To date, phosphorus recovery as struvite in wastewater treatment plants has been mainly implemented on water phases resulting from dewatering processes of the sludge line. However, it is possible to recover struvite directly from sludge phases. Besides minimising the return loads of phosphorus from the sludge line to the water line, placing such a process within the sludge line is claimed to offer advantages such as a higher recovery potential, enhanced dewaterability of the treated sludge, and reduced speed of scaling in pipes and dewatering devices. In the wastewater treatment plant at Leuven (Belgium), a full-scale struvite recovery process from digested sludge has been tested for 1 year. Several monitoring campaigns and experiments provided indications of the efficiency of the process for recovery. The load of phosphorus from the sludge line returning to the water line as centrate accounted for 15% of the P-load of the plant in the reference situation. Data indicated that the process divides this phosphorus load by two. An improved dewaterability of 1.5% of dry solids content was achieved, provided a proper tuning of the installation. Quality analyses showed that the formed struvite was quite pure.

INTRODUCTION

Future depletion of high-grade phosphorus rock mines is increasingly present on the political agenda of European spheres (European Commission 2013a, b). As a consequence, recovery and recycling of phosphorus is being encouraged (ESPP 2013). Studies evaluated that about 15–20% of the annual quantity of mined phosphorus eventually ends up in human excreta (Cordell et al. 2009; Clift & Shaw 2012). This provides an order of magnitude of the recovery potential in the municipal sewage sector, and makes this sector one of the most important for phosphorus depletion mitigation, after animal manure and agricultural efficiency improvement (Cordell et al. 2009). The most straightforward way to recycle phosphorus from wastewater treatment plants (WWTPs) is to valorise the sludge as fertiliser on agricultural surfaces (Linderholm et al. 2012). Nevertheless, in some countries, like Belgium (Flanders), spreading of municipal sludge on agricultural land has been forbidden. Therefore, other phosphorus recycling strategies have to be developed. Existing alternatives range from recovery from concentrated wastewater streams (filtrates, centrates, sludge, etc.) to recovery from fly ashes from incineration of sewage sludge. The first option usually uses a precipitation principle of orthophosphate with Mg (struvite) or Ca (apatite) after a pre-concentration step or not (e.g. electrodialysis or exchange ion resin). The precipitation can possibly be aided by electrochemistry (Cusick et al. 2014; Kruk et al. 2014). The second option (recovery from ashes) consists of either acidic and/or alkaline leaching (Petzet et al. 2012) or high-temperature thermal processes (Weigand et al. 2013).

But one of the most commonly implemented strategies so far remains the precipitation of orthophosphorus as struvite (MgNH4PO4·6H2O), which, once recovered, can be used in agriculture as a slow release fertiliser, often mixed with common fertiliser.

This struvite has been recovered in sewage plants at full scale, mostly from the supernatant phase resulting from the dewatering process, in other words from water phases (Cornel & Schaum 2009). But this same process can also be implemented directly on sludge phases of the sludge line. Indeed, high free orthophosphate concentrations are usually met after an anaerobic step, like digestion, on plants where enhanced biological phosphorus removal (EBPR) is employed. It should be underlined that EBPR is a prerequisite to the direct implementation of this technique. If phosphorus is removed chemically with iron or aluminium salts on the water line, orthophosphates remain bound to these metals and are not readily available for struvite precipitation on the sludge line (Parsons & Smith 2008).

Compared to implementation on water phases, such an alternative is claimed to offer different advantages at full-scale, further referred to as ‘operational benefits’, that could compensate the investment of a phosphorus recovery installation. Theoretically, the operational benefits that can be expected are the following:

  • (1) An enhanced sludge dewaterability: if the dewatering of the treated sludge can achieve a higher dry solids percentage of just a few points, this would mean important energy savings when the sludge gets transported, mono-incinerated or dried afterwards (due to the diminished quantity of water to transport or evaporate).

