Back to: Projects

Phosphate Recovery

Home | Introduction | The phosphates file | Current research projects | Bibliography | Contact us


Mark C.M. Van Loosdrecht
Kluyver Institute for Biotechnology
Delft University of Technology
Julianalaan 67
2628 BC Delft
Fax: 31-15-2782355


Phosphate is the limiting component for growth in most ecosystems. Emission of phosphate in surface waters leads to eutrophication and blooming of algae. This has negative impacts on nature conservation, recreation, and drinking water production. Therefore it is essential to control the emission of phosphates from discharges of waste water. Governments have implemented regulation on phosphate emissions which in turn pushed the development of techniques for phosphate removal. With a stronger emphasis on a more sustainable society nowadays, also the recovery of phosphates from wastewater gets more attention.

Most wastewater is diluted and contains only low amounts of phosphorus (several milligrams) it is therefore not easy to recover phosphate in one step. Phosphate has concentrations of typically 10 mgP/l or below in most European wastewater. Effluent standards are typically in the range of 0.5-1 mgP/l, i.e. removal efficiencies of over 90 % have to be achieved. Chemical processes have as disadvantage that they are not selective. Since wastewater contains many different ions, these will always disturb the processes for chemical P-recovery. The lower the concentration needed to achieve (or the higher the percentage recovery) the more side-processes will occur. This either leads to a product with a high amount of contaminants or a train of pre-treatment techniques. Biological processes can be highly selective, and can achieve easily low concentrations. Disadvantage is that the phosphate is only concentrated in the biomass. If the phosphate is set free from the biomass in a small volume a concentrated phosphorus solution is obtained where physical-chemical techniques can be used for recovery. A combination of a biological process for concentrating the phosphates and a physical-chemical process for recovery seems therefore to be the best option for phosphorus recovery.

In this contribution an overview of the state of art of biological P-removal is given. This process is discussed in the context of recovery of phosphates from wastewater and an optimal "sustainable" wastewater treatment process.




The first indication of biological phosphate removal occurring in a wastewater treatment process was described by Srinath et al. (1959) from India. The effluent of a treatment plant was used as fertiliser in rice paddys. At a certain moment the rice wouldnt grow due to a deficiency in phosphate. Shrinath et al. observed that sludge from this treatment plant exhibited excessive (more then needed for cell growth) phosphate uptake when aerated. It was shown that the phosphate uptake was a biological process (inhibition by toxic substances, oxygen requirement), and could be prevented when the initial stage of the plug flow process was properly aerated. Later, in more (plug flow) wastewater treatment plants this so-called enhanced phosphate removal was noticed.

Based on work by different groups biological Phosphate removal has become a full grown technology which can be applied on large scale. At this moment however, still no pure culture representative for the biological P-removal process has been isolated (Van Loosdrecht et al. 1997). This leads to much confusion because papers can have conflicting information. One of these is the term "luxury uptake" or "overplus effect" in the context of biological P-removal. These terms are used for the P-uptake exhibited by many (if not almost all) micro-organisms. When micor-organisms are after an anaerobic period, or a period without phosphate present, subject to an aerobic period with phosphate in the medium usually polyphosphate is accumulated. In these cases the P-uptake is a kind of transient response and not part of the normal metabolism. The polyphosphate accumulated by these organisms is only used as phosphorus reserve. The organisms involved in biological P-removal, however, accumulate polyphosphate as part of their normal physiology, and moreover can use the stored polyphosphate as energy storage (Van Groenestijn et al. 1989).

The physiology of the organisms involved had to be fully inferred from measurements on enhanced cultures. In these laboratory cultures bacteria are cultivated in such a manner that the bio-P organisms are favoured. This done in a cyclically operated batch reactor (SBR proces) where acetate from the influent is accumulated by the sludge in absence of oxygen or nitrate, after which oxygen or nitrate is dosed to the system allowing for cell growth and phosphate uptake. It is assumed that no other organisms then the bio-P organisms can grow in this system, since they only can derive energy from polyphosphate needed in the uptake of acetate. The physiological processes are further derived from these cultures by measuring conversion of all kind of compounds and by techniques such as in-vivo NMR.


