This article was downloaded by: [Northeastern University] On: 04 November 2014, At: 19:25 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Environmental Technology Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tent20

Biological NOx removal by denitrification process in a jet-loop bioreactor: system performance and model development a

a

a

b

a

Serdar Durmazpinar , Nalan Ilhan , Gonca Demir , Güçlü Insel , Nadir Dizge , Pınar a

a

a

Ergenekon , Elif Erhan & Bülent Keskinler a

Environmental Engineering Department, Gebze Institute of Technology, Gebze, Kocaeli TR-41400, Turkey b

Environmental Engineering Department, Istanbul Technical University, Maslak, Istanbul TR-34469, Turkey Published online: 06 Jan 2014.

To cite this article: Serdar Durmazpinar, Nalan Ilhan, Gonca Demir, Güçlü Insel, Nadir Dizge, Pınar Ergenekon, Elif Erhan & Bülent Keskinler (2014) Biological NOx removal by denitrification process in a jet-loop bioreactor: system performance and model development, Environmental Technology, 35:11, 1358-1366, DOI: 10.1080/09593330.2013.868529 To link to this article: http://dx.doi.org/10.1080/09593330.2013.868529

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions

Environmental Technology, 2014 Vol. 35, No. 11, 1358–1366, http://dx.doi.org/10.1080/09593330.2013.868529

Biological NOx removal by denitrification process in a jet-loop bioreactor: system performance and model development Serdar Durmazpinara , Nalan Ilhana , Gonca Demira , Güçlü Inselb , Nadir Dizgea , Pınar Ergenekona∗ , Elif Erhana and Bülent Keskinlera a Environmental

Engineering Department, Gebze Institute of Technology, Gebze, Kocaeli TR-41400, Turkey; b Environmental Engineering Department, Istanbul Technical University, Maslak, Istanbul TR-34469, Turkey

Downloaded by [Northeastern University] at 19:25 04 November 2014

(Received 3 June 2013; accepted 18 November 2013 ) Nitrogen monoxide (NO) and nitrogen dioxide referred as NOx are one of the most important air pollutants in the atmosphere. Biological NOx removal technologies have been developing to reach a cost-effective control method for upcoming stringent NOx emission standards. The BioDeNOx system was seen as a promising biological NOx control technology which is composed of two reactors, one for absorbing of NO in an aqueous Fe(II)EDTA2− solution and the other for subsequent reduction to N2 gas in a biological reactor by the denitrification process. In this study, instead of two discrete reactors, only one jet-loop bioreactor (JLBR) was utilized as both absorption and denitrification unit and no chelate-forming chemicals were added. In other words, the advantage of better mass transfer conditions of jet bioreactor was used instead of Fe(II)EDTA2− . The process was named as Jet-BioDeNOx. The JLBR was operated for the removal of NOx from air streams containing 500–3000 ppm NOx and the results showed that the removal efficiency was between 81% and 94%. The air to liquid flow ratio (QG /QRAS ) varied in the range of 0.07–0.12. Mathematical modelling of the system demonstrated that the removal efficiency strongly depends on this ratio. The high mass transfer conditions prevailed in the reactor provided a competitive advantage on removing NO gas without any requirement of chelating chemicals. Keywords: nitric oxide (NOx ) removal; jet-loop bioreactor; denitrification; air pollution control technology; BioDeNOx

1. Introduction The chemical element nitrogen (N) can form several different oxides having valence states from plus one to plus six. Nitrogen monoxide (NO) and nitrogen dioxide (NO2 ) are among the nitrogen oxide gases known as NOx . Main anthropogenic sources for NOx are fuel combustion and nitric acid production in various chemical and agricultural industries. Flue gas from coal combustion contains about 90% of NOx in NO form. The NOx can lead to urban air pollution, acid deposition, and visibility impairment, upon entering the atmosphere.[1] Other than these atmospheric effects, the epidemiological studies have also shown associations between NO2 concentrations and daily mortality.[2] The NOx control technologies have been developed as a result of concern for environmental and health issues related to NOx pollution. Conventional post-combustion NOx controls are mainly selective non-catalytic reduction (SNCR) and selective catalytic reduction (SCR). The SNCR systems use gas phase, homogeneous chemical reactions to chemically reduce a portion of the NOx emissions to N2 in a high temperature zone of the combustion system. The SCR systems use ammonia to react with NOx in a complex set

∗ Corresponding

author: Email: [email protected]

