method, 2540 E (1995) American Public Health Association, American Waste Works Association, Water Pollution Control Federation (APHA-AWWA-WPFC) (1995) Standard Methods for Examination of Water and Wastewater, 19th ed., America Public Health Association, Washington DC, USA.

Benitez, J., Rodriguez, A., and Malaver, R. (1995) ‘Stabilisation and dewatering of wastewater using hollow fibre membranes’, Wat. Res., Vol. 29, pp.2281–2286.


Chin, P.S., Singh, M.M., John, C.K., Karim, M.Z.A., Bakti, N.A.K., Sethu, S. and Yong, W.M. (1978) ‘Effluents from natural rubber processing factories and their abatement in Malaysia’, Paper presented at Int. Conf. on Water Pollution Control in Developing Countries, 21–25 February, Bangkok, Thailand.

Cho, B.D. and Fane, A.G. (2002) ‘Fouling transients in nominally sub-critical flux operation of a membrane bioreactor’, J. Membrane Sci., Vol. 209, pp.391–403.

Choo, K.H. and Stensel, H. (2000) ‘Sequencing batch membrane bioreactor treatment: nitrogen removal and membrane fouling evaluation’, Wat. Env. Res., Vol. 72, No. 4, pp.490–498.

Davis, M.L. and Cornwell, D.A. (1998) Environmental Engineering, 3rd ed., McGraw-Hill, New York.

Defrance, L. and Jaffrin, M.Y. (1999) ‘Reversibility of fouling foamed in activated sludge filtration’, J. Membrane Sci., Vol. 157, pp.73–84.

Department of Environment (1995) Industrial Processes and the Environment (Handbook No. 3): Raw Natural Rubber Industry, Aslita Sdn Bhd, Kuala Lumpur.

Field, R.W., Wu, D., Howell, J.A. and Gupta, B.B. (1995) ‘Critical flux concept for microfiltration fouling’, J. Membrane Sci., Vol. 100, pp.259–272.

Harunsyah, Y. and Sulaiman, N.M. (2002) ‘Treatment of skim latex serum using gas sparged ultrafiltration’, Proc. Regional Symp. on Chem. Engineering (RSCE Chem. Engineers –SOMChe 2002), 28–30 October, Malaysia.

Hasar, H., Kinaci, C. and Umlu, A. (2002) ‘Viability of microbial mass in a submerged membrane bioreactor’, Desalination, Vol. 150, pp.263–268.

Howell, J.A. (1995) ‘Sub-critical flux operation of microfiltration’, J. Membrane Sci., Vol. 107, pp.165–171.

Isa, Z., Ibrahim, A., Bakti, N.A.K. and Karim, M.Z.A. (1988) ‘Treatment of rubber factory effluent – a survey’, Notes on 4th Workshop on Rubber and Palm Oil Effluent Testing, 9–10 March, Ayer Keroh, Melaka.

John, C.K. and Ong, C.T. (1979) ‘Ponding as a treatment system for effluent from rubber factories in Malaysia’, Proc. 2nd Symp. on Our Environment, Singapore, p.400.

Kubota (2004) Submerged Membrane Bioreactor Kubota Trade Brochure. Lau, C.M. (1994) ‘Concentration of natural rubber serum by reverse osmosis’, J. Nat. Rubb. Res.,

Vol. 9, No. 4, pp.226–240. Le-Clech, P., Jefferson, B., Chang, I.S. and Judd, S. (2003) ‘Critical flux determination by the

flux-step method in a submerged membrane bioreactor’, J. Membrane Sci., Vol. 227, pp.81–93.

Lim, L.P. and Ujang, Z. (2004a) ‘Critical flux of membrane bioreactor for digested palm oil mill effluent (POME) treatment’, Abstract book Asiawater Conf., 30–31 March, Kuala Lumpur, p.64.

Lim, L.P. and Ujang, Z. (2004b) The Operating Strategy of Flat Sheet Membrane Bioreactor for Palm Oil Mill Effluent Treatment, Institute of Environment and Water Resource Management (IPASA) Internal paper.

Madaeni, S.S. (1997) ‘The effect of operating conditions on critical flux in membrane filtration of latexes’, Trans. IChemE, Vol. 75B, pp.266–269.

Muller, E.B., Stouthamer, A.H., van Verseveld, H.W. and Eijelboom, D.H. (1995) ‘Aerobic domestic wastewater treatment in a pilot plant with complete sludge retention by cross-flow filtration’, Wat. Res., Vol. 29, No. 4, pp.1179–1189.

Nah, Y.M., Ahn, K.H. and Yeom, I.T. (2000) ‘Nitrogen removal in household wastewater treatment using an intermittently aerated membrane bioreactor’, Env. Tech., Vol. 21, pp.107–114.

Ong, S.A., Toorisaka, E., Hirata, M. and Hano, T. (2004) ‘Effects of nickel (II) addition on the activity of activated sludge microorganisms and activated sludge process’, J. Hazardous Materials, Vol. B113, pp.111–121.


Pollice, A., Laera, G. and Blonda, M. (2004) ‘Biomass growth and activity in a membrane bioreactor with complete sludge retention’, J. Wat. Res., Vol. 38, pp.1799–1808.

Rosenburger, S., Kraume, M. and Szewzyk, U. (1999) ‘Operation of different membrane bioreactors experimental results and physiological state of the microorganism’, Proc. IWA Conf. Membrane Tech. in Environmental Management, Tokyo, pp.310–316.

Shamel, M.M. (1999) Ultrafiltration Studies on Pectin-Glucose Solution with Pulsatile Flow in-situ Cleaning, MEngSc Thesis, Dept. of Chemical Engineering, University of Malaya.

Ujang, Z. and Le-Clech, P. (2003) ‘Membrane technology for water and wastewater treatment’, Notes from Specilized Workshop on Membrane Technology for Water and Wastewater Treatment, 10–11 July, UTM Kuala Lumpur,

van-Houten, R.T. and Eikelboom, D.H. (1997) ‘High performance membrane bioreactors: a physiological approach’, Proc. MBR1 – First Int. Meeting on MBRs for Wastewater Treatment, 5 March, Cranfield University, UK.

Veerasamy, D., Sulaiman, N., Nambir, J. and Yusof, A. (2003) ‘Environmental friendly natural rubber latex concentration by membrane separation technology’, Proc. Int. Membrane Science and Technology Conference (IMSTEC 03), 10–14 November, University of New South Wales, Sydney, Australia.

Documents

Nik Meriam Nik Sulaiman* Shaliza Ibrahim Sarah Lim Abdullah · Membrane Bioreactor for the treatment of natural rubber wastewater 95 The objectives are: i to adapt available laboratory