Influence of Mixed Liquor Properties and Aeration Intensity on Membrane Fouling

in a Submerged Membrane Bioreactor at High Mixed Liquor Suspended Solids

Concentrations

by

R. Shane Trussell1, Ph.D., P.E.

Rion P. Merlo2, Ph.D.

Slawomir W. Hermanowicz3, Ph.D.

and David Jenkins4, Ph.D.

1 Principal, Trussell Technologies, Inc., 232 N. Lake Ave., Ste. 300, Pasadena, CA 91101

shane.trussell@trusselltech.com, P:626-486-0560, F:626-486-0571 2 Senior Engineer, Brown and Caldwell, 201 N. Civic Dr., Ste. 115, Walnut Creek, CA

94596 3 Associate Professor, Department of Civil and Environmental Engineering, University of

California, Berkeley, CA 94720 4 Professor in the Graduate School, Department of Civil and Environmental Engineering,

University of California, Berkeley, CA 94720

2

ABSTRACT

This paper presents the results of 195 days of pilot-scale submerged membrane bioreactor

(SMBR) experiments on settled municipal wastewater. Short-term and long-term

thickening experiments were performed at a constant membrane flux of 30 LMH to

determine the impact of the following mixed liquor properties: colloidal material, soluble

COD, soluble microbial products, extracellular polymeric substances, and viscosity along

with aeration intensity on membrane fouling at high mixed liquor suspended solids

(MLSS) concentrations. The normalized specific flux declined with increasing MLSS

concentrations in all experiments and increasing the coarse bubble aeration intensity

increased the specific flux at a given MLSS concentration. Using a dynamic approach,

this work demonstrates the importance of mixed liquor viscosity, which impacts the

efficacy of the coarse bubble aeration, in sustaining membrane permeability. Over an

extended thickening time period, a small increase in MLSS concentration and mixed

liquor viscosity becomes more prevalent and leads to greater specific flux decline at a

given MLSS concentration.

KEYWORDS

Submerged membrane bioreactor (SMBR), aeration intensity, membrane fouling,

viscosity, soluble microbial products (SMP), extracellular polymeric substances (EPS)

3

INTRODUCTION

The membrane bioreactor (MBR) process can operate at high MLSS concentrations

because a membrane, rather than a gravity sedimentation basin, is used for solid-liquid

separation (Chiemchaisri and Yamamoto 1994; Trussell et al. 2005; Yamamoto et al.

1989). Increased MLSS concentrations allow reduced hydraulic residence times (HRTs)

at a given mean cell residence time (MCRT) resulting in a more compact treatment

process (Trussell et al. In Press). For a given HRT, higher MLSS concentrations allow

operation at increased MCRTs resulting in reduced sludge production and other

advantages (Adham et al. 2001; Innocenti et al. 2002; Metcalf&Eddy 2003; Muller et al.

1995; Scholzy and Fuchs 2000; Wei et al. 2003; Yoon 2003; Yoon and Lee 2005). High

MLSS concentrations can have detrimental effects on membrane performance (Le Clech

et al. 2003; Yamamoto et al. 1989), but the available literature data does not adequately

describe the complex factors influencing membrane permeability at high MLSS

concentrations. Previous research has developed empirical relationships between the

aeration intensity, membrane flux, and MLSS concentrations (Liu et al. 2003; Shimizu et

al. 1996), but the influence of mixed liquor properties, such as bulk mixed liquor

viscosity and colloidal content, have not been incorporated. Research has shown that

increasing bulk viscosity will decrease the mass transfer coefficient of material from the

membrane surface to the bulk solution (Pritchard et al. 1995) and bulk viscosity has been

shown to increase with increasing MLSS concentration (Dick and Ewing 1967; Krampe

and Krauth 2003). Recent work has also demonstrated the importance of the organic

colloidal mixed liquor fraction on long-term membrane fouling rates and the membrane

critical flux (Fan et al. 2006; Rosenberger et al. 2006; Trussell et al. In Press). The goal

4

of this research was to examine the effects of aeration intensity and mixed liquor

properties on membrane permeability at high and changing MLSS concentrations.

MATERIALS AND METHODS

Pilot-Scale Submerged Membrane Bioreactor (SMBR): A pilot-scale SMBR,

designed to operate at a range of variable hydraulic retention times (HRTs) while

maintaining a constant membrane flux of 30 L/m2-h (LMH) and a working volume of

1,514 L was custom built by ZENON Environmental Services Inc., Ontario, Canada. A

complete description of this pilot unit is provided in (Trussell et al. In Press). ZW500C

ultrafiltration modules provided a total membrane surface are of 61.3 m2 (660 ft2) with a

vertical hollow fiber configuration.

Feedwater Characteristics: The SMBR was fed with primary effluent from the

Southeast Water Pollution Control Plant (SEP), San Francisco, CA and the median COD,

TSS, and TKN concentrations were 345, 99, and 30 mg/L respectively. A centrifugal

pump was immersed in the SEP primary effluent channel and operated continuously.

