Reactor design for minimizing product inhibition during enzymatic lignocellulose hydrolysis: II. Quantification of inhibition and suitability of membrane reactors

Product inhibition of cellulolytic enzymes affects the efficiency of the biocatalytic conversion of lignocellulosic biomass to ethanol and other valuable products. New strategies that focus on reactor designs encompassing product removal, notably glucose removal, during enzymatic cellulose conversion are required for alleviation of glucose product inhibition. Supported by numerous calculations this review assesses the quantitative aspects of glucose product inhibition on enzyme-catalyzed cellulose degradation rates. The significance of glucose product inhibition on dimensioning of different ideal reactor types, i.e. batch, continuous stirred, and plug-flow, is illustrated quantitatively by modeling different extents of cellulose conversion at different reaction conditions. The main operational challenges of membrane reactors for lignocellulose conversion are highlighted. Key membrane reactor features, including system set-up, dilution rate, glucose output profile, and the problem of cellobiose are examined to illustrate the quantitative significance of the glucose product inhibition and the total glucose concentration on the cellulolytic conversion rate. Comprehensive overviews of the available literature data for glucose removal by membranes and for cellulose enzyme stability in membrane reactors are given. The treatise clearly shows that membrane reactors allowing continuous, complete, glucose removal during enzymatic cellulose hydrolysis, can provide for both higher cellulose hydrolysis rates and higher enzyme usage efficiency (kg_product/kg_enzyme). Current membrane reactor designs are however not feasible for large scale operations. The report emphasizes that the industrial realization of cellulosic ethanol requires more focus on the operational feasibility within the different hydrolysis reactor designs, notably for membrane reactors, to achieve efficient enzyme-catalyzed cellulose degradation.     

Synthesis of isomaltooligosaccharides and oligodextrans in a recycle membrane bioreactor by the combined use of dextransucrase and dextranase

A recycle ultrafiltration membrane reactor was used to develop a continuous synthesis process for the production of isomaltooligosaccharides (IMO) from sucrose, using the enzymes dextransucrase and dextranase. A variety of membranes were tested and the parameters affecting reactor stability, productivity, and product molecular weight distribution were investigated. Enzyme inactivation in the reactor was reduced with the use of a non-ionic surfactant but its use had severe adverse effects on the membrane pore size and porosity. During continuous isomaltooligosaccharide synthesis, dextransucrase inactivation was shown to occur as a result of the dextranase activity and it was dependent mainly on the substrate availability in the reactor and the hydrolytic activity of dextranase. Substrate and dextranase concentrations (50-200 mg/mL(-1) and 10-30 U/mL(-1), respectively) affected permeate fluxes, reactor productivity, and product average molecular weight. The oligodextrans and isomaltooligosaccharides formed had molecular weights lower than in batch synthesis reactions but they largely consisted of oligosaccharides with a degree of polymerization (DP) greater than 5, depending on the synthesis conditions. No significant rejection of the sugars formed was shown by the membranes and permeate flux was dependent on tangential flow velocity.     

Hydrolysis of cellulose in a cellulase-bead fluidized bed reactor

Cellulase was immobilized in a collagen fibril matrix, and no leakage of cellulase from the collagen fibril matrix was observed. The immobilized cellulase was more stable than the native cellulase. The substrate cellulose was hydrolyzed quantitatively with immobilized cellulase. The final reaction product was identified as glucose. Immobilized cellulase was used in a fluidized bed reactor where the pressure drop of the fluidized bed reactor was low and constant. Cellulose was hydrolyzed to glucose by the cellulase-bead fluidized bed reactor. The minimum flow velocity (U_mf) was 0.5 cm/sec and the optimum flow velocity of the cellulose hydrolysis was 1 cm/sec.     

Enzymatic hydrolysis of sodium-hydroxide-pretreated sallow in an ultrafiltration membrane reactor

An ultrafiltration membrane reactor was used to investigate the recovery of biocatalysts during enzymatic hydrolysis of pretreated sallow. Product inhibition could be eliminated by continuous removal of products through the ultrafiltration membrane, thus retaining the macromolecular substrate and enzymes. In this way, the degree of conversion was improved from 40% in a batch hydrolysis to 95% (within 20 h), and the initial hydrolysis rate was increased up to seven times. The recovery studies were focused on mechanical deactivation and irreversible adsorption on to the nonconvertible fraction of the substrate. Cellulase deactivation during mechanical agitation was not significant, and the loss of activity was attributed mainly to strong adsorption of the enzymes onto undigested material. This process was studied in semicontinuous hydrolyses, where fresh substrate was added intermittently. The amount of reducing sugars produced in this experiment was 25.7 g/g enzyme, compared to 4.7 g/g enzyme in a batch hydrolysis.     

