The effect of mixed liquor suspended solids - Universiti Teknologi ...

1701 & IWA Publishing 2011 Water Science & Technology 9 63.8 9 2011 

The effect of mixed liquor suspended solids (MLSS) on 

biofouling in a hybrid membrane bioreactor for the 

treatment of high concentration organic wastewater 

A. Damayanti, Z. Ujang, M. R. Salim and G. Olsson 

ABSTRACT 

Biofouling is a crucial factor in membrane bioreactor (MBR) applications, particularly for high 

organic loading operations. This paper reports a study on biofouling in an MBR to establish a 

relationship between critical flux, Jc, mixed liquor suspended solids (MLSS) (ranging from 

5to20gL1 ) and volumetric loading rate (6.3 kg COD m 3 h 1 ) of palm oil mill effluent (POME). 

A lab-scale 100 L hybrid MBR consisting of anaerobic, anoxic, and aerobic reactors was used with 

flat sheet microfiltration (MF) submerged in the aerobic compartment. The food-to-microorganism 

(F/M) ratio was maintained at 0.18 kg COD kg 1 MLSSd 1 . The biofouling tendency of the 

membrane was obtained based on the flux against the transmembrane pressure (TMP) 

behaviour. The critical flux is sensitive to the MLSS. At the MLSS 20 g L 1 the critical flux 

is about four times lower than that for the MLSS concentration of 5 g L 1 . The results showed 

high removal efficiency of denitrification and nitrification up to 97% at the MLSS concentration 

20 g L 1 . The results show that the operation has to compromise between a high and a low MLSS 

concentration. The former will favour a higher removal rate, while the latter will favour a higher 

critical flux. 

Key words 9 biofouling, flux, membrane bioreactor, MLSS, TMP 

INTRODUCTION 

Palm oil mill has significantly contributed to the generation of 

wastewater in Malaysia. As the world’s largest producer and 

exporter of palm oil, Malaysia has more than three million 

hectares of oil palm cultivated area. More than 500 palm oil 

mills have been built and operated for the past 40 years 

(Damayanti et al. 2010). Palm oil processing is carried out 

in mills where oil is extracted from the oil palm fruit bunch 

(Vijayaraghavan et al. 2007). Extraction of crude palm oil 

needs large quantities of water, contributing to about 50% 

of the POME. POME is characterized as a thick brownish 

liquid containing high amounts of total solids (E40 g L 1 ), 

oil and grease (E8 gL 1 ), COD (E43 g L 1 ), BOD 

(E25 g L 1 ), ammonia nitrogen (E0.1 g L 1 ), and pH of 

3.570.1 (Ahmad et al. 2008). 

doi: 10.2166/wst.2011.338 

A. Damayanti 

Z. Ujang (corresponding author) 

M. R. Salim 

Institute of Environmental & Water 

Resource Management (IPASA), 

Universiti Teknologi Malaysia (UTM), 

81310 Skudai, 

Johor Bahru, 

Malaysia 

E-mail: zaini@utm.my 

G. Olsson 

Industrial Electrical Engineering and 

Automation (IEA), 

Lund University, 

Box 118, SE-221 00 Lund, 

Sweden 

In recent years, membrane technology has been used 

basically for low strength wastewater like domestic wastewater 

in aerobic conditions similar to the conventional 

activated sludge process (Hamdzah et al. 2005; Ujang et al. 

2007). The membrane bioreactor (MBR) has been successfully 

used for treatment of municipal and industrial wastewaters 

(Alaboud 2009). The research reported in this paper has been 

conducted under different operating conditions using a 

hybrid MBR (HMBR) with a combination of anaerobic, 

anoxic, and aerobic reactors to treat POME. A sustainable 

process performance in the last aerobic tank was obtained at 

different levels of MLSS concentration 5 to 20 g L 1. The 

membrane replaces the function of the settling tanks that is 

conveying the supernatant to the water body as well as

1702 A. Damayanti et al. 9 The effect of mixed liquor suspended solids (MLSS) on biofouling in a hybrid membrane bioreactor 

Water Science & Technology 9 63.8 9 2011 

returning the solids back to the reactor (Qin et al. 2007). 

Internal biofouling due to the adsorption of dissolved matter 

into the membrane pores and pore blocking is considered 

irreversible and is generally only removed by chemical cleaning. 

