An additional simple denitrification bioreactor using packed gel envelopes applicable to industrial...

ARTICLE

An Additional Simple Denitrification BioreactorUsing Packed Gel Envelopes Applicable toIndustrial Wastewater Treatment

Masahiko Morita, Hiroaki Uemoto, Atsushi Watanabe

Environmental Science Research Laboratory, Central Research Institute of Electric Power

Industry (CRIEPI), 1646 Abiko, Abiko-shi, Chiba 270-1194, Japan;

telephone: þ81-4-7182-1181; fax: þ81-4-7183-3347; e-mail: [email protected]

Received 31 October 2006; accepted 12 January 2007

Published online 24 January 2007 in Wiley InterScience (www.interscience.wiley.com
). DOI 10.1002/bit.21349
ABSTRACT: A simple denitrification bioreactor for nitrate-containing wastewater without organic compounds wasdeveloped. This bioreactor consisted of packed gel envelopesin a single tank. Each envelope comprised two plates of gelscontaining Paracoccus denitrificans cells with an internalspace between the plates. As an electron donor for deni-trification, ethanol was injected into the internal space andnot directly into the wastewater. P. denitrificans cells in thegel reduced nitrate to nitrogen gas by using the injectedethanol. Nitrate-containing desulfurization wastewaterderived from a coal-fired thermal power plant was con-tinuously treated with 20 packed gel envelopes (size,1,000� 900� 12 mm; surface area, 1.44 m2) in a reactortank (volume 1.5 m3). When the total nitrogen concentra-tion in the inflow was around 150 mg-N�LS1, the envelopesremoved approximately 60–80% of the total nitrogen, andthe maximum nitrogen removal rate was 5.0 g-N�dayS1 persquare meter of the gel surface. This value correspondedto the volumetric nitrogen removal performance of0.109 kg-N�mS3�dayS1. In each envelope, a high utilizationefficiency of the electron donor was attained, although morethan the double amount of the electron donor was empiri-cally injected in the present activated sludge system toachieve denitrification when compared with the theoreticalvalue. The bioreactor using the envelopes would be extre-mely effective as an additional denitrification system becausethese envelopes can be easily installed in the vacant spaces ofpreinstalled water treatment systems, without requiringadditional facilities for removing surplus ethanol and sludge.

Biotechnol. Bioeng. 2007;97: 1439–1447.

� 2007 Wiley Periodicals, Inc.

KEYWORDS: nitrogen removal; wastewater treatment; im-mobilized bacteria; denitrification; Paracoccus denitrificans

Correspondence to: H. Uemoto

� 2007 Wiley Periodicals, Inc.

Introduction

Nitrogen compounds that are derived from fertilizers,livestock excreta, and domestic and industrial effluentscause environmental pollution such as eutrophication(Kuenen and Robertson, 1988). Therefore, strict regulationshave been applied for the nitrogen content of industrialwastewater in Japan. In addition, the environmental qualitystandard value for the nitrate and nitrite contents in publicwaters is set below 10 mg-N�LS1 in Japan.

In various industries that use combustion furnaces, forexample, thermal power plants and chemical plants, a part ofthe desulfurization wastewater contains nitrates derivedfrom NOX, which is formed from the oxidation of nitrogengas in the air or from the nitrogen element in the fuel of thecombustion furnace. A characteristic of this desulfurizationwastewater is that it contains little organic carbon. Nitrate-containing desulfurization wastewater occurs in coal-firedthermal power plants that do not have a fuel gas denitrationunit. At present, biological denitrification by bacteria iswidely used for nitrogen removal in comparison to chemicaland physical methods (Shrimali and Singh, 2001). Thenitrate-containing desulfurization wastewater is mostlytreated by the following three steps: anoxic denitrification,aerobic treatment to remove the surplus electron donoradded to the wastewater in the denitrification step, andsettling to segregate the activated sludge containing bacteriafrom wastewater (Tchobanoglous and Burton, 1991). Sincethere are many steps in the present nitrogen removalsystems, a complicated sequence of operations is necessary,and a large installation area is required.

To simplify the present nitrogen removal systems, a novelimmobilized-cell bioreactor using packed gel envelopes wasproposed and investigated; each envelope comprised two

Biotechnology and Bioengineering, Vol. 97, No. 6, August 15, 2007 1439

plates of polymeric gels with an internal space between theplates for ethanol injection (Uemoto and Saiki, 2000). Thisbioreactor does not require the aerobic step because electrondonors are not directly added to the wastewater but areadded to the internal space of the envelopes. The bioreactormay probably attain a high utilization efficiency of electrondonors for denitrification and decrease the quantity ofsurplus sludge. Thus, the bioreactor can remove nitrogenfrom wastewater in a single step. In addition, the envelopescan be easily installed in vacant spaces of preinstalledwastewater treatment systems; thus, the bioreactor would beextremely effective as an additional denitrification system.Since the research experiments were performed underlaboratory conditions, the scale-up of the bioreactor isnecessary for application to industrial wastewater. In thisstudy, a large-scale bioreactor using packed gel envelopeswas investigated and nitrate-containing wastewater from acoal-fired thermal power plant was treated.

