By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Many countries strive to reduce the emissions of nitrogen compounds (ammonia, nitrate, NOx) to the surface waters and the atmosphere. Since mainstream domestic wastewater treatment systems are usually already overloaded with ammonia, a dedicated nitrogen removal from concentrated secondary or industrial wastewaters is often more cost-effective than the disposal of such wastes to domestic wastewater treatment. The cost-effectiveness of separate treatment has increased dramatically in the past few years, since several processes for the biological removal of ammonia from concentrated waste streams have become available. Here, we review those processes that make use of new concepts in microbiology: partial nitrification, nitrifier denitrification and anaerobic ammonia oxidation (the anammox process). These processes target the removal of ammonia from gases, and ammonium-bicarbonate from concentrated wastewaters (i.e. sludge liquor and landfill leachate). The review addresses the microbiology, its consequences for their application, the current status regarding application, and the future developments.
The three ammonium removal processes that are the focus of this review make use of (combinations of) three groups of chemolithoautotrophic bacteria: the well known ‘aerobic’ ammonia and nitrite oxidizers (Eqs. 1–3) and anaerobic ammonia oxidizers (Eq. 4). They all derive energy for microbial growth (CO2 fixation) from the oxidation of inorganic nitrogen compounds:
For each of these three processes, microbiological aspects important for wastewater treatment are reviewed in the next sections.
2Phylogeny and ecological niche
The evolutionary relationships (based on 16S rDNA phylogeny) of the chemolithoautotrophs of the nitrogen cycle are relevant to wastewater treatment, because 16S rDNA-based probing of populations of these organisms has been remarkably successful. Such 16S rDNA probes have been used to quantify the amounts of nitrifiers or anaerobic ammonia oxidizers in wastewater treatment plants [1–4]. In several cases probing resolved the mechanism of not-understood nitrogen conversions [5–7]. The probes were used to measure the in situ growth rates of relevant organisms in the actual plant [8,9]. The characterization of populations of nitrifiers using probes does not yet enable solid conclusions to improve plant management, but in view of the rapidly accumulating genomic data (http://www.arb.de), phylogenetic (16S) gene-probing may become part of the standard tools for daily plant management and optimization in the coming years.
2.1Proteobacterial ammonia oxidizers
These ammonia oxidizing bacteria form two monophyletic groups, one within the beta- and one within the gamma-proteobacteria . They are generally considered as aerobic chemolithoautotrophs, but recently organic compounds have been described that can serve them as carbon and energy source (see below). The beta-ammonia oxidizers comprise the well known genera Nitrosomonas and Nitrosospira, Nitrosococcus is the gamma-proteobacterial genus , but does not include Nitrosococcus mobilis, that is related to Nitrosomonas. Different members of these genera have been found to dominate different wastewater treatment plants or natural ecosystems [2,4,8,12–17], but general relationships between the ecological niche and evolutionary position are often still obscure. The SHARON process (single reactor system for high-rate ammonium removal over nitrite; discussed below) is carried out largely by Nitrosomonas eutropha. The same bacterium is also one of the most capable denitrifiers (among nitrifiers) and was found to dominate the nitrifier denitrification (NOx process, see below). Salty wastewaters were found to be dominated by N. mobilis. The genome project of Nitrosomonas europaea nears completion. Although the relevance of this organism for wastewater treatment is disputable, it will still provide an invaluable source of information.
2.2Aerobic nitrite oxidizers
The second step of nitrification, the oxidation of nitrite to nitrate, is performed by nitrite oxidizing bacteria, e.g. members of the genera Nitrobacter, Nitrococcus and Nitrospira. The first two genera are part of the alpha-proteobacteria, while Nitrospira is phylogenetically unrelated to any other cultivated species and forms a separate division .