There are principally three strategies for limiting biofouling 

in submerged MBR process systems: maintaining the 

permeate flux below the critical flux, Jc, increasing the membrane 

aeration, and employing physical or chemical cleaning 

(Razak et al. 2007). Under constant-TMP or constant-flux 

conditions, the fouling mechanism will increase hydraulic 

resistance and the trans membrane pressure will be increased 

as permeate flux will be declined, as further studied by Qin 

et al. (2007). 

Many studies about biofouling at MBR in different MLSS, 

for several example Chang & Kim (2005) using synthetic 

wastewater, showed which cake resistance (Rc) decreased 

as MLSS concentration (between 0.09–3.7 g L 1 ) decreasing. 

Another study compared attached growth (MLSS between 

0.1–2 g L 1 ) and suspended growth (MLSS 3 g L 1 )microorganisms 

in MBR using municipal wastewater showed the 

rate of attached growth system 7 times higher affected the 

biofouling rate compare with suspended growth (Lee et al. 

2001). Therefore this study was carried out to investigate 

performance and biofouling tendency at four different 

MLSS concentrations between 5 to 20 g L 1 in a hybrid 

MBR treating high strength wastewater, like POME. Thus 

the study has two aims: (1) to observe the biofouling through 

the critical flux, J c, which was determined using the flux step 

method and (2) to investigate the performance of the organic 

and nitrogen removal in the system at different MLSS levels. 

MATERIALS AND METHODS 

POME samples and pretreatment process 

The characteristics of raw POME collected from Felda Bukit 

Besar Palm Mill, Johor, Malaysia are shown in Table 1. Upon 

sampling the temperature of POME was measured at 

approximately 801C751C. POME samples used in this 

study were stored at 41C. The operating temperature was 

251C (room temperature in a tropical country). The POME 

was mixed in the reactor using a stirrer according to Hofvendahl 

et al. (1999) for anaerobic compartment and Ramalho 

et al. (2004) for anoxic compartment, for agitation speed 

120 rpm or Newton number 4 to maximize the process. 

Finally the POME was collected in the feed tank to study 

the performance of different levels of MLSS. The F/M was 

calculated using Equation (1) based on measurements taken 

Table 1 9 Characteristics of POME compared to municipal sewage 

Symbol POME Values a Sewage b,c 

pH 5.6 7 

Tot COD 45 0.60 

TS (Total Solid) 43 0.80 

TDS (Total Dissolved Solid) 35 0.23 

SS (Suspended Solid) 8.2 0.56 

NVSS (Non Volatile 

Suspended Solid) 

from the experiment. The values of F/M range from 0.7 to 

1kgBOD5 kg MLSS 1 day 1 (Metcalf and Eddy 2004). 

F QS0 

¼ 

M VX 

where F/M ¼ food to microorganism ratio 

S0 ¼ influent concentration (ML 3 ) 

Q ¼ flowrate (L3T 1 ) 

V ¼ volume (L3 ) 

X ¼ biomass concentration (ML 3 ). 

Laboratory scale trials 

4.0 0.25 

NO3 0.10 0.01 

NO2 0.30 0.06 

NH 4 þ 

0.20 0.03 

Total Nitrogen 0.50 0.01 

*Note: 

a This study, 

b Nor-Anuar et al. (2007), 

cUjang et al. (2008) 

All parameter’s units in gL 1 except pH. 

The schematic process flow diagram for the lab scale HMBR 

is shown in Figure 1. A specification of the HMBR system is 

given in Table 2. For acclimatization the HMBR system was 

seeded with biological sludge taken from the wastewater 

treatment plant of the Felda Bukit Besar Palm Mill, Johor 

Bahru. The experiments were carried out at room temperature 

251C. The system was operated continuously (24-h) 

during this study, with acclimatization taking placed within 

two months and a total study duration of six months. Trials 

for different MLSS levels were conducted to obtain biofouling 

conditions. The HRT of the anaerobic, anoxic, and aerobic 

reactors, as shown in Table 2, were 11, 7, and 8 hours 

respectively. The flat sheet membrane (15cm 10cm) used 

ð1Þ



Figure 1 9 Schematic process flow diagram. 

in this study had been kindly donated by Kubota, Inc. and was 

vertically submerged in the bioreactor. The size of the membrane 

bioreactor is 24cm 10cm 90cm. The height between 

the membrane module and the base of the reactor is 30 cm. 

