- Email: [email protected]
Denitrification potential of industrial wastewaters
ARTICLE IN PRESS
Water Research 39 (2005) 3715–3726 www.elsevier.com/locate/watres
Denitrification potential of industrial wastewaters A. De Lucas, L. Rodrı´ guez, J. Villasen˜or, F.J. Ferna´ndez Department of Chemical Engineering, University of Castilla-La Mancha, Avenida Camilo Jose´ Cela S/N E-13071, Ciudad Real, Spain Received 21 February 2005; received in revised form 24 June 2005; accepted 29 June 2005
Abstract Nitrate utilisation batch experiments have been carried out with different agro-food industrial wastewaters. The results obtained were modelled using a modification of the ASM2d to determine the main kinetic and stoichiometric parameters: denitrification rate with fermentation products (qNSA) and with fermentable substrates (qNSF); fermentation products and fermentable substrates consumption rates, (qSA) and (qSF), respectively; yield coefficient with fermentation products (YHSA) and with fermentable substrates (YHSF) and saturation coefficient for growth on fermentation products (KHSA). It was found that fermentation products (SA) and fermentable substrates (SF) consumption rates can be assumed to be both constant and independent of the agro-food industrial wastewater used, with values of 14.6 and 5.4 –1 (mg SCOD g COD-X–1 OHO h ), respectively. The mean denitrification rates obtained with each agro-food industrial –1 wastewater ranged from 1.5 to 2.5 (mg SN g COD-X–1 OHO h ), values slightly lower to the obtained with domestic –1 wastewater (2.7 mg SN g COD-X–1 h ). Given the substrate consumption rate and the denitrification rate, the yields OHO with fermentation products and fermentable substrates for each wastewater were determined. The values obtained for the yields ranged from 0.50 to 0.69 on using SA and from 0.50 to 0.70 on using SF. Having determined these parameters, the denitrification potential (DNP) was evaluated for each agro-food wastewater. Wastewater from slaughterhouses gave the lowest DNP (22.6 mg SN g S–1 COD), whereas wastewater from potato and tomato processing plants gave the highest values (126.5 and 128.2 mg SN g S–1 COD, respectively), which are very similar to that obtained with domestic wastewater (130.3 and 128.2 mg SN g S–1 COD). Taking into account the low N/COD ratio presented in the agro-food wastewaters, the last wastewaters are the most indicated to enhance the nitrogen removal in the WWTP. Finally, a relationship between the denitrification rates of several agro-food industrial wastewaters and their respective SA/(SA+SF) ratio was obtained. This relationship can be used to estimate the mean denitrification rate of an agro-food industrial wastewater. r 2005 Elsevier Ltd. All rights reserved. Keywords: Industrial wastewater; Activated sludge; Denitrification potential; Denitrification rate; Nitrite acumulation
1. Introduction
Corresponding author. Tel.: +34 926 295300;
fax: +34 926 295242. E-mail address: [email protected] (F.J. Ferna´ndez).
The excessive application of fertilizers, intensive exploitation of farms and the significant contribution from industry have increased the nitrogen load discharged to receiving waterways (Vitousek et al., 1997). This has led to a decrease in water quality and has caused health problems related to oxidised forms of
0043-1354/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2005.06.024
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3716
Nomenclature bH KHSA KHNO3 qSA qSF qend qNend qNSA qNSF SA SALK SCa SCOD SF
decay rate (d–1) saturation coefficient for growth on fermentation products (g COD m–3) saturation/inhibition coefficient for nitrate (g COD m–3) fermentation products consumption rate –1 (g COD g COD-X–1 OHO h ) fermentable substrates consumption rate –1 (g COD g COD-X–1 OHO h ) endogenous substrates degradation rate –1 (g COD g COD-X–1 OHO h ) denitrification rate with endogenous sub–1 strates (g N g COD-X–1 OHO h ) denitrification rate with fermentation pro–1 ducts (g N g COD-X–1 OHO h ) denitrification rate with fermentable sub–1 strates (g N g COD-X–1 OHO h ) fermentation products concentration (g COD m–3) alkalinity concentration (mol m–3) calcium concentration (g Ca m–3) COD concentration (g COD m–3) fermentable substrates (g COD m–3)
nitrogen. As a result, more stringent regulations have been approved in an effort to deal with this problem. In Europe, directive 91/271 prescribes the nitrogen standards for treated wastewater discharged into sensitive areas and establishes a timetable to reach an acceptable level of effluent quality. The most economic options for the removal of nitrogen from wastewater are biological processes and several methods already exist to achieve this goal: nitrification–denitrification, single reactor system for high ammonium removal over nitrite (SHARON) (Hellinga et al., 1998), anaerobic ammonium oxidation (ANAMMOX) (Jetten et al., 2001) and completely autotrophic nitrogen removal over nitrite (CANON) (Sliekers et al., 2002). Each of these processes exploits one of the different available pathways in the nitrogen cycle (van Loosdrecht and Jetten, 1998). Currently, the most developed system is the nitrification–denitrification process, which consists of an initial nitrification stage, accomplished by autotrophic bacteria, in which ammonia is oxidized to nitrite by Nitrosomonas sp. The nitrite is subsequently oxidised to nitrate by Nitrobacter sp., this process is outlined in Eqs. (1) and (2). A second denitrification stage (Eq. (3)) is then carried out by numerous facultative heterotrophic bacteria. In this second process, nitrate obtained in the nitrification stage is reduced to molecular nitrogen in four steps, using the substrates contained in the wastewater as an electron donor. These steps involve the
SI SK SMg SN SN2NO3 SN2NO2 SNa SPT SP2PO4 XPAO XGAO XAUT XOHO XPT XSS YHSA YHSF
inert COD concentration (g COD m–3) potasium concentration (g K m–3) magnesium concentration (g Mg m–3) nitrogen concentration (g N m–3) nitrate–nitrogen concentration (g N m–3) nitrite–nitrogen concentration (g N m–3) sodium concentration (g Na m–3) phosphorus total concentration (g P m–3) orthophosphate concentration (g P m–3) phosphate accumulating Organism concentration (g COD-XPAO m–3) glycogen accumulating Organism concentration (g COD-XGAO m–3) autotrophic organisms concentration (g COD-XAUT m–3) ordinary heterotrophic Organisms concentration (g COD-XOHO m–3) phosphorus content in the sludge concentration (g P g VSS) suspended solids concentration (g m–3) yield coefficient with fermentation products (g COD g–1 COD) yield coefficient with fermentable substrates (g COD g–1 COD)
reduction of nitrate to nitrite, which is further reduced to one or more of the gaseous nitrogen products (nitric oxide, nitrous oxide and nitrogen gas) that will be finally released to the atmosphere (Henze et al., 1995a). þ 2NHþ 4 þ 3O2 ! 2NO2 þ 2H2 O þ 4H ,
(1)
2NO 2 þ O2 ! 2NO3 ,
(2)
NO 3 ! NO2 ! NOðgÞ ! N2 OðgÞ ! N2ðgÞ .
(3)
The overall rate and extension of the denitrification process depends mainly on the biodegradability characteristics of the electron donor used (Bolzonella et al., 2001) and on the final N/COD ratio in the bioreactor (Chiu and Chung, 2003). For these reasons, the denitrification process can be enhanced by adding readily biodegradable carbon sources—such as acetic acid and methanol—or carbon sources presenting high COD/N ratios to the wastewater. However, the addition of these carbon sources is expensive and increases the cost of treating the wastewater. The cost involved in these processes has led to carry out studies using waste carbon sources such as industrial wastes and organic fraction of municipal wastes (Bilanovic et al., 1999). Due to EU regulations there are, at present, many wastewater treatment plants (WWTP) working with this nitrogen removal process. These plants are designed to treat typical domestic wastewater by removing the
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
nitrogen from it biologically. However, in many cases, these systems would receive wastewater with high nitrogen content or with different characteristics, mainly biodegradability, due to industrial discharges (Rodrı´ guez et al., 2004). These changes in the wastewater composition can either increase the nitrate concentration in the effluent, due to the absence of an efficient denitrification process caused by inhibition phenomena or the absence of a suitable electron donor, or decrease the nitrate concentration. This latter effect can be caused by a high biodegradability or low ratio N/COD of the industrial wastewater. One of the best ways to describe the process behaviours and to predict its evolution is the mathematical modelization (Harder and Roels, 1982). In the case of the denitrification process, it has been described in several models (Henze et al., 1987, 1995b, 1999). In the ASM2d (Henze et al., 1999) the denitrification process is carried out by both ordinary heterotrophic organisms (OHO) and phosphate accumulating organisms (PAOs). The OHO denitrification process is modelled as two parallel processes, which consume nitrates, as electron acceptor, to oxidise the two biodegradable COD fractions, SA and SF. For both processes the same kinetics and stoichiometrics coefficients are assumed. The PAO denitrification process is carried out at expense of the internal organic storage products. In this context, the aim of the work described here was to study the effects of agro-food industrial discharges on the denitrification stage of a biological nutrient removal process. The results obtained in the nitrate utilisation experiments were modelled using a modified ASM2d model. As a result of the modelisation, the values of the most important parameters for the denitrification process (i.e., qSA, qSF, qNSA, qNSF, qNend, YHSA, YHSF, and KHSA) were evaluated. Finally, the denitrification potential (DNP) of each agro-food industrial wastewater was evaluated and a relationship between the mean denitrification rate for each agro-food industrial wastewater and their respective SA/(SA+SF) ratio was proposed.
2. Materials and methods 2.1. Activated sludge The activated sludge used in this work was taken from sequence batch reactors (SBRs) working with the A/O process for biological nutrient removal (BNR). Further details concerning the SBRs experimental set-up can be found in previous publications (Can˜izares et al., 2000; De Lucas et al., 2000). The wastewater used for the cultivation of microorganisms was prepared daily and was composed of real domestic wastewater—taken from the primary clarifiers
3717
of a full-scale wastewater treatment plant—modified by mixing with synthetic wastewater. The composition of the synthetic wastewater and the characteristics of the resulting wastewater are shown in Table 1 (sections a and b, respectively). The criteria for reaching steady state in the SBRs were (i) to reach a stable biomass concentration and (ii) to attain stable effluent treated water characteristics. 2.2. Pre-treatment of activated sludge The activated sludge used in this work was acclimatized to BNR. For this reason, activated sludge was enriched on PAO and glycogen accumulating organisms (GAO). Under anoxic conditions, PAO and GAO microorganisms could use intracellularly stored substrates to denitrify (Kuba et al., 1996; Zeng et al., 2003), thus consuming nitrates but not COD from the bulk liquid. Pre-treatment of the sludge was necessary in order to avoid this process, which would negatively influence the determination of the stoichiometry and the kinetics parameters of the denitrification process carried out by ordinary heterotrophic organisms (OHO). This pre-treatment consisted of an excessive aeration stage to exhaust all of the intracellular polymers accumulated on PAO and GAO organisms. Once this pre-treatment had been performed, only the OHO of the sludge were able to denitrify. Prior to all batch experiments, the sludge was washed with osmotized water in aerobic conditions in order to remove the absorbed substrates from the surface of the flocks. 2.3. Industrial wastewater The synthetic agro-food industrial wastewaters used in this work were prepared according to the procedure Table 1 Concentration (a) Composition of the synthetic wastewater Component 322 Sodium acetate (mg L1): (NH4)2SO4 (mg L1): 74.2 KH2PO4 (mg L1): 44.5 115.0 NaHCO3 (mg L1): MgSO4 7H2O(mg L1): 50.0 CaCl2 (mg L1): 30.0 3.0 (NH4)2 Fe(SO4)2 (mg L1): (b) Characteristics of the resulting wastewater used for the cultivation of microorganisms Parameter 340 (46) COD (mg L1): BOD5 (mg L1): 181 (22) 9.1 (2.6) P (mg L1): N (mg L1): 15.8 (3.7) pH 7.3 (0.1)
ARTICLE IN PRESS 3718
A. De Lucas et al. / Water Research 39 (2005) 3715–3726
described in literature (Table 2). Agro-food wastewaters are characterised by its very low N/COD and P/COD ratios, because of this, nutrients were added synthetically in order to avoid nutrients limitations. The wastewaters used in this work were: wastewaters from cheese industries, milk bottling industries, slaughterhouses, potato processing industries, beet sugar processing industries, tomato processing industries and winery industries.