Materials and Methods

Bacterial Strain and Culture Medium

The denitrifier Paracoccus denitrificans JCM-6892 was usedin this study. P. denitrificanswas aerobically cultured at 308Cin a medium (pH 7.2) containing (LS1) 10.0 g peptone(Difco), 10.0 g meat extract (Difco), and 5.0 g NaCl. Allnutrients were dissolved in distilled water.

Characteristics of the Actual Wastewater

Desulfurization wastewater from which heavy metals wereremoved by an antecedent process was used. The organiccarbon content of this desulfurization wastewater was verylow. The total organic carbon (TOC) concentration inthe wastewater ranged between 1.8 and 10.9 mg-C�LS1. Thewastewater contained variable concentrations of nitrate(11.6–167.3 mg-N�LS1), whereas a low concentration ofnitrite (maximum value: 1.4 mg-N�LS1) and no ammonia(<0.2 mg-N�LS1) were detected during batch and con-tinuous operations. Phosphates (<0.1 mg�LS1) were notdetected in the wastewater, although it contained greatamount of sulfate (731–3,157 mg�LS1).

Figure 1. Schematic diagram of the packed gel envelope.

Packed Gel Envelope

P. denitrificans cells were harvested by centrifugation (20,000g,10 min, 48C) and washed three times with phosphate buffer(pH 7.5) containing (LS1) 9.0 g Na2HPO4 � 12H2O and1.5 g KH2PO4. P. denitrificans cells were suspended in thephosphate buffer (11 mg�dry weight�mLS1). The suspensionwas then mixed with a photo-crosslinkable polymer,namely, PVA-SbQ (SPP-H-13, Toyo Gosei Kogyo Co.,Chiba, Japan) in the ratio of 1:3.

1440 Biotechnology and Bioengineering, Vol. 97, No. 6, August 15, 2007

Each packed envelope (1,000� 900� 12 mm) containeda frame composed of vinyl chloride and polyester nets(UX-Screen 100-040/255PW, Toray Co., Chiba, Japan)that were attached to both sides of the frame (Fig. 1). Theabovementioned suspension containing P. denitrificanswas applied to the double-sided polyester nets of thepacked envelope at 0.5 mm thickness and solidified into agel by irradiation with metal halide lamps for 20 min(1,000 mmol�mS2�sS1). The area of the double-sided gelsurface of each packed gel envelope was 1.44 m2. Sincethe polyester nets of the packed envelope were tensioned,the packed gel envelope contained an internal spaceafter solidification. Three holes were made at the topof each envelope, and polyurethane tubes (outerdiameter, 8 mm; inner diameter, 6 mm) were attachedto the holes. Of the three holes, twoweremade for the exitof nitrogen gas after denitrification, and one central holewas made for the injection of an electron donor fordenitrification.

Bioreactor System

A large-scale bioreactor system was installed in Ishikawacoal-fired thermal power plant (Uruma City, OkinawaPrefecture, Japan). The system consisted of a reactor tank(1.0� 1.0� 1.5 m); an overflow tank; packed gel envelopes;pumps for wastewater circulation; a feeder for the electrondonor; meters and recorders for temperature, pH, and thedissolved oxygen concentration in the reactor tank; a controlpanel, etc. The interval between each envelope was 10 mm(Fig. 2a). In the first method, the wastewater was pumpedinto the reactor tank and treated in a batch operation. In thesecond method, the wastewater was continuously pumpedinto the reactor tank and overflowed into the overflow tank.The treated wastewater was then removed using a pump.Figure 2b shows the schematic diagram of the system undercontinuous operation.

Batch Operation

Batch operations were carried out overnight using 10 packedgel envelopes containing P. denitrificans. The desulfurization

DOI 10.1002/bit

Figure 2. a: Interior of the reactor tank. b: Schematic flow diagram of the bioreactor system under continuous operation.

wastewater (containing mostly nitrate) that originatedfrom the thermal power plant was treated with theabovementioned bioreactor system (water volume1.16 m3). At the start of the operation, 10% ethanol(220 mL per day) was injected into the internal spaceof each envelope. The phosphate source (concentrationLS1; 0.09 g Na2HPO4 � 12H2O and 0.015 g KH2PO4)and trace elements (concentration LS1; 2 mgMgSO4 � 7H2O, 0.1 mg CaCl2 � 2H2O, 5 mg NaHCO3,0.05 mg EDTA-Fe, 1 mg ZnSO4 � 7H2O, 0.3 mgMnCl2 � 4H2O, 3 mg H3BO3, 2 mg CoCl2 � 6H2O, 0.1 mgCuCl2 � 2H2O, 0.2 mg NiCl2 � 6H2O, and 0.3 mg Na2MoO4 � 2H2O) were added to the ethanol solution. Thewastewater was sampled for analysis at the start time, after3 h, and at the end time.