  • (2) A better recovery potential: as polymer solution is added to the sludge for the dewatering, causing considerable dilution, the concentration would be higher in the sludge before dewatering than in the centrate. As precipitation processes are limited to residual concentrations of the equilibrium, the higher the initial concentration, the higher the yield.

  • (3) A reduced scaling: natural struvite precipitation in digested sludge lines is known to cause operational problems like pipe clogging and valve freezing, requiring regular and time-consuming pipe maintenance (Borgerding 1972; Munch & Barr 2001; Neethling & Benisch 2004). Slowing down this process would allow some cost savings.

  • (4) A reduced phosphorus content in the dried sludge pellets in case of subsequent drying: the dried sludge in Belgium can be valorised by the cement industry, which requires the P-content to be as low as possible to guarantee good hardening properties of the cement. Therefore, a diminution of the phosphorus content would be considered as an improvement (Husillos Rodriguez et al. 2013).

  • (5) A reduction of the P and N loads in the rejection waters from the dewatering that gets recycled to the water line: the withdrawn P and N as struvite on the sludge line would provoke a reduction of and concentrations in rejection waters of the centrifuges. This would diminish the P and N load on the water line, allowing a decrease in aeration (needed for nitrification) and carbon source consumption (sometimes needed for the bio-P removal and/or for denitrification).

Operational benefits (1–4) are specific to implementation on sludge, whereas benefit (5) is also obtained with an implementation from rejection water from dewatering.

To our knowledge, very few full-scale pilot experiences have been published to date to explore the technical feasibility of struvite recovery from sludge and to confirm or deny these expected operational benefits. A struvite recovery process has been built and implemented in the municipal WWTP of Leuven city (Belgium, 120,000 people equivalents) from March 2013 onwards. This paper presents in a first part the results of the first year of full-scale operation as well as a quantification of the above-mentioned operational benefits, thanks to numerous monitoring campaigns. Also, in order to screen the potential valorisation market of the formed struvite, the elemental composition and the purity of the product were determined. This return on experience allows to identify the future technical challenges of struvite recovery from sludge. In a second part of this article (Geerts et al. 2015), all operational benefits quantified in part I will be used as inputs for a cost–benefit and sensitivity analysis dealing with the economic opportunities and risks of struvite recovery from digestate. A theoretical comparison with recovery from centrate will also be performed in part II.

MATERIALS AND METHODS

Water line

The raw municipal wastewater comes from combined sewers at an approximate rate of 40,000 m³/day. It first gets pre-treated by a coarse grid filtration followed by a sand and fat trap. There is no primary sedimentation. The biological steps consist of anoxic contact tanks, oxidation ditches and settling tanks. The anoxic/anaerobic tanks, necessary for the EBPR, have a retention time of around 3 hours. The oxidation ditches provide alternate aeration and non-aeration periods, whose durations are optimised based on continuous online oxygen, ammonium and nitrate monitoring. The hydraulic retention time in the oxidation ditches is about 15 hours.

Monitoring campaigns

Three monitoring campaigns on the sludge line of the WWTP Leuven were performed in 2013 during spring, summer and autumn. Figure 1 shows a scheme of the sludge line with the seven locations of the daily sampled flows. Points 1, 2, 5 and 7 were grab-sampled while points 3 and 4 were averaged over a half day (composite samples). The analyses for each sample are described in Table 1. Online flow meters constantly registered the flow at points 2, 3 and 5 and the dry solids content at point 5 and at the inlet of the thickening tables. Flows at points 1 and 7 were calculated based on mass balances on dry solids and water, respectively, at thickening tables and centrifuge.