Reactor Operation: The reactor was operated for at least 2 MCRTs prior to steady-state

data collection at each MCRT. The mixed liquor dissolved oxygen (DO) concentration

was always more than 2 mg/L and sodium bicarbonate was added to the feed wastewater

to elevate its alkalinity and maintain pH in the SMBR above 6.5.

5

Membrane Performance: Membrane fouling at high MLSS concentrations was

investigated with short-term and long-term experiments at 10-d, 20-d and 30-d MCRTs.

The short-term experiments (each lasting less than 1 hour) were performed after

achieving steady state MCRTs of 10, 20 and 30 d at an MLSS concentration of

approximately 15 g/L. Long-term (no-wasting) experiments were performed over several

days using the previously obtained, steady-state conditions as starting points (10, 20 and

30-d MCRTs).

In the short-term experiments, the wastewater feed to the SMBR was stopped. While the

permeate flow was maintained, mixed liquor was pumped from the aeration basin to the

membrane tank and the membrane was used as a dewatering device, increasing the MLSS

concentration in the membrane tank linearly with time. Membrane performance was

observed at a constant flux of 30 LMH (18 gfd) while monitoring MLSS concentration

increase. In addition to MLSS concentration, viscosity and soluble COD (sCOD)

concentrations were determined on all collected samples. Soluble microbial products

(SMP) and colloidal material content were determined on the initial and final mixed

liquor samples and extracellular polymeric substance (EPS) content on the initial mixed

liquor sample. The experiment was repeated at each MCRT with coarse bubble aeration

intensities of 1.5x10-4, 2.3x10-4, and 3.0x10-4 m3/m2.s (m/s). The aeration intensity was the

calculated by dividing the coarse bubble aeration rate (9.4, 14.2, and 18.9 L/s) by the total

membrane area (61.3 m2). At the end of each experiment, the mixed liquor was diluted to

the initial MLSS concentration with membrane permeate and membrane permeability

was restored without chemical cleaning.

6

Extended short-term experiments were performed at the 20-d and 30-d MCRT conditions

by recycling 50% of the membrane permeate back to the system. With the permeate

recycle, the rate of MLSS concentration increase was half of that in the previously

described short-term experiments. The same analytical measurements were performed

while observations of membrane performance were made. The membrane flux and coarse

bubble aeration intensity were constant at 30 LMH and 2.3x10-4 m/s, respectively.

The long-term experiments were performed at the conclusion of the short-term

experiments at each MCRT. The established steady-state condition was used as a starting

point. The SMBR was treating wastewater but sludge wasting was stopped to allow the

MLSS concentration to increase; the mixed liquor was characterized daily (i.e. MLSS,

viscosity, colloidal material, soluble COD, SMP, EPS) until the maximum vacuum

pressure (18 in. Hg) was reached. The membrane flux and coarse bubble aeration rate

were constant at 30 LMH and 2.3x10-4 m/s, respectively.

Membrane performance was measured by a temperature corrected specific flux to 20oC

using the following equation:

!

LP

20oC =

1

µW

20oCR

=J " e

-0.0239 T -20( )( )

#P Equation 1

where,

!

LP

20oC

= specific flux at 20oC, L/m2 .h.bar (LMH/bar)

T = temperature of water, oC

7

!

µW

20oC = absolute viscosity of water at 20oC, kg/m.s

R = total resistance, m-1

J = membrane flux, L/m2 .h (LMH)

ΔP = transmembrane pressure, Pa

It is important to note that the filtration resistance, R, can be obtained by multiplying the

specific flux by the absolute viscosity of water at 20oC. For ease of comparison, the

specific flux data for each experiment was divided by the initial specific flux to determine

a normalized specific flux.

Analytical Methods: MLSS and mixed liquor volatile suspended solids (MLVSS) were

measured according to Standard Methods 2540 D/E (APHA 1998). Mixed liquor

viscosity measurements were made using a rotational disk viscometer (Viscometers UK

Ltd, London – LV spindle #2) for all samples collected during high MLSS experiments.

The rheological properties of the sludge were that of a power-law fluid, or “shear-

thinning”. Mixed liquor sCOD was determined by analyzing the mixed liquor filtrate

from a 0.45 µm nitrocellulose membrane (Fisherbrand Water-Testing Membrane Filters,

Fisher Scientific L.L.C., Hampton, NH) using Standard Method 5220 D (APHA 1998).

Extracellular polymeric substances (EPS) concentrations were measured as carbohydrate

and protein using a cation exchange resin (CER) (Dowex Marathon C, Na+ form,

Sigma-Aldrich, Bellefonte, PA) extraction method (Frolund et al. 1996). A mixed liquor

sample was immediately cooled to 4oC to minimize microbial activity. The exchange

resin (70 g of CER/g VSS) was added to a 50-mL sample and mixed at 600 rpm using a

8

single blade paddle for 2 h at 4oC. The mixture (50 mL) was centrifuged for 15 min at

12,000 g to remove MLSS. Supernatant carbohydrate and protein concentrations were

measured colorimetrically by the methods of Dubois et al. (1956) and Lowry et al.

