Water Research 212 (2022) 118102 Contents lists available at ScienceDirect Water Research journal homepage: www.elsevier.com/locate/watres Review A review on the integration of mainstream P-recovery strategies with enhanced biological phosphorus removal Congcong Zhang, Albert Guisasola *, Juan Antonio Baeza GENOCOV. Departament d’Enginyeria Química, Biològica i Ambiental. Escola d’Enginyeria. Universitat Autònoma de Barcelona, Bellaterra (Barcelona) 08193, Spain A R T I C L E I N F O A B S T R A C T Keywords: Phosphorus (P), an essential nutrient for all organisms, urgently needs to be recovered due to the increasing Configurations demand and scarcity of this natural resource. Recovering P from wastewater is a feasible and promising way Enhanced biological phosphorus removal widely studied nowadays due to the need to remove P in wastewater treatment plants (WWTPs). When enhanced (EBPR) biological P removal (EBPR) is implemented, an innovative option is to recover P from the supernatant streams Mainstream obtained in the mainstream water line, and then combine it with liquor-crystallisation recovery processes, being Phosphorus recovery Precipitation the final recovered product struvite, vivianite or hydroxyapatite. The basic idea of these mainstream P-recovery strategies is to take advantage of the ability of polyphosphate accumulating organisms (PAO) to increase P concentration under anaerobic conditions when some carbon source is available. This work shows the main­ stream P-recovery technologies reported so far, both in continuous and sequenced batch reactors (SBR) based configurations. The amount of extraction, as a key parameter to balance the recovery efficiency and the main­ tenance of the EBPR of the system, should be the first design criterion. The maximum value of P-recovery ef­ ficiency for long-term operation with an adequate extraction ratio would be around 60%. Other relevant factors (e.g. COD/P ratio of the influent, need for an additional carbon source) and operational parameters (e.g. aera­ tion, SRT, HRT) are also reported and discussed. 1. Introduction rock (e.g. cadmium or uranium), which can be highly toxic (Cornel and Schaum, 2009), or to avoid the need for external dependence on a The new paradigm in wastewater treatment proclaims that waste­ strategic resource such as phosphate rock. In addition, P in urban water treatment plants (WWTPs) should evolve to water resource re­ wastewaters has to be removed anyway to avoid eutrophication pro­ covery facilities (WRRFs), where resources and energy are recovered cesses, making P-recovery a win-win scenario. from water. Amongst all the potential resources to be recovered, P arises WWTPs seem to be an adequate place to recover P, since 1.3 Mt of P as a perfect candidate since: i) in the current context of population and are removed annually worldwide through wastewater treatment. Some food demand growth, P is essential for fertilizers production (Chowd­ estimates indicate that global demand for P could be covered by 15–20% hury et al., 2017; Cieślik and Konieczka, 2017; Roy, 2017), ii) P is with this recovery (Li and Li, 2017; Wu et al., 2019; Yuan et al., 2012). currently a pollutant of our wastewater which may promote eutrophi­ Other studies indicate that in Central Europe, P-recovery from municipal cation in water bodies if it is not removed and iii) P is mainly obtained wastewater would satisfy 40–50% of the mineral P fertilizer applied from non-renewable sources such as phosphate rocks (Van Vuuren et al., annually in agriculture (Egle et al., 2016). 2010), which are expected to be depleted in only 50–100 years (Cordell Furthermore, from a life cycle assessment (LCA) point of view, it has et al., 2009; Desmidt et al., 2015; Rittmann et al., 2011). For the been concluded that P-recovery from municipal wastewater sludge, li­ full-scale application of P-recovery processes, one of the main current quor or ash of sludge mono-incineration would provide environmental limitations is that the market price (Table 1) of rock phosphate is lower benefits (Remy and Jossa, 2015). Although differences appear in the than recovered products such as struvite. However, if the additional amount of P recovered, requirements of energy, chemicals and fuels, or benefits of P-recovery are accounted for, significant incentives to sup­ in side effects in the process, P-recovery processes can reduce environ­ port it emerge: for example, the need to treat the impurities in phosphate mental impacts, mainly on global warming potential, freshwater * Corresponding author. E-mail address: albert.guisasola@uab.cat (A. Guisasola). https://doi.org/10.1016/j.watres.2022.118102 Received 20 September 2021; Received in revised form 5 January 2022; Accepted 17 January 2022 Available online 18 January 2022 0043-1354/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). C. Zhang et al. Water Research 212 (2022) 118102 Table 1 anaerobic digestion and P-recovery is not implemented (D.Wild et al., The market values of P rock and P-recovery products. 1996; Doyle and Parsons, 2002). The repair of this clogging is costly due P related species Market price Reference to the need to shut down and clean the clogged equipment. In a modelling study, Lizarralde et al. (2019) constructed a plant-wide model P rock ~120 $/ton (Mew, 2016) Struvite ~500 €/ton (Wu et al., 2019) to understand the effect of several operational parameters on the un­ Vivianite ~10,000 €/ton (Wu et al., 2019) controlled struvite precipitation when EBPR was implemented. They Hydroxyapatite ~10,000 $/ton (Annisa et al., 2021) concluded that struvite recovery has many benefits for the plant not only for its commercial value, but also for the reduction of sludge production, ferric chloride dosage and the prevention of uncontrolled struvite eutrophication of and fossil energy demand (Remy and Jossa, 2015). precipitation. P-recovery as struvite seems to be especially attractive for urban areas, to satisfy the growing trend of local crop production linked to the large 1.2. Most studied technologies for P-recovery in WWTPs volumes of wastewater treated in centralized WWTPs (Rufí-Salís et al., 2020). In this recent LCA study, three recovery technologies were ana­ Different technologies have been proposed for P-recovery from flows lysed for large WWTPs, showing that were able to recover 5–30 times the containing phosphorus in WWTP (i.e. effluent, sewage sludge, sludge amount of P required to fertilize the agricultural area of the studied ash, dewatering liquor and digester supernatant), providing different region annually (36.5 t). Although there was an increase of chemicals operability, costs and potential uses of the recovered material (Cieślik and energy requirements per m3 of wastewater due to struvite recovery, and Konieczka, 2017; Desmidt et al., 2015; Egle et al., 2016; Law and it resulted in lower eutrophication potential. Pagilla, 2018). Egle et al. (2016) indicated that an ideal technology should provide good P-removal from wastewater, maximize P-recovery 1.1. Fate of P in WWTPs rate, destroy hazardous substances and produce a material that could be applied with low environmental risks, good fertilizing effects and high The fate of the influent P in a WWTP depends on its configuration. economic efficiency. They also studied all these characteristics from 19 Part of the influent P is removed with the primary sludge (around 11%) different P-recovery technologies and concluded that the choice of a according to Cornel and Schaum (2009). If the plant does not implement specific recovery technology is a compromise amongst these factors. enhanced biological phosphorus removal (EBPR) in the secondary step, Furthermore, the LCA analysis by Amann et al. (2018) showed wide approximately 15–30% of the P in the influent becomes part of the range of changes in gaseous emissions, energy demand and P-recovery biomass (a conventional WWTP sludge contains around 1–2% of P ac­ potential of different technologies based on liquid phase, sewage sludge cording to mass basis) and removed with the purged sludge (Henze and ash, and concluded that LCA should be combined with additional et al., 2008; Parsons and Smith, 2008). Then, if complete P-removal is environmental criteria to put the performance of the technologies into mandatory, the remaining P should be chemically removed with addi­ perspective. tion of Fe(III) and Al(III) salts in different dosing points such as the Considering the wide range of existing P-recovery processes, this primary sedimentation, before and/or following biological treatment or section presents the most studied ones classified in two groups, in a tertiary treatment (Tchobanoglous et al., 2014). However, the depending on whether they can be obtained from primary/secondary chemical sludge produced is not recommended as fertiliser since these effluents or from the sludge line, and includes another subsection salts could reduce the mobility of P in soils where this chemical sludge is showing the crystallisation processes that could be applied to P-enriched applied (Desmidt et al., 2015; Sartorius et al., 2012). The sludge streams to make P-recovery effective. Section 2 then presents the P-re­ generated is often accumulated in landfills or burnt in incinerators, and covery processes related to mainstream strategies that are the focus of may end up contaminating water bodies (such as aquifers, rivers or sea) this review. (Mainstone et al., 2000). Alternatively, the use of ashes from sewage sludge incinerators is proposed for P-recovery (Ma and Rosen, 2021) and 1.2.1. P-recovery from primary or secondary effluents to produce fertilizers via a chemical extraction (Fang et al., 2021; J. Li P-recovery from primary or secondary effluents using chemical et al., 2018). precipitation does not seem to be a cost-effective alternative because it The fate of P in WWTPs that include EBPR is very different. EBPR is would require a large amount of chemicals (Cornel and Schaum, 2009; based on the proliferation of polyphosphate accumulating organisms Desmidt et al., 2015; Egle et al., 2016). Several alternative methodolo­ (PAO), which can accumulate phosphate internally as polyphosphate gies have been described previously (Egle et al., 2015; Kabdaşlı and (poly-P) under aerobic (or anoxic) conditions after being exposed to Tünay, 2018). For instance, Liberti et al. (1979) reported