© 2014 Taylor & Francis

of high-temperature heterogeneous reactions on the surface of a catalyst. The SCR has some disadvantages such as high-energy consumption, high capital cost, ammonia use, and related ammonia emission risks. The SNCR operations, on the other hand, require very high temperature and have also a risk of appearance of the ammonium in the exhaust gas if the reaction temperature is below 730◦ C. Hence, the efficient and low-cost biological NOx abatement technologies have been seeking as an alternative to current control technologies.[3] In recent years, many researchers have investigated the removal of NOx using conventional gasphase bioreactors such as biofilters, biotrickling filters, and bioscrubbers.[4,5] In addition, several alternative processes have been investigated such as using membrane bioreactors (MBRs), rotating drum biofilters, and rotating biological contactors.[6–8] There are also combined systems, such as BioDeNOx, which is an integrated physico-chemical and biological process consisting of two discrete reactors, namely a scrubber and a biological reactor.[9–11] A chelating agent, Fe(II)(EDTA), is used to improve the transfer of gaseous NO into the scrubbing liquid. Hence, the absorption of NOx into aqueous phase is achieved in the first reactor. In the second biological reactor, the iron chelate

Downloaded by [Northeastern University] at 19:25 04 November 2014

Environmental Technology is regenerated while NO is reduced to nitrogen by denitrification. The BioDeNOx system was shown to be operated continuously by van der Maas et al.[12] During denitrification process, EDTA (NO absorbed Fe(II)EDTA−2 and also oxidized form of Fe(III)EDTA− due to O2 in the flue gas) is regenerated so as to be used again in the absorption stage. The activated sludge modelling has become as an indispensable tool for engineers since it enables in-depth analysis of biological nitrogen and phosphorus removal processes in bioreactors.[13,14] The activated sludge models were initially formulated to describe the electron acceptor/donor conditions among substrate, O2 , and nitrate (NO− 3 ). As a result, the standard models often consider both nitrification and denitrification as single-step processes. However, increasing awareness on emissions of greenhouse gases (N2 O, NO, etc.) during biological treatment has motivated modellers to include partially oxidized nitrogen components in biological process models. Consequently, the standard models were later extended to multi-step nitrification and denitrification processes in order to calculate the greenhouse gas emissions.[15–17] Low solubility of NO is an issue in developing an efficient biological NOx control technology. It is well known that jet-loop bioreactors (JLBRs) are extensively used in the field especially as aerobic biological reactors and provide high mass transfer conditions compared to conventional

1359

biological reactors. In this study, a compact NOx control system was developed by employing a JLBR as both NOx absorber and denitrifying bioreactor. Therefore, in the proposed system higher transfer efficiency of gaseous NO into liquid phase was achieved using only one reactor and without adding a chelate-forming compound as done in the BioDeNOx process. The proposed process was named as Jet-BioDeNOx. The performance of the system was evaluated with a multi-component-activated sludge model to elucidate the nitrogen transformations in the process. 2. Materials and methods 2.1. JLBR configuration The schematic illustration of the pilot Jet-BioDeNOx reactor configuration used in this study is shown in Figure 1. The inlet synthetic flue gas was prepared by using three mass flow controllers (GFC-17-AALBOR6) calibrated for Ar, NO, and O2 , respectively, and mixed in a stainless steel gas-mixing chamber for ensuring the homogeneous composition. The Ar was used as the balance gas to be able to monitor NOx conversion efficiency in terms of N2 concentration in the outlet stream. The influent gas entered into the reactor from the air inlet (numbered 3 in the figure) via a single tube with a diameter of 35 mm. The flow rate of the inlet gas stream (QG ) was continuously monitored by

Figure 1. Experimental set-up of the Jet-BioDeNOx system ((1) Mass flow controllers (for NO, Ar, and O2 gases), (2) gas-mixing chamber, (3) air/gas inlet, (4) JLBR, (5) liquid drain valve, (6) cooler, (7) liquid flow meter, (8) dissolved oxygen (DO), oxidation–reduction potential (ORP), pH, and temperature probes, (9) pump, (10) inlet feeding, (11) membrane system, (12) air/gas outlet, (13) GC, and (14) MRU VarioPlus Flue Gas Analyser and SerinusNOx Analyser, V, valve).

