Back to: Projects

Phosphate Recovery

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

Full scale phosphate recovery at sewage treatment plant Geestmerambacht, Holland

Ir. Simon Gaastra, Ir. Ruud Schemen, Piet Bakker and Ing. Mathieu Bannink
Waterboard "Uitwaterende Sluizen in Hollands Noorderkwartier"
Post-box 14, NL - 1135 ZH Edam, Holland

Telephone : ** 31 (0) 299 391 424
Fax : ** 31 (0) 299 391 189

Email :

1. Introduction; waterboard US

The waterboard "Uitwaterende Sluizen in Hollands Noorderkwartier" (US) was founded in 1544. Since 1993 it is an all-in waterboard in the north-western part of Holland, which is responsible for:

  • quality control of all surface waters;
  • water quantity (level) control;
  • maintenance of dikes and dunes (in order to keep "dry feet" in the area);
  • maintenance of the roads in several communities.

Its a semi-public organisation, a democratic institution. That means that it has a general board (48 members, the "hoofdingelanden"), which is chosen by all who have interests in the work of the waterboard (inhabitants, industries, communities, farmers). The chairman ("dijkgraaf"), however, is appointed by the Queen. The general board takes all decisions. Uitwaterende Sluizen has about 425 employees.

In order to control the quality surface waters (canals, lakes, ponds, streams) US fulfils several tasks: the formulation of the policy, the planning of actions and projects, the legislation (discharge of sewages is forbidden without a permit), and the treatment of sewage and by-products. Until now the communities collect the sewage from households and industries, the waterboard transports and treats it.

US operates 20 central sewage treatment plants (STPs), 125 pumping stations, 350 kilometres of pressure pipelines and a sludge drying installation. The STPs have a total capacity of 1,76 million population equivalents (pe; 1 pe equals 136 g TOD). We serve a area of about 200.000 hectares , with circa 1,3 million inhabitants (430.000 households) and 25.500 industries. With the planning and operation of the facilities about 80 people are employed.

Based on the Dutch Water Pollution Act the waterboard collects a fee from all inhabitants and industries in the area, based on the principle: who pollutes, pays. At the moment the fee is 100 / pe ( £28,50 or 45 euros).

Since US is responsible for water quality control and is spending public money, important drives with the operation of the treatment plants are:

  • meeting the effluent standards, and
  • meeting the average reduction value for Ptotal (and starting in 2003 for Ntotal as well), and
  • cutting the costs (public money!).

2. P-removal from sewage at US

Based on the Water Pollution Act waterboard US handles a set of effluent standards for each of its STPs. For instance for total phosphorus (Ptotal) it is 1 or 2 mgP/l in the effluent (depending on the capacity of the plant; averaged over 10 measurements).

Since 1995 all waterboards in Holland have to meet a reduction of 75% for Ptotal, as a yearly average over all STPs operated (influent ---> effluent). In a few years this reduction of 75% has to be achieved for total nitrogen (Ntotal) as well.

To meet the mentioned reduction of Ptotal waterboard US chose about ten years ago to implement in principal chemical phosphate removal (simultaneous or preprecipitation), because its a simple and reliable process. So at 7 STPs chemical installations were built and are in operation now.

However, for STP Geestmerambacht biological P-removal in a sidestream process plus crystallizers was chosen. Reasons were:

  • the potentials for phosphate recovery in an separate stream, thus enabling recycling of phosphate, and
  • co-operation in the development of a new process: crystallisation of calciumphosphate from sewage.
Due to the potentials of the process the waterboard received governmental financial support for the project.

In 1997 waterboard US treated 795 ton of Ptotal; about 65% was treated via chemical removal, 10% was treated at STP Geestmerambacht and 25% at the other 12 STPs. The overall removal of Ptotal performed was 77%.

3. Biological P-removal and crystallizers at STP Geestmerambacht; experiences

STP Geestmerambacht: sewage and sludge treatment
STP Geestmerambacht was built in 1973 and was extended and upgraded two times; in 1994 the existing plant was ready and started up. The STP treats the sewage from inhabitants and industries (80 and 20%, respectively). The design capacity of STP Geestmerambacht is 231.600 pe (equals with 340 kg Ptotal/d) and 5.000 m3/h (RWF). For capacities and loads in 1997 see table 1 (STP Geestmerambacht; process data 1997 (average values)).

