The occurrence and removal of dissolved organic nitrogen (DON) is an issue of increasing importance for the reclamation of treated wastewater. Effluent DON may act as a precursor of disinfection by-products during wastewater disinfection and may contribute to eutrophication of receiving surface waters. The aim of this study was to understand the effect of the post-denitrification process on final effluent DON (organic nitrogen filtered by 0.45 μm pore size) concentration to further gain knowledge on how to optimize denitrifying filtration, in order to reach the required discharge standards. To evaluate DON variation, denitrification batch experiments were carried out with suspended and attached biomass under different shear conditions. For both conditions, with suspended and attached biomass, DON concentration did not increase or decrease during the denitrification process with addition of an external carbon source. Moreover, the increase of shear rate did not affect the DON concentration. Apparently, there is no direct link between DON evolution and the denitrification process itself.

INTRODUCTION

Understanding the occurrence and removal of dissolved organic nitrogen (DON) is of increasing importance for the reclamation of treated wastewater and/or treatment up to the most stringent discharge criteria for open surface water discharge. Although the exact nature of DON has not been fully revealed yet, the effluent DON may act as a disinfection by-product precursor during wastewater disinfection (Pehlivanoglu-Mantas & Sedlak 2006) and may also contribute to eutrophication of receiving surface waters (Seitzinger & Sanders 1997; Pehlivanoglu-Mantas & Sedlak 2006). Constituents of DON have been identified as proteins, amino acids and nucleic acids (Nam & Amy 2008), humic substances (Pehlivanoglu-Mantas & Sedlak 2006) and low molecular weight compounds (Pehlivanoglu-Mantas & Sedlak 2008). Inorganic nitrogen is removed in biological nutrient removal (BNR) treatment plants through biomass synthesis and sludge wasting as well as via the nitrification–denitrification pathway, whereas particulate organic nitrogen and part of colloidal organic nitrogen (CON) are removed through solid–liquid separation (Mekinia et al. 2009). This leaves DON as the predominant (up to 95%) form of organic nitrogen (ON) in the final effluent (Pagilla et al. 2008). According to Chen et al. (2011), DON control will be an important focus for improving performance of wastewater treatment plants (WWTPs). Current tertiary treatment processes are able to perform full denitrification (higher than 90%), but the presence of ON hampers final N concentrations reaching required limits, as low as 2.2 mg/L of total nitrogen (TN) according to European targets.

In this paper, DON is defined as ON filtered by 0.45 μm (DON0.45) but the definition of DON or CON may vary among different authors. Thus, DON or CON will be further expressed, when it is relevant, as DONX or CONX, where X indicates the pore size (or range of pore sizes) of the filter used to fractionate ON. Liu et al. (2012a) have shown that DON0.45 removal and production occurred during biofiltration of drinking water, and both phenomena were observed during wastewater biofiltration in a Dutch reclamation system (results not published). Likewise, increases of CON1.2 were observed during tertiary treatment filtration by Sattayatewa et al. (2010). Increases of DON in drinking water biofilters have been suggested to be related to different phases between backwashes (Liu et al. 2012a), but until now the effect of tertiary denitrifying filters on final effluent DON concentration has not been well known. To understand the effect of denitrifying filters on final effluent DON, it is necessary to firstly know the effect of denitrification on DON occurrence and disappearance. Mekinia et al. (2009) observed production of DON0.1 resulting from CON0.1–0.45 and CON0.45–1.2 hydrolysis during anoxic conditions in batch and at full scale, but results from another plant did not confirm such ON conversion. Czerwionka et al. (2012) also observed a decrease in CON0.45. The observation was based on ON measurements carried out before, in the middle and after a period of anoxic conditions in batch reactors but the measurement of DON0.45 along with the denitrification process was never analysed continuously. This way, the DON dynamics along the process of denitrification is still not clear. The decrease or conversion of CON0.45 observed by Mekinia et al. (2009) or Czerwionka et al. (2012) and its increase in a tertiary treatment biofilter, as shown by Sattayatewa et al. (2010) or observed in the aforementioned Dutch treatment plant, indicate that processes other than denitrification might be responsible for DON production in a denitrifying filter. In fact, the stable structure of the biofilm is a result of the interactive strength between aggregates and hydrodynamic shear force (Liu & Tay 2002). It has been generally observed that a high shear force leads to high production of extrapolymeric substances (EPS) (Trinet et al. 1991; Ohashi & Harada 1994; Pratt & Kolter 1999), although the exact mechanism by which hydrodynamic shear forces stimulate the production of exopolymers is not yet clear. Such production of EPS may be responsible for an increase in ON. In this research, experiments were carried out using batch-fed reactors inoculated with suspended biomass and batch-fed reactors inoculated with an inert carrier medium containing a denitrifying biofilm, from a Dutch municipal WWTP, in order to evaluate variations in DON0.45 concentrations during the denitrification process. Moreover, the batch tests with attached biomass were performed at different stirring velocities (shear rates) in order to test the effect of shear stress on DON behaviour. Whereas other studies have shown DON concentration profiles during activated sludge treatment under various oxygen conditions (Sattayatewa et al. 2009; Mekinia et al. 2009; Czerwionka et al. 2012) or during aeration (Parkin & McCarty 1981b), this study presents a DON concentration profile during denitrification. To our knowledge, this paper presents the first results of DON measurement during denitrification in batch with attached biomass and at different shear rates.