2.4. Batch experiments The denitrification processes for the agro-food industrial wastewaters and the domestic wastewater were studied by carrying out several batch tests in closed reactors, using the pre-treated biomass from the SBRs. Each batch reactor had a total volume of 260 mL. The reactors were filled with 250 mL of the pre-treated biomass with nutrients and industrial wastewater to give the following approximate nutrients and COD final concentrations: S PT ¼ 10 mg L1 , SN2NH4 ¼ 15 mg L1 , SK ¼ 12 mg L–1, SNa ¼ 14 mg L–1, SMg ¼ 5 mg L–1, SCa ¼ 10 mg L–1, SCOD ¼ 300 mg L–1. In this way, the dead gas volume inside the closed reactor is minimised and this reduces the adverse effect of oxygen on the denitrification process. In order to maintain the sludge suspended, avoiding oxygen transfer from the gas phase, a very slow stirring rate was applied. Experiments were carried out with an SCOD/XOHO ratio of 0.18 g COD g COD-X–1 OHO. Nitrate was added continuously to avoid substrate inhibition due to very high nitrate concentrations, but ensuring the continuous presence of it (Wachtmeister et al., 1997). The pH in the batch reactors oscillated between 7.2 and 7.4 and the temperature was continuously maintained at 20 1C by means of an incubator (ISCO FTD 220). Two reference tests were simultaneously carried out. One of these tests consisted of a batch reactor filled with 250 mL of the pre-treated biomass and nutrients
Table 2 References for the preparation of the synthetic agro-food industrial wastewater Industrial wastewater
Reference
Potato processing industries Tomato processing industries Beet-sugar processing industries Cheese industries Milk bottling industries Winery industries Slaughterhouses
Contreras et al. (2001) Niementowski and Nelson (1977) Nemerow and Dasgupta (1991) Herbert and Yu (2001) Alvarez-Mateos et al. (2000) Nemerow and Dasgupta (1991) Masse and Masse (2000)
(blank test). From the results obtained in this experiment the endogenous denitrification and the decay rate can be obtained. The other reference test (maximum test) consisted of a batch reactor filled with 250 mL of the pre-treated biomass, nutrients and sodium acetate to reach the nutrient and COD final concentrations outlined above. Sodium acetate is known to be a very easily biodegradable substrate that offers the highest denitrification rates (Gerber et al., 1986; Somiya et al., 1988; Tam et al., 1992). The results obtained in this experiment enabled the maximum denitrification rates to be evaluated. 2.5. Sampling procedure and analytical methods Samples were taken by syringe and were immediately filtered through a Millipore glass fibre filter. After filtration, samples were analysed for the following parameters: S COD ; SN2NO2 ; S NNO3 ; SPPO4 and SN2NH4 : In addition, XSS, volatile suspended solids (VSS) and XPT were determined in the activated sludge. SP2PO4 , XPT and SN2NH4 were analysed to confirm the absence of denitrifying phosphate accumulation and nitrification. All analyses of wastewater and activated sludge were performed as described by APHA (1998). The biodegradable COD fractions, SA and SF, were determined using electrolytic respirometry (Spanjers and Vanrolleghem, 1995). 2.6. Mathematical model A large number of models can be used to evaluate the results obtained in the anoxic experiments. These mathematical models usually have complex structures and a very large number of equations and parameters. The model used in this work is a simplification of the ASM2d model (Henze et al., 1999), in which a couple of modifications were made. These modifications are as follows: 1. ASM2d does not take into account the possible accumulation of nitrites in the bulk liquid as the reduction of nitrite to nitrogen gas (Eq. (3)) is supposed to be very fast. However, nitrite accumulation has been reported in numerous studies (Henze, 1986; So¨zen and Orhon, 1999). Henze proposed in 1986 that nitrite accumulation could occur due to at least two different causes: excessive growth of biomass that is able to reduce nitrate to nitrite and/ or the inhibition of the nitrite reductase enzyme. It is important to take the possibility of this accumulation into account for two reasons; firstly, neglecting it would increase the real denitrification rate obtained if nitrates are supposed to be completely reduced and, secondly, the real denitrification rate will be
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
decreased if oxidized nitrogen were measured. The corrected value for the amount of nitrate–nitrogen completely reduced during the anoxic experiment can be evaluated by taking into account the variations in nitrates and nitrites, Eq. (4), (So¨zen and Orhon, 1999). DS NCorrected ¼ ðDSNNO3 0:6DSNNO2 Þ.
(4)
2. In this equation, SNCorrected represents the theoretical amount of nitrates that, when completely reduced to nitrogen gas, will supply the same amount of electron acceptor as consumed in the batch reactor. 3. In the ASM2d model the maximum specific carbon degradation rate due to the denitrification process is supposed to be the same regardless of the substrate used. However, studies carried out with different substrates have shown that the rate depends on the characteristics of the substrate (Alleman and Irvine, 1980; Christensson et al., 1994; Bolzonella et al., 2001). For this reason, we changed the maximum specific substrate consumption rate expression from ASM2d to allow two rates: one for the utilisation of SA and the other for the utilisation of SF, qSA and qSF, respectively. The experimental conditions (anaerobic environment, excess concentrations of ammonium, alkalinity, phosphorus, nitrate and fermentable substrates during the experiments) and the low sensitivity of certain parameters allows the ASM2d model to be simplified to a set of dynamic equations as follows, Eqs. (5)–(11): Denitrification with SA SNO3 dSN ¼ qN2SA dt SA K H2NO3 þ SNO3 SA SA X OHO . ð5Þ K H2SA þ SA S F þ S A Denitrification with SF SNO3 dSN SF ¼ qN2SF X OHO . dt SF K H2NO3 þ S NO3 S F þ S A
(7)
Consumption of SA S NO3 dS A ¼ qSA dt K H2NO3 þ SNO3 SA SA X OHO . K H2SA þ SA SF þ SA
Consumption of SF S NO3 dSF SF ¼ qSF X OHO . dt K H2NO3 þ SNO3 SF þ SA
(9)
Anoxic growth dX H ¼ dt
Y H2SA qSA
SA SA K H2SA þ SA SF þ SA K H2NO3 þ SNO3 SF bH þY H2SF qSF SF þ SA SNO3 S NO3 X OHO . ð10Þ K H2NO3 þ SNO3
Relationship between variables qN2SF ¼ qSF
1 Y H2SF , 2:86
and qN2SA ¼ qSA
1 Y H2SA . 2:86
ð11Þ
The blank test enables the endogenous denitrification rate (qN–end) and the decay rate (bH) to be obtained. Having obtained these values, the remaining equations can be solved. The simultaneous estimation of qSA, qSF, qN–SA, qN–SF and KH–SA involved assuming an initial set of parameter values and calculating theoretical nitrate, COD and growth of the biomass profiles in the batch reactor by solving simultaneously the Eqs. (5)–(11) using the Newton method. A minimal residual sum of squared errors, associated with the difference between these theoretical curves and the experimental results, have to be reached. The values of qSA, qSF, qN–SA, qN–SF and KH–SA associated with that minimum constitute the best estimates of these parameters. In an effort to ensure a correct estimation, the initial supposed values for the parameters ranged from those corresponding to the blank reference test to those for the maximum reference test. 2.7. DNP
(6) Denitrification with endogenous substrate dSN ¼ qN2end X OHO . dt end
3719
ð8Þ
Once the kinetic and stoichiometric values had been obtained for the denitrification process with the wastewaters under investigation, the DNPs were determined (Kujawa and Klapwijk, 1999). The DNP was calculated using Eqs. (12) and (13), in which DNP is expressed as the amount of nitrates that can be denitrified in the SA and SF fractions of COD. DNPSA ¼
DS N2Corrected , DSA
(12)
DNPSF ¼
DSN2Corrected . DS F
(13)
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3720
The overall DNP of the substrates contained in each industrial wastewater depends on the composition of the wastewater. This value can be calculated using Eq. (14), where SA, SF and SCOD are the concentrations of the different COD fractions in the influent wastewater. DNP ¼
S A DNPSA þ SF DNPSF . SCOD
(14)
3. Results and discussion 3.1. Wastewater characterisation The quality of predictions achieved by activated sludge models depends on the accuracy of the wastewater characterisation and on the calibration of the parameters. Therefore, the influent wastewater was characterised using a reliable method proposed by the Dutch Foundation of Applied Water Research (STOWA) (Roeleveld and van Loosdrecht, 2002). The wastewater characteristics for each batch experiment are presented in Table 3.
3.2. Biomass characterisation The activated sludge used in this work was assumed to be composed mainly of PAO, GAO, OHO and autotrophic organisms (AUT). The concentration of each of these microorganisms was determined (Ferna´ndez, 2004) using a metabolic model (Smolders et al., 1995) and formulae proposed in the literature (Rodrigo et al., 1999; Cortacans, 2000). The concentrations of bacteria in the activated sludge were as follows: PAO ¼ 275 (mg COD-XPAO L–1), GAO ¼ 410 (mg COD-XGAO L–1), AUT ¼ 175 (mg COD-XAUT L–1) and OHO ¼ 1767 (mg COD-XOHO L–1).
3.3. Endogenous denitrification rate and maximum denitrification rate The maximum specific denitrification rate using the endogenous substrate (qNend) and decay rate (bH) were determined from the results of the blank experiment (Fig. 1). During this experiment, the accumulation of nitrite was negligible and, as a result, nitrite was not taken into account in Fig. 1. The maximum specific rate obtained for endogenous denitrification was 1 0:3 ðmg SN-Corrected g COD-X 1 This value is OHO h Þ. slightly lower than that reported in the literature for similar conditions (Kujawa and Klapwijk, 1999). The rate constant for decay (bH) obtained from the experimental results was 0.39 d–1, a value very similar to that corresponding to the ASM2d model (0.40 d–1). The maximum specific denitrification rate and the maximum specific COD consumption rate were obtained using the results from the reference test carried out using acetate as the organic substrate (Fig. 2). In this case, nitrite was accumulated in the bulk liquid and it was therefore necessary to take its effect into account. The nitrate utilisation rate was thus calculated using Eq. (4). The maximum specific substrate con–1 sumption rate was 24.7 (mg SCOD g COD-X–1 OHO h ) and the maximum specific denitrification rate was –1 3.4 (mg SNCorrected g COD-X–1 OHO h ). Both values are slightly higher than those reported by Kujawa and Klapwijk (1999) for similar experiments. This result could be due to the acclimatization of the sludge to acetate because this substrate was fed into the SBRs from which the activated sludge was taken. The heterotrophic yield, YH, was determined by means of the COD and nitrogen balances using (Eq. (15)). dSCOD 2:86 dS N2NO3 ¼ . 1 YH dt dt
(15)
The value obtained was 0.60 (mg COD-XOHO mg S–1 COD).
Table 3 Industrial wastewater characterisation according to STOWA guidelines Parameter
Wastewater
Symbol
Unit
AF1
AF2
AF3
AF4
AF5
AF6
AF7
DWW
SF SA SN2NH4 SN2NO3 SP2PO4 SI SALK
g COD m3 g COD m3 g N m3 g N m3 g P m3 g COD m3 mol m3
208 58 17 0 13.5 85 1.43
207 29 18.5 0 9.5 68 1.53
56 7 22 0 9 238 1.37
183 33 18.4 0 10.2 57 1.35
165 48 15 0 11.2 86 1.39
201 96 19 0 10.1 67 1.85
180 73 19.5 0 9.7 59 1.85
173 41 20.2 0 9.4 21 4.12
AF1: cheese industries; AF2: milk bottling industries; AF3: slaughterhouses; AF4: potatoes processing; AF5: beet sugar processing; AF6: tomatoes processing; AF7: winery industries; DWW: domestic wastewater.