Partial acid hydrolysis of cellulosic materials as a pretreatment for enzymatic hydrolysis

Partial acid hydrolysis was studied as a per treatment to enhance enzymatic hydrolysis, such a pretreatment was carried out in a continuous flow reactor on oak corn Stover, newsprint, and Solka Floc at temperatures ranging from 160 to 220°C, acid concentration ranging from 0 to 1.2%, and a fixed treatment time of 0.22 min. The resulting slurries and solids were than hydrolyzed with Trichoderma ressei QM 9414 cellulase at 50°C for 48 hr. For all substrates except Solka Floc, increased glucose yields were achieved during enzymatic hydrolysis of the pretreated materials as compared to hydrolysis of the original substrate. In several cases, after pretreatment, 100° of the potential glucose content of the substrate was converted to glucose after 24hr of enzymatic hydrolysis. It is felt that the increased glucose yields achieved after this pretreatment are due to acid's removal of hemicellulose, reduced degree of polymerization, and possibly due to a change in the crystal structure of the cellulose.     

Enzymatic saccharification of sugar cane bagasse by continuous xylanase and cellulase production from cellulomonas flavigena PR‐22

Cellulase (CMCase) and xylanase enzyme production and saccharification of sugar cane bagasse were coupled into two stages and named enzyme production and sugar cane bagasse saccharification. The performance of Cellulomonas flavigena (Cf) PR‐22 cultured in a bubble column reactor (BCR) was compared to that in a stirred tank reactor (STR). Cells cultured in the BCR presented higher yields and productivity of both CMCase and xylanase activities than those grown in the STR configuration. A continuous culture with Cf PR‐22 was run in the BCR using 1% alkali‐pretreated sugar cane bagasse and mineral media, at dilution rates ranging from 0.04 to 0.22 1/h. The highest enzymatic productivity values were found at 0.08 1/h with 1846.4 ± 126.4 and 101.6 ± 5.6 U/L·h for xylanase and CMCase, respectively. Effluent from the BCR in steady state was transferred to an enzymatic reactor operated in fed‐batch mode with an initial load of 75 g of pretreated sugar cane bagasse; saccharification was then performed in an STR at 55°C and 300 rpm for 90 h. The constant addition of fresh enzyme as well as the increase in time of contact with the substrate increased the total soluble sugar concentration 83% compared to the value obtained in a batch enzymatic reactor. This advantageous strategy may be used for industrial enzyme pretreatment and saccharification of lignocellulosic wastes to be used in bioethanol and chemicals production from lignocellulose. © 2016 American Institute of Chemical Engineers Biotechnol. Prog., 32:321–326, 2016     

Adherence of bacteria, yeast, blood cells, and latex spheres to large-porosity membrane filters.

Strong adherence of bacteria, yeast, erythrocytes, leukocytes, platelets, spores, and polystyrene spheres to membrane filter materials was noted during filtration through membranes with pore size diameters much larger than the particles themselves. Quantitative recovery on the membrane filters of these particles from low-concentration suspensions was achieved during gravity- or vacuum-assisted filtration through membranes with pore diameters as much as 30 times that of the filtered particles. Mechanical sieving was not responsible. The phenomenon was judged to be electrostatic. It could be partially blocked by pretreating the filter with a nonionic surfactant (Tween 20), and elution of adherent particles was achieved with 0.05% Tween 20. Gram-positive cocci were removed from suspension more efficiently than gram-negative rods. The commonly used cellulose membranes adsorbed more bacteria, blood cells, and other particles than did polycarbonate filters. Of lesser adsorptive capacity were vinyl acetate, nylon, acrylic, and Teflon membranes. Backwashing with saline, serum, 6% NaCl, dextran solutions, or phosphate buffers of varying molality and pH removed only a fraction of adherent particles. Tween 20 (0.05%) eluted up to 45% of adherent particles in a single back-filtration. Selected filters quantitatively removed the particles tested, which then could be washed and subjected to reagents for a variety of purposes. It is important to anticipate the removal of particles during membrane filtration, since it is not a simple mechanical event.     