The physiology is described in several reference papers such as by Arun et al. 1987, Comeau et al. 1986, Kuba et al. 1993, Maurer et al. 1997, Mino et al. 1997, Perreira et al. 1996, Smolders et al. 1994 a,b. The physiology is typically divided in two phases: (I) an anaerobic phase in which substrate (fatty acids) are accumulated and stored inside the cells, the energy for this process is derived from polyphosphate and (II) an aerobic phase where the internally stored substrate is used for growth and phosphate uptake.

Under anaerobic conditions the bacteria use stored poly-phosphate as energy source for ATP production with the aid of the enzyme Poly-P:AMP-phosphotransferase (Van Groenestijn et al. 1987). ATP is used for the uptake of volatile fatty acids (VFA) and subsequent formation of polyhydroxyalkanoates (PHA). The reduction equivalents needed for the reduction VFA to PHA is derived from the conversion of glycogen to PHA (Arun et al. 1987, Smolders et al. 1994a). Since the transport energy for VFA and phosphate over the cell membrane is strongly influenced by the pH, the pH has a strong effect on the ratio between VFA uptake and phosphate release (Smolders et al. 1994a).

When oxygen, nitrate or nitrite are present in the absence of substrate, PHA is used as substrate. Under these conditions the bacteria not only produce new biomass but also restore the storage pools of polyphosphate and glycogen. This leads to a net uptake of phosphate in the overall process. If substrate as well as electron acceptors are present the substrate is predominantly converted into PHA instead of being used for growth (Kuba et al. 1994, Brdjanovic et al 1997).

The presence of three storage polymers in the cells (PHA, Glycogen and polyphosphate) makes the microbiology of the bio-P organisms extremely complex. A further complicating factor is that the growth of the cells is not directly coupled to substrate oxydation, as in normal microbiological processe. Growth results from the difference between PHA consumption rate and PHA use for glycogen and polyphosphate formation (Murnleitner et al. 1997).


For normal bacteria their presence or absence in a treatment process is directly coupled to the sludge age of the process. For bio-P organisms this is not the case. The substrate they prefer - volatile fatty acids - can be used by hundreds of other normal heterotrophic bacteria which will under normal conditions do this more rapid and efficient then the phosphate accumulating bacteria. If one desires to accumulate the bio-P organisms for removal of phosphate from the waste water it is therefore needed to give them a selective advantage by properly using their unique metabolism. The possession of polyphosphate as energy reserve makes that in the absence of oxygen or nitrate the bio-P bacteria can accumulate substrates such as VFA inside their cells. By introducing an anaerobic period at the first stage of a treatment process in which sludge (bacteria) and wastewater are mixed fermentative bacteria will convert the organic carbon compounds into VFA. These can then be taken up by the bio-P organisms and stored inside their cells. If after this process has been complete the sludge is supplied with nitrate or oxygen the bio-P organisms have substrate inside their cells, whereas other bacteria do not access to this substrate and cannot grow. The bio-P organisms can grow and become in this way the dominant bacterium in the treatment process.

Process design

Any process for biological P-removal will also have to remove Nitrogen compounds. The latter is performed by an ammonium oxidation followed by a nitrate reduction process (nitrification-denitrification). Therefore a wastewater treatment process consists generally of three stages through which the sludge flows:

  • Anaerobic compartment: needed for the selection of bio-P bacteria. It is crucial that no oxygen or nitrate is introduced in this compartment. The retention time in this compartment will generally depend on the rate of fermentation of the complex organic carbon to VFA.
  • Anoxic compartment: needed for denitrification and P-uptake. It is advantageous to remove phosphorus by organisms that also denitrify. This will save significant amounts of organic carbon (Kuba et al. 1993,1996,1997, Sorm 1996). Accumulation of these kind of denitrifying bio-P organisms can best be achieved in UCT-like process configurations, in which sludge is continuously recycled between the anaerobic and the anoxic stage. The retention time in the anoxic reactor is determined by the denitrification rate or the hydrolysis rate by which particulate organic matter is converted into soluble substrate available for denitrification.
  • Aerobic compartment: needed for nitrification and P-uptake. The aerobic compartment is mainly needed for the conversion of ammonium to nitrate. The formed nitrate is then recycled with the sludge to the anoxic compartment. The retention time in the aerobic compartment is determined by the nitrification rate.
From above description it becomes clear that the phosphate removal process is usually not determining the size of the treatment process. This makes that a biological P-removal process can be easily introduced in an existing treatment process. The main criterion for obtaining a good removal efficiency is the amount of bio-P bacteria formed. These are needed to accumulate the phosphate. The cells can maximal contain approx. 12% P on dry weight basis. This means that in the influent there needs to be enough organic carbon relative to phosphorus. As a rule of thumb approx. 20 gCOD is needed per gP which need to be removed as poly-P (remember that aprox. 30 % of the P removed already by normal cell growth). In order to optimise the process it is needed to realy ensure anaerobic conditions in the anaerobic zone. In a correct process configuration this is relatively easy to obtain by a control scheme based on redox electrode measurements (Van Loosdrecht et al. 1997, 1998).


Biological P-removal has to be implemented often in already existing process configurations. Sometimes this is directly possible in the existing reactor volume. However often the anaerobic compartment has to be added as an extension. Upgrading the treatment plant for bio-P removal needs to done cost effective and with a stable process, including a stable sludge volume index. Taking this into account an upgrading strategy has been developed and a specialised reactor has been designed which can easily be integrated in an existing treatment plant (Van Loosdrecht et al. 1998).



Regularly there is not enough COD present in the waste water to ensure a stable biological P-removal. In this case the biological process can be supported by a chemical precipitation. Hereto it is essential that the chemical precipitant does not accumulate in the activated sludge. This will lead to a lower sludge age and consequently higher sludge production and lower nitrogen removal efficiencies. Therefore a range of different processes have been developed in which the P is precipitated in a separate tank or reactor. The most applied variant is the pho-strip process. In this process all or a fraction of the return sludge is brought to an anaerobic tank where eventually some acetic acid is added. The sludge will release phosphorus which can then be precipitated. The precipitation can be done after sludge/water separation. In this case the phosphate can be recovered by e.g. crystallisation (Eggers et al. 1991). Since P can be precipitated at a higher concentration the required amount of chemicals is minimised. An elegant other alternative is retrieving P-rich supernatant at the end of the anaerobic phase and add recover the P from this flow (van Loosdrecht etal. 1998). This alternative doesnt require an extra reactor, and there is no need for acetate addition.

It could be realised that when bio-P organisms are only used to concentrate the phosphate much less bio-P bacteria are needed, i.e. less COD is needed. This is particularly of interest if there is a need for COD addition. It can be calculated that a normal bio-P process requires approx. 20 mgCOD/mgP removed compared to 2 mgCOD/mgP when all the P-is eventually precipitated by a pho-strip like process (Smolders et al. 1996).




Biological phosphorus removal is a grown up technology which can be applied. The only disadvantage might be its complexity, skilled people are needed for the design and supervision of such treatment plants. Biological can be used as a concentration step for the phosphate in diluted waste water. In a separate tank the phosphate can be released from the sludge after which the concentrated phosphate can be subjected to e.g. a crystallisation process.

Alternative process options for P-recovery can however also be evaluated. Burning and Gasification of biological materials is becoming more and more popular. If the P is accumulated in the sludge it will after these processes be accumulated in the ashes. It s worthwhile to evaluate whether the phosphorus can be recovered from these ashes.

A completely different alternative can be found when one realises that 80 % of the phosphate in the waste water originates from urine where the P is present in a high concentration. If an innovative collecting technique can be developed it would become relatively easy not only to recover phosphorus but also ammonium and other minerals.