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
35
2000
XOHO (mg COD·L-1)
SN (mg·L-1)
28 21 14 7 0
3721
0
1
2
3
4
1500
1000
500
0
5
0
1
2
4
3
t (h)
5
t (h)
Fig. 1. Evolution of SN2NO3 (m) and biomass (K) during the blank test. Solid lines indicate the model fit.
2000
500
35
1500
300 1000 200 SCOD
500
100
0
0 0
1
2
3
4
5
t (h)
28 SN (mg/L)
SCOD (mg·L-1)
400
XOHO (mg COD·L-1)
XOHO
21 14 SN-Corrected 7
SN-NO3
SN-NO2 0
0
1
3
2
4
5
t (h)
Fig. 2. Evolution of SN2NO3 (n), SN2NO2 (&), SN–Corrected (K), XOHO expressed as COD (’) and SCOD (m) during the acetate reference test. Solid lines indicate the model fit.
3.4. Denitrification with agro-food industrial wastewater Denitrification with agro-food wastewaters from cheese industries, milk bottling industries, slaughterhouses, potato processing industries, beet sugar processing industries, tomato processing industries and winery industries were studied. The proposed model was employed to simultaneously determine the maximum specific substrate consumption rates for SA and SF during the denitrification process in all the experiments in order to give comparable values. These values were 14.6 (mg SCOD g COD–1 X–1 OHO h ) for the utilisation of SA and 5.4 (mg SCOD g –1 COD-X–1 OHO h ) for the utilisation of SF. The experimental results and the results predicted by the model are shown in Fig. 3. Data predicted by the model fit very well to the experimental results. Consequently, the agro-food industrial wastewater studied can be taken—at least in terms of COD removal—to be a mixture in different
percentages of SA and SF. As can be seen, the maximum specific substrate consumption rate for SA –1 (14.6 mg SCOD g COD-X–1 OHO h ) is lower than that –1 obtained for acetate (24.7 mg SCOD g COD-X–1 OHO h ). This can be explained because SA concerns not only acetate but also by other fermentation products. In all the experiments carried out with agro-food industrial wastewater, nitrite was accumulated in the bulk liquid and Eq. (4) was therefore used to determine the maximum specific denitrification rates. The values obtained for this parameter ranged from 1.58 to 2.55 –1 (mg SNCorrected g SCOD-X–1 OHO h ) on using SA and –1 from 0.57 to 0.92 (mg SN–Corrected g SCOD-X–1 OHO h ) on using SF. These denitrification rates were used in conjunction with the agro-food industrial wastewater characterisation to give the mean denitrification rates for each wastewater. The mean denitrification rates obtained were very similar to that reported previously for domestic wastewater (Mattsson, 1997; Bolzonella et al., 2001).
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3722
Cheese industries 500
2000
1250
300
1000 200
750
SCOD
500
100
50 SN (mg·L-1)
1500
60 XOHO (mg COD·L-1)
SCOD (mg·L-1)
1750
XOHO
400
0
2
4
0 8 10 12 14 16 18 20 t (h)
6
SN-Corrected
30 SN-NO3
20 10
250 0
40
SN-NO2
0 0
2
4
6
8
10 12 14 16 18 20 t (h)
Milk bottling industries 2000 1500 1250
300
1000 200
750
SCOD
500
100
60 50 SN (mg·L-1)
SCOD (mg·L-1)
1750
XOHO
400
XOHO (mg COD·L-1)
500
0
2
4
6
8
SN-Corrected
30 20
SN-NO3 SN-NO2
10
250 0
40
0 10 12 14 16 18 20 t (h)
0 0
2
4
6
8 10 12 14 16 18 20 t (h)
Slaughter houses 2000
SCOD (mg·L-1)
1500 1250
300 SCOD
1000
200
750 500
100
60 50 SN (mg·L-1)
XOHO 1750
400
XOHO (mg COD·L-1)
500
1
0
3 t (h)
2
4
6
5
30
SN-Corrected
20
SN-NO3
10
250 0
40
0
0
SN-NO2 1
0
3 t (h)
2
4
6
5
Potatoes processing waste water 500
2000 1500 1250
300
1000 200
750 SCOD
100
500 250
0 0
2
4
6
8
0 10 12 14 16 18 20 t (h)
60 50 SN (mg·L-1)
SCOD (mg·L-1)
1750
XOHO (mg COD·L-1)
XOHO
400
40 SN-Corrected
30 20
SN-NO3 SN-NO2
10 0 0
2
4
6
8
10 12 14 16 t (h)
18 20
Fig. 3. Evolution of SN2NO3 (n), SN2NO2 (&), SN–Corrected (K), XOHO expressed as COD (’) and SCOD (m) during the experiments carried out with agro-food industries wastewater. Solid lines indicate the model fit.
The maximum specific substrate consumption rates and the maximum specific denitrification rates were used to determine the yields of the process. In addition, the value for KHSA was determined (5 mg SCOD L–1), the
value obtained is very similar to that proposed by ASM2d, (4 mg SCOD L–1). A summary of the values for the kinetics and stoichiometrics parameters of the denitrification process is given in Table 4.
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3723
Beet-sugar processing wastewater 500
2000
1000 200
750 SCOD
100
500
50 SN (mg·L-1)
SCOD (mg·L-1)
1500 1250
300
60
1750 XOHO (mg COD·L-1)
XOHO
400
2
4
6
SN-Corrected
30 20
SN-NO3
10
250 0 0
40
0
0 8 10 12 14 16 18 20 t (h)
SN-NO2 0
2
4
6
8
10 12 14 16 t (h)
18 20
Tomatoes processing wastewater 2000
SCOD (mg·L-1)
1500 1250
300
1000 200
750 SCOD
100
60
1750
500
50 SN (mg·L-1)
XOHO
400
XOHO (mg COD·L-1)
500
0
2
4
6
2000
1250 1000
200
750 SCOD
500
2
4
6
4
6
8
10 12 14 16 18 20 t (h)
40 SN-Corrected
30 20
SN-NO3
10
250 0
SN-NO2 2
50 SN (mg·L-1)
SCOD (mg·L-1)
1500
300
0
SN-NO3
60
1750 XOHO (mg COD·L-1)
XOHO
100
SN-Corrected
20
0 0 0 8 10 12 14 16 18 20 t (h) Winery industries
500 400
30
10
250 0
40
0 8 10 12 14 16 18 20 t (h)
0
SN-NO2 0
2
4
6
8
10 12 14 16 t (h)
18 20
Fig. 3. (Continued)
The DNP was calculated on the basis of these results (Table 4). The industrial wastewaters with the highest DNPs were those arising from tomato and potato processing, with potential values of 128.3 and 126.5 (mg SN–Corrected g S–1 COD), respectively. This result can be explained in terms of the composition and characteristics of these wastewaters, mainly related to their fermentation products and fermentable substrates concentration. The lowest DNP was obtained with slaughterhouse wastewater, with a value of 22.6 (mg SN–Corrected g S–1 COD). Parallel to these experiments carried out with agrofood wastewater, an experiment was carried out with domestic wastewater. The objective of this experiment is to establish a reference to compare the results
obtained with agro-food wastewater. The experimental results were modelled (Fig. 4) obtaining the values for the kinetics and stoichiometrics parameters shown in Table 4. The yields and the half-saturation constant obtained were similar to that obtained with agro-food wastewaters. However, the consumption rates of the COD fractions, SA and SF, and the denitrification rates associated to these fractions were slightly higher than the obtained with agro-food wastewater. The DNP obtained with domestic wastewater 130 (mg SN–Corrected g S–1 COD) was higher than the obtained with agro-food wastewater, only wastewaters from tomatoes and potatoes processing industries presented similar values. These results indicate that the
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3724
Table 4 Kinetic and stoichiometric coefficients for the denitrification process Wastewater
Parameter Symbol
Unit
AF1
qSA qSF qN–end qN–SA qN–SF YH–SA YH–SF KH–SA
1 mg SCOD g COD-X1 OHO h 1 mg SCOD g COD-X1 h OHO 1 mg SN–Corrected g COD-XOHO h1 1 mg SN–Corrected g COD-X1 OHO h 1 mg SN–Corrected g COD-XOHO h1 mg COD-XOHO mg S1 COD mg COD-XOHO mg S1 COD 1 mg SCOD L
14.6 5.4 0.3 2.00 0.66 0.60 0.65 5.0
DNP
mg SN–Corrected g S1 COD
95.1
AF2
AF3
14.6 5.4 0.3 2.55 0.76 0.50 0.60 5.0
AF4
14.6 5.4 0.3 1.94 0.57 0.62 0.70 5.0
112.6
22.6
AF5
14.6 5.4 0.3 2.55 0.85 0.50 0.58 5.0
AF6
14.6 5.4 0.3 1.58 0.87 0.69 0.54 5.0
126.5
106.3
AF7
14.6 5.4 0.3 1.88 0.92 0.63 0.51 5.0 128.2
DWW
14.6 5.4 0.3 2.14 0.81 0.58 0.57 5.0
20.5 8.4 0.3 3.26 1.17 0.65 0.60 4.0
120.7
130.3
AF1: cheese industries; AF2: milk bottling industries; AF3: slaughterhouses; AF4: potatoes processing; AF5: beet sugar processing; AF6: tomatoes processing; AF7: winery industries; DWW: domestic wastewater.
500
2000 XOHO
1500
300 1000 200 500
100
0
2
4
8
6
10
40 SN-Corrected
30 20
SN-NO3
10
SCOD 0
50 SN (mg/L)
SCOD (mg·L-1)
400
XOHO (mg COD·L-1)
60
12
SN-NO2
0 14
0
t (h)
0
2
4
6
8
10
12
14
t (h)
Fig. 4. Evolution of SN2NO3 (n), SN2NO2 (&), SN–Corrected (K), XOHO expressed as COD (’) and SCOD (m) during the experiment carried out with domestic wastewater. Solid lines indicate the model fit.
discharge of these agro-food wastewaters would not significantly influence the denitrification rate in a WWTP. However, the discharge of the agro-food wastewaters would increase the extension of the denitrification process, because the agro-food wastewaters supply considerable amounts of readily biodegradable substrates presenting a very low N/COD ratio. The enhancement of the denitrification process would reduce the nitrate and nitrite concentrations, reducing therefore, the filament proliferation caused by these compounds (Musvoto et al., 1999). The mean denitrification rates obtained for each wastewater mainly depended on the SA and SF concentrations. The mean denitrification rates vs. the SA/(SA+SF) ratio obtained for each agro-food wastewater are shown in Fig. 5 along with the limiting values in each case; the theoretical maximum specific denitrification rate for pure SF wastewater
(obtained in the modelisation) and the experimental maximum specific denitrification rate for pure SA wastewater (reference test carried out with acetate). These data were fitted to an empirical equation (Eq. (16)) using a non-linear regression method. The values obtained previously were used as initial guesses for the parameters: –1 qNSA ¼ 3.40 (mg SNCorrected g SCOD-X–1 OHO h )
maximum specific denitrification rate (obtained in the maximum reference test). qNSF ¼ 0.70 (mg SN–Corrected g SCOD-XOHO–1 h–1) average maximum specific denitrification rate (obtained with the SF fraction of the agro-food wastewaters). A ðq qN2SF Þ K S SþS qN ¼ qN2SA 1 N2SA (16) e A F . qN2SA
ARTICLE IN PRESS
mean qN (mg SN·g COD-XOHO-1·h-1)
A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3.5 3.0
Domestic wastewater
Winery industries
SA
Tomatoes 2.5 Potatoes processing processing Beet-sugar 2.0 Milk bottling processings industries Cheese 1.5 industries
1.0
Slaughter houses SF
0.5 0.0
0.0
0.2
0.4
0.6
0.8
1.0
SA/(SA+SF) Fig. 5. Mean denitrification rate for the studied agro-food wastewater. Solid line indicates the predictions of the empirical equation used.