(1951), respectively. At the same time, 50 mL of untreated mixed liquor was centrifuged

for 15 min at 12,000 g, and the protein and carbohydrate concentrations were determined

on the supernatant to represent the soluble fraction (Soluble Microbial Products, SMP).

The centrifuge supernatant of the mixed liquor sample represented the SMP

concentration, and the centrifuged supernatant of the sample after CER addition

represented the sum of SMP and EPS concentrations. The difference between these

measurements is the EPS concentration. Bovine serum albumin (BSA) was used as a

protein standard, and dextrose was used as a carbohydrate standard.

Mixed liquor colloidal material was measured by a method modified from Wilen et al.

(2000). Mixed liquor from each reactor was centrifuged for 2 min. at 1000g and then the

supernatant turbidity was measured using a Hach 2100N nephelometer (Hach Company,

Loveland, Co).

9

RESULTS

The SMBR pilot plant was operated for 415 d to study the effects of organic loading on

membrane fouling (Trussell et al. In Press). Following completion of these experiments,

the SMBR pilot was operated for an additional 195 d to perform the experiments

presented here.

Mixed Liquor Suspended Solids: The MLSS concentration was initially maintained at

approximately 14 g/L (Figure 1). After completing the 10-d MCRT experiments, the

MLSS concentration was reduced during the 20-d MCRT to approximately 12 g/L by

decreasing the reactor HRT. The MLSS was reduced to ensure that SMBR operation

prior to the high-solids experiments did not result in significant membrane fouling. The

MLSS concentrations for the long-term high solids experiments are shown in Figure 1. In

all experiments, the MLSS concentration approached 18 g/L at the end of the experiment.

Membrane Performance: Membrane performance data for the entire 195 d of operation

are presented in Figure 1. The system was operated at the target flux of 30 LMH a

majority of the time. However, during the 10-d MCRT phase and for part of the 20-d

MCRT phase, the mixed liquor filterability was poor and continuous operation at the

target membrane flux of 30 LMH was not possible. When we attempted to maintain the

membrane flux near the target flux of 30 LMH for a 2-week period (Days 9 to 24) and the

specific flux declined rapidly (due to TMP increase). Because the goal of our high-solids

experiments was to start them after reaching high-solids steady-state conditions, the

10

membrane flux was reduced to 13 LMH to prevent further membrane fouling. The actual

experiments described in this work were conducted at 30 LMH as described in the

previous section. Although SMBRs operated at higher MCRTs have been shown to

reduce membrane fouling (Lee et al. 2003; Ng et al. 2006; Trussell et al. 2005), the

authors believe that the reduced sludge filterability was a result of large rainwater inflows

to the San Francisco combined sewer system serving the treatment plant (Trussell et al.

2006) and did not result from the differences in MCRT. Trussell et al. (In Press) presents

steady state 10-d MCRT conditions where the same SMBR system was capable of

sustaining a membrane flux of 30 LMH for 87 d without requiring a chemical clean

treating the same wastewater.

The mixed liquor filterability improved during the 20-d MCRT and the membrane flux

remained at 30 LMH for the remaining test periods with no significant decline in specific

flux. The long-term (no wasting) experiments are also presented on Figure 1; rapid

membrane fouling was observed at the conclusion of each experiment. The membrane

was chemically cleaned between each operating condition.

Short-term experiments: Figures 2 and 3 show the results of the short-term, high-solids

experiments at coarse bubble aeration intensities of 1.5x10-4

, 2.3x10-4

and 3.0x10-4

m/s

for all MCRT conditions tested. In all experiments, the normalized specific flux declined

as the membrane tank MLSS concentration increased. At an MCRT of 10-d and an

aeration intensity of 1.5x10-4

m/s, the specific flux declined notably with increasing

MLSS concentrations with a 10% specific flux decline occurring at 18.2 g/L (Figure 2).

11

The specific flux declined to a lesser extent with increasing MLSS concentrations at the

20-d and 30-d MCRTs with a 10% specific flux decline occurring at an MLSS

concentration of 21.5 g/L (Figure 2).

Higher aeration rates have been previously demonstrated to allow operation at increased

membrane flux values and higher MLSS concentrations before significant specific flux

decline occurred (Bouhabila et al. 2001; Hwang et al. 2003; Le Clech et al. 2003; Liu et

al. 2003; Liu et al. 2000; Tardieu et al. 1998; Tardieu et al. 1999; Ueda et al. 1997).

