WO2008093287A1 - Treatment of wastewaters using dual-stage membrane bioreactors - Google Patents

Abstract

A wastewater treatment process including the steps of providing an acclimation membrane bioreactor system (2,4) for developing a desired microganism population inoculum. The system is in selectirely interruptible fluid flow connection with an effluent hydrolysis system (3). This allow to flow of microorganisms from the bioreactor system to be regulated independently from the effluent flow through the hydrolysis system.

Description

TREATMENT OF WASTEWATERS USING DUAL-STAGE MEMBRANE BIOREACTORS

THIS INVENTION relates to water treatment. More particularly, this invention relates to a process for treating industrial, domestic, or agri-sector wastewaters, or a combination thereof, using a dual stage membrane bioreactor.

BACKGROUND TO THE INVENTION

The combination of membrane technology with biological reactors is fast achieving considerable success for the treatment of domestic and industrial effluent. This is due mainly to the small footprint requirement, reduced sludge production, and the decrease in overall capital costs as a result of the availability of economically competitive membranes and modules. The operational characteristics of MBRs, however, are as diverse as the number of different module configurations in use.

Numerous conventional wastewater treatment facilities known to the Applicant employ aerobic activated sludge (AS) systems for the majority of wastewater treatment applications. These systems are used in conjunction with downstream separation via decantation. The quality, however, of the final effluents generated by conventional activated sludge systems is significantly dependent on the settling characteristics of the sludge as well as the hydrodynamic conditions within the sedimentation/decantation tanks. This necessitates large volume tanks enabling relatively long residence times, thereby to facilitate adequate liquid/solid separation. Effective operation of activated sludge processes however, requires sedimentation of the aggregated mixed microbial floes. This process is facilitated by the physico-chemical interactions between the microorganisms, mineral particles, multivalent cations, and microbial exopolysaccharide production enabling the binding of the microorganisms thus resulting in sufficient floe size development to promote settling. The exopolysaccharide-producing microorganisms are, however, usually not the microorganisms responsible for the majority of COD reduction within these processes. As a result, activated sludge plants usually operate at significantly lower efficiencies due to this intrinsic requirement to maintain the exopolysaccharide-producing populations. Furthermore, the overall biological process may be limited due to diffusional constraints within the flocculated biological mass. In contrast, MBRs do not have this requirement due to the fact that microfiltration (MF) and ultrafiltration (UF) membranes prevent the loss of biological solids and high molecular weight solutes. As a result, the maintenance of a high biomass concentration within the MBR potentially allows for complete mineralisation of influent organic matter. Furthermore, most wastewater treatment plants experience peak flows of at least twice the normal average flow, with diurnal flow variations being even higher. Compensation of these daily variations in flow can be achieved by construction of large tanks to facilitate flow equalization. This option is, however, impractical both economically and in terms of footprint requirements. The versatility of membrane bioreactor separation systems enables these processes to be designed to cope with the design dry weather flow. The modular nature of MBR systems then compensates for both the minimum plant inflow by shutdown of some modules as well as the maximum plant inflows by raising the permeate flux due to maximum plant inflow periods being usually in the order of a few hours.

The membrane bioreactor constitutes a combination of biological and physical processes for solid/liquid separation thus combining conventional suspended growth systems with membrane filtration. Membrane processes provide reliable, high rate filtration as small footprint units. Furthermore, membrane filtration processes are easily retrofitted to existing wastewater treatment facilities. The associated small footprint requirement of membrane processes is also facilitated by the fact that both primary and final settling tanks are not required for the effective operation of membrane systems. MBRs are most suitable for processes where relatively long residence times are required to achieve sufficiently high biomass concentrations in order to achieve desirable pollutant removal efficiencies.

Typically, mixed liquor suspended solids (MLSS) concentrations as high as 20 000mg/L can be maintained by MBRs in comparison to 4000mg/L typical of conventional activated sludge/secondary clarifier systems. This results in significantly lower reactor volume requirements with minimum sludge wastage as well as less sludge production from longer sludge retention times (SRT). Furthermore, the operational characteristics of MBRs allows for the complete uncoupling of the hydraulic retention time (HRT) with that of the SRT.

Two main configurations of solid/liquid separation membrane processes are employed in the operation of MBRs. These are defined as immersed (submerged) membrane process operation or in-series membrane separation units operated in the external circuit of the bioreactor. At steady-state conditions, MBRs are capable of greater than 90% removal of influent COD. The production of high quality permeates at high organic loading rates is also associated with significantly greater MLSS concentrations in comparison with conventional biological wastewater treatment systems.

Loading rates of between 0.2 kgCODm^"3.d^"1 for aerobic systems and as high as 19.7 kgCODm^{" 3}.d^"1 for anaerobic systems have been recorded with MLSS concentrations approaching 28.7 and 113.3kg/m³ for aerobic and anaerobic systems, respectively. Typically this variability is also reflected in the operating conditions associated with MBRs with HRTs ranging from 1.8 h for aerobic processes to as much as 240 h for anaerobic processes. Of the various membrane types, materials, and configurations used in MBRs, the steady-state operating flux of these systems also varies considerably from 2.1 to 70 Lm^"2h^"1.

Two main configurations of solid/liquid separation membrane processes are employed in the operation of MBRs. These are defined as immersed (submerged) membrane process operation or in-series membrane separation units operated in the internal circuit of the bioreactor. At steady-state conditions, MBRs are capable of greater than 90% removal of influent COD. The production of high quality permeates at high organic loading rates is also associated with significantly greater MLSS concentrations in comparison with conventional biological wastewater treatment systems.

It is an object of the invention to provide a process for treating wastewaters using a dual-stage membrane bioreactor.

SUMMARY OF THE INVENTION

According to one aspect of the invention, in a dual-stage membrane bioreactor system, there is provided a wastewater treatment process for removing undesired impurities from an effluent feed, the process including the steps of: providing a discrete acclimation bioreactor system for developing a desired microorganism inoculum, the system being in selectively interruptible fluid flow connection with a discrete effluent hydrolysis system, such that the flow of desired microorganisms from the bioreactor system to the effluent hydrolysis system may be regulated independently from the effluent feed flow through the hydrolysis system; providing, into the hydrolysis system, a supply of effluent flow from which undesired impurities are to be removed; and regulating the supply of inoculum via a discrete operating strategy applied to the flow contacting the inoculum from the acclimation bioreactor system, thereby to maintain sufficient levels of a desired microbial population to effect COD removal of between 50% and 100% from the hydrolysis system.

As such, the invention provides a wastewater treatment process for removing undesired impurities from an effluent feed, the process including the steps of: providing a discrete acclimation bioreactor system for developing a desired acclimated microbial population inoculum, the system being in selectively interruptible fluid flow connection with a discrete effluent hydrolysis system, such that the flow of desired microorganisms from the acclimation bioreactor system to the effluent hydrolysis system may be regulated independently from the effluent feed flow through the hydrolysis system via an integrated process-controlled intermediate membrane bioreactor system; and providing, into the hydrolysis system, a regulated supply of acclimated inoculum from the acclimation bioreactor system to maintain sufficient levels of a desired microbial population to effect COD removal of between 50% and 100% from the supply of effluent flow from which undesired impurities are to be removed within the hydrolysis system.

The inoculum may be retained by a semi-permeable membrane in the bioreactor system. The acclimation bioreactor system may be a solid-liquid retention membrane bioreactor and may be fed discretely with the effluent to seed and acclimatize the inoculum to the particular effluent to be treated, prior to release of inoculum into the hydrolysis system, thereby to provide a waste- specific microbial population.

The bioreactor system may be seeded with inocula consisting of MLSS at between 1000 and 20 000 mg.L^"1, typically about 10 000 mg.L^"1.

A feed from the bioreactor system to the hydrolysis system may be opened intermittently to provide fresh microorganisms in the form of a mixed liquor to the hydrolysis system, as dictated by effluent feed characteristics and conversion rates.

Alternatively, or additionally, the feed from the bioreactor system may be trickle-fed into the hydrolysis system on a continuous basis. The feed from the bioreactor may trickle into the hydrolysis system at a rate of 1% v/v to 20% v/v, preferably about 5%. The microorganisms may be cultured in the bioreactor system at a concentration of 0.01 g/L to 20 g/L, dependant on the initial inoculum size and effluent source. A typical concentration may be 5 g/L.

