Luận văn thạc sĩ khoa học OILY WASTEWATER TREATMENT BY MEMBRANE BIOREACTOR PROCESS COUPLED WITH BIOLOGICAL ACTIVATED CARBON PROCESS

OILY WASTEWATER TREATMENT BY MEMBRANE BIOREACTOR PROCESS COUPLED WITH BIOLOGICAL ACTIVATED CARBON PROCESS Title page by Phan Thanh Tri A thesis submitted in partial fulfillment of the re

OILY WASTEWATER TREATMENT BY MEMBRANE BIOREACTOR PROCESS COUPLED WITH BIOLOGICAL ACTIVATED CARBON PROCESS

Title page

by

Phan Thanh Tri

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering

Examination Committee: Prof C Visvanathan (Chairman)

Dr Josef Trankler

Dr Preeda Parkpian

Nationality: Vietnamese

Previous Degree: Bachelor of Chemical Engineering

Ho Chi Minh City University of Technology

Ho Chi Minh City, Vietnam Scholarship Donor: Japan - Asian Development Bank

Asian Institute of Technology School of Environment, Resources and Development

Thailand August 2002

Acknowledgement

The author wishes to express his deepest respect and profound gratitude to his advisor, Prof C Visvanathan, for the valuable guidance, advice and encouragement throughout this study

Sincere thanks are extended to Dr J Trankler and Dr P Parkpian who served as members of the thesis examination committee Their suggestive and stimulating comments are highly treasured

The author gratefully acknowledges the Asian Development Bank and the Government of Japan for the scholarship grant which allowed the author to pursue his study at AIT

Grateful thanks are given to the staffs of the Environmental Engineering Laboratory, AIT for the precious assistances during this study

The author would like to acknowledge the Research and Development Center for Petroleum Safety and Environment, PetroVietnam, for providing favorable condition for the author to pursue his study at AIT

The author is very grateful to Dr N.P Dan and other friends at AIT for their helps and encouragement Special appreciations are given to Mrs T.T.H Nhien, who devotedly stimulated and helped the author to pursue graduate study

The author is most grateful to his darling wife, who encouraged and shared with the author the difficulty times during the study Finally, the author wishes to dedicate this achievement to his respected and beloved mother, who has sacrificed her life for the growing-up of her children

Abstract

Preliminary study of biodegradation of the oily wastewater was evaluated by activated sludge (AS) process and biological activated carbon (BAC) process operated in sequencing batch reactor (SBR) mode In BAC process, powdered activated carbon (PAC) was added into the reactors at different doses of 100 – 500 mg/L The feed wastewater had COD of 550 mg/L and Oil & Grease (O&G) concentration of 150 mg/L The reactors were experimented at hydraulic retention time (HRT) of 3, 6, 10, 13 and 16 h The results showed that the removal efficiency of AS process reached 80% for COD and 90% for O&G at HRT=13h There was a tendency of effluent quality improvement of the BAC process at PAC dose = 500 mg/L and HRT = 3h in comparison with the AS process

Lab scale submerged membrane bioreactor (MBR) system with microfiltration membrane module was used to investigate the treatment of oily wastewater The MBR system provided excellent effluent quality with COD in range of 11.2 to 85.9 mg/L and O&G in range of 0.2 – 7.3 mg/L when influent COD varied from 495 – 1835 mg/L and O&G varied from 150 – 600 mg/L The removal efficiency was 89.9 – 99.9 mg/L for COD and 97.6 – 99.9 mg/L for O&G The MBR system provided stable effluent quality against shock loading The system could stand for the COD loading of 10.5 g COD/L.day and oil loading of 3.5 g oil/L.day without deteriorating the effluent quality

Combination of MBR process and BAC process to treat oily wastewater was investigated by adding PAC to the reactor at the dose of 2 g/L The hybrid system showed little improvement in effluent quality Membrane clogging of the hybrid system occurred more quickly in comparison with the same system without PAC addition about 5 days The reason would be due to the plugging of PAC particles in the pores of the membrane

The membrane fouling in the MBR system was found reversible while that of the hybrid system with PAC addition was partially irreversible The resistance of the cake layer was found the highest component contributing to the total resistance of the clogged membrane It is suggested that during operation there was a critical point when the blockage of the membrane pores reached a level that made the air backwash ineffective, the formation of the cake layer became intensive leading to the membrane clogging

1.2 Objectives of the study 2

1.3 Scopes of the study 2

2.1.1 Sources of car wash wastewater 3 2.1.2 Characteristics of oil/water emulsion 3 2.1.3 Pollution of oil/water emulsion 5 2.1.4 Oily wastewater treatment system 5 2.2 Application of MBR process in wastewater treatment 7

2.2.2 Advantages of MBR process 8 2.2.3 Membrane fouling 10 2.2.4 Application of MBR process in oily wastewater treatment 11 2.3 Biological activated carbon process 12

2.3.1 Activated carbon adsorption 12 2.3.2 Biological activated carbon process 13 2.3.3 MBR process coupled with BAC process 14

3.1 Materials and microorganisms 16

3.1.1 Feed wastewater 16 3.1.2 Powdered activated carbon 17 3.1.3 Microorganisms 17 3.2 Overall experimental course 17

3.3 Adsorption isotherm experiments 19

3.4 Batch experiments 19

3.5 Membrane bioreactor experiments 20

3.5.1 Membrane bioreactor set-up 20 3.5.2 Experimental conditions 20 3.5.3 Measured parameters in the MBR experiments 23 3.5.4 Membrane cleaning 23 3.5.5 Membrane resistance measurement 24 3.6 Analytical methods 24

4 Results and Disscussions 26

4.1 Adsorption isotherm of the PAC 26

4.2 Batch experiments 27

4.3.1 Membrane resistance measurement 31 4.3.2 Optimum HRT run 32 4.3.3 Shock loading run 35 4.3.4 Long run A and B 37 4.4 Conceptual idea for application of MBR system in