Na, k, and nonspecific solute requirements for induction and function of galactose active transport in an antarctic psychrophilic marine bacterium

Strong adherence of bacteria, yeast, erythrocytes, leukocytes, platelets, spores, and polystyrene spheres to membrane filter materials was noted during filtration through membranes with pore size diameters much larger than the particles themselves. Quantitative recovery on the membrane filters of these particles from low-concentration suspensions was achieved during gravity- or vacuum-assisted filtration through membranes with pore diameters as much as 30 times that of the filtered particles. Mechanical sieving was not responsible. The phenomenon was judged to be electrostatic. It could be partially blocked by pretreating the filter with a nonionic surfactant (Tween 20), and elution of adherent particles was achieved with 0.05% Tween 20. Gram-positive cocci were removed from suspension more efficiently than gram-negative rods. The commonly used cellulose membranes adsorbed more bacteria, blood cells, and other particles than did polycarbonate filters. Of lesser adsorptive capacity were vinyl acetate, nylon, acrylic, and Teflon membranes. Backwashing with saline, serum, 6% NaCl, dextran solutions, or phosphate buffers of varying molality and pH removed only a fraction of adherent particles. Tween 20 (0.05%) eluted up to 45% of adherent particles in a single back-filtration. Selected filters quantitatively removed the particles tested, which then could be washed and subjected to reagents for a variety of purposes. It is important to anticipate the removal of particles during membrane filtration, since it is not a simple mechanical event.     

Continuous ethanol production by immobilized yeast reactor coupled with membrane pervaporation unit

A system comprised of an immobilized yeast reactor producing ethanol, with a membrane pervaporation module for continuously removing and concentrating the produced ethanol, was developed. The combined system consisted of two integrated circulation loops: In one the sugar-containing medium is circulated through the membrane pervaporation module. The two loops were interconnected in a way allowing for separate parameter optimization (e.g., flow rate, temperature, pH) for each loop.The fermentation unit was 2.0 L bioreactor with five equal segments, packed with 5-mm beads of immobilized yeasts. The bead matrix was a crosslinked polyacrylamide hydrazide gel coated with calcium alginate. The fast circulation loop of the bioreactor allowed for efficient liberation of CO(2) at the top of the immobilized yeast reactor. Continuous operation of the uncoupled reactor for over 50 days with inflowing defined medium or dilute molasses at a residence time of 1.25 h yielded ethanol at a rate of about 10 g/L h.The pervaporation unit was constructed from four 60-cm-long tubular membranes of silicone composite on a polysulfone support. The output from the fermentor was circulated through the inside of the tubes of a unit with a total surface area of 800 cm(2), having an average flux of 150 mL/h, and selectivities to ethanol vs. water up to 7. A vacuum of 30 mb was applied to the outside of the tubes, removing 20-30 g of ethanol per hour, which was collected in condensors. The continuous removal of ethanol, avoiding inhibition of the fermentation process, resulted in an improved productivity and allowed the use of high sugar concentrations (40% wt/vol) offering the potential of a compact system with reduced stillage.The combined system of ethanol production and removal enabled an operative steady state at which the liquid volume of the system, and the concentrations of ethanol within the reactor ( 4% wt/vol), as well as within the flux crossing the pervaporation membrane (17%-20% wt/vol) were kept constant. At the steady state, a 40% wt/vol sugar solution could be continuously added to the fermentor when 12%-20% wt/vol clear ethanol solution was continuously removed by the pervaporation unit. Membrane fouling was reversed by short washing steps, and continuous step operation was maintained by working with two different modules that were interchanged. In this manner, long term continuous operation (over 40 days) was achieved with a productivity of 20-30 g/L h, representing over a twofold increase relative to the continuously operated reactor uncoupled from the membrane and a fivefold increase in comparison with the value obtained fro a corresponding batch fermentation.     

Study of microencapsulated urease in a continuous feed,stirred tank reactor

Urease, (urea amidohydrolase, EC 3.5.1.5) co-encapsulated with haemoglobin in cellulose nitrate membranes was found to exhibit apparent Michaelis-Menten kinetics; however, a steadily increasing apparent Michaelis-Menten constant over the lifetime of the preparation was observed. The activity of the enzyme in a continuous feed stirred tank reactor (CSTR) was investigated and correlated with a mathematical model derived from basic Michaelis-Menten kinetics. Plots relating substrate conversion to feed substrate concentration and tank reactor capacity were constructed and found to be accurate to less than 15% error under the experimental conditions studied.     

