A

qlicuoS

rt©

K

1

haonrcDwim[tvfln

0d

Enzyme and Microbial Technology 40 (2007) 1627–1634

Kinetics of cross-inhibited denitrification of a high load wastewater

Oscar Soto a, Estrella Aspe b, Marlene Roeckel b,∗a Escuela de Ciencias Forestales, Universidad Catolica de Temuco, Casilla 15-D, Temuco, Chile

b Departamento de Ingenierıa Quımica, Universidad de Concepcion, Casilla 160-C, Concepcion, Chile

Received 22 December 2005; received in revised form 16 November 2006; accepted 16 November 2006

bstract

Batch denitrification of a synthetic saline medium at 37 ◦C, pH 7.5 was studied. A kinetic expression and an inhibition function which ade-uately expresses inhibition and cross-inhibition by nitrate and nitrite in denitratation and denitritation steps occurring in high organic and nitrateoad wastewater treatment was determined. Denitratation and denitritation rates were measured independently. Monod- and Haldane-type kinet-cs were used for denitratation and denitritation, respectively. The maximum specific rate, kNO3 = 12.7 mg NO3-N/g VSS h, and the saturationonstant, KSNO3

= 0.47 mg NO3-N/L were calculated for denitratation. The maximum specific rate, kNO2 = 13.8 mg NO2-N/g VSS h, the sat-ration constant KSNO2

= 0.36 mg NO2-N/L and the inhibition constant KI-NO2 = 906 mg NO2-N/L were calculated for denitritation. Amongther functional forms the Levenspiel inhibition model for denitratation and denitritation kinetics was proposed. The Levenspiel constants were

NO2-M = 1149.4 mg NO2-N/L, SNO3-M = 34.7 mg NO3-N/L, β1 = 1.578 and β2 = 1.005, respectively. The fitted experimental values show a goodepresentation of the denitrification of effluents with high nitrogen load containing both nitrate and nitrite. Parametric sensitivity analysis showhat inhibition functions (fi) and dimensionless constants (βi) greatly affect predicted nitrite and nitrate consumption rates.

2006 Elsevier Inc. All rights reserved.

m

Mtpuc

laiii

1

r

eywords: Denitrification; Denitratation; Denitritation; Kinetics; Saline mediu

. Introduction

Wastewaters from salmon and other fish processes featureigh salt, proteins, nitrate and nitrite concentrations [1,2], whichre treated stepwise anaerobically and aerobically to reducerganic matter, producing nitrate and nitrite. Often, resultingitrite and nitrate concentrations exceed levels imposed by envi-onmental legislation. In this context, biological denitrificationould be used to reduce nitrogen oxides to acceptable levels.enitrification is a process of anoxic microbial respiration inhich nitrate is reduced to molecular nitrogen with nitrite as an

ntermediate product [3]. The electron donor is generally organicatter, which may be obtained from the treated wastewater

4]. Denitrification systems could be used to treat anaerobicallyreated fishery wastewaters. Such effluents contain significantolatile fatty acids loads [5], which are known electron donors

or denitrification [6]. Denitrification includes two major bio-ogical processes, namely, denitratation (i.e. nitrate reduction toitrite) and denitritation (i.e. nitrite reduction to gaseous N2).

∗ Corresponding author. Tel.: +56 41 2203663; fax: +56 41 2243750.E-mail address: mroeckel@udec.cl (M. Roeckel).

nnpfiit

141-0229/$ – see front matter © 2006 Elsevier Inc. All rights reserved.oi:10.1016/j.enzmictec.2006.11.014

oreover, nitrate and nitrite have been found to inhibit deni-rification systems [3,7–13]. Although detailed information isrovided in the literature for the inhibition on each individ-al denitrification step, little attention has been paid to possibleross-inhibition of reactants.

The objective of the present study is to experimentally ana-yze the inhibitory effects of nitrate and nitrite, at high organicnd nitrate load and saline conditions. Denitratation and den-tritation rates are assessed separately in order to determinenhibition and cross-inhibition by nitrate and nitrite. A suitablenhibition model is proposed.

.1. Mathematical modeling

Denitrification processes involve the initial nitrate (NO3−)

eduction to nitrite (NO2−), followed by further reduction to

itric oxide (NO), nitrous oxide (N2O) and finally molecularitrogen (N2). Although denitratation and denitratation are com-lex biological processes, the overall denitrification pathway

rom nitrate to gaseous nitrogen may be simplified as illustratedn Fig. 1. Based on experimental findings, nitrite and nitratenhibition mechanisms proposed in this work are included inhis figure.