wastewater treatment and reuse at gas station 41

5 Conclusions and Recommendations 43

References 46 Appendix A: Experimental results 49

Appendix B: Membrane resistance measurement data 56

Appendix C: Photographs of experimental works 68

List of Figures

2.1 Process flow diagram of car wash process (Awas, 1997) 4

2.2 Oil/water emulsion conventional treatment systems 6

2.3 Different configurations of MBR process 9

2.5 Schematic diagram of dynamic porous PAC layer in cross flow

membrane filtration (Pirbazari et al., 1996) 11

3.1 Relative contributions of COD and BODfrom the oil, glucose and

emulsifier of a typical feed wastewater sample having oil

concentration of 150 mg/L 17

3.2 Overall experimental course 18

3.3 Membrane bioreactor setup 21

3.4 Laboratory scale MBR system to treat oily wastewater 22

4.1 Variation of equilibrium COD vs different contact times (PAC dose =

100 mg/L; T = 25oC; agitation speed = 100 rpm) 26

4.2 Experimental data plotted according to Freundlich isotherm 27

4.3 Experimental data plotted according to Langmuir isotherm 27

4.4 Variation of COD removal efficiency vs HRT 29

4.5 Variation of O&G removal efficiency vs HRT 29

4.6 Comparison of COD removal efficiency between AS process and

BAC process (PAC = 500 mg/L) 30

4.7 Comparison of COD removal efficiency between AS process and

BAC process (PAC = 500 mg/L) 30

4.8 Resistance of the membrane before used and after chemical cleaning 32

4.9 Variation of COD of the influent and effluent during Optimum HRT

run 33

4.10 Variation of O&G in the reactor and in the effluent during Optimum

4.11 O&G mass balance of Optimum HRT run 34

4.12 Variation of transmembrane pressure and MLSS during Optimum

4.13 Variation of COD of the influent and effluent of Shock loading run 36

4.14 Variation of O&G in the reactor and in the effluent during Shock

4.15 Variation of COD of the influent and effluent of Long run A and B 37

4.16 Variation of O&G in the reactor and in the effluent during Long run

4.17 Biofilm development on carbon surface after 14 days of Long run B 39

4.18 BAC sludge after 14 days of Long run B (x580) 39

4.19 Variation of transmembrane pressure and MLSS during Long run A

4.20 Conceptual diagram for a wastewater treatment and reuse system

using MBR process at gas station 42

List of Tables

2.1 Characteristics of wastewater from various gas stations in Bangkok,

Thailand (Limpananon and Yodsang, 1997) 4

2.2 Composition of lubricant oil 5

2.3 Comparison of operation parameters between industrial-scale applied

ultrafiltration system treating oil-contaminated wastewater from

machine factoring and the membrane bioreactor system (Scholz and

3.1 Composition of the feed wastewater 16

3.2 The major properties of the PAC 17

3.3 Experimental conditions for equilibrium adsorption time test 19

3.4 Characteristics of the membrane module 20

3.5 Operational condition of the MBR experiments 23

4.1 Average results of the batch experiments 28

4.2 Membrane resistances of the new membrane and the membrane after

4.3 Membrane resistances after each step of chemical cleaning course

4.4 Oil loading of Optimum HRT run 34

4.5 Oil loading of Shock loading run 37

4.6 Comparison of EPS content in the sludge of Long Run A and Long

4.7 Various components of membrane resistance contributing to

A.1 Results of equilibrium adsorption time test 50

A.2 Results and calculation of adsorption isotherm test 50

A.3 Results of the batch experiments 51

A.4 Experimental results of Optimum HRT run 52

A.5 Experimental results of Shock loading run 53

A.6 Experimental results of Long run A 54

A.7 Experimental results of Long run B 55

List of Abbreviations

AC Activated Carbon

API American Petroleum Institute

AS Activated Sludge

BAC Biological Activated Carbon

BOD Biochemical Oxygen Demand [mg/L]

COD Chemical Oxygen Demand [mg/L]

CPI Corrugated Plate Interceptors

DAF Dissolved Air Floatation

Eff Effluent

EPS Extracellular Polymeric Substance

GAC Granular Activated Carbon

HRT Hydraulic Retention Time [h]

Inf Influent

MBR Membrane Bioreactor

MLSS Mixed Liquor Suspended Solid [mg/L]

O&G Oil and Grease concentration [mg/L]

PAC Powdered Activated Carbon

PAH Polyaromatic Hydrocarbon

PPI Parallel Plate Interceptors

Rc Cake layer resistance [m-1]

Rf Fouling resistance [m-1]

Rm Intrinsic membrane resistance [m-1]

Rt Total membrane resistance [m-1]

SBR Sequencing Batch Reactor

SRT Solid Retention Time [day]

T Temperature [oC]

TMP Transmembrane Pressure [mmHg, kPa]

1.1 General

Oil contaminated wastewater has been recognized as one of the most concerned pollution sources This kind of wastewater comes from variety of sources such as crude oil production, oil refinery, petrochemical industry, metal processing, compressor condensates, lubricant and cooling agents, car washing The oily wastewater is considered

as hazardous industrial wastewater because it contains toxic substances such as phenols, petroleum hydrocarbons, polyaromatic hydrocarbons which are inhibitory to plant and animal growth and also are mutagenic and carcinogenic to human being

Physical treatment of oily wastewater such as API gravity separator, dissolved air floatation (DAF), ultrafiltration etc does not remove the pollutants completely but just transfer them to a more concentrated waste (Schols and Fuchs, 2000) Moreover, the physical treatment cannot remove soluble fraction of the wastewater (Seo, 1997) Currently conventional biological treatment techniques are mostly incapable of complete elimination

of hydrocarbon in stable emulsion Removal efficiency of the conventional biological processes is also low due to inhibitive effects of toxic substances and hydrophobic characteristics of oil components (Schols and Fuchs, 2000) Therefore, there is a need to develop a more efficient treatment technique based on biological process to treat oily contaminated wastewater

Membrane bioreactor (MBR) process is a novel technology in wastewater treatment

in recent years It is a modification of conventional activated sludge process in which solid/liquid separation is accomplished by membrane filtration instead of using a secondary sedimentation tank One of the advantages of MBR process is that it can be operated at very high sludge retention time (SRT) in comparison with conventional activated sludge process This will create favorable conditions for the growth of slow-growing microorganisms which can degrade recalcitrant and toxic compounds such as petroleum hydrocarbons (Fuchs and Braun, 2001) Other advantages include stability against shock loading, low rate sludge production, compact size and high effluent quality which is attractive for water reuse

However, membrane fouling which leads to high-energy consumption and high cleaning chemical requirement has limited the application of the MBR process due to high operation cost Researches have been conducted to find out various causes relating to the membrane fouling phenomenon The presence of extracellular polymeric substances (EPSs) and the characteristics of the polarization cake layer such as porosity and particle size are among the factors affecting the membrane clogging process

Recent efforts have been made to modify the MBR process to make it widely applicable in wastewater treatment and reclamation in term of high removal efficiency and reduction of membrane fouling Among the modifications is the addition of filter aid into the mixed liquor to control the membrane fouling by increasing the porosity of the cake

layer Some researchers have proposed the addition of iron oxide particles (Chang et al., 1998) and powdered activated carbon (PAC) (Pirbazari et al., 1996; Kim et al., 1998)

Among the filter aids, PAC has gained special interests because it also develops biological

Chapter 1 Introduction

activated carbon (BAC) process which is expected enhancing biodegradation and producing less EPSs (Kim et al., 1998)

The concept of biological activated carbon (BAC) process has been applied in water treatment as well as wastewater treatment In the BAC process, activated carbon either in powdered form (PAC) or in granular form (GAC) is added into the mixed liquor The removal of contaminants is achieved by a combination of biodegradation and adsorption The use of PAC in MBR process has been shown to enhance the biodegradation and also maintain a high permeate flux for a long duration (Pirbazari et al., 1996)

This study focuses on the development of a MBR system to treat the oily wastewater from gas station effectively to meet the effluent standards for car wash wastewater and to meet the requirements for water reuse Effect of PAC addition into the reactor to the performance of the system is also investigated

1.2 Objectives of the study

1 to investigate the oil adsorption isotherm of powdered activated carbon

2 to investigate the biodegradation of oily wastewater in the activated sludge process and biological activated carbon (BAC) process

3 to evaluate the performance of a MBR system to treat oily wastewater in term of membrane fouling, removal efficiency and process stability

4 to investigate the effect of PAC addition on the performance of the MBR system in term of membrane fouling and removal efficiency

1.3 Scopes of the study

1 Synthetic wastewater containing lubricant oil, surfactant and nutrient salts, representing for oily carwash wastewater, was used in this study as feed wastewater

2 Mixed bacterial culture acclimated to oily wastewater was used in the biological process

3 Batch experiments were used to determine the adsorption isotherm and biodegradation of the feed wastewater in the AS and BAC processes

4 Laboratory scale submerged MBR with hollow fiber membrane module was used

to evaluate the performance of MBR system to treat oily wastewater

5 Effluent quality and removal efficiency were mainly addressed in term of COD and O&G concentration

2.1.1 Sources of car wash wastewater

Typical car washing operation consists of four consecutive steps as shown in Figure 2.1 Each stage in the process is simply described as follows (Panpanit, 2001):

ƒ Stage (1) Dust cleaning: high-pressure water is injected on the car body and engine to wash away dirty particles The wastewater in this stage contains a high concentration

of sludge, clay and free oil In addition, the mechanical force from the high pressure of the water breaks the oil into oil/water emulsion

ƒ Stage (2) Foaming: The emulsifier solution is sprayed on the car body The wastewater

in this stage consists mainly of emulsifier at high concentration Later the wastewater from this stage is combined with oily wastewater from stage 1 and then the stabilization of oil/water emulsion takes place

ƒ Stage (3) Presoak: This stage is to remove the emulsifier from the car body by spraying with fresh water The wastewater generated in this stage mostly contains low emulsifier concentration because it is diluted by fresh water

ƒ Stage (4) Spot free rinse and dryer: The objective of this stage is to polish the car color High quality water is sprayed on the car body, which is later dried by hot air The fresh water contains high impurity such as hardness and organic are avoided in this stage because it can create an unpolished appearance problems such as spot on the car’s surface color