Table 1

Description of the measured physico–chemical parameters for each sampling point

Sampling pointsMeasured parametersMeasured parameters after filtration (pore size 0.45 μm)Analyses set 1Analyses set 2Analyses set 3
1. Thickened sludge Analyses set 1 Analyses set 2 
  • pH

  • Dry matter

  • Ash rest on dry matter

  • Nitrogen Kjeldahl

  • Phosphorus total

  • Iron total

  • Magnesium total

  • Calcium total

  • Aluminium total

 
  • Nitrogen Kjeldahl

  • Ammonium

  • Phosphorus total

  • Orthophosphate

  • Iron total

  • Magnesium total

  • Calcium total

  • Aluminium total

 
  • pH

  • Suspended solids

  • Ash rest on suspended solids

  • Nitrogen Kjeldahl

  • Phosphorus total

  • Iron total

  • Magnesium total

  • Calcium total

  • Aluminium total

  • COD, BOD

 
2. After buffer 1 before digester Analyses set 1 Analyses set 2 
3. Just before struvite process Analyses set 1 Analyses set 2 
4. Just after struvite process Analyses set 1 Analyses set 2 
5. Just before centrifuge (before polyelectrolyte addition) Analyses set 1 Analyses set 2 
6. Centrate Analyses set 3 Analyses set 2 
7. Dewatered sludge Analyses set 1 Analyses set 2 
Sampling pointsMeasured parametersMeasured parameters after filtration (pore size 0.45 μm)Analyses set 1Analyses set 2Analyses set 3
1. Thickened sludge Analyses set 1 Analyses set 2 
  • pH

  • Dry matter

  • Ash rest on dry matter

  • Nitrogen Kjeldahl

  • Phosphorus total

  • Iron total

  • Magnesium total

  • Calcium total

  • Aluminium total

 
  • Nitrogen Kjeldahl

  • Ammonium

  • Phosphorus total

  • Orthophosphate

  • Iron total

  • Magnesium total

  • Calcium total

  • Aluminium total

 
  • pH

  • Suspended solids

  • Ash rest on suspended solids

  • Nitrogen Kjeldahl

  • Phosphorus total

  • Iron total

  • Magnesium total

  • Calcium total

  • Aluminium total

  • COD, BOD

 
2. After buffer 1 before digester Analyses set 1 Analyses set 2 
3. Just before struvite process Analyses set 1 Analyses set 2 
4. Just after struvite process Analyses set 1 Analyses set 2 
5. Just before centrifuge (before polyelectrolyte addition) Analyses set 1 Analyses set 2 
6. Centrate Analyses set 3 Analyses set 2 
7. Dewatered sludge Analyses set 1 Analyses set 2 

COD: chemical oxygen demand; BOD: biochemical oxygen demand.

Figure 1

Phosphorus mass balance of the sludge line (loads are expressed in kg of phosphorus per day; percentages as percentage of the total load received in the plant from the sewers).

Figure 1

Phosphorus mass balance of the sludge line (loads are expressed in kg of phosphorus per day; percentages as percentage of the total load received in the plant from the sewers).

Dewaterability tests

Four dewatering tests were performed, one at each season during the weeks with the struvite process in operation (‘ON-weeks'), by means of a mobile filter press. 1 m³ of sludge was sampled at the inlet of the struvite process (point 3) and dewatered with three repetitions in spring and summer and two repetitions in autumn and winter. For each dewatering repetition, the obtained percentage of dry solids of the dewatered sludge was analysed for three different dewatered sludge grab samples. One retention time of the struvite process later, the procedure was repeated for the sludge at the outlet of the struvite process (point 4).

Struvite recovery process

The struvite process is a two-tank reactor consisting of (1) a CO2 stripper tank and (2) a reactor, mixed by a double propeller, in which MgCl2 (and also possibly NaOH) is dosed. The harvester, composed of a cyclone, allows a partial separation of the crystals from the sludge. The crystals retained by the cyclone can be either recirculated in the reactor or harvested. The technology provider is NuReSys®. A scheme is presented in part II of this article (Geerts et al. 2015).