The microorganisms may be held in the hydrolysis system at a concentration of 0.01 g/L to 20 g/L, typically about 5 g/L, dependant on the initial inoculum size and effluent source. The microorganisms may be held in the hydrolysis system at a concentration of about 10 000 mg.L^"1, dependent on the inoculum size and effluent source.

The effluent feed may be provided to the acclimation bioreactor system at a rate of 1 % v/v/h to 100% v/v/h, dependant on the effluent source and system volume. The effluent feed may also be provided to the hydrolysis system at a rate of 1% v/v/h to 100% v/v/h , dependant on the effluent source and system volume.

The process may include the step of monitoring the composition of the microorganism population in the bioreactor system at set intervals, say every 24 hours to 20 days using molecular techniques. Monitoring of the microorganism population may include the steps of obtaining a sample of the microorganism population from the bioreactor system and subjecting said samples to: DNA analysis, lipid analysis, protein analysis. Typically, this may include phospholipid fatty acid analysis (PLFA) for identifying the relative proportions of the major microbial groups. This may also include ribosomal DNA fingerprinting using 16S/18S rDNA for indicating variation at the specific microbial strain level. This technique can be used to discern bacterial 16S rDNA from fungal 18S rDNA fragments. The analyses may also include techniques such as community DNA hybridization, %G+C profiling, restriction digestion and sequence comparisons, amplified ribosomal DNA restriction analysis (ARDRA), terminal restriction fragment length polymorphism (T-RFLP), ribosomal intergenic spacer analysis (RISA), as well as various combinations of these techniques. These techniques can be combined with electrophoretic separation based on melting (TGGE) or denaturing (DGGE) behaviour.

The process includes the step of optimizing operation of the bioreactor system and hydrolysis system in terms of residence time, cross-flow velocity, transmembrane pressure (TMP), and loading rates, where optimal microbial populations associated with the greatest productivity and process efficiency for a given effluent may be maintained. The bioreactor may include a polymeric or ceramic membrane and may be either an ultrafiltration or microfiltration membrane. The membrane may be in the form of a tubular, hollow fiber, or flat sheet configuration, or a combination thereof..

The membrane may therefore have an average pore size of between 0.01 μm and 5.0 μm, preferentially characterized by a narrow pore size distribution of ±0.01 μm

The process further may include the step of recirculating microorganisms which may be present in the hydrolysis system.

Additionally, the process may include the steps of clearing the membranes by backpulsing.

Further aspects of the invention will now be illustrated in a non-limiting fashion with reference to the accompanying description and drawings

DRAWINGS

In the drawings:

Figure 1 shows as schematic diagram of an MBR operating configuration in accordance with one aspect of the invention, including: (1 ) SGL feed tank; (2) MBR recirculation tank; (3) MBR effluent receiving tank; (4) MBR membrane modules;

Figure 2.2 shows the average flux for the SGL-MBR over a period of 280 days (1700 operating hours). Data is represented as the average flux for the five membrane modules operated in-series;

Figure 2.3 shows COD removal efficiency of the MBR. Phenol and 2,4-dichlorophenol addition is indicated with arrows and respective concentrations in ppm (mg.L^"1);

Figure 2.4 shows COD removal efficiency of the AS process;

Figure 2.5 shows a dendrogram showing relationships between the various samples analysed by DGGE;

Figure 2.6 shows examples of DGGE profiles of 16S rDNA (A) and 18S rDNA (B) fragments in MBR tank and reactor samples (The letters (A, B, H and J) in Figure 3.6. (A) show the bands that were the marker positions used. A total of 23 operational taxonomic units (OTU's) were analyzed for presence/absence as well as band intensity data. Figure 3.6. (B) represents DGGE profiles of 18S rDNA. The profiles were less complex and diverse than the bacterial 16S profiles with a total of only 6 OTU's resolved and analyzed. The numerical data sets of both 16S and 18S rDNA profiles were further analyzed for OTU diversity using Shannon-Weaver biodiversity index (Figure 2.7.);

Figure 2.7 shows a comparison of community changes between bacterial and fungal OTUs in MBR tank and reactor modules, using the Shannon-Weaver diversity index;

Figure 2.8 shows ratios of Shannon-Weaver 18S/16S indices indicating the ratios of fungal and bacterial species in the tank and membrane reactor compartments of the MBR system;

Figure 2.9 shows the relative proportion of the major phospholipid fatty acid groups (mol %) within the MBR;

Figure 2.10 shows the relative proportion of the major phospholipid fatty acid groups (mol %) within the activated sludge system;

Figure 3.1 shows a piping and instrumentation diagram of a seeding and hydrolysis reactor in accordance with once aspect of the invention;

Figure 3.2 shows the initial operating flux of a submerged-type MBR system in accordance with an aspect of the invention;

Figure 3.3 shows the initial operating flux of external circuit MBR system;

Figure 3.4 shows COD removal associated with external circuit MBR system;

Figure 3.5 shows COD removal associated with submerged MBR system;

Figure 3.6 shows COD removal efficiency of the combined dual-stage system in accordance with an aspect of the invention. This graph also indicates the initial performance efficiency of the seeding tank system, MBR biofilm retention system and the final hydrolysis process; Figure 3.7 shows a comparison of the average COD concentrations associated with the unit operations of the combined dual-stage system of the invention, indicating the overall (250-days) performance efficiency of the seeding tank system, MBR biofilm retention system and the final hydrolysis process; and

Figure 3.8 shows a comparison of the average COD concentrations associated with the unit operations of the combined dual-stage system of the invention, indicating the overall (250-days) performance efficiency of the seeding tank system, MBR biofilm retention system and the final hydrolysis process.

DETAILED DESCRIPTION OF THE INVENTION

1. INTRODUCTION

Advances over the past 40 years in membrane bioreactor (MBR) technology have legitimised the application of MBRs as alternatives to conventional activated sludge (AS) processes. As with all wastewater treatment (WWT) technology options, including AS and particulate biofilm reactors, MBR treatment efficiency as an alternative WWT process is dependant on effective operational management. In MBRs, this equates to limiting the long-term fouling propensity of the membranes by using numerous fouling alleviation strategies. In this invention, focus on high permeate flow maintenance through the membranes whilst mitigating the fouling trend is obviated instead for a MBR-facilitated microbial ecology management strategy. Integration of the biomass retention characteristics of MBRs creates the ideal ecological management tool for exploiting the self-assemblage characteristics of biofilms and tapping into the phenotypic potential of the microbial community. Structural and functional diversity of biofilm communities derived from nascent effluent populations can be effectively managed using MBRs because of the ease of control of the hydraulic retention time, the potential for infinite sludge age and hence, retention of the critical slow-growing microbial population fraction (μ < 0.25/day), and the emerging genomics tools applicable to microbial community analysis enabling the quantification of community structure and phenotypic potential. 2. SEEDING REACTOR TREATMENT OF WASTEWATER

2.1 INTRODUCTION

2.1.1. MBR treatment of wastewater

Studies on MBR, activated sludge and anaerobic digestion systems have shown how the microbial population adapt to waste composition and that shock loading or changes in the operational parameters are inevitably detrimental to the 'remediation active' population. This may have increased cost implications for operating a MBR system. Monitoring the microbial population dynamics is thus essential for optimizing and operating such systems. With industrial effluents in particular, the use of micro- or ultrafiltration-based biomass retention enables the build-up of a waste-specific microbial population. Due to the biomass retention characteristics of membranes these populations do not necessarily include organisms responsible for flocculation as in traditional AS systems. This circumvents the necessity for maintaining microorganisms responsible for producing high levels of exopolysaccharide - as is necessary for floe formation in conventional AS systems. The resultant microbial population within the developed biofilm is therefore composed of a highly 'remediation active' population. Furthermore, the ability of biofilms to support a variety of different populations at various locations within the biofilm enables fixed film processes to be more amenable to the degradation of xenobiotics than the conventional activated sludge process.

2.1.2. Analysis and characterization of biofilm-associated microbial communities

The accurate assessment of the viable fraction of microbial populations using culture-dependent techniques such as Biolog^® and MIDI is limited by the lack of technology presently available to culture more than 90% of the microorganisms present in most complex communities such as those associated with biofilms and soil. Utilisation of non-culture dependent molecular based techniques circumvents this problem. As a result, direct assessment of non-culturable viable microorganisms present in complex interactive communities is possible by direct extraction of cellular components such as the nucleic acid and lipid biomembrane fractions. The combination of nucleic acid analysis with lipid biomarker identification and quantification can provide complementary information with each approach substantiating the results of the other.