Characteristics of wastewater from various gas stations in Bangkok, Thailand are presented in Table 2.1

2.1.2 Characteristics of oil/water emulsion

Oil/water emulsion is oil dispersed in water phase The emulsion is stabilized at the presence of emulsifier Generally, an emulsifier consists of a molecule with hydrophilic and hydrophobic ends The emulsifier plays the role of a bridging agent which helps to lower the interfacial tension of the oil, resulting in very small oil droplet size (Althers,

Chapter 2 Literature Review

1998) The oil droplet size plays an important role in the stability of the oil/water emulsion

The oil/water emulsion formed at the presence of emulsifier may have as small oil droplet

as 5 µm, resulting in rising velocity of these oil droplets is negligible to the Brownian

movement Therefore, the emulsion cannot be removed by gravity force

Figure 2.1 Process flow diagram of car wash process (Awas, 1997)

Table 2.1 Characteristics of wastewater from various gas stations in Bangkok, Thailand

(Limpananon and Yodsang, 1997)

Gas station O&G

(mg/L)

BOD (mg/L)

Suspended solid (mg/L) pH Caltex, Bangkaen 31.6 42 113 7.60

The formation of oil/water emulsion increases the bioavailability of oil components

to biodegradation due to the increased solubility and the greater oil-water interface

Volkering et al (1995) reported the effects of dissolution process to the increased

biodegradation rates of naphthalene and phenanthrene when surfactants were added Some

researchers have also confirmed evidences that the greater the oil-water interface, the faster

the rate of decomposition by bacteria (Reisfeld et al., 1972; Oberbremer et al., 1990;

Rosenberg, 1993)

Dryer (high quality)

Dirty

car

Dust cleaning Foaming Presoak

Rinse water

Clean car

Fresh water

Emulsifier Fresh water

Fresh water (high quality)

Oily wastewater Treated wastewater Free oil

Sludge API separator

Dryer (high quality)

Dirty

car

Dust cleaning Foaming Presoak

Rinse water

Clean car

Fresh water

Emulsifier Fresh water

Fresh water (high quality)

Oily wastewater Treated wastewater Free oil

Sludge API separator

2.1.3 Pollution of oil/water emulsion

Lubricant oil and emulsifier are the main polluting components in carwash

wastewater Physical properties of lubricant oil are very low volatility and less water

solubility The lubricant oil is not volatile under normal environmental conditions Table

2.2 shows the composition of a type of lubricant oil (Scholz and Fuchs, 2001) The

lubricant oil contains mainly aliphatic compounds such as paraffin and cycloparaffins,

aromatics and some amounts of olefins approximately C25 – C40 (Cole, 1994; James and

Speight, 1991) The composition of the lubricant oil makes it highly stable to chemical and

biological oxidation The lubricant oil also contains polyaromatic hydrocarbons (PAHs)

such as naphthalene, phenanthrene, anthracene etc which are toxic substances and some

are listed as priority pollutants by US EPA Due to their lipophilic behavior and low water

solubility, PAHs are often dissolved in the dispersed phase of oil/water emulsion or in

hydrophobic organic materials (Kornmuller and Wiesmann, 1999) These compounds are

thought to contribute to the toxic and inhibitive effects in biological processes for oily

Emulsifier, which sometimes is called surfactant, is one of the major constituents in

carwash detergent Emulsifier is also found in wastewater from oil industry, rising baths

from metal processing, compressor condensates (Scholz and Fuchs, 2000) Environmental

problems associated with emulsifier discharge are objectionable foaming and low

biodegradability of some types of emulsifiers Their biodegradability depends greatly upon

chemical structure (Sawyer et al., 1994) There are two main types of emulsifiers used in

industries for cleaning purposes: anionic and nonionic emulsifiers About 90% of the

industrial wastewater streams contains nonionic emulsifiers (Althers, 1998) The nonionic

surfactants formed from polymers of ethylene oxide appear susceptible to biological

degradation (Sawyer et al., 1994)

2.1.4 Oily wastewater treatment system

Treatment methods for oily wastewater can be classified into three categories

namely, primary or gravity treatment units, secondary processes and tertiary processes

(Figure 2.2) The primary treatment of oily wastewater is usually conducted by the API

separator which treats the free oil effectively by gravity force Modifications of the API

separator include the use of parallel plate interceptors (PPI) or corrugated plate interceptors

(CPI) These modifications are to improve the removal efficiency and reduce the space

requirement of API separator However, the API separator and its modifications have no

effect on the soluble fraction of the oil and on the oil/water emulsion Because the densities

of the oil/water emulsion and water are approximate, the oil/water emulsion cannot be

removed by simple gravity separation Therefore, a variety of secondary and tertiary

treatment methods have been applied to treat the soluble fraction and the oil/water

emulsion of the oily wastewater

Oily

wastewater

API Separator

Corrugate Plate Interceptors

Parallel Plate Interceptors

Dissolved Air Flotation

Membrane Filtration

Coalescence

High Rate Filtration

Adsorption

Ozonation

Treated wastewater

Primary Secondary Tertiary

Oily

wastewater

API Separator

Corrugate Plate Interceptors

Parallel Plate Interceptors

Dissolved Air Flotation

Membrane Filtration

Coalescence

High Rate Filtration

Adsorption

Ozonation

Treated wastewater

Primary Secondary Tertiary

Figure 2.2 Oil/water emulsion conventional treatment systems

Among the physical-chemical processes, dissolved air flotation (DAF) is the one of the most common processes used to treat the oily wastewater after gravity separation The DAF process includes two different stages: (a) the chemical stage, which consists of emulsion breaking and creation of solid aggregates (coagulation and flocculation) by chemical addition, (b) the physical stage, which separates the aggregates from the liquid phase using air bubbles released under pressure This process has been reported to achieve oil removal of 80% to 90% (ASTM, 1969; Donal, 1987 and Meyers, 1984) Galil and Wolf (2001) also reported the removal efficiency of dissolved organic mater at 40% of the DAF process

Coalescence, filtration and adsorption are the other alternatives for oily wastewater treatment Coalescence method, which create the microdroplet coalesce to form large droplet size to be easily separated by decantation (Liem and Woods, 1974), is normally suitable for oil emulsion with low concentration of emulsion Filtration process can be carried out by various types of filter media ranging from coarse granular to fine filtration such as sand, anthrasilt and diatomaceous earth It has been reported that the influent oil concentrations of 130 mg/L to 500 mg/L can be reduced to the range of 12 mg/L to 77

mg/L by filtration system (Panpanit, 2001) Campos et al (2002) used microfiltration as

pretreatment of oilfield wastewater before biological process The removals of 35%, 25%, 92% and 35% were reported for COD, TOC, O&G and phenols respectively Adsorption process is normally used as tertiary treatment to remove trace amount of oil and emulsifier Among the adsorbents used are silica base material, organic polymer, bentonite, activated carbon etc

Most of the physical-chemical treatment methods for oily wastewater require intensive washing and replacement of filtration or adsorption materials that lead to difficulty in operation and increase in treatment cost Moreover, the common disadvantage

of all physical-chemical treatment methods is that they do not remove the oil completely but just transfer them to a more concentrated waste Therefore, waste disposal is always problematic for physical-chemical treatment of oily wastewater because the oily waste is considered as hazardous waste

Biological processes have also been applied in oily wastewater treatment These processes are mainly to remove soluble organic matters remaining after physico-chemical treatment Biological treatment is only effective for highly diluted oily wastewater since oil components are adsorbed on microorganism faster than they can be metabolized Subsequently the adsorbed oil tends to damage sludge settling characteristics and cause system failure (Rhee et al., 1987) Zilverentant (1997) studied the treatment of oily wastewater from road and rail car cleaning using SBR process indicating that COD removal in range of 80 – 90 % and oil removal of less than 100 mg/L can be achieved