1628 O. Soto et al. / Enzyme and Microbial T

Nomenclature

f1 inhibition function on denitratation for nitritef2 inhibition function on denitritation for nitratei subunitKI-NO2 Haldane inhibition constant in denitritation

(mg NO2-N/L)kNO2 maximum specific nitrite reduction rate

(mg NO2-N/g VSS h)kNO3 maximum specific nitrate reduction rate

(mg NO3-N/g VSS h)KSNO2

half-saturation constant for nitrite (mg NO2-N/L)KSNO3

half-saturation constant for nitrate (mg NO3-N/L)rNO2 specific nitrite reduction rate (mg NO2-N/g VSS h)rNO3 specific nitrate reduction rate (mg NO3-N/g VSS h)r∗

NO2reference specific nitrite reduction rate(mg NO2-N/g VSS h)

r∗NO3

reference specific nitrate reduction rate(mg NO3-N/g VSS h)

SI-NOx non-competitive inhibition constant (mg NOx-N/L)SNOx-M maximum concentration at which specific

reduction rate is zero (mg NOx-N/L)SNOx nitrate or nitrite concentration (mg NOx-N/L)X biomass concentration (g VSS/L)

Greek symbolsβ inhibition dimensionless constant for Eq. (3)

bMsnmi

iAt

Fs

dtatinaianp

vTfw

r

r

wdtHtmtw

•

•

i

γ i inhibition dimensionless constant for Eq. (4)

Several attempts have been made to model the inhibition ofoth steps. Nitrate reduction has usually been described by aonod-type kinetics which means substrate limiting, but no sub-

trate inhibition [7,14,15]. Several authors have demonstrateditrite inhibition in the denitratation stage [7,8,16]; affectingicrobial growth. These experimental evidences should be

ncluded in the denitratation kinetics.On the other hand, Haldane kinetics, representing substrate

nhibition, has been proposed for nitrite reduction [10,13,17].lmeida et al. [7] and Wang et al. [13] proposed a nitrate inhibi-

ion model in denitritation for Pseudomonas fluorescens and P.

ig. 1. Denitrification process with inhibition of denitritation and denitratationteps.

•

N

amLarripm

echnology 40 (2007) 1627–1634

enitrificans, respectively. These models represent adequatelyhe processes, but each step – denitratation and denitritation –re separately studied. Glass and Silverstein [10] establishedhe existence of two different maximum denitritation rates, onen the presence of nitrate (k′

2), and a second one (k2) once theitrate has been completely reduced. Huang et al. [11] presented“distributed fraction of nitrate-reductase enzyme” (f), which

s an empirical solution similar to the Glass and Silverstein [10]pproach. Finally, Schonharting et al. [12] applied a “two-siteon-competitive” model without showing experimental kineticarameter evaluation.

Our preliminary work showed that these models representery well nitrite and nitrate consumption at low concentrations.herefore, this work proposes the following integrated model

or the combined denitratation and denitratation of species in aide range of concentrations:

NO3 =[

kNO3SNO3

KSNO3+ SNO3

]f1 (1)

NO2 =[

kNO2SNO2

KSNO2+ SNO2 + S2

NO2/KI-NO2

]f2 (2)

here f1 and f2 represent the inhibition functions of nitrite onenitratation, and of nitrate on the denitritation kinetics, respec-ively. According to the previously cited literature, Monod andaldane kinetic models were used for denitratation and denitri-

ation kinetics, respectively. Thus, this work will focus on theathematical corrections that the inhibition of species poses on

hese models. The following inhibitory models were used in thisork [18–20]:

Levenspiel’s model:

fi =(

1 − SNOx

SNOx-M

)βi

(3)

Luong’s model:

fi = 1 −(

SNOx

SNOx-M

)γi

(4)

Non-competitive model:

fi =(

SI-NOx

SI-NOx + SNOx

)(5)

In Eqs. (3)–(5) when i = 1, NOx is NO2; when i = 2, NOx isO3.Most of the kinetic models have an empirical nature and

re based on models describing enzymatic inhibition. Luong’sodel [18] is a modification of the general model proposed byevenspiel [19], while the non-competitive model constitutesn extension of the non-competitive inhibition of enzymaticeactions [20]. The goal of these models is to correct the

ate of substrate consumption by a factor that depends on thenhibitor concentration; however, due to their simplicity, therevious models present some limitations. The non-competitiveodel, although based on reaction enzyme mechanisms has the

bial T

dciwewlobita

tttd

2dim

2

2

fitTwws3eKyeC7

a8wda

2

coee

c

sitoibwepubwwsw

fl5Ni

tfit

ta(c

emE

eswioa9Ettf1caopt

2

scsoaTEt

O. Soto et al. / Enzyme and Micro

rawback that predicts substrate consumption at any inhibitoroncentration, a situation that is far from real as it will be provedn this work that exists a limit concentration of the inhibitor athich substrate consumption stops. Although Luong’s and Lev-

nspiel’s model overcome this limitation, as they predict thathen SNOx = SNOx-M the substrate consumption is halted, γ

acks any physical significance. And, because the equations aref an exponential type, a wide inhibitor concentration range muste covered, even larger than the extremes ones actually foundn the industry, to reach reliable values for the constants. Thesehree models were chosen comparing the inhibition functionsnd the rate decay in preliminary essays (data not shown).