Quality analyses of the crystals

The elemental composition and the crystalline structure were determined on the struvite samples, dried at 40 °C overnight. A relatively low drying temperature was maintained to avoid degradation of the crystalline structure. The elements N and H were determined according to the Dumas method using an element analyser (Truspec CHN, LECO) at a temperature of 850 °C. The heavy metals, together with Mg and P, were determined with inductively coupled plasma-atomic emission spectroscopy (ICP-AES, Optima 3000, Perkin Elmer) after HNO3:HCl digestion using a hot block system (DigiPrep, SPC Science). The digester was slowly heated up to a temperature of 105 °C and maintained at that temperature for 2 hours. The total carbon content was measured by means of a TOC analyser (Multi EA 400, Analytik Jena) at a combustion temperature of 1,300 °C. Besides these specific elemental analyses, all samples were screened using a high-performance energy dispersive X-ray fluorescence spectrometer with polarised X-ray excitation geometry (EDXRF, XEPOS HE, Spectro), as it performs a qualitative and semi-quantitative determination of all elements present in a fast and non-destructive manner. Furthermore, X-ray diffraction spectrometry (XRD, X'Pert Pro MPD, PANalytical) was applied to characterise the crystalline structure and therefore, the purity of the formed struvite. Finally, a thermo-gravimetrical determination (TGA-DSC STA 449, Netzsch) was conducted on a wet sample, providing information on the change in weight through a heating process from 20 to 1000 °C at a rate of 20 °C min−1.

RESULTS AND DISCUSSION

P and N mass balances – impact of the struvite process on the water line

Recoverable amount

Figure 1 represents the mass balances of phosphorus. In reality, the mass balances undergo important seasonal variations. That is why the figures displayed are based on the average of the mass balances obtained in the three intensive measurement campaigns. We can see in Figure 1 that the struvite process allows conversion of about 15% of the phosphorus load (30 kgP/day) received by the plant via the sewers (200 kgP/day). This divides by two the phosphorus load in the centrate (it decreases from 15 to 7.5% of the plant load), which is non-negligible regarding the functioning of the water line. Conversely, the impact of the struvite process on the nitrogen mass balance is minor. The withdrawn nitrogen under the form of struvite is about 1–2% of the N load of the plant and the decrease of the N load in the centrate was too small to be measurable.

The influence of the external sludge on the mass balance is limited because they come from other plants where EBPR is not implemented. Therefore, the phosphorus that they bring is not recoverable (not under the form of free orthophosphates). That's why only the influent from the sewers is kept as a reference and the external sludge is ignored from the recovery percentage calculation. It is still unclear whether this recovery potential would be different or the same with an implementation on centrate. Theoretical calculations with a simulated implementation in centrate (at point 6 on Figure 1) leads to a similar struvite removal if the residual orthophosphate concentration after the struvite process is assumed to be 20 mgP-PO4/L. But this residual concentration was never tested and a real implementation would be the only way to provide evidence of this statement.

Diminution of the phosphorus content in the dried sludge pellets

From Figure 1, we can estimate that the withdrawal of 30 kgP/d at the struvite process will induce a maximum of 10% diminution of the P-content of the dewatered sludge (320 kgP/d), and thus of the pellets. In the specific case of Leuven, where about the same quantity of external dewatered sludge is added, the theoretical diminution of the P-content in the dried sludge pellets would be of about 5% maximum.

Advantages of a direct treatment on the digestate

Improved dewaterability

The divalent cation bridging theory, as explained by Sobeck & Higgins (2002) states that flocculation, which is strongly linked to dewaterability, is driven by the ratio of divalent cation concentrations (Ca2+, Mg2+) over monovalent cations (Na+, K+, ). Divalent cation create bridges between particles whereas monovalent cations tend to deteriorate flock structures. Therefore, an improved dewaterability can be expected if the addition of magnesium divalent cations surpasses the effect of sodium hydroxide dosing. The molar ratio of divalent over monovalent cations of the sludge is then an important parameter for the dewaterability. Besides, the dewaterability of sludge is most probably not correlated to the concentration in orthophosphate (STOWA 2012).

