Nitrogen removal from wastewater has become the focus of the regulating agencies. In the US, some local restrictions in the southeast, the Chesapeake Bay around Washington, DC, and the Long Island Sound North of New York, impose nitrogen discharges below 5mg TN/l.

During biological wastewater treatment, a significant part of nitrogen is bound in the biomass, and gets released during anaerobic digestion and returned to the head of the plant. This return liquor can have concentrations around 1,000mg N/l and add 10-20% of the influent load, or more if the treatment plant is a centralised solids treatment centre and receives sludge imports from other facilities.

To reach low nitrogen concentrations reliably, it can be economical to use specific nitrogen removal treatment on sludge streams before returning them to the head of the plant. To treat high ammonia concentrations in sludge return liquors, several treatment technologies have been applied on a large scale:

  • magnesium ammonium phosphate (MAP) precipitation,
  • stripping at high pH with either steam or air,
  • sequencing batch reactor biological treatment,
  • Sharon (ammonia removal over nitrate),
  • de-ammonification.

The relative cost of each option, based on theoretical calculations, are listed in Table 1.

Chemical options such as precipitation with magnesium ammonium phosphate or steam stripping, tend to be more expensive than biological options. One exception is air stripping, and a handful of installations were installed in Scandinavia about ten years ago – but of which only Oslo in Norway still operating. One of the new promising approaches developed for nitrogen removal from warm and concentrated streams is single reactor high activity ammonia removal over nitrite (Sharon), which allows a short-cut in the traditional nitrogen-removal pathways, reducing energy consumption by 25% and methanol dosage by 40%.

Conventional biological nitrogen removal comprises two steps, each with two separate reactions. Ammonia nitrification will need ammonia oxidisers (nitrosomonas) to convert ammonium to nitrite, and then nitrite oxidisers to (nitrobacter) convert nitrite to nitrate:





Denitrification is an anoxic process where facultative anaerobic micro-organisms use the oxygen bound to nitrate or nitrite for respiration:





As nitrification and denitrification are carried out under different conditions and by different micro-organisms, the processes have to be separated in time or space, creating aerobic conditions first with low carbon concentration, and subsequently achieving anoxic conditions with a sufficient carbon dosage.

At higher temperatures, nitrosomonas prove to have higher maximum growth rates than nitrobacter. Therefore, in a system without sludge retention, they can survive at shorter residence times. In a reactor with a temperature of 30-40ûC and a short residence time, only the first ammonia oxidation to nitrite occurs because nitrobacter is washed out of the system. Therefore the second-step nitrification/denitrification are suppressed to nitritation/denitritation in order to save oxygen and carbon.

This is the principle of the Sharon process, designed to operate at high temperatures, around 35ûC. In one reactor, an aerated zone is followed by a mechanically mixed zone. The high temperature and substrate concentration makes it possible to operate without any sludge recirculation. Results from full-scale plants show that ammonia can be reduced from around 1,000 to less than 1mg N/l and the effluent NOx can be adjusted to a value that will be compatible with the raw influent, typically around 100mg N/l.

The process was developed in the Netherlands by Delft Technical University, is designed by Grontmij, and has a handful of larger references, some of which have been running for more than five years. A Sharon system is currently under construction for the City of New York, and has been evaluated on a few other US sites. But, as shown in Table 2, overall investment and operation for the nitrogen removal infrastructure seem high, and not significantly different from conventional approaches.

Once other cost factors are added in:

  • amortization over 20 years at 4% interest = £0.60/kg N,
  • maintenance at 3% or the investment = £0.4/kg N
  • Personnel cost at twice the maintenance cost = £0.72/kg N
  • The total cost per kg N will be £2.59/kg N similar to the estimate in Table 1.

Anammox the paradox

To substantially reduce the cost of nitrogen conversion, a new step can be added after nitrite generation: anaerobic ammonium oxidation, named anammox. A mixed culture of nitrosomonas and aerobic denitrifiers can eliminate ammonium to a high extent. This ammonia reduction under anoxic conditions and the reaction can simplified be written as:

Anammox: NH4+N02-‘N2+2H20

Anammox is considered to be autotrophic so there is no need for external carbon to support the formation of di-nitrogen (elementar gas). As nitrite is used as electron acceptor, it is interesting to make a process configuration of a pre-nitritation followed by an Anammox. The operational conditions for the process are around 8pH, no oxygen and a high temperature (36ûC). Biomass development is very slow and therefore favoured by biofilms.