Several strains of Nitrobacter and one strain of Nitrospira are the only nitrite oxidizers that are not restricted to marine environments [21,22]. There is some evidence that Nitrospira is the more specialized nitrite oxidizer. The other genera are more versatile, being facultative autotrophs and anaerobes, able to grow on heterotrophic substrates such as pyruvate and also capable of the first step of denitrification (the reduction of nitrate to nitrite) . It appears that the genomes of nitrite oxidizers will not become available in the near future.
2.3Anaerobic ammonia oxidizers
Anaerobic ammonia oxidation (anammox) is mediated by a group of planctomycete bacteria , two of which have been named provisionally (‘Candidatus Brocadia anammoxidans’ and ‘Candidatus Kuenenia stuttgartiensis’). Retrieval of 16S rDNA sequences from anammox wastewater treatment has revealed several relatives of both species [7,25] and at least one other distinct group (Schmid et al., unpublished results). The molecular biodiversity of anammox bacteria is much larger than the diversity of their proteobacterial counterparts . It is not yet known if or how the niche differentiation correlates with the phylogeny. This issue is important for application because the long start-up times of anammox reactors could be reduced significantly if it could be predicted how to seed a new reactor. The lack of pure cultures of anammox bacteria makes a genomic approach less straightforward in the near future.
3.1Proteobacterial ammonia oxidizers
The physiology of conventional, ‘aerobic’ ammonia oxidizers is not completely understood. Only recently, it was discovered that these organisms also have an anaerobic metabolism (see below).
The proteobacterial ammonia oxidizers can obtain their energy for growth from both aerobic or anaerobic ammonia oxidation. Most likely ammonia (NH3) and not ammonium (NH4+) is the substrate for the oxidation process . The main products are nitrite under oxic conditions and dinitrogen, nitrite and nitric oxide under anoxic conditions [27,28]. Aerobic (Eq. 1) and anaerobic ammonia oxidation (Eq. 2) is initiated by the enzyme ammonia monooxygenase (AMO), that oxidizes ammonia to hydroxylamine. Oxygen and dinitrogen tetroxide (dimer of NO2) are the most likely electron acceptors for this enzyme [27–33] (Eqs. 4 and 5).
The hydroxylamine resulting from ammonia oxidation is further oxidized to nitrite (Eq. 7) by the hydroxylamine oxidoreductase (HAO) [34–36].
The four reducing equivalents derived from this reaction enter the AMO reaction (Eqs. 5 and 6), the CO2 assimilation, and the respiratory chain . The reducing equivalents are transferred to the terminal electron acceptors O2 (oxic conditions) or nitrite (anoxic conditions) [33,37]. The reduction of nitrite under anoxic conditions leads to the formation of N2 resulting in a N-loss of 45±15%. Under anoxic conditions the ammonia oxidation activity is relatively low (2.5 nmol NH3 (g protein)−1 min−1). The doubling time is about 30 days at best and the biomass yield is 0.13±0.019 g dry weight (g NH3-N)−1. The Ks value for the substrate ammonia is about 20 μM at pH values between 6.7 and 8.3. These organisms are reversibly or irreversibly inhibited by various carbon compounds [38,39]. In contrast to aerobic ammonia oxidation , ammonia oxidation under anoxic conditions is not inhibited by acetylene .
In the presence of oxygen, the produced NO can be oxidized to NO2. Therefore, only small amounts of NO are detectable in the gas phase of N. eutropha cell suspensions . According to Eq. 2 N2O4 is the oxidizing agent also under oxic conditions . Hydroxylamine and NO are produced as intermediates. While hydroxylamine is further oxidized to nitrite (Eq. 7), NO is (re)oxidized to NO2 (N2O4) (Eq. 8).
Recently, a model was developed to explain the role of NOx in the metabolism of the ammonia oxidizers .