Table 2 shows the results of the filter press tests during the four seasons. Statistically, there were no significant difference between dewaterabilities before and after the struvite pilot as P-values are all above 0.05 (results of a one-way ANOVA test). Dewaterability was worse in spring and summer, but better in autumn and winter. The worse dewaterability in spring is explained by the use of a type of polymer which was not adapted to the sludge treated by the struvite process. Indeed, two different types of polymers are required for the sludge before and after the struvite process. In the experiment a more cationic polymer is required for the treated sludge, supposedly due to the dosing of the monovalent chloride anion in the struvite process. In summer, proper polymers were used for both sludges before and after the process, but the NaOH setpoint (pH control in the reactor) was particularly high and lower stripping intensity was used, inducing a particularly low divalent over monovalent ratio in the sludge. On the contrary, in autumn and winter, the NaOH dosing could be limited relative to the dosing of MgCl2 dosing thanks to an enhanced CO2 stripping. MgCl2 dosing was higher in autumn than in winter, explaining the more pronounced improvement of +2.5% of dry solids in autumn. The free Mg2+ concentration increases from 5 mg/L before the struvite process to about 50 mg/l after treatment, which significantly increases the molar divalent over monovalent cation ratio in the sludge. In addition, the struvite process increases the concentration of free Ca2+, Fe3+ and Al3+, which also contributes to the improved dewaterability. These increased concentrations have been observed during all three seasonal measurement campaigns and with analyses performed at the occasion of the filter press tests. The release of Ca2+, Fe3+ and Al3+ can be explained by the dissociation of some aluminium, calcium and iron phosphate salts in response to the sharp decrease of orthophosphate concentration inside the reactor (which eliminates the supersaturation of these salts). Lee et al. (2013) have indeed demonstrated that the very high thermodynamic stability of struvite (pKs = 12.6 at 25 °C) can result in the re-dissolution of other phosphate salts like calcium phosphate if a sufficient contact time is provided.

Table 2

Results of the dewaterability tests in spring, summer, autumn and winter

Dry solids content achieved after dewatering (%DS)Before the struvite process (point 3 on Figure 1)After the struvite process (point 4 on Figure 1)P-value
Spring 21.9 20.8 0.23 
Summer 30.4 29.1 0.12 
Autumn 24.1 27.4 0.08 
Winter 23.4 24.3 0.15 
Dry solids content achieved after dewatering (%DS)Before the struvite process (point 3 on Figure 1)After the struvite process (point 4 on Figure 1)P-value
Spring 21.9 20.8 0.23 
Summer 30.4 29.1 0.12 
Autumn 24.1 27.4 0.08 
Winter 23.4 24.3 0.15 

Bergmans et al. (2014) also measured on a laboratory scale an improved dewaterability of digested sludge after the same kind of treatment (CO2 stripping and MgCl2 addition). The evolution of the dewaterability was also positively correlated to the increase in Mg/P dosing ratio. They also observed that the gain in dewaterability diminishes with a pH increase between 7 and 8. Therefore the operating pH inside the reactor is very important, and should be determined balancing the gain in struvite recovery (which increases with pH) and the dewaterability (the gain slightly decreases with pH).

In WWTP Leuven, this influence of the struvite process on the dewaterability (deterioration in spring and winter; improvement from autumn and winter) were also observed in the dewatered sludge from the centrifuges at full scale, confirming the conclusions of the filter press tests.

A moderation of the use of caustic solution (which brings Na+) by a sufficiently powerful aeration is fundamental to both dewaterability and the limitation of the consumables costs. Fattah et al. (2008) achieved a phosphorus removal over 85% without caustic addition providing a proper aeration on supernatant. This project demonstrates at full scale that it is also possible on a digested sludge fraction.

Prevention of natural scaling

As shown in Table 3, the struvite process operation strongly diminishes the orthophosphate concentration and to a small amount the ammonium concentration. In theory, this should reduce the speed of scaling downstream in the process. For the rest, it is still too early to state if, in the case of WWTP Leuven, it will allow significant economic savings due to less pump and centrifuge maintenance.