As shown in Fig. 1, nitrite inhibits both steps: denitrata-ion and denitritation; additionally, nitrate’s inhibitory effect onhe denitritation step is included. Considering Eqs. (1) and (2),he variation of nitrate and nitrite concentrations during batchenitrification is calculated as follows

dSNO3

dt= −rNO3X (6)

dSNO2

dt= (rNO3 − rNO2 )X (7)

Chilean fishery wastewaters feature a saline content around.4% (w/w). Experimental results presented here were con-ucted here at such salt level and no explicit term for anynhibitory effect [21–23] was included in the proposed kinetic

odel.

. Materials and methods

.1. Inoculum and culture media

A synthetic culture medium was used, which was optimal for denitri-cation [24]. This media contains most compounds found in anaerobically

reated fishing effluents with high organic and nitrogen load (unpublished data).herefore, to mimic the anoxic treatment of fishery effluents this mediumas modified by the addition of 24 g/L NaCl, similar to the fish processingastewater concentration. Acetic acid was added as an electron donor, and potas-

ium nitrate as an electron acceptor. The C/N ratio was established between0 and 300, in order to assure excess in electron donor capacity in all thexperiments. The medium composition was as follows: KH2PO4: 3.00 g/L;

2HPO4: 3.00 g/L; NaHCO3: 0.40 g/L; MgSO4·7H2O: 0.005 g/L; NaCl: 24 g/L;east extract: 0.05 g/L; peptone: 0.12 g/L; glacial acetic acid: 1.22 mL/L; tracelements: 1.00 mL/L (FeSO4·7H2O: 1.096 g/L; CaCl2: 0.5 g/L; KCl: 0.5 g/L;oCl2: 0.1 g/L); KNO3: 11–290 mg/L. The medium pH value was adjusted to.5. The reactor bottles were incubated at 37 ◦C.

The inoculum was a sludge obtained from an anaerobic pilot fishery wastew-ter treatment plant. This sludge was adapted to anoxic conditions over an-month period, prior to initiation of experimentation. The adaptation reactoras fed continuously, at a hydraulic residence time of 1-day using the above-escribed medium and a NO3-N concentration of 100 mg/L. The pH was 7.5nd temperature 37 ◦C during adaptation.

.2. Batch culture conditions

Kinetic parameter determination was made using data from batch reactorsontaining suspended cultures and fed with nitrate or nitrite. The discontinu-

us pulse method was chosen to preserve the culture history and to be able tostablish the diverse effects of the species participating in the process during thexperimental period.

All denitrification assays were carried out in 300-mL Erlenmeyer flasks, eachontaining 250-mL of reaction mixture, and agitated at 200 rpm on an orbital

T

fma

echnology 40 (2007) 1627–1634 1629

haker (New Brunswick Scientific Co., USA, model G-24). All essays were donen triplicate. To determine individual parameters in denitratation and denitrita-ion, the sludge concentration was 900–1800 mg VSS/L and the inhibitor effectf nitrite and nitrate was avoided using concentrations lower than 50 mg/L inndividual essays. For model verification in the complete denitrification process,iomass concentrations of up to 6 g VSS/L were employed. Anoxic conditionsere maintained in all assays, with an initial N2 gassing for 5 min, and sealing

ach flask with a butyl stopper. The temperature of all assays was 37 ◦C and theH was initially adjusted to 7.5 with NaOH addition. This temperature level wassed in order to obtain standard kinetic parameters and to compare with otheriological related treatment processes, such as anaerobic digestion at 37 ◦C. pHas monitored during the process at 1 h intervals. The acetic acid concentrationas maintained in excess (1200 mg COD/L). Given the nature of the carbon

ource and the predetermined range (C/N = 30–300), complete denitrificationas assured without ammonification [25,26].

Nitrite (NO2-N) and nitrate (NO3-N) concentrations were measured usingow injection analysis (FIA) (Foss Tecator, Sweden). Nitrite is measured at40 nm as the colored product after reaction with sulfanilamide diazotized with-(1-naphtyl)-ethylenediamine dihydrochloride; and nitrate is reduced to nitrite

n a cadmium column, and measured as described for nitrite [27].The maximum rates (k) were initially determined to minimize the data fit-

ing error following the proposed numerical methods. Once the value of k wastted, the half-saturation constants KS were estimated. Denitritation parame-

ers (kNO2 and KSNO2), were obtained using initial nitrite concentrations of less

han 40 mg NO2-N/L in order to avoid substrate inhibition (S2/K1 ≈ 0 in Eq. (2))nd without nitrate (f2 = 1 in Eq. (2)). For the determination of KI-NO2 in Eq.2), previously determined constants kNO2 and KSNO2

were used. Initial nitriteoncentrations were 25, 170, 390, 564, 624, 762, 396, and 1192 mg NO2-N/L.