Downloaded by [Northeastern University] at 19:25 04 November 2014

1360

S. Durmazpinar et al.

an air rotometer (0–5 L/min). The composition of influent and effluent gas was determined by a gas analyser (MRU VarioPlus Flue Gas Analyser) and a gas chromatography with thermal conductivity detection (GC-TCD) method. The JLBR used in this study is made of Plexiglas and has a working volume of 20 L. The Jet-BioDeNOx system consists of a downflow jet-loop reactor connected to a membrane unit. The reactor consists of a cylindrical vessel (height 780 mm and inner diameter 102 mm) having a draft tube inside with a diameter of 35 mm. The jet located at the top of the reactor creates a downward directed two-phase flow inside the draft tube and disperses NO-containing gas sucked in through the gas tube located within the liquid jet. The liquid and the gas inside the draft tube first move downwards and then following the striking to the bottom of the reactor, rise within the annulus between the wall of the reactor and the draft tube. At the upper end of the draft tube, a part of the rising fluid is recycled back into the draft tube resulting in a re-dispersion of the bubbles and the biomass produced in the biological reaction.[18] Feeding medium was pumped with a peristaltic pump from the feed tank into the reactor. The internal recirculation flow rate (QRAS ) to the bioreactor was measured by a flow meter. The Jet-BioDeNOx reactor was operated in batch and continuous mode. The effect of system parameters (recirculation flow, mixed liquor-suspended solids (MLSS) concentrations, and levels of NO and O2 in the influent gas) on the NOx removal efficiency was studied in batch operation. For continuous operation of the reactor, flat sheet membrane was placed in the reactor line just before the JLBR to ensure a constant hydraulic retention time (HRT). Membrane module had two flat sheet membranes (15 × 45 cm2 area). The MP005 membrane obtained from MicroDyn-Nadir Co was used. 2.2. Starting-up bioreactor and feed composition The JLBR was inoculated with activated sludge obtained from the secondary sedimentation basin of Gebze Wastewater Treatment Plant in Turkey. In the first 30 days of adaptation, the JLBR was operated as a repeated-batch process to increase the amount of MLSS concentration by feeding prepared medium. The synthetic wastewater was prepared using sodium acetate, potassium nitrate, ammonium chloride, and potassium dihydrogen phosphate (KH2 PO4 ) as a source of carbon, nitrogen, and phosphorous, respectively. In the starting period, ammonium nitrate was also added to the medium to enhance the denitrifying population. The influent stock substrate solution contained 40,000 mg/L sodium acetate, 15,200 mg/L potassium nitrate, 7000 mg/L ammonium chloride, and 5000 mg/L KH2 PO4 . Required trace metals derived from CaCl2 (10 mg/L), MgSO4 ·7H2 O (50 mg/L), NaCl (50 mg/L), KCl (10 mg/L), CuSO4 ·5H2 O (0.1 mg/L), FeCl3 ·6H2 O (10 mg/L), ZnCl2 (0.25 mg/L), CoCl2 ·6H2 O (0.45 mg/L), and MnSO4 ·7H2 O (7 mg/L)

Table 1.

Operating conditions of JLBR.

System parameters

Unit

Value

Bioreactor volume Effective bioreactor volume Solids retention time HRT MLSS concentration Temperature Air to liquid ratio

L L days hour mg/L ◦C –

20 15 30 72 2700 ± 300 21 ± 0.5 0.07–0.12

were provided to the system. All reagents used were of analytical grade. The HRT was kept at 72 h and solids retention time (SRT) was 30 days. Excess sludge was withdrawn daily from the bioreactor (discharge rate = 0.5 L/day) to maintain the concentration of MLSS at 2700 ± 300 mg/L. The temperature of bioreactor was maintained around 21 ± 0.5◦ C by circulating distilled water through a stainless steel heat exchanger using a recirculating water bath (VWR-Circulator System 9006V12 V). Detailed operational conditions of the Jet-BioDeNOx system are given in Table 1. After 50 days, the system reached the steady-state conditions and the NOx removal experiments were started. 2.3. Analytical methods 2.3.1. Analyses in aqueous solution Measurements of chemical oxygen demand (COD), nitrite − (NO− 2 -N), nitrate (NO3 -N), and MLSS concentrations were performed according to the procedures described in the American Public Health Association Standard Methods.[19] Samples for soluble COD were obtained by filtration of the mixed liquor through filter paper (cellulose acetate) with mean pore size of 0.45 μm. The temperature, pH, dissolved oxygen, and oxidation–reduction potential in the bioreactor were measured online with a multi-parameter measurement device (Hamilton Electrochemical Sensors). 2.3.2. Analyses in gas streams The NO, NO2 , CO2 , and O2 gas concentrations of inlet and off-gases were determined by a multi-gas analyser (MRU VarioPlus Flue Gas Analyser), while N2 for off-gas was measured by the GC-TCD system (Agilent, 6890N Network). The N2 was separated with the Molesieve (MS) 5A Column (30 m × 530 μm × 25 μm) after the CO2 in the stream was trapped in a PLOT Q column and eluted from the system by using a column isolation valve to protect the MS column from CO2 . The separation was achieved by a fast temperature programme increasing from 50◦ C to 100◦ C by 30◦ C/min and from 100◦ C to 150◦ C by 10◦ C/min. 2.4.

Evaluation of NOx removal performance The NOx removal performance of the Jet-BioDeNOx process was investigated by conducting experiments at various

Environmental Technology

Downloaded by [Northeastern University] at 19:25 04 November 2014

operational conditions with the duration of 1–2 hours. Prior to introducing inlet gas to the reactor, formed N2 was swept away from the system by passing Ar gas and 1 L of liquid nutrient medium was supplied to the bioreactor with a peristaltic pump. The effects of various system parameters, such as liquid recycle flow rate, MLSS concentration, O2 content, and inlet NO concentration, on the removal performance were evaluated. 2.5. Species diagnosis The sludge sample was diluted and spread on nutrient agar plates and after incubating for 2 days the colonies with different morphologies were selected and isolated. The 16S rRNA genes of the isolates were amplified by using the eubacterial primers. Forward and reverse primers were used for sequencing the polymerase chain reaction products. Gene amplification and DNA sequencing were carried out at REFGEN Company (Ankara, Turkey). The deduced nucleotide sequence of the data was compared with the National Centre for Biotechnology Information (NCBI) database using the BLAST search available through the centre’s website (http://www.ncbi.nlm.nih.gov/BLAST).