a methodology anaerobic conditions with available carbon source. Then, P is removed based on the combination of ion exchange (IX) for phosphate selective as poly-P inside cell biomass by waste sludge (Tchobanoglous et al., exchange and P-recovery from the concentrated stream by struvite 2014). PAO can incorporate up to 0.38 mg P/mg VSS compared to the precipitation, which was further evaluated in subsequent works (Liberti 0.02 mg P/mg VSS incorporated by the ordinary heterotrophic organ­ et al., 2001; Petruzzelli et al., 2003). Williams et al. (2015) studied P and isms (OHO) (Gebremariam et al., 2011; Henze et al., 2008; Yuan et al., N removal with media based on Fe, Cu, and Al and then precipitation 2012). EBPR-based WWTPs can typically remove more than 85% of the columns with clinoptilolite IX. They achieved almost 84% of P-recovery P from the influent that is stored within the biomass (Bunce et al., 2018; and a regeneration eluate with concentration up to 560 mg P⋅L− 1. Gebremariam et al., 2011). The reported range of the P content in Beaudry and Sengupta (2021) used a novel pyridine-based polymeric EBPR-based sludge is around 0.06–0.15 mg P/mg VSS (Henze et al., ion exchanger and three different fixed-bed columns, producing an 2008). However, the use of this sewage sludge in agriculture is nowa­ effluent concentration with P < 6 μg/L and recovering about 90% of the days limited in many areas because of the ecological and health threats phosphorus. involved. For example, it is estimated that only about 20% of the sewage The hybrid anion exchanger (HAIX), which consists of a hybrid sludge meets the EU regulations for fertilizers in terms of concentration polymeric base dispersed with hydrated ferric oxide (HFO) nano­ of heavy metals or radioactive elements (e.g. radon and uranium) (Eu­ particles, has also been reported for P removal (Martin et al., 2018, ropean Commission, 2019; Pettersson et al., 2008; Weigand et al., 2009; Sendrowski and Boyer, 2013). Phosphate uptake is based on the 2013). selective sorption sites with HFO nanoparticles, which form sphere There is another reason to implement P-recovery strategies. In EBPR- complexes with Fe(III) and phosphate (Blaney et al., 2007). After based WWTPs, the clogging of tubes and pumps due to undesired stru­ long-term sorption, phosphate breakthrough occurs, but the exhausted vite precipitation is observed when the bio-P sludge undergoes HAIX can be reused with a regeneration phase with NaOH/NaCl solution 2 C. Zhang et al. Water Research 212 (2022) 118102 and, finally, a rinse with CO2 sparged filtered water (Blaney et al., 2007). Table 2 A recent work by Guida et al. (2021) evaluated a large (10 m3/d) HAIX Stoichiometry and solubility for the most usual P-recovery chemical compounds. system for 2.5 years using treated municipal wastewater, showing high P Compound Stoichiometry pKsp Reference removal (95%, i.e. from 6 to <0.3 mg PO43− -P/L) and recovery effi­ 2+ 3− struvite Mg + NH4+ + PO4 + 6 H2O → 13.26 (Ohlinger et al., ciency (95% of phosphate from the wastewater recovered as hydroxy­ MgNH4PO4⋅6H2O 1998) apatite). In subsequent work, the regeneration of the media and the K-struvite 2+ Mg + K + PO43− + 6 H2O → + 11.7 (Luff and Reed, modelling of this system were also studied (Pinelli et al., 2022). How­ MgKPO4⋅6H2O 1980) ever, the feasibility of this methodology at full-scale has not been proved vivianite 3 Fe2+ + 2 PO43− + 8 H2O → 29.03 (Chen and Fe3(PO4)2⋅8H2O Faust, 1974) yet (Egle et al., 2015; Rufí-Salís et al., 2020; Siciliano et al., 2020). hydroxyapatite 5 Ca2++ 3 PO43− +OH− → 58.5 (Mcdowell et al., Apart from the above-mentioned approach, microalgae cultivation Ca5(PO4)3OH 1977) has been proposed as a promising technology for P-recovery from sec­ ondary effluents due to its low cost, eco-friendliness and low technology dependence (Yang et al., 2017). P is taken up biologically by algal and high pH (around 9–10) are needed so that struvite crystallisation growth by assimilation, accompanied by oxygen generation that sup­ becomes thermodynamically favourable. Struvite precipitation in ports aerobic bacteria growth. Algae can also be used either as a feed­ streams with low concentration of phosphate or ammonium or low pH stock for biofuel production or directly as a biofertilizer (Chisti, 2013, (as the influent or effluent of WWTPs) becomes an unaffordable option 2008). El Hamouri (2009) showed that 63% of influent P could be because of the requirement of adding high amount of magnesium (e.g. removed and incorporated into algae as a tertiary treatment using a see Fig. 1 at the condition of pH = 7). high-rate algal pond. Xu et al. (2016, 2015) also showed that algae could In any case, there are successful reports of P-recovery as struvite remove about 60% of influent phosphorus for secondary wastewater under full-scale conditions using side-streams obtained from the sludge effluent. However, the application of algae for P-recovery is still limited line (Desmidt et al., 2015; Jabr et al., 2019; Law and Pagilla, 2018; Le in large-scale wastewater treatment (Gebremariam et al., 2011). Corre et al., 2009; Mayer et al., 2016) and successfully marketed tech­ nologies for the recovery of struvite include (amongst others) Ostara®, 1.2.2. P-recovery from the sludge line Airprex™, Multiform™, DHV Crystalactor® and NuReSys™ (Law and As mentioned above, the maximum P-recovery potential is from Pagilla, 2018). However, struvite crystallization also has some draw­ EBPR-based sludge. However, this option is not straightforward. For backs in its practical application, for example, an unsatisfactory recov­ instance, the agricultural application of stabilized sludge with high P ery efficiency related to the total P load in the raw sludge, in the range content (Hester and Harrison, 2013) has to consider the strict national 45–80% in some full-scale projects for P-recovery from digested sludge health and safety restrictions on nutrients and heavy metals. This hin­ or liquor, as reported by Remy and Jossa (2015). ders the agricultural application of sludge and, thus, this pathway of Vivianite (Fe3(PO4)2⋅8H2O) is an alternative valuable product for P- P-recovery (Law and Pagilla, 2018). Another option is the incinerated recovery. It has a high market value (Table 1) as it is a relatively slow- sludge ashes, which may contain 4 to 11% of P (Biswas et al., 2009; release fertilizer widely found in soils, lake sediments, hydrothermal Donatello et al., 2010; Franz, 2008; Ottosen et al., 2013). However, as a deposits or bogs (Wu et al., 2019). Moreover, the precipitating agent, result of the organic matter oxidation, the heavy metal content in these iron, is a common dissolved mineral in natural water and Fe(III) salts are ashes may be even higher than the sludge itself. Therefore, a chemical commonly used for improving sludge settleability (Li et al., 2014) and (Franz, 2008; Ottosen et al., 2013) or thermochemical (Adam et al., odour control (Zhang et al., 2009). The reaction to precipitate vivianite 2007) process is needed to separate the P from the heavy metals of the and its solubility is presented in Table 2. ashes. Wet chemical treatment by acid, alkaline or both have been The theoretical molar ratio of Fe:P to form vivianite is 1.5:1, but the proposed for the dissolution of P (Law and Pagilla, 2018); acid leaching optimal stoichiometric ratio range applied is 1.5–2 considering some of (sulphuric or nitric acid) is proposed for Fe-rich ashes (Ottosen et al., Fe(II) could be transformed to Fe(OH)2 or oxidized to Fe(III) (Priam­ 2013), while the combination of both acid and alkaline treatment is bodo et al., 2017). The formation condition of vivianite is less dependant proposed for Al-rich ashes (Petzet et al., 2012). on pH (i.e. adequate range is 6–9) compared to struvite (8–9.5). Viv­ The most reported location to recover P is a liquid stream from the ianite can be precipitated from the digested sewage sludge (Prot et al., sludge line, either from the sludge treatment or the anaerobic digestion 2020; Wilfert et al., 2018), wastewater (Wu et al., 2019) and aqueous (Egle et al., 2015). The treatment of the purged secondary sludge re­ solutions (Liu et al., 2018). Some researchers have shown that about leases soluble phosphate and ammonium, resulting in a supernatant 62% of P could be precipitated to vivianite from an anaerobic digestate with increased concentrations of these compounds. These streams can with a pH around 8 (Li, Li and Li, 2018; Priambodo et al., 2017). contain up to 30% of the influent P loads. Levlin and Hultman (2003) However, the P-recovery process as vivianite also has disadvantages: i) reviewed several side-stream based P-recovery strategies and concluded the formation of vivianite in the sludge produces small crystals or ag­ that P-recovery efficiency about 60–65% is achieved from these streams. gregates (10–150 μm) resulting in a complex separation from sludge and ii) vivianite usually contains calcium or magnesium impurities (Robles 1.2.3. Crystallization processes for P-recovery et al., 2020). Impure vivianite is less stable, since it can be oxidized Crystallization is nowadays the most accepted strategy for recov­ within 48 h, in contrast to the several weeks of stability reported for pure ering P from these enriched liquors, being struvite the preferred pre­ vivianite (Wilfert et al., 2018). In addition, separation and purification cipitation product (Le Corre et al., 2009; Li et al., 2019; Rittmann et al., methods for vivianite still need further development (Robles et al., 2011). Struvite (MgNH4PO4⋅6H2O) can be applied directly to the field as 2020). a slow release fertiliser, also providing nitrogen (N) and magnesium Another alternative to struvite is hydroxyapatite (HAP, (Mg) to the soil, and has been reported to be a good option for agri­ Ca5(PO4)3OH). HAP can be directly used as a fertilizer in the field of cultural uses (Hu et al., 2016). Sometimes, ammonium in struvite can be agriculture (Johansson et al., 2017; Mayer et al., 2016) and as an replaced by potassium and then the precipitate is known as K-struvite adsorbent for removing dyes, emerging pollutants and heavy metals (Ronteltap et al., 2007). The reactions to precipitate struvite and (Maity et al., 2018; Zhang et al., 2018). It is of great potential for K-struvite and their corresponding solubility are presented in