Continuous Operation

The wastewater was continuously treated using thebioreactor system (water volume 1.32 m3) with 20 envelopesat the inflow rate of 0.055 m3�hS1 (average hydraulicretention time of 24 h) for 2 months. Ten percent

ethanol (80 mL) was injected into each envelopeevery 12 h (160 mL per day per envelope). Thewastewater was sampled for analysis every 4 h from9:00 AM to 9:00 PM.

Analytical Methods

Ammonia and nitrite concentrations in the wastewater werecolorimetrically measured according to previously describedmethods (Callaway, 1992). The nitrate concentration wasdetermined using an ion-chromato analyzer (DX-AQ,Dionex Co., Sunnyvale, CA) with an IonPac AS12A column.The TOC concentration was measured using a TOC analyzer(TOC-650, Toray Engineering Co.).

Results

Batch Operation Using Nitrate-Containing Wastewater

When the wastewater containing various concentrations ofnitrate was treated with ethanol as an electron donor in the

Morita et al.: An Additional Denitrification Bioreactor 1441

Biotechnology and Bioengineering. DOI 10.1002/bit

Figure 4. Relationship between the initial total nitrogen concentration and the

nitrogen removal rate in the bioreactor under batch operations.

large-scale bioreactor system with 10 packed gel envelopes,the nitrate concentration in the wastewater decreasedgradually (Fig. 3), whereas a very low concentration ofnitrite (maximum value: 3.2 mg-N�LS1) was detected.Ammonia was not detected (<0.2 mg-N�LS1) at any timepoint during the treatment duration. The pH of thewastewater increased gradually and ranged between 7.0 and8.3. The temperature of the wastewater ranged between29.0 and 33.0-C.

The nitrogen removal rate (the transformation of nitrateto nitrogen gas) by the 10 gel envelopes was calculated basedon the change in total nitrogen (the sum of nitrate, nitrite,and ammonia) concentrations at the start and end time ofthe batch operation. The bioreactor yielded nitrogenremoval rates between 2.2 and 6.5 g-N�dayS1 per squaremeter of the gel surface at various initial concentrations ofnitrate. In the batch operation, the surface area of allenvelopes per water volume was 12.4 m2�mS3. Therelationship between the initial total nitrogen concentrationand the nitrogen removal rate in the bioreactor systemunder the batch operations is shown in Figure 4. Thenitrogen removal rate was proportional to the totalnitrogen concentration, thereby yielding the followingrelationship:

R ¼ X1C þ Y1; r2 ¼ 0:926 (1)

where C is the total nitrogen concentration in the reactortank (g-N�mS3), R is the nitrogen removal rate persquare meter of the gel surface (g-N�mS2 � dayS1), X1 is aconstant (m3�mS2 � dayS1), and Y1 is a constant (g-N�mS2 �dayS1). In this case, X1 and Y1 were 0.042 m3�mS2�dayS1

and 1.144 g-N�mS2 � dayS1, respectively.

Figure 3. Typical time-dependent change in nitrate concentration in the actual

wastewater in the bioreactor under batch operations. The initial nitrate concentrations

are as follows: Run 1, 118.4 mg-N�LS1; Run 2, 96.3 mg-N�LS1; Run 3, 79.4 mg-N�LS1; Run

4, 59.4 mg-N�LS1; Run 5, 32.6 mg-N�LS1.


Prediction of Time-Dependent Changes in the TotalNitrogen Concentration

The material balance on nitrogen in the bioreactor can beexpressed as follows:

VdC

dt

� �¼ FCin � FCout � AallR (2)

where Aall is the surface area of all packed gel envelopes inthe reactor tank (m2), Cin is the total nitrogen concentra-tion in the inflow (g-N�mS3), Cout is the total nitrogenconcentration in the outflow (g-N�mS3), F is the wastewaterflow rate (m3�dayS1), t is time (day), and V is the wastewatervolume in the reactor tank (m3). Based on Eqs. (1) and (2),the total nitrogen concentration in the outflow in thecontinuous operation was predicted under various condi-tions. Since the bioreactor could be regarded as a continuousstirred tank-type bioreactor, Cout was equal to C. Here, thecalculation was carried out by considering that the totalnitrogen concentration in the inflow was 100 mg-N�LS1. Theeffects of the wastewater flow rate and the number of packedgel envelopes on the total nitrogen concentration in theoutflow are shown in Figure 5. In any case, a steady state wasachieved in the reactor tank after 60 h. For 20 envelopes, thetotal nitrogen concentrations in the outflow were calculatedto be 39.2, 60.0, 76.3, 86.9, and 95.3 mg-N�LS1 under thesteady state when the wastewater flow rates were 0.055, 0.11,0.22, 0.44, and 1.32 m3�hS1, respectively (Fig. 5a). For thewastewater flow rate of 0.055 m3�hS1, the total nitrogenconcentrations in the outflow were calculated to be 39.2,26.3, 17.7, 11.4, and 6.7 mg-N�LS1 under the steady state

DOI 10.1002/bit

Figure 5. Simulation of the time-dependent changes in the total nitrogen

concentration in the outflow from the bioreactor under continuous operations:

(a) effect of the wastewater flow rate on 20 packed gel envelopes, (b) effect of

the number of packed gel envelopes on the wastewater flow rate of 0.055 m3�hS1.

when the number of envelopes used was 20, 30, 40, 50, and60, respectively (Fig. 5b).

Figure 6. Time-dependent changes in (a) nitrogen concentrations in the inflow

and outflow, (b) TOC concentrations in the inflow and outflow, (c) temperature, (d) pH,

and (e) dissolved oxygen (DO) concentration in the reactor tank under continuous

operations using 20 envelopes at the wastewater flow rate of 0.055 m3�hS1. Symbols:

nitrate concentration in the inflow (&) and in the outflow (&), nitrite concentration in

the inflow (~) and in the outflow (~), and total nitrogen concentration in the inflow

(*) and in the outflow (*) in (a); TOC concentration in the inflow (&) and in the

outflow (&) in (b).

Nitrogen Removal in Continuous Operation UsingNitrate-Containing Wastewater

The wastewater was continuously treated in the large-scalebioreactor system with 20 packed gel envelopes at the flowrate of 0.055 m3�hS1 (average hydraulic retention time of24 h) for 2 months. The time-dependent changes inthe nitrogen concentrations in the inflow and outflow areshown in Figure 6a. Sudden variations in the nitrateconcentrations in the inflow resulted in abrupt fluctuationsin the total nitrogen concentration. A very low concentra-tion of nitrite (maximum value: 1.4 mg-N�LS1) was detectedin the inflow. The nitrate, nitrite, and total nitrogenconcentrations in the outflow varied according to the abruptfluctuations in the total nitrogen concentration in theinflow. Ammonia was not detected (<0.2 mg-N�LS1) in theinflow and outflow during the continuous operation. Whenthe total nitrogen concentration in the inflow was low(below 50 mg-N�LS1), the accumulation of nitrite in thereactor tank did not occur, and the total nitrogenconcentration in the outflow ranged between 3.0 and7.5 mg-N�LS1. On the other hand, when the totalnitrogen concentration in inflow was high (greater than150 mg-N�LS1), the accumulation of nitrite in the reactortank occurred (the concentration ranged between 16.2 and

26.8 mg-N�LS1). In this case, the nitrate concentration in theoutflow ranged between 35.1 and 49.5 mg-N�LS1, and thetotal nitrogen concentration in the outflow was between51.3 and 75.6 mg-N�LS1. The total nitrogen concentrationin the outflow was maintained below 50 mg-N�LS1,when the total nitrogen concentration in the inflow wasless than 100 mg-N�LS1.

The TOC concentrations in the inflow and outflow,temperature, pH, and dissolved oxygen concentration in thereactor tank under the continuous operation are shown inFigure 6. The TOC concentrations in the inflow andoutflow ranged between 1.8 and 10.9 mg-C�LS1 and between17.0 and 150.5 mg-C�LS1, respectively (Fig. 6b). When thetotal nitrogen concentration in the inflow was high (greater



than 150 mg-N�LS1), the TOC concentration in the outflowwas relatively low (between 23.3 and 38.5 mg-C�LS1). Thetemperature ranged between 29.1 and 34.4-C, andthe average temperature of 2 months was 32.2-C(Fig. 6c). The pH ranged between 7.0 and 7.6 (Fig. 6d).The dissolved oxygen concentration ranged between 0.0 and1.4 mg-O2�LS1, and it was maintained at a low level duringthe continuous operation (Fig. 6e).