Under oxic conditions (>0.8 mg O2 l−1) aerobic nitrifiers convert ammonia to nitrite (see above). At an oxygen concentration below 0.8 mg O2 l−1 they use small amounts of the produced nitrite as terminal electron acceptors producing NO, N2O, and N2. In the absence of nitrogen oxides, up to 15% of the converted ammonia can be denitrified . N. eutropha was shown to nitrify and simultaneously denitrify under fully oxic conditions in the presence of NO2 or NO. Interestingly, there is no fixed stoichiometry measurable between ammonia and NO2 (NO) consumption under oxic conditions. The ratio of ammonia to NOx consumption range between 1000:1 and 5000:1. Obviously, nitrogen oxides have a regulatory function in the metabolism of nitrifiers under oxic conditions, stimulating the denitrification activity . Influenced by nitrogen oxides, ammonia oxidizers convert ammonia to gaseous dinitrogen (about 60% of the converted ammonia) and nitrite (just about 40% of the converted ammonia) . The specific aerobic ammonia oxidation activity is stimulated by NO2, with values increasing from 33 μmol NH3 (g protein)−1 min−1 without NOx addition to 280 μmol NH3 (g protein)−1 min−1 and a denitrification activity of 150 μmol NO2− (g protein)−1 min−1 in the presence of 50 ppm NO2. The biomass yield and the affinity for ammonia remain unchanged. Control experiments with N. europaea and Nitrosolobus multiformis have yielded similar results. The reaction mechanism is the same, but the activities vary. Nitrogen oxides are toxic for many other microorganisms (nitrite oxidizers, heterotrophic bacteria) . Reducing the cell number and the activity of the nitrite oxidizers by adding NOx can be desirable in wastewater treatment, because the nitrite formed by the ammonia oxidizers is not further oxidized to nitrate (i.e. nitrite oxidizers). This is important since the nitrite is needed for the denitrification by the ammonia oxidizers.
3.2Aerobic nitrite oxidizers
As mentioned above, nitrite oxidizers are often more versatile than ammonia oxidizers. When growing autotrophically with nitrite, the biomass yield is 0.036 g dry weight (g nitrite-N)−1, at a maximum growth rate of 0.04 h−1. The apparent activation energy of nitrite oxidation is 44 kJ mol−1. Like the ammonia oxidizers, these bacteria can have high substrate affinities (around <70 μM for nitrite and <25 μM for oxygen). It has been reported that hydroxylamine, ammonia and NO can inhibit nitrite oxidizers , but a mechanism for such inhibitions has not yet been proposed.
The key enzyme of nitrite oxidizing bacteria is the membrane-bound nitrite oxidoreductase  which oxidizes nitrite with water as the source of oxygen to form nitrate . The electrons released from this reaction are transferred via a- and c-type cytochromes to a cytochrome oxidase of the aa3-type. However, the mechanism of energy conservation in nitrite oxidizers is still unclear. Neither Hollocher et al.  nor Sone et al.  was able to find any electron transport chain-linked proton translocation, which is necessary to maintain a proton motive force for ATP regeneration. Thus, NADH is thought to be produced as the first step of energy conservation . Nitrite oxidizers are generally lithoautotrophic organisms . Higher growth rates are obtained when the cells are growing mixotrophically [55,56]. Several strains of Nitrobacter are capable of heterotrophic growth under oxic as well as anoxic conditions [57,58]. Heterotrophic growth is significantly slower than lithoautotrophic growth , although 10–50-fold higher cell densities are obtained . Some strains of Nitrobacter were shown to be denitrifying organisms as well. Under anoxic conditions, nitrate can be used as an acceptor for electrons derived from organic compounds to promote anoxic growth . Since the oxidation of nitrite is a reversible process, the nitrite oxidoreductase can reduce nitrate to nitrite in the absence of oxygen . Nitrite oxidation occurs obligatory under oxic conditions. The involved organisms are much more sensitive to oxygen limitation than ammonia oxidizers are. Already at dissolved oxygen concentrations of about 0.5 mg l−1 nitrite oxidation is completely inhibited . Additionally, Nitrobacter is inhibited at high oxygen concentrations . Thus, the oxygen content of a nitrite oxidizing nitrification vessel has to be maintained carefully to avoid accumulation of nitrite. With sufficient oxygen supply nitrite oxidation proceeds at a faster rate than conversion of ammonia to nitrite. Therefore, high nitrite concentrations are rarely found neither in natural environments nor in wastewater treatment plants .