2.1.2.1. Molecular techniques - DGGE analysis

Community genomic analysis using techniques such as community DNA hybridization, %G+C profiling, restriction digestion and sequence comparisons, amplified ribosomal DNA restriction analysis (ARDRA), terminal restriction fragment length polymorphism (T-RFLP), ribosomal intergenic spacer analysis (RISA), as well as various combinations of these techniques can reveal considerable information regarding the richness, evenness, and composition of the community. An increasingly popular tool within microbial ecology has been the PCR-based amplification of bacterial 16S rDNA and fungal 18S rDNA fragments combined with electrophoretic separation based on melting (TGGE) or denaturing (DGGE) behaviour. In combination with hybridisation, sequence analysis, community substrate utilisation, PLFA analysis, and various other culture methods, TGGE and DGGE profiling gives a reasonable estimation of the dominant in situ microbial community structure. Furthermore, 16S and 18S rDNA-based techniques also provide a wealth of information concerning specific microbial population activity and more importantly, community stability.

2.1.2.2. Molecular techniques - PLFA analysis

The analysis of phospholipid fatty acids is a useful measure of the viable biomass content, microbial community population structure, and the physiological status of the community. Although the analysis of PLFA profiles does not enable the identification of microorganisms at a species or strain level, it does, however, provide a comprehensive description of the microbial community based on the different functional groupings within the PLFA profiles. Furthermore, the identification of specific genera is possible by the presence of certain signature fatty acids (for example, c/s-octadec-8-enoic is associated with Type Il methanotrophs). The identification of physiological stress is also possible by monitoring increases in frans-mono-unsatu rated fatty acids. Through the use of different functional groupings from PLFAs, characterisation of the microbial community structure is possible using a community level approach by evaluating shifts in PLFA profiles. Changes in the PLFA profiles would therefore be indicative of changes or stress responses as well as the in situ nutritional/metabolic status within the microbial community.

By combining 16S and 18S rDNA-based nucleic acid analysis with lipid biomarker identification and quantification, stress responses experienced by the membrane-associated biofilm and planktonic tank communities may be evaluated and correlated with process operational characteristics. Furthermore, as PLFAs are associated with rapid turnover within the cell, immediate stress-related responses to shock or toxic loading effects may be correlated with shifts in the dominant in situ microbial community structure determined using 16S/18S rDNA- based DGGE analysis and subsequent sequence identification. Thus, specific MBR operating conditions under which the metabolic status and stability of the membrane- and tank-associated populations is subject to shifting may therefore be identified. The operation of the MBR may therefore be optimized in terms of residence time, cross-flow velocity, TMP, and loading rates, where optimal microbial populations associated with the greatest productivity and process efficiency may be maintained. 2.2. Membrane Bioreactor Process 2.2.1. Membrane bioreactor tank

The membrane bioreactor system is housed in a tank, used in association with the tubular ceramic membrane modules, had a working volume of 6OL. The MBR tank was maintained at 25-27⁰C for the duration of the experimental period. The dissolved oxygen concentration was maintained between 4 and 6mg.L-1 with the pH of the system remaining stable between 7.5- 8.5. The system was initially operated for a period of 7-days under batch operating conditions and thereafter operation followed a continuous feed-and-bleed approach for a total period of 20- days. Thereafter, the membrane module component of the membrane bioreactor system was started up. The feed-and-bleed approach entailed the removal of 20L.day-1 mixed liquor thereby maintaining a hydraulic retention time in the tank of 3-days. The removed mixed liquor was replaced daily with 2OL of fresh SGL make-up feed corrected to a C:N:P ratio of 100:10:1 using KH₂PO₄ZK₂HPO₄ and urea.

2.2.2. Membrane modules

Each single-membrane reactor module comprises an effective membrane area of 0.0067m². The membrane module construct, comprising three in-series modules, have a total membrane area of 0.02m² with each MBR assembled (comprising 5 constructs in series) having an effective total membrane area of 0.133m² for the membrane bioreactor (refer to Fig. 1 ). Fluid reticulation within the MBR system was achieved using a magnetic drive centrifugal pump with a recirculation flow rate of 150L.h^"1 from the MBR tank and a membrane module influent flow rate of 10L.h^"1. At this flow rate, the cross flow velocity (V) was 0.035m.s^"1 and the transmembrane pressure was <2kPa.

The hydrodynamics of the membrane module units enabled the maintenance of laminar flow conditions at this cross flow velocity with a resultant Reynolds number (Re) of 350. Periodic displacement of the accumulated biofilm in each membrane module for PLFA and 16S rDNA analysis was achieved by increasing the cross flow velocity to 0.21 m.s^"1 to induce turbulence within the system (Re=2100) thus facilitating removal of the accumulated biomass.

2.2.3. Activated Sludge Process

Activated sludge reactors comprising aerated CSTR-type vessels coupled to in-series clarifiers were used. The working aerated reactor volume was 3.4L and the temperature of the reactor was maintained at 20⁰C in a thermostat-controlled room. Air sparging of the reactor facilitated dissolved oxygen maintenance within the aerated reactor. Mixing of the aerated reactor was facilitated using Rushton-type impeller-based agitation. The pH of the reactors was monitored using a pH/salinity probe. The SGL feed solution was fed using a peristaltic pump maintaining the hydraulic retention time of the system at 4 days. The sludge purge volume of the system was kept constant resulting in a SRT of 30 days. The airlift sludge return mechanism of the clarifier facilitated sludge recirculation back to the aerated reactor at a return flow rate of 0.18L.min^"1. The AS system was initially operated for a period of 7-days under batch operating conditions and thereafter operation followed a continuous feed approach for the duration of the operating period.

2.2.4. Effluent

The industrial effluent was obtained as Stripped Gas Liquor (SGL). SGL is the condensate (gas liquor) that is generated by a coal gasification process. This condensate then undergoes a sedimentation process for the removal of tar products. Once the removal of the tar products is complete, the condensate is treated via the Phenolsolvan process for the removal of phenolics and then undergoes steam-stripping for the removal of ammonia and to facilitate a decrease in the pH of the condensate. The resultant effluent is then referred to as SGL.

The SGL feed composition, as used in these experiments, is as follows: COD 1500-2500mg/L; NH₃ 550mg/L; pH 8-8.5; Conductivity 4500μS/cm; Phenolics 200mg/l; Fluoride 250mg/L. KH₂PO₄/K₂HPO₄ and urea were added to the both the membrane bioreactor tank and activated sludge feed to achieve a molar C:N:P ratio of 100:10:1. Both the activated sludge and membrane bioreactor tank were seeded with mixed liquor suspended solids (MLSS) obtained from a waste activated sludge plant. Both reactors were seeded with a 20% (v/v) inoculum consisting of MLSS at 10 OOOmg.L^"1. This mixed culture was then acclimatized to the hydrodynamic and biological conditions by operating both the activated sludge and membrane bioreactor CSTR for a period of 7-days under batch conditions. Thereafter the AS system was operated continuously as described in section 2.2.3. and the MBR tank was operated as described in section above (2.2.2.).