Figure 7 shows the time-dependent changes in thenitrogen removal efficiency and nitrogen removal rate in thecontinuous operation using 20 gel envelopes. The nitrogenremoval efficiency, which was defined as the ratio of thedifference between the total nitrogen concentrationsin the inflow and outflow to the total nitrogen concentra-tion in the inflow, was calculated from the totalnitrogen concentrations in the inflow and outflow. Thenitrogen removal rate was also calculated based on thedifference between the total nitrogen concentrations inthe inflow and outflow. With the exception when there wassome abrupt decrease in the nitrogen removal efficiency, thebioreactor using 20 gel envelopes removed approximately60–80% of the total nitrogen in the wastewater (Fig. 7a). Thenitrogen removal rate varied abruptly, and some suddendecrease in the nitrogen removal rate was observed (Fig. 7b).When the total nitrogen concentration in the inflow was147.4 mg-N�LS1, the maximum nitrogen removal rate of5.0 g-N�dayS1 per square meter of the gel surface wasattained. In the continuous operation, the surface area of allenvelopes per water volume was 21.8 m2�mS3.

Figure 7. Time-dependent changes in (a) the nitrogen removal efficiency and

(b) the nitrogen removal rate in the bioreactor using 20 envelopes under continuous

operations at the wastewater flow rate of 0.055 m3�hS1.


Relationship Between Total Nitrogen Concentrationand Nitrogen Removal Rate

The relationship between the total nitrogen concentra-tion in the inflow and the nitrogen removal rate is shown inFigure 8. The data obtained when the total nitrogenconcentration in the inflow varied abruptly was not reliablebecause some time was consumed during the restoration ofthe conditions in the reactor tank to the steady state.Therefore, the data that was obtained when the totalnitrogen concentration in the inflow varied by greaterthan 30 mg-N�LS1 during 24 h was excluded. The nitrogenremoval rate was proportional to the total nitrogenconcentration in the inflow, thereby yielding the followingrelationship:

R ¼ X2Cin þ Y2; r2 ¼ 0:948 (3)

where X2 is a constant (m3�mS2�dayS1) and Y2 is aconstant (g-N�mS2�dayS1). In this case, X2 and Y2were 0.028 m3�mS2�dayS1 and 0.109 g-N�mS2�dayS1,respectively. The higher the total nitrogen concentrationin the inflow, the larger the nitrogen removal rate.

Discussion

Scale-Up of the Packed Gel Envelope

The bioreactor (water volume 0.25 L) using 8 packed gelenvelopes (height, 100 mm; length, 48 mm; thickness,0.5 mm; surface area, 0.0096 m2) yielded the maximumnitrogen removal rate of 5.0 g-N�dayS1 per square meter ofthe gel surface (Uemoto and Saiki, 2000). In the condition,the surface area of all envelopes per water volume was332 m2�mS3. The packed gel envelopes should be scaled upin order to treat a large amount of actual wastewater. In thescale-up of the packed gel envelopes, the maintenance of

Figure 8. Relationship between the total nitrogen concentration in the inflow

and the nitrogen removal rate in the bioreactor under continuous operations.

DOI 10.1002/bit

nitrogen removal rate per square meter of the gel surface ismost important. For that, the wastewater should be flowedsmoothly so as to diffuse the nitrate in the wastewater intothe packed gel envelope and the injected ethanol should beextended to the whole internal space of the packed gelenvelope. The functional size of the packed gel envelopeshould be decided from these viewpoints. When the scale-upof the packed gel envelopes was examined, large intervalbetween each envelope was kept and each polyurethane tubefor injection of ethanol had five holes at the tip, which faceddifferent directions.

The size of the packed gel envelopes (height, 500 mm;length, 1,000 mm; thickness, 16 mm; surface area, 0.72 m2)was increased and the nitrogen removal performance wasinvestigated in preliminary experiments. When the artificialnitrate-containing wastewater was treated using a laboratorybatch bioreactor system (water volume 0.35 m3) with twopacked gel envelopes that were half the size of the envelopesused in this study, the maximum nitrogen removal rate of6.5 g-N�dayS1 per square meter of the gel surface wasachieved (data not shown). In this condition, the surfacearea of all envelopes per water volume was 4.1 m2�mS3.Since the bigger packed gel envelopes could achieve identicalnitrogen removal performance, a packed gel envelope offunctional size (height, 1000 mm; length, 900 mm;thickness, 12 mm; surface area, 1.44 m2) was developedfor the scale-up of the bioreactor system to treat actualwastewater in this study. The maximum nitrogen removalrate of 6.5 g-N�dayS1 per square meter of the gel surface wasattained in the bioreactor system under the batch operation.In the condition, the surface area of all envelopes per watervolume was 12.4 m2�mS3. The results showed that thepacked gel envelopes of functional size used in this studyattained a nitrogen removal rate that was equivalent tothose in the beaker scale and preliminary laboratoryexperiments.