Table 2. P-recovery since calcium phosphate has a large market (Driver et al., Fig. 1 presents a graphical representation of the minimum magne­ 1999) and, thus, it is expected to be of essential importance for the sium (Mg2+) concentration needed to reach the struvite solubility closure of P cycle (Leinweber et al., 2018). The reaction to precipitate product at different pH as a function of phosphate and ammonium HAP and its solubility is presented in Table 2. concentration. As observed, significant concentrations of P, Mg2+, NH4+ HAP is thermodynamically the most stable phase for orthophosphate 3 C. Zhang et al. Water Research 212 (2022) 118102 Fig. 1. Minimum magnesium (Mg2+) concentration to reach the Ksp value for struvite at different pH and as function of phosphate and ammonium concentration. Addition of magnesium beyond this value will drive struvite precipitation. and calcium in solution within a broad range of pH of 4–12 (Cichy et al., be a favourable option for struvite precipitation. However, these systems 2019; Oubagha et al., 2017). When dealing with the common problem of also bring disadvantages such as undesired precipitation in the anaer­ low P concentration in municipal wastewater, HAP is a more attractive obic digester or in the downstream tubing and instrumentation (Doyle option than struvite since struvite demands higher enriched P. For and Parsons, 2002). On the other hand, the high concentration of some example, although this is highly specific depending on the pH and ions in the anaerobic digestion sludge (e.g. calcium) could hinder the species concentration, for a typical wastewater composition (8 precipitation of P as struvite, since when P is released it can precipitate mgPO43− -P/L and 40 mgNH4+-N/L), and based on the HAP and struvite as other compounds such as HAP, brushite and other calcium phos­ solubility, the addition of 10 mgCa2+/L and 10 mgMg2+/L at pH = 8 or phates. In any case, to determine the actual precipitation competition in higher leads to almost complete precipitation of Ca to produce HAP, a given anaerobic digestion scenario, not only solubility has to be taken whereas struvite is not formed even at a higher pH. Thus, some works into account, but also kinetics (D.Wild et al., 1996). have exploited this facility of HAP to precipitate and evaluated P-re­ Apart from the digestate, the other location with the highest P con­ covery strategies in lab-scale systems (Berg et al., 2005; Zou and Wang, centration in an EBPR plant is the anaerobic reactor and, thus, it can be 2016). However, HAP could lose advantage over struvite precipitation an alternative for P-recovery. During the anaerobic phase, PAO have when aiming at a simultaneous recovery of N and P in real industrial taken up most of the readily biodegradable COD and most of the P has wastewater treatment (Cichy et al., 2019). In addition, HAP has lower been released into the bulk liquid, reaching levels from 30 to 70 mg P/L potential for fertilizer use (Robles et al., 2020). (Acevedo et al., 2015; Izadi et al., 2021, 2020; Zhang et al., 2020; Zou and Wang, 2016). Therefore, the anaerobic reactor appears as a good 2. Potential P-recovery from mainstream point to implement mainstream P-recovery whilst the digestate seems to be the ideal point for side-stream P-recovery. The integration of EBPR P-recovery via precipitation (e.g. struvite, K-struvite, HAP, vivianite) and mainstream P-recovery as struvite would be beneficial to prevent requires a high concentration of P for displacing the solubility equilib­ undesired precipitation throughout the plant and to reduce N load when rium to the precipitation side (Fig. 1). As described before, in EBPR- treating wastewater with high ammonium concentration (1 g of P based WWTPs, digestate is a common option to precipitate/recover P removed as struvite yields 7.92 g of struvite containing 0.45 g of N). since the sludge entering anaerobic digestion contains a high amount of Thus, it shows a good potential for improving system performance, and internally stored P as poly-P that can be easily released inside the for this reason it has recently received the attention of researchers (e.g., digester and results in a high concentration of P into the digestate. The Acevedo et al., 2015; Larriba et al., 2020; Shi et al., 2016; Zou and Lu, digestate also contains high ammonia concentration and, thus, seems to 2016). 4 C. Zhang et al. Water Research 212 (2022) 118102 The first proposal on the potential to obtain a P-rich liquor from the PAS mainstream anaerobic reactor was reported by Van Loosdrecht et al. PR = ∗ 100% (3) PINF (1998): the Biological-chemical P and N removal (BCFS®) process (Fig. 2a). BCFS was a modification of the UCT design where a stream where POUT represents the P output per day in the effluent (PEFF) plus in from the anaerobic reactor (for instance, an average flow of 10% of the the purged biomass (PBIO), PAS the P in the anaerobic supernatant (thus influent flow) was sent to a sludge thickener and a phosphate stripper available for P-recovery), PINF the total input of P in the influent to the where FeCl3 was dosed for P precipitation. Thus, a continuous EBPR secondary treatment per day and PR the P-recovery efficiency. PINF is operation and a P chemical precipitation system were combined. highly influenced by the configuration of the primary treatment. Typical Modelling was used to better understand this process and to assess its P removal values around 11% have been reported for conventional performance under different scenarios (Barat and van Loosdrecht, 2006; primary treatments (Cornel and Schaum, 2009), but these values can be Hao and van Loosdrecht, 2006). Hao and van Loosdrecht (2006) simu­ increased considerably to values of 40–99% if chemical dosing is used in lated the effect of anaerobic stripping on the P-removal and recovery the primary settler, as in the chemically enhanced primary treatment performance within the BCFS® frame and reported an optimal stripping (CEPT) processes (Shewa and Dagnew, 2020). flow rate ratio (flow rate of supernatant/flow rate of influent) of 0.2 to The rest of the manuscript is organized as follows. The studies related favour P-recovery. Higher values resulted in decreased PAO activity and to mainstream P-recovery currently available in the literature are briefly lower values in low P-recovery (see Section 3.2 for the discussion of this described in the following two subsections, classified by their operation effect). Moreover, they showed that the system could operate with a mode: Section 2.1 for sequencing batch reactors (SBRs) and Section 2.2 lower influent COD/P ratio without hindering PAO activity. The ratio for continuous systems. This classification separates SBR configurations, could be decreased from 20 to 10 g COD/g P and the system could still which offer greater configurability than continuous systems, although meet the discharge requirements of P (1 mg P/L) and maintain 36% of they are typically used for lab-scale evaluation of different operating P-recovery. strategies, from continuous systems, which offer more stable operation This first experimental attempt of mainstream P-recovery already and are more suitable for full-scale implementation, although the showed that there is a maximum threshold value for the amount of P achievable P concentration is typically lower. After this schematic extracted. Barat and van Loosdrecht (2006) implemented the model of description, Section 3 critically discusses the most important factors the Hardenberg WWTP and simulated different control strategies to affecting P-recovery: SBR vs. continuous configurations, extraction explore the P-recovery potential of the BCFS® configuration. They ratio, influent COD/P ratio, concentration of P in the recovered stream, stated that the maximum potential P-recovery without affecting PAO carbon source needs and future outlook. activity was 60% of the influent P. Following this first report, several other works have presented configurations taking advantage of the PAO capability to increase P concentration under anaerobic conditions 2.1. SBR configurations (Baeza et al., 2017; Shi et al., 2016; Wong et al., 2013). The P-recovery efficiencies of the evaluated works were calculated Most of the configurations for mainstream P-recovery studied in according to Eqs. (1), (2) and (3): previous works have been based on SBR operation (Table 3, Fig. 3). The POUT = PEFF + PBIO (1) supernatant after the anaerobic phase is used for P-recovery since it has the highest P concentration in all the cycle. SBRs provide greater flexi­ PAS = PINF − POUT (2) bility than continuous systems for research studies, and the fact that particular actions can be taken in a single cycle (such as extra COD dosing or supernatant extraction) facilitates the understanding of P-re­ covery performance under different operation modes. For example, Fig. 2. Schematic representations of the main continuous configurations with mainstream P-recovery found in the literature. a) BCFS process implemented in Hardenberg (Hao and van Loosdrecht, 2006), b) A2N-IC process (Shi et al., 2012), c) EBP2R system (Valverde-Pérez et al., 2015), d) trickling filter with alternative anaerobic and aerobic phases and sludge fermentation (Kodera et al., 2013). Please note that the symbols in the legend include those used in Figs. 2 and 3. 