Denitratation kinetic constants (kNO3 and KSNO3 ) were determined fromxperiments featuring NO2-N initial concentrations below 50 mg/L, in order toinimize the inhibitory effect of nitrite produced on nitrate reduction (f1 ≈ 1 inq. (1)). Results obtained were fitted to the numerical solution of Eq. (6).

The cross-inhibition pattern was determined by performing batch-pulsexperiments at different initial NO3

− concentrations and following the con-umption rates for different injected NO2

− concentrations. Pulse experimentsere carried out in order to determine f1 in Eq. (1) by injection of a nitrite solution

nto the batch assays during nitrate reduction, measuring the change in the slopef the nitrate concentration-time curve. Initial concentrations of nitrate werepproximately 50 and 200 mg NO3-N/L, and concentrations of 38, 73, 117, 397,06, and 1149 mg NO2-N/L were used to obtain the inhibition constants (β1 forq. (3); γ1 for Eq. (4) and SI-NO2 for Eq. (5)). To determine f2 in Eq. (2), data were

aken from the complete denitratation–denitritation process. Initially and whenhe nitrate was depleted, the denitritation rate was measured (SNO3 = 0). Then,

2 was determined using both rates. Initial nitrate concentrations of 1, 10, 13, 15,8, 21, 23, 25, 35, 68 and 115 mg NO3-N/L were used to obtain the inhibitiononstants (β2 for Eq. (3); γ2 for Eq. (4) and SI-NO3 for Eq. (5)). Maximum nitratend nitrite concentrations (SNO3-M and SNO2-M) for Eqs. (3) and (4) were directlybtained from experimental data. Solving Eqs. (6) and (7) using the determinedarameters, nitrite and nitrate concentration were estimated and compared withhe experimental data, thus the model of the combined process was verified.

.3. Kinetic parameters estimation

Table 1 summarizes the experiments to obtain the kinetic constants. Con-tants for denitratation and denitritation were determined using different initialoncentrations of substrates and measuring initial consumption rates. Initiallopes were plotted against initial substrate concentration. Afterwards, the resultsbtained from the batch assays were used to obtain best-fit solutions to Eqs. (6)nd (7); therefore, the denitratation and denitritation constants were confirmed.he fixed path 4–4 method of Runge–Kutta was used to resolve the differentialqs. (6) and (7). Fitting of the experimental data to the corresponding solu-

ions was carried out using the non-linear method of Levenberg–Marquardt.

he algorithms were programmed using MATLAB software, version 4.2.

Reaction rates are relatively low and the sensibility of the analysis methodsor the determination of nitrogen compound using FIA are the same order ofagnitude than the saturation constants (error 3%, sensibility, 0.1 mg/L). In

ddition, the history of the culture is known and constant, as inocula was obtained

1630 O. Soto et al. / Enzyme and Microbial Technology 40 (2007) 1627–1634

Table 1Summary of experimental conditions for the parameters determination

Assay Conditions Obtained parameter

Nitrite oxidation without inhibition Initial nitrite concentrations between 1.7 and40 mg NO2-N/L (S2/K1 ≈ 0 in Eq. (2)); nitrateconcentration = 0 (f2 = 1 in Eq. (2))

kNO2 and KSNO2

Inhibition by nitrite of nitrateoxidation

kNO2 and KSNO2values already determined; initial

nitrite concentrations: 25, 170, 390, 564, 624, 762, 396,and 1192 mg NO2-N/L

KI-NO2

Nitrate oxidation without inhibition Initial concentrations of NO3-N between 2.9 and15 mg/L (f1 ≈ 1 in Eq. (1))

kNO3 and KSNO3

Inhibition by nitrite of nitrateoxidation

Assays with nitrite pulses; initial concentrations ofnitrate: 50–200 mg NO3-N/L; nitrite concentrations: 38,73, 117, 397, 906, and 1149 mg NO2-N/L

Inhibition factor f1; β1 for Eq. (3); γ1 for Eq. (4) andSI-NO2 for Eq. (5)

Inhibition by nitrate of nitriteoxidation

Data from the complete nitrate oxidation–nitriteoxidation process; initial nitrate concentrations of 1, 10,13, 15, 18, 21, 23, 25, 35, 68 and 115 mg NO3-N/L; thenitrate consumption rate, rNO3 , was determined; theinitial nitrite consumption rate, rNO2(t=0), was calculatedfrom Eq. (7); the rate of nitrite consumption,rNO2(t),.was calculated when nitrate was completelyconsumed (SNO3 = 0 at time t); the inhibition factor, f2,was calculated as: f2 = (rNO2(t=0)/rNO )