However, these experiments were all performed at constant MLSS concentrations while,

varying the membrane flux and aeration intensity. Our approach was to hold to the

membrane flux and aeration intensity constant while the MLSS concentration is

increased, either by mechanical dewatering or microbial growth. This approach

emphasizes that the aeration intensity, while an important parameter, cannot be

interpreted in isolation without considering the operational changes of MLSS and its

properties ultimately affecting sludge viscosity. To assess these interactions, we

considered the changes of MLSS from the steady state (i.e. at start of each experiment)

yielding 10% decrease from the initial specific flux. Comparing the MLSS concentrations

at 10% specific flux decline, the increase in aeration intensity from the lowest value of

1.5x10-4

to 2.3x10-4

m/s resulted in a more significant increase in the allowed MLSS

concentration at the 20-d and 30-d MCRT compared to conditions of poor filterability

sludge at the 10-d MCRT. The 20-d and 30-d MCRT MLSS concentrations increased by

3.1 g/L from 21.5 to 24.6 g/L at 10% specific flux decline, while the 10-d MCRT MLSS

concentration increased by 1.6 g/L from 18.2 to 19.8 g/L at 10% specific flux decline

12

(Figure 2). However, further increasing the aeration intensity from 2.3x10-4

to 3.0x10-4

m/s resulted in similar increases in the MLSS concentration at the 10% specific flux

decline for all sludge conditions. The 20-d and 30-d MCRT MLSS concentrations

increased by 2.3 g/L from 24.6 to 26.9 g/L at 10% specific flux decline, while the 10-d

MCRT MLSS concentration increased by 2.2 g/L from 19.8 to 22.0 g/L at 10% specific

flux decline (Figures 2 and 3).

At the conclusion of all short-term experiments, the mixed liquor was returned to the

initial MLSS concentration by returning membrane permeate collected during the short-

term experiments to the biological reactor. For all experiments, the membrane

permeability was recovered within 3 h by agitating the membranes with 14.2 L/s of

coarse bubble aeration and with no permeate flow (Table 1). This indicates that the flux

decline was caused by the reversible accumulation of filtration cake at the membrane

surface. Chemical cleaning was not required, demonstrating that membrane fouling was

not caused by adsorption of fouling components or by pore plugging. At the 10-d MCRT,

the mixed liquor produced a “stickier” cake that required the full recovery period (3 h) to

restore the specific flux to its initial value (Table 1). Table 1 shows that the specific flux

was often restored to within 1-2% of its initial value immediately following the dilution

of the mixed liquor to its initial MLSS concentration at the 20-d and 30-d MCRT

conditions.

Long-Term Experiments: The long-term experiments (with no sludge wasting) at initial

MCRT conditions of 10-d, 20-d and 30-d (Figure 3) showed that regardless of differences

13

in sludge filterability, the specific flux declined by 10% at approximately 16 g/L MLSS –

a value that is significantly lower than that for the short-term experiments.

Mixed Liquor Properties: Table 2 summarizes the mixed liquor properties for the short-

term experiments where the initial (t0) and final (tF) mixed liquor samples were analyzed.

The mixed liquor properties are significantly different between the sludge with poor

filterability (e.g. 10-d MCRT) and sludges with normal filterability (e.g. 20-d and 30-d

MCRTs).

While mixed liquor colloidal material can affect the filterability of a sludge cake and

colloids have been implicated in MBR fouling, these effects were not observed in these

experiments (Itonaga et al. 2004; Ma et al. 2005; Ng and Hermanowicz 2005). The initial

colloidal material concentration was significantly higher at the 10-d MCRT than at the

20-d and 30-d MCRTs. During the short-term experiments, the colloidal material

concentration varied significantly. In general, the colloidal material concentration

increased through the short-term experiments, but on two occasions it decreased. Thus,

the observed membrane fouling could not be attributed unambiguously to colloids.

Recent work has found that soluble carbohydrate (e.g. SMPc) is the primary cause of

increased membrane fouling rates (Rosenberger et al. 2005; Rosenberger et al. 2006).

However, as Table 2 shows, the SMPc concentration remained relatively constant around

20 mg/L for the conditions tested, ranging from 12 mg/L to 27 mg/L (Table 2). The short-

term experiments did not increase the SMPc concentration significantly regardless of the

14

sludge properties or aeration intensity. In fact, one of the poor filterability sludge samples

exhibited the lowest SMPc concentration in any mixed liquor tested. However, the SMPp

concentration varied more significantly and ranged from 10 to 140 mg/L. The poor

filterability sludge (10-d MCRT) was influenced by the aeration intensity and the highest

SMPp concentration was observed at the highest aeration intensity, indicating that this

sludge was more sensitive to the additional shear force provided by the increased aeration

intensity, resulting in released SMPp from the biological floc. The average total SMP

concentration was higher for the poor filterability sludge and was lowest for the 30 d

MCRT. The average ratio of SMPp to SMPc was higher for the poor filterability sludge,

not lower, which supports the importance of the total soluble organic concentration as the

best measure of membrane fouling. This is in agreement with recent work that

demonstrated the importance of the colloidal organic content on the membrane critical

flux (Fan et al. 2006). It is important that the soluble organic content be determined by a

centrifugation method or possibly paper filtration and not by bench membrane filtration

(≤0.5µm) because this method does not correlate with membrane fouling rates (Itonaga et

al. 2004; Rosenberger et al. 2005; Trussell et al. In Press).