2.3. Denaturing Gradient Gel Electrophoresis (DGGE)

2.3.1. Sample collection

Duplicate samples (50 ml) were collected from the feed, inoculum, membrane bioreactor tank, the membrane bioreactors and the activated sludge reactor. The 50 ml samples were centrifuged (4000xg) for 5 minutes and 35 ml of supernatant carefully discarded. The remaining 15 ml were used as the microbial community samples from which total community DNA were extracted. 2.3.2. DNA extraction and purification

The DNA extraction and purification protocol encompassed physical and chemical lysis in a high-salt extraction buffer followed by phenol/chloroform/isoamyl alcohol, chloroform/isoamyl alcohol extraction and ethanol precipitation. The method is detailed as follows:

One ml of this sample was placed in a weighed 1.5 ml microfuge tube and centrifuged at full speed for 5 minutes. The supernatant was discarded and the pellet weighed. The pellet was resuspended in 100 μl of TE buffer, lysozyme added to a final concentration of 50 μg/ml and incubated for 30 minutes at 37°C. Hot (65°C) 2X CTAB [hexadecyltrimethylammonium bromide] isolation buffer (10OmM Tris pH 8.0, 20 mM EDTA, 2% w/v CTAB, 1.4 M NaCI, 0.2 % v/v β-mercaptoethanol) [β-mercaptoethanol added immediately before use] was added and further incubated at 65°C for a further 30 minutes gently inverting the tubes 3 or 4 times (every 3 minutes). One freeze-thaw step involved 5 minutes on ice then incubation at -85°C for 15 minutes followed by addition of hot (65°C) PVP solution (final concentration of 0.5 %) and incubation 65°C for further 30 minutes inverting the tubes 3 or 4 times (every 3 minutes). This was followed by hot (65°C) buffered phenol:chloroform:isoamyl alcohol (pH 8.0) (25:24:1 ) extraction for 10 minutes at 65°C (gently inverting the tubes 3 or 4 times (every 3 minutes). The phases were separated by centrifugation and the aqueous phase re-extracted with an equal volume buffered chloroform:isoamyl alcohol (pH 8.0) (24:1 ) for 10 minutes at room temperature gently inverting the tubes 3 or 4 times. After centrifugation (13,500 rpm for 5 minutes) the DNA in the aqueous phase was precipitated with NaCI, ice cold 95 % Ethanol and incubation at -80⁰C for at least 1 hour. The precipitating DNA mixture was then immediately centrifuged at 4°C (full speed in microfuge) and the pellet washed with ice cold 70 % ethanol. The DNA was vacuum dried and reconstituted by adding 50 μl dH₂O and incubating at 65°C (1 hour).

2.3.3. DNA quantification

The extracted DNA was quantified spectrophotometrically using 260 nm values and diluted accordingly. The 260/280 nm ratios were used to determine the purity of the DNA.

2.3.4. Polymerase Chain Reaction (PCR) conditions

The eubacterial primer combination GM5F and 907R was used for PCR amplification of community DNA. The sequence of forward primer (GM5F) was 5'-GCC CGC CGC GCC CCG CGC CCG TCC CGC CGC CCC CGC CCG CCT ACG GGA GGC AGC AG-3' [the GC - clamp. is underlined] and the reverse primer (907R) 5'-CCG TCA ATT CCT TTG AGT TT-3'.

PCR amplification was performed in a Hybaid thermal cycler and the conditions for the euabacterial primers consisted of 5 min at 94°C followed by 35 cycles consisting of 30 sec at 94°C, 30 sec at 65°C and 1 min at 72°C. The final cycle was followed by a 4 minute extension at 72°C. The total reaction volume of 25-μl contained 1.0 U of Taq DNA polymerase (Super- therm Taq, AB Biolabs), 1 X Roche PCR master mix (10 mM Tris-HCI pH 8.0 , 50 mM KCI, 1.5 mM MgCI , each deoxyribonucleotides at a concentration of 200 μM), 4M MgCI [additional], 25 pmol of each primer, 100 ng of the extracted DNA, 100 ng of BSA overlaid with 25 μl of mineral oil (AB Biolabs). Amplification of PCR products of the proper size was confirmed by electrophoresis through a 1.0 % w/v agarose gel containing ethidium bromide (50 μg/ml) in TAE buffer. Electrophoresis was performed for 45 minutes at 100 V using standard electrophoretic techniques. A molecular size marker (1 kb ladder, Roche DNA molecular weight marker XIV) was used as the reference. A gel image analysis system (Syngene Gene Genius Biolmager) was used to document and analyse the gels.

2.3.5. DGGE analysis

DGGE analysis was performed using a D-Code universal mutation detection system (BioRad) and electrophoresis reagents from BioRad. PCR products (15 μl) were loaded onto a 6.0% (w/v) polyacrylamide gel cast in 1 x TAE (40 mM Tris, 20 mM acetic acid, 1 mM EDTA, pH 8.0). The polyacrylamide gels (acrylamide:bisacrylamide, 37.5:1 ) were made with denaturing gradients ranging from 30 to 60% (100% denaturant contained 7 M urea and 40% formamide). Gels were equilibrated to 60⁰C before loading the PCR products. Electrophoresis was at 60⁰C for 16.5 h at 100 V. After electrophoresis, the gels were stained for 20 min with ethidium bromide (50 μg/ml) and destained for 15 minutes in 1X TAE. A gel image analysis system (Syngene Gene Genius Biolmager) was used to document the gels. Syngene Gene Tools and Gene Directory software were used to analyse the gels.

Manual statistical analysis using data obtained from the software was also performed estimating species richness & diversity within microbial populations analysed. Since each band is likely to be derived from a single phylogenetically distinct clade/species, species richness was calculated by total number of bands on the gel. The Shannon-Weaver species diversity index H, which is computed according to the following equation:

where Pi is the portion of the total sample occupied by species / and n is the number of species, was used to estimate species richness.

In addition pair-wise similarity indices were compiled and from these dendrograms were derived that graphically demonstrated the relatedness of the various community DGGE profiles.

2.4. Fatty Acid Methyl Ester (FAME) analysis

2.4.1. Sample collection

Triplicate samples (50 ml) were collected from the feed, inoculum, MBR tank, the membrane bioreactors and the activated sludge reactor. The 50 ml samples were frozen immediately, freeze dried (Dura Dry™, FTS Systems, Inc., N.Y.), and stored at -84⁰C until extraction procedures were initiated.

2.4.2. Phospholipid extraction

All glassware used for extraction, fractionation, and analysis was washed with phosphate-free soap, air dried, and heated at 450⁰C in a muffle furnace for 4 hours prior to use. Total community phospholipid was extracted from 0.05g lyophilized biofilm and planktonic samples using a modified Bligh and Dyer procedure.

2.4.3. Phospholipid separation

The total community lipid extract was subjected to fractionation using salicic acid column chromatography to separate the glyco-, neutral, and polar lipids. The resultant polar lipid fraction was then transesterified using mild alkali methanolysis resulting in the recovery of the PLFA as methyl esters (FAMEs) in hexane.

2.4.4. Phospholipid analysis

The transesterified PLFAs (FAMEs) were separated, quantified, and identified using capillary gas chromatography (GC) with flame ionization detection. The system hardware consisted of a HP 6890 series Il chromatograph (Hewlett-Packard, Wilmington, DE, USA) fitted with a 60m SPB-1 capillary column (0.25mm I. D.; 0.25μmfilm). Methylnonadecanonate (C19:0) was used as the internal standard and FAMEs were expressed as equivalent peak responses to the internal standard. Definitive peak identification was facilitated using a HP 6890 series Il chromatograph interfaced with a HP 5973 series mass selective detector (Hewlett-Packard, Wilmington, DE, USA).

2.5. RESULTS

2.5.1. Membrane Bioreactor System

2.5.1.1. Hydraulic performance

After initial characterization of the membrane modules, the flux of the membrane bioreactor was monitored continuously for a period of ±300 days (7200 hours). After inoculation, the MBR tank system was operated initially as a batch system for 7 days and then for a further period of 13 days under continuous feed-and-bleed conditions as described above. Thereafter, the membrane modules were started up and operated using the mixed liquor in the MBR tank as the influent feed. During this 20-day start-up period prior to the membrane modules being operated, the MLSS in the MBR tank dropped from 2000mg.L^"1 to approximately 300 mg.L^"1. This showed minimal fluctuation over the full 300-day MBR tank operating period.

The main objective of the seeding reactor was the establishment of an acclimated biofilm for maintenance of a constant inoculation regime to enhance the productivity of the hydrolysis reactor. Maintenance of a high flux operating regime was therefore secondary in terms of operational requirements. Due to the nature of the influent feed, rapid flux decline (65% loss in initial flux) was observed within the first 24-hours after start-up of the membrane modules (see Figure 2.2). Over the operating period of 1700 hours, slight flux recovery was observed at days 8-12, 28-37, and 40-43, however, the average steady-state flux was 2L.nT².h^"1. This operating flux enabled the formation of a loose, easily dislodged biofilm that was removed periodically for analysis using low pressure backflushing pulses. A similar biofilm management strategy was envisaged for the coupled seeding and hydrolysis MBR configuration.

2.5.1.2. Performance of the MBR and AS systems

The initial operation of the MBR tank was as a batch process for a period of 7-days after inoculation, and then as a feed-and-bleed approach for a further 13-day period prior to circulation of the MLSS and SGL feed through the ceramic membrane modules. During this acclimation period a maximum of 25 - 35% COD removal was observed (results not shown). Thereafter, during the MBR operating period up to 180-days, an average COD removal of 45% was observed (Refer to Figure 3.4).