2.2 Application of MBR process in wastewater treatment

2.2.1 MBR process

Membrane bioreactor (MBR) process is a novel technology in wastewater treatment

in recent years Investigation and innovations to develop MBR process in wastewater treatment have been conducted intensively during the last decade To meet the requirements of water reuse and stricter effluent standards in the future, application of MBR process has become an attractive alternative in comparison with other conventional

treatment techniques due to its wide range of applicability and performance characteristics (Visvanathan et al., 2000) There are 500 commercial MBRs in operation worldwide, with

many more proposed or currently under construction (Stephenson et al., 2000)

Generally, MBR process is a modification of conventional activated sludge process

in which solid/liquid separation is accomplished by membrane filtration unit instead of using a secondary sedimentation tank MBR systems can be categorized as two different configurations by the allocation of membrane modules (Figure 2.3)

The first configuration is the externally pressurized cross-flow MBR in which the membrane module is installed separately from the aeration tank The mixed liquor is pumped to the membrane module where it is filtrated by cross-flow filtration through the membrane The excess flow is circulated to the aeration tank

The second configuration is the submerged MBR in which membrane module is submerged inside the bioreactor and the permeate is suctioned directly by dead-end filtration The submerged MBR is superior to an externally pressurized cross-flow MBR in regard to power consumption and the simplicity of the installation due to the absence of

recirculation pump (Yamamoto et al., 1989)

2.2.2 Advantages of MBR process

MBR process has been proved to have many advantages in comparison with conventional biological processes The main advantages are high quality of treated water, small size of treatment unit, less sludge production and flexibility of operation

(Visvanathan et al., 2000)

ƒ Treated water quality: In conventional activated sludge process, effluent quality

strongly depends on the settling of sludge in sedimentation tank In MBR process, solid/liquid separation is conducted by membrane filtration Therefore, the final effluent does not contain suspended matter, this enables the direct discharge of the final effluent into the surface water and the reuse of the effluent for cooling, toilet flushing, lawn watering, or with further polishing, as process water

ƒ Flexibility in operation: Solid retention time (SRT) can be controlled completely

independent from hydraulic retention time (HRT) Therefore, the system can be run at very long SRT providing favorable conditions for the growth of slow-growing microorganisms which are able to degrade biorefractory compounds

ƒ Compact plant size: Because the MBR process is independent upon sludge settling

quality, high biomass concentration can be maintained up to 30 g/L in the system (Yamamoto et al., 1989) Therefore, the system can stand for high volumetric loading rate resulting in the reduced size of the bioreactor In addition, secondary settling tank, sludge thickener or post treatment for further BOD and SS removal are not necessary

in MBR process, thus the plant becomes more compact

ƒ Low rate sludge production: Studies on MBR show that the sludge production rate is

very low Excess sludge from MBR process is much lower than conventional activated sludge process Low F/M ratio and longer sludge age in the reactor may be the reason for this low sludge production rate

Membrane module Air

Air diffuser

Effluent

Circulation pump

Influent

Bioreactor

Membrane module Air

Air diffuser

Effluent

Circulation pump

Influent

Bioreactor

(a) externally pressurized cross flow MBR

Membrane module

2.2.3 Membrane fouling

Despite the advantages of the MBR process, the present cost of treatment by MBR

is unfortunately higher than that of the conventional treatment One of the major reasons for this higher cost of treatment is attributable to membrane fouling Membrane fouling causes the increase in filtration resistance resulting in the decline in permeate flux

Membrane fouling can result from the formation of a polarization cake layer and the plugging of membrane pores (Figure 2.4)

Membrane Cake layer

Figure 2.4 Membrane fouling

Effects of membrane fouling on the decline of permeate flux can be explained using the resistance-in-series model According to this model, the relationship between permeate flux and transmembrane pressure (TMP) is described in by Equation 2.1

∆P : transmembrane pressure (Pa)

µ : viscosity of the permeate (Pa.s)

Rt : total resistance for filtration (1/m)

Rt = Rm + Rc + Rf Eq 2.2

Rm : intrinsic membrane resistance

Rc: cake layer resistance

Rf: fouling resistance due to irreversible and pore plugging Intensive researches have been conducted to understand the mechanisms and causes

of membrane fouling Current trends of controlling membrane fouling are focused on (1) controlling the production of extracelular polymeric substances (EPSs) in the bioreactor; (2) reducing the cake layer resistance

Extracelular polymeric substances (EPSs) are the substances excreted by microorganism These compounds comprises of protein, polysaccharide, nucleic acid and lipid EPSs in microbial flocs have been reported as a major foulant in the membrane

bioreactor system (Chang et al., 1996; Nagaoka et al., 1996) as they occupy the pores of

the membrane Among different approaches to control of EPS production have been investigated is control of nutrient composition in the reactor and development attached

growth for MBR system (Kim et al., 1998)

Characteristics of the cake layer play an important role in membrane fouling Effects of cake layer characteristics can be described by the Carman-Kozeny equation as

follows (Liew et al., 1995):

of the cake layer by the addition of filter aids such as metal-based coagulants (Chang and

Benjamin, 1996), PAC (Pirbazari et al., 1996) into the reactor These filter aids are

expected to form a dynamic cake layer on the membrane surface The permeability of the dynamic cake layer is thought to be higher due to larger particle size and porosity Schematic diagram of the dynamic cake layer is illustrated in Figure 2.5 The porous layer also plays as a filter layer to retain soluble organic compounds preventing them to contact and plug in the membrane pores Dan (2002) reported a different approach as developing a yeast culture for MBR system Due to larger size of yeast cells in comparison with bacteria cells, the yeast cells play as a porous layer and therefore permeate flux can be enhanced

Biofilm

Carbon particle

Tangential Flow

Membrane Permeate Flux

Soluble organic compound

Dynamic porous cake layer Biofilm

Carbon particle

Tangential Flow

Membrane Permeate Flux

Soluble organic compound

Dynamic porous cake layer

Figure 2.5 Schematic diagram of dynamic porous PAC layer in cross flow membrane

filtration (Pirbazari et al., 1996)

2.2.4 Application of MBR process in oily wastewater treatment

Some researchers have investigated the feasibility of using MBR process to treat

oily wastewater (Scholz and Fuchs, 2000; Seo et al., 1997; Fuchs and Braun, 2001)

Removal of fuel oil, lubricant oil, surfactant, residual oil-originated organic matters as well

as polyaromatic compounds by the MBR process was studied

Scholz and Fuchs (2000) investigated the application of MBR process to treat oily

wastewater External membrane module configuration was used Synthetic wastewater

containing either fuel oil or lubricant oil and a surfactant was used in the study Influent

concentration was in range of 500 – 1000 mg/L in term of hydrocarbon Biomass

concentration in the system was maintained up to 48 g/L Removal efficiency of 99.9% for

fuel oil as well as lubricant oil could be achieved at the hydraulic retention time of 13.3 h

The maximum biodegradation of fuel oil could be observed at 0.82 g hydrocarbons/(g

MLVSS.day) with the average values ranging from 0.26 – 0.54 TOC and COD removal

efficiencies were also reported at 94 – 96 % for fuel oil and 97 – 98 % for lubricant oil

The removal efficiency of surfactant was in range of 92.9 – 99.3 % The permeate quality

was as high as the effluent can be reused in industrial process The results when compared

with those from plain membrane filtration showed clearly that the implementation of a

biological stage enhanced permeate quality due to degradation of pollutants (Table 2.3)

However, membrane fouling and decline of permeate flux over operating time were not

reported

Table 2.3 Comparison of operation parameters between industrial-scale applied

ultrafiltration system treating oil-contaminated wastewater from machine

factoring and the membrane bioreactor system (Scholz and Fuchs, 2000)

Parameter Membrane bioreactor Membrane application COD (mg/L) 129 – 131 373

Oil (mg/L) 0.036 - 0.35 1

COD removal efficiency (%) 97 85.6

Oil removal efficiency (%) 99.9 99.2

Seo et al (1997) investigated the treatment of residual organic matters from oily

wastewater from an automobile engine manufacturing plant by MBR process The system

was run at HRT of 3.67 day, TMP of 1.0 kg/cm2, SRT of 300 day Removal efficiency of