5 C. Zhang et al. Table 3 Summary of P-recovery from mainstream and the related operation parameters based on sequencing batch configurations. Reference Configuration Volume Nitrification SRT HRT Temperat- DO pH Recovery Wastewater InfluentCOD/ Influent [P] in the Extraction P-recovery P-removal Sporadic ure species P Pa recovery percentageb efficiencyc efficiencyd external stream carbon addition (L) (d) (h) (◦ C) (mg/L) (gCOD/gP) (mg/L) (mg/L) (%) (%) (%) (Xia et al., anaerobic/ 8 no (ATUe) 60 24 18–22 1.5–2.5 7.0 - Mg3(PO4)2 synthetic 22 9 240f 12.5 79 100 acetate 640 2014) aerobic 8.0 mg COD/L (Acevedo anaerobic/ 7 no (ATU) 10 12 20 1.5–2.5 7.0 - struvite synthetic 13 7.5 60–72 59h 100 acetate 100 et al., aerobic 8.5 110–200 85g 81h 100 mg COD/L 2015) acetate 350 mg COD/L (Wong et al., anaerobic/ano- 10 no (ATU) n.d. n.d. 22 ± 2 5–8 6.8 - n.d. synthetic 22 8 >90 8 80 79 ± 6 acetate 370 2013) xic + biofilm 7.2 mg COD/L (Shi et al., A2O + induced – yes 15 n.d. 20 n.d. 8.2 - HAP synthetic 41 6.1 20 20 33 100 no 2016) crystallization 8.6 /raw (Lv and SBR+EBPR 10 n.d. 11 16 20 ± 1 >2 7.1 - vivianite synthetic 40 10 39 50 50 95- 97 no Yuan, 8.3 2015) (Zou and anaerobic/ 6 yes 20 n.d. 16–21 n.d. >8.5 HAP synthetic 40 5 15.8 20 59.3 91.6 no Wang, anoxic 2016) (Baeza et al., anaerobic/ 10 yes 15 n.d. n.d. 3 n.d. struvite urban 33 9 57 5 63 100 no 2017) aerobic (model) c (Guisasola anaerobic/ 10 no (ATU) 10 12 25 2.5–3.5 7.50 struvite synthetic 15 20 40–60 10 54 98 ± 1 no et al., aerobic ± 10 30 71–123 66c 67± 10 2019) 0.05 10 30 52–95 69c 82 ± 14 6 8 40 59–78 33c 77± 6 (Larriba two sludge 7800 yes 15 24–36 20 ± 6 3 n.d. struvite raw 63 3.8 35 8.5 63 93±9 no et al., system 2020) (Yu et al., anaerobic/ 5.4 yes 10 24 16 ± 0.5 1 n.d. struvite synthetic 400/8 8 26–19 8, 11, 17 10, 24, 50 >90 no 2021) aerobic 0.6 25–16 4, 18,40 >90 0.5 21–6.5 0, 0 60 a Influent entering the secondary treatment. b Calculated as the fraction of the volume of the anaerobic supernatant extracted with respect to the volume of the reactor. c Calculated as the percentage of P in the anaerobic supernatant extracted with respect to the total input of P (per day). d Calculated as the fraction of P in the influent that is removed. e ATU: Allylthiourea. f P concentration of a parent batch reactor with extra dose of acetate. The anaerobic P concentration is around 80 mg/L. g Average. h Weekly average. n.d. Not described. Water Research 212 (2022) 118102 C. Zhang et al. Water Research 212 (2022) 118102 Fig. 3. Schematic representations of the phases used in SBR configurations with mainstream P-recovery found in the literature. a) anaerobic/aerobic SBR EBPR2 process (Baeza et al., 2017; Guisasola et al., 2019), b) anaerobic/anoxic/nitrifying-induced crystallization SBR process (Shi et al., 2016), c) two-step biofilm system alternatively exposed to a dilute wastewater stream to facilitate P uptake (under anoxic condition) and to a recovery stream to facilitate P release (under anaerobic condition with external carbon source) (Wong et al., 2013), d) mainstream shortcut enhanced phosphorus and PHA recovery (mainstream SCEPPHAR) configuration (Larriba et al., 2020). Please note that the symbols in the legend include those used in Figs. 2 and 3. some authors have evaluated the effects of sporadic extractions of cycle. The authors applied this new P-recovery sequence for nearly one anaerobic supernatant and others have tested the automated extraction month in only one out of four cycles although recommended to evaluate every single cycle. Long-term operation is in general required to this approach in a longer period. demonstrate the ability of these two different approaches to maintain Acevedo et al. (2015) investigated a comparable EBPR-SBR config­ stable PAO activity despite the extractions, as a decrease in P availability uration but the extra carbon was dosed in situ in the anaerobic phase. may hinder PAO metabolism. The frequency of this extra carbon addition and subsequent anaerobic Wong et al. (2013) reported a novel two-step biofilm (Fig. 3c) pro­ extraction for potential P-recovery was varied in different sets of ex­ cess to facilitate anoxic P-recovery. In the first step, the PAO-enriched periments: two times per week with low P extraction or once per week biofilm was immersed into a diluted wastewater stream (F1) so that but with high P extraction. Their results were promising since they ob­ the biomass could take up P and nitrate to remove P and N of the tained 59% recovery of the influent P when extracting P two times per wastewater. In a second step, the biofilm was subjected to a week and 81% when they extracted once per week. In the latter case, carbon-enriched stream (F2) under anaerobic conditions, where acetate they used an extra VFA addition of 9% of the usual load. No P-removal was taken up and P was released. Then, the biofilm acted under aerobic deterioration was observed in the cycle after both extractions, but the condition storing P from wastewater and under anaerobic condition long-term impact of a P-recovery strategy with continuous extraction of releasing P and hence allowing the generation of an enriched P-recovery P was not investigated. stream. The anaerobic supernatant volume (1.8 L) was four times lower Lv et al. (2014) studied a 10 L EBPR-SBR system to investigate the than the initial volume of wastewater (7.2 L) and achieved a P con­ effect of the extraction of 5 L of supernatant enriched in P for subsequent centration almost 4 times higher (28 vs 7 mg P/L). An additional chemical precipitation, supernatant that was obtained after an addi­ experiment was conducted to improve the system efficiency. They tional anaerobic settling period. The supernatant was treated with FeCl3 applied a highly enriched carbon stream after 10 cycles of normal for P-precipitation and the treated supernatant was returned to the operation and they obtained an anaerobic supernatant (around 100 mg aerobic stage. This extraction was applied to the SBR once every 3 cycles P/L) which is theoretically high enough for an efficient P-recovery by for a long term and observed the decrease of P-release, P-uptake and the struvite (Rittmann et al., 2011). Nevertheless, they did not test the effect concentration of intracellular poly-P. EBPR activity was deteriorated, of extracting this P on PAO survival neither in a short-term nor a and P-removal efficiency decreased to about 75%. They disabled the long-term basis. P-recovery strategy and tried to recover the EBPR activity by operating Xia et al. (2014) studied a configuration where a fraction of the under conventional anaerobic/aerobic conditions, but the system did anaerobic mixed liquor from a parent EBPR-SBR was treated in a sepa­ not recover. After that work, Lv and Yuan (2015) demonstrated that the rate batch reactor, and intermittent external carbon source was dosed to same system could have a successful short-term performance (40 d) stimulate P-release in view of P-recovery. The sludge from this batch when 50% of anaerobic supernatant (with a concentration of 39.4 mg reactor was separated with a non-woven cloth and recycled to the parent P/L) was extracted. The recovery process was operated in another SBR, whereas the supernatant was treated with MgCl2⋅6H2O to recover P separated reaction-sedimentation basin with FeCl3 as a precipitating as Mg3(PO4)2. Thus, dosing 10% of additional carbon source in the batch agent, and half of the anaerobic supernatant was removed for P-recov­ reactor resulted in the recovery of 79% of the P in the influent in single ery. EBPR activity was strongly limited because of the shortage of poly-P 7 C. Zhang et al. Water Research 212 (2022) 118102 as a continuous stripping of 40 d, and it could not be recovered even anaerobic supernatant extraction ratios (0, 1/12, 1/9, 1/6) and low DO though the stripping was stopped for 60 d conditions (0.6 and 1 mg/L). More than 40% of P could be recovered as In the same direction, Guisasola et al. (2019) proposed the operation struvite due to these extractions. The reactor began to deteriorate and with an automated extraction of anaerobic supernatant every cycle to the efficiency of P-removal decreased to 60% when operating the system understand the P-recovery capacity in a long-term basis but without any with the lowest DO of 0.2 mg/L and a relatively high extraction ratio external dose of additional organic carbon (Fig. 3a). Thus, the anaerobic (1/9), and no P-recovery was found afterwards. They proposed a DO of P concentration was lower than in the previous reports since it was the 0.6 mg/L and an extraction ratio of 1/6 in view of saving energy and an routine anaerobic P concentration of the system. They demonstrated optimal condition of DO of 1 mg/L and extraction ratio of 1/9 to that a successful P-recovery strategy could be maintained with 1 L (10% maximise the process performance. of reactor volume, 20% of the influent) of automated anaerobic Some modelling studies have been reported about mainstream P- extraction per cycle during more than 120 days without any deleterious recovery with SBR systems. Baeza et al. (2017) simulated an EBPR-based effect on EBPR activity. They tried to push the system into the limits and SBR configuration for P-recovery named EBPR2. A typical EBPR anae­ operated with other COD/P ratios and different extraction conditions in robic/aerobic cycle configuration was modified including, after the order to investigate two different aspects: the minimum initial COD/P anaerobic phase, a settling and decanting phase where the P-enriched ratio and the maximum amount of P extraction that the system could stream was extracted for P-recovery (Fig. 3a). They simulated different tolerate without the loss of the PAO activity. They demonstrated that up scenarios and showed that it was possible to obtain a stable operation to 60% of the influent P could be recovered (when extracting 10% of the with a high amount of P recovered (up to 63%) when the supernatant anaerobic volume per cycle) without any harmful effect on PAO activity extraction in each cycle was 5% of the reactor volume. Moreover, total and that the system could be operated under low COD/P ratios (i.e. 7.5 P-removal was achieved regardless of COD fractionation in the waste­ gCOD/gP). water. In addition, they proposed the inclusion of anaerobic sludge (Zu, Nan, He and Liu, 2020) also used a lab-scale SBR to study the purging, which could be a good option in view of a higher energy re­ effect of the volume of anaerobic extraction on the biological C and covery (i.e. increasing biochemical methane potential per gram of vol­ P-removal performance. Different extraction ratio of anaerobic super­ atile suspended solid). Sludge obtained from the anaerobic phase natant (30%, 60% and 90%) were tested for P-recovery as vivianite. contains a higher concentration of poly‑hydroxyalkanoate (PHA), which Extraction was carried out every fourth cycle. P-removal efficiency was would boost the production of methane during anaerobic digestion improved when increasing from 30 to 60% of the extraction volume. (Chan et al., 2020a). The highest PHA production was obtained by However, the phosphorus balance and the PAO population severely extracting 4.3% of the reactor volume per cycle. deceased under the extraction ratio of 90%. The harm was irreversible Taking into account the previous simulations and the pilot plant even