Table 3

  Observed modification of concentrations between inlet (IN) and outlet| (OUT) of the struvite process

 PO43− IN (mgP/L)PO43− OUTNH4+ IN (mgN/L)NH4+ OUTMg2+ IN (mg/L)Mg2+ OUTCa2+ IN (mg/L)Ca2+ OUT
Average concentration 179 29 1036 1018 52 43 68 
Standard deviation 14 16 68 31 35 10 13 
Number of daily samples 10 10 10 10 10 10 
 PO43− IN (mgP/L)PO43− OUTNH4+ IN (mgN/L)NH4+ OUTMg2+ IN (mg/L)Mg2+ OUTCa2+ IN (mg/L)Ca2+ OUT
Average concentration 179 29 1036 1018 52 43 68 
Standard deviation 14 16 68 31 35 10 13 
Number of daily samples 10 10 10 10 10 10 

Struvite purity

Table 4 summarises the analytical results from the dried struvite sample. By analysing the elements N, H, Mg and P the ratio of Mg:NH4:PO4 could be calculated and compared with the theoretical ratio of struvite. It is clear that the produced struvite has a high purity, as the measured values approach the theoretical values. Moreover, the presence of possible contaminants is low, as reflected by the total carbon and heavy metal content. A total carbon content of 0.3% C indicates that during the process of struvite formation almost no pollution occurred by organic matter. The ICP-AES results showed that the struvite is not contaminated by heavy metals, as the values for all these elements are very low. Whether the heavy metals content in the struvite of Leuven are low enough for a safe agricultural application should be evaluated by the relevant authorities after further analyses on a longer period. Also, the EDXRF measurements did not show any other contaminants present, and moreover, provided information on the elemental content of Al, Ca, Fe, Na and Si, indicating the presence of sand particles.

Table 4

Results of struvite sample of SFR – comparison with theoretical values (all values in %) in legal limits for soil applications

ComponentTheoretical valuePercentage in sample
5.7 5.1 
Mg 9.9 9.1 
12.6 12.2 
O from PO4  25.2 
O from H2 36.4 
 6.0 
 0.3 
Ca  0.2 
As  0.000125 
Cd  0.0000312 
Cr  0.00013 
Cu  0.0011 
Pb  0.00039 
Ni  0.00012 
Zn  0.0022 
Si, Al, Fe, Na  
Total   96.5 
ComponentTheoretical valuePercentage in sample
5.7 5.1 
Mg 9.9 9.1 
12.6 12.2 
O from PO4  25.2 
O from H2 36.4 
 6.0 
 0.3 
Ca  0.2 
As  0.000125 
Cd  0.0000312 
Cr  0.00013 
Cu  0.0011 
Pb  0.00039 
Ni  0.00012 
Zn  0.0022 
Si, Al, Fe, Na  
Total   96.5 

An XRD spectrum of the struvite was produced. The detected peaks correspond to the mineral struvite. Only a small fraction of silicium oxide (quartz, green line) is noticed, which confirms the detected Si content measured by EDXRF. No other products with a crystalline structure are formed during the process of struvite formation.

Finally, the thermo-gravimetrical curve of the struvite sample shows that from a temperature of 40 °C the struvite starts to dehydrate and degrade. Throughout the heating process ammonia (NH3) and water (H2O) are released. An overall mass loss of 54% is registered at the end. This observation corresponds with the TGA curve of a pure magnesium ammonium phosphate product. In summary, all analyses confirmed that the struvite obtained was quite pure, and calcium is present only in trace amounts. Fertiliser quality tests on plants with the struvite produced in Leuven were not conducted given that several other studies have already showed the good bioavailability of the phosphorus leaching from struvite (Rahman et al. 2014).

Future challenges of struvite recovery from digested sludge

In-depth quality screening

Further investigations to check the innocuousness of this product could also focus on micropollutants (pesticides, hormones, pharmaceuticals, etc.) and microbiological pathogens (e.g. Salmonella).