METHODS

Denitrification batch tests using suspended biomass

An 18 L batch reactor was filled with activated sludge from a Dutch municipal WWTP. The reactor was kept anoxic by sparging with N2 gas when necessary. Dissolved oxygen (DO) was checked through a flow-through cell connected to the reactor and to a DO meter (WTW FDO 925). The liquid was mixed with a magnetic overhead stirrer (IKA RW16 basic) at a constant speed to keep the biomass in suspension. Sodium nitrate was added to the activated sludge in order to obtain a final concentration of 10 mg NO3-N/L, and 65 mg/L of sodium acetate trihydrate was added in order to provide enough chemical oxygen demand for full denitrification. Since the aim of this study is to compare DON evolution during denitrification by suspended and attached biomass, the concentration of 10 mg/L of NO3-N was chosen as it is representative of an initial concentration of NO3 in tertiary filters. Samples were collected every 4 minutes until denitrification was completed. Each sample was analysed for inorganic N and, to examine DON variation with denitrification, at least 10 samples were analysed for DON. The same samples were measured for pH and temperature. Volatile suspended solids (VSS) were measured at the beginning and the end of the experiment. The denitrification test was repeated three times.

Sample treatment

Part of each sample was immediately cooled in dry ice for 3.5 minutes and transferred to the fridge immediately until centrifugation and further analyses. The other part was stored in the fridge for subsequent measurement of VSS. The supernatant of the centrifuged samples was frozen for further measurement. After melting the samples they were filtrated using a 1.2 μm pore size membrane filter (Whatman) followed by a 0.45 μm filter (Sartorius). Part of the filtrate was immediately analysed for NO3-N, NO2-N and NH4-N and the remainder was acidified and frozen for further DON measurement according to Liu et al. (2012b), i.e., based on the measurement of inorganic N and total dissolved nitrogen (TDN) after NO3 removal by ion exchange.

Analytical measurements

For measurement of NO3-N, NO2-N and NH4-N immediately after filtration, spectrophotometric MERCK tests were used and VSS were determined following Standard Methods (APHA, AWWA, WEF 1998). Before measurement of TDN, NO3 was removed by means of ion exchange, following the same procedure as Liu et al. (2012b). TDN was measured in triplicate by persulphate digestion following Standard Methods (APHA, AWWA, WEF 1998). After NO3 removal, NO2-N and NH4-N were measured by Hach tests and a possible residual of NO3-N was measured with ion chromatography. DON was calculated in triplicate by the difference between each of the three measured TDN concentrations and the inorganic N-forms NO2-N, NH4-N and NO3-N, measured after NO3 removal.