The physiology of the anaerobic ammonia oxidizer ‘Candidatus Brocadia anammoxidans’ has been studied in detail. The bacterium is a chemolithoautotroph, has a doubling time of 11 days and the biomass yield is 0.13 g dry weight (g NH3-N)−1. It has a very high affinity for the substrates ammonia and nitrite . It is reversibly inhibited by oxygen and irreversibly by nitrite (at concentrations in excess of 70 mg N l−1 for several days) and phosphate (>60 mg P l−1 for several days) [64–66]. The apparent activation energy is approximately the same as for aerobic ammonia oxidation: 70 kJ (mol NH3)−1. ‘Candidatus Kuenenia stuttgartiensis’ has a higher, but still low, tolerance to nitrite (180 mg N l−1) and phosphate (600 mg P l−1) . Both bacteria have a similar temperature (37°C) and pH (8) optima.
Anammox bacteria do not consume ammonia and nitrite in a ratio of 1:1, as might be expected from their catabolism (Eq. 4), but in a ratio of 1:1.3. The excess nitrite (0.3 mol of nitrite per mol of ammonia) is oxidized anaerobically to nitrate. The electrons derived from this oxidation are probably used for the fixation of CO2.
The biochemistry of the anammox bacteria is not yet completely resolved. It is known that the anaerobic oxidation of ammonia proceeds via hydrazine (N2H4), a volatile and toxic intermediate [67,68]. An enzyme that resembles HAO from aerobic ammonia oxidizers is responsible for the oxidation of hydrazine to dinitrogen gas (ΔG°′=−288 kJ mol−1) . It has been postulated that the electrons from this oxidation are channelled to nitrite leading to the production of hydroxylamine (ΔG°′=−22.5 kJ mol−1). Hydroxylamine and ammonia could yield hydrazine in a condensation reaction (ΔG°′=−46 kJ mol−1), which completes the catalytic cycle.
The ultrastructure of B. anammoxidans has many features in common with previously described planctomycetes (Fig. 1). These microorganisms have a proteinaceous cell wall lacking peptidoglycan and are thus insensitive to ampicillin. Anammox catabolism is at least partly located in a membrane-bound intracytoplasmic compartment, known as the anammoxosome . All anammox cells have exactly one anammoxosome . Anammoxosomes can be isolated intact from anammox cells . They contain little or no RNA or DNA  and are surrounded by a dedicated membrane that is very impermeable because it consists of ladderane lipids . The bacterial nucleoid is located on the outside of the anammoxosome membrane; it is extremely condensed as is the case for the other planctomycetes. Fig. 1 shows the ultrastructure of the anammox bacterium Candidatus‘Brocadia anammoxidans’. Interestingly, B. anammoxidans as well as aerobic ammonia oxidizers like Nitrosomonas develop internal membrane systems.
The oxidation of ammonia , hydroxylamine  or organic nitrogen compounds, e.g. oximes , to nitrite and nitrate by various chemoorganotrophic microorganisms is called heterotrophic nitrification. The latter is a co-metabolism which is not coupled to energy conservation . Heterotrophic nitrifiers are found among algae , fungi  and bacteria . Compared to those of autotrophic nitrifiers nitrification rates of heterotrophic nitrifiers are low . Therefore, heterotrophic nitrification was thought to occur preferentially under conditions which are not favorable for autotrophic nitrification, e.g. acidic environments .
