The BiOPhree technology is based on phosphorus adsorption using industrially produced iron oxide granules with a large specific surface area. The effluent from a WWTP containing around 1 mg P/l (mostly dissolved P) is passed over an adsorption column containing these granules. The column becomes saturated with phosphorus over time. The phosphorus can be recovered, and the adsorbent reused, by regenerating the column with caustic soda. The regenerant (caustic soda together with the recovered dissolved phosphorus) is treated with a nanofiltration membrane, also to recover as much caustic soda as possible for subsequent regenerations.
The pilot project addressed two main questions:
Experimental design
The pilot project consisted of a prefiltration step for removing suspended solids from effluent (Sand filter) followed by three parallel adsorption columns. The system also featured a caustic soda tank for regeneration (Figure 1). The pilot project had a capacity of 3 m3/hour, achieved with a maximum of two columns operating simultaneously and one column on standby.
Samples were collected from the system every day to determine the effectiveness of phosphorus removal. Regenerations were performed after saturation of the adsorption column.
A batch of the regenerant collected from the adsorption column regenerations was tested to establish whether the caustic soda could be recovered using nanofiltration.
Lab tests were carried out to gain a better understanding of the regeneration strategy and phosphorus precipitation from the regenerant. After the first two regenerations it turned out that only a limited amount of P was desorbed and too much P adhered to the iron granules. The pilot system was therefore adapted to allow regeneration with more caustic soda. The pilot project ran with the improved regeneration strategy from May until December 2024. The remainder of this article relates to this improved strategy.
Figure 1. Process diagram: BiOPhree system for P removal from WWTP effluent
Results
Phosphorus removal and regeneration
The average P level in the effluent inflow was 0.9 mg P/l, but the levels varied from 0.5-2 mg/l with outliers up to 3.5 mg/l and even 8 mg/l in one instance (not visible in the graph). Despite these strong fluctuations in influent concentrations, high levels of P were removed (Figure 2) and the peaks were virtually shaved off. During the period shown, the adsorbent was regenerated six times. The P levels consistently dropped to around 0.1 mg P/l or even lower.
During the test period, the adsorbent became saturated faster and more frequent regeneration was needed. The reasons for this were a suboptimal and an increasing P load to the adsorption column over time. After two regenerations, the amount of caustic soda was therefore increased from 3 to 5 bed volumes per regeneration.
Timely regeneration was found to be an important condition to keep the P concentration in the effluent of the column consistently <0.1 mg P/l since phosphorus concentrations in the effluent of the column increase rapidly after saturation of the adsorbent.
Figure 2. BiOPhree input and output phosphorus levels, pilot project from May until December 2024 (vertical dotted lines: the six regeneration moments; solid lines: seven-day moving average.
On average, around 4 g of P was removed per kg of adsorbent before breakthrough of the column was observed. Regeneration of the adsorption column with sodium hydroxide resulted in an average P-desoprtion of 82%. As the pilot project progressed and operations improved, the desorption efficiency rate increased to 88%.
Nanofiltration of the regenerant
Nanofiltration of the regenerant shows that a flux of 25–30 l/m2 per hour is possible at a pressure of 38 bar. Of the phosphorus, 99% remained in the concentrate. Of the caustic soda, 98% entered the permeate, allowing the majority to be recovered and reused. The total NaOH-recovery that could be achieved by nanaofiltration of the regenerant is 70-90%.
In addition to phosphorus, organic matter (expressed as chemical oxygen demand, or COD) is also adsorbed to the column, and desorbed upon regeneration. This was also effectively removed from the regenerant (88% COD removal on average). The COD consists mainly of humic acids, which could potentially also be recovered as a useful raw material.
Lab tests showed 70–90% P precipitation with calcium from the regenerant. The experiments were performed with beakers in which the regenerant and concentrate from nanofiltration were mixed with CaCl2 with different Ca/P molar ratios. At a ratio of 2 mol Ca/mol P, 70% P was precipitated from the concentrate and 90% from the regenerant. It is not clear why a better result was achieved with regenerant, but the precipitation of phosphorus with calcium seems to take place fast and is expected not to be very complex.
Nanofiltration of the regenerant and phosphorus recovery using calcium precipitation therefore appears to be a suitable method for recovering chemicals and raw materials in order to ensure a sustainable process.