Henry’s constants (KH ) for gas components.

Components

KH (M/atm)

NO2 Nitric oxide (NO) Dinitrogen oxide (N2 O) Nitrogen gas (N2 )

Modelling approach for NOx removal 2.6.1. Gas transfer

0.01200 0.00190 0.02500 0.00061

The ‘Ci ’ in the equation represents the gas concentration in the liquid.   Pi dCi (3) = KL a · − Ci dt KHe 2.6.2. Biological processes The nitrogen removal was formulated using a four-step denitrification process.[16] The model matrix is given in Table 3. Basically, nitrogen was denitrified by heterotrophic biomass using organic carbon (sole acetate, SS ) as an electron donor. The oxidized nitrogen serves as an electron acceptor. In the presence of organic carbon the oxidized nitrogen is reduced to a lower oxidation step as given in Equation (4).[15] The numbers given in brackets show the oxidation step of the nitrogen in each compound. − NO− 3 (+5) → NO2 (+3) → NO(+2)

2.6.

(4)

→ N2 O(+1) → N2 (0)

The model simulating the JLBR consists of two sub-models: (1) the gas transfer model and (2) the biological denitrification model. Basically, the gas transfer model was built on the basis of Henry’s law where the solubility of the gases in liquid merely depends on the partial pressure of the gas in the air (Equation (1)): Pi = kHe · Ci

Table 2.

1361

(1)

where Pi is the partial pressure of the gas (atm) and Ci is the concentration in the liquid. The Henry’s constants (KH ) specified for respective NO2 , NO, N2 O, and N2 components (Table 2) were made unitless using Equation (2) in order to calculate the interaction between gas and liquid phases during simulation. The conversion of the Henry’s equation into unitless form was carried out by using the expression as given below: KH kHe = (2) R·T where kHe is the unitless Henry’s constant (–); R is the universal gas constant (0.0820578 L atm/K mol), and T is the process temperature (K). Finally, the transfer rate was defined in accordance with the two-film theory suggested by Whitman.[20] The rate was basically controlled by the volumetric mass transfer rate KL a (1/day) as given in Equation (3). The KL a parameter is regarded as a process-tuning parameter; its adjustment was controlled by QG /QRAS mixing ratios at the jet ejector system. The mass transfer was modelled by using Equation (3).

In the availability of organic carbon, NO− 3 is reduced to nitrogen gas after the denitrification process via the consecutive reduction steps of heterotrophic denitrification. − According to the denitrification pathway, NO− 3 and NO2 are converted to NO via nitrate and nitrite reductase enzymes. Then the heterotrophic biomass denitrifies NO to N2 O and terminal nitrogen gas, sequentially.[17] Thus, the model is able to calculate denitrification using any oxidized nitrogen form in the availability of organic carbon (SS ) and active biomass (XH ). In this respect, the model given in Table 3 has 10 state variables, namely the SS , SNO3 , SNO2 , SNO , SN2 O , SN2 , SNH , XH , XS, and XP together with six biological processes. The biological processes consist of four denitrification (growth) processes and two additional processes representing the decay of heterotrophs (XH ) together with the enzymatic hydrolysis process. After the decay of XH , slowly biodegradable organic matter (XS ) which enables the COD turnover within the model is generated. The growth rates with respect to each denitrification step are calculated by the same rate equation controlled by maximum growth rate (μ) ˆ and half saturation constant for growth (KS ) parameters. However, only difference is that each reaction has a different correction factor (η) and affinity constant governed by Monod-type equations. The reactions of oxygen, such as (1) aerobic heterotrophic growth, (2) nitrification, (3) NO oxidation to NO2 , and (4) N2 O5 (via HNO) generation, were neglected in the model since the bulk is nearly anaerobic (at ORP< −300 mV).

1362

S. Durmazpinar et al.

Table 3.

Matrix representation of the biological model.

Process

SS

Growth on NO3 Growth on NO2 Growth on NO Growth on N2 O Hydrolysis

1 YH 1 − YH 1 − YH 1 − YH −

SNO3

SNO2

–α

α –β

SNO

SN 2 O

SN 2

β –β

β –β

β

–iXB

1

–iXB

1

–iXB

1

–iXB

1

XS

XP

iXB

–1

Rate SS SNO3 XH KS + SS KNO3 + SNO3 SS SNO2 nNO2 μˆ ANX XH KS + SS KNO2 + SNO2 SS SNO nNO μˆ ANX XH KS + SS KNO + SNO SS SN 2 O nN2 O μˆ ANX XH KS + SS KN2 O + SN2 O XS /XH kh XH KX + XS /XH nNO3 μˆ ANX