Denitrification is the reduction of oxidized nitrogen compounds like nitrite or nitrate to gaseous nitrogen compounds. This process is performed by various chemoorganotrophic, lithoautotrophic, and phototrophic bacteria and some fungi [80,81], especially under oxygen-reduced or anoxic conditions . Denitrification can be described as a kind of anoxic respiration. Electrons originated from e.g. organic matter, reduced sulfur compounds, or molecular hydrogen are transferred to oxidized nitrogen compounds instead of oxygen in order to build up a proton motive force usable for ATP regeneration. Enzymes involved are the nitrate reductase, the nitrite reductase, the nitric oxide reductase, and finally the nitrous oxide reductase [83,84]. Dinitrogen is the main end product of denitrification while the nitrogenous gases (nitric oxide and nitrous oxide) are occurring as intermediates in low concentrations . However, these gases are also released as end products when denitrification enzymes are not completely expressed, e.g. when the concentration of dissolved oxygen is too high . Denitrification also occurs in the presence of oxygen. The range of oxygen concentrations permitting aerobic denitrification is broad and differs from one organism to another . The onset of aerobic denitrification is not depending on oxygen sensitivity of the corresponding enzymes, but rather on regulation of oxygen- or redox-sensing factors involved in the regulation on a transcriptional level.
4Processes for N-removal
The newly discovered anaerobic ammonia oxidizing planctomycetes and the anaerobic metabolism of proteobacterial ammonia oxidizers open up new possibilities for nitrogen removal from wastewater. More specifically, the paradigm that the only way to biologically convert wastewater ammonium to dinitrogen gas necessitates the complete oxidation to nitrate followed by heterotrophic denitrification, has become obsolete. In this section the application of the new microbial possibilities is discussed.
The oxidation of nitrite to nitrate can be prevented in at least two ways. First, by making use of the difference in activation energy between ammonia and nitrite oxidation (68 kJ mol−1 and 44 kJ mol−1, respectively). The high activation energy of ammonia oxidation makes the rate of this process more dependent on temperature. The SHARON process (Fig. 2) makes use of the different growth rates of ammonia and nitrite oxidizers at sufficiently high temperatures (more than 26°C) [48,87]. It works at a hydraulic retention time higher than the growth rate of nitrite oxidizers but lower than ammonia oxidizers (about 1 day). Because this process has no sludge retention nitrite oxidizers are not able to remain in the SHARON reactor and they are washed out. Because SHARON depends on high temperature, it is not suitable for all wastewaters (but many wastewaters high in ammonium also have a high temperature, such as sludge liquor). Furthermore, because there is no sludge retention and the hydraulic retention time is fixed, the volumetric ammonium reactor loading depends on the ammonium concentration. Thus, the process costs also depend on the ammonium concentration (rising costs with decreasing ammonium concentration). Aeration is not only necessary for oxygen supply, but also to strip CO2 from the reactor to control the pH. SHARON still makes use of denitrification (with added methanol) to reduce the nitrite to dinitrogen gas. Methanol is supplied periodically while the aeration is switched off. The stripping of CO2 combined with the addition of methanol neutralizes all the protons formed in Eq. 1– if bicarbonate is the counter-ion for the wastewater ammonium. SHARON has been scaled-up and applied successfully at the Rotterdam wastewater treatment plant, for the treatment of sludge liquor. The 1500-m3 reactor is in operation for 2 years and treats 1000 kg N day−1.
A variation on the SHARON process does make use of sludge retention. Instead of the hydraulic retention time, here the sludge age is controlled (in SHARON, the hydraulic retention time equals the sludge age) . This allows higher ammonium loading rates and more efficient aeration. The process also makes use of a second principle to prevent nitrite oxidation; at low oxygen concentrations (less than 0.4 mg l−1 or 5% air saturation) and with surplus ammonium, nitrite oxidizers are unable to grow, and nitrite becomes the stable end product of nitrification. It is unclear why nitrite oxidizers are inhibited; inhibition of nitrite oxidizers by ammonia and a lower affinity for oxygen and/or nitrite have been suggested as possible explanations, but we still lack mechanistic evidence. This process has not yet been applied at full scale.