The total EPS concentration was lowest for the poor filterability sludge (e.g. 10-d

MCRT) and this may have contributed to the poor flocculation indicated by high

colloidal material content at this condition. Similar to the SMP, the carbohydrate fraction

(EPSc) was relatively constant for all conditions tested, ranging from 24 to 29 mg/gVSS.

The EPSp varied significantly from 57 to 88 mg/gVSS and primarily influenced changes

15

in the total EPS concentration and the ratio of EPS P/C. These data show that measuring

EPS P/C does not provide significant information about sludge filterability.

Table 2 shows that the sCOD concentrations were higher for the poor filterability sludge

and increased during the short-term experiment. However, the sCOD concentration did

not increase for all the short-term experiments as shown in Figure 4. The sCOD at the 10-

d MCRT condition increased with increasing MLSS concentration while the sCOD at the

20-d and 30-d MCRTs did not consistently increase during the short-term experiments.

The increase in sCOD for the 10-d MCRT and not the 20-d and 30-d MCRTs indicates

that organics, smaller than 0.45µm, are being retained as the poor filterability sludge was

being thickened. In contrast, these organics were readily passing through the membrane

while the 20-d and 30-d MCRT sludges were thickened.

The mixed liquor properties were measured daily in the long-term experiments (Figure

5). The mixed liquor properties did not increase in the long-term experiments with the

exception of the mixed liquor viscosity and 20-d MCRT colloidal content. However, the

30-d MCRT colloidal content did not increase in the long-term experiment and the only

mixed liquor property that consistently increased for all three experiments was the mixed

liquor viscosity. Figure 5 presents the mixed liquor viscosity for all the samples, from

both the short-term and long-term experiments, as a function of MLSS concentration, and

clearly shows that the 10-d MCRT sludge was more viscous than the measured 20-d and

30-d MCRT sludges. It is important to note that activated sludge is a non-Newtonian

fluid whose viscosity depends on the applied shear rate and the same device, speed, and

16

spindle are required to compare viscosity data with those presented here (Defrance et al.

2000; Dick and Ewing 1967).

DISCUSSION

Membrane Filtration in the SMBR: The SMBR process uses direct membrane

filtration of mixed liquor to provide solid-liquid separation and the membrane

permeability declines with increases in filtrate volume or operational time. Agitation and

cross-flow velocity provided by coarse bubble aeration produce shear forces that re-

suspend the rejected materials into the bulk solution and maintain membrane

permeability. A balance must be maintained between the flux of materials toward the

membrane surface and the re-suspending flux provided by the coarse bubble aeration

intensity.

For the experiments presented in this paper, the membrane hydraulic flux was held

constant while the concentration of MLSS increased, thus increasing the solids flux

towards the membrane. The coarse bubble aeration intensity was also constant, but as the

MLSS concentration increased, the mixed liquor viscosity increased exponentially and

changed the flow regime of the coarse bubble aeration and therefore its efficacy (Cui et

al. 2003). Figure 6 illustrates the effect of mixed liquor viscosity on the specific flux for

all of the short-term experiments where the specific flux began to decline above a

viscosity of 190 cP. For each coarse-bubble aeration intensity, the flux decline curves

collapse to a single line when plotted as a function of viscosity (Figure 6) regardless of

17

sludge properties and operating conditions. Thus, the membrane performance was

primarily affected by changes in viscosity for the high MLSS concentrations in the short-

term experiments. Based upon the data and samples collected, no other mixed liquor

property could have impacted membrane performance so dramatically (Table 2). This

finding underscores the importance of the external mass transfer in the SMBR operating

at high MLSS concentrations. For a particular reactor design, the resuspending solids flux

is an increasing function of the aeration intensity and the counteracting effects of mixed

liquor viscosity. The viscosity is in turn affected by the MLSS concentration and the

biomass composition and structure.

Effect of Time Exposure on Specific Flux Decline: The length of time that the SMBR

process was exposed to the increasing MLSS concentrations affected the degree of

membrane permeability decline. Figure 7 shows the effect of filtration time on membrane

permeability with MLSS concentrations for the 20-d and 30-d MCRTs. In the short-term

experiments, the MLSS concentration increased more rapidly (reaching 25 g/L and a 10%

specific flux decline in 30 min) than in the extended short-term experiment (reaching 22

g/L and a 10% specific flux decline in 1 h). In the long-term experiment, the MLSS

concentration increased gradually over a one-week period and reached a concentration of

16 g/L at 10% specific flux decline. The effect of filtration time on the specific flux

decline at increasing MLSS concentrations is expected since small differences in

equilibrium between MLSS transport to and from the membrane will be exaggerated over

extended periods of time.