Improve the separation of the crystals from the sludge

The greatest remaining challenge of struvite recovery from digested sludge remains the separation of the crystals from the sludge. When the process is applied to the water phase, the separation is easy, and a simple settlement allows recovery of most of the formed struvite (Moerman et al. 2009). By contrast, on a digested sludge fraction, the high viscosity of the media and the small size of the formed crystals did not allow them to be recovered by mere settlement in Leuven. Strategies are currently under development to increase the crystal size and therefore improve the recovery rate and make the washing step easier. To date, the harvester managed to recover only about 25% of the formed struvite, bringing down the maximum recovery potential of 15 to 4% of effective phosphorus recovery from the influent of the plant. As for every crystal formation, struvite formation in digested sludge consists of two steps: (1) nucleation and (2) crystal growth. It is the relative speed of these two steps that will determine the number and the size of these crystals. It is still unclear why the obtained crystals are smaller when the process is applied to digested sludge compared to a water phase. It might be linked to the local turbulences that are lower in digestate than in centrate, as the mixing speed has been observed to influence significantly the crystal growth (Wang et al. 2006; Ariyanto et al. 2014). Further research could be undertaken to have an in-depth theoretical understanding of the conditions that have to be achieved with the control parameters as well as the optimum design in order to obtain crystals with a maximum size. Studies have already been carried out in that sense on water fractions and synthetic solutions (Hanhoun et al. 2013; Capdevielle et al. 2014; Rahaman et al. 2014; Galbraith & Schneider 2014; Ariyanto et al. 2014), but the challenge for a usable nucleation-growth model on real digested sludge phases remains.

CONCLUSION

The full-scale struvite process achieved, during the first year of operation, an efficiency of 80% of orthophosphate removal in the digested sludge between inlet and outlet of the struvite reactor, with a Mg/P dosing ratio of around 1.75 and a pH of 7.5 in the reactor. Such an implementation of the process allows a maximum recovery potential of 15% of the total phosphorus load of the plant of Leuven. The major obstacle to implementation on the sludge remains the difficulty of separating the crystals from the sludge. To date, the harvester could effectively recover 25% of the formed struvite in Leuven, being equivalent to an actual recovery of 4% of the phosphorus from the influent of the plant. The process divides the load of phosphorus in the centrate back to the water line by 2, which normally accounts for 15 to 20% of the P-load of the water line. The extraction of this P load from the centrate is suspected to support the biological P removal on the water line but more in-depth demonstration is necessary there. There is a diminution of the scaling speed downstream the process but beneficial impact on pumps and dewatering devices should be evaluated over a longer term. It is reckoned that a dewaterability improvement of +1.5% DS is achievable provided a sufficient magnesium dosing and the implementation of a powerful stripping to minimise the resort to caustic solution. The dewatering improvement is attributed mainly to the ‘bridging’ effect of the magnesium divalent cations whose concentration gets multiplied by 10 by the struvite process due to an overdosing of around 1.75 molar ratio Mg:P.

The obtained struvite is quite pure. Whether it is safe to use in food or biofuel agriculture should still be evaluated by the competent authorities with further analyses over a longer period. Remaining technical challenges of struvite recovery from sludge include the separation efficiency of the formed struvite crystals from the sludge.

To conclude, such a full-scale experience should contribute to determine the best way to recover phosphorus from municipal wastewater: on centrate, or on digestate or at the end of sludge life cycle from incinerated sludge ashes or by land spreading. This latter point is currently under discussion in the wastewater sector and should be discussed on the basis of technical performance evaluations, cost–benefit analysis, and comparative life cycle analyses with other non-recycling phosphorus production means.