The anammox process (Fig. 2) is the denitrification of nitrite with ammonia as the electron donor [90,91]. Anammox needs a preceding partial nitrification step, that converts half of the wastewater ammonium to nitrite. A modified SHARON process has been applied successfully in the laboratory to generate such ammonium/nitrite mixtures [91,92]. By simply not supplying any methanol and removing the anoxic periods, a SHARON reactor yields the desired ammonium/nitrite mixture, without the need for feedback control. This is possible because after 50% of the ammonium is oxidized, the decrease in pH (to 6.7) prevents the oxidation of the remaining ammonium. By limiting the oxygen supply to a nitrification reactor with sludge retention, the same result can be obtained, although feedback control might be necessary .
The first full-scale anammox reactor is currently being built in Rotterdam, The Netherlands, as an addition to the SHARON reactor that is already in place. The reactor is estimated to have a return on investment of less than 7 years, because addition of methanol (currently used to sustain the denitrification) will no longer be required.
Laboratory experiments and design calculations have shown that anammox reactors will be extremely compact with volumetric ammonium loading rates of more than 15 kg N m−3 day−1. Depending on the ammonium concentration of the wastewater and the reactor design, the dinitrogen gas produced by the process could at least partially mix the reactor (analogous to upflow anaerobic sludge blanket process (UASB) reactors), leading to very low power consumption. Additional mixing could be provided by recycling part of the produced dinitrogen gas. Due to the low growth rate of the responsible bacteria, sludge retention is extremely important. The reactor should be well mixed (to keep the redox potential in the ‘denitrification zone’ and prevent formation of toxic sulfide) and should not be overloaded, because high nitrite concentrations (more than 70 mg N l−1 NO2−‘Candidatus Brocadia anammoxidans’, more than 180 mg N l−1 NO2−‘Candidatus Kuenenia stuttgartiensis’) for extended periods are also detrimental to the process [64,66].
On laboratory scale, anammox has been tested in different reactors: fixed bed , fluidized bed , sequencing batch , and gas-lift reactors (unpublished results) all appeared to be suitable, although the economics of the process differ for the different reactor configurations (depending on existing reactors already in place that could be adapted to the process). One of the main challenges of the anammox process is the long start-up time. Because the anammox planctomycetes grow so slowly (see above) it takes between 100 and 150 days before an anammox reactor inoculated with activated sludge reaches full capacity . Experience in anaerobic wastewater treatment (with UASB reactors) has shown that this problem may be overcome once the first full-scale anammox plants are in operation and seeding will become possible.
Canon is an acronym for ‘completely autotrophic nitrogen removal over nitrite’. This concept (Fig. 2) is the combination of partial nitrification and anammox in a single, aerated reactor [89,93,94]. The name ‘Canon’ also refers to the way the two groups of bacteria cooperate: they perform two sequential reactions (Eqs. 1 and 4) simultaneously. The nitrifiers oxidize ammonia to nitrite, consume oxygen and so create anoxic conditions the anammox process needs. Canon has been tested extensively on laboratory scale. The volumetric loading rate (1.5 kg N m−3 day−1 in a gas-lift reactor)  is lower than for anammox and also somewhat lower than has been achieved with high-end dedicated nitrification reactors. However, because only one reactor is required, the economics might still be advantageous when the daily ammonium load is low. Canon would need process control, to prevent nitrite build-up by oxygen excess.
The Canon concept has not been purposefully tested on pilot or full scale, but is known to occur accidentally in sub-optimally functioning full-scale nitrification systems [25,95–97]. Such systems combine three processes (Eqs. 1, 3 and 4), and convert ammonium to mixtures of nitrate and dinitrogen gas. The stoichiometry of the released products is influenced by e.g. the bacterial population and the physical parameters.
Controlling and stimulating the denitrification activity of Nitrosomonas-like microorganisms by adding nitrogen oxides offers new possibilities in wastewater treatment . In the presence of NOxNitrosomonas-like microorganisms nitrify and denitrify simultaneously even under fully oxic conditions with N2 as main product (Fig. 2). Just about 40% of the ammonia load is converted to nitrite. Besides a 50% lower oxygen demand in the nitrification step (since nitrite is used as terminal electron acceptor), the subsequent denitrification step consumes less COD. The N-conversion in a combined nitrification/denitrification without NOx supply is shown in Eqs. 9–11 and N-conversions influenced by nitrogen oxides in Eqs. 12–14. The [H] represents the reducing equivalents (e.g. supplied by an external C-source). These results might vary depending on the composition of the wastewater.