18

CONCLUSIONS

• All experiments showed that membrane permeability decreased as MLSS

concentrations increased

• Increasing the coarse bubble aeration intensity increased the MLSS concentration

that could be obtained before a 10% decline in specific flux occurred

• Highest colloidal material, total SMP and soluble COD concentrations were

obtained at the 10-d MCRT when the mixed liquor filterability was the poorest.

The lowest total EPS concentration was obtained at these conditions

• Poor mixed liquor filterability (10-d MCRT) affected the influence of MLSS

concentration on specific flux

• Increasing the MLSS concentration in the long-term experiments resulted in

similar specific flux declines regardless of mixed liquor properties

• Virtually complete specific flux recovery could be obtained by diluting to the

initial MLSS concentration and aerating the membrane without permeate

production

• Poor mixed liquor filterability (10-d MCRT) required a longer aeration period to

restore the original specific flux, indicating the formation of a “stickier” cake

layer

• Mixed liquor viscosity, not other mixed liquor properties, controlled the specific

flux decline for these experiments

• A longer exposure time to increasing MLSS conditions resulted in a greater

specific flux decline with MLSS

19

ACKNOWLEDGEMENTS

The project was funded in part by Water Environment Research Foundation (WERF)

Project No. 01-CTS-19-UR. The authors thank Zenon Environmental Services, Inc. for

the pilot SMBR and technical support; the City and County of San Francisco SEP

Engineering, Operations, Maintenance, and Laboratory staff for support and technical

assistance; and Mr. Namjung Jang for research assistance. At the time this work was

conducted Shane Trussell and Rion Merlo were doctoral candidates at the Department of

Civil and Environmental Engineering, University of California, Berkeley.

REFERENCES

Adham, S., Gagliardo, P., Boulos, L., Oppenheimer, J., and Trussell, R. (2001).

"Feasibility of the membrane bioreactor process for water reclamation." Water

Science and Technology, 43(10), 203-209.

APHA. (1998). Standard Methods for the Examination of Water and Wastewater,

American Public Health Association, American Water Works Association, Water

Pollution Control Federation, Washington, D.C.

Bouhabila, E., Ben Aim, R., and Buisson, H. (2001). "Fouling characterisation in

membrane bioreactors." Separation and Purification Technology, 22-3(1-3), 123-

132.

20

Chiemchaisri, C., and Yamamoto, K. (1994). "Performance of membrane separation

bioreactor at various temperatures for domestic wastewater treatment." Journal of

Membrane Science, 87(1-2), 119-129.

Cui, Z. F., Chang, S., and Fane, A. G. (2003). "The use of gas bubbling to enhance

membrane processes." Journal of Membrane Science, 221, 1-35.

Defrance, L., Jaffrin, M. Y., Gupta, B., Paullier, P., and Geaugey, V. (2000).

"Contribution of various constituents of activated sludge to membrane bioreactor

fouling." Bioresource Technology, 73(2), 105-112.

Dick, R. I., and Ewing, B. B. (1967). "Rheology of activated sludge." Journal Water

Pollution Control Federation, 39(4), 543-&.

Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A., and Smith, F. (1956).

"Colorimetric method for determination of sugars and related substances."

Analytical Chemistry, 28(3), 350-356.

Fan, F. S., Zhou, H. D., and Husain, H. (2006). "Identification of wastewater sludge

characteristics to predict critical flux for membrane bioreactor processes." Water

Research, 40(2), 205-212.

Frolund, B., Palmgren, R., Keiding, K., and Nielsen, P. H. (1996). "Extraction of

extracellular polymers from activated sludge using a cation exchange resin."

Water Research, 30(8), 1749-1758.

Hwang, E. J., Sun, D. D., and Tay, J. H. (2003). "Operational factors of submerged

inorganic membrane bioreactor for organic wastewater treatment: sludge

concentration and aeration rate." Water Science and Technology, 47(1), 121-126.

21

Innocenti, L., Bolzonella, D., Pavan, P., and Cecchi, F. (2002). "Effect of sludge age on

the performance of a membrane bioreactor: influence on nutrient and metals

removal." Desalination, 146(1-3), 467-474.

Itonaga, T., Kimura, K., and Watanabe, Y. (2004). "Influence of suspension viscosity and

colloidal particles on permeability of membrane used in membrane bioreactor

(MBR)." Water Science and Technology, 50(12), 301-309.

Krampe, J., and Krauth, K. (2003). "Oxygen transfer into activated sludge with high

MLSS concentrations." Water Science and Technology, 47(11), 297-303.

Le Clech, P., Jefferson, B., and Judd, S. J. (2003). "Impact of aeration, solids

concentration and membrane characteristics on the hydraulic performance of a

membrane bioreactor." Journal of Membrane Science, 218(1-2), 117-129.