after a long-term normal operation without stripping process. experiments of a two-sludge configuration with EBPR and via-nitrite N- In addition to P-removal and P-recovery, some researchers have removal (Marcelino et al., 2011), a new two-sludge process including studied the simultaneous P and N recovery from mainstream in a batch P-recovery was designed (Larriba et al., 2020). This novel configuration system configuration. Shi et al. (2016) proposed an (Fig. 3d) was named shortcut enhanced P and PHA recovery (main­ anaerobic/anoxic/nitrifying-induced crystallization (IC) SBR configu­ stream SCEPPHAR) and was tested in a large demo pilot-scale (total ration (Fig. 3b, named as A2N-IC-SBR) for N, P-removal and P-recovery volume of 7.8 m3) fed with real primary influent (Larriba et al., 2020). It which was composed of three different SBRs (A/A/O-SBR, N-SBR, and consisted of a two-sludge SBR system: HET-SBR was operated as a IC-SBR, with 10, 10 and 1.3 L). Nitrifiers oxidised ammonium to nitrate, conventional anaerobic/anoxic/aerobic (A2O) EBPR-SBR and AUT-SBR which was denitrified by denitrifying PAO (DPAO) linked to P uptake. was aerobic and aimed at a complete or partial nitrification. A third During this process, 1.3 L of anaerobic supernatant extraction was sent crystallization reactor (PRE-SBR) was designed so that the supernatant to the IC column with calcite seeds for P-recovery. The percentage of extracted after the anaerobic phase of HET-SBR could be used for P-re­ chemical P-removal with respect to the total P-removal was 14.9–17.1%. covery. An additional interchange vessel was required to interchange Similarly, Zou and Wang (2016) implemented a lab-scale EBPR-­ supernatants amongst these reactors, in order to avoid idle phases and to P-recovery process aiming at removing COD, nutrients and recovering P, ensure that all three reactors could operate simultaneously rather than which consisted of a DPAO-enriched SBR operated with sequentially. This configuration was operated successfully for a long anaerobic-aerobic/anoxic mode and an IC P-recovery column. The IC term for C, N and P-removal (86 ± 12%, 93 ± 9% and 79 ± 6%) with column received part of the supernatant obtained after a post-anaerobic either full nitrification or nitrite shortcut. P-recovery as struvite was settling phase and the IC supernatant was returned to the anoxic phase carried out in this third reactor by Mg2+ dosing at a pH of 8.5, allowing with external nitrate solution dosage. The system was operated with the recovery of an average of 45% of the influent P as struvite precipitate either synthetic (around 250 mg COD /L) or real domestic wastewater (with a maximum peak of 63%). (COD concentration between 153 and 223 mg COD/ L). They achieved successful N and P-removal and P-recovery efficiency via HAP precipi­ 2.2. Continuous configurations tation: the average removal efficiencies were 82.6%, 87.5% and 91.6% for COD, P and N, respectively. The flow ratio of anaerobic supernatant This section details the reported alternatives for P-recovery from to influent (from 0 to 0.5) was varied and the maximum P-recovery as mainstream with continuous configurations, in addition to the seminal HAP (up to 59%) was achieved at a ratio of 0.3. BCFS process discussed above (van Loosdrecht et al., 1998). Continuous Salehi et al. (2019) explored the possibility of a simultaneous nitri­ systems are often used to test the P-recovery strategies previously fication, denitrification and P-removal (SNDPR) process for nutrient developed in SBR mode. Under these conditions, part of the anaerobic removal and P-recovery. A lab-scale anaerobic/aerobic SBR was supernatant is diverted to a P-recovery equipment and, in some cases, employed with synthetic wastewater feeding. They achieved anaerobic the flow is returned to the system after P is recovered. liquor with high P concentration (i.e. around 100 mg/L) in granular For instance, an anaerobic-anoxic/nitrifying (A2N) continuous SNDPR process with an influent P of 10 mg/L and a low COD = 200 configuration combined with an induced P-crystallisation (IC) column mg/L. was proposed by Shi et al. (2012) for mainstream P-recovery in addition Yu et al. (2021) used an anaerobic-aerobic lab-scale SBR (5.4 L) to to removal of COD, N and P (Fig. 2b). Part of the P enriched anaerobic promote biological C, N and P-removal. Biological nitrogen removal supernatant was extracted to precipitate calcium phosphate crystals in (BNR) was achieved with low DO during the aerobic phase and thus the IC column. The conventional A2N system performance was simultaneous nitrification and denitrification was promoted. Successful compared to that of A2N-IC. A2N-IC showed a more stable performance nutrient removal was stable for a long period (310 d) under different than A2N. The authors used a wide range of COD/P ratios (from 17.7 to 8 C. Zhang et al. Table 4 Summary of P-recovery from mainstream and the related operation parameters based on continuous configurations. Reference Configuration Volume Nitrificat- SRT HRT Temper- DO Recovery Wastewater Influent Influent [P] in the Extraction P-recovery P-removal Sporadic ion ature species COD/P Pa recovery percentageb efficiencyc efficiencyd external stream carbon addition (L) (d) (h) (◦ C) (mg/ (gCOD/ (mg/L) (mg/L) (%) (%) (%) L) gP) (Barat and van BCFS-TUDe model 1.215⋅107 yes n.d. 3.8 10–20 0.6 P Stripper model 54 14.2 60 - 70 kg P/ 45 ~60 > 90 no Loosdrecht, calibrated for and (stripping) d 2006) Hardenberg WWTP 2.8 (Shi et al., 2012) BNR-ICf (A2O- 31 yes 15 n.d. n.d. 2–3.8 HAP synthetic 19.9 ± 1.5 12.7 ± 0.7 n.d. 20 8 > 99 no crystallization) 31.0 ± 4.2 7.82 ± 9.1 42.4 ± 4.4 0.68 10.8 67.4 ± 5.73 ± 12.6 10.2 0.51 3.75± 0.51 (Kodera et al., PAOs-enriched biofilm 2.4 n.d. n.d. 0.5 20 0.9 L/ struvite synthetic 200/5 5 125 10 ~60 50–60 no 2013) within a modified h airg trickling filter (Valverde-Pérez EBP2R green 5.1⋅106 yes 5 4.25 n.d. 0–3 green model 79 9 n.d. 30 70 n.d. no et al., 2015) microalgae cultivation microalgae 9 (Zou and Lu, 2016) BNR-IC 12.4 yes n.d. n.d. n.d. >2 HAP synthetic 25 10 28 35 70 95.2 ± 1.0 no real 20–30 7.08–7.69 12 35 74 90.7 ± 1.2 domestic (Dai et al., 2017) BNR-IC 316 yes 12–16 n.d. n.d. n.d. HAP synthetic 50 5 24 30 83 95 no a Concentration of P for the EBPR process. b Calculated as the ratio of the flow rate of anaerobic supernatant diverted to the P-recovery unit and flow rate of the influent. c Calculated as the percentage of P in the anaerobic supernatant extracted for P-recovery per day in respect of the total input of P in the influent per day. d Calculated as the removal fraction of phosphate in the effluent with respect to the phosphate in the influent. e BCFS-TUD: Biological-chemical phosphorus and nitrogen removal system-Model system of Delft University Technology. f BNR-IC: An anaerobic-anoxic/nitrifying (A2N) continuous configuration for biological nutrients removal (BNR) combined with an induced P-crystallisation (IC) column. g Aeration flow rate. n.d. Not described. Water Research 212 (2022) 118102 C. Zhang et al. Water Research 212 (2022) 118102 67.4 g COD/gP) and found that the P-removal efficiency was consis­ recovery can be achieved with different configurations and, thus, several tently maintained at 99.2% for A2N-IC, regardless of the influent COD/P operation parameters are essential for an efficient P-recovery perfor­ ratio tested. However, the A2N process reduced P-removal efficiency mance. Tables 3 and 4 summarize the different reported proposals for from 93.0% to 65.7% when the influent COD/P ratio was decreased the implementation of P-recovery from mainstream and lists, for each from 67.4 to 17.7 gCOD/gP. P-recovery in the A2N-IC process ranged case, the applied equipment and recovery methods, basic operation from 8.9 to 12.9% when the influent COD/P ratio was reduced from 67.4 parameters (SRT, HRT, operation temperature, aeration, pH, extraction to 19.9 gCOD/gP. Later, Zou and Wang (2017) used the same configu­ volume and frequency) and influent characteristics. ration and studied the effect of various operational conditions on the IC performance and optimized the crystal dose acting as a seed, pH, tem­ 3.1. SBR vs continuous configurations for mainstream P-recovery perature, DO, molar Ca/P ratio and reaction time in view of boosting P-recovery. They showed that on the optimal conditions (80 min, pH of Mainstream P-recovery under SBR conditions is based on extracting 8.5, molar Ca/P ratio of 2.0, 60 g/L seed crystal loads and 4.0 L/min of part of the supernatant once the anaerobic phase is finished, whereas it aeration flow) the system reached an excellent P-recovery of 92.3%. In consists of diverting part of the anaerobic liquid to a P-recovery unit addition, they found that the temperature had a slight effect on induced under continuous conditions. Most reports on mainstream P-recovery HAP crystallization, thus benefiting the integration of EBPR with employed lab-scale SBRs (Table 3), as SBRs have several advantages induced HAP crystallization. compared to continuous systems. On the one hand, it is easier to obtain a In a similar configuration, Dai et al. (2017) evaluated the influence of biomass-free anaerobic supernatant as a settling phase can be easily the flow of the anaerobic supernatant to IC in a long run (over 180 d) and included after the anaerobic phase, whereas in the case of continuous found that the optimal ratio of anaerobic supernatant extracted to system a membrane or a combined settling reactor+biomass recycle influent was 0.3. Under these conditions, P-recovery efficiency reached would be needed (Sekine et al., 2018). Moreover, SBRs show a lower 72.4% at the optimal HRT = 0.5 h in IC reactor. Higher extraction ratios space requirement and lower configuration and installation costs, as could disrupt system stability without enhancing P-recovery and could well as the flexibility to face operational or influent variations (Jiang induce a change of the activated sludge characteristics since it favoured et al., 2016; Tomei et al., 2016). However, it can also lead to higher the percentage of extracellular polymeric substances of the sludge. The operational costs and capital expenditures due to its discontinuous same authors reported a modelling study about the continuous A2N-IC operation, as reported in (Larriba et al., 2020) when comparing the aiming at finding the optimal conditions to maximize energy savings mainstream SCEPPHAR configuration to a conventional continuous A2O and minimize emission reduction (Dai et al., 2019). Their results indi­ configuration. cated that P-recovery decreased with the reduction of the extraction ratio and the IC volume. 