Separation of proteases from yellowfin tuna spleen by ultrafiltration

Separation of protease, trypsin and chymotrypsin from yellowfin tuna spleen extract by ultrafiltration (UF) using regenerated cellulose membranes with molecular weight cut off (MWCO) 30 and 100 kDa was studied. The 100 kDa membrane had a higher transmission of enzymes than that of the 30 kDa membrane. The enzyme transmission varied from 0.01 to 0.18 and from 0.6 to 0.8 for the 30 kDa membrane and 100 kDa membrane, respectively. The protein transmission was about 0.8 for both membranes. Increasing cross-flow rate and transmembrane pressure (TMP) increased permeate flux. The limiting fluxes at cross-flow rate 120, 240 and 360 L/h for the 30 kDa membrane were 17.3, 43.9 and 54.7 L/m2h, respectively and the limiting fluxes at the same flow rate for 100 kDa membrane were 34.1, 51.1 and 68.4 L/m2h, respectively. The separation of these proteases was achieved using the 30 kDa membrane. The purities of proteases were increased more than ten times at TMP 1.5 bar and cross-flow rate 360 L/h by diafiltration using 30 kDa membrane.     

Performance of a membrane reactor for cellobiose hydrolysis

A continuous enzymatic hollow fiber reactor (HFR), obtained by immobilizing cellobiose active cells into the shell side of hollow-fiber modules, was studied. The HFR yield was monitored by glucose analysis resulting from hydrolysis of cellobiose. The residence time of substrate in the bioreactor to obtain convenient hydrolysis yields was calculated from tests carried out by varying the reactor dilution rate in the range 0.001-0.004 L/min. The glucose yield was measured for 300 h (continuous substrate flux). The yield decreased from 40 to 15%. This decrease was due to the loss of specific activity in the operating conditions and to the pressure drop increase from 0.2 to 1.7 atm. The pressure drop increase is in turn dependent on the cell loading (0.2-2.1 g dry cell) and the substrate flux.     

One-pot bioethanol production from cellulose by co-culture of Acremonium cellulolyticus and Saccharomyces cerevisiae

ABSTRACT: BACKGROUND: While the ethanol production from biomass by consolidated bioprocess (CBP) is considered to be the most ideal process, simultaneous saccharification and fermentation (SSF) is the most appropriate strategy in practice. In this study, one-pot bioethanol production, including cellulase production, saccharification of cellulose, and ethanol production, was investigated for the conversion of biomass to biofuel by co-culture of two different microorganisms such as a hyper cellulase producer, Acremonium cellulolyticus C-1 and an ethanol producer Saccharomyces cerevisiae. Furthermore, the operational conditions of the one-pot process were evaluated for maximizing ethanol concentration from cellulose in a single reactor. RESULTS: Ethanol production from cellulose was carried out in one-pot bioethanol production process. A. cellulolyticus C-1 and S. cerevisiae were co-cultured in a single reactor. Cellulase producing-medium supplemented with 2.5 g/l of yeast extract was used for productions of both cellulase and ethanol. Cellulase production was achieved by A. cellulolyticus C-1 using Solka-Floc (SF) as a cellulase-inducing substrate. Subsequently, ethanol was produced with addition of both 10%(v/v) of S. cerevisiae inoculum and SF at the culture time of 60 h. Dissolved oxygen levels were adjusted at higher than 20% during cellulase producing phase and at lower than 10% during ethanol producing phase. Cellulase activity remained 8--12 FPU/ml throughout the one-pot process. When 50--300 g SF/l was used in 500 ml Erlenmeyer flask scale, the ethanol concentration and yield based on initial SF were as 8.7--46.3 g/l and 0.15--0.18 (g ethanol/g SF), respectively. In 3-l fermentor with 50--300 g SF/l, the ethanol concentration and yield were 9.5--35.1 g/l with their yields of 0.12--0.19 (g/g) respectively, demonstrating that the one-pot bioethanol production is a reproducible process in a scale-up bioconversion of cellulose to ethanol. CONCLUSION: A. cellulolyticus cells produce cellulase using SF. Subsequently, the produced cellulase saccharifies the SF, and then liberated reducing sugars are converted to ethanol by S. cerevisiae. These reactions were carried out in the one-pot process with two different microorganisms in a single reactor, which does require neither an addition of extraneous cellulase nor any pretreatment of cellulose. Collectively, the one-pot bioethanol production process with two different microorganisms could be an alternative strategy for a practical bioethanol production using biomass.     