The estimated values for the parameters are as follows:
3725
wastewater. Only wastewater from a winery and tomato processing plant gave similar denitrification rates, which could be explained because of the higher SA concentration in these wastewaters. These results indicate that the discharge of the industrial wastewater studied would not increase the denitrification rate in a WWTP. However, the denitrification process could be enhanced taking into account the combined effect of the denitrification rate and the N/COD ratio of the agro-food wastewaters. The different mean denitrification rates of agro-food wastewater can be calculated using the proposed equation (Eq. (16)), which predicts the denitrification rate for agro-food industrial wastewater using only one variable—the SA/(SA+SF) ratio. The industrial wastewaters with the highest DNP values were those from tomato and potato processing industries, each of which has a potential of about 127 mg SN–Corrected g S–1 COD. The combination of results, DNP and maximum specific denitrification rate, indicate that the most suitable wastewater for nitrogen removal is that arising from tomato processing industries.
1 qN2SA ¼ 3:49 ðmg S N2Corrected g COD-X1 OHO h Þ, 1 qN2SF ¼ 0:72 ðmg SN2Corrected g COD-X1 OHO h Þ,
K ¼ 3:01 ðg S COD g S 1 COD Þ. This equation relates the mean denitrification rate obtained for the agro-food industrial wastewater studied with its SA/(SA+SF) ratio. This equation can be used to predict the mean denitrification rate for agro-food industrial wastewaters. In Fig. 5, the mean denitrification rate for the domestic wastewater is also represented. The mean denitrification rate obtained with domestic wastewater is higher than the obtained with the agrofood wastewaters, which could be explained taking into account that the biomass was acclimatised to the substrates contained in this wastewaters and because of the more complex structure of the substrates contained in the agro-food wastewaters.
4. Conclusions The substrate utilisation rate is not the same for fermentation products (SA) and fermentable substrates (SF), with the maximum specific utilisation rate for each of these parameters having values of 14.6 and –1 5.4 (SCOD g COD-X–1 OHO h ) regardless of the industrial agro-food wastewater used. Owing to the different yields of the substrates contained in the industrial wastewaters, the maximum specific denitrification rate for each fraction of the COD was different. These rates were in –1 the range 1.6–2.6 (mg SN–Corrected g COD-X–1 OHO h ) for –1 –1 SA and 0.6–0.9 (mg SN–Corrected g COD-XOHO h ) for SF. The mean denitrification rates for the studied wastewaters were lower than the obtained with domestic
References Alleman, J.E., Irvine, R.L., 1980. Storage induced denitrification using Sequencing Batch Reactor operation. Water Res. 14, 1483–1488. A´lvarez-Mateos, P., Pereda-Marı´ n, J., Carta-Escobar, F., Dura´nBarrantes, M.M., Guille´n-Jime´nez, E., 2000. Influence of inoculum and initial pH on dairy effluent biodegradation and mineralization. Chem. Biochem. Eng. Q. 14 (3), 101–106. A.P.H.A.-A.W.W.A.-W.P.C.F., 1998. Standard Methods for the Examination of Water and Wastewater, 20a% edicio´n. American Public Health Association/American Water Works Association/Water Environmental Federation, Washington, DC. Bilanovic, D., Battistoni, P., Cecchi, F., Pavan, P., MataAlvarez, 1999. Denitrification under high nitrate concentration and alternating anoxic conditions. Water Res. 33 (15), 3311–3320. Bolzonella, D., Innocenti, L., Pavan, P., Cecchi, F., 2001. Denitrification potential enhancement by addition of anaerobic fermentation products from the organic fraction of municipal solid waste. Water Sci. Technol. 44 (1), 187–194. Can˜izares, P., De Lucas, A., Rodriguez, L., Villasen˜or, J., 2000. Anaerobic uptake of different organic substrates by enhanced biological phosphorus removal sludge. Environ. Technol. 21, 397–405. Chiu, Y.C., Chung, M.S., 2003. Determination of optimal COD/nitrate ratio for biological denitrification. Int. Biodeterior. Biodegrad. 51, 43–49. Christensson, M., Lie, E., Welander, T., 1994. A comparison between ethanol and methanol as carbon sources for denitrification. Water Sci. Technol. 30 (6), 83–90. Contreras, E., Bertola, N., Zaritzky, N., 2001. The application of different techniques to determine activated sludge kinetic parameters in a food industry wastewater. Water SA 27 (2), 169–176.
ARTICLE IN PRESS 3726
A. De Lucas et al. / Water Research 39 (2005) 3715–3726
Cortacans, J.A., 2000. Activated Sludge. Biological Nutrients Removal. (Fangos activos. Eliminacio´n biolo´gica de nutrientes). Civil Engineering publication service, Madrid, Spain. De Lucas, A., Can˜izares, P., Rodriguez, L., Villasen˜or, J., 2000. Determination of biodegradation kinetics of bacterial storage organic substrates through electrolytic respirometry. Environ. Technol. 21, 1111–1118. Ferna´ndez, F.J., 2004. Effect of industrial discharges on Biological nutrient removal from domestic wastewaters (Efecto de vertidos industriales sobre la eliminacio´n biolo´gica de nutrientes de aguas residuales urbanas). Ph.D. Thesis, Faculty of Sciences, University of CastillaLa Mancha, Ciudad Real, Spain. Gerber, A., Mostert, E.S., Winter, C.T., de Villiers, R.H., 1986. The effect of acetate and other short chain carbon compounds on the kinetics of biological nutrient removal. Water SA 12 (1), 7–11. Harder, A., Roels, J.A., 1982. Application of simple structured models in bioengineering. In: Fiechter, A. (Ed.), Advances in Biochemical Engineering and Biotechnology 21. Springer, Berlin, p. 55. Hellinga, C., Schellen, A.A.J.C., Mulder, J.W., van Loosdrecht, M.C.M., Heijnen, J.J., 1998. SHARON process: an innovative method for nitrogen removal from ammoniumrich waste water. Water Sci.Technol. 37 (9), 135–142. Henze, M., 1986. Nitrate versus oxygen utilization rates in wastewater and activated sludge systems. Water Res. 18, 115–122. Henze, M., Grady, C.P.L., Gujer, W., Marais, G.v.R., Matsuo, T., 1987. Activated sludge Model No. 1, IAWPRC Scientific and Technical Report No. 1. International Association on Water-Pollution Research and Control, London. Henze, M., Harremoe¨s, P., Arvin, E., la Cour Jansen, J., 1995a. Wastewater Treatment: Biological and Chemical Processes. Springer, Heidelberg, Berlin. Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M.C., Marais, G.v.R., 1995b. Activated Sludge Model No. 2. IAWQ Scientific and Technical Report No. 3. International Association on Water Quality, London. Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M., Marais, G., van Loosdrecht, M., 1999. Activated Sludge Model No. 2d. Water Sci. Technol. 39 (1), 165–182. Herbert, H.P.F., Yu, H.Q., 2001. Acidification of lactose in wastewater. J. Environ. Eng. 127 (9), 825–831. Jetten, M.S., Wagner, M., Fuerst, J., van Loosdrecht, M., Kuenen, G., Strous, M., 2001. Microbiology and application of the anaerobic ammonium oxidation (‘anammox’) process. Curr. Opin. Biotechnol. 12 (3), 283–288. Kuba, T., van Loosdrecht, M.C.M., Heijnen, J.J., 1996. Phosphorus and nitrogen removal with minimal COD requirement by integration of denitrifying dephosphatation and nitrification in a two sludge system. Water Res. 30 (7), 1702–1710. Kujawa, K., Klapwijk, B., 1999. A method to estimate denitrification potential for predenitrification systems using NUR batch test. Water Res. 33 (10), 2291–2300. Masse, D.I., Masse, L., 2000. Characterisation of wastewater from hog slaughterhouses in Eastern Canada and evaluation of their in-plant wastewater treatment systems. Can. Agric. Eng. 42 (3), 139–146.
Mattsson, A., 1997. Denitrification in a non-nitrifying activated sludge system employing recirculation from a tertiary nitrification unit. Ph.D. Thesis, Civil Engineering Department, Chalmers University of Technology. Musvoto, E.V., Lakay, M.T., Casey, T.G., Wentzel, M.C., Ekama, G.A., 1999. Filamentous organism bulking in nutrient removal activated sludge systems. Paper 8: the effect of nitrate and nitrite. Water SA 25 (4), 397–407. Nemerow, N., Dasgupta, A., 1991. Industrial and Hazardous Waste Treatment. Ed. Van Nostrand Reinhold, New York. Niementowski, C.C., Nelson, P.E., 1977. Effects of temperature, pH and reaction time on the biodegradation of tomato waste in tank-type reactors. J. Food Sci. 41, 339–445. Rodrigo, M.A., Seco, A., Ferrer, J., Penya-Roja, 1999. The effect of sludge age on the deterioration of enhanced biological phosphorus removal process. Environ. Technol. 20, 1055–1065. Rodrı´ guez, L., Villasen˜or, J., Ferna´ndez, F.J., 2004. Long term effect of industrial wastewater discharges on a biological nutrient removal activated sludge. In: International Conference on Wastewater Treatment for Nutrient Removal and Reuse, vol. 1, pp. 172–179. Roeleveld, P.J., van Loosdrecht, M.C.M., 2002. Experience with guidelines for wastewater characterisation in The Netherlands. Water Sci. Technol. 45 (6), 77–87. Sliekers, A.O., Derwort, N., Gomez, J.L., Strous, M., Kuenen, J.G., Jetten, M.S., 2002. Completely autotrophic nitrogen removal over nitrite in one single reactor. Water Res. 36 (10), 2475–2482. Smolders, G.J.F., van der Meij, J., van Loosdrecht, M.C.M., Heijnen, J.J., 1995. A structured metabolic model for anaerobic and aerobic stoichiometry and kinetics of the biological phosphorus removal process. Biotechnol. Bioeng. 47, 277–287. Somiya, I., Tsuno, H., Matsumoto, M., 1988. Phosphorus release storage reaction and organic substrates behaviour in biological phosphorus removal. Water Res. 22 (1), 49–58. So¨zen, S., Orhon, D., 1999. The effect of nitrite correction on the evaluation of the rate of nitrate utilization under anoxic conditions. J. Chem. Technol. Biotechnol. 74, 790–800. Spanjers, H., Vanrolleghem, P.A., 1995. Respirometry as a tool for rapid characterization of wastewater and activated sludge. Water Sci. Technol. 31 (1), 105–114. Tam, N.F.Y., Wong, Y.S., Leung, G., 1992. Effect of exogenous carbon sources on removal of inorganic nutrient by nitrification-denitrification proceses. Water Res. 26 (9), 1229–1236. van Loosdrecht, M.C.M., Jetten, M.S.M., 1998. Microbiological conversions in nitrogen removal. Water Sci. Technol. 38 (1), 1–7. Vitousek, P.M., Mooney, H.A., Lubchenco, J., Melillo, J.M., 1997. Human domination of Earth’s ecosystem. Science 277, 494–499. Wachtmeister, A., Kuba, T., van Loosdrecht, M.C.M., Heijnen, J.J., 1997. A sludge characterisation assay for aerobic and denitrifiying phosphorus removing sludge. Water Res. 31 (3), 471–478. Zeng, R.J., Yuan, Z., Keller, J., 2003. Enrichment of denitrifying glycogen-accumulating organisms in anaerobic/anoxic activated sludge system. Biotechnol. Bioeng. 81 (4), 397–404.