One of the most interesting new processes is when nitrogen reduction takes place under aerobic conditions and without any need of carbon with the total reaction called aerobic de-ammonification (De-Amon process). This means transformation of ammonium into dinitrogen under oxygen limitation and without using carbon, carried out by nitrosomonas. It is thought that the ammonia is oxidised to nitrite, which is then denitrified to di-nitrogen – which escapes to the atmosphere:

NH4+ + NO2- _ N2 + 2H2O + NO3

With this process, more nitrogen is removed than that the BOD:N-ratio makes possible, according to the traditional theory of biological nitrogen removal. The growth of the micro-organisms involved in this reaction seems to be favoured by anaerobic pre-treatment, and by wastewater with a very high ratio of N/COD, for instance return liquors from sludge dewatering of anaerobically digested solids or leachate from landfills. But growth is slow and therefore De-Amon has been developed as a biofilm process.

The process reduces nitrogen in sludge water without the need for any carbon source, since only small amounts of nitrate are formed. This results in a very low oxygen demand, and makes the running cost for the process lower than other types of nitrogen removal processes. Fig 1 compares the relative needs for carbon and oxygen of conventional, Sharon and de-ammonification processes for nitrogen removal.

The first full-scale plant using the de-ammonification pathway was treating leachate from a landfill in biological contactors, and has been operational for almost ten years at Mechernich near Bonn in Germany.

This plant was designed by the Technical University of Hannover (TUH) as a multi-stage process: the biological pre-treatment with De-Amon is followed by two-stage reverse osmosis and concentrate treatment through evaporation and drying.

Initially, the throughput of the plant was about 150m3/d, with ammonia reduction efficiency higher than 80%. It was observed that nitrogen losses were not the result of a possibly prevailing inhibition of the oxygen diffusion, that is of a ‘classical’ simultaneous nitrification/denitrification in anoxic zones in the interior of the biofilm. Complementary laboratory tests with homogenised biofilm confirmed that the process is really aerobic de-ammonification and the main end product of the conversion was elementary nitrogen.

To prepare an extension of landfill area used (about 25ha), capacity was increased to around 300m3/d with a new Kaldnes Moving Bed Bioreactor (MBBR) in 1998. This process uses small cylindrical plastic carriers, with a surface of 750m2/m3. Typically, only the inner sheltered surface is considered, and as the fill ratio is 40%, the typical surface considered per reactor volume is 200m2/m3.

Before the moving-bed plant was initiated, de-ammonifying biomass was transferred from the rotating biological contactor into the first nitrification tank, mixed with the moving bed material, and aerated.

After approximately one week, the influent flow was gradually increased in proportion to the nitrification and denitrification performance. The complete

integration into the existing plant was effected at the end of November 1998 through the initiation of the recirculation to the preliminary denitrification unit.

At average surface degradation rates of 2.7g NH4-N/(m2d), the nitrification performance was consistently above 99%. From the very beginning, nitrogen losses could be observed in the nitrification, which from the middle of January 1999 could be increased and stabilized at a de-ammonification rate of about 40-50% in relation to the load of inorganic nitrogen in the moving-bed influent.

Sludge centrate

Following the reliable de-ammonification in Mechernich, a similar system was implemented in early 2001 to treat the sludge return liquors from the solids dewatering in Hattingen near Essen, Germany, using MBBR.

For an average flow of 200m3/d, the total volume of the plant is 319m3, divided into five tanks:

  • a settling tank with a volume of 58m3,
  • three biological reactors, one of 104m3 and two of 67m3,
  • a 24m3 degassing basin, is available to prevent oxygen returns in the recycle.

All reaction basins can be fed individually by bypass, can be aerated, and/or operated under anoxic conditions using agitators, as de-ammonification is taking place under low oxygen concentrations. Each reactor is equipped with a pH control and a chemical dosing system, as shown in Fig 2.

The operating experience in the last two years shows a maximum nitrogen removal of 80% of the influent load between 100 and 160kg N/d. The total investment costs in the De-Ammon demonstration were close to £569,000, of which today almost £126,000 would not be deemed necessary, for instance for carbon dosing or odour control. Compared with the eliminated load of 100kg N/d, the specific investment cost over 20 years is £33,000/(100 x 365) = £0.90/kg N elim.

Very comprehensive operating costs were established by the operator, Ruhrverband, one of the largest public sewage districts in Germany. After process optimisation, pure consumables cost was reduced from £0.49/kg N to £0.39/kg N, or considerably lower than the expected £0.63-0.82/kg N with Sharon.

When all maintenance and personnel cost, as well as amortization is included, operating costs have been reduced from £2.64 to £1.96/kg N , and are therefore close to the value of £1.83/kg N estimated by the Swiss Eawag Institute (Table 1). The ecological advantages of DeAmon in terms of sustainability compared with conventional processes are summarised in Table 3.

Action inspires action. Stay ahead of the curve with sustainability and energy newsletters from edie