The TMP was measured using an online system by installing a 

pressure transducer. For the aeration, the compressed air 

supplied individually and the flowrate has been used is 12 

LMH. Recirculation flow rate in this study was 1:3. 

The system is operated for acclimatization to the new 

media was about one a half months before starting experiments. 

The sludge divided into 4, for MLSS study 5, 10, 15 

and 20 g L 1 . After that, the sludge placed in cool room. 

Then the system operated using lowest MLSS first, 5 g L 1 , 

acclimatized again around one to two weeks until steady state 

phase, and finished it in one month. After that, the sludge was 

replaced from reactor, using new sludge (from cool room) 

which have acclimatized before, but it is still need acclimatized 

again around one to two weeks until steady state phase, 

using MLSS 10 g L 1 , until finished in one month. The same 

step as MLSS 15 and 20 g L 1 . We excess sludge withdrawal 

and stationary MLSS for same SRT, 70 days. All step study 

finished in four months excluded acclimatization in two 

Table 2 9 The Specifications of the HMBR system 

Item Specifications 

Membrane type Flat sheet 

Membrane material Chlorinated Polyethylene 

Membrane pore size 0.4 mm 

Membrane area per module 0.1 m 2 

Number of modules 3 

Effective volume of anaerobic reactor 50 L 

Effective volume of anoxic tank 30 L 

Effective volume of aeration tank 20 L 

months. Flux 11 Lm 2 h 1 applied for operating condition 

in HMBR. 

During the studies, the fouled membranes were removed 

from the aeration reactor and cleaned with a sponge followed 

by water flush cleaning. The membranes were then soaked in 

a 0.5% ppm NaOCl and left overnight (Ujang et al. 2005). 

After this, the cleaned membranes were ready to be used 

again in the process (Ujang et al. 2005). 

Critical flux evaluation 

The flux step method is used to determine the critical flux 

(Le-Clech et al. 2003). The critical flux measured at the start 

for each period. This carried out by increasing values of 

permeate flux and recording the relative TMP variations. 

For each flux step two TMP values were considered: the 

initial TMP, corresponding to the initial sudden increase of 

the filtration resistance, and the final TMP, i.e. the TMP at the 

end of the step (Le-Clech et al. 2003). From these two TMP 

values two parameters connected to fouling can be evaluated, 

the average TMP and the rate of change of the TMP (dTMP/ 

dt). Any flux with dTMP/dtZ0.5 is defined as the critical flux 

(Bottino et al. 2009). By using resistance series model as in 

Equation (2), Rtot or total filtration resistance then could be 

calculated (Chang & Kim 2005): 

Rtot ¼ Rm þ Rc þ R f 

where Rtot is the total filtration resistance, Rm is the intrinsic 

membrane resistance (obtained from pure water permeability 

measurements), Rc is the cake resistance, and Rf is the fouling 

resistance due to irreversible adsorption and pore plugging 

(Bottino et al. 2009). Rm and Rtot were calculated according to 

Darcy’s law below: 

R ¼ TMPm 1 J 1 

ð2Þ 

ð3Þ



J (Lm –2 h –1 ) 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

where m is the fluid viscosity and J is the flux. Permeate flux (J) 

applied from 0 to 40 Lm 2 h 1 for critical flux test, 15 minute 

for one value constant. The critical flux measured at the start 

each test period pf MLSS. Each MLSS concentration used 

new membrane. 

Samples analysis 

Laboratory experiments were carried out in the Environmental 

Engineering Laboratory, UTM. All experimental analysis 

was conducted according to Standard Methods (APHA 1998) 

within three hours after sampling. MLSS concentrations were 

regularly measured. Chemical Oxygen Demand, NO 3-N, 

NO2-N, NH4-N, and PO4-P were analysed using a spectrophotometer 

(HACH/DR 4000). The measurements were 

carried out in duplicate according to HACH testing procedures. 