3.2. Extraction ratio Kodera et al. (2013) studied a configuration based on a PAO-enriched biotrickling filter, exposed alternatively to anaerobic and As mentioned before, the key operational parameter for efficient P- aerobic conditions (Fig. 2d). P-uptake by PAO was promoted under recovery is the extraction ratio, i.e. the percentage of anaerobic volume aerobic conditions, when the filter treated by recirculation a synthetic extracted for an SBR or the flow diverted from the anaerobic reactor to stream emulating an urban wastewater, which contained a low con­ the P-recovery unit for a continuous system. High anaerobic extraction centration of P (5 mg P/L). For the anaerobic phase, the system required volumes should be chosen to maximise P-recovery, but too high values a stream with VFA to enhance P-release, with up to 2000 mg COD/L may hinder P-storage activity of PAO. The rationale is that the effect of which could be generated by primary sludge fermentation, although in internal poly-P limitations is not observed with single periodic extrac­ this work a synthetic stream was used. The effluent from the anaerobic tion or in the short-term, but after several cycles, high P extraction phase had a high concentration of P (125 mg P/L) that was adequate for values would deplete the internal poly-P reserves and PAO would lose P-recovery. The system operated for a long period of time (250 d) with the competitive advantage versus other microorganisms such as this configuration, obtaining a P-recovery of almost 60% of the influent glycogen accumulating organisms (GAO), which are able to uptake VFA P. as PHA without requiring poly-P reserves. For this reason, many authors In the frame of the cultivation of microalgae for P-recovery, Valve­ have opted for sporadic extraction of anaerobic liquid. Thus, the re­ rde-Pérez et al. (2015) proposed a novel model-based design of P ported P-recovery efficiencies for single/short-term experiments can be removal and recovery system consisting of a conventional continuous very high, particularly if an extra dose of carbon is added under A2O operation where nitrification was suppressed at low DO (Fig. 2c). anaerobic conditions to boost P-release. In this sense, as observed in The objective was to produce a nutrient medium for microalgae culti­ Table 3, the P-recovery efficiency ranges from 30 to 90% as a function of vation to produce biofuel or a natural fertiliser for direct use in agri­ the SBR operational conditions. For example, Acevedo et al. (2015) re­ culture. The P recovered for algae cultivation was obtained from the ported P-recovery efficiencies around 80% in a single cycle with anaerobic effluent as a stream with high P concentration. They inves­ extraction values of nearly 75% and an external carbon dosage of 350 tigated the potential P-recovery under different scenarios of DO and SRT mg COD/L as acetate. However, it must be carefully considered that and obtained a maximum of 70% P-recovery of influent P, when 30% of even if 100% P-recovery from one out of four cycles was achieved, the the influent flow was diverted as a P-stream and the system was operated average extraction per treated cycle in the plant would only be about at SRT = 5 d. At DO = 1.5 mg/L, nitrification was still being inhibited 25% of the P input. Moreover, one should realise that these high P ex­ due to this low DO and P uptake was favoured, increasing P-recovery up tractions are sporadic and could not be sustained in a routine basis due to 75%. to loss of PAO activity. For example, Lv et al. (2014) operated an EBPR-SBR with an extraction ratio of 0.5 once every 3 cycles. After 83 3. Discussion days of P stripping, P-removal performance decreased, observing a metabolic shift from PAO to GAO and PAO declined to less than 1% of Amongst the different locations for P-recovery in EBPR-based the population. In a subsequent study, Lv and Yuan (2015) extracted half WWTPs, the supernatant of the anaerobic reactor has recently gained of the anaerobic supernatant every cycle and reached a P-recovery ef­ attention since mainstream P-recovery would i) mitigate undesired ficiency of 50% in 40 days of operation without additional VFA dosage precipitation of P in the sludge line, ii) reduce the COD requirements for in the anaerobic stage, but again the internal poly-P levels decreased and P-removal, iii) enable the treatment of wastewater with lower COD/P the system failed in the next 20 days ratio and iv) depending on the recovery agent, decrease the N load to the On the other hand, if the extraction ratio is properly selected to avoid plant and, thus, its operational costs. As detailed above, mainstream P- over-extraction, PAO activity can be maintained for a longer period of 10 C. Zhang et al. Water Research 212 (2022) 118102 time. In this sense, a maximum automated cycle extraction ratio of 10% influent ammonium (1:1 molar P:N ratio) would be directly recovered was suggested to ensure the stability of PAO activity and thus the without nitrification/denitrification needed. This would decrease the N feasibility of a long-term P-recovery process (Guisasola et al., 2019), load to the plant and, thus, the operation costs (e.g. those related to which allowed the recovery of 60% of influent P without any harmful aeration or biomass handling). To the best of our knowledge, an effect on PAO. experimental study that analyses these interactions has not been re­ In a modelling study, Baeza et al. (2017) evaluated the impact of the ported yet. When aiming at simultaneous N and P-removal, a higher anaerobic supernatant extraction ratio on the performance of an COD/P ratio is needed in order to divert some electron donor to BNR. EBPR-SBR P-recovery system. If the objective was achieving the highest BNR results in a certain entrance of nitrate to the anaerobic reactor and, concentration of P in the supernatant to facilitate precipitation, an if COD limitations exist, that would hinder the amount of P to be extraction ratio lower than 5% of the reactor volume (10% of the released and, thus, the amount of P to be recovered. Guerrero et al. influent) was recommended, for which a concentration of 61 mg P/L was (2011) already demonstrated that PAO can outcompete denitrifiers obtained. Lower ratios did not increase P concentration and resulted in a under certain operational conditions. Salehi et al. (2019) showed that lower amount of P redirected to the P-recovery system, due to the lower P-recovery could be achieved together with BNR with synthetic waste­ volume. On the other hand, higher ratios led to a decrease of PAO ac­ water of only 200 mg COD/L and a COD/P ratio of 20 gCOD/gP. tivity and hence lower P concentration, hindering P-precipitation. Regarding the experimental studies, the range of extraction ratios 3.4. Concentration of P in the recovered stream that led to a successful long-term P-recovery was 8–15% of reactor volume in batch mode and 20 to 35% of flowrate under continuous A high P concentration is critical to guarantee a high P-recovery mode. Moreover, lower extraction ratios lead to higher anaerobic P performance: the Phostrip and BCFS® processes report a minimum concentrations, which from a practical point of view is very interesting phosphate concentration in the stripper stream of 25 mg P/L for an since precipitation becomes more feasible (i.e. with less chemical dosage efficient P-recovery during the crystallization process (Barat and van requirements). On the other side, higher extraction ratios enable a Loosdrecht, 2006; Hao and van Loosdrecht, 2006). Other works report a higher amount of P-recovery but at expenses of higher costs since the higher minimum P concentration (more than 50 mg/L) for the struvite anaerobic P concentration is lower. In summary, choosing the optimal precipitation (Cordell et al., 2009; Kodera et al., 2013). Most of the extraction ratio requires a balance between a) optimal PAO activity, mainstream P-recovery processes listed in Tables 3 and 4 report P con­ high anaerobic P concentration and a low amount of P recovered vs. b) centrations higher than 50 mg/L in the recovery stream. Even so, some hindering PAO activity, low anaerobic P concentration and a high works report more than 100 mg/L of P in the enriched stream by amount of P recovered. Thus, it is difficult to suggest a general optimal repeated exposure of the biofilm to external COD source (acetate or value but, based on the results reported in the literature, we propose propionate). The typical reported anaerobic P concentrations in EBPR 10% as a first initial guess for the extraction ratio when designing batch systems are in the range of 30 to 70 mg P/L, so they are in the bottom systems and 25% when designing continuous systems. limit and this is why some of these reports use an external carbon dosage under anaerobic conditions to increase anaerobic P-release. In a real 3.3. Influent COD/P ratio and N-removal environment, external COD dosage under anaerobic conditions boosts P-release at expenses of higher operating cost. Moreover, this addition Another significant parameter is the influent COD/P ratio. The should be sporadic or, otherwise, the internal poly-P reserves will be theoretical studies on mainstream P-recovery indicate that wastewater depleted and, thus, this operation would not be sustained in a long-term with a low COD/P ratio could be successfully treated in systems where basis. mainstream P-recovery is implemented due to a much better utilisation of the influent C. The range of influent COD/P ratios in Tables 3 and 4 is 3.5. Carbon source needs very wide (from 7.5 to 80 gCOD/gP), since few reports aimed at finding the minimum COD requirements for a successful mainstream P- Most of the works in Table 3 and 4 were conducted at lab-scale and removal/recovery. In general, continuous systems are reported with a used synthetic wastewater to mimic urban wastewater. However, some higher COD/P ratio. In addition, some of the reports include simulta­ of them prepared a highly loaded wastewater with an extra C inlet, neous N and P-removal and, hence, a higher