Denitrification batch tests using attached biomass

The previously described 18 L batch reactor set-up, equipped with the same DO probe and mixer was used, now filled with 8 L of Kaldnes k1 carriers and 16 L of Harnaschpolder WWTP effluent. The methodological approach of the batch assay with the bio-carrier was identical to that of the free suspended biomass. The test was carried out at two different stirring speeds, 43 rpm and 108 rpm. The velocities were chosen according to the stirrer limitations and in order to have a difference of 2.5 times between each other. The procedure including both stirring speeds was repeated three times. VSS were measured three times, i.e., before starting each set of two tests using different stirring speeds.

Bio-carrier medium preparation

Ten litres of Kaldnes carriers were submerged in activated sludge from a Dutch urban WWTP and kept in suspension for a day. The Kaldnes was removed from the activated sludge and fed every 2 days with synthetic wastewater containing CaCl2·2H2O, FeCl3·6H2O, MgSO4·7H2O, CoCl2·6H2O, NaSiO3·5H2O, Al2(SO4)3·18H2O, MnCl2·4H2O, C2H9NaO5, NaNO3 and K2HPO4.

Sample treatment

The same sample treatment was used as for the denitrification tests with suspended biomass.

Analytical measurements

The same analytical measurements were used as for the denitrification tests with suspended biomass. For the determination of VSS, 200 Kaldnes pieces were collected from the reactor and submitted to 6 hours of sonication for detachment of biofilm from the carriers.

Figure 1 shows a scheme of the performed denitrification batch tests.
Figure 1

Performed denitrification tests.

Figure 1

Performed denitrification tests.

RESULTS AND DISCUSSION

The first four columns of Table 1 show the characteristics of the three denitrification tests carried out with suspended biomass (tests A, B and C). NOX represents the sum of NO3-N and NO2-N.

Table 1

Results of denitrification tests with suspended biomass (tests A, B and C) and attached biomass (tests D, E, F, G, H and I)

Tests with suspended biomass
 
Tests with attached biomass and stirring speed of 43 rpm
 
Tests with attached biomass and stirring speed of 108 rpm
 
  Denitrification rate VSS   Denitrification rate VSS   Denitrification rate VSS 
Test [gNOx-N/(kgVSS.d)] [g/L] [0C] Test [gNOx-N/(kgVSS.d)] [g/L] [0C] Test [gNOx-N/(kgVSS.d)] [g/L] [0C] 
71 2.58 20 ± 0.6 159 1.84 21 ± 0.5 255 1.84 21 ± 0.5 
22 2.79 21 ± 0.5 131 1.34 21 ± 0.5 180 1.34 21 ± 0.5 
28 2.28 19 ± 0.5 165 1.50 21 ± 0.5 112 1.50 21 ± 0.5 
Tests with suspended biomass
 
Tests with attached biomass and stirring speed of 43 rpm
 
Tests with attached biomass and stirring speed of 108 rpm
 
  Denitrification rate VSS   Denitrification rate VSS   Denitrification rate VSS 
Test [gNOx-N/(kgVSS.d)] [g/L] [0C] Test [gNOx-N/(kgVSS.d)] [g/L] [0C] Test [gNOx-N/(kgVSS.d)] [g/L] [0C] 
71 2.58 20 ± 0.6 159 1.84 21 ± 0.5 255 1.84 21 ± 0.5 
22 2.79 21 ± 0.5 131 1.34 21 ± 0.5 180 1.34 21 ± 0.5 
28 2.28 19 ± 0.5 165 1.50 21 ± 0.5 112 1.50 21 ± 0.5 