Simulation of Total Nitrogen Concentration

Based on the prediction of time-dependent changes in thetotal nitrogen concentration, 20 packed gel envelopes wereused in the bioreactor with the wastewater flow rate of0.055 m3�hS1 in order to achieve a total nitrogenconcentration of below 50 mg-N�LS1 in the outflowwhen the total nitrogen concentration in the inflow wasless than 100 mg-N�LS1. When 20 envelopes were used inthe bioreactor with the wastewater flow rate of 0.055 m3�hS1

and the total nitrogen concentration in the inflow was100 mg-N�LS1, the total nitrogen concentration in theoutflow was calculated to be 39.2 mg-N�LS1. In fact, the totalnitrogen concentration in the outflow could be maintainedbelow 39.1 mg-N�LS1 in the continuous operation. Thegood agreement between the calculated and experimentalresults shows that the simulation method is quite effective topredict the total nitrogen concentration in the outflowand to determine the appropriate operation method for

attaining the set point. The simulation method is extremelyuseful for estimating the required number of packed gelenvelopes and installation area of the treatment system whenthe entire wastewater in a thermal power plant is to betreated.

Ethanol Injection for Denitrification

An electron donor (e.g., ethanol) is necessary fordenitrification. The theoretical maximum nitrogen removalrate per square meter of the gel surface was calculated asfollows:

Rmax ¼a� b� r� c

A(4)

where A is the surface area of each packed gel envelope(m2), a is the absolute ethanol volume injected into eachenvelope per day (mL�dayS1), b is the purity of used ethanol(S), c is the nitrogen mass capable of denitrifying perethanol mass (g-N�gS1), Rmax is the theoretical maximumnitrogen removal rate per square meter of the gel surface(g-N�mS2�dayS1), and r is the relative density of ethanol at20-C (g�mLS1). In this case, A, b, and r were 1.44 m2, 0.995,and 0.789 g�mLS1, respectively. In the continuous opera-tion, since 160 mL of a 10% ethanol solution was injectedinto each envelope per day, a was 16 mL�dayS1. Thechemical reaction in the denitrification step using ethanol isas follows:

12NO�3 þ 5C2H5OH

! 6N2 þ 10CO2 þ 9H2Oþ 12OH� (5)

Thus, one mole of ethanol can denitrify 2.4 mol of nitrate,and c was 0.73 g-N�gS1. As a result, the theoretical maximumnitrogen removal rate per square meter of the gel surface wascalculated to be 6.4 g-N�mS2�dayS1. In this study, when themaximum nitrogen removal rate of 5.0 g-N�dayS1 persquare meter of the gel surface was attained, the maximumethanol utilization efficiency for denitrification was 78.1%.More than the double amount of the electron donor wasempirically injected in the present activated sludge system toachieve denitrification when compared with the theoreticalvalue. In an advanced sewage treatment system using ahollow fiber pilot-scale bioreactor, methanol was used as anelectron donor at an amount that was greater than threetimes the theoretical value (Hatanaka et al., 2005). Thebioreactor with the packed gel envelopes could use ethanol(electron donor) effectively when the total nitrogenconcentration in the inflow was high.

During the continuous operation, the TOC concentrationin the outflow was always higher than that in the inflow. Theincrease in the TOC concentration in the outflow was due tothe injection of excess ethanol. When the total nitrogenconcentration in the inflow was greater than 150 mg-N�LS1,



the TOC concentrations in the inflow and outflow werebetween 3.0 and 8.0 mg-C�LS1 and between 23.3 and38.5 mg-C�LS1, respectively. Since the wastewater flow ratewas 0.055 m3�hS1, the average carbon leakage derived fromethanol was calculated as 33.3� 6.6 g-C�dayS1. On the otherhand, the injected carbon in the ethanol per day wascalculated as follows:

M ¼ a� b� r� d � n (6)

where M is the injected carbon in the ethanol per day(g-C�dayS1), d is the ratio of carbon in the molecule ofethanol (g-C�gS1), and n is the number of packed gelenvelopes under continuous operation (–). In this case, a, b,d, n, and r were 16 mL�dayS1, 0.995, 0.522 g-C�gS1, 20, and0.789 g�mLS1, respectively. As a result, the injected carbon inthe ethanol per day was calculated to be 131.1 g-C�dayS1.Thus, the average ratio of carbon leakage was 25.4� 5.1%when the total nitrogen concentration in the inflow wasgreater than 150 mg-N�LS1. On the other hand, since theaverage nitrogen removal rate per square meter of the gelsurface was 4.3� 0.3 g-N�dayS1, the average ratio of unusedethanol was 32.7� 4.8%. In the case of an entire leakageof the unused ethanol, the amount of carbon leakagecorresponds to 42.9� 6.2 g-C�dayS1. However, the actualamount of carbon leakage was 33.3� 6.6 g-C�dayS1. Theseresults showed that 79.9� 22.8% of the injected ethanol waseffectively used for denitrification. The remaining carbonwas probably used for the growth of microorganisms andconsumption of molecular oxygen in the wastewater. Thebioreactor could use the electron donor effectively anddecrease the quantity of surplus sludge when the totalnitrogen concentration in the inflow was greater than150 mg-N�LS1. However, a high level of TOC was existed inthe outflow when the total nitrogen concentration wasrelatively low. The quantitative control of injected ethanolwill be necessary according as the change of the totalnitrogen concentration in the inflow. As the otherpracticable countermeasure, moderate aeration in thereactor tank seems to be quite effective for the decreaseof TOC concentration in the outflow though the aerationsystem could not be installed in the bioreactor for facilitycondition in this pilot test.