Water Research 39 (2005) 3715–3726 www.elsevier.com/locate/watres
Denitrification potential of industrial wastewaters A. De Lucas, L. Rodrı´ guez, J. Villasen˜or, F.J. Ferna´ndez Department of Chemical Engineering, University of Castilla-La Mancha, Avenida Camilo Jose´ Cela S/N E-13071, Ciudad Real, Spain Received 21 February 2005; received in revised form 24 June 2005; accepted 29 June 2005
Abstract Nitrate utilisation batch experiments have been carried out with different agro-food industrial wastewaters. The results obtained were modelled using a modification of the ASM2d to determine the main kinetic and stoichiometric parameters: denitrification rate with fermentation products (qNSA) and with fermentable substrates (qNSF); fermentation products and fermentable substrates consumption rates, (qSA) and (qSF), respectively; yield coefficient with fermentation products (YHSA) and with fermentable substrates (YHSF) and saturation coefficient for growth on fermentation products (KHSA). It was found that fermentation products (SA) and fermentable substrates (SF) consumption rates can be assumed to be both constant and independent of the agro-food industrial wastewater used, with values of 14.6 and 5.4 –1 (mg SCOD g COD-X–1 OHO h ), respectively. The mean denitrification rates obtained with each agro-food industrial –1 wastewater ranged from 1.5 to 2.5 (mg SN g COD-X–1 OHO h ), values slightly lower to the obtained with domestic –1 wastewater (2.7 mg SN g COD-X–1 h ). Given the substrate consumption rate and the denitrification rate, the yields OHO with fermentation products and fermentable substrates for each wastewater were determined. The values obtained for the yields ranged from 0.50 to 0.69 on using SA and from 0.50 to 0.70 on using SF. Having determined these parameters, the denitrification potential (DNP) was evaluated for each agro-food wastewater. Wastewater from slaughterhouses gave the lowest DNP (22.6 mg SN g S–1 COD), whereas wastewater from potato and tomato processing plants gave the highest values (126.5 and 128.2 mg SN g S–1 COD, respectively), which are very similar to that obtained with domestic wastewater (130.3 and 128.2 mg SN g S–1 COD). Taking into account the low N/COD ratio presented in the agro-food wastewaters, the last wastewaters are the most indicated to enhance the nitrogen removal in the WWTP. Finally, a relationship between the denitrification rates of several agro-food industrial wastewaters and their respective SA/(SA+SF) ratio was obtained. This relationship can be used to estimate the mean denitrification rate of an agro-food industrial wastewater. r 2005 Elsevier Ltd. All rights reserved. Keywords: Industrial wastewater; Activated sludge; Denitrification potential; Denitrification rate; Nitrite acumulation
1. Introduction
Corresponding author. Tel.: +34 926 295300;
fax: +34 926 295242. E-mail address: [email protected] (F.J. Ferna´ndez).
The excessive application of fertilizers, intensive exploitation of farms and the significant contribution from industry have increased the nitrogen load discharged to receiving waterways (Vitousek et al., 1997). This has led to a decrease in water quality and has caused health problems related to oxidised forms of
0043-1354/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2005.06.024
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3716
Nomenclature bH KHSA KHNO3 qSA qSF qend qNend qNSA qNSF SA SALK SCa SCOD SF
decay rate (d–1) saturation coefficient for growth on fermentation products (g COD m–3) saturation/inhibition coefficient for nitrate (g COD m–3) fermentation products consumption rate –1 (g COD g COD-X–1 OHO h ) fermentable substrates consumption rate –1 (g COD g COD-X–1 OHO h ) endogenous substrates degradation rate –1 (g COD g COD-X–1 OHO h ) denitrification rate with endogenous sub–1 strates (g N g COD-X–1 OHO h ) denitrification rate with fermentation pro–1 ducts (g N g COD-X–1 OHO h ) denitrification rate with fermentable sub–1 strates (g N g COD-X–1 OHO h ) fermentation products concentration (g COD m–3) alkalinity concentration (mol m–3) calcium concentration (g Ca m–3) COD concentration (g COD m–3) fermentable substrates (g COD m–3)
nitrogen. As a result, more stringent regulations have been approved in an effort to deal with this problem. In Europe, directive 91/271 prescribes the nitrogen standards for treated wastewater discharged into sensitive areas and establishes a timetable to reach an acceptable level of effluent quality. The most economic options for the removal of nitrogen from wastewater are biological processes and several methods already exist to achieve this goal: nitrification–denitrification, single reactor system for high ammonium removal over nitrite (SHARON) (Hellinga et al., 1998), anaerobic ammonium oxidation (ANAMMOX) (Jetten et al., 2001) and completely autotrophic nitrogen removal over nitrite (CANON) (Sliekers et al., 2002). Each of these processes exploits one of the different available pathways in the nitrogen cycle (van Loosdrecht and Jetten, 1998). Currently, the most developed system is the nitrification–denitrification process, which consists of an initial nitrification stage, accomplished by autotrophic bacteria, in which ammonia is oxidized to nitrite by Nitrosomonas sp. The nitrite is subsequently oxidised to nitrate by Nitrobacter sp., this process is outlined in Eqs. (1) and (2). A second denitrification stage (Eq. (3)) is then carried out by numerous facultative heterotrophic bacteria. In this second process, nitrate obtained in the nitrification stage is reduced to molecular nitrogen in four steps, using the substrates contained in the wastewater as an electron donor. These steps involve the
SI SK SMg SN SN2NO3 SN2NO2 SNa SPT SP2PO4 XPAO XGAO XAUT XOHO XPT XSS YHSA YHSF
inert COD concentration (g COD m–3) potasium concentration (g K m–3) magnesium concentration (g Mg m–3) nitrogen concentration (g N m–3) nitrate–nitrogen concentration (g N m–3) nitrite–nitrogen concentration (g N m–3) sodium concentration (g Na m–3) phosphorus total concentration (g P m–3) orthophosphate concentration (g P m–3) phosphate accumulating Organism concentration (g COD-XPAO m–3) glycogen accumulating Organism concentration (g COD-XGAO m–3) autotrophic organisms concentration (g COD-XAUT m–3) ordinary heterotrophic Organisms concentration (g COD-XOHO m–3) phosphorus content in the sludge concentration (g P g VSS) suspended solids concentration (g m–3) yield coefficient with fermentation products (g COD g–1 COD) yield coefficient with fermentable substrates (g COD g–1 COD)
reduction of nitrate to nitrite, which is further reduced to one or more of the gaseous nitrogen products (nitric oxide, nitrous oxide and nitrogen gas) that will be finally released to the atmosphere (Henze et al., 1995a). þ 2NHþ 4 þ 3O2 ! 2NO2 þ 2H2 O þ 4H ,
(1)
2NO 2 þ O2 ! 2NO3 ,
(2)
NO 3 ! NO2 ! NOðgÞ ! N2 OðgÞ ! N2ðgÞ .
(3)
The overall rate and extension of the denitrification process depends mainly on the biodegradability characteristics of the electron donor used (Bolzonella et al., 2001) and on the final N/COD ratio in the bioreactor (Chiu and Chung, 2003). For these reasons, the denitrification process can be enhanced by adding readily biodegradable carbon sources—such as acetic acid and methanol—or carbon sources presenting high COD/N ratios to the wastewater. However, the addition of these carbon sources is expensive and increases the cost of treating the wastewater. The cost involved in these processes has led to carry out studies using waste carbon sources such as industrial wastes and organic fraction of municipal wastes (Bilanovic et al., 1999). Due to EU regulations there are, at present, many wastewater treatment plants (WWTP) working with this nitrogen removal process. These plants are designed to treat typical domestic wastewater by removing the
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
nitrogen from it biologically. However, in many cases, these systems would receive wastewater with high nitrogen content or with different characteristics, mainly biodegradability, due to industrial discharges (Rodrı´ guez et al., 2004). These changes in the wastewater composition can either increase the nitrate concentration in the effluent, due to the absence of an efficient denitrification process caused by inhibition phenomena or the absence of a suitable electron donor, or decrease the nitrate concentration. This latter effect can be caused by a high biodegradability or low ratio N/COD of the industrial wastewater. One of the best ways to describe the process behaviours and to predict its evolution is the mathematical modelization (Harder and Roels, 1982). In the case of the denitrification process, it has been described in several models (Henze et al., 1987, 1995b, 1999). In the ASM2d (Henze et al., 1999) the denitrification process is carried out by both ordinary heterotrophic organisms (OHO) and phosphate accumulating organisms (PAOs). The OHO denitrification process is modelled as two parallel processes, which consume nitrates, as electron acceptor, to oxidise the two biodegradable COD fractions, SA and SF. For both processes the same kinetics and stoichiometrics coefficients are assumed. The PAO denitrification process is carried out at expense of the internal organic storage products. In this context, the aim of the work described here was to study the effects of agro-food industrial discharges on the denitrification stage of a biological nutrient removal process. The results obtained in the nitrate utilisation experiments were modelled using a modified ASM2d model. As a result of the modelisation, the values of the most important parameters for the denitrification process (i.e., qSA, qSF, qNSA, qNSF, qNend, YHSA, YHSF, and KHSA) were evaluated. Finally, the denitrification potential (DNP) of each agro-food industrial wastewater was evaluated and a relationship between the mean denitrification rate for each agro-food industrial wastewater and their respective SA/(SA+SF) ratio was proposed.
2. Materials and methods 2.1. Activated sludge The activated sludge used in this work was taken from sequence batch reactors (SBRs) working with the A/O process for biological nutrient removal (BNR). Further details concerning the SBRs experimental set-up can be found in previous publications (Can˜izares et al., 2000; De Lucas et al., 2000). The wastewater used for the cultivation of microorganisms was prepared daily and was composed of real domestic wastewater—taken from the primary clarifiers
3717
of a full-scale wastewater treatment plant—modified by mixing with synthetic wastewater. The composition of the synthetic wastewater and the characteristics of the resulting wastewater are shown in Table 1 (sections a and b, respectively). The criteria for reaching steady state in the SBRs were (i) to reach a stable biomass concentration and (ii) to attain stable effluent treated water characteristics. 2.2. Pre-treatment of activated sludge The activated sludge used in this work was acclimatized to BNR. For this reason, activated sludge was enriched on PAO and glycogen accumulating organisms (GAO). Under anoxic conditions, PAO and GAO microorganisms could use intracellularly stored substrates to denitrify (Kuba et al., 1996; Zeng et al., 2003), thus consuming nitrates but not COD from the bulk liquid. Pre-treatment of the sludge was necessary in order to avoid this process, which would negatively influence the determination of the stoichiometry and the kinetics parameters of the denitrification process carried out by ordinary heterotrophic organisms (OHO). This pre-treatment consisted of an excessive aeration stage to exhaust all of the intracellular polymers accumulated on PAO and GAO organisms. Once this pre-treatment had been performed, only the OHO of the sludge were able to denitrify. Prior to all batch experiments, the sludge was washed with osmotized water in aerobic conditions in order to remove the absorbed substrates from the surface of the flocks. 2.3. Industrial wastewater The synthetic agro-food industrial wastewaters used in this work were prepared according to the procedure Table 1 Concentration (a) Composition of the synthetic wastewater Component 322 Sodium acetate (mg L1): (NH4)2SO4 (mg L1): 74.2 KH2PO4 (mg L1): 44.5 115.0 NaHCO3 (mg L1): MgSO4 7H2O(mg L1): 50.0 CaCl2 (mg L1): 30.0 3.0 (NH4)2 Fe(SO4)2 (mg L1): (b) Characteristics of the resulting wastewater used for the cultivation of microorganisms Parameter 340 (46) COD (mg L1): BOD5 (mg L1): 181 (22) 9.1 (2.6) P (mg L1): N (mg L1): 15.8 (3.7) pH 7.3 (0.1)
ARTICLE IN PRESS 3718
A. De Lucas et al. / Water Research 39 (2005) 3715–3726
described in literature (Table 2). Agro-food wastewaters are characterised by its very low N/COD and P/COD ratios, because of this, nutrients were added synthetically in order to avoid nutrients limitations. The wastewaters used in this work were: wastewaters from cheese industries, milk bottling industries, slaughterhouses, potato processing industries, beet sugar processing industries, tomato processing industries and winery industries.