Arun V, Mino T, Matsuo T (1987) Biological mechanisms of acetate uptake mediated by carbohydrate consumption in excess phosphate removal systems. Wat. Res. 22:565-570.

Comeau Y, Hall KJ, Hancock REW, Oldham WK, (1986) Biochemical model for enhanced biological phosphorus removal. Wat.Res. 20:1511-1521.

Eggers E, Dirkzwager AH, Van den Honing H (1991) Full scale phosphate crystallisation in a crystalactor. Wat. Sci. Tech. 24:333-334

Kuba T, Smolders GJF, Van Loosdrecht MCM, Heijnen JJ (1993) Biological phosphorus removal from wastewater by anaerobic-anoxic sequencing batch reactor. Wat. Sci. Tech. 27:241-252.

Kuba T, Van Loosdrecht MCM, Heijnen JJ (1996) Phosphorus and nitrogen removal with minimal COD requirement by integration of denitrifying dephosphatation and nitrification in a two-sludge system. Wat. Res. 30:1702-1710.

Kuba T, Van Loosdrecht MCM, Brandse F, Heijnen JJ (1997) Occurrence of denitrifying phosphorus removing bacteria in modified UCT-type wastewater treatment plants. Wat. Res. 31:777-787.

Maurer M, Gujer W, Hany R, Bachmann S (1997) Intracellular carbon flow in phosphorus accumulating organisms in activated sludge systems. Wat. Res. 31:907-917.

Mino T, Van Loosdrecht MCM, Heijnen JJ (1997) Microbiology and biochemistry of the enhanced biological phosphate removal process. Wat. Res. Accepted.

Perreira H, Lemos PC, Reis MA, Crespo JPSG, Carrondo MJT, Santos H (1996) Model for carbon metabolism in biological phosphorus removal processes based on in-vivo 13C-NMR labeling experiments. Wat. Res. 30:2128-2138.

Smolders GJF, Van der Meij J, Van Loosdrecht MCM, Heijnen JJ (1994a) Model of the anaerobic metabolism of the biological phosphorus removal process; stoichiometry and pH influence. Biotechn. Bioeng. 43:461-470.

Smolders GJF, Van Loosdrecht MCM, Heijnen JJ (1994b) A metabolic model for the biological phosphorus removal process. Wat.Sci.Techn. 31:79-93.

Smolders GJF, Van Loosdrecht MCM, Heijnen JJ (1996) Steady state analysis to evaluate the phosphate removal capacity and acetate requirement of biological phosphorus removing mainstream and sidestream process configurations. Wat. Res. 30:2748-2760.

Sorm R, Bortone G, Saltarelli R, Jenicek P, Wanner J, Tilche A (1996) Phosphate uptake under anoxic conditions and fixed film nitrification in nutrient removal activated sludge system. Wat. Res. 30:1573-1585.

Srinath EG, Sastry CA, Pillai SC (1959) Rapid removal ofnphosphorus from sewage by activated sludge. Experienta XV:339-340.

Van Groenestijn JW, Bentvelsen MMA, Deinema MH, Zehnder AJB (1989) Polyphosphate degrading enzymes in Acinetobacter spp. and activated sludge. Appl. Environm. Microbiol. 55:219-223.

Van Loosdrecht MCM, Smolders GJF, Kuba T, Heijnen JJ (1997b) Metabolism of microorganisms responsible for biological phosphorus removal from wastewater. Ant. v. Leeuwenhoek 71:109-116.

Van Loosdrecht MCM, Kuba T, Van Veldhuizen HM, Brandse FA, Heijnen JJ (1997c) Environmental impacts of nutrient removal processes: case study. J. Env.Eng. 123:33-40.

Van Loosdrecht MCM, Brandse FA, De Vries AC (1998) Upgrading of wastewater treatment processes for integrated nutrient removal - the BCFS process Wat. Sci. Technol. Submitted.