Table I. Comparison of the volumetric nitrogen removal rates of various nit

Reactor system Polymeric gel for imm

Water-jacketed glass column (Kokufuta et al., 1987) Polyelectrolyte comple

calcium alginate

Reactor using an immobilized pellet

(Mori et al., 1993)

Polyethylene glycol re

Reactor using packed gel envelopes

(Uemoto and Saiki, 2000)

Photo-crosslinkable po

Reactor using packed gel envelopes (this study) Photo-crosslinkable po


Limiting Factor of Nitrogen Removal Performance

The nitrogen removal rate was proportional to the totalnitrogen concentration in both batch and continuousoperations. Since P. denitrificans requires an anaerobicenvironment for denitrification, it appeared that denitrifi-cation was carried out in the inner part of each envelope.Thus, for denitrification, nitrate should enter the inner partof the packed gel envelope by diffusion. Since the diffusionof nitrate into the packed gel envelope was proportional toits concentration, that is, the total nitrogen concentration(the concentration of nitrite was low in the wastewater in theinflow), the nitrogen removal performance of P. denitrificanswas thought to be proportional to the total nitrogenconcentration. In the bioreactor, the diffusion of nitrate intothe packed gel envelope was a limiting factor of the nitrogenremoval performance. In fact, nitrate was not present in theinternal space of the envelope when the pilot test undercontinuous operation was finished. The intensity of thewastewater flow may decrease the thickness of the interfacialfilm. This may improve the diffusion that was previouslylimited and increase nitrogen removal performance.

Volumetric Nitrogen Removal Performance

The packed gel envelopes attained the maximum nitrogenremoval rate of 5.0 g-N�dayS1 per square meter of the gelsurface. This value corresponded to the volumetric nitrogenremoval performance of 0.109 kg-N�mS3�dayS1. Thevolumetric nitrogen removal rates of various nitrogenremoval systems using immobilized cells are summarized inTable I. Since the water-jacketed glass column reactor couldremove 80 mg of nitrogen (200 mL of medium containing400 mg-N�LS1) in approximately 100 h, the volumetricnitrogen removal rate of this system was calculated to be0.61 kg-N�mS3�dayS1 (Kokufuta et al., 1987). On the otherhand, the nitrogen removal system using immobilizedmicroorganisms achieved a nitrogen removal rate of0.066 kg-N�mS3�dayS1 (Mori et al., 1993). In the actualthermal power plants, the nitrogen removal systems usingactivated sludge showed nitrogen removal rates rangingfrom 0.087 to 0.18 kg-N�mS3�dayS1 (unpublished data).The bioreactor using packed gel envelopes yielded avolumetric nitrogen removal rate of 1.60 kg-N�mS3�dayS1

under laboratory conditions (Uemoto and Saiki, 2000). In

rogen removal systems using immobilized cells.

obilization

Water

volume [L]

Volumetric nitrogen

removal rate [kg-N�dayS1 for 1 m3

reactor volume]

xes-stabilized 0.0314 0.61

sin 750,000 0.066

lymer (PVA-SbQ) 0.25 1.60

lymer (PVA-SbQ) 1,320 0.109

DOI 10.1002/bit

this study, the volumetric nitrogen removal rate of thebioreactor system using packed gel envelopes in a pilot testwas 0.109 kg-N�mS3�dayS1. This difference in the volu-metric nitrogen removal performances of the bioreactorusing packed gel envelopes under laboratory conditions andin a pilot test resulted from the surface area of all envelopesper water volume. Since the thickness of the envelope was3 mm and the interval between each envelope was 3 mmunder laboratory conditions, the surface area of all envelopesper water volume was 332 m2�mS3. On the other hand, inthe large-scale bioreactor system, the surface area of allenvelopes per water volume was 21.8 m2�mS3 because thethickness of and interval between each envelope in thissystem are greater than those of the laboratory system andthere were vacant spaces in the reactor tank (Fig. 2a). If thepacked gel envelopes were added to the vacant spaces of thereactor tank at equal intervals, 45 envelopes could be placedin the tank. When the packed gel envelopes attained thenitrogen removal rate of 5.0 g-N�dayS1 per square meterof the gel surface, the volumetric nitrogen removalperformance of the bioreactor system was estimated to be0.245 kg-N�mS3�dayS1. In this condition, the surface area ofall envelopes per water volume was 49.1 m2�mS3. In futurestudies, the lamellation of the packed gel envelopes will beessential in order to further increase the volumetric nitrogenremoval performance.