2.4. Batch experiments The denitrification processes for the agro-food industrial wastewaters and the domestic wastewater were studied by carrying out several batch tests in closed reactors, using the pre-treated biomass from the SBRs. Each batch reactor had a total volume of 260 mL. The reactors were filled with 250 mL of the pre-treated biomass with nutrients and industrial wastewater to give the following approximate nutrients and COD final concentrations: S PT ¼ 10 mg L1 , SN2NH4 ¼ 15 mg L1 , SK ¼ 12 mg L–1, SNa ¼ 14 mg L–1, SMg ¼ 5 mg L–1, SCa ¼ 10 mg L–1, SCOD ¼ 300 mg L–1. In this way, the dead gas volume inside the closed reactor is minimised and this reduces the adverse effect of oxygen on the denitrification process. In order to maintain the sludge suspended, avoiding oxygen transfer from the gas phase, a very slow stirring rate was applied. Experiments were carried out with an SCOD/XOHO ratio of 0.18 g COD g COD-X–1 OHO. Nitrate was added continuously to avoid substrate inhibition due to very high nitrate concentrations, but ensuring the continuous presence of it (Wachtmeister et al., 1997). The pH in the batch reactors oscillated between 7.2 and 7.4 and the temperature was continuously maintained at 20 1C by means of an incubator (ISCO FTD 220). Two reference tests were simultaneously carried out. One of these tests consisted of a batch reactor filled with 250 mL of the pre-treated biomass and nutrients
Table 2 References for the preparation of the synthetic agro-food industrial wastewater Industrial wastewater
Reference
Potato processing industries Tomato processing industries Beet-sugar processing industries Cheese industries Milk bottling industries Winery industries Slaughterhouses
Contreras et al. (2001) Niementowski and Nelson (1977) Nemerow and Dasgupta (1991) Herbert and Yu (2001) Alvarez-Mateos et al. (2000) Nemerow and Dasgupta (1991) Masse and Masse (2000)
(blank test). From the results obtained in this experiment the endogenous denitrification and the decay rate can be obtained. The other reference test (maximum test) consisted of a batch reactor filled with 250 mL of the pre-treated biomass, nutrients and sodium acetate to reach the nutrient and COD final concentrations outlined above. Sodium acetate is known to be a very easily biodegradable substrate that offers the highest denitrification rates (Gerber et al., 1986; Somiya et al., 1988; Tam et al., 1992). The results obtained in this experiment enabled the maximum denitrification rates to be evaluated. 2.5. Sampling procedure and analytical methods Samples were taken by syringe and were immediately filtered through a Millipore glass fibre filter. After filtration, samples were analysed for the following parameters: S COD ; SN2NO2 ; S NNO3 ; SPPO4 and SN2NH4 : In addition, XSS, volatile suspended solids (VSS) and XPT were determined in the activated sludge. SP2PO4 , XPT and SN2NH4 were analysed to confirm the absence of denitrifying phosphate accumulation and nitrification. All analyses of wastewater and activated sludge were performed as described by APHA (1998). The biodegradable COD fractions, SA and SF, were determined using electrolytic respirometry (Spanjers and Vanrolleghem, 1995). 2.6. Mathematical model A large number of models can be used to evaluate the results obtained in the anoxic experiments. These mathematical models usually have complex structures and a very large number of equations and parameters. The model used in this work is a simplification of the ASM2d model (Henze et al., 1999), in which a couple of modifications were made. These modifications are as follows: 1. ASM2d does not take into account the possible accumulation of nitrites in the bulk liquid as the reduction of nitrite to nitrogen gas (Eq. (3)) is supposed to be very fast. However, nitrite accumulation has been reported in numerous studies (Henze, 1986; So¨zen and Orhon, 1999). Henze proposed in 1986 that nitrite accumulation could occur due to at least two different causes: excessive growth of biomass that is able to reduce nitrate to nitrite and/ or the inhibition of the nitrite reductase enzyme. It is important to take the possibility of this accumulation into account for two reasons; firstly, neglecting it would increase the real denitrification rate obtained if nitrates are supposed to be completely reduced and, secondly, the real denitrification rate will be
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
decreased if oxidized nitrogen were measured. The corrected value for the amount of nitrate–nitrogen completely reduced during the anoxic experiment can be evaluated by taking into account the variations in nitrates and nitrites, Eq. (4), (So¨zen and Orhon, 1999). DS NCorrected ¼ ðDSNNO3 0:6DSNNO2 Þ.
(4)
2. In this equation, SNCorrected represents the theoretical amount of nitrates that, when completely reduced to nitrogen gas, will supply the same amount of electron acceptor as consumed in the batch reactor. 3. In the ASM2d model the maximum specific carbon degradation rate due to the denitrification process is supposed to be the same regardless of the substrate used. However, studies carried out with different substrates have shown that the rate depends on the characteristics of the substrate (Alleman and Irvine, 1980; Christensson et al., 1994; Bolzonella et al., 2001). For this reason, we changed the maximum specific substrate consumption rate expression from ASM2d to allow two rates: one for the utilisation of SA and the other for the utilisation of SF, qSA and qSF, respectively. The experimental conditions (anaerobic environment, excess concentrations of ammonium, alkalinity, phosphorus, nitrate and fermentable substrates during the experiments) and the low sensitivity of certain parameters allows the ASM2d model to be simplified to a set of dynamic equations as follows, Eqs. (5)–(11): Denitrification with SA SNO3 dSN ¼ qN2SA dt SA K H2NO3 þ SNO3 SA SA X OHO . ð5Þ K H2SA þ SA S F þ S A Denitrification with SF SNO3 dSN SF ¼ qN2SF X OHO . dt SF K H2NO3 þ S NO3 S F þ S A
(7)
Consumption of SA S NO3 dS A ¼ qSA dt K H2NO3 þ SNO3 SA SA X OHO . K H2SA þ SA SF þ SA
Consumption of SF S NO3 dSF SF ¼ qSF X OHO . dt K H2NO3 þ SNO3 SF þ SA
(9)
Anoxic growth dX H ¼ dt
Y H2SA qSA
SA SA K H2SA þ SA SF þ SA K H2NO3 þ SNO3 SF bH þY H2SF qSF SF þ SA SNO3 S NO3 X OHO . ð10Þ K H2NO3 þ SNO3
Relationship between variables qN2SF ¼ qSF
1 Y H2SF , 2:86
and qN2SA ¼ qSA
1 Y H2SA . 2:86
ð11Þ
The blank test enables the endogenous denitrification rate (qN–end) and the decay rate (bH) to be obtained. Having obtained these values, the remaining equations can be solved. The simultaneous estimation of qSA, qSF, qN–SA, qN–SF and KH–SA involved assuming an initial set of parameter values and calculating theoretical nitrate, COD and growth of the biomass profiles in the batch reactor by solving simultaneously the Eqs. (5)–(11) using the Newton method. A minimal residual sum of squared errors, associated with the difference between these theoretical curves and the experimental results, have to be reached. The values of qSA, qSF, qN–SA, qN–SF and KH–SA associated with that minimum constitute the best estimates of these parameters. In an effort to ensure a correct estimation, the initial supposed values for the parameters ranged from those corresponding to the blank reference test to those for the maximum reference test. 2.7. DNP
(6) Denitrification with endogenous substrate dSN ¼ qN2end X OHO . dt end
3719
ð8Þ
Once the kinetic and stoichiometric values had been obtained for the denitrification process with the wastewaters under investigation, the DNPs were determined (Kujawa and Klapwijk, 1999). The DNP was calculated using Eqs. (12) and (13), in which DNP is expressed as the amount of nitrates that can be denitrified in the SA and SF fractions of COD. DNPSA ¼
DS N2Corrected , DSA
(12)
DNPSF ¼
DSN2Corrected . DS F
(13)
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3720
The overall DNP of the substrates contained in each industrial wastewater depends on the composition of the wastewater. This value can be calculated using Eq. (14), where SA, SF and SCOD are the concentrations of the different COD fractions in the influent wastewater. DNP ¼
S A DNPSA þ SF DNPSF . SCOD
(14)
3. Results and discussion 3.1. Wastewater characterisation The quality of predictions achieved by activated sludge models depends on the accuracy of the wastewater characterisation and on the calibration of the parameters. Therefore, the influent wastewater was characterised using a reliable method proposed by the Dutch Foundation of Applied Water Research (STOWA) (Roeleveld and van Loosdrecht, 2002). The wastewater characteristics for each batch experiment are presented in Table 3.
3.2. Biomass characterisation The activated sludge used in this work was assumed to be composed mainly of PAO, GAO, OHO and autotrophic organisms (AUT). The concentration of each of these microorganisms was determined (Ferna´ndez, 2004) using a metabolic model (Smolders et al., 1995) and formulae proposed in the literature (Rodrigo et al., 1999; Cortacans, 2000). The concentrations of bacteria in the activated sludge were as follows: PAO ¼ 275 (mg COD-XPAO L–1), GAO ¼ 410 (mg COD-XGAO L–1), AUT ¼ 175 (mg COD-XAUT L–1) and OHO ¼ 1767 (mg COD-XOHO L–1).
3.3. Endogenous denitrification rate and maximum denitrification rate The maximum specific denitrification rate using the endogenous substrate (qNend) and decay rate (bH) were determined from the results of the blank experiment (Fig. 1). During this experiment, the accumulation of nitrite was negligible and, as a result, nitrite was not taken into account in Fig. 1. The maximum specific rate obtained for endogenous denitrification was 1 0:3 ðmg SN-Corrected g COD-X 1 This value is OHO h Þ. slightly lower than that reported in the literature for similar conditions (Kujawa and Klapwijk, 1999). The rate constant for decay (bH) obtained from the experimental results was 0.39 d–1, a value very similar to that corresponding to the ASM2d model (0.40 d–1). The maximum specific denitrification rate and the maximum specific COD consumption rate were obtained using the results from the reference test carried out using acetate as the organic substrate (Fig. 2). In this case, nitrite was accumulated in the bulk liquid and it was therefore necessary to take its effect into account. The nitrate utilisation rate was thus calculated using Eq. (4). The maximum specific substrate con–1 sumption rate was 24.7 (mg SCOD g COD-X–1 OHO h ) and the maximum specific denitrification rate was –1 3.4 (mg SNCorrected g COD-X–1 OHO h ). Both values are slightly higher than those reported by Kujawa and Klapwijk (1999) for similar experiments. This result could be due to the acclimatization of the sludge to acetate because this substrate was fed into the SBRs from which the activated sludge was taken. The heterotrophic yield, YH, was determined by means of the COD and nitrogen balances using (Eq. (15)). dSCOD 2:86 dS N2NO3 ¼ . 1 YH dt dt
(15)
The value obtained was 0.60 (mg COD-XOHO mg S–1 COD).
Table 3 Industrial wastewater characterisation according to STOWA guidelines Parameter
Wastewater
Symbol
Unit
AF1
AF2
AF3
AF4
AF5
AF6
AF7
DWW
SF SA SN2NH4 SN2NO3 SP2PO4 SI SALK
g COD m3 g COD m3 g N m3 g N m3 g P m3 g COD m3 mol m3
208 58 17 0 13.5 85 1.43
207 29 18.5 0 9.5 68 1.53
56 7 22 0 9 238 1.37
183 33 18.4 0 10.2 57 1.35
165 48 15 0 11.2 86 1.39
201 96 19 0 10.1 67 1.85
180 73 19.5 0 9.7 59 1.85
173 41 20.2 0 9.4 21 4.12
AF1: cheese industries; AF2: milk bottling industries; AF3: slaughterhouses; AF4: potatoes processing; AF5: beet sugar processing; AF6: tomatoes processing; AF7: winery industries; DWW: domestic wastewater.