REFERENCES

REFERENCES
Bergmans
B. J. C.
Veltman
A. M.
van Loosdrecht
M. C. M.
van Lier
J. B.
Rietveld
L. C.
2014
Struvite formation for enhanced dewaterability of digested wastewater sludge
.
Environmental Technology
35
(
5
),
549
555
.
Borgerding
J.
1972
Phosphate deposits in digestion systems
.
Journal of Water Pollution Control Federation
44
,
813
.
Capdevielle
A.
Sykorova
E.
Beline
F.
Daumer
M.-L.
2014
Kinetics of struvite precipitation in synthetic biologically treated swine wastewaters
.
Environmental Technology
35
(
10
),
1250
1262
.
Clift
R.
Shaw
H.
2012
An industrial ecology approach to the use of phosphorus
.
Procedia Engineering
26
,
39
44
.
Cordell
D.
Drangert
J.-O.
White
S.
2009
The story of phosphorus: global food security and food for thought
.
Global Environmental Change
19
(
2
),
292
305
.
Cornel
P.
Schaum
C.
2009
Phosphorus recovery from wastewaters: needs, technologies and costs
.
Water Science Technology
59
(
6
),
1069
1076
.
ESPP
2013
European Sustainable Phosphorus Platform website: http://www.phosphorusplatform.eu/
.
European Commission
2013a
COM (2013) 517 Final. Consultative Communication on the Sustainable Use of Phosphorus.
European Commission
2013b
Science for Environment Policy
.
In-Depth Report: Sustainable Phosphorus Use, Issue 7
.
Fattah
K. P.
Sabrina
N.
Mavinic
D. S.
Koch
F. A.
2008
Reducing operating costs for struvite formation with a carbon dioxide stripper
.
Water Science and Technology
58
(
4
),
957
962
.
Geerts
S.
Marchi
A.
Weemaes
M.
2015
Full-scale phosphorus recovery from digested waste water sludge in Belgium – part II: economic opportunities and risks
.
Water Sci. Technol. – Special Issue on Full Scale Investigations
71
(
4
),
495
502
.
Hanhoun
M.
Montastruc
L.
Azzaro-Pantel
C.
Biscans
B.
Freche
M.
Pibouleau
L.
2013
Simultaneous determination of nucleation and crystal growth kinetics of struvite using a thermodynamic modelling approach
.
Chemical Engineering Journal
215
,
903
912
.
Husillos Rodriguez
N.
Martinez-Ramirez
S.
Teresa Blanco-Varela
M.
Donatello
S.
Guillem
M.
Puig
J.
Fos
C.
Larrotcha
E.
Flores
J.
2013
The effect of using thermally dried sewage sludge as an alternative fuel on Portland cement clinker production
.
Journal of Cleaner Production
52
,
94
102
.
Kruk
D. J.
Elektorowicz
M.
Oleszkiewicz
J. A.
2014
Struvite precipitation and phosphorus removal using magnesium sacrificial anode
.
Chemosphere
101
,
28
33
.
Lee
S.-H.
Yoo
B.-H.
Kim
S.-K.
Lim
S. J.
Kim
J. Y.
Kim
T.-H.
2013
Enhancement of struvite purity by re-dissolution of calcium ions in synthetic wastewaters
.
Journal of Hazardous Materials
261
,
29
37
.
Linderholm
K.
Tillman
A.-M.
Mattsson
J. E.
2012
Life cycle assessment of phosphorus alternatives for Swedish agriculture
.
Resources, Conservation and Recycling
66
,
27
39
.
Moerman
W.
Carballa
M.
Vandekerckhove
A.
Derycke
D.
Verstraete
W.
2009
Phosphate removal in agro-industry: pilot- and full-scale operational considerations of struvite crystallization
.
Water Research
43
(
7
),
1887
1892
.
Neethling
J.
Benisch
M.
2004
Struvite control through process and facility design as well as operation strategy
.
Water Science Technology
49
(
2
),
191
199
.
Rahman
M. M.
Salleh
M. A. M.
Rashid
U.
Ahsan
A.
Hossain
M. M.
Ra
C. S.
2014
Production of slow release crystal fertilizer from wastewaters through struvite crystallization – a review
.
Arabian Journal of Chemistry
7
,
139
155
.
STOWA
2012
Trends in Slibontwatering
.
Report 46
.
Wang
J.
Burken
J.
Zhang
X.
2006
Effect of seeding materials and mixing strength on struvite precipitation
.
Water Environment Research
78
(
2
)
125
132
.
Weigand
H.
Bertau
M.
Huebner
W.
Bohndick
F.
Bruckert
A.
2013
Recophos: full-scale fertilizer production from sewage sludge ash
.
Waste Management
33
,
540
544
.