In contrast, the MBR membrane module-associated COD removal efficiency increased rapidly from an initial 40% to 80% over the 17-day operating period. Thereafter the removal efficiency stabilized for the following 180-day operating period with an average of 70-80% COD removal observed (refer to Figure 2.3 and Figure 2.4). Thereafter (day 197), additional COD in the form of I OOppm phenol and IOOppm 2,4-dichlorophenol was added to both the AS and MBR feed (see Figure 2.3).

The phenol and 2,4-dichlorophenol concentration was then increased to 250ppm (Day 235), 500ppm (Day 257), and IOOOppm (Day 293). A comparative evaluation of the MBR process with conventional AS system revealed that during these incremental loading periods, no significant decrease in operating efficiency in terms of COD removal was observed at concentrations of 250ppm (an accumulative concentration of 500ppm aromatic load). Thereafter, from 500 to IOOOppm addition, the average COD removal of 75% for the MBR decreased to 20% (refer to Fig. 2.4) and from 55% for the AS reactor to 20%.

As the design and control of MBR systems is usually based on the operational stability and ability of the MBR system to treat different wastewaters under varying operating conditions, the stability of the MBR system over the initial 256-day operating period (0-250ppm) is observed as an essential component of the seeding reactor. Substrate assimilation may therefore be quantitatively described as the amount of substrate utilized (r_s) per unit biomass (X) for both the synthesis of cellular material (P) as well as cellular maintenance energy (E).

2.5.2. Community profiling - DGGE

2.5.2.1. Community profile analysis using DGGE

The DNA profiles of the various MBR- and AS-associated communities after DGGE analysis are indicative of microbial community changes in the MBR tank and the conventional activated sludge reactor (AS) as determined using DGGE analysis from day 0 to day 18 (profiles not shown). Numerical analysis of these DGGE profiles is displayed in dendrogram format in Figure 2.5. The dendrogram was obtained by the neighbour joining method and topology of the tree as well as the branch lengths are informative.

Four different clusters A to D were observed. These clusters are based on track similarities based on band matching data. The topology of the dendrogram showed four major clusters labelled A to D. Samples that were taken and analysed for the first four days of the operation of the reactors were in cluster A and B. The profiles of these samples were very similar. On the other hand, clusters C and D included the samples that showed differences in the DGGE profiles that could be correlated with bacterial population structure. Branch lengths of these neighbour joining dendrograms are also informative and represent similarity between tracks - the shorter the lengths the more similar the lanes. In the first three weeks where the SGL-MBR and SGL-AS processes were analysed and compared, 12 potentially different clades or species (referred to as operational taxonomic units (OTUs)) were observed. The numerical analysis of these profiles showed graphical (dendrogram) relationships of bacterial communities that were reflected in the visual comparison of the DGGE profiles.

Further analysis of the population profile up to the point during operation when additional COD was added in the form of 100-1000ppm phenol and 2,4-dichlorophenol indicated considerable variation in the population based upon 16S and 18S rDNA analysis. As much as 23 individual operational taxonomic units (OTU's) (determined using DGGE) were observed to exist at various operating periods during the 305-day course of analysis.

Figure 2.6. (A and B) depicts examples of the 16S and 18S DGGE profiles. The 16S rDNA gel examples depicted here show complex banding patterns (melting types or operational taxonomic units).

From Figure 2.7. it is clear that there were considerable changes in bacterial OTU diversity in the feed-tank during the initiation of the experiment. The first samples for the reactor modules were analysed 41 days after the system was inoculated, thus the first results for the reactors were on day 41. The dynamics indicated by the 18S rDNA data were less variable but showed an increase during the initial stages of the experiment with subsequent continuous fluctuation over the course of the experiment. It gradually decreased over the 305 days. On the last day of the experiment (day 305), the 18S diversity was higher in the tank than in the reactors.

The Shannon-Weaver indices of MBR tanks, based on 16S rDNA data, were also compared to the reactors. It shows that bacterial diversity in the reactors was greater to that of the tanks on days 41 , 60, 105, 235 and 305. On day 82 the opposite was observed ie. the bacterial diversity in the tank was less than the diversity in the reactors. The bacterial diversity after day 198 was also lower in both the compartments of the MBR (reactors and tank) than the diversity between days 41 and day 198. After day 198 the phenol was added to the system starting with 100 ppm at day 198. At day 235 the phenol concentration was increased to 250 ppm, and sequentially thereafter from 500 ppm to I OOOppm after day 258. The decline in species diversity (both bacterial and fungal) was attributed to the addition of increasing concentrations of phenol to the system.

In the analysis of the fungal/bacterial diversity ratio (Shannon-Weaver 18S/16S, Figure 2.8.) show that this ratio was higher in the reactors than in the MBR tank only on day 60 and day 258. On the other days the ratio was higher in the MBR tank than in the reactors. This trend in fungal/bacterial ratio could be linked to operational efficiency fluctuations described earlier. Day 60 samples were associated with a decline in COD removal efficiency (Figure 2.4.). The samples for the other days were taken when the COD removal efficiency was recovering.

The changes in Shannon diversity indices could be related to changes in operational conditions of the system as well as efficiency as measured by the COD removal. The results further showed that ratio of 18S/16S indices could be more clearly linked to these observed changes.

The data presented here demonstrate that the DGGE technique is a powerful tool for resolving more than just the dynamics of the microbial community in biofilms. In this study DGGE analysis showed that microbial communities in the two compartments of the MBR system were differently affected by the duration of the experiments as well as differences in operational conditions and shock loading. The DGGE showed that the operational taxonomic units in both the bacterial and fungal components were dynamic over time but that the dynamics are also influenced by changes in operational conditions. Furthermore, species diversity indices may be useful for comparative studies such as those between various MBR and, activated sludge systems.

2.5.3.Community profiling - PLFAs

2.5.3.1. PLFA analysis of microbial community structure

PLFA data for the initial 82 days of MBR operation (refer to Figure 2.9) indicated that the major PLFA group observed consisted of the monounsaturated PLFAs (indicative of Gram-negative bacteria) with the second major grouping being the terminally branched saturated PLFAs (indicative of Gram-positive bacteria). Relative proportions of these two major groupings were also consistent within the MBR over this operating period.

In contrast, relative proportions of these major bacterial groups were less consistent within the first 82 days of operation associated with the AS reactor (see Figure 2.10.). Coupled with the DGGE data, significant insight was gained concerning the impact of COD loading stress on the major microbial population groupings. Furthermore, although the results of the PLFA analysis indicate relatively consistent proportions of both Gram negative and Gram positive groups (with Gram negative comprising the majority in both the AS and MBR systems), the data obtained using DGGE indicate that the relative diversity within these two major groupings is still consistently high. Sequence data based on individual species within the microbial consortia within the MBR and AS systems will enable the identification of specific organisms at a species level. The combination of DGGE, PLFA, and sequence data enables the identification of the organisms at a species level and therefore conclusions to be reached in terms of whether the frequency of species succession and level of structural diversity is greatest in the Gram negative fraction of the population or the Gram positive fraction. These data may then be correlated with published identifications of similar organisms in other wastewater treatment processes.

2.6. DISCUSSION AND CONCLUSIONS

Although the efficiency of the seeding MBR has been shown to be more consistent compared with the conventional AS process, it is useful to determine at what point would the decrease in productivity outweigh the overall efficiency of the MBR system in order to ascertain the stability, or lack thereof, of the associated microbial population using DGGE and PLFA analysis. The assessed process efficiency can then be correlated in terms of COD removal coupled with PLFA and 16S/18S rDNA analysis to enable correlations to be made as to any stress response within the microbial biofilm population as a result of changes in the MBR operational characteristics (for example, as increased COD loading by pollutant addition) being implemented within the system. The structural diversity succession within the system can then be correlated back to the overall performance within these systems, enabling the identification of which major microbial groupings are contributing to a lesser or greater extent to the overall process efficiency. Addition of these populations to the hydrolysis reactor can then be directly correlated with expected efficiency and process predictability.