The average values were taken for different MLSS 

concentrations between 5 and 20 g L 1 .Dissolvedoxygen 

concentration, temperature and pH were monitored using an 

on-line measurement system, and the biotransformation was 

observed using on-line respirometry. The DO probe used in 

this work was the WTW Oxil oxygen meter. It was calibrated 

with pure water before being used in the membrane bioreactor. 

RESULTS AND DISCUSSION 

MLSS 5 gL –1 




0 100 200 300 400 500 

TMP (mbar) 

Figure 2 9 TMP and J behavior during ascending and descending runs at different MLSS 

concentrations with Y error bar 5%. 

Behaviour of membrane critical flux in MBR operation 

Figures 2 and 3, andTable 3 show the influence of different 

MLSS concentrations on the TMP and biofouling behaviour. 

dTMP/dt (mbar/sec) 

18 

16 

14 

12 

10 

8 

6 

4 

2 

0 

0 5 10 15 20 25 

Jc (Lm –2h –1 MLSS 5 gL 

) 

–1 




Figure 3 9 Trend of dTMP/dt (mbar/sec) as a function of the flux at MLSS concentrations 5– 

20 gL 1 with Y-error bar 5%. 

The values in Table 3 are calculated using Equation (3). The 

results clearly demonstrate that the membrane biofouling rate 

increases with an increase in the MLSS concentration. The 

results obtained indicate that the MLSS concentration has a 

significant influence on the trans membrane pressure (TMP). 

Figure 2 shows the flux vs. TMP behaviour, during the 

ascending and descending run of membrane with POME at 

different MLSS concentrations. The differences between two 

runs were significant. 

During the ascending run, the flux first proportionally 

increased with an increase in the TMP and leveled off, more 

quickly at higher MLSS concentrations, due to membrane 

biofouling. The detrimental effect of fouling on the membrane 

flux is also apparent by observing the behaviour of the 

membrane flux as a function of the applied TMP during the 

descending run. 

Figure 3 shows the trend of dTMP/dt as a function of the 

flux determined from the ascen-ding runs of various hysteresis 

loops of Figure 2. From the reported trend the critical 

flux, i.e. the flux corresponding to dTMP/dt ¼ 0.5 can 

be easily calculated. The critical flux values are reported in 

Table 3, along with the values of various resistances. Figure 3 

also shows that by increasing the MLSS concentration from 5 

to 20 g L 1 the critical flux becomes four times lower. 

Although the Rm value of the membrane was very low, 

Table 3 9 Effect of MLSS concentration on critical flux and resistances of POME 

No MLSS (g/L) Jc (Lm 2 h 1 ) Rtot (10 12 m 1 ) Rm (10 12 m 1 ) Rc þ Rf (10 12 m 1 ) 

1 5 9.2 1.73 0.19 1.54 

2 10 7.1 2.04 0.19 1.85 

3 15 4.7 2.83 0.19 2.64 

4 20 2.5 3.90 0.19 3.71

cake formation and fouling occurred during the treatment 

process. Table 3 demonstrates that higher MLSS will lead to 

higher R tot. TheR m value led to a significant increase of the 

total resistance, since biofouling and cake formation take 

place, especially at high MLSS concentration (Bottino et al. 

2009). Unlike Bottino et al. (2009), Chang & Kim (2005) 

showed that the Rm value not always led to a significant 

increase of the total resistance, especially at low MLSS 

concentration. This is in line also with Chang & Kim (2005) 

that found that R c, cake resistance, will decrease as MLSS 

between 3.7 to 0.09 g L 1 decrease. 

Chang & Kim (2005) also showed that higher MLSS 

concentrations lead to higher higher Rtot. The rate of TMP 

increase was found to generally increase with increasing 

MLSS concentration at each flux tested because of higher 

suspended solids in the reactor (Bottino et al. 2009). Furthermore, 

at each MLSS concentration, dTMP/dt increased with 

increasing flux-step. In addition, Rosenberger & Kraume 

(2002) who found that higher SS concentration affected 

lower F/M ratio and also lower EPS concentration, led to 

better filterability. 

COD, NO 3 ,NO 2 ,andNH 4 þ removal efficiency at 

different flux and MLSS 

COD, NO3 ,NO2 ,andNH4 þ removal efficiencies at different 

MLSS are shown in Figure 4. Towards the end of the 

experimental period the feed COD concentration had an 

average of 40.5 g L 1 , while the effluent concentration was 

kept in the range 200 to 350 mg L 1 depending on the POME 

source quality and MLSS used in the MBR. These mean 

values here is continuous with acclimatized sludge from lab 

scale hybrid MBR. This effluent quality need to be treated 

further to meet effluent quality standard (Environmental 

Quality Regulation 2009). 