–1 1–fEX

fEX

bh X H

1 − YH 1 − YH ; β, . 1.717 · YH 0.571 · YH

In the simulation study, first the system was run with carbon feed (acetate) and NO− 3 addition for 2 months. After obtaining steady-state biomass concentration in the simulations for Jet-BioDeNOx, the batch experiments conducted with NO gas were simulated. The bulk COD, off-gas NOx , biomass concentration in terms of mixed liquor volatile suspended solids (MLVSS) were calculated using AQUASIM software developed by Reichert et al.[21] 3. Results and discussion 3.1. NOx removal performance 3.1.1. Effect of recycle flow rate and MLSS The reactor was operated at different QRAS values and MLSS concentration levels for inlet NO concentration of 1000 ppm. Increasing QRAS resulted in an increase in the NO removal efficiency. Figure 2 illustrates that the removal efficiency increased from 49% to 83% when the QRAS increased from 1.0 to 1.8 m3 /h for the MLSS level of 5400 mg/L since this decrease in QG /QRAS ratio enhances the mass transfer conditions (higher KL a values) in the JLBR. Both the variations in MLSS and recycle flow rate seem to affect the NO removal efficiency. As seen from Figure 2, increase in MLSS from 2520 to 5400 mg/L results in an increase in the NO removal efficiency for both QRAS values. System reaches up to 90% of NO removal efficiency with QRAS 1.8 m3 /h at a high MLSS concentration. Increase in this removal efficiency might be stimulated by the unique hydrodynamic conditions of JLBR which produce micro-organism flocs with smaller sizes. 3.1.2. Effect of oxygen concentration The effect of O2 content in the inlet gas was evaluated by preparing synthetic waste gas streams containing initially 1150 ppm NO and different O2 contents (0%, 2%, and 8% O2 ). Depending on the O2 content, NO and NO2

90 80 NO removal efficiency (%)

Downloaded by [Northeastern University] at 19:25 04 November 2014

XH

1

Decay Note: α,

SNH

Q = 1.0 m3/h Q = 1.8 m3/h

70 60 50 40 30 20 10 0 2000

2500

3000 3500 4000 4500 5000 MLSS concentration (mg/L)

5500

Figure 2. The effect of liquid recycle flow rate and MLSS concentrations on NO removal efficiency.

concentrations of the inlet gas changed due to reaction of NO oxidation to NO2 with O2 . Maximum NO2 concentration was observed as 123 ppm in the inlet gas containing 8% O2 . The NO removal efficiency was calculated to be between 89% and 92% for three different O2 contents. This suggests that the NO removal performance of the system was not affected significantly by the O2 content of the influent gas. In other words, the Jet-BioDeNOx system was not sensitive to O2 concentration within the tried concentration range compared with the BioDeNOx systems which are very sensitive to O2 content in the inlet gas due to its reaction with Fe(II)EDTA−2 .[22,23] The JLBR surpasses the other reactors with two main features: (1) creating smaller size of micro-organism flocs within the bioreactor and (2) unique hydrodynamic conditions favouring high mass transfer rate. The results clearly show that these characteristics of JLBR might be acting as a barrier to inhibition of denitrification by O2 content up to 8%.

Environmental Technology 3.1.3. Effect of inlet NO concentration The Jet-BioDeNOx system’s performance on the basis of the different NO loadings was evaluated by introducing 500, 1000, 2000, and 3000 ppm NO as inlet gas. The NO removal

1363

efficiencies and corresponding N2 concentration in the offgas for these experiments are shown in Figure 3. Although the NO removal efficiency varied between 80% and 92%, no correlation was observed between removal efficiency

100

1600 NO N2

1400

1000

60

800 40

600 400

Off gas N2 concentration

NO removal efficiency (%)

1200

20 200 0

0 500

1000

1500

2000

2500

3000

Inlet NO concentration Figure 3.

NO removal efficiency and outlet N2 concentrations for different inlet NO contents.

(a) 25 Volume (%)

20 15 10 5 0 0.1

1

10 100 Diameter (nm)

1000

10000

(b) 200000 Intensity (kcounts)

Downloaded by [Northeastern University] at 19:25 04 November 2014

80

150000 100000 50000 0 -200

Figure 4.

-100

0 Zeta Potential (mV)

The floc size (a) and the zeta potential (b) of the Jet-BioDeNOx sludge.