Factors Affecting the Activity of Cellulases Isolated from the Rumen Digesta of Sheep

Sodium phosphate buffer was used to extract cellulases from the plant solids fraction of rumen contents. The mixed cellulase preparation had maximal activity at pH 6.9 and 49°C. The V_max and the apparent K_m for wheaten hay cellulose were 19.8 glucose units/min and 6.35 mg/ml, respectively, and for microcrystalline cellulose (Sigmacell) at the same enzyme concentration, they were 33 glucose units/min and 27.5 mg/ml, respectively. For these assays a glucose unit was defined as nanomoles of glucose plus twice the nanomoles of cellobiose. Consideration of thermodynamic and kinetic data suggested that the hydrolysis of a relatively labile arabino-xylan comprising 3% of the wheaten hay cellulose was dependent on prior removal of the protecting β-1,4-glucose chains at the outer surface of the cellulose preparation. Sequential removal of structural polysaccharides from the plant cell wall rendered the latter more susceptible to cellulase activity. Cellulase activity was stimulated by increasing the concentration of phosphate from 5 to 50 mM. The stimulation was magnified in the presence of cell-free rumen fluid. Cellulase activity was not stimulated by calcium, magnesium, iron, zinc, manganese, copper, or cobalt ions and was unaffected by the chelators ethylenediaminetetraacetic acid and ethyleneglycol-bis (β-aminoethyl ether)-N,N′-tetraacetic acid. O-phenanthroline inhibited activity by 30 to 50%, but this may have been due to nonchelate properties. Anaerobic conditions or thiol protective agents were not essential for either the activity or stability of the cellulases during assay. An ultrafiltrable inhibitor of cellulase activity was detected in cell-free rumen fluid.     

Enzymatic saccharification of pretreated corn stover in a fed-batch membrane bioreactor

Enzymatic hydrolysis of corn stover was performed in an integrated membrane bioreactor (MBR) incorporating a 10 kDa flat sheet polysulfone membrane to increase cellulose conversion and to reduce enzyme dosage. Several pretreatment methods and semi-continuous MBR were examined to investigate their effect on the glucose yield and enzyme utilization efficiency. Compared with conventional batch reactor (CBR), cellulose conversion increased by 5% in a MBR because of the removal of glucose and cellobiose inhibitors. More than 15% increment in cellulose conversion was obtained using fed-MBR, and the reaction rate improved significantly. Enzyme utilization efficiency in a fed-batch MBR were 1.94-fold of CBR and 1.34-fold of fed-CBR for corn stover pretreated by soaking in aqueous ammonia and 3.31-fold of CBR and 1.32-fold of fed-CBR for corn stover pretreated by diluted sulfuric acid?Csodium hydroxide.     

Retention and regeneration of native NAD(H) in noncharged ultrafiltration membrane reactors: application to L-lactate and gluconate production

NAD(H) was retained in a noncharged ultrafiltration membrane reactor for the simultaneous and continuous production of L-lactate and gluconate with coenzyme regeneration. Polyethyleneimine (PEI), a 50-kDa cationic polymer, achieved coenzyme retentions above 0.8 for PEI/NAD(H) molar ratios higher than 5. The ionic strength of the inlet medium caused a decrease of NAD(H) retention that can be counterbalanced by an initial addition of 1% bovine serum albumin (BSA). Continuous reactor performance in the presence of PEI and BSA showed that NAD(H), glucose dehydrogenase, and lactate dehydrogenase were retained by 10-kDa ultrafiltration membranes; L-lactate and gluconate were produced at conversions higher than 95%. PEI enhanced the thermal stability of the enzymes used and increased the catalytic efficiency of glucose dehydrogenase, while no effect was found on the kinetic parameters of lactate dehydrogenase. A model that implements the kinetic equations of the two enzymes describes the reactor behavior satisfactorily. In brief, the use of PEI to retain NAD(H) is a new interesting approach to be widely applied in continuous synthesis with the large number of known dehydrogenases.     

Continuous cultivation of Trichoderma viride on cellulose

Using ball milled cellulose as the only carbon source Trichoderma viride was grown in a continuous flow culture at pH = 5.0 and T = 30°C. Steady-state values for cell protein, cellulose, and cellulase for different substrate concentrations (4–11 g/liter) and dilution rates (0.033–0.080 hr^?1) were obtained. Under steady-state conditions, 50–75% of the cellulose was consumed indicating a critical dilution rate on 0.17 hr^?1. Cellulase activity (U/ml) in the fermentation broth increased slightly with increasing substrate concentration and decreased with increasing dilution rate, while the specific cellulase productivity (U/mg cell protein·hr) was fairly independent of the dilution rate, with a maximum around D = 0.05 hr^?1. Following step changes in substrate concentration and dilution rate, new steady-state values were reached after three to five residence times (cell protein and cellulose) and four to six residence times (celullase activity).     