The main parts of Geestmerambacht are shown in figure 1 (Flow scheme). The water treatment process mainly takes place in an activated sludge system, consisting of:

  • Anaerobic selector (selection of sludge with good settling properties);
  • Denitrification tank (in the new part of the STP);
  • Two oxidation tanks (Carrousels ® with point aerators; also fine bubble aeration in old tank);
  • Six sedimentation tanks.

The main parts of the sludge treatment are:

  • Three gravity thickeners;
  • Homogenizer tank;
  • Two dewatering decanters;
  • Two dewatered sludge bunkers.

The dewatered sludge is transported to and dried at a central sludge drying installation (SDI). The dried sludge is still tipped; tests are conducted for co-incineration at an energy plant.

Phosphate removal at Geestmerambacht
The phosphate removal at Geestmerambacht in fact is a combination of biological and physico-chemical techniques. Phosphate (orthophosphate plus organic phosphate) is transported to the STP with the influent. The biomass in the activated sludge process is using P for cellgrowth, which is quite natural for bacteria to do. However, by subjecting the sludge to a sequence of aerobic and anaerobic conditions, bacteria are grown which use polyphosphate as an energy-carrier. Here they are called the P-bacteria.

Under the aerobic conditions in the oxidation tanks the phosphate from the influent is incorporated by the P-bacteria in polyphosphates. A part of the return sludge is pumped to the (anaerobic) strippertanks in the sidestream (see figure 2: scheme of sidestream).

By splitting off orthophosphate ions the P-bacteria are able to use fatty acids as carbon source under these conditions. The orthophosphate is going into solution. To stimulate this process acetic acid is dosed (solutions of 70 to 85%). The pH-level is controlled at about 7,3 by dosing of sodium hydroxide (NaOH).

Next the stripped sludge and the phosphate-rich water (50 80 mgP/l) are separated in a gravity thickener. The separation process is controlled by a sludge blanket measuring device. The sludge is returned to the activated sludge system; the supernatant goes to the physico-chemical process.

First, in order to reduce the formation of calciumcarbonates, the carbonates are is stripped in a cascade stripper. The pH-value should be controlled at about 5 to shift the equilibria to CO2 (in practice about 3,5, due to a technical problem), by the dosage of sulphuric acid (96% solution). The stripped supernatant is collected in the buffer tanks before the crystallizers.

Figure 3 shows the principle of the crystallizer reactor. The process and reactor are developed by DHV Water BV, Amersfoort, Holland, under the name Crystalactor ®. The reactor consists of a vertical cylindrical vessel, partially filled with a fluidised bed of seeding material, in this case filter sand. The water is pumped through nozzles into the reactor (vertical velocity about 40 m/h). Also in the bottom Ca(OH)2-solution is added via separate nozzles, to provide an elevation of the pH-level and Ca-ions. Through the high turbulence a good mixing of P-rich water and solution is obtained. Crystallisation of calciumphosphate takes place mostly on the seeding sand, by which the pellets grow. Dosing of calcium hydroxide is controlled by pH-measurement (pH about 8). The crystallisation efficiency is enhanced by recirculation over the pump wells; by recirculation 2,5 3 times the crystallisation efficiency is circa 70%.

The bed height is controlled at 4 4,5 metre. The height is adjusted by extracting pellets near the bottom and subsequently adding new seeding sand.

The side stream process, including the crystallizers, is working well at the moment, and can be controlled by the people at STP Geestmerambacht. Since the operation of the sidestream process (dosing of several chemicals to keep setpoints in a narrow range; finding the right level of setpoints; little biomass to be pumped to crystallizers) differs substantially from the processes normally applied at STPs, getting familiar with the process took some time. Anyhow: the team at STP Geestmerambacht gathered lots of experience, also by solving some problems!

The process was started up in 1994: first the biological part and next the cascade stripper and crystallizers. Also since application of the process in the sidestream was new, and the plant was built as a demo site, some technical problems had to be overcome. Not all of them had a direct relation with the sidestream process. Lots of problems in the control system of the STP had to be solved. In the first year the plant had to be operated, without all measurements at the STP being ready.