100

200

1364

S. Durmazpinar et al.

Downloaded by [Northeastern University] at 19:25 04 November 2014

and NO inlet concentration at the selected operating conditions of the reactor (COD: 5000 mg/L and QRAS : 1 m3 /h). The N2 concentration in the off-gas accordingly increases as the inlet NO concentration and the removed NO amount increase. We can conclude that the system can tolerate concentration variation in the inlet gas and exhibits a stable performance in contrast to some other biological NOx control systems reporting significant decrease in the efficiency as the inlet NO load increases.[24] 3.2. Characterization of sludge Reactor biomass was investigated in terms of volumetric particle size distribution (Figure 4(a)) and zeta potential (Figure 4(b)). The particle volume distribution has a mean diameter of approximately 1000 nm. It is well known that the size of micro-organism flocs is different for the jetloop reactor and the activated sludge process. The previous studies indicate that it is difficult to measure the size of activated sludge flocs since they are heterogeneous and the size range is very broad.[25] However, the activated sludge has been measured to have a bimodal particle size distribution with one group in the range of 0.5–5 μm (also known as un-flocculated dispersed cells) and another group in 25– 2000 μm range.[26,27] In this study, floc size in the jet-loop reactor was measured in the range of 500–2000 nm (0.5– 2 μm), which is much smaller than floc size in the classical activated sludge. However, this smaller size of the biomass provides higher surface area which in turn implies a positive effect on the NO removal efficiency. In other words, this lower size generated in the Jet-BioDeNOx process contributes to a high NOx removal efficiency without adding chelate-forming chemicals. The zeta potential is in the range of 0 ± 60 mV (Figure 4(b)). The surface charge of the bacteria is generally negative depending on the exopolymeric substances. In the Jet-BioDeNOx process, on the other hand, it varied between negative and positive regions due to ionizable groups found in the exopolymeric substances such as carboxylic, sulphate, phosphate, and amphoter structures like proteins. 3.3. Species diagnosis There are many micro-organism species known to be denitrifiers (Pseudomonas, Paracoccus, Bacillus, and Thiobacillus) under anaerobic conditions.[28] To investigate the denitrifying species adapted to the harsh hydrodynamic environment of the Jet-BioDeNOx process, species diagnosis was performed on the samples taken from the reactor. According to the 16S rRNA test, Alcaligenes faecalis, Staphylococcus hominis, Staphylococcus haemolyticus, and Corynobacterium glutamicum species were identified with phylogenetic similarity above 99%. Alcaligenes are known as denitrifying bacteria.[29] S. hominis was reported to convert the nitrate to nitrogen gas.[28] Harbi et al. [30] showed that Staphylococcus epidermis

possess the genes which synthesize reductase for nitrogen compounds. C. glutamicum was also reported as a facultative anaerobe and reduces the nitrate to nitrite.[31] These results show that bacteria species found in the reactor biomass are the species that contribute to the denitrification process and formed an effective consortium in NO reduction to N2 gas. Reached MLSS levels and a high NOx removal efficiency support that this consortium has a high specific production rate at the reactor operating temperatures of 20–22◦ C and were not adversely affected from the hydrodynamic conditions of the JLBR.

3.4. Modelling and simulation results In the modelling part of the study, first, a steady-state simulation was carried out in order to obtain a stable biomass concentration characterizing the conditions of long-term operation of the Jet-BioDeNOx reactor. The simulation study revealed that the COD in the bulk remained around 7000 mg/L after the consumption of NO− 3 by yielding MLSS concentration of 3000 mgVSS/L in the bioreactor. These results are compatible with the experiments evidenced during steady-state operation. The second set of simulations was devoted to mimic the NO gas removal rate for two different KL a values of 250/day and 530/day, respectively.

Figure 5. Modelling of off-gas NO for (a) experiment KLa = 250/day and (b) experiment KLa = 530/day.

Environmental Technology

Downloaded by [Northeastern University] at 19:25 04 November 2014

Figure 6. Dependency of KL a on gas to mixed liqour recycle ratio (QG /QRAS ).

Basically, the experiments were dedicated to simulate off-gas NO during sparkling the reactor with NO gas at steady-state conditions. During the simulation, the concentration of NO gas was adjusted to 1100 ppm (1.47 mg/L) at the inlet gas stream of the JLBR. Figures 5(a) and 5(b) illustrate the measured and simulated NO and N2 concentrations in the off-gas. It should be noted that the KL a parameter was adjusted by setting the QRAS to 1.0 and 1.8 m3 /h for KL a of 250/day and 530/day, respectively. The volumetric mass transfer coefficients are in agreement with the clean water tests conducted previously (results not shown). On the basis of modelling results, the removal efficiency of NO in the gas phase was increased from 65% to 81% at a higher KL a level around 530/day. The dependency of KL a with respect to various QG /QRAS ratios can be characterized as shown in Figure 6. Increase in QG /QRAS ratio resulted in a decrease in transfer rate. The model parameters used in simulation were given in Table 4. The parameters used in simulations indicate that the default values resulted in a reasonable model fit (Figure 5) without any further requirement for recalibration. The only difference is that much lower affinity constants (KNO and KN2 O ) were required to maintain biological reactions in the liquid phase at low concentration of NO. It is argued that Table 4.