amount of COD is needed. normally acetic acid or propionic acid (Acevedo et al., 2015; Guisasola Other reports somehow prevent nitrification from occurring either with et al., 2019; Wong et al., 2013) and there was no need for a sporadic a low SRT, with allylthiourea (ATU) addition, low DO setpoint or with a carbon dosage. In the case of the reports with real wastewater, the very low influent nitrogen load. For example, Valverde-Pérez et al. additional carbon source was only supplemented when the influent COD (2015) fixed a DO setpoint of 1.5 mg/L and P-recovery efficiency was relatively low (Zou and Lu, 2016). The only work with external COD improved up to 75% of the influent P because the activity of nitrifiers amendment and a real environment is presented by Larriba et al. (2020). was being limited at low oxygen concentrations and P uptake was fav­ They treated the effluent from the primary settler of a full sale WWTP by oured. In any case, the choice of the DO setpoint is not straightforward the mainstream SCEPPHAR configuration aiming at P-recovery in a since too low DO values may also hinder PAO activity. demo scale and studied different strategies (such as nitrite pathway) to Guisasola et al. (2019) prevented nitrifying activity using ATU, improve the efficiency of C usage in view of decreasing the carbon tested different influent COD/P ratio and showed that the minimum footprint and the operational costs. initial COD/P ratio (7.5 gCOD/gP) allowed 60% of the P-recovery This external carbon can also be obtained in a more sustainable way without any harmful effect on PAO activity, which is the value that from the own internal plant resources though the fermentation of waste agrees with the first experiments conducted with the BCFS® method­ sludge, which is an environmental way to obtain the VFA required to ology (Barat and van Loosdrecht, 2006). This value also agrees with the boost anaerobic P-release. In the work of Kodera et al. (2013), an modelling work of Baeza et al. (2017), where a supernatant extraction of additional fermentation tank was conceptually proposed to obtain 5% of the reactor volume in each SBR cycle led to a stable operation with enriched VFA by primary sludge fermentation, which was supposed to up to 63% of P recovered. The rest of P removed was incorporated to the provide enough COD for anaerobic P-release and subsequent P-recovery. purged biomass, where additional P-recovery strategies reported for the Nowadays, the use of side-stream sludge fermenters to provide COD to sludge line (i.e. as sewage, sewage ash or digestate) could also be applied fulfil the carbon needs for biological nutrient removal is becoming a in view of boosting the global P-recovery efficiency. trend (Ali et al., 2021; Li, Li and Li, 2018; Wang et al., 2019). Thus, it The interaction between mainstream P-recovery and nitrogen could also be used as a source for extra carbon dosage in view of removal deserves more attention. If P is recovered as struvite, part of the P-recovery. 11 C. Zhang et al. Water Research 212 (2022) 118102 3.6. Effect of other operational parameters - Configurations combining EBPR with mainstream P-recovery are still at an early stage and their successful full-scale application has not yet Regarding the effect of other operational parameters on the observed been reported. performance, the main general recommendation would be to select them - The most important parameter to describe these systems is the to ensure good PAO activity. The same conditions that would maintain extraction ratio: the amount of anaerobic supernatant to be extracted PAO activity in a conventional SBR or A2O continuous system would be or derived to recover P. The choice of this value is a compromise recommended in a mainstream P-recovery configuration. For example, between i) having a low amount of P recovered with high PAO ac­ (Baeza et al., 2017) reported the effect of SRT in an SBR configuration tivity and anaerobic P concentration (i.e. easing precipitation) and ii) with P-recovery and found a minimum of SRT = 4 d was required to having a high amount of P recovered with a potential deleterious achieve complete P-removal, but to achieve maximum PHA production, effect to PAO activity and low anaerobic P concentration. SRT = 10 d was recommended. Therefore, for the specific wastewater - A maximum value for the P-recovery efficiency of 60% of influent P composition used and the other particular experimental conditions (e.g. seems the most sensitive goal for a long-term operation of these T, DO, pH, volume exchange ratio, operation objective) these results systems. However, higher values can be used for sporadic or single were obtained, but they are case-specific, and hence, for each scenario, extractions. they should be determined experimentally or using a proper modelling - Regarding the preferred agent for P-recovery, most of the works approach. In fact, mainstream P-recovery is adding an additional limi­ opted for struvite, while vivianite or HAP are less common at this tation to the typical needs that PAO microorganisms also require, such point. as a minimal T-dependent SRT (Brdjanovic et al., 1998; Chan et al., 2020b), enough DO and aerobic phase length, and typical aerobic and Declaration of Competing Interest anaerobic HRT (Tchobanoglous et al., 2014). The authors declare that they have no known competing financial 3.7. Future outlook interests or personal relationships that could have appeared to influence the work reported in this paper. P-recovery from the anaerobic supernatant of the mainstream EBPR process is a novel and promising process, as in addition to the advan­ Acknowledgements tages of other P-recovery configurations such as alleviating the problem of undesired P precipitation and recovering a valuable resource, it has This work was supported by the Spanish Ministerio de Economía y additional advantages that help to reduce the COD requirements for P- Competitividad (CTQ2017–82404-R) with funds from the Fondo removal, allowing the treatment of wastewaters with lower COD/P Europeo de Desarrollo Regional (FEDER). Congcong Zhang would like to ratio, and maintaining high P-removal activity. However, most of the thank the financial support from China Scholarship Council. The authors investigations concentrated on lab-scale SBR and almost all recovery are members of the GENOCOV research group (Grup de Recerca Con­ processes were based on P crystallization. solidat de la Generalitat de Catalunya, 2017 SGR 1175, www.genocov. In order to optimize these systems, the extraction ratio is the most com). basic and important operating parameter to achieve high P-recovery without impairing the long-term stability of the system. Anyway, proper References simulation studies should be performed, as the operating parameters are case-specific and highly dependant on the operating objective. Acevedo, B., Camiña, C., Corona, J.E., Borrás, L., Barat, R., 2015. The metabolic To demonstrate that this is a reliable technology for P-recovery, versatility of PAOs as an opportunity to obtain a highly P-enriched stream for further P-recovery. Chem. Eng. J. 270, 459–467. https://doi.org/10.1016/j. future works should focus on continuous configurations, which seem cej.2015.02.063. more suitable for full-scale application, and using real municipal Adam, C., Kley, G., Simon, .F.-G., 2007. Thermal treatment of municipal sewage sludge wastewater. The objective is challenging, as it is necessary to obtain an aiming at marketable P-fertilisers. Mater. Trans. 48, 3056–3061. https://doi.org/ 10.2320/matertrans.MK200707. enriched stream with high P concentration to facilitate P-recovery, but Ali, P., Zalivina, N., Le, T., Riffat, R., Ergas, S., Wett, B., Murthy, S., Al-Omari, A., without requiring the addition of an external carbon source that would deBarbadillo, C., Bott, C., De Clippeleir, H., 2021. Primary sludge fermentate as compromise the economics of the process. Moreover, LCA and cost carbon source for mainstream partial denitrification–anammox (PdNA). Water analysis should be used to demonstrate the overall positive effect of P- Environ. Res. 93, 1044–1059. https://doi.org/10.1002/wer.1492. Amann, A., Zoboli, O., Krampe, J., Rechberger, H., Zessner, M., Egle, L., 2018. recovery systems in WRRFs. In any case, policy decisions will be of Environmental impacts of phosphorus recovery from municipal wastewater. Resour. paramount importance for the implementation of P-recovery, as it has Conserv. Recycl. 130, 127–139. https://doi.org/10.1016/j.resconrec.2017.11.002. been shown that the costs are affordable, but like many environmental Annisa, T., Azkiya, A., N Fauzi, R., B D Nandiyanto, A., Hofifah, S.N., 2021. Cost analysis and economic evaluation for manufacturing hydroxyapatite nanoparticles from technologies, it is not cost-effective unless it is driven by policy (Nättorp eggshell waste. Int. J. Res. Appl. Technol. 1, 211–226. https://doi.org/10.34010/ et al., 2017). injuratech.v1i1.5669. Baeza, J.A., Guerrero, J., Guisasola, A., 2017. Optimising a novel SBR configuration for enhanced biological phosphorus removal and recovery (EBPR2). Desalin. Water 4. Conclusions Treat 68, 319–329. https://doi.org/10.5004/dwt.2017.20468. Barat, R., van Loosdrecht, M.C.M., 2006. Potential phosphorus recovery in a WWTP with This paper critically reviews for the first time the performance of the the BCFS process: interactions with the biological process. Water Res 40, 3507–3516. https://doi.org/10.1016/j.watres.2006.08.006. current reported strategies for mainstream P-recovery in EBPR WWTPs Beaudry, J.W., Sengupta, S., 2021. Phosphorus recovery from wastewater using pyridine- and provides the main points to be addressed for its successful imple­ based ion-exchange resins: role of impregnated iron oxide nanoparticles and mentation. Furthermore, it shows that full-scale mainstream P-recovery preloaded Lewis acid (Cu 2+). Water Environ. Res. 93, 774–786. https://doi.org/ 10.1002/wer.1469. is a medium-term possibility. Berg, U., Donnert, D., Ehbrecht, A., Bumiller, W., Kusche, I., Weidler, P.G., Nüesch, R., 2005. “Active filtration” for the elimination and recovery of phosphorus from waste - It is based on the PAO capability of producing a P-enriched super­ water. Colloids Surf. A Physicochem. Eng. Asp. 265, 141–148. https://doi.org/ natant that enables P-recovery, and SBR and continuous configura­ 10.1016/j.colsurfa.2004.10.135. Biswas, B.K., Inoue, K., Harada, H., Ohto, K., Kawakita, H., 2009. Leaching of phosphorus tions at different scales have been reported. from incinerated sewage sludge ash by means of acid extraction followed by - It can improve P-removal performance and enable a successful adsorption on orange waste gel. J. Environ. Sci. 21, 1753–1760. https://doi.org/ treatment of influents with a relatively low COD/P ratio. 