For test A, the calculated denitrification rate was 71 g NO3-N/(kg MLVSS.d), which is in the range of 40 to 420 g NO3-N/(kg MLVSS.d) observed for pre-anoxic tanks in full-scale installations (Metcalf & Eddy Inc. 2003) or close to the range of 72 to 720 g NO3-N/(kg MLVSS.d), observed for anoxic batch tests (Ekama et al. 1986). For tests B and C, the denitrification rates were 22 and 28 g NO3-N/(kg MLVSS.d) which is between 11 and 42 g N/(kg MLVSS.d), observed by Gerber et al. (1986) in denitrification batch tests. The difference between the denitrification rate of the test A and the consecutive tests B and C can probably be attributed to the nature of the activated sludge, since the concentration of the solids was similar for all tests. It should be noted that the behaviour of inorganic N species was similar in all three batch tests; i.e., while NO3 decreased from 10 to 0 mg/L, NO2 and NH4 remained close to 0 mg/L (results not shown).

The DON concentration during the denitrification batch tests A, B and C, is shown in Figure 2. The vertical bars represent the standard deviation obtained from the three calculated DON values per sample. In the present set-up, the fast cooling of samples with dry ice allowed filtration to be delayed by some hours, which facilitated multiple sampling during the denitrification of 10 mg NO3-N/L. In fact, the dry-ice cooling procedure was found to be indispensable, since cooling down samples in a 4 °C refrigerator was not enough to stop the denitrification occurring (at such small nitrate concentration). Moreover, the time between two sample collections was not enough to filter a sample. Thus, dry-ice cooling resulted in an effective conservation method for denitrification batch samples.
Figure 2

DON concentration during denitrification batch tests with suspended biomass: tests A, B and C.

Figure 2

DON concentration during denitrification batch tests with suspended biomass: tests A, B and C.

The concentration of DON during denitrification varied from 1.07 mg/L to 1.61 mg/L; 0.60 mg/L to 1.48 mg/L and 0.04 mg/L to 0.85 mg/L for tests A, B and C, respectively. This resulted in averages of 1.30 ± 0.80 mg/L, 1.20 ± 0.25 mg/L and 0.33 ± 0.27 mg/L. The averages of the first two tests (A and B) are similar to the 1.5 mg/L of average filtered effluent ON concentration (CON + DON), presented by Parkin & McCarty (1981a). They are also comparable with the sum of DON0.1 and CON0.1–1.2 concentrations between 1.9 and 2.4, observed by Czerwionka et al. (2012) in two full-scale biological nutrient removal plants.

The variation in DON0.45 found during denitrification (Figure 2) did not present a clear increasing or decreasing trend, suggesting that DON was not, or only marginally, affected by the batch denitrification process, although a somewhat decreasing tendency might be observed towards the end of the test. The more or less steady DON concentration is in agreement with Pagilla et al. (2008), whose studies in BNR WWTPs in the USA and Poland suggest that effluent CON and DON concentrations are independent of the influent TN concentrations. Conversely, Czerwionka et al. (2012) observed that the largest reductions of the sum of CON0.1–0.45 and CON0.45–1.2 occurred in the anaerobic and anoxic compartments of two full-scale BNR bioreactors whereas DON0.1 slightly increased. Measuring ON concentration profiles in a full-scale four-stage Bardenpho bioreactor, Sattayatewa et al. (2009) observed ON release in the anoxic compartment. Contrary to full-scale results, the results observed by Czerwionka et al. (2012) during batch tests showed ambiguous ON behaviour. If for one plant production of DON0.1 resulted from an evident CON0.1–0.45 and CON0.45–0.1.2 hydrolysis, in another plant the CON slightly decreased or even increased during anoxic conditions. The reported results were based on three measurements: before, in the middle and at the end of an anoxic phase of 4 hours. Before this study, only Parkin & McCarty (1981b, c) had analysed the DON0.45 behaviour in activated sludge, measuring ON during aeration using the Kjeldahl method. The authors attributed the excretion of DON during aeration to concentration gradients, starvation conditions, addition of exogenous substrate, and changes in phase and rate of growth. Subsequent utilization of DON0.45 by the microorganisms would lead to a minimum DON0.45 concentration, after which a gradual increase would occur due to the starvation. Likewise, Czerwionka et al. (2012) observed a minimum DON0.1 concentration at the end of the anoxic phase during bench-scale experiments. During the current experiment, starvation conditions were in principle not achieved as enough carbon and nitrate were added in the beginning of the batch test. The observed tendency for a constant DON is possibly a result of simultaneous DON production due to addition of exogenous substrate, and degradation due to hydrolysis of DON>0.45, which is accompanied by DON utilization by the microorganisms. Profiles similar to the ones observed in Figure 2 for different fractions of ON would help to understand whether there is simultaneous consumption and production of DON, or whether the DON concentration is stagnant. Similar tests in other plants could give valuable information for understanding variations in effluent DON concentration that are dependent on the BNR plant characteristics as stated by Pagilla et al. (2008).