3. DUAL-STAGE SEEDING/HYDROLYSIS MEMBRANE BIOREACTOR TREATMENT OF WASTEWATERS

3.1. INTRODUCTION

3.1.1. Reactor Configurations

The initial reactor design for the preliminary seeding MBR (SGL-treatment) was based on a feed-and-bleed approach where the operating conditions based on the Biological Oxygen Demand to Chemical Oxygen Demand (BOD₅/COD) ratios were facilitated by the bleed ratio of the seeding reactor coupled with the HRT of the solid/liquid separation membrane module. The preliminary results enabled the assessment of population shifts within microbial groups and population stability within these groups over extended operating periods to ascertain the applicability of using these populations as a continuous inoculum source for the hydrolysis reactor.

The biodegradability of a compound can be described by the BOD5/COD ratio maintained within the contact phase of the microbial population and the substrate (pollutant) of interest. An easily biodegradable compound such as citric acid has a BOD₅/COD ratio of 0.95, while a less easily biodegradable compound such as triethanolamine will have a BOD₅/COD ratio of 0.54. Compounds highly resistant to biodegradation can even have BOD₅/COD ratios as low as 0.1. By operating the seeding reactor at a low BOD₅/COD ratio i.e. a high solids retention time and low hydraulic retention time to ensure a high feed-to-microorganism ratio, the selection of microbial populations capable of degrading the highly persistent compounds will be achieved. The maintenance of the low BOD5/COD ratio will be facilitated by, firstly, adopting a high feed flow rate into the seeding reactor, thus maintaining high COD loading rates. Due to the usually slow growth rates of the bacterial population of interest (pollutant degraders) this will result in a desirable low BOD₅/COD ratio for pollutant degradation, however, the high loading rate will eventually favour an increase in the predatory population resulting in an undesirable high BOD5/COD ratio as a result of the easily degradable COD fraction being preferentially consumed. Secondly, by operating the coupled MBR at high sludge recirculation rates, the usual associated cell removal in the effluent is avoided due to the cell retention characteristics of the membrane, thus facilitating the continued maintenance of the pollutant-degrading population within the seeding reactor. Under these operating conditions, the BOD₅/COD ratio will increase gradually until at steady-state operating conditions the main component of the resident microbial community will constitute the pollutant-degrading fraction.

Utilisation of this microbial community as a continuous source of inoculum for the hydrolysis reactor which will operate at conventional high BOD₅/COD ratios i.e. a high hydraulic retention time and high solids retention time to maintain a low feed-to-microorganism ratio, will facilitate continued operation at these ratios at high loading rates but with alleviated conditions of predator-prey interactions within the present microbial consortia. Balancing the sludge bleed ratio with that of the BOD5/COD ratio within the seeding reactor will facilitate the maintenance of the correct operating BOD₅/COD ratio conditions for optimal pollutant degradation rates at significantly higher loading rates in comparison with conventional AS systems. Furthermore, due to the solids retention time (SRT) being independent of hydraulic retention time (HRT), in terms of the biofilm component, the MBR will be characterized by higher biomass concentrations (MLSS) than conventional activated sludge systems. Consequently, higher maintenance energy requirements mean that continuous biosynthesis and cell growth necessitates high oxygen concentrations within the membrane bioreactor.

3.1.2. Seeding and Hydrolysis Membrane Bioreactor Design

The seeding-hydrolysis hybrid dual-stage MBR system was designed to provide a seeding MBR configuration capable of maintaining a high solids retention time (SRT) and low hydraulic retention time (HRT), to ensure a high feed-to-microorganism ratio enabling the selection of microbial consortia capable of degrading most persistent compounds in the wastewater. Membrane-based biomass retention facilitated the utilisation of this microbial community as a continuous source of inoculum supporting the design of the hydrolysis reactor aimed at operating at conventional high BOD₅/COD ratios i.e. a high hydraulic retention time and high solids retention time to maintain a low feed-to-microorganism ratio. The piping and instrumentation diagram (see Figure 3.1 ), shows the combination seeding and hydrolysis MBR system with associated feed, seeding, hydrolysis, and permeate (or product) collection vessels. The operating configuration comprises two membrane bioreactor systems (described below) operating as biomass retention/biofilm development processes for two seeding reactor tank systems. Backflushing of the membrane-associated biofilms facilitated continuous inoculum introduction into a single hydrolysis reactor operated as described above. In this Figure, reference numeral 1 indicates the hydrolysis system in a hydrolysis tank, reference numeral 2 indicates the external circuit MBR seeding feed tank, reference numeral 3 indicates the internal submerged MBR seeding feed tank, and reference numeral 4 indicates the raw effluent feed tank.

3.1.3. Submerged-type MBR system configuration

Each membrane reactor module comprised an effective membrane area of 0.0265m². The membrane module constructs, comprising three separate modules, had a total membrane area of 0.0795m². Fluid reticulation within the MBR system is achieved using a magnetic drive centrifugal pump with a recirculation flow rate of 100Lh^"1 from the MBR tank and a membrane module influent flow rate of 100Lh^"1. At this flow rate the transmembrane pressure was greater than 1OkPa. Air sparging was implemented during backflush sampling and intermittent operation to maintain aerobic conditions and facilitate biofilm dislodging during hydrolysis tank seeding. 3.1.4. External circuit membrane modules

Each single-membrane reactor module comprised an effective membrane area of 0.0067m². The membrane module constructs comprising three in-series modules have a total membrane area of 0.02m² with each MBR assembled (comprising five constructs in series) having an effective total membrane area of 0.133m² for the membrane bioreactor (refer to Fig. 3.1 ). Fluid reticulation within the MBR system was achieved using a magnetic drive centrifugal pump with a recirculation flow rate of 150Lh^"1 from the MBR tank and a membrane module influent flow rate of 10Lh^"1. At this flow rate, the cross flow velocity was 0.035m.s^"1 and the transmembrane pressure was <2kPa.

3.2. HYDRAULIC PERFORMANCE

After inoculation, the MBR seeding tank system was operated initially as a batch system for 40- days. Thereafter, the membrane module start-up was initiated and these were operated using the mixed liquor in the seeding reactor tanks as the influent feed to initiate biofilm accumulation and microbial consortia acclimation/selection. During initial characterization of the membrane modules, the flux of the membrane bioreactor was monitored continuously to determine variation or stability in flux and the onset of biofilm growth-associated transmembrane pressure (TMP) increase.

3.3. OPERATING CONDITIONS

The membrane bioreactor tank, used in association with the tubular ceramic membrane modules, had a working volume of 6OL The MBR tank temperature varied between 25-26⁰C and did not exceed 3O⁰C for the 250-day operating period. The dissolved oxygen concentration was maintained between 4 and 5 mg.L^"1 by using diffused air sparging, with the pH of the system ranging between 7.5-8.7 over the same period. The membrane modules were operated at a transmembrane pressure greater than 10 kPa and the system was initially operated for a period of 40-days under batch operating conditions. Thereafter, the membrane module component of the membrane bioreactor system was started up. The feed approach entailed the removal of 30L.day^"1 mixed liquor thereby maintaining a hydraulic retention time in the tank of 3- days. The removed mixed liquor was replaced daily with 3OL of fresh SGL make-up feed corrected to a C:N:P ratio of 100:10:1 using KH₂PO₄ZK₂HPO₄ and urea.

3.4. RESULTS

3.4.1. Membrane Bioreactor System

Figures 3.4 and 3.5 indicate the initial (30 days) COD removal performance of the external circuit membrane bioreactor modules and the simulated submerged membrane bioreactor modules respectively after the initial batch operation start-up of the seeding tanks (40 days). Significantly better performance was observed in the external circuit modules and this was attributed to longer residence times of the wastewater within the membrane-associated biofilms due to the lower flux performance of these membranes.

3.4.2. Seeding/Hydrolysis Hybrid Bioreactor System

Initially operated as a batch process for approximately 40-days, the seeding reactor performance peaked at 40% COD removal efficiency with a gradual performance decline to approximately 10-15% COD removal efficiency over a three month operating period. This was attributed to the removal of biomass from the system (retained as biofilm within the membrane bioreactor system). This is typically the same scenario associated with many AS processes where, operated as continuous systems, COD removal efficiencies (for industrial effluents) are typically low and recovery periods are lengthy when no re-inoculation or re-seeding procedures are implemented to compensate for washout phenomena associated with slow growing microbial community fractions. From an economic perspective, these procedures are more costly in terms of the associated plant downtime than the actual reseeding process itself.