Inhibition factor f2; β2 for Eq. (3); γ2 for Eq. (4) andSI-NO3 for Eq. (5)

Inhibition constants for the Luongand Levenspiel models

The nitrate or nitrite concentration werrNO2 or rNO3 values = 0, respectively

Table 2Kinetic parameters for denitritation obtained from batch data without nitrateinhibition

Parameter Units Estimatedvalue

S.D. Correlation(r2)

kNO2 mg NO2-N/g VSS h 13.8 2.4 0.994KSNO2 mg NO2-N/L 0.36 0.09 0.996KI-NO2 mg NO2-N/L 906 190 0.991

Table 3Kinetic parameters for denitratation obtained from batch data without nitriteinhibition

Parameter Units Estimatedvalue

S.D. Correlation(r2)

k

K

fpc

bmpTc

3

3

d

mFfitd

raestablished maximum nitrate consumption rate values (kmax) inthe range of 9–94 mg N/g VSS h, and of 0.06–0.2 mg N/L forthe saturation constant (KS). For conditions similar to this work,

NO3 mg NO3-N/g VSS h 12.7 1.4 0.996

SNO3mg NO3-N/L 0.47 0.14 0.997

rom a steady state continuous culture reactor. Also since all the experiments forarameter estimation were run for less than 5 h, it was assumed that the biomassoncentration remained constant for purposes of parameter estimation [28].

Parametric sensitivity analysis was conducted using methodology proposedy Zaror and Pyle [29]. Initial denitratation and denitritation rates were esti-ated at different parameter values, and compared with baseline results. Each

arameter was set within −50% and +50% around its baseline value, shown inables 2 and 3. The sensitivity was quantified as the ratio between calculatedonsumption rate and the baseline value.

. Results and discussion

.1. Single step kinetics

Tables 2 and 3 present the experimental denitratation andenitritation kinetic parameters, respectively. Corresponding

Ft

2(t)

e registered at SNO3-M and SNO2-M

odel predictions and experimental results are compared inigs. 2 and 3. As observed from Figs. 2 and 3, experimental datattings were acceptable, with a high degree of correlation with

he proposed models (larger than r2 = 0.994) and low standardeviation (smaller than 5%).

Kinetic values obtained here are in agreement with thoseeported in the literature. Pala and Bolukbas [30] reported valuess low as 0.34 mg N/g VSS h. Tchobanoglous and Burton [31]

ig. 2. Denitritation batch experiments using different initial nitrite concentra-ion and model fitting.

O. Soto et al. / Enzyme and Microbial Technology 40 (2007) 1627–1634 1631

Ft

Dnc

fn[n[t

3

fdtfe

TEa

M

L

L

N

Ft

pfiLp

r

r

ct

ig. 3. Denitratation batch experiments using different initial nitrate concentra-ion and model fitting.

incer and Kargi [32] found values for the maximum specificitrate consumption rate of 9.6 mg N/g VSS h, and a saturationonstant of 0.3 mg N/L.

The Haldane inhibition constant (KI-NO2 ) in Eq. (2) accountsor the nitrite inhibition effect. This parameter cannot beeglected in the case of effluents featuring high nitrogen loads2]. When changing the C/N ratio from 30 to 300, there wereot variations in the kinetic results, confirming literature findings4]; this result can also be extended to industrial scale biologicalreatments.

.2. Cross-inhibition

For the denitratation and denitritation cross-inhibition, f1 and2 were estimated. Fig. 4 shows one pulse experiment used for the

etermination of the f1 function for nitrite inhibition on denitrata-ion. As shown in Table 4, the values of the inhibition constantsor the Levenspiel, Luong and Non-competitive models werestimated from Eqs. (3)–(5), respectively. Fig. 5A and B com-

able 4stimated values for denitratation and denitritation inhibition models by nitritend nitrate, respectively

odel Constant Estimated value Correlation(r2)

evenspiel β1 1.578 ± 0.065 0.990SNO2-M (mg NO2-N/L) 1149.4 ± 34.5β2 1.005 ± 0.098 0.954SNO3-M (mg NO3-N/L) 34.7 ± 1.0

uong γ1 0.776 ± 0.045 0.986SNO2-M (mg NO2-N/L) 1149.4 ± 34.5γ2 0.950 ± 0.053 0.926SNO3-M (mg NO3-N/L) 34.7 ± 1.0

on-competitive SI-NO2 (mg NO2-N/L) 43.7 ± 10.7 0.807SI-NO3 (mg NO3-N/L) 10.4 ± 3.9 0.535

bftibrcairn

ishtaa

h

ig. 4. Denitratation inhibition experiment. Black arrow shows the nitrite injec-ion point.

are model predictions and experimental results. As seen in thesegures, a good fit was obtained in the case of Levenspiels anduong models. However, the non-competitive model showed aoor fitting for both denitratation and denitritation.