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
35
2000
XOHO (mg COD·L-1)
SN (mg·L-1)
28 21 14 7 0
3721
0
1
2
3
4
1500
1000
500
0
5
0
1
2
4
3
t (h)
5
t (h)
Fig. 1. Evolution of SN2NO3 (m) and biomass (K) during the blank test. Solid lines indicate the model fit.
2000
500
35
1500
300 1000 200 SCOD
500
100
0
0 0
1
2
3
4
5
t (h)
28 SN (mg/L)
SCOD (mg·L-1)
400
XOHO (mg COD·L-1)
XOHO
21 14 SN-Corrected 7
SN-NO3
SN-NO2 0
0
1
3
2
4
5
t (h)
Fig. 2. Evolution of SN2NO3 (n), SN2NO2 (&), SN–Corrected (K), XOHO expressed as COD (’) and SCOD (m) during the acetate reference test. Solid lines indicate the model fit.
3.4. Denitrification with agro-food industrial wastewater Denitrification with agro-food wastewaters from cheese industries, milk bottling industries, slaughterhouses, potato processing industries, beet sugar processing industries, tomato processing industries and winery industries were studied. The proposed model was employed to simultaneously determine the maximum specific substrate consumption rates for SA and SF during the denitrification process in all the experiments in order to give comparable values. These values were 14.6 (mg SCOD g COD–1 X–1 OHO h ) for the utilisation of SA and 5.4 (mg SCOD g –1 COD-X–1 OHO h ) for the utilisation of SF. The experimental results and the results predicted by the model are shown in Fig. 3. Data predicted by the model fit very well to the experimental results. Consequently, the agro-food industrial wastewater studied can be taken—at least in terms of COD removal—to be a mixture in different
percentages of SA and SF. As can be seen, the maximum specific substrate consumption rate for SA –1 (14.6 mg SCOD g COD-X–1 OHO h ) is lower than that –1 obtained for acetate (24.7 mg SCOD g COD-X–1 OHO h ). This can be explained because SA concerns not only acetate but also by other fermentation products. In all the experiments carried out with agro-food industrial wastewater, nitrite was accumulated in the bulk liquid and Eq. (4) was therefore used to determine the maximum specific denitrification rates. The values obtained for this parameter ranged from 1.58 to 2.55 –1 (mg SNCorrected g SCOD-X–1 OHO h ) on using SA and –1 from 0.57 to 0.92 (mg SN–Corrected g SCOD-X–1 OHO h ) on using SF. These denitrification rates were used in conjunction with the agro-food industrial wastewater characterisation to give the mean denitrification rates for each wastewater. The mean denitrification rates obtained were very similar to that reported previously for domestic wastewater (Mattsson, 1997; Bolzonella et al., 2001).
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3722
Cheese industries 500
2000
1250
300
1000 200
750
SCOD
500
100
50 SN (mg·L-1)
1500
60 XOHO (mg COD·L-1)
SCOD (mg·L-1)
1750
XOHO
400
0
2
4
0 8 10 12 14 16 18 20 t (h)
6
SN-Corrected
30 SN-NO3
20 10
250 0
40
SN-NO2
0 0
2
4
6
8
10 12 14 16 18 20 t (h)
Milk bottling industries 2000 1500 1250
300
1000 200
750
SCOD
500
100
60 50 SN (mg·L-1)
SCOD (mg·L-1)
1750
XOHO
400
XOHO (mg COD·L-1)
500
0
2
4
6
8
SN-Corrected
30 20
SN-NO3 SN-NO2
10
250 0
40
0 10 12 14 16 18 20 t (h)
0 0
2
4
6
8 10 12 14 16 18 20 t (h)
Slaughter houses 2000
SCOD (mg·L-1)
1500 1250
300 SCOD
1000
200
750 500
100
60 50 SN (mg·L-1)
XOHO 1750
400
XOHO (mg COD·L-1)
500
1
0
3 t (h)
2
4
6
5
30
SN-Corrected
20
SN-NO3
10
250 0
40
0
0
SN-NO2 1
0
3 t (h)
2
4
6
5
Potatoes processing waste water 500
2000 1500 1250
300
1000 200
750 SCOD
100
500 250
0 0
2
4
6
8
0 10 12 14 16 18 20 t (h)
60 50 SN (mg·L-1)
SCOD (mg·L-1)
1750
XOHO (mg COD·L-1)
XOHO
400
40 SN-Corrected
30 20
SN-NO3 SN-NO2
10 0 0
2
4
6
8
10 12 14 16 t (h)
18 20
Fig. 3. Evolution of SN2NO3 (n), SN2NO2 (&), SN–Corrected (K), XOHO expressed as COD (’) and SCOD (m) during the experiments carried out with agro-food industries wastewater. Solid lines indicate the model fit.
The maximum specific substrate consumption rates and the maximum specific denitrification rates were used to determine the yields of the process. In addition, the value for KHSA was determined (5 mg SCOD L–1), the
value obtained is very similar to that proposed by ASM2d, (4 mg SCOD L–1). A summary of the values for the kinetics and stoichiometrics parameters of the denitrification process is given in Table 4.
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3723
Beet-sugar processing wastewater 500
2000
1000 200
750 SCOD
100
500
50 SN (mg·L-1)
SCOD (mg·L-1)
1500 1250
300
60
1750 XOHO (mg COD·L-1)
XOHO
400
2
4
6
SN-Corrected
30 20
SN-NO3
10
250 0 0
40
0
0 8 10 12 14 16 18 20 t (h)
SN-NO2 0
2
4
6
8
10 12 14 16 t (h)
18 20
Tomatoes processing wastewater 2000
SCOD (mg·L-1)
1500 1250
300
1000 200
750 SCOD
100
60
1750
500
50 SN (mg·L-1)
XOHO
400
XOHO (mg COD·L-1)
500
0
2
4
6
2000
1250 1000
200
750 SCOD
500
2
4
6
4
6
8
10 12 14 16 18 20 t (h)
40 SN-Corrected
30 20
SN-NO3
10
250 0
SN-NO2 2
50 SN (mg·L-1)
SCOD (mg·L-1)
1500
300
0
SN-NO3
60
1750 XOHO (mg COD·L-1)
XOHO
100
SN-Corrected
20
0 0 0 8 10 12 14 16 18 20 t (h) Winery industries
500 400
30
10
250 0
40
0 8 10 12 14 16 18 20 t (h)
0
SN-NO2 0
2
4
6
8
10 12 14 16 t (h)
18 20
Fig. 3. (Continued)
The DNP was calculated on the basis of these results (Table 4). The industrial wastewaters with the highest DNPs were those arising from tomato and potato processing, with potential values of 128.3 and 126.5 (mg SN–Corrected g S–1 COD), respectively. This result can be explained in terms of the composition and characteristics of these wastewaters, mainly related to their fermentation products and fermentable substrates concentration. The lowest DNP was obtained with slaughterhouse wastewater, with a value of 22.6 (mg SN–Corrected g S–1 COD). Parallel to these experiments carried out with agrofood wastewater, an experiment was carried out with domestic wastewater. The objective of this experiment is to establish a reference to compare the results
obtained with agro-food wastewater. The experimental results were modelled (Fig. 4) obtaining the values for the kinetics and stoichiometrics parameters shown in Table 4. The yields and the half-saturation constant obtained were similar to that obtained with agro-food wastewaters. However, the consumption rates of the COD fractions, SA and SF, and the denitrification rates associated to these fractions were slightly higher than the obtained with agro-food wastewater. The DNP obtained with domestic wastewater 130 (mg SN–Corrected g S–1 COD) was higher than the obtained with agro-food wastewater, only wastewaters from tomatoes and potatoes processing industries presented similar values. These results indicate that the
ARTICLE IN PRESS A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3724
Table 4 Kinetic and stoichiometric coefficients for the denitrification process Wastewater
Parameter Symbol
Unit
AF1
qSA qSF qN–end qN–SA qN–SF YH–SA YH–SF KH–SA
1 mg SCOD g COD-X1 OHO h 1 mg SCOD g COD-X1 h OHO 1 mg SN–Corrected g COD-XOHO h1 1 mg SN–Corrected g COD-X1 OHO h 1 mg SN–Corrected g COD-XOHO h1 mg COD-XOHO mg S1 COD mg COD-XOHO mg S1 COD 1 mg SCOD L
14.6 5.4 0.3 2.00 0.66 0.60 0.65 5.0
DNP
mg SN–Corrected g S1 COD
95.1
AF2
AF3
14.6 5.4 0.3 2.55 0.76 0.50 0.60 5.0
AF4
14.6 5.4 0.3 1.94 0.57 0.62 0.70 5.0
112.6
22.6
AF5
14.6 5.4 0.3 2.55 0.85 0.50 0.58 5.0
AF6
14.6 5.4 0.3 1.58 0.87 0.69 0.54 5.0
126.5
106.3
AF7
14.6 5.4 0.3 1.88 0.92 0.63 0.51 5.0 128.2
DWW
14.6 5.4 0.3 2.14 0.81 0.58 0.57 5.0
20.5 8.4 0.3 3.26 1.17 0.65 0.60 4.0
120.7
130.3
AF1: cheese industries; AF2: milk bottling industries; AF3: slaughterhouses; AF4: potatoes processing; AF5: beet sugar processing; AF6: tomatoes processing; AF7: winery industries; DWW: domestic wastewater.
500
2000 XOHO
1500
300 1000 200 500
100
0
2
4
8
6
10
40 SN-Corrected
30 20
SN-NO3
10
SCOD 0
50 SN (mg/L)
SCOD (mg·L-1)
400
XOHO (mg COD·L-1)
60
12
SN-NO2
0 14
0
t (h)
0
2
4
6
8
10
12
14
t (h)
Fig. 4. Evolution of SN2NO3 (n), SN2NO2 (&), SN–Corrected (K), XOHO expressed as COD (’) and SCOD (m) during the experiment carried out with domestic wastewater. Solid lines indicate the model fit.
discharge of these agro-food wastewaters would not significantly influence the denitrification rate in a WWTP. However, the discharge of the agro-food wastewaters would increase the extension of the denitrification process, because the agro-food wastewaters supply considerable amounts of readily biodegradable substrates presenting a very low N/COD ratio. The enhancement of the denitrification process would reduce the nitrate and nitrite concentrations, reducing therefore, the filament proliferation caused by these compounds (Musvoto et al., 1999). The mean denitrification rates obtained for each wastewater mainly depended on the SA and SF concentrations. The mean denitrification rates vs. the SA/(SA+SF) ratio obtained for each agro-food wastewater are shown in Fig. 5 along with the limiting values in each case; the theoretical maximum specific denitrification rate for pure SF wastewater
(obtained in the modelisation) and the experimental maximum specific denitrification rate for pure SA wastewater (reference test carried out with acetate). These data were fitted to an empirical equation (Eq. (16)) using a non-linear regression method. The values obtained previously were used as initial guesses for the parameters: –1 qNSA ¼ 3.40 (mg SNCorrected g SCOD-X–1 OHO h )
maximum specific denitrification rate (obtained in the maximum reference test). qNSF ¼ 0.70 (mg SN–Corrected g SCOD-XOHO–1 h–1) average maximum specific denitrification rate (obtained with the SF fraction of the agro-food wastewaters). A ðq qN2SF Þ K S SþS qN ¼ qN2SA 1 N2SA (16) e A F . qN2SA
ARTICLE IN PRESS
mean qN (mg SN·g COD-XOHO-1·h-1)
A. De Lucas et al. / Water Research 39 (2005) 3715–3726
3.5 3.0
Domestic wastewater
Winery industries
SA
Tomatoes 2.5 Potatoes processing processing Beet-sugar 2.0 Milk bottling processings industries Cheese 1.5 industries
1.0
Slaughter houses SF
0.5 0.0
0.0
0.2
0.4
0.6
0.8
1.0
SA/(SA+SF) Fig. 5. Mean denitrification rate for the studied agro-food wastewater. Solid line indicates the predictions of the empirical equation used.