1365

lower MLSS concentrations to a definite level favour much better mass transfer conditions due to smaller floc structure and lower density of mixed liquor.[32,33] The results found in this paper are compatible with the literature where much lower affinity constants are applicable for MBR systems operated at lower MLSS concentrations.[34] Operation of the MBR system at lower ML(V)SS concentration around 3000 mg/L yielded lower half saturation constants, in turn having a better transfer conditions into biological flocs. Nevertheless, the removal of the poorly soluble NO gas could be achieved efficiently in the MBR by enhanced mass transfer conditions adjusted by floc size density. As shown in Figure 4(a), the mean diameter of floc size distribution was in the order of 1000 nm (1 μm) which was 100 folds of the conventional activated sludge systems. In this respect, the high mass transfer conditions prevailed in the MBR provided a competitive advantage on removing the poorly soluble NO gas in the availability of substrate.

4. Conclusion A new biological NOx control technology based on denitrification of the nitrogen oxide gases in a jet-loop reactor has been designed. This Jet-BioDeNOx system can achieve over 90% of NO removal efficiency at the temperature of 21 ± 0.5◦ C. The removal efficiency increases with the ratio of liquid recycle flow rate to gas flow rate. The system is not sensitive to O2 content (up to 8%) in the inlet gas. Modelling studies reveal that the Jet-BioDeNOx process enables to remove the poorly soluble NO gas efficiently, using the enhanced mass transfer conditions. In this respect, the high mass transfer conditions prevailed in the reactor provided a competitive advantage on removing NO gas in the availability of substrate without any requirement of chelating chemicals as in the conventional BioDeNOx process. In addition, the mass transfer conditions could be adjusted by the ratio of gas flow rate to recycled liquid flow rate.

Model parameters used in simulations.

Model Parameters Heterotrophic yield coefficient of denitrifiers (YHNO ) Lysis rate for heterotrophs (bH ) Maximum growth rate for heterotrophs(μˆ Hmax ) Half saturation constant for growth of heterotrophs(KS ) Anoxic correction factor for growth on NO3 (ηNO3 ) Anoxic correction factor for growth on NO2 (ηNO2 ) Anoxic correction factor for growth on NO (ηNO ) Anoxic correction factor for growth on N2 O (ηN2 O ) Half saturation constant for growth of autotrophs(KNH ) Half saturation constant of NO3 for denitrifiers(KNO3 ) Half saturation constant of NO2 for denitrifiers(KNO2 ) Half saturation constant of NO for denitrifiers(KNO ) Half saturation constant of N2 O for denitrifiers(KN2 O )

Unit

This study

Range

gCOD/gCOD 1/day 1/day mgCOD/L – – – – mgN/L mgN/L mgN/L mgN/L mgN/L

0.6 0.45 5.0 5 1.0 1.0 1.0 0.8 0.01 0.2 0.05 0.01 0.01

0.57–0.65 0.3–0.6 3.0–6.0 5–20 0.28–1.0 0.16–1.0 0.35–1.0 0.35–1.0 0.1–1.0 0.2 0.2 0.05 0.05

Reference

[13] [34]

[14]

1366

S. Durmazpinar et al.

However, more studies should be devoted to the optimization of the Jet-BioDeNOx process by using real NO source for the long-term operation and economic analyses of the system should be performed considering the actual process conditions. Acknowledgements This study was financially supported by the TUBITAK, the Scientific and Technological Research Council of Turkey [Project No: 110Y031]. One of the authors (Dr. Nadir Dizge) would like to thank Ms. Ceyda Ozaksoy for valuable support and encouragement through out this study.

Downloaded by [Northeastern University] at 19:25 04 November 2014

References [1] Graedel TE, Crutzen PJ. The changing atmosphere. Sci Am. 1989;261:58–68. [2] Burnett RT, Cakmak S, Brook JR. The effect of the urban ambient air pollution mix on daily mortality rates in 11 Canadian cities. Can J Public Health. 1998;89:152–156. [3] Niu H, Lend DYC. A Review on the removal of nitrogen oxides from polluted flow by bioreactors. Environ Rev. 2010;18:175–189. [4] Shareefdeen Z, Singh A. Biotechnology for odor and air pollution control. Berlin: Springer; 2008. [5] Guo B, Ren A, Zhao D, Dongyue Y, Wenxia Z. Removal of NO(x) emitted from nitric acid production with biological packing tower. 3rd International conference on Bioinformatics and Biomedical Engineering; 2009. Vol. 1–11; p. 3633–3636. [6] Jiang R, Huang SB, Chow AT, Yang J. Nitric oxide removal from flue gas with a biotrickling filter using Pseudomonas putida. J Hazard Mater. 2009;164:432–441. [7] Kumar A, Ergas SJ, Yuan X, Fitch M, Min KN, Dewulf J, Van Langenhove H. Modeling of a hollow fiber membrane biofilm reactor for nitric oxide removal: model development and experimental validation. J Chem Technol Biotechnol. 2010;85:423–428. [8] Liang W, Huang SB, Yang YY, Jiang R. Removal of nitric oxide in a biotrickling filter under thermophilic condition using Chelatococcus daeguensis. J Air Waste Manage Assoc. 2012;62:509–516. [9] Buisman CJ, Dijkman H, Verbaak PL, Den Hartog AJ. Process for purifying flue gas containing nitrogen oxides. United States patent US 5891408. 1999. [10] Li W, Wu CZ, Shi Y. Metal chelate absorption coupled with microbial reduction for the removal of NOx from flue gas. J Chem Technol Biotech. 2006;81:306–311. [11] Li W, Wu CZ, Zhang SH, Shao K, Shi Y. Evaluation of microbial reduction of Fe(III)EDTA in a chemical absorption— biological reduction integrated NOx removal system. Environ Sci Technol. 2007;41:639–644. [12] Van der Maas P, Van den Bosch P, Klapwijk B, Lens P. NOx removal from flue gas by an integrated physicochemical absorption and biological denitrification process. Biotechnol Bioeng. 2005;90:433–441. [13] Dold PL, Ekama GA, Marais GR. A general model for the activated sludge process. Prog Wat Tech. 1980;12:47–54. [14] Henze M, Gujer W, Mino T, Van Loosdrecht MCM. Activated sludge models: ASM1, ASM2, ASM2d and ASM3. London (UK): IWA Publishing; 2000. (Scientific and Technical Report No: 9) [15] Wild D, Schulthess R, Gujer W. Structured modelling of denitrification intermediates. Water Sci Technol. 1995;31: 45–54.