Eqs. (1) and (2) can be completed as follows

NO3 =[

kNO3SNO3

KSNO3+ SNO3

] [1 − SNO2

SNO2-M

]β1

(8)

NO2 =[

kNO2SNO2

KSNO2+ SNO2 + S2

NO2/KI-NO2

] [1 − SNO3

SNO3-M

]β2

(9)

Experimental data presented in Fig. 5A and B reveal that theross-effect of the nitrate on the denitritation rate is much largerhan the effect of nitrite on the denitratation rate, as total inhi-ition is observed at concentrations near to 34.7 mg NO3-N/Lor denitritation, and near to 1149 mg NO2-N/L for denitrata-ion. This is also reflected in the large differences between thenhibition constants, SNO3-M and SNO2-M, shown in Table 4. Inhi-ition by nitrate starts very early; indeed, 10% inhibitions areeached when nitrite concentration is 74.2 mg N/L and nitrateoncentration is 3.4 mg/L for denitratation inhibition by nitritend denitritation inhibition by nitrate, respectively. Also, thenhibition constant for denitritation is very large, 906 mg/L, cor-oborating the fact that nitrite inhibition is less important thanitrate inhibition.

In our work, no nitrate inhibition was found when work-ng at nitrate concentrations up to 1000 mg NO3-N/L (data nothown). Other researchers have found nitrate inhibition duringydrogenotrophic denitrification where hydrogen is used as elec-ron donor for the nitrate reduction [33]. Therefore, these results

re limited to denitrification with mixed cultures where aceticcid is used as electron donor.

Fishery effluents used in this work are saline and the literatureas reported inhibition of denitrification at salt concentrations

1632 O. Soto et al. / Enzyme and Microbial Technology 40 (2007) 1627–1634

Fig. 5. (A) Cross-inhibition in denitrification process. Effect of nitrite concentra-to

ortnitTnpie

3

ab

Fig. 6. Kinetic model of batch denitrification; points and lines show experimen-t

sc

cdncOtbcotcmbtph

fm(pcawS

tha

ion on denitratation rate. (B) Cross-inhibition in denitrification process. Effectf nitrate concentration on denitritation rate.

f 24 g NaCl/L [22]. In our case, as the constants are in theange reported by the literature, it can be assumed that the cul-ure is adapted to these conditions; therefore, salt inhibition wasot considered in the models. Preliminary assays performedn the absence of salt showed a decrease in the nitrate reduc-ion rate when compared to the saline media (data not shown).his reveals the fact that the mixed community is adapted andot inhibited by a 2.4% salt content. Further experiments areresently being carried out to quantify the effect of increas-ng salt concentrations on both steps of denitrification in saltynvironments.

.3. Model application and sensitivity analysis

Fig. 6 compares model predictions and experimental nitritend nitrate concentrations over time. Predictions were obtainedy simultaneously solving Eqs. (6) and (7), using parameters

acph

al data and model, respectively.

hown in Tables 2 and 3. A good fitting is observed in bothases.

As shown in Figs. 5A and 6, the high concentration of nitritean inhibit nitrate consumption; however, even under these con-itions, nitrite consumption is postponed in order to consumeitrate. Nitrite accumulates in the liquid media when the nitrateonsumption rate is larger than the nitrite consumption rate.nce the nitrate present has been consumed, nitrite consump-

ion begins. This sequential scheme is not entirely representedy the mathematical models existing in the literature on denitrifi-ation and would be in agreement with the physical arrangementf denitrifying enzymes in the cellular structure [34,35]. Whilehe nitrate reductase enzyme is located at the internal side of theytoplasmic membrane, the nitrite reductase enzyme is periplas-atic; this distribution could explain the denitrification kinetic

ehavior with high nitrite and nitrate loads. Since nitrite reduc-ase is located on the external cell membrane, it would lackhysical protection barriers, which makes it more sensitive toigh nitrite and nitrate concentrations than the nitrate reductase.

Since both nitrite and nitrate are usually present in effluentsrom anaerobic treatment of fishery wastewater, the case of aixture of nitrate and nitrite is analyzed below. As shown in Eqs.