Lee, W., Kang, S., and Shin, H. (2003). "Sludge characteristics and their contribution to

microfiltration in submerged membrane bioreactors." Journal of Membrane

Science, 216(1-2), 217-227.

Liu, R., Huang, X., Sun, Y. F., and Qian, Y. (2003). "Hydrodynamic effect on sludge

accumulation over membrane surfaces in a submerged membrane bioreactor."

Process Biochemistry, 39(2), 157-163.

Liu, R., Huang, X., Wang, C. W., Chen, L. J., and Qian, Y. (2000). "Study on hydraulic

characteristics in a submerged membrane bioreactor process." Process

Biochemistry, 36(3), 249-254.

Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951). "Protein

measurement with the folin phenol reagent." Journal of Biological Chemistry,

193(1), 265-275.

22

Ma, L., Li, X. F., Du, G. C., Chen, J., and Shen, Z. S. (2005). "Influence of the filtration

modes on colloid adsorption on the membrane in submerged membrane

bioreactor." Colloids and Surfaces a-Physicochemical and Engineering Aspects,

264(1-3), 120-125.

Metcalf&Eddy. (2003). Wastewater Engineering Treatment, Disposal, and Reuse-

Fourth Edition, McGraw-Hill, New York, NY.

Muller, E. B., Stouthamer, A. H., Vanverseveld, H. W., and Eikelboom, D. H. (1995).

"Aerobic domestic wastewater treatment in a pilot-plant with complete sludge

retention by cross-flow filtration." Water Research, 29(4), 1179-1189.

Ng, H. Y., and Hermanowicz, S. W. (2005). "Specific resistance to filtration of biomass

from membrane bioreactor reactor and activated sludge: Effects of exocellular

polymeric substance and dispersed microorganisms." Water Environment

Research, 77(2), 187-192.

Ng, H. Y., Tan, T. W., and Ong, S. L. (2006). "Membrane fouling of submerged

membrane bioreactors: impact of mean cell residence time and the contributing

factors." Environmental Science & Technology, 40(8), 2706-2713.

Pritchard, M., Howell, J. A., and Field, R. W. (1995). "The ultrafiltration of viscous

fluids." Journal of Membrane Science, 102, 223-235.

Rosenberger, S., Evenblij, H., Poele, S. T., Wintgens, T., and Laabs, C. (2005). "The

importance of liquid phase analyses to understand fouling in membrane assisted

activated sludge processes - six case studies of different European research

groups." Journal of Membrane Science, 263(1-2), 113-126.

23

Rosenberger, S., Laabs, C., Lesjean, B., Gnirss, R., Amy, G., Jekel, M., and Schrotter, J.

C. (2006). "Impact of colloidal and soluble organic material on membrane

performance in membrane bioreactors for municipal wastewater treatment."

Water Research, 40(4), 710-720.

Scholzy, W., and Fuchs, W. (2000). "Treatment of oil contaminated wastewater in a

membrane bioreactor." Water Research, 34(14), 3621-3629.

Shimizu, Y., Okuno, Y., Uryu, K., Ohtsubo, S., and Watanabe, A. (1996). "Filtration

characteristics of hollow fiber microfiltration membranes used in membrane

bioreactor for domestic wastewater treatment." Water Research, 30(10), 2385-

2392.

Tardieu, E., Grasmick, A., Geaugey, V., and Manem, J. (1998). "Hydrodynamic control

of bioparticle deposition in a MBR applied to wastewater treatment." Journal of

Membrane Science, 147(1), 1-12.

Tardieu, E., Grasmick, A., Geaugey, V., and Manem, J. (1999). "Influence of

hydrodynamics on fouling velocity in a recirculated MBR for wastewater

treatment." Journal of Membrane Science, 156(1), 131-140.

Trussell, R. S., Adham, S., and Trussell, R. R. (2005). "Process limits of municipal

wastewater treatment with the submerged membrane bioreactor." Journal of

Environmental Engineering, 131(3), 410-416.

Trussell, R. S., Merlo, R. P., Cheng, S., Jolis, D., Sukapanpotharam, P., Hermanowicz, S.

W., and Jenkins, D. "Treating wet weather flows in a membrane bioreactor:

Changes in mixed liquor properties cannot be neglected." WEFTEC, Dallas, TX.

24

Trussell, R. S., Merlo, R. P., Hermanowicz, S. W., and Jenkins, D. (In Press). "The effect

of organic loading on process performance and membrane fouling in a submerged

membrane bioreactor treating municipal wastewater." Water Research.

Ueda, T., Hata, K., Kikuoka, Y., and Seino, O. (1997). "Effects of aeration on suction

pressure in a submerged membrane bioreactor." Water Research, 31(3), 489-494.

Wei, Y. S., Van Houten, R. T., Borger, A. R., Eikelboom, D. H., and Fan, Y. B. (2003).