The columns 5 to 8 of Table 1 show the characteristics of the denitrification tests carried out with attached biomass, at stirring speed of 43 rpm (tests D, E and F).

The obtained denitrification rates of 159, 131 and 165 g NOx/(kg VSS.d)) are comparable with the results of Zafarzadeh et al. (2010), who quantified an average and maximum rate of 40.1 and 157 g NOx/(kg VSS.d), respectively. The DON variation obtained for the tests with stirring speed of 43 rpm – tests D, E and F – is shown in Figure 3.

From Figure 3, it can be observed that the DON concentration is almost constant between 0.5 and 1.0 mg/L for the tests D and E. In contrast, test F showed that after an initial DON increase from approximately 1 to 2 mg/L, a big decrease to 0.5 mg/L occurred followed by a new increase up to above 1 mg/L. Such change on the DON curve was not related to any of the characteristics presented by columns 5 to 8 of Table 1.
Figure 3

DON concentration during denitrification batch tests with attached biomass, at stirring speed of 43 rpm – tests D, E and F; and stirring speed of 108 rpm – tests G, H and I.

Figure 3

DON concentration during denitrification batch tests with attached biomass, at stirring speed of 43 rpm – tests D, E and F; and stirring speed of 108 rpm – tests G, H and I.

Parkin & McCarty (1981a) identified the soluble organic nitrogen (SON) produced (SONp) during activated sludge treatment as a combination of SON generated due to growth (SONg), SON generated due to decay (SONd) and SON generated to attain an equilibrium concentration between the organisms and the medium (SONeq). Studying batch aeration, the author suggested that SON excreted due to non-limited substrate conditions can't be experimentally distinguishable from a DONd (decay) peak due to a rapid utilization of the excreted SON by the organisms. Conversely, the same study showed that the contribution of SONd to SONp varied with culture characteristics. In most cases, SONg production is small in comparison with SONeq and SONd (Parkin & McCarty 1981a). For the tests with attached biomass, an increased shear force was applied in order to induce biofilm disturbance, possibly leading to increased microorganism decay and subsequent DON release. Nevertheless, similar to the results of Figure 2, and the results of the tests D and E of Figure 3, the graphs of the tests G, H and I of Figure 3 show that DON concentrations do not tend to decrease or increase during the denitrification process with attached biomass.

The columns 9 to 12 of Table 1 list the characteristics of the denitrification rates carried out with attached biomass, at stirring speed of 108 rpm (tests G, H and I). Contrary to the tests G and I, where the DON concentration is quite constant within the interval of 0.5 to 1.25 mg/L, test H showed a decrease from around 1 to 0 mg/L followed by an increase to close to the initial concentration. The difference between the DON curve of the test H and the curves of the tests G and I cannot be related to any of the parameters presented by columns 9 to 12 of Table 1. The comparison between the graphs of tests D, E, F and the tests G, H, I suggests that the increased stirring speed, and thus the increased shear rate, did not affect the DON concentration during denitrification. Our current analysis on DON concentrations during denitrification with addition of external carbon source for both suspended and attached growth shows a more or less constant value throughout the test. However, as the literature reports an increase and decrease of ON during denitrification under dissimilar experimental conditions, and also provides evidence that different fractions of ON exist, further study should address and compare these fractions in more detail under various growth and decay conditions, applying a wider range of biomass concentrations than those used in our assays. It should be noted that the specific biomass N-loading in denitrifying filters varies considerably within the operational cycle between two back washes. The latter may impact DON dynamics depending on filtration cycle time.