The bioreactor with the packed gel envelopes was capableof operating stably and continuously for removing nitrogenfrom wastewater for 2 months even though the totalnitrogen concentration in the inflow showed abruptfluctuations. The bioreactor using the packed gel envelopeswould be extremely effective as an additional denitrificationsystem because these envelopes can be easily installed in thevacant spaces of preinstalled water treatment systems,without requiring additional facilities for removing surplusethanol and sludge.

Nomenclature

A su
rface area of each packed gel envelope [m2]
Aall su
rface area of all packed gel envelopes [m2]
a a
bsolute ethanol volume injected into each packed gel envelope
per day [mL�dayS1]

b p
urity of used ethanol [–]
c n
itrogen mass capable of denitrifying per ethanol mass [g-N�gS1]
C to
tal nitrogen concentration in the reactor tank [g-N�mS3]
Cin to
tal nitrogen concentration in the inflow [g-N�mS3]
Cout to
tal nitrogen concentration in the outflow [g-N�mS3]
d ra
tio of carbon in the molecule of ethanol [g-C�gS1]
F w
astewater flow rate [m3�dayS1]
M in
jected carbon in the ethanol per day [g-C�dayS1]
n n
umber of packed gel envelopes under continuous operation [S]
R n
itrogen removal rate per square meter of the gel surface
[g-N�mS2�dayS1]

Rmaxth
eoretical maximum nitrogen removal rate per square meter of
the gel surface [g-N�mS2�dayS1]

t ti
me [day]
V w
astewater volume in the reactor tank [m3]
X1 c
onstant [m3�mS2�dayS1]
X2 c
onstant [m3�mS2�dayS1]
Y1 c
onstant [g-N�mS2�dayS1]
Y2 c
onstant [g-N�mS2�dayS1]
r re
lative density of ethanol at 208C [g�mLS1]
The authors thank YokoWatanabe for her assistance with the analysis,

and Hiroshi Saiki, Ichiro Saitoh, Yoshiyuki Matsuki, Tomohiko

Yoshii, Tatsunori Yoshida, Hiroyuki Nakui, Tsuyoshi Matsuda, Taka-

shi Sakamoto, and Toshinori Baba for their valuable advice and

comments.

References

Callaway JO. 1992. Ammonia and nitrite. In: Greenberg AE, Clesceri LS,

Eaton A, editors. Standard methods for the examination of water and

wastewater, 18th edition. Washington DC: American Public Health

Association. pp 4–75–4–87.

Hatanaka C, Mizuno K, Hatate Y. 2005. Advanced sewage treatment with

hollow fiber bioreactor. Resour Process 52:167–171 (in Japanese).

Kokufuta E, Yukishige M, Nakamura I. 1987. Coimmobilization of Nitro-

somonas europaea and Paracoccus denitrificans cells using polyelec-

trolyte complex-stabilized calcium alginate gel. J Ferment Technol

65:659–664.

Kuenen JG, Robertson LA. 1988. Ecology of nitrification and denitrification.

In: Cole JA, Ferguson SJ, editors. The nitrogen and sulphur cycles.

Cambridge: Cambridge University Press. pp 162–218.

Mori N, Sumino T, Shirai M. 1993. Advanced wastewater treatment using

immobilized nitrifiers in synthetic resin. J Resour Environ 29:1057–

1062 (in Japanese).

Shrimali M, Singh KP. 2001. New methods of nitrate removal from water.

Environ Pollut 112:351–359.

Tchobanoglous G, Burton FL. 1991. Removal of nitrogen by biological

nitrification/denitrification. In: Clark BJ, Morris JM, editors. Waste-

water engineering-treatment, disposal, and reuse-, 3rd edition. Singa-

pore: McGraw-Hill, Inc. pp 711–726.

Uemoto H, Saiki H. 2000. Nitrogen removal reactor using packed gel

envelopes containing Nitrosomonas europaea and Paracoccus deni-

trificans. Biotechnol Bioeng 67:80–86.



An additional simple denitrification bioreactor using packed gel envelopes applicable to industrial...

Documents

Transcript of An additional simple denitrification bioreactor using packed gel envelopes applicable to industrial...