8) and (9), SNO2-M is the nitrite concentration that inhibits com-letely denitratation (rNO3 = 0), whereas SNO3-M is the nitrateoncentraction where denitritation stops (rNO2 = 0). SNO2-Mnd SNO3-M values indicate that the process starts preferablyith nitrate consumption. The lower SNO3-M value compared toNO2-M shows that the nitrate effect on denitritation is higher

han the nitrite effect on denitratation. In anoxic processes withigh organic and nitrogen load, denitritation is inhibited forny instantaneous nitrite concentration, at nitrate concentration

bove SNO3-M. Therefore, during application of denitrifying pro-esses, such as in anaerobic post-treatment of fishing or meatrocessing effluents, the probability of nitrite accumulation isigh as initial nitrate concentration is higher than the SNO3-M

O. Soto et al. / Enzyme and Microbial T

Fig. 7. (A) Sensitivity analysis of denitratation using parameters determinedfor high load processes (SNO3 = 100 mg NO3-N/L) at 37 ◦C and pH 7.5. (B)Sp

vd

pta1ooiasr

ct

1aeribcci

4

•

•

•

•

•

A

CC

R

[9] Glass C, Silverstein J, Oh J. Inhibition of denitrification in activated sludge

ensitivity analysis of denitritation using parameters determined for high loadrocesses (SNO2 = 100 mg NO2-N/L) at 37 ◦C and pH 7.5.

alue. This fact was also observed in laboratory experimentsuring batch and continuous reaction processes.

A sensitivity analysis of all the kinetic parameters waserformed. The baseline values of denitratation and denitri-ation rates, r∗

NO3and r∗

NO2, were estimated using Eqs. (8)

nd (9), respectively, for nitrate and nitrite concentrations of00 mg NOx-N/L. These concentrations are in the high rangef experimental values. Parameters were varied using a rangef 0.5 and 1.5 times their original value in order to determinets effect on the consumption rates, rNO3 and rNO2 . Fig. 7And B shows the denitratation and denitritation parameters sen-itivity expressed as the ratios (rNO3/r∗

NO3) and (rNO2/r∗

NO2),

espectively.

As shown from Fig. 7A and B, for high nitrate and nitrite

oncentrations, SNOx-M and βi values strongly influence deni-ratation and denitritation rates. For denitratation, an error of

[

echnology 40 (2007) 1627–1634 1633

0% in the determination of SNO2-M or β1 means a variation ofbout 15% in rNO3 ; whereas for denitritation, a 20% and 40%ffect can be observed in rNO2 if β2 and SNO3-M varied 10%,espectively. Such variations could influence the design of den-trifying reactors since the reactor size is directly influencedy the rate value. Therefore, inhibitory interactions should bearefully evaluated in order to establish design and operationalonditions for a denitrifying treatment system of a high-loadndustrial effluent.

. Conclusions

It is possible to model the denitrification process of a highnitrogen and organic load saline wastewater using a kineticmodel that considers the nitrite effect on nitrate consumptionand the nitrate effect on nitrite reduction.The cross-inhibition model adequately predicts the overallprocess.Nitrite and nitrate show inhibitory effect on denitratation anddenitritation, respectively, which can be represented by Lev-enspiel’s model.The cross-inhibition scheme favors that nitrate consumptionat the start of the process.Cross-inhibition is relevant for the design of denitrificationprocesses of effluents with high load and the inhibition param-eters exert a sensitive effect on the denitrification rates.

cknowledgements

This work was made possible through the grants FONDE-YT (Chile) No. 2000129, ECOS/CONICYT (France) No.00B06, and FONDECYT (Chile) No. 1040495.

eferences

[1] Griffiths A. Meat and animal by-products: processes and effluents. In:Herzka A, Booth RG, editors. Food industry wastes: disposal and recovery.Washington, DC: Applied Science Publishers; 1981. p. 171–83.

[2] Aspe E, Martı MC, Roeckel M. Anaerobic treatment of fishery wastewaterusing a marine sediment inoculum. Water Res 1997;31:2147–60.

[3] Knowles R, Denitrification. Microbiol Rev 1982;46:43–70.[4] Akunna JC, Bizeau C, Moletta R. Denitrification in anaerobic digester: pos-

sibilities and influence of wastewater COD/N-NOx ratio. Environ Technol1992;13:825–36.

[5] Aboutboul Y, Arviv R, van Rijn J. Anaerobic treatment of intensive fishculture effluents: volatile fatty acid mediates denitrification. Aquaculture1995;133:21–32.

[6] Fass S, Ganaye V, Urbain V, Manem J, Block JC. Volatile fattyacids as organic carbon sources in denitrification. Environ Technol1994;15:459–67.

[7] Almeida JS, Reis MAM, Carrondo MJT. Competition between nitrate andnitrite reduction in denitrification by Pseudomonas fluorescens. BiotechnolBioeng 1995;46:476–84.

[8] Christensen HM, Harremoes P. Biological denitrification of sewage: a lit-erature review. Prog Water Technol 1977;8:509–55.

by nitrite. Water Environ Res 1997;69:1086–93.10] Glass C, Silverstein J. Denitrification kinetics of high nitrate concentra-

tion water: pH effect on inhibition and nitrite accumulation. Water Res1998;32:831–9.