Plant with NOx supply
NOx (NO/NO2) is the regulatory signal inducing the denitrification activity of the ammonia oxidizers and it is only added in trace amounts (NH4+/NO2 ratio about 1000/1 to 5000/1) . As a consequence about 50% of the reducing equivalents [H] are now transferred to nitrite as terminal electron acceptor (Eq. 12) instead of oxygen. Therefore the oxygen consumption of the process is reduced (Eqs. 12 and 14).
The new method for N-removal was developed in a laboratory scale reactor system allowing a performance increase as well as a decrease of the operating costs. The new process offers the possibility to be integrated into existing wastewater treatment plants with minimal financial and technical efforts.
One 2-m3 pilot plant for the treatment of wastewater from intensive fish farming and a pilot plant of the company Nitra GmbH (Germany) at a municipal wastewater treatment plant (sludge liquor) were tested . We will present data of a 3.5-m3 plant. The installation is working for 22 months, treating highly loaded wastewater. Exhaust fume containing ammonia is processed via a washer and the effluent is treated in a nitrification/denitrification system equipped with the NOx method. The highly loaded wastewater (about 2 kg NH4+-N m−3) is fed into the nitrification reactor (3 m3), which is connected to a stirred but not aerated denitrification tank (0.5 m3). The contact between the sewage and biomass is mediated via membrane surfaces (cross-flow filtration). The nitrification step is aerated with about 65 l air min−1 supplied with 200 ppm NO2.
The performance data of the nitrification step (without denitrification step) are presented in Fig. 3. The volume load of the plant was increased from about 2 kg NH4+-N m−3 day−1 to about 4.7 kg NH4+-N m−3 day−1. The denitrification activity of the nitrifying biomass was very sensitive towards the NO2 supply. When the NO2 supply in the nitrification step was turned off between days 100 and 112, this led on short-term to an increased ammonia concentration caused by a reduced ammonia oxidation activity. The long-term effect was more interesting: a significantly increased nitrite concentration was detectable between days 112 and 150. Obviously, the denitrification activity of the ammonia oxidizers decreased when NO2 was not present (days 100–112). Under the influence of NO2, the ammonia oxidizers again increased their denitrification activity (days 112–150).
During the 22-month operation time, the ammonia consumption in the nitrification step was about 3.5 times higher than the nitrite production. Since hardly any nitrate (<1 g NO3−-N m−3) was formed and NO and N2O were only detectable in small amounts (in waste gas <40 ppm), strong evidence is given that the N2 production by ammonia oxidizers is mainly responsible for the average N-loss of about 67% (Fig. 4). Control tests under laboratory conditions confirmed these results [28,42]. The remaining N-load (nitrite) was removed in the small denitrification step with methanol as carbon source. Including the denitrification step into the analysis of the nitrogen balance of the treatment plant, the N-loss was about 97% (Fig. 4). The costs to equip an existing sewage plant with a system for the NOx supply were calculated with EUR 10.000–55.000 depending on the dimension of the plant. The method causes additional operating costs of EUR 0.05–0.08 per kilogram ammonia-N (NO2 supply). The following savings have to been taken into account: the external C-source can be reduced up to 80%, and the supply of oxygen can be reduced up to 50%.
The OLAND process (oxygen-limited nitrification and denitrification) is described as a new process for one-step ammonium removal without addition of COD . Apart from the basic fact that nitrifiers are involved, the mechanism is not yet understood and the ammonium loading rates are low. It seems possible that OLAND will be based on either the Canon concept (a combination of aerobic and anaerobic ammonia oxidizers) or the NOx process (nitrifier denitrification in the presence of NOx).