In contrast, the seeding reactor-associated MBR system consistently operated during this time within a 60-80% efficiency range similar to the performance reported in the seeding reactor study summarised hereinbefore. Figure 3.6 is a summarised representation of the initial performance of the hybrid dual-stage seeding/MBR/hydrolysis process (first three months of operation).

After approximately 80 days of operation (40 days as seeding reactor batch operation and 40 days with continuous operation of the membrane bioreactor modules), the hydrolysis reactor startup was initiated. No acclimation period was initiated (i.e. operation in batch mode) and the initial seeding of the hydrolysis bioreactor was started immediately using backflush transfer of the membrane module-associated biofilm. COD removal efficiency reached 60% within 10 days. In contrast, the seeding reactor reached a maximum COD removal efficiency of 40% after 40 days of operation in batch mode.

To ascertain the robustness and recoverability of the hybrid MBR system, an aromatic organic pollutant (phenol) and halogenated aromatic (2,4-dichlorophenol), in combination, were added to the system to simulate high shock loading. In this scenario the simulation facilitated observation of the effect of a high organic load perturbation (as rapidly increased total COD) and as a rapid pollutant load increase. Between days 217 and 238, both MBR systems and the hydrolysis reactor were supplemented with increasing concentrations of phenol and 2,4- dichlorophenol. The effect on total COD concentration increase at the different concentrations of supplemented phenols is indicated in Table 3.1 below.

Table 3.1. The effect of aromatic compound addition on total COD concentration (range average indicated)

Following pollutant addition, significant decreases in the operating efficiency (in terms of COD removal) were observed for all systems (see Figure 3.8). The percentage COD removal associated with the MBR tank initially ranged between 20% and 40% - this decreased to 0% with the shock load simulation. Recovery of this system once the pollutant was slowly removed was the slowest when compared with both the hydrolysis tank and the MBR modules. This was expected and confirmed observations of AS systems where washout after shock loading is a common problem. Membrane modules operating efficiency was also severely affected, however rapid recovery was observed once the pollutant was removed from the system. Rapid recovery was also observed within the hydrolysis system. The maximum recovery that this system could potentially achieve was not realised as operations were discontinued after approximately 250 days of operation. At that point in the experiment, the overall performance trend with the hydrolysis reactor was still increasing, as opposed to the MBR tank system which levelled off and never recovered significantly or reached equivalent performance levels observed within the first 40-days of operation.

3.5. DISCUSSION AND CONCLUSIONS

A long-term (250-days) operating process with imposed shock loading treatment using a hybrid system for industrial wastewater treatment comprising a seeding and hydrolysis reactor coupled to intermediate membrane module operation was tested. Continuous inoculum development in the seeding MBR and intermittent continuous transfer of that acclimated retained biofilm to the hydrolysis reactor facilitated significant improvements in the enhancement of population adaptability and performance efficiency as well as drastically decreasing the usual adaptation time necessary (75% faster), associated with the majority of conventional wastewater treatment processes. A 75% improvement in microbial population adaptability translated to accelerated efficiency improvement in terms of COD removal capabilities. Furthermore, and more importantly, the coupling of the hybrid system enabled rapid recovery of a treatment process when perturbed by a shock load. In contrast, systems such as conventional activated sludge processes are not capable of microbial population retention and are thus significantly more susceptible to washout risks associated with either variations in hydraulic load (such as summer/winter influent volume variations) or with shock loading (as rapid COD load elevation or the introduction pollutant compounds).

In this example of the invention, the application of solid-liquid retention membrane bioreactors is shown herein to be a highly efficient system for the treatment of high-strength industrial effluents containing recalcitrant pollutants. In comparison to activated sludge systems, the long- term operation of this membrane bioreactor process treating high-strength effluents was characterised by more stable microbial populations significantly less susceptible to deleterious shifts in the community dynamics resulting in enhanced process efficiency due to less process variability. In addition, the dual-stage hybrid membrane bioreactor enabled the intermittent transfer of acclimated retained biofilm developed in a seeding reactor to a hydrolysis reactor. This facilitated significant improvements in the enhancement of microbial population adaptability and performance efficiency as well as drastically decreasing the usual acclimation time necessary by 75% when compared with activated sludge wastewater treatment processes.

Results obtained confirmed that the MBR system was more stable in terms of operational efficiency and recovery when subjected to various loading rate challenges than was a conventional activated sludge process. The advantages of modularity associated with commercial membrane-based systems are also applicable to the dual-stage hybrid membrane bioreactor included herein. The significant advantage is therefore that this process design can be retrofitted to existing infrastructure with relatively low capital expenditure. Because the reticulation and process flow control of this system is relatively simple insofar as hardware requirements are concerned, automated control of the process would be significantly less complex and relatively inexpensive compared with the current systems presently considered when new activated sludge processes and associated infrastructure are designed.

Coupled with the advantages of this modular, small footprint, easily retro-fitted system, the potential process intensification facilitated by this system design in the treatment of a variety of industrial, domestic, and agri-sector origin effluents is clear. Significant improvements to the current wastewater treatment options available with minimal infrastructure changes can potentially be implemented using this process design. The significant improvements in operational efficiency of the dual-stage hybrid membrane bioreactor process, coupled with the improved adaptation time, recovery and overall robustness, signify that this process has a positive impact on the overall process efficiency of existing activated sludge processes known to the Applicant and other conventional wastewater treatment processes. The portability of modular units is also advantageous and this advantage is exploitable especially if several effluent source treatment facilities are incorporated.

The Applicant is of the opinion that the integrated dual-stage MBR process of the invention provides a unique operations strategy employing membrane bioreactors for the treatment of wastewaters of industrial origin. The process facilitates a continuous development and acclimation design strategy for generating groups or consortia of microorganisms capable of degrading industrial, domestic, and agri-sector origin wastewaters. These adapted consortia are then harvested to be used in the continuous operation of hydrolysis reactors. The hydrolysis reactors are operated under similar conditions to conventional wastewater treatment tank facilities, however, the continuous addition of adapted microbial populations developed within the so-called seeding reactor configurationfacilitates, firstly, a significant decrease in adaptation periods associated with conventional treatment strategies and, secondly, an inherent robustness facilitated by obviating the requirement for adaptation within the hydrolysis reactor configuration.

The invention provides improved methods: to design and construct the seeding and hydrolysis integrated membrane bioreactors incorporating an appropriate reticulation system and operating configuration; to analyse the performance of the membrane bioreactor (MBR) seeding reactors for treating industrial, domestic, and agri-sector origin wastewaters whilst generating stable, adapted microbial consortia; to evaluate the performance of the MBRs via comparison with conventional activated sludge (AS) bioreactors and/or other particulate biofilm bioreactor types; to monitor population changes within the MBR and AS systems using molecular techniques (phospholipid fatty acid (PLFA) analysis for identifying the relative proportions of the major bacterial groups coupled with ribosomal DNA 'fingerprinting' (16S/18S rDNA) which gives insight into any variation at the specific microbial strain level) for determining the microbial population stability of the seeding reactors; to incorporate the seeding MBR process into the overall dual-stage hybrid MBR process and evaluate the process efficiency of the system.

In summary, the initial studies described herein involved the design and construction of the seeding MBR for the treatment of Stripped Gas Liquor (SGL) industrial effluent (COD of ± 2000mg.L^"1). To facilitate evaluating the performance of the MBR system in terms of process efficiency against a familiar benchmark, a conventional AS system comprising an aerated CSTR-type vessel coupled to an in-series clarifier was operated under similar hydraulic conditions. The seeding MBR was operated for approximately 300 days with an average COD removal of between 70-80% maintained after an initial 17 day acclimation period. In contrast, the AS system was operated for a period of 300 days with a maximum COD removal of only 60% observed. Furthermore, fluctuations in COD removal efficiency were minimal in the MBR compared with the AS process. This indicated greater stability of the MBR system in terms of operational performance. The combination of PLFA and 16S/18S rDNA analysis confirmed this by revealing that although the proportions of the major microbial groups during the 300 day operating period were less variable (i.e. more stable) in the MBR than the AS system, 16S rDNA analysis revealed that within those major groupings, greater variation at the bacterial strain level was observed in terms of relative strain abundance. Thus, the MBR generated and maintained a highly stable microbial group structure (with the major group being Gram negative) within the biofilm population that was sufficiently dynamic at the strain level to adapt at an accelerated rate to compensate for hydraulic load variations as well as to simulated pollutant shock loading.