1 bial T

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

634 O. Soto et al. / Enzyme and Micro

11] Huang JS, Her JJ, Jih CG. Kinetics of denitritification and denitratificationin anoxic filters. Biotechnol Bioeng 1998;59:52–61.

12] Schonharting B, Rehner R, Metzger JW, Krauth K, Rizzi M. Release ofnitrous oxide (N2O) from denitrifying activated sludge caused by H2S con-taining wastewater: quantification and application of a new mathematicalmodel. Water Sci Technol 1998;38:237–46.

13] Wang JH, Baltzis BC, Lewandowski GA. Fundamental denitrificationkinetics studied with Pseudomonas denitrificans. Biotechnol Bioeng1995;47:26–41.

14] Dincer AR, Kargı F. Kinetics of sequential nitrification and denitrificationprocesses. Enzyme Microbial Technol 2000;27:37–42.

15] Wicht H. A model for predicting nitrous oxide production during denitri-fication in activated sludge. Water Sci Technol 1996;34:99–106.

16] Wild D, Von Schulthess R, Gujer W. Structured modelling of denitrificationintermediates. Water Sci Technol 1995;31:45–54.

17] Betlach MR, Tiedje JM. Kinetic explanation for accumulation of nitrite,nitric oxide, and nitrous oxide during bacterial denitrification. Appl EnvironMicrobial 1981;42:1074–84.

18] Mulchandani A, Luong J. Microbial inhibition kinetics revisited. EnzymeMicrobial Technol 1989;11:66–73.

19] Levespiel O. The chemical reactor omnibook. Corvallis: Chemical Engi-neering Dept., Oregon State University; 1979. p. 82–6 [Distributed by OSUBook Stores, Inc., Corvallis, USA].

20] Segel IH. Simple intersecting linear non-competitive inhibition. In: Behav-ior and analysis of rapid equilibrium and steady-state enzyme systems.Davis, CA, USA: John Wiley & Sons; 1975. p. 125.

21] Smythe G, Matelli G, Bradford M, Rocha C. Biological treatment of saltywastewater. Environ Prog 1997;16:179–83.

22] Kargi F, Dincer AR. Effect of salt concentration on biological treatment

of saline wastewater by fed-batch operation. Enzyme Microbial Technol1996;19:529–37.

23] Uygur A, Kargi F. Salt inhibition on biological nutrient removal fromsaline wastewater in a sequencing batch reactor. Enzyme Microbial Technol2004;34:313–8.

[

[

echnology 40 (2007) 1627–1634

24] Akunna JC, Bizeau C, Moletta R. Nitrate reduction by anaerobicsludge using glucose at various nitrate concentrations: ammonification,denitrification and methanogenic activities. Environ Technol 1994;15:41–9.

25] Akunna JC, Bizeau C, Moletta R. Nitrate and nitrite reductions with anaer-obic sludge using various carbon sources: glucose, glycerol, acetic acid,lactic acid and methanol. Water Res 1993;27:1303–12.

26] Delanghe F, Nakamura F, Myoga H, Magara Y. Biological denitrifica-tion with ethanol in a membrane bioreactor. Environ Technol 1994;15:61–70.

27] American Public Health Association (APHA). Standard methods for theanalysis of water and wastewater. 16th ed. Washington, DC: AmericanPublic Health Assoc; 1985.

28] Bailey JE, Ollis DF. Kinetics of substrate utilization, product for-mation, and biomass production in cell cultures. In: Biochemicalengineering fundamentals. 2nd ed. New York: McGraw-Hill, Inc; 1986.p. 384–5.

29] Zaror CA, Pyle DL. Competitive reactions model for the pyrolysisof lignocellulose: a critical study. J Anal Appl Pyrolysis 1986;10:1–12.

30] Pala A, Bolukbas O. Evaluation of kinetic parameters for biological CNPremoval from a municipal wastewater through batch tests. Process Biochem2005;40:629–35.

31] Tchobanoglous G, Burton FL. Wastewater engineering: treatment, disposaland reuse. 3rd ed. NewYork, NY: Metcalf and Eddy Inc./McGraw-Hill Inc;1991.

32] Dincer AR, Kargi F. Kinetics of sequential nitrification and denitrificationprocesses. Enzyme Microbial Technol 2000;27:37–42.

33] Vasiliadou I, Pavlou S, Vayenas DV. A kinetic study of hydrogenotrophic

denitrification. Process Biochem 2006;41:1401–8.

34] Richardson D. Bacterial respiration: a flexible process for a changing envi-ronment. Microbiology 2000;146:551–71.

35] Zumft WG. Cell biology and molecular basis of denitrification. MicrobiolMol Biol Rev 1997;61:533–616.