"Minimization of excess sludge production for biological wastewater treatment."

Water Research, 37(18), 4453-4467.

Wilen, B. M., Keiding, K., and Nielsen, P. H. (2000). "Anaerobic deflocculation and

aerobic reflocculation of activated sludge." Water Research, 34(16), 3933-3942.

Yamamoto, K., Hiasa, M., Mahmood, T., and Matsuo, T. (1989). "Direct solid-liquid

separation using hollow fiber membrane in an activated sludge aeration tank."

Water Science and Technology, 21(4-5 -5 pt 1), 43-54.

Yoon, S. H. (2003). "Important operational parameters of membrane bioreactor-sludge

disintegration (MBR-SD) system for zero excess sludge production." Water

Research, 37(8), 1921-1931.

Yoon, S. H., and Lee, S. (2005). "Critical operational parameters for zero sludge

production in biological wastewater treatment processes combined with sludge

disintegration." Water Research, 39(15), 3738-3754.

25

TABLES

Table 1. Summary of Membrane Permeability Recovery for High Solids Experiments

Percent Post Experiment

Initial After Change Recovery Time, min

9.4 70.4 62.4 11% 0

9.4 same 69.7 1% 180

14.2 69.6 59.3 15% 5

14.2 same 68.6 1% 120

18.9 71.1 67 6% 2

18.9 same 70.5 1% 180

9.4 66.2 66.7 -1% 0

14.2 65.5 67 -2% 180

18.9 65.4 64.4 2% 0

9.4 79.8 75.7 5% 0

14.2 81.3 80.5 1% 0

18.9 77.2 76.4 1% 0

30

20

Coarse Aeration

Rate, L/s

Specific Flux@20oC, LMH/bar

MCRT, d

10

26

Table 2. Summary of Mixed Liquor Properties For Short-Term Experiments

1.5x10-4

2.3x10-4

3.0x10-4

1.5x10-4

2.3x10-4

3.0x10-4

1.5x10-4

2.3x10-4

3.0x10-4

t0 98 72 73 23 20 27 26 29 21

tF 86 77 - 27 32 23 46 38 48

Total SMP 29 45 105 41 30 46 33 32 35

SMPc 12 26 26 19 20 19 21 21 21

SMPp 17 19 79 22 10 27 12 11 14

SMP P/C - 1.4 0.7 3.0 1.2 0.5 1.4 0.6 0.5 0.7

Total SMP 45 53 166 45 35 48 33 35 39

SMPc 16 27 26 20 22 19 21 22 20

SMPp 29 26 140 25 13 29 12 13 19

SMP P/C - 1.8 1.0 5.4 1.3 0.6 1.5 0.6 0.6 1.0

Total EPS 81 87 102 115 104 114 103 103 105

EPSc 24 24 27 26 27 27 29 28 29

EPSp 57 62 74 88 77 87 74 75 76

EPS P/C - 2.4 2.6 2.7 3.4 2.8 3.2 2.6 2.6 2.6

t0 73 82 75 47 50 50 37 74 63

tF 101 115 113 70 62 57 71 59 92

Units

10-d MCRT

NTU

Aeration Intensity, m/s

20-d MCRT 30-d MCRT

Sample*

Mixed

Liquor

Property

mg/g VSS

sCOD mg/L

mg/L

mg/L

t0

Colloidal

Material

tF

t0

*t0 – mixed liquor sample collected at start of short-term thickening experiment and tF – final mixed liquor sample collected

27

FIGURES

Figure 1. Daily MLSS concentrations and membrane performance for 10-d, 20-d and 30-

d MCRT high solids experiments

28

Figure 2. Effect of MLSS on specific flux in short-term experiments at aeration

intensities of 1.5x10-4

and 2.3x10-4

m/s

29

Figure 3. Effect of MLSS on specific flux in short-term experiments at an aeration

intensity of 3.0x10-4

m/s and long-term experiments at an aeration intensity of 2.3x10-4

m/s

*Sludge wasting was stopped after achieving steady state at the specified condition

30

Figure 4. Changes in mixed liquor sCOD with MLSS for short-term experiments

31

Figure 5. Changes in all mixed liquor properties for the duration of the long-term experiments

32

y = 6.357e0.1726x

R2 = 0.9393

y = 5.3776e0.2142x

R2 = 0.9348

y = 7.4013e0.1656x

R2 = 0.9885

0

100

200

300

400

500

600

700

10 12 14 16 18 20 22 24 26 28

MLSS, g/L

Vis

cosi

ty,

cP

10-d MCRT 20-d MCRT 30-d MCRT

LV Spindle #2, 100 rpm

10-d MCRT:

20-d MCRT:

30-d MCRT:

Figure 6. Effects of MLSS on sludge viscosity for LV spindle #2 at 100 rpm

33

Figure 7. Effect of sludge viscosity and aeration intensity on specific flux

34

Figure 8. Effect of exposure time to high MLSS on specific flux