The ‘aerobic deammonification’ process is another one-step ammonium removal process that does not depend on COD . It has been tested on pilot plant and full scale, and converts part of the ammonium to dinitrogen gas and part to nitrate. Recently, it was discovered that this process is based on the Canon concept, with nitrifiers and anaerobic ammonia oxidizers cooperating under oxygen limitation. Because the ‘aerobic deammonification’ process evolved in reactors designed for conventional nitrification, the process design (a rotating biological contactor) is not optimal and the nitrogen loading rates and removal efficiency are low (Table 1).
Table 1. Operational characteristics of wastewater treatment processes for nitrogen removal
Conventional nitrification denitrification
Aerobic ammonia oxidizers
unknown salt tolerant ammonia oxidizer
Aerobic nitrite oxidizers
Anaerobic ammonia oxidizers
B. anammoxidans, K. stuttgartiensis
B. anammoxidans, K. stuttgartiensis
Biofilms or suspension
NH4+ loading (kg N m−3 reactor day−1)
separate oxic and anoxic compartments or periods, methanol dosing
separated oxic and anoxic compartments, methanol dosing, membrane for sludge retention
aeration needs to be tuned to ammonia loading
separate oxic and anoxic compartments or periods, methanol dosing
preceding partial nitrification needed
aeration needs to be tuned to ammonia loading
aeration needs to be tuned to ammonia loading
two full-scale plants
full scale initiated
two full-scale plants
Over the past 25 years, a significant amount of resources have been invested to construct wastewater treatment plants. Unfortunately, the performance of many of these facilities has not fulfilled the requirements of the discharge permits. In many cases, newly constructed plants have had to be retrofitted or modified at considerable expense to meet the discharge requirements and to provide more reliable performance. The need to conserve energy and resources is well documented, and therefore more attention is being paid to the selection of processes that conserve energy and resources. Operation and maintenance costs plus reliable process control are extremely important to operating agencies. Thus, the operability of treatment plants is receiving more attention. To design and operate a wastewater treatment system (activated sludge system) efficiently, it is necessary to understand the biochemistry of the involved microorganisms and basic research is the keystone to optimize established processes and to invent new and innovative systems. Discovering the group of anammox microorganisms opened new ways in nitrogen removal. Processes like SHARON and Canon have been developed, meeting the needs of treatment plants to handle e.g. high nitrogen-loaded wastewater. Also the discovery of the versatility of aerobic ammonia oxidizers led to the development of new treatment processes (e.g. NOx process). In the future the combination of the different groups of nitrogen converting microorganisms and the optimization of the process management (adaptation according to the wastewater composition; design of the treatment plants, temperature, oxygen and NOx supply) will improve the nitrogen removal. One of these options is a complete nitrogen removal (ammonia to N2) by a mixed population of ‘aerobic’ ammonia oxidizers and anammox bacteria under anoxic conditions in the presents of NO2.
The interest in small treatment systems has often been overshadowed by concern over design, construction, and operation of large regional systems. Small systems were often designed as small-scale models of large plants. As a consequence, many are operationally energy and resource intensive. New and innovative techniques like those described in the review might offer a solution for many treatment plants. The activated sludge process has been used extensively in its original form as well as in many modified forms. In the method used and the design of the process, consideration must be given to selection of the reactor type, loading criteria, sludge production, oxygen requirements and transfer, nutrient requirements, control of filamentous organisms, and effluent characteristics. More specific characteristics for the biological part are the operation factors like reaction kinetics, oxygen transfer, nature of the wastewater, local environmental conditions, construction, operation mode, and maintenance costs.
In view of these considerations, we believe there is no single best process for ammonium removal from wastewater. In each case it has to be evaluated which process is most suitable. Table 1 compares the different characteristics of new and established processes to allow a proper evaluation which method is best for a specific application.
The authors gratefully acknowledge the support of the EU in the 5th framework project ICON (EVK1-CT2000-00054). This is CWE (Center for Wetland Ecology) publication 2002-271.