Figure 4 also demonstrates a COD removal efficiency of 

up to 99%. The NO 3-N effluent concentration depends on the 

Removal efficiency (%) 



100 

99.5 

99 

98.5 

98 

97.5 

97 

96.5 

5 10 15 20 

MLSS (gL –1 ) 

COD 

NO 3 – 

NO 2 – 

NH 4 + 

Figure 4 9 COD, NO3 ,NO2 ,andNH4 þ removal efficiency at different MLSS concentrations of 

the POME with Y error bar 0.2%, respectively. 

POME source quality and the average removal rate is up to 

97%. This is comparable with other studies, where NO3 

removal using aerobic treatment was up to 95% (Kumbasar 

2009) and 95% NH4 þ removal (Sarioglu et al. 2009). In Figure 

4 it is shown that the rate of removal efficiency generally 

increases with increasing MLSS concentration at each flux 

tested because of higher biomass activity working in the 

reactor (Alaboud 2009) Furthermore, at each MLSS concentration, 

dTMP/dt increased with increasing flux-step. This is 

why balancing the biofouling tendency and effluent quality is 

needed to reduce operation and maintenance costs in a 

wastewater treatment plant (Ujang et al. 2008). It is not a 

big difference but it is important for the effluent quality at 

high MLSS concentrations. However it is also important to 

consider the biofouling effect at high MLSS concentrations. 

CONCLUSIONS 

This study has indicated that the membrane biofouling rate 

increases with an increasing MLSS concentration, as 

observed through the critical flux analysis and extensive 

experimental results for POME. During the experiments of 

increasing MLSS the flux first increased proportionally with 

the increase in TMP and levelled off, more quickly at higher 

MLSS concentration, due to membrane biofouling. The detrimental 

effect of fouling on the membrane flux is also apparent 

by observing the behaviour of the membrane flux as a function 

of the applied TMP during the descending run. The rate 

of change in the TMP was found to generally increase with 

increasing MLSS concentration at each flux tested. The study 

also showed that the removal increased with increasing 

MLSS concentrations. In general, the results from this 

paper indicate the need to balance the biofouling tendency 

and the effluent quality to reduce operation and maintenance 

costs in wastewater treatment plants, for high strength wastewater 

such as POME. 

ACKNOWLEDGEMENTS 

This study was made possible by the financial support from 

the Ministry of Science of Technology (MOSTI) Malaysia 

through eScience funding (Vot 79006) and Islamic Development 

Bank for providing scholarship. Thanks to Mr. Zulkifli 

Ahmad for special assistance and discussions, and thanks to 

Institute Technology of Sepuluh Nopember for supporting the 

first author. Membrane modules were kindly donated by 

Kubota Inc.



REFERENCES 

Ahmad, Z., Ujang, Z., Yunus, S., Damayanti, A. & Olsson, G. 2008 A 

Novel Approach for Palm Oil Mill Effluent (POME) Treatment 

And Recovery Using Hybrid Membrane Bioreactor. Asia Water 

Conference. Kuala Lumpur. 

Alaboud, S. T. M. 2009 Membrane Bioreactor for Wastewater Reclamation-Pilot 

Plant Study in Jeddah, Saudi Arabia. J. of Env. Sci. 3, 

267–277. 

APHA 1998 Standard Methods for the Examination of Water and 

Wastewater, American Public Health Association: Washington 

DC, USA. 

Bottino, A., Capannelli, G., Comite, A. & Mangano, R. 2009 Critical flux 

in submerged membrane bioreactors for municipal wastewater 

treatment. Desalination 245(1–3), 748–753. 

Chang, I. S. & Kim, S. N. 2005 Wastewater treatment using membrane 

filtration—effect of biosolids concentration on cake resistance. 

Proc. Biochem. 40, 1307–1314. 

Damayanti, A., Ujang, Z., Salim, M. R., Olsson, G. & Sulaiman, A. Z. 