Severe were the problems with the water nozzles in the crystallizers. Two times heavy abrasion of these nozzles occurred. In the beginning of 1996 we installed new ones of another material and with 3 instead of 12 borings, and since then we dont see abrasion of them anymore.

Two times we had problems with the formation of columns of "glued" pellets in the reactors, probably related to the problems with the nozzles and the control op the sytsem. The reactors had to be emptied and were started again.

Algae grew on the rims in the top the reactors; this growth wasnt seen as a big problem.

Clogging problems with the return sludge pumps, which caused shut down of feeding pumps. By changing the control system the problem was solved.

From the "outside world": problems with an industry. The sewage, rich of phosphate, was discharged in batches. The P-loads in the influent of STP exceeded the design capacity (1000 kgPtotal/d versus capacity 340 kgPtotal/d). Since 1997 this factory is performing chemical P-removal, and the sewage is transported to another location.

Investment and costs; production of pellets
The total investment for the sidestream process was about 9,3 million (£ 2,7 million or 4,2 million euros). The Dutch government supported US for this project with 2,0 million.

The costs of phosphate-removal from the sewage are about 16 / kgPremoved (£4,6 or 7,3 euro / kgPremoved ; in 1996). Capital costs are firm: about 55% of total. The build-up of the variable costs (45% of total) is given in figure 4 (Variable yearly costs (%)).

Production of phosphate pellets in 1998 will be 180 240 tons. The pellets consist for about 11% of P (average). At the moment the pellets are re-used in the production of chicken feed.

4. Possibilities for optimisation at STP Geestmerambacht

Several possibilities are seen to optimise the process of phosphate removal at STP Geestmerambacht. At the moment we are optimising the dosing of acetic acid; the specific dosage is about 15 g/kg ds. Parallel with the reduction of HAc-dosage and with a chosen pH-level the dosing of sodiumhydroxide is reduced. These chemicals form about 44% of the variable costs.

Due to technical problems the dosing of sulphuric acid for CO2-stripping is higher than necessary. By solving the usage of this chemical and parallel of calciumhydroxyde can be diminished; at the moment both make about 13% of the variable costs.

Further reduction of costs could be obtained if we get hold of alternative carbon sources, in stead of the batches HAc-solutions we buy at the moment. A very good variant would be to use specific industrial effluents (we receive these sewages, diluted in the influent, anyhow), or to produce acetic acids from the influent. This has to be looked at.

An alternative process, the introduction of denitrifying phosphate-removal, will be studied.

Optimisations of the side stream process with crystallizer
The plant at Geestmerambacht was designed in 1991, and was built as a demo location. Anno 1998 we have gathered a lot of experience, with which we could design a simpler and more effective plant. As a result the investment and capital costs can be reduced, with about 5 - 10%. The specific costs of phosphate removal would go down with about 2,5 - 5 %.

STP Geestmerambacht is underloaded at the moment (60 65% of design capacity), and so is the sidestream process. At the moment one crystallizer is shut down, and acts as a reserve. For that reason the (specific) costs of phosphate removal are relatively high. If the capacity of the process would be used completely, the specific costs could be reduced.

With the above mentioned optimisations the costs of the process are estimated at about 13 15 / kg P removed (£3,7 4,3 or 5,9 6,8 euro /kg Premoved ).

5. Conclusions

The conclusions, after gathering first-hand experience with the sidestream process, including crystallizers, are:

  • The process can be operated satisfactorily, after solving some operational and technical problems;
  • The process should be operated carefully, like in a physico-chemical industry;
  • Costs of P-reduction at STP Geestmerambacht are 16 / kgPremoved ( £ 4,6 or 7,3 euro / kgPremoved );
  • By optimisation costs could be about 13 15 / kg P removed (£3,7 4,3 or 5,9 6,8 euro /kg Premoved ).
  • The phosphate pellets can be re-used !

Piekema, P.G. and Gaastra, S.B.: Upgrading of a wastewater treatment plant in the Netherlands: combination of several nutrient removal processes. European Water Pollution Control, 3 (1993), nr. 3, 21 - 26.