The estimated values for the parameters are as follows:
3725
wastewater. Only wastewater from a winery and tomato processing plant gave similar denitrification rates, which could be explained because of the higher SA concentration in these wastewaters. These results indicate that the discharge of the industrial wastewater studied would not increase the denitrification rate in a WWTP. However, the denitrification process could be enhanced taking into account the combined effect of the denitrification rate and the N/COD ratio of the agro-food wastewaters. The different mean denitrification rates of agro-food wastewater can be calculated using the proposed equation (Eq. (16)), which predicts the denitrification rate for agro-food industrial wastewater using only one variable—the SA/(SA+SF) ratio. The industrial wastewaters with the highest DNP values were those from tomato and potato processing industries, each of which has a potential of about 127 mg SN–Corrected g S–1 COD. The combination of results, DNP and maximum specific denitrification rate, indicate that the most suitable wastewater for nitrogen removal is that arising from tomato processing industries.
1 qN2SA ¼ 3:49 ðmg S N2Corrected g COD-X1 OHO h Þ, 1 qN2SF ¼ 0:72 ðmg SN2Corrected g COD-X1 OHO h Þ,
K ¼ 3:01 ðg S COD g S 1 COD Þ. This equation relates the mean denitrification rate obtained for the agro-food industrial wastewater studied with its SA/(SA+SF) ratio. This equation can be used to predict the mean denitrification rate for agro-food industrial wastewaters. In Fig. 5, the mean denitrification rate for the domestic wastewater is also represented. The mean denitrification rate obtained with domestic wastewater is higher than the obtained with the agrofood wastewaters, which could be explained taking into account that the biomass was acclimatised to the substrates contained in this wastewaters and because of the more complex structure of the substrates contained in the agro-food wastewaters.
4. Conclusions The substrate utilisation rate is not the same for fermentation products (SA) and fermentable substrates (SF), with the maximum specific utilisation rate for each of these parameters having values of 14.6 and –1 5.4 (SCOD g COD-X–1 OHO h ) regardless of the industrial agro-food wastewater used. Owing to the different yields of the substrates contained in the industrial wastewaters, the maximum specific denitrification rate for each fraction of the COD was different. These rates were in –1 the range 1.6–2.6 (mg SN–Corrected g COD-X–1 OHO h ) for –1 –1 SA and 0.6–0.9 (mg SN–Corrected g COD-XOHO h ) for SF. The mean denitrification rates for the studied wastewaters were lower than the obtained with domestic
References Alleman, J.E., Irvine, R.L., 1980. Storage induced denitrification using Sequencing Batch Reactor operation. Water Res. 14, 1483–1488. A´lvarez-Mateos, P., Pereda-Marı´ n, J., Carta-Escobar, F., Dura´nBarrantes, M.M., Guille´n-Jime´nez, E., 2000. Influence of inoculum and initial pH on dairy effluent biodegradation and mineralization. Chem. Biochem. Eng. Q. 14 (3), 101–106. A.P.H.A.-A.W.W.A.-W.P.C.F., 1998. Standard Methods for the Examination of Water and Wastewater, 20a% edicio´n. American Public Health Association/American Water Works Association/Water Environmental Federation, Washington, DC. Bilanovic, D., Battistoni, P., Cecchi, F., Pavan, P., MataAlvarez, 1999. Denitrification under high nitrate concentration and alternating anoxic conditions. Water Res. 33 (15), 3311–3320. Bolzonella, D., Innocenti, L., Pavan, P., Cecchi, F., 2001. Denitrification potential enhancement by addition of anaerobic fermentation products from the organic fraction of municipal solid waste. Water Sci. Technol. 44 (1), 187–194. Can˜izares, P., De Lucas, A., Rodriguez, L., Villasen˜or, J., 2000. Anaerobic uptake of different organic substrates by enhanced biological phosphorus removal sludge. Environ. Technol. 21, 397–405. Chiu, Y.C., Chung, M.S., 2003. Determination of optimal COD/nitrate ratio for biological denitrification. Int. Biodeterior. Biodegrad. 51, 43–49. Christensson, M., Lie, E., Welander, T., 1994. A comparison between ethanol and methanol as carbon sources for denitrification. Water Sci. Technol. 30 (6), 83–90. Contreras, E., Bertola, N., Zaritzky, N., 2001. The application of different techniques to determine activated sludge kinetic parameters in a food industry wastewater. Water SA 27 (2), 169–176.
ARTICLE IN PRESS 3726
A. De Lucas et al. / Water Research 39 (2005) 3715–3726
Cortacans, J.A., 2000. Activated Sludge. Biological Nutrients Removal. (Fangos activos. Eliminacio´n biolo´gica de nutrientes). Civil Engineering publication service, Madrid, Spain. De Lucas, A., Can˜izares, P., Rodriguez, L., Villasen˜or, J., 2000. Determination of biodegradation kinetics of bacterial storage organic substrates through electrolytic respirometry. Environ. Technol. 21, 1111–1118. Ferna´ndez, F.J., 2004. Effect of industrial discharges on Biological nutrient removal from domestic wastewaters (Efecto de vertidos industriales sobre la eliminacio´n biolo´gica de nutrientes de aguas residuales urbanas). Ph.D. Thesis, Faculty of Sciences, University of CastillaLa Mancha, Ciudad Real, Spain. Gerber, A., Mostert, E.S., Winter, C.T., de Villiers, R.H., 1986. The effect of acetate and other short chain carbon compounds on the kinetics of biological nutrient removal. Water SA 12 (1), 7–11. Harder, A., Roels, J.A., 1982. Application of simple structured models in bioengineering. In: Fiechter, A. (Ed.), Advances in Biochemical Engineering and Biotechnology 21. Springer, Berlin, p. 55. Hellinga, C., Schellen, A.A.J.C., Mulder, J.W., van Loosdrecht, M.C.M., Heijnen, J.J., 1998. SHARON process: an innovative method for nitrogen removal from ammoniumrich waste water. Water Sci.Technol. 37 (9), 135–142. Henze, M., 1986. Nitrate versus oxygen utilization rates in wastewater and activated sludge systems. Water Res. 18, 115–122. Henze, M., Grady, C.P.L., Gujer, W., Marais, G.v.R., Matsuo, T., 1987. Activated sludge Model No. 1, IAWPRC Scientific and Technical Report No. 1. International Association on Water-Pollution Research and Control, London. Henze, M., Harremoe¨s, P., Arvin, E., la Cour Jansen, J., 1995a. Wastewater Treatment: Biological and Chemical Processes. Springer, Heidelberg, Berlin. Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M.C., Marais, G.v.R., 1995b. Activated Sludge Model No. 2. IAWQ Scientific and Technical Report No. 3. International Association on Water Quality, London. Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M., Marais, G., van Loosdrecht, M., 1999. Activated Sludge Model No. 2d. Water Sci. Technol. 39 (1), 165–182. Herbert, H.P.F., Yu, H.Q., 2001. Acidification of lactose in wastewater. J. Environ. Eng. 127 (9), 825–831. Jetten, M.S., Wagner, M., Fuerst, J., van Loosdrecht, M., Kuenen, G., Strous, M., 2001. Microbiology and application of the anaerobic ammonium oxidation (‘anammox’) process. Curr. Opin. Biotechnol. 12 (3), 283–288. Kuba, T., van Loosdrecht, M.C.M., Heijnen, J.J., 1996. Phosphorus and nitrogen removal with minimal COD requirement by integration of denitrifying dephosphatation and nitrification in a two sludge system. Water Res. 30 (7), 1702–1710. Kujawa, K., Klapwijk, B., 1999. A method to estimate denitrification potential for predenitrification systems using NUR batch test. Water Res. 33 (10), 2291–2300. Masse, D.I., Masse, L., 2000. Characterisation of wastewater from hog slaughterhouses in Eastern Canada and evaluation of their in-plant wastewater treatment systems. Can. Agric. Eng. 42 (3), 139–146.
Mattsson, A., 1997. Denitrification in a non-nitrifying activated sludge system employing recirculation from a tertiary nitrification unit. Ph.D. Thesis, Civil Engineering Department, Chalmers University of Technology. Musvoto, E.V., Lakay, M.T., Casey, T.G., Wentzel, M.C., Ekama, G.A., 1999. Filamentous organism bulking in nutrient removal activated sludge systems. Paper 8: the effect of nitrate and nitrite. Water SA 25 (4), 397–407. Nemerow, N., Dasgupta, A., 1991. Industrial and Hazardous Waste Treatment. Ed. Van Nostrand Reinhold, New York. Niementowski, C.C., Nelson, P.E., 1977. Effects of temperature, pH and reaction time on the biodegradation of tomato waste in tank-type reactors. J. Food Sci. 41, 339–445. Rodrigo, M.A., Seco, A., Ferrer, J., Penya-Roja, 1999. The effect of sludge age on the deterioration of enhanced biological phosphorus removal process. Environ. Technol. 20, 1055–1065. Rodrı´ guez, L., Villasen˜or, J., Ferna´ndez, F.J., 2004. Long term effect of industrial wastewater discharges on a biological nutrient removal activated sludge. In: International Conference on Wastewater Treatment for Nutrient Removal and Reuse, vol. 1, pp. 172–179. Roeleveld, P.J., van Loosdrecht, M.C.M., 2002. Experience with guidelines for wastewater characterisation in The Netherlands. Water Sci. Technol. 45 (6), 77–87. Sliekers, A.O., Derwort, N., Gomez, J.L., Strous, M., Kuenen, J.G., Jetten, M.S., 2002. Completely autotrophic nitrogen removal over nitrite in one single reactor. Water Res. 36 (10), 2475–2482. Smolders, G.J.F., van der Meij, J., van Loosdrecht, M.C.M., Heijnen, J.J., 1995. A structured metabolic model for anaerobic and aerobic stoichiometry and kinetics of the biological phosphorus removal process. Biotechnol. Bioeng. 47, 277–287. Somiya, I., Tsuno, H., Matsumoto, M., 1988. Phosphorus release storage reaction and organic substrates behaviour in biological phosphorus removal. Water Res. 22 (1), 49–58. So¨zen, S., Orhon, D., 1999. The effect of nitrite correction on the evaluation of the rate of nitrate utilization under anoxic conditions. J. Chem. Technol. Biotechnol. 74, 790–800. Spanjers, H., Vanrolleghem, P.A., 1995. Respirometry as a tool for rapid characterization of wastewater and activated sludge. Water Sci. Technol. 31 (1), 105–114. Tam, N.F.Y., Wong, Y.S., Leung, G., 1992. Effect of exogenous carbon sources on removal of inorganic nutrient by nitrification-denitrification proceses. Water Res. 26 (9), 1229–1236. van Loosdrecht, M.C.M., Jetten, M.S.M., 1998. Microbiological conversions in nitrogen removal. Water Sci. Technol. 38 (1), 1–7. Vitousek, P.M., Mooney, H.A., Lubchenco, J., Melillo, J.M., 1997. Human domination of Earth’s ecosystem. Science 277, 494–499. Wachtmeister, A., Kuba, T., van Loosdrecht, M.C.M., Heijnen, J.J., 1997. A sludge characterisation assay for aerobic and denitrifiying phosphorus removing sludge. Water Res. 31 (3), 471–478. Zeng, R.J., Yuan, Z., Keller, J., 2003. Enrichment of denitrifying glycogen-accumulating organisms in anaerobic/anoxic activated sludge system. Biotechnol. Bioeng. 81 (4), 397–404.