10.1016/S1001-0742(08)62484-5. Blaney, L., Cinar, S., Sengupta, A., 2007. Hybrid anion exchanger for trace phosphate removal from water and wastewater. Water Res. 41, 1603–1613. https://doi.org/ 10.1016/j.watres.2007.01.008. 12 C. Zhang et al. Water Research 212 (2022) 118102 Brdjanovic, D., van Loosdrecht, M.C.M., Hooijmans, C.M., Alaerts, G.J., Heijnen, J.J., nitrogen and phosphorus removal. Water Res 45, 4793–4802. https://doi.org/ 1998. Minimal aerobic sludge retention time in biological phosphorus removal 10.1016/j.watres.2011.06.019. systems. Biotechnol. Bioeng. 60, 326–332. https://doi.org/10.1002/(SICI)1097- Guida, S., Rubertelli, G., Jefferson, B., Soares, A., 2021. Demonstration of ion exchange 0290(19981105)60:3<326::AID-BIT8>3.0.CO;2-J. technology for phosphorus removal and recovery from municipal wastewater. Chem. Bunce, J.T., Ndam, E., Ofiteru, I.D., Moore, A., Graham, D.W., 2018. A review of Eng. J. 420, 129913 https://doi.org/10.1016/j.cej.2021.129913. phosphorus removal technologies and their applicability to small-scale domestic Guisasola, A., Chan, C., Larriba, O., Lippo, D., Suárez-Ojeda, M.E., Baeza, J.A., 2019. wastewater treatment systems. Front. Environ. Sci. https://doi.org/10.3389/ Long-term stability of an enhanced biological phosphorus removal system in a fenvs.2018.00008. phosphorus recovery scenario. J. Clean. Prod. 214, 308–318. https://doi.org/ Chan, C., Guisasola, A., Baeza, J.A., 2020a. Correlating the biochemical methane 10.1016/j.jclepro.2018.12.220. potential of bio-P sludge with its polyhydroxyalkanoate content. J. Clean. Prod. 242, Hao, X.-.D., van Loosdrecht, M.C.M., 2006. Model-based evaluation of struvite recovery 118495 https://doi.org/10.1016/j.jclepro.2019.118495. from an in-line stripper in a BNR process (BCFS). Water Sci. Technol. 53, 191–198. Chan, C., Guisasola, A., Baeza, J.A., 2020b. Living on the edge: prospects for enhanced https://doi.org/10.2166/wst.2006.092. biological phosphorus removal at low sludge retention time under different Henze, M., van Loosdrecht, M.C.M., Ekama, G.A., Brdjanovic, D., 2008. Biological temperature scenarios. Chemosphere 258, 127230. https://doi.org/10.1016/j. Wastewater Treatment. IWA Publishing. chemosphere.2020.127230. Hester, Harrison, 2013. Waste as A Resource. Royal Society of Chemistry. Chen, P.-.J., Faust, S.D., 1974. The solubility product of ferrous phosphate. Environ. Lett. Hu, M., Fan, B., Wang, H., Qu, B., Zhu, S., 2016. Constructing the ecological sanitation: a 6, 287–296. https://doi.org/10.1080/00139307409437371. review on technology and methods. J. Clean. Prod. 125, 1–21. https://doi.org/ Chisti, Y., 2013. Constraints to commercialization of algal fuels. J. Biotechnol. https:// 10.1016/j.jclepro.2016.03.012. doi.org/10.1016/j.jbiotec.2013.07.020. Izadi, Parnian, Izadi, Parin, Eldyasti, A., 2021. A review of biochemical diversity and Chisti, Y., 2008. Biodiesel from microalgae beats bioethanol. Trends Biotechnol. 26, metabolic modeling of EBPR process under specific environmental conditions and 126–131. https://doi.org/10.1016/j.tibtech.2007.12.002. carbon source availability. J. Environ. Manage. 288, 112362 https://doi.org/ Chowdhury, R.B., Moore, G.A., Weatherley, A.J., Arora, M., 2017. Key sustainability 10.1016/j.jenvman.2021.112362. challenges for the global phosphorus resource, their implications for global food Izadi, Parnian, Izadi, Parin, Eldyasti, A., 2020. Design, Operation and Technology security, and options for mitigation. J. Clean. Prod. 140, 945–963. https://doi.org/ Configurations For Enhanced Biological Phosphorus Removal (EBPR) Process: A 10.1016/j.jclepro.2016.07.012. Review, Reviews in Environmental Science and Biotechnology. Springer Cichy, B., Kużdżał, E., Krztoń, H., 2019. Phosphorus recovery from acidic wastewater by Netherlands. https://doi.org/10.1007/s11157-020-09538-w. hydroxyapatite precipitation. J. Environ. Manage. 232, 421–427. https://doi.org/ Jabr, G., Saidan, M., Al-Hmoud, N., 2019. Phosphorus recovery by struvite formation 10.1016/j.jenvman.2018.11.072. from Al Samra municipal wastewater treatment plant in Jordan. Desalin. Water Cieślik, B., Konieczka, P., 2017. A review of phosphorus recovery methods at various Treat. 146, 315–325. https://doi.org/10.5004/dwt.2019.23608. steps of wastewater treatment and sewage sludge management. The concept of “no Jiang, Y., Wei, L., Zhang, H., Yang, K., Wang, H., 2016. Removal performance and solid waste generation” and analytical methods. J. Clean. Prod. 142, 1728–1740. microbial communities in a sequencing batch reactor treating hypersaline phenol- https://doi.org/10.1016/j.jclepro.2016.11.116. laden wastewater. Bioresour. Technol. 218, 146–152. https://doi.org/10.1016/j. Cordell, D., Drangert, J.-.O., White, S., 2009. The story of phosphorus: global food biortech.2016.06.055. security and food for thought. Glob. Environ. Chang. 19, 292–305. https://doi.org/ Johansson, S., Ruscalleda, M., Colprim, J., 2017. Phosphorus recovery through 10.1016/j.gloenvcha.2008.10.009. biologically induced precipitation by partial nitritation-anammox granular biomass. Cornel, P., Schaum, C., 2009. Phosphorus recovery from wastewater: needs, technologies Chem. Eng. J. 327, 881–888. https://doi.org/10.1016/j.cej.2017.06.129. and costs. Water Sci. Technol. 59, 1069–1076. https://doi.org/10.2166/ Kabdaşlı, I., Tünay, O., 2018. Nutrient recovery by struvite precipitation, ion exchange wst.2009.045. and adsorption from source-separated human urine – a review. Environ. Technol. Wild, D., Kisliakova, A., Siegrist, H., 1996. P-fixation by Mg, Ca and zeolite a during Rev. 7, 106–138. https://doi.org/10.1080/21622515.2018.1473504. stabilization of excess sludge from enhanced biological P-removal. Water Sci. Kodera, H., Hatamoto, M., Abe, K., Kindaichi, T., Ozaki, N., Ohashi, A., 2013. Phosphate Technol. 34, 391–398. https://doi.org/10.1016/0273-1223(96)00528-8. recovery as concentrated solution from treated wastewater by a PAO-enriched Dai, H., Chen, W., Peng, L., Wang, X., Lu, X., 2019. Modeling and performance biofilm reactor. Water Res. 47, 2025–2032. https://doi.org/10.1016/j. improvement of an anaerobic–anoxic/nitrifying-induced crystallization process via watres.2013.01.027. the multi-objective optimization method. Environ. Sci. Pollut. Res. 26, 5083–5093. Larriba, O., Rovira-Cal, E., Juznic-Zonta, Z., Guisasola, A., Baeza, J.A., 2020. Evaluation https://doi.org/10.1007/s11356-018-3971-1. of the integration of P recovery, polyhydroxyalkanoate production and short cut Dai, H., Dai, Z., Peng, L., Wu, Y., Zou, H., Lu, X., 2017. Metagenomic and metabolomic nitrogen removal in a mainstream wastewater treatment process. Water Res. 172, analysis reveals the effects of chemical phosphorus recovery on biological nutrient 115474 https://doi.org/10.1016/j.watres.2020.115474. removal system. Chem. Eng. J. 328, 1087–1097. https://doi.org/10.1016/j. Law, K.P., Pagilla, K.R., 2018. Phosphorus recovery by methods beyond struvite cej.2017.07.119. precipitation. Water Environ. Res. 90, 840–850. https://doi.org/10.2175/ Desmidt, E., Ghyselbrecht, K., Zhang, Y., Pinoy, L., Van der Bruggen, B., Verstraete, W., 106143017X15131012188006. Rabaey, K., Meesschaert, B., 2015. Global phosphorus scarcity and full-scale p- Le Corre, K.S., Valsami-Jones, E., Hobbs, P., Parsons, S.A., 2009. Phosphorus recovery recovery techniques: a review. Crit. Rev. Environ. Sci. Technol. 45, 336–384. from wastewater by struvite crystallization: a review. Crit. Rev. Environ. Sci. https://doi.org/10.1080/10643389.2013.866531. Technol. 39, 433–477. https://doi.org/10.1080/10643380701640573. Donatello, S., Tong, D., Cheeseman, C.R., 2010. Production of technical grade phosphoric Leinweber, P., Bathmann, U., Buczko, U., Douhaire, C., Eichler-Löbermann, B., acid from incinerator sewage sludge ash (ISSA). Waste Manag 30, 1634–1642. Frossard, E., Ekardt, F., Jarvie, H., Krämer, I., Kabbe, C., Lennartz, B., Mellander, P. https://doi.org/10.1016/j.wasman.2010.04.009. E., Nausch, G., Ohtake, H., Tränckner, J., 2018. Handling the phosphorus paradox in Doyle, J.D., Parsons, S.A., 2002. Struvite formation, control and recovery. Water Res. 36, agriculture and natural ecosystems: scarcity, necessity, and burden of P. Ambio 47, 3925–3940. https://doi.org/10.1016/S0043-1354(02)00126-4. 3–19. https://doi.org/10.1007/s13280-017-0968-9. Driver, J., Lijmbach, D., Steen, I., 1999. Why recover phosphorus for recycling, and how? Levlin, E., Hultman, B., 2003. Phosphorus recovery from phosphate rich side-streams in Environ. Technol 20. https://doi.org/10.1080/09593332008616861. Volume. wastewater treatment plants. Polish Swedish Semin. Gdansk March 47–56. Egle, L., Rechberger, H., Krampe, J., Zessner, M., 2016. Phosphorus recovery from Li, B., Huang, H.M., Boiarkina, I., Yu, W., Huang, Y.F., Wang, G.Q., Young, B.R., 2019. municipal wastewater: an integrated comparative technological, environmental and Phosphorus recovery through struvite crystallisation: recent developments in the economic assessment of P recovery technologies. Sci. Total Environ. 571, 522–542. understanding of operational factors. J. Environ. Manage. 248, 109254 https://doi. https://doi.org/10.1016/j.scitotenv.2016.07.019. org/10.1016/j.jenvman.2019.07.025. Egle, L., Rechberger, H., Zessner, M., 2015. Overview and description of technologies for Li, J., Chen, Z., Wang, Q., Fang, L., Xue, Q., Cheeseman, C.R., Donatello, S., Liu, L., recovering phosphorus from municipal wastewater. Resour. Conserv. Recycl. 105, Poon, C.S., 2018. Change in re-use value of incinerated sewage sludge ash due to 325–346. https://doi.org/10.1016/j.resconrec.2015.09.016. chemical extraction of phosphorus. Waste Manag 74, 404–412. https://doi.org/ El Hamouri, B., 2009. Rethinking natural, extensive systems for tertiary treatment 10.1016/j.wasman.2018.01.007. purposes: the high-rate algae pond as an example. Desalin. Water Treat 4, 128–134. Li, R., Li, B., Li, X., 2018. An integrated