COD was greater than 95% Removal of oil substances was 76.1% resulting in average

effluent concentration of 4.4 mg/L as n-Hexane extract MLSS concentration increased

gradually to more than 2000 mg/L that was twice the initial concentration Membrane was

cleaned when the flux decreased to about 2 L/m2/hr after 7 days

Fuchs and Braun (2001) studied the degradation of phenanthrene in a membrane

bioreactor Synthetic wastewater containing phenanthrene in suspension form as sole

carbon source was fed at concentration up to 1900 ppm Mixed culture of bacteria was

adapted with phenanthrene before the continuous run of 160 days The effluent was

completely free of phenanthrene and the system was proved very stable against shock

loading However, about 5 % of phenanthrene was not mineralized completely, indicated

by TOC values of the effluent ranging from 40 – 70 ppm The metabolites from the

biodegradation of phenanthrene were not biodegraded further under biodegradability test

No membrane fouling phenomenon was reported Sorption of phenanthrene onto biomass

and membrane were not investigated

2.3 Biological activated carbon process

2.3.1 Activated carbon adsorption

Activated carbon has been used for a long time in water and wastewater treatment

The surface properties of activated carbon depend upon both the initial material used to produce the carbon and the exact production procedure (Metcalf and Eddy, 1991) Generally, granular carbon from bituminous coal has small pore size, large surface area and the highest bulk density Lignite carbon has the largest pore size, least surface area and the lowest bulk density (Eckenfelder, 2000)

The ability of activated carbon to adsorb pollutants is a function of both the characteristics and concentration of the pollutants and the temperature Generally, the amount of material adsorbed is determined as a function of concentration at a constant temperature, and the resulting function is called an adsorption isotherm The most common adsorption isotherm equations used in water and wastewater treatment are Freundlich isotherm and Langmuir isotherm

Freundlich isotherm:

x

⎯ = Kf Ce1/n Eq 2.4

m Where: x/m : amount adsorbate adsorbed per unit of carbon

Ce : equilibrium concentration of adsorbate in solution after adsorption

Kf, n : empirical constants Langmuir isotherm:

x ab Ce

⎯ = ⎯⎯⎯⎯ Eq 2.5

m 1 + bCe

Where: x/m : amount adsorbate adsorbed per unit of carbon

Ce : equilibrium concentration of adsorbate in solution after adsorption

a, b : empirical constants Activated carbon can be classified into two types: powdered activated carbon (PAC) which has a diameter of less than 200 mesh and granular activated carbon (GAC) which has a diameter greater than 0.1 mm GAC is usually used in fixed bed column while PAC is added to the effluent from secondary biological treatment to serve as a polishing process Another application of PAC is direct addition of PAC to aeration basin of the activated sludge process

2.3.2 Biological activated carbon process

Rice and Robinson (1982) used the term Biological Activated Carbon (BAC) for the water and wastewater treatment system in which aerobic microbial activity is deliberately promoted in an activated carbon adsorber system Both types of activated carbon, PAC and GAC, can be used in the BAC process In this study, the term BAC process is referred to the process in which PAC is added directly to the bioreactor

Many researchers have suggested that a synergy exists between activated carbon and microorganism Therefore, the BAC process could remove organic compounds more efficiently than what would be expected from either biodegradation or adsorption alone

The addition of PAC has served several process advantages including:

(1) system stability during shock loads, (2) reduction of refractory pollutants, (3) color and ammonia removal and (4) improved settleability (Metcalf and Eddy, 1991)

The dosage of PAC and the mixed-liquor-PAC-suspended-solids concentration are related to the sludge age as follows:

Xi θc

Xp = ⎯⎯⎯ Eq 2.6

θ Where : Xp : equilibrium PAC MLSS content, mg/L

Xi : PAC dosage, mg/L

θc : solid retention time, d

θ : hydraulic retention time, d Carbon dosage ranges typically from 20 to 200 mg/L With higher sludge ages, the organic removal per unit of carbon is enhanced, thereby improving the process efficiency Reasons cited for this phenomenon include (Eckenfelder, 2000):

(1) additional biodegradation due to decreased biological toxicity or inhibition via activated carbon,

(2) degradation of normally nondegradable substances due to increased exposure time to the biomass through adsorption on the carbon,

(3) replacement of low-molecule-weight compounds with high-molecule-weight compounds, resulting in improved adsorption efficiency and lower toxicity

Pirbazari et al (1996) explained the removal of contaminants in BAC process was

a combination of biodegradation and adsorption Biodegradation of organics occurred both

in suspended form in the liquid form and attached form at the biofilm developed on the carbon particles The major factor for achieving high organic removal was conceivably the retention time of enzymes capable of degrading the xenobiotic compounds The extended contact time between the microbial biota and organic solutes enhanced the biodegradation and/or mineralization of the contaminants via microbial enzyme attack

2.3.3 MBR process coupled with BAC process

The combination of BAC process and MBR process has gained special attention from researchers because the PAC addition in this hybrid system would play the two roles simultaneously: enhancement of pollutant removal and reduction of membrane fouling

Seo et al (1997) investigated the sorption characteristics of biological powdered activated carbon (BPAC) in a hybrid membrane process that was accomplished by introducing PAC into a MBR with external cross flow microfiltration The system was used for reclamation of secondary sewage effluent PAC dose used in the system was 0.5 g/L The results showed that a removal of 66% was achieved for refractory organic materials The sorption capacity of the BPAC was almost four times higher than that of fresh PAC The removal efficiency tends to increase when increasing PAC concentration

No information about membrane fouling was reported in this study

Pirbazari et al (1996) studied the treatment of landfill leachate using a hybrid technology known as the ultrafiltration-biologically active carbon (UF-BAC) process The system consisted of a bioreactor and a cross-flow ultrafiltrtion unit The wastewater was fed continuously into the reactor along with PAC slurry The contents of slurry were continuously pumped through the filtration unit This configuration was similar to the externally pressurized cross-flow MBR configuration The process removal efficiencies were in range of 95 - 98% in term of COD and exceeded 97% for specific organic pollutants The process compared well with the powdered activated carbon treatment (PACT) process in terms of organic removal and produced higher quality effluent in terms

of suspended solids (100% removal) The study demonstrated that addition of 1% PAC mitigated the permeate flux deterioration Without PAC addition, the permeate flux reduced to zero within the first hour of process operation The mechanism for mitigation of the permeate flux deterioration was explained by the formation of a PAC particular layer

on the membrane surface This permeable PAC particular layer would filter out microbial cells and colloids, preventing them reaching the membrane surface to cause a biofilm on the membrane surface (Figure 2.5)

Kim et al (1998) conducted a study to compare ultrafiltration characteristics of activated sludge and biological activated carbon (BAC) sludge The activated sludge reactor was fed with synthetic wastewater using glucose as carbon source The BAC sludge was prepared by adding PAC to activated sludge reactor for one month The biomass suspended solids of the two systems were maintained equal at 3000 mg/L The PAC dose varied in range of 0.5 – 1.8 mg/L The two kinds of sludge were then experimented ultrafiltration characteristics using a batch stirred cell unit at constant transmembrane pressure of 1 bar and stirring speed of 200 rpm The results showed that the addition of PAC increase the permeate flux in spite of the increase of suspended solid and reduced floc size This finding was explained by the fact that BAC sludge have lower compressibility and higher porosity than normal activated sludge and BAC sludge have lower content of EPSs inside the floc than activated sludge The authors also confirmed the PAC addition as

a mere filter aid resulting in flux decline

It is noted that membrane filtration experiment of the above studies were conducted

as crow-flow ultrafiltration No study of flux enhancement effect of PAC addition in submerged MBR and microfiltration has been found in the literature so far Because the mechanisms of the formation of dynamic cake layer in the two cases would be quite different, effect of PAC addition in submerged MBR to membrane fouling is still questionable

3.1 Materials and microorganisms

3.1.1 Feed wastewater

In this study, synthetic wastewater was used as feed wastewater Lubricant oil was

used to represent the oil fraction in the oily synthetic wastewater Data from a survey for

various gas stations in Thailand (Table 2.1) showed that O&G concentration and BOD of

carwash wastewater were in range of 7.8 – 280.5 mg/L and 10 – 1220 mg/L respectively