2010 Respirometric analysis of activated sludge models from palm 

oil mill effluent. Biores. Technol. 101(1), 144–149. 

Environmental Quality (Industrial Effluent) Regulation 2009 Malaysia. 

Hamdzah, M., Ujang, Z. & Ozaki, H. 2005 Reverse Osmosis (RO) for 

Micropollutant Rejection: The Effect of Low Pressure System. 

Water Malaysia No: 10 (August 2005). 

Hofvendahl, K., van Niel, W. J. & Hahn-HaÈ gerdal, B. 1999 Effect of 

temperature and pH on growth and product formation of Lactococcus 

lactis ssp. lactis ATCC 19435 growing on maltose. J. Appl 

Microbiol Biotechnol. 51, 669–672. 

Kumbasar, R. A. 2009 Separation and concentration of cobalt from zinc 

plant acidic thiocyanate leach solutions containing cobalt and 

nickel by an emulsion liquid membrane using triisooctylamine as 

carrier. J. Mem. Sci. 333(1–2), 118–124. 

Le-Clech, P., Jefferson, B., Chang, I. S. & Judd, S. J. 2003 Critical flux 

determination by flux-step method in a submerged membrane 

bioreactor. Journal Membrane Science 227(1–2), 81–93. 

Lee, J., Ahn, W. Y. & Lee, C. H. 2001 Comparison of the filtration 

characteristics between attached and suspended growth microorganisms 

in submerged membrane bioreactor. Wat. Res. 35(10), 

2435–2445. 

Metcalf. and Eddy 2004 Wastewater Engineering: Treatment and Reuse. 

Singapore, McGraw-Hill. 

Nor-Anuar, A., Ujang, Z., van Loosdrecht, M. C. M. & de Kreuk, M. K. 

2007 Settling behaviour of aerobic granular sludge. Wat. Sci. Tech. 

56(7), 55–63. 

Qin, J. J., Oo, M. H., Tao, G. & Kekre, K. A. 2007 Feasibility study on 

petrochemical wastewater treatment and reuse using submerged 

MBR. J. Mem. Sci. 293(1–2), 161–166. 

Ramalho, P. A., Cardoso, M. H., Paulo, A. C. & Ramalho, T. 2004 

Characterization of Azo Reduction Activity in a Novel Ascomycete 

Yeast Strain. J. Appl. Env. Microbiology 70(4), 2279–2288. 

Razak, A. R. A., Ujang, Z. & Ozaki, H. 2007 Removal of endocrine 

disrupting chemicals (EDCs) using low pressure reverse osmosis 

membrane. Wat. Sci. Tech. 56(8), 161–168. 

Rosenberger, S. M. & Kraume 2002 Filterability of activated sludge in 

membrane bio-reactors. Desalination 146, 373–379. 

Sarioglu, M., Insel, G., Artan, N. & Orhon, D. 2009 Model evaluation of 

simultaneous nitrification and denitrification in a membrane 

bioreactor operated without an anoxic reactor. J. Mem. Sci. 

337(1–2), 17–27. 

Ujang, Z., Ng, K. S., Hamzah, T. H. T., Roger, P., Ismail, M. R., 

Shahabudin, S. M. & Abdul Hamid, M. H. 2007 Application of 

immersed MF (IMF) followed by reverse osmosis membrane for 

wastewater reclamation: The case of Malaysia. Wat. Sci. Tech. 

56(9), 103–108. 

Ujang, Z., Ng, S. S. & Nagaoka, H. 2005 Package plant of extended 

aeration membrane reactors: A study on aeration intensity and 

biofouling control. Wat. Sci. Tech. 51(10), 335–342. 

Ujang, Z., Talib, J. & Sairan, F. M. 2008 Rationalization of Small-sizes 

Wastewater Treatment Plant: The Options and Strategic Policy for 

Johor Bahru Catchments, Malaysia. Proceedings of the International 

Water Association (IWA) Conference on Sanitation, 

November 2008, Hanoi, Vietnam. 

Vijayaraghavan, K., Ahmad, D. & Aziz, M. E. A. 2007 Aerobic treatment 

of palm oil mill effluent. J. Env. Man. 82(1), 24–31.

The effect of mixed liquor suspended solids - Universiti Teknologi ...

Create successful ePaper yourself

Delete template?

Save as template?