Polymerization of glucans by enzymatically active membranes

Conventional enzyme membrane reactors are not appropriate for a continuous synthesis of macromolecules and simultaneous product release. By immobilizing the enzyme in sufficiently large pores of a membrane an ensemble of miniaturized bioreactors is created. Product molecules are continuously removed from the enzyme by the flow of the reaction mixture across the membrane. Additionally, by varying the flow rate, it ought to be possible to influence the substrate as well as the enzyme-product residence times and thereby the product macromolecule's size. In this paper we present the first results of experiments involving enzymatic 1,4-alpha-glucan synthesis, using sucrose as substrate, maltooligosaccharides (DP 3-6) as primers, and membrane-immobilized amylosucrase. Epoxy groups for a covalent enzyme immobilization were generated on polypropylene microfiltration membranes by heterogeneous photoinitiated graft polymerization of glycidyl methacrylate. The influence of primer concentration and flow rate through the enzyme-membrane on amylosucrase activity, molecule growth, and coupling efficiency for glucose (% of coupled glucose versus free glucose) were investigated. The enzymatically mediated chain elongation of maltooligosaccharides by the successive addition of glucose units was achieved for the first time in a transmembrane process utilizing amylosucrase membranes.     

The pentose phosphate pathway leads to enhanced succinic acid flux in biofilms of wild-type <Emphasis Type="Italic">Actinobacillus succinogenes</Emphasis>

Increased pentose phosphate pathway flux, relative to total substrate uptake flux, is shown to enhance succinic acid (SA) yields under continuous, non-growth conditions of Actinobacillus succinogenes biofilms. Separate fermentations of glucose and xylose were conducted in a custom, continuous biofilm reactor at four different dilution rates. Glucose-6-phosphate dehydrogenase assays were performed on cell extracts derived from in situ removal of biofilm at each steady state. The results of the assays were coupled to a kinetic model that revealed an increase in oxidative pentose phosphate pathway (OPPP) flux relative to total substrate flux with increasing SA titre, for both substrates. Furthermore, applying metabolite concentration data to metabolic flux models that include the OPPP revealed similar flux relationships to those observed in the experimental kinetic analysis. A relative increase in OPPP flux produces additional reduction power that enables increased flux through the reductive branch of the TCA cycle, leading to increased SA yields, reduced by-product formation and complete closure of the overall redox balance.     

Enhanced COD and nitrogen removals for the treatment of swine wastewater by combining submerged membrane bioreactor (MBR) and anaerobic upflow bed filter (AUBF) reactor

In order to enhance performances of organics removal and nitrification for the treatment of swine wastewater containing high concentration of organic solids and nitrogen than conventional biological nitrogen removal process, a submerged membrane bioreactor (MBR) was followed by an anaerobic upflow bed filter (AUBF) reactor in this research (AUBF–MBR process). The AUBF reactor is a hybrid reactor, which is the combination of an anoxic filter for denitrification and upflow anaerobic sludge blanket (UASB) for acid fermentation. In the AUBF–MBR process, it showed a considerable enhancement of the effluent quality in terms of COD removal and nitrification. The submerged MBR could maintain more than 14,000 mg VSS/L of the biomass concentration. Total nitrogen (T-N) removal efficiency represented 60% when internal recycle ratio was three times of flow-rate (Q), although the nitrification occurred completely. Although the volatile fatty acids produced in AUBF reactor can enhance denitrification rate, but the AUBF–MBR process showed reduction of overall removal efficiency of the nitrogen due to the reduction of carbon source by methane production in the AUBF reactor compared to that of theoretical nitrogen removal efficiency.

Long-term operation of the submerged MBR showed that the throughputs of the submerged MBR were respectively 74, 63, and 31 days at 10, 15, and 30 L/m² h (LMH) of permeate flux. Resistance to filtration by rejected solid is the primary cause of fouling, however the priority of cake resistance (R_c) and fouling resistance (R_f) with respect to filtration phenomenon was different according to the amount of permeate flux. The submerged MBR, here, achieved a steady-state flux of 15 LMH at 0.4 atm. of trans-membrane pressure (TMP) but the flux can be enhanced in the future because shear force by tangential flow will be greater when multi-layer sheets of membrane were used.