Membrane-based biomass retention facilitated the utilisation of this microbial community as a continuous source of inoculum supporting the design of the hydrolysis reactor aimed at operating at conventional high BOD₅/COD ratios i.e. a high hydraulic retention time and high solids retention time to maintain a low feed-to-microorganism ratio.

In terms of long term stability, the operations data obtained from the hybrid dual-stage system over a period of 250 days indicated that whereas a conventional system represented by the seeding bioreactor tank approached a maximum of 40% COD removal, the hydrolysis reactor configuration, supplemented on a continuous basis with acclimated biofilm developed and retained by the seeding MBR, showed a start-up time of less than 10 days to achieve 55-60% COD removal efficiency. Compared with 40 days acclimation for the AS system, this translated into a 75% improvement in acclimation efficiency. This substantiated the initial proposal hypothesis describing the nature and effectiveness of the dual-stage approach to industrial, domestic, and agri-sector origin wastewater treatment using the MBR configuration in order to enhance the performance and increase the long-term adaptability and stability of the developed and retained microbial populations within the system.

In this example of the invention, the application of solid-liquid retention membrane bioreactors was shown to be a highly efficient system for the treatment of high-strength industrial effluents containing recalcitrant pollutants. In comparison to activated sludge systems, the long-term operation of this membrane bioreactor process treating high-strength effluents was characterised by more stable microbial populations significantly less susceptible to deleterious shifts in the community dynamics resulting in enhanced process efficiency due to less process variability.

The dual-stage hybrid membrane bioreactor enabled the intermittent transfer of acclimated retained biofilm developed in a seeding reactor to a hydrolysis reactor. This facilitated significant improvements in the enhancement of microbial population adaptability and performance efficiency as well as drastically decreasing the usual acclimation time necessary by 75% when compared with activated sludge wastewater treatment processes.

Claims

1. A wastewater treatment process for removing undesired impurities from an effluent feed, the process including the steps of: providing a discrete acclimation bioreactor system for developing a desired acclimated microbial population inoculum, the system being in selectively interruptible fluid flow connection with a discrete effluent hydrolysis system, such that the flow of desired microorganisms from the acclimation bioreactor system to the effluent hydrolysis system is regulated independently from the effluent feed flow through the hydrolysis system via an integrated process-controlled intermediate membrane bioreactor system; and providing, into the hydrolysis system, a regulated supply of acclimated inoculum from the acclimation bioreactor system to maintain sufficient levels of a desired microbial population to effect COD removal of between 50% and 100% from the supply of effluent flow from which undesired impurities are to be removed within the hydrolysis system.

2. A wastewater treatment process as claimed in claim 1 , wherein the acclimation bioreactor system is a solid-liquid retention membrane bioreactor and is fed with the effluent to seed and acclimatize the inoculum to the particular effluent to be treated, prior to release of inoculum into the hydrolysis system.

3. A wastewater treatment process as claimed in claim 1 or claim 2, wherein the bioreactor system is seeded with inocula consisting of MLSS at between 1000 and 20 000 mg.L^"1, dependent on the effluent source and system volume.

4. A wastewater treatment process as claimed in claim 3, wherein the bioreactor system is seeded with inocula consisting of MLSS at about 10 000 mg.L^"1.

5. A wastewater treatment process as claimed in any one of claims 1 to 4, wherein a feed from the membrane bioreactor system to the hydrolysis system is intermittently opened to seed fresh microorganisms from the filtration area, as required by effluent feed characteristics.

6. A wastewater treatment process as claimed in claim 5, wherein the effluent feed is provided to the acclimation bioreactor system at a rate of 1 % v/v/h to 100% v/v/h , dependant on the effluent source and system volume.

7. A wastewater treatment process as claimed in claim 5 or claim 6, wherein the effluent feed is provided to the hydrolysis system at a rate of 1% v/v/h to 100% v/v/h , dependant on the effluent source and system volume.

8. A wastewater treatment process as claimed in any one of claims 1 to 7, wherein a mixed liquor from the bioreactor acclimation system is transferred continuously or intermittently to provide fresh microorganisms to the hydrolysis system, as dictated by effluent feed characteristics and conversion rates.

9. A wastewater treatment process as claimed in any one of claims 1 to 8, wherein the feed from the acclimation bioreactor system is trickle-fed as a mixed liquor into the hydrolysis system on a continuous or intermittent basis.

10. A wastewater treatment process as claimed in claim 9, wherein the feed from the bioreactor is trickled into the hydrolysis system at a rate of 1 % v/v to 20% v/v, dependant on the effluent source and system volume..

11. A wastewater treatment process as claimed in claim 10, wherein the feed from the bioreactor is trickled into the hydrolysis system at a rate of about 5% v/v.

12. A wastewater treatment process as claimed in any one of claims 1 to 11 , wherein the microorganisms are cultured in the acclimation bioreactor system at a concentration of 0.01 g.L^{" 1} to 20 g.L^"1 but typically dependant on the effluent source and system volume.

13. A wastewater treatment process as claimed in claim 12, wherein the microorganisms are cultured in the acclimation bioreactor system at a concentration of about 10 000 mg.L^"1.

14. A wastewater treatment process as claimed in any one of claims 1 to 13, wherein the microorganisms are held in the hydrolysis system at a concentration of about 10 mg.L^"1 to 20 000 mg.L^"1.

15. A wastewater treatment process as claimed in claim 14, wherein the microorganisms are held in the hydrolysis system at a concentration of about 10 000 mg.L^"1.

16. A wastewater treatment process as claimed in any one of claims 1 to 15, which includes the step of monitoring the composition of the microorganism population in the bioreactor system at set intervals of 24 hours to 20 days using molecular techniques.

17. A wastewater treatment process as claimed in claim 16, which includes the step of obtaining a sample of the microorganism population from the bioreactor system and subjecting said samples to DNA analysis, lipid analysis, and/or protein analyses.

18. A wastewater treatment process as claimed in claim 17, wherein the analyses include phospholipid fatty acid analysis (PLFA) for identifying the relative proportions of the major microbial groups.

19. A wastewater treatment process as claimed in claim 17, wherein the analyses include ribosomal DNA fingerprinting using 16S/18S rDNA for indicting variation at the specific microbial strain level.

20. A wastewater treatment process as claimed in claim 17, wherein the analyses include techniques selected from the group consisting of: community DNA hybridization, %G+C profiling, restriction digestion and sequence comparisons, amplified ribosomal DNA restriction analysis (ARDRA), terminal restriction fragment length polymorphism (T-RFLP), ribosomal intergenic spacer analysis (RISA), and various combinations of these techniques.

21. A wastewater treatment process as claimed in claim 17, wherein the analyses include electrophoretic separation based on melting (TGGE) or denaturing (DGGE) behaviour.

22. A wastewater treatment process as claimed in any one of claims 1 to 21 , which includes the step of optimizing operation of the acclimation bioreactor system and hydrolysis system in terms of residence time, cross-flow velocity, TMP, and loading rates, where optimal microbial populations associated with the greatest productivity and process efficiency for a given effluent are maintained.

23. A wastewater treatment process as claimed in any one of claims 1 to 22, wherein the bioreactor is provided with at least one ceramic or polymeric membrane type.

24. A wastewater treatment process as claimed in claim 23, wherein the or each membrane provided has an average pore size of between 0.01 μm and 5.0 μm.

25. A wastewater process as claimed in claim 24, wherein the or each membrane provided has a narrow pore size distribution of about ±0.01 μm.

26. A wastewater treatment process as claimed in any one of claims 23 to 25, wherein the or each membrane provided is used as a substratum for the seeding inoculum retention.

27. A wastewater treatment process as claimed in claim 26, wherein the specific surface area of the or each membrane provided is system volume-dependant and calculated based on maintaining a permeate flux ranging between 1 L.m^"2.h^"1 and 50 Lm^"2.h^"1.

28. A wastewater treatment process, wherein the or each membrane bioreactor is operable with the mixed liquor contacting the or each membrane surface internally or externally.

29. A wastewater treatment process as claimed in any one of claims 23 to 28, wherein the or each membrane is a tubular, hollow fibre, or flat sheet membrane.

30. A wastewater treatment process as claimed in claim 1 , substantially as herein described and illustrated.