Practical application
The BiOPhree process (Figure 1) can be scaled up to a full scale system. The design is based on a retention time of only five minutes allowing the columns to remain compact which is significantly reducing the investment costs. Operating several columns in parallel means that one column can always be regenerated while the others remain in operation. Regeneration is done with diluted caustic soda (1 mol/l), which is recovered with a nanofiltration membrane. The removed phosphorus ends up in the concentrate and can be recovered locally by precipitation with calcium, or centrally by processing it in an existing struvite system at a large WWTP in the region.
Although the iron granules will be regenerated multiple times, they do not last indefinitely. It is expected that around 10–20% of the granules need to be replaced every year.
Cost and sustainability
The pilot project results have been extrapolated for a business case on a practical scale (100,000 P.E.). Two scenarios were assessed:
For scenario A, a BiOPhree system with a capacity of 750 m3/hour is sufficient; this is 1.2x the dry weather flow (DWF). In this case 70% of the yearly WWTP effluent will be treated, removing phosphorus to 0.1 mg P/l. The remaining 30% of the effluent is not treated (during rainy weather) but bypassing the BiOPhree system, resulting in an annual average of 0.3 mg P/l. For scenario B, 100% of the annual flow must be treated, also during rain weather events. This will require a much larger system, with a capacity of 3,750 m3/hour (6x DWF rate).
The cost ranges from €0.18/m3 (scenario A) to €0.53/m3 (scenario B).
The costs of scenario A were calculated using the standardised cost estimate system (Standaard Systematiek Kostenramingen, SSK) and compared with the costs of reference techniques, such as cloth and sand filtration with coagulant dosing. The costs of these techniques per volume treated is in the same range: 0.17–0.22 euros/m3 [1]. However, these techniques are unable to produce effluent with phosphorus concentrations below 0.25 mg P/l [1].
This means that there is no known reference technology for scenario B that has been applied on a practical scale, so no cost comparison could be made.
The carbon emissions (derived from key figures from the EcoInvent database) of BiOPhree (scenario A) are around 19 g CO2/m3, similar to cloth and sand filtration: 18–22 g CO2/m3 [1].
Conclusions
The following conclusions can be drawn from the research described here:
However, BiOPhree achieves a higher removal rate and is more efficient in terms of chemical consumption and sludge production. At DWF, BiOPhree achieves 0.1 mg P/l, whereas reference techniques max out at 0.25 mg P/l. Since WWTPs are subject to DWF 75% of the time, this alone is a major benefit for the receiving surface water – particularly in summer, when many water bodies are fed largely by WWTP effluent.
If BiOPhree is designed for a larger flow rate (>1.2x DWF), the annual average P level can drop even further, from 0.3 mg P/l towards 0.1 mg P/l.
Next steps
BiOPhree technology is still being optimised, which is expected to reduce costs further. Regardless, it is also important to optimise existing processes in WWTPs. The lower the P levels in the supplied effluent, the fewer regenerations needed, and the smaller the proportion of effluent that needs to be treated to meet the discharge standard.
The next step will be scaling up to a demonstration system on a practical scale. Conversations with interested water authorities are currently underway.
This study was funded by the Foundation for Applied Water Research STOWA, the Zuiderzeeland, Brabantse Delta and Rijn and IJssel water authorities, Waternet, Aquacare and Haskoning.
Thanks to EMI-Twente for its support with the membrane tests.
BiOPhree technology for extensive phosphorus removal from WWTP effluent has been tested as part of a pilot project at Dronten WWTP. The pilot project shows that phosphorus can be removed down to <0.1 mg P/l, provided the adsorption material (iron oxide granules) is regenerated with caustic soda on time. The caustic soda used can be reused after nanofiltration. In the pilot project, the granules were regenerated six times, with the P level subsequently falling again each time to <0.1 mg P/l. This makes BiOPhree an interesting post-treatment technique for WWTPs that must meet stringent phosphorus requirements. BiOPhree’s carbon emissions are comparable to reference techniques such as sand and cloth filtration, but these techniques max out at 0.25 mg P/l. Depending on the degree of P removal required, the costs associated with BiOPhree are the same as or lower than those of reference techniques. An additional advantage of BiOPhree is that most of the phosphorus removed can be recovered.