CONCLUSIONS

  • The fast dry-ice cooling of samples containing biomass with a concentration between 2 and 3 g VSS/L allows the measurement of a profile of DON concentration during the denitrification of 10 mg NO3-N/L.

  • For both conditions, with suspended and attached biomass, DON concentration during batch feeding denitrification does not tend to increase or decrease.

  • The increase in shear rate introduced by the 2.5 times increase in the stirring speed of the biological carriers did not influence DON concentration during denitrification with attached biomass.

REFERENCES

REFERENCES
APHA, AWWA, WEF
1998
Standard Methods for the Examination of Water and Wastewater
.
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington, DC
.
Chen
B.
Kim
Y.
Westerhoff
P.
2011
Occurrence and treatment of wastewater-derived organic nitrogen
.
Water Research
45
(
15
),
4641
4650
.
Ekama
G. A.
Dold
P. L.
Marais
G. V. R.
1986
Procedures for determining influent COD fractions and the maximum specific growth rate of heterotrophs in activated sludge systems
.
Water Science and Technology
18
(
6
),
91
114
.
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
.
Liu
B.
Gu
L.
Yu
X.
Yu
G.
Zhang
H.
Xu
J.
2012a
Dissolved organic nitrogen (DON) profile during backwashing cycle of drinking water biofiltration
.
Science of the Total Environment
414
,
508
514
.
Metcalf & Eddy Inc.
2003
Wastewater Engineering
.
McGraw-Hill
,
New York
.
Ohashi
A.
Harada
H.
1994
Characterization of detachment mode of biofilm developed in an attached-growth reactor
.
Water Science and Technology
30
(
11
),
35
45
.
Pagilla
K. R.
Czerwionka
K.
Urgun-Demirtas
M.
Makinia
J.
2008
Nitrogen speciation in wastewater treatment plant influents and effluents—the US and polish case studies
.
Water Science and Technology
57
(
10
),
1511
1517
.
Parkin
G. F.
McCarty
P. L.
1981a
Sources of soluble organic nitrogen in activated sludge effluents
.
Journal (Water Pollution Control Federation)
53
(
1
),
89
98
.
Parkin
G. F.
McCarty
P. L.
1981b
Production of soluble organic nitrogen during activated sludge treatment
.
Journal (Water Pollution Control Federation)
53
(
1
),
99
112
.
Pehlivanoglu-Mantas
E.
Sedlak
D. L.
2006
Wastewater-derived dissolved organic nitrogen: Analytical methods, characterization, and effects – a review
.
Critical Reviews in Environmental Science and Technology
36
(
3
),
261
285
.
Pratt
L. A.
Kolter
R.
1999
Genetic analyses of bacterial biofilm formation
.
Current Opinion in Microbiology
2
(
6
),
598
603
.
Sattayatewa
C.
Pagilla
K.
Sharp
R.
Pitt
P.
2010
Fate of organic nitrogen in four biological nutrient removal wastewater treatment plants
.
Water Environment Research
82
(
12
),
2306
2315
.
Trinet
F.
Heim
R.
Amar
D.
Chang
H. T.
Rittmann
B. E.
1991
Study of biofilm and fluidization of bioparticles in a three-phase liquid-fluidized-bed reactor
.
Water Science and Technology
23
(
7–9
),
1347
1354
.
Zafarzadeh
A.
Bina
B.
Nikaeen
M.
Movahedian Attar
H.
Hajian Nejad
M.
2010
Performance of moving bed biofilm reactors for biological nitrogen compounds removal from wastewater by partial nitrification-denitrification process
.
Iranian Journal of Environmental Health Science and Engineering
7
(
4
),
353
364
.