[16] Hiatt WC, Grady CPL. Application of the activated sludge model for nitrogen to elevated nitrogen conditions. Water Environ Res. 2008;80:2134–2144. [17] Hiatt WC, Grady CPL. An updated process model for carbon oxidation, nitrification and denitrification. Water Environ Res. 2008;80:2145–2156. [18] Farizoglu B, Keskinler B, Yildiz E, Nuhoglu A. Simultaneous removal of C, N, P from cheese whey by jet loop membrane bioreactor (JLMBR). J Hazard Mater. 2007;146:399–407. [19] APHA. Standard methods for the examination of water and wastewater. 21st ed. Washington, DC: American Public Health Association/American Water Works Association/ Water Environment Federation; 2005. [20] Whitmann WG. The two-film theory of absorption. Chem Met Eng. 1923;29:146–148. [21] Reichert P, Ruchti J, Simon W. AQUASIM 2.0. Swiss Federal Institute for Environmental Science and Technology (EAWAG), CH-8600 Duebendorf; 1998. Software available from: http://www.eawag.ch/forschung/siam/ software/aquasim/index [22] Apel WA, Barnes JM, Barret KB. Removal of nitrogen oxides from gas streams using biofiltration. J Hazard Mater. 1995;41:315–326. [23] Van der Maas P, Van den Bosch P, Klapwijk B, Lens P. NOx removal from flue gas by an integrated physicochemical adsorption and biological denitrification process. Biotechnol Bioeng. 2004;90:433–441. [24] Juna C, Yifenga J, Haoleia S, Jianmenga C. Effect of key parameters on nitric oxide removal by an anaerobic rotating drum filter. Environ Technology. 2008;29:1241–1247. [25] Wilen BM, Balmer P. The effect of dissolved oxygen concentration on the structure, size and size distribution of activated sludge flocs. Water Res. 1999;33:391–400. [26] Parker DS, Kaufman WJ, Jenkins D. Physical conditioning of activated-sludge floc. J Water Pollut Control Fed. 1971;43:1817–1833. [27] Wu J, Jiang X, Wheatley A. Characterizing activated sludge process effluent by particle size distribution, respirometry and modeling. Desalination. 2009;249:969–975. [28] Constantin H, Raoult S, Montigny W. Denitrification of concentrated industrial wastewater: microorganism selection and kinetic studies. Environ Technol. 1996;17: 831–840. [29] Maier RM. Biogeochemical cycling in environmental microbiology. In: Maier RM, Pepper IL, Gerba PC, editors. Academic Press; 2009. p. 309. [30] Harbi B, Chaieb K, Jabeur C, Mahdouani K, Bakhrouf A. PCR detection of nitrite reductase genes (nirK and nirS) and use of active consortia of constructed ternary adherent staphylococcal cultures via mixture design for a denitrification process. World J Microbiol Biotechnol. 2010;26:473–480. [31] Nishimura T, Vertès A, Shinoda Y, Inui M, Yukawa H. Anaerobic growth of Corynebacterium glutamicum using nitrate as a terminal electron acceptor. Appl Microbiol Biotechnol. 2007;75:889–897. [32] Manser R, Gujer W, Siegrist H. Consequences of mass transfer effects on the kinetics of nitrifiers. Water Res. 2005;39:4633–4642. [33] Insel G, Hocaoglu SM, Çokgor EU, Orhon D. Modelling the effect of biomass induced oxygen transfer limitations on the nitrogen removal performance of membrane bioreactor. J Membrane Sci. 2011;368:54–63. [34] Barker PS, Dold PL. General model for biological nutrient removal activated sludge systems: model presentation. Water Environ Res. 1997;69:969–984.