The oil concentration in this study was varied in range of 150 - 600 mg/L Glucose was

supplemented to the feed wastewater to simulate the carbon source other than the oil in the

real wastewater BOD contributed by glucose in the feed wastewater was maintained at

100 mg/L throughout all the experiments

Table 3.1 Composition of the feed wastewater

Emulsifier concentration was varied proportional to oil concentration

(4) Glucose concentration was maintained constant at all experiments, providing a BOD

value of approximate 100 mg/L

The oil fraction, provided in form of oil/water emulsion, was prepared by adding

1.0 g of oil and 100 mL of emulsifier solution (0.1% volume) to tap water to make one liter

of the stock oil/water emulsion solution The mixture was blended by the Heidolph Stirrer

RZR1 (Germany) at the speed of 1500 rpm for 4 hours Commercial lubricant Performa

SAE 20W-50, manufactured by Petroleum Authority of Thailand (PTT), was used in the

study The emulsifier used was Ter 10 12A, a non-ionic emulsifier, which is principally the

main component of commercial car wash detergent The stock oil/water emulsion was then

diluted to desired oil concentration and was added glucose and nutrient salts to ensure

favorable conditions for microorganism growth Typical composition of the feed

wastewater is presented in Table 3.1

Chapter 3 Methodology

Relative contributions of COD and BOD5 (20 C) from the oil, glucose and

emulsifier of a typical feed wastewater sample having oil concentration of 150 mg/L are

shown in Figure 3.1

Figure 3.1 Relative contributions of COD and BODfrom the oil, glucose and emulsifier of

a typical feed wastewater sample having oil concentration of 150 mg/L

3.1.2 Powdered activated carbon

The activated carbon used in this study was untreated powdered activated carbon

(PAC) supplied by Sigma Chemical Company Major properties of the PAC are presented

in Table 3.2 Nominal size distribution of the PAC was in range of 100 – 400 mesh that is

equivalent to 37 – 150 µm The PAC was boiled and washed several times with distilled

water to remove impurity and then dried at 105oC and kept in desiccator before used

Table 3.2 The major properties of the PAC

Parameter Value Surface area 750 m2/g

Mesh size 100 – 400 mesh

Methylene blue adsorption 10 –15 g/100g

Activated sludge collected at aeration tank of a wastewater treatment plant was

used as microorganism seeding The sludge was acclimatized to the oily wastewater at oil

concentration of 150 mg/L in a batch reactor with fill-and-draw process for two months

3.2 Overall experimental course

The experimental works can be divided into three main parts: PAC adsorption

isotherm experiment, batch experiment and MBR experiment Overall experimental course

is outlined in Figure 3.2

Feed wastewater

COD = 552.6 mg/L BOD = 261 mg/L

From emulsifier (13.6 mg/L)

COD = 28.6 mg/L BOD = 20 mg/L

PAC Adsorption Isotherm Experiment

Batch Experiment

BAC Process Run

AS Process Run

MBR Experiment

Optimum HRT run

Shock loading run

Long run A

Long run B (with PAC addition)

PAC Adsorption Isotherm Experiment

Batch Experiment

BAC Process Run

AS Process Run

MBR Experiment

Optimum HRT run

Shock loading run

Long run A

Long run B (with PAC addition)

Figure 3.2 Overall experimental course

3.3 Adsorption isotherm experiments

This experiment was to determine the adsorptability of the PAC on the oil fraction

of the feed wastewater The experiment was conducted at the temperature of 25 ± 0.2o

C using a series of 500-mL flasks containing 250 mL of sample The flasks were agitated by

a rotatory shaker at 100 rpm The adsorptability was measured in term of COD First, the equilibrium time experiment was carried out to determine the time needed for the adsorption process reaching the equilibrium state In this experiment, initial adsorbate concentration and PAC dose were maintained constant while agitation time was varied The equilibrium time (4h) was used as agitation time in the adsorption isotherm experiment which was conducted at different initial adsorbate concentrations and PAC doses The experimental conditions are presented in Table 3.3

Table 3.3 Experimental conditions for equilibrium adsorption time test

Initial COD (mg/L)

PAC dosage (mg/L)

Contact time (h) Equilibrium time experiment 343 100 1, 2, 3, 5, 8, 12 Adsorption isotherm experiment 124; 240 20, 50, 100, 200 4

3.4 Batch experiments

These experiments were conducted to determine preliminary parameters of the biodegradation of the oily wastewater through activated sludge process (AS) and biological activated carbon (BAC) process

The batch reactors were operated at sequencing batch reactor (SBR) mode in which each cycle went through the four main steps: fill, react (aeration), settle and draw Two-liter-glass-bottles were used as model reactors There were four reactors: one was run as activated process (AS) and to the remaining reactors PAC was added at different doses of

100, 300 and 500 mg/L

The reactors were mixed and aerated by compressed air through stone air diffusers

to maintain suitable mixing and to supply sufficient dissolved oxygen for microorganism growth (DO = 2 - 4 mg/L) The pH of the reactors was controlled within the range of 6.2 – 8.0 by adding NaOH when needed The reactors were aerated at different periods of time

to achieve various HRTs Calculation of HRT is followed Equation 3-1 (Grady et al,

1999)

ζ V HRT = ⎯⎯⎯ Eq 3.1

Q Where:

V: reactor volume (2 L)

Q : volume of effluent withdrawn per cycle = volume of influent filled per cycle =1L

ζ : aeration time per cycle (h)

The reactors were then let to settle for 2 hours After the settling step, one liter of the supernatant from each reactor was decanted for COD and O&G analyses Each reactor was then filled with one liter of the feed wastewater and aerated to start the next cycle

The reactors were run at five different HRTs of 3, 6, 10, 13 and 16 hours Each

HRT run lasted for four days The feed wastewater used in the batch experiments had the

oil concentration of 150 mg/L The reactors were operated without sludge wastage except

for sampling purpose

3.5 Membrane bioreactor experiments

3.5.1 Membrane bioreactor set-up

Laboratory scale membrane bioreactor was used in this study Schematic diagram

of the experimental set-up for the membrane bioreactor experiments is presented in Figure

3.3 Photograph of the lab-scale reactor is presented in Figure 3.4 The reactor was made of

transparent acrylic pipe having diameter of 100 mm The working volume of the reactor is

6 L A U-shaped hollow fiber microfiltration membrane was immersed in the reactor The

membrane was manufactured by Sterapore® Company (Japan) having a pore size of 0.1

µm Characteristics of the membrane module are given in Table 3.4

Table 3.4 Characteristics of the membrane module

Characteristic Membrane material polyethylene

Membrane type hollow fiber Pore size 0.1 µm Surface area 0.42 m2

The influent was fed at the bottom of the reactor from the level control tank This

tank operated based on an automatic valve that controlled the flow-in by a floating buoy

This set-up allowed the feed wastewater flow into the reactor at the same flow rate as the

influent being withdrawn The effluent was withdrawn through the membrane module by a

roller pump Flow rate of the permeate was adjusted by a pump speed controller The

reactor was continuously aerated by compressed air through stone diffusers placed at the

bottom of the reactor

The reactor was operated with periodic air backwashing The alternative operation

of filtration and air backwashing was controlled by an intermittent controller (timer) and

solenoid valves Each operational cycle consisted of 25 minutes of filtration followed by 3

minutes of air backwashing at air pressure of 300 kPa and 1 minute for releasing the

pressure in the system before starting next cycle

3.5.2 Experimental conditions

The MBR experiments were carried out through four runs named as Optimum HRT

Run, Shock Loading Run, Long Run A and Long Run B Operating conditions of the four

runs are presented in Table 3.5

(1) Optimum HRT Run: The aim of this run was to determine appropriate HRT

for the MBR system The oil concentration of the feed wastewater in this run was 150

mg/L Based on the results from the batch experiments, four HRT values of 8, 6, 4 and 2 h

were experimented

Effluent tank

Level control tank

Air diffuser

Pump

Solenoid valve

Air pressure Regulator F

Influent

Effluent tank

Level control tank

Air diffuser

Pump

Solenoid valve

Air pressure Regulator F

Influent

Figure 3.3 Membrane bioreactor setup

(2) Shock Loading Run: In this run, stability as well as capability of the treatment

system against shock loading was experimented The system was run at the HRT = 4h as selected from the previous run with increasing oil loading The oil concentration in the feed wastewater was increased stepwise from 300 mg/L to 600 mg/L with an increment of

150 mg/L

(3) Long Run A: This run was to investigate the membrane clogging course under

long term basis The reactor was run at HRT = 4 h and oil concentration of 300 mg/L This run was ceased when there was a sharp increase in transmembrane pressure

(4) Long Run B: In this run, PAC was added to the reactor at dose of 2 g/L to

develop BAC sludge in the MBR reactor The run was to evaluate the effect of the PAC addition on the performance of the MBR reactor in terms of reduction of membrane clogging rate and permeate quality improvement This run was conducted at the same operational conditions as the Long Run A., i.e at HRT = 4 h and oil concentration of 300 mg/L This run was ceased when there was a sharp increase in transmembrane pressure This made the comparison between the two runs possible

Figure 3.4 Laboratory scale MBR system to treat oily wastewater

There was no sludge wastage except being sampled 10ml daily during Optimum HRT run and Shock Loading Run because the operational conditions changed in a fairly short time The solid retention time (SRT) was maintained at 50 days through the Long Run A and B by wasting 120 mL of the mixed liquor daily Calculation of the SRT is as follows:

Qw : wastage sludge flow rate = 0.12 L per day

Xw : MLSS in the wastage sludge flow (X = Xw)

Qe : Effluent flow rate

Xe : MLSS in the effluent = 0 For Long Run B, virgin PAC was supplemented to the reactor daily to compensate the PAC being withdrawn with the sludge

The pH of the reactor was maintained in range of 6.2 to 8 by adding NaOH when needed DO was adjusted in range of 2 – 4 mg/L After each run, the membrane underwent chemical cleaning to restore its permeability

Table 3.5 Operational condition of the MBR experiments

Run

Influent oil concentration (mg/L)

HRT (h)

SRT (day)

PAC dose

in reactor (g/L)

Parameters to be measured

Long Run A 300 4 50 0

Influent COD, effluent COD and O&G, reactor O&G, MLSS, TMP, EPS,

Rm, Rc, Rf

Long Run B 300 4 50 2

Influent COD, effluent COD and O&G, reactor O&G, MLSS, TMP, EPS,

Rm, Rc, Rf

3.5.3 Measured parameters in the MBR experiments

COD of the influent was analyzed to monitor the strength of the prepared feed wastewater COD and O&G of the permeate were determined to evaluate the effluent quality MLSS analysis was performed to check the microorganism growth O&G concentration of the mixed liquor was also determined to inspect the accumulation of the oil in the reactor if any Variation of transmembrane pressure was recorded to monitor membrane-clogging process

For Long Run A and B, content of extracellular polymeric substances (EPS) of the activated sludge was additionally determined to inspect the relationship between EPS and membrane clogging process if any Various components of membrane resistance contributing to membrane clogging were also measured as described in details in the next section

3.5.4 Membrane cleaning

The membrane after each run was taken out of the reactor for cleaning The membrane was first flushed with tap water to remove the cake layer attaching on the membrane surface Next, the membrane was soaked in NaOH 2% solution for 12 hours The membrane was then washed with tap water and soaked in HNO3 1% solution for 4 hours The membrane subsequently washed with tap water again and was measured membrane resistance before used in the next run

3.5.5 Membrane resistance measurement

Membrane resistance was measured based on the resistance-in-series model (Choo

and Lee, 1996) according Equation 2.1 and 2.2

Applying this model, membrane resistance was measured by filtrating with pure

water at different filtration fluxes and recording the corresponding transmembrane

pressures Membrane resistance was derived from the slope of the linear curve of ∆P

versus J as described by the following equation:

∆P = Rt.µ J + ∆Po Eq 3.3 Where:

∆Po is the initial pressure to overcome the membrane set-up system resistance

For further understanding of various components of membrane resistances causing

the membrane clogging, detailed membrane resistance measurements after Long Run A

and B were conducted in accordance with the procedure suggested by Kim et al (1998)

The total resistance (Rt) was measured right after the run with the membrane still in its

clogging condition Rm+ Rf was obtained by measuring the resistance of the membrane

after being washed with tap water to remove the cake layer The membrane resistance at

the beginning of the run (after chemical cleaning) was considered as Rm Value of Rc was

derived from Rt, Rm and Rf using Equation 2.2

3.6 Analytical methods

The analyses of COD, BOD, O&G and MLSS were followed the procedures

recommended by Standard Methods (APHA et al., 1995) List of these methods is

presented in Table 3.6

Table 3.6 Analytical methods

COD Closed reflux titrimetric method Standard methods, 5220 C

Oil and grease n-Hexane extraction

Partition-Gravimetric method Standard methods, 5520 C Total suspended solid

(MLSS) Dried at 103 – 105

o

C Standard methods, 2540 D

For determination of the oil concentration in the reactor, 100 mL of the mixed liquor

sample was filled in a one-liter extraction funnel and was let for settling in 30 minutes The

settled sludge was withdrawn from the bottom of the funnel The supernatant was analysed

for O&G by Partition-Gravimetric method

Thermal extraction method was used to extract EPS (Morgan et al., 1990) A

measured volume of sludge solid was concentrated by a centrifuge at 2000 rpm for 10

minutes The sludge was then resuspended in distilled water before being heated at 80oC

for 1 h The solution that contains the extracted EPS was harvested by centrifuging first at

2000 rpm for 10 minutes and then at 5000 rpm for 20 minutes to remove the sludge solids

The extracted EPS solution was then analysed for total EPS and EPS composition

(carbohydrate and protein) Total EPS was determined by precipitated out from the

solution by adding three parts of solvent mixture of acetone and ethanol (3 vol: 1 vol) to one part of the extracted EPS solution and then kept at 4oC overnight The quantity of the extracted EPS was determined as suspended solid using filtration method described in Standard Method 2540 D Contents of carbohydrate and protein in the extracted solution were determined using spectrophotometric method at wave lengths of 480 nm and 570 nm respectively

4.1 Adsorption isotherm of the PAC

Equilibrium adsorption time of PAC to the oil was determined by varying contact times of the adsorption tests The results are presented in Table A.1 and illustrated in Figure 4.1 The results showed that the adsorption process reached the equilibrium status somewhere between two and four hours The COD values of the supernatant at the contact time longer than two hours were changed slightly in range of 164 mg/L to190 mg/L Therefore, it is suitable to take 4 hours as the equilibrium time for the adsorption isotherm experiments

0 50 100 150 200 250 300 350 400

Figure 4.1 Variation of equilibrium COD vs different contact times

(PAC dose = 100 mg/L; T = 25oC; agitation speed = 100 rpm) The adsorption isotherm was determined at initial COD of 124 mg/L and 240 mg/L and at PAC doses ranging from 20 mg/L to 200 mg/L The results were presented in Table A.2 The experimental data were plotted in accordance with Freundlich and Langmuir isotherm models (Figure 4.2 and 4.3) For Freundlich isotherm, the 1/n value of 0.22 was within the magnitude of significance of reported isotherms The reported 1/n values for various priority pollutants are in range of 0.16 – 2 (Montgomery, 1985) Comparison of the empirical values for Langmuir isotherm was impossible because of the lack of similar work in the literature

Based on the R-square values, the Langmuir isotherm was more appropriate model for the adsorption of lubricant oil onto PAC The R-square value for Langmuir isotherm was 0.952, which is approaching 1, while that for Freundlich isotherm was only 0.689 It is noted that lubricant oil mainly consists of long chain compounds which have large molecule size This would be the reason why the adsorption of lubricant oil onto PAC follow the Langmuir isotherm model which is based on the assumption that the adsorbate only can be adsorbed a monolayer onto the surface of the adsorption medium

Chapter 4 Results and Discussions

Kf = 225.84 1/n = 0.2244

Figure 4.2 Experimental data plotted according to Freundlich isotherm

(T = 25oC, agitation speed = 100 rpm, contact time = 4 h)

C e /(x/m) = 0.0341 + 0.0012C e

R2 = 0.952 0.00

0.05 0.10 0.15 0.20 0.25 0.30 0.35

(x/m) ab a

a = 833

b = 0.352

Figure 4.3 Experimental data plotted according to Langmuir isotherm

(T = 25oC; agitation speed = 100 rpm, contact time = 4 h)

4.2 Batch experiments

Batch experiments were conducted using four 2-liter reactors Reactor 1 was purely activated sludge process In the reactors 1, 2 and 3, PAC was added at doses of 100, 300 and 500 mg/L respectively The results are presented in Table A.3 Averaged values of the experimental data is summarized in Table 4.1 and illustrated in Figure 4.4 and 4.5

For AS process (Reactor 1), the removal efficiency of COD increased from 64% to 86% when HRT increased from 3h to 16h Similarly, the removal efficiency of O&G varied from 67% to 92% with HRT increasing from 3h to 16h At HRT=13h, the removal efficiencies of COD and O&G reached 80% and 90% respectively At this HRT, the COD

of the effluent was 109 mg/L and that of O&G was 15 mg/L The results revealed that biodegradation of oily wastewater took place at significant level and treatment of oily wastewater through biological process is possible

Table 4.1 Average results of the batch experiments (Inf COD = 550 mg/L; Inf.O&G =

150 mg/L)

HRT

(h)

MLSS (mg/L)

CODeff (mg/L)

O&Geff.

(mg/L)

COD removal (%)

O&G removal (%)

Note: MLSS including PAC and biomass

There was no distinguished difference between the performance of BAC process (Reactors 2, 3 and 4) and AS process at HRTs longer than 10h At HRT = 13h, the COD removal efficiency of BAC process was in range of 80–81% in comparison with that of AS process of 80% The corresponding values for O&G were 87–91% and 90% The same situation was observed at HRT = 16h This would be due to the effect of biodegradation process was predominant over the effect of adsorption process at long contact times The predominant effect of biodegradation could also be the reason why the performance of Reactor 2, which had PAC dose of 100 mg/L, was not improved in comparison with AS process at HRT of 10, 6 and 3h

Figure 4.4 Variation of COD removal efficiency vs HRT (Influent COD = 550 mg/L; Influent O&G = 150 mg/L)

Figure 4.5 Variation of O&G removal efficiency vs HRT (Influent COD = 550 mg/L; Influent O&G = 150 mg/L) There was a tendency of effluent quality improvement of the BAC process with increasing PAC dose The removal efficiencies of COD and O&G of Reactor 3 (PAC =

300 mg/L) and Reactor 4 (PAC = 500 mg/L) were almost higher than those of Reactor 1 The COD removal efficiency of Reactor 4 increased from 71% to 80 % when HRT increased from 3h to 10h The corresponding values for Reactor 3 and Reactor 1 were 68%

- 77% and 64% - 77%, respectively Similar pattern could be found for O&G The most pronounced enhancement was observed at PAC dose = 500 mg/L and HRT = 3h (Figure 4.6 and 4.7) It was learned from the adsorption isotherm experiment that adsorption process reach equilibrium at contact time longer than 2h This would suggest that the adsorption of PAC help to remove the oil from the supernatant and therefore improve the performance of the BAC process at short HRTs

78.4 109

126

161

142

77.4 111

165 197

10 6

Figure 4.6 Comparison of COD removal efficiency between AS process

and BAC process (PAC = 500 mg/L)

12 15

20

34 39

12 16

24 37

10 6

Figure 4.7 Comparison of COD removal efficiency between AS process

and BAC process (PAC = 500 mg/L)

The results from the batch experiments suggested that the MBR experiments should

be conducted at HRTs shorter than 10h because the removal efficiency of MBR process is expected higher than AS system Another factor should be considered was the PAC/MLSS ratio For Reactor 4 (PAC = 500 mg/L), the averaging value of MLSS was about 3500 mg/L, therefore, the PAC/MLSS ratio was roughly 1/7 It is suggested that PAC should be experimented at higher dose because MLSS in the MBR is much higher than that of AS process

4.3 MBR experiments

4.3.1 Membrane resistance measurement

Before used in the experiments, the membrane was measured membrane resistance Membrane resistance of the membrane after chemical cleaning were measured as described

in Section 3.5.4 before the next run Detail data of membrane resistance measurements are presented in Appendix B Membrane resistances of the new membrane and the membrane after undergoing chemical cleaning are given in Table 4.2 and illustrated in Figure 4.8 Table 4.2 Membrane resistances of the new membrane and the membrane

after chemical cleaning

Membrane resistance

(1/m) New membrane 8.95x1011

After Optimum HRT run 8.73x1011

After Shock loading run 8.98x1011

After Long run A 8.60x1011

After Long run B 10.55x1011

With an exception of the Long run B, the membrane resistances were almost recovered to initial value of the new membrane The membrane resistances after chemical cleaning for Optimum HRT run, Shock loading run and Long run A were in range of 8.60x1011 - 8.98x1011 1/m, less than ± 3.9% different in comparison with the initial value

of the membrane Therefore, these values could be considered as the same magnitude of significant as the resistance of the new membrane of 8.95x1011 1/m, i.e the membrane was almost recovered to its original conditions after chemical cleaning Visually, the cake layer attaching on the membrane surface was removed mostly by flushing with tap water When the membrane was soaked in sodium hydroxide solution, the solution became turbid and brownish yellow particles could be observed in the solution This means that sodium hydroxide solution took effect on removing the particles out of the pores of the membrane The acid nitric solution remained clear after membrane soaking time This suggests that the acid solution dissolved the substances attaching on the membrane and/or the foulants had been removed mostly by the alkalis solution However, the used membrane could not be retrieved back to the pure white color of the new membrane The used membrane became slightly yellowish white

For Long run B, the only run that PAC was added to the reactor, the membrane resistance could not be retrieved back to the new membrane resistance by chemical cleaning although additional cleaning was applied In addition to the normal chemical cleaning procedure, the membrane after Long run B was undergone an additional soaking time of 12h in nitric acid solution in an attempt to recover its membrane resistance Membrane resistances after each step of chemical cleaning course after Long run B are presented in Table 4.3 The data showed that with only the alkalis solution, the membrane resistance (12.58x1011 1/m) was still 40.5% higher that of the new membrane Acid solution soaking significantly reduced the resistance to 10.55x1011 1/m, about 17.9% higher than that of the new membrane However, in spite of the long contact time for acid solution soaking, the membrane could not be retrieved back to its initial membrane resistance This would suggest that the membrane pores in the case of Long run B be

plugged by submicron carbon particles in addition to the organic components and the carbon particles be hard to be removed by NaOH and HNO3 solutions This would be the reason why the membrane resistance could not be recovered in the case of Long run B In future works, composition of the substances attaching on the membrane surface should be determined somehow for better understanding of the cause of membrane fouling

12

New membrane

After Optimum HRT run

After Shock loading run

After Long run A

After Long run B

After soaking time of 4h in HNO3 solution 11.56x1011

After additional soaking time of 12h in HNO3 solution 10.55x1011

4.3.2 Optimum HRT run

The optimum HRT run was conducted to determine suitable HRT for the MBR system to treat the oily wastewater The run was carried out at four HRTs of 8, 6, 4 and 2h O&G concentration in the influent was maintained at 150 mg/L through out this run The experimental results are presented in Table A.4 and illustrated in Figure 4.9, 4.10, 4.11 and 4.12

The MBR system produced excellent quality effluent at all experimented HRTs With the COD in the influent ranged from 495 – 607 mg/L, the effluent COD was in range

of 26.3 – 59.8 mg/L, providing the removal efficiency about 89.9% - 95.5% In term of O&G, the removal efficiency was 97.8% - 99.9%, with O&G concentration measured by n-Hexane partition-gravimetric method was in range of 0.2 – 3.2 mg/L Visually, the effluent was very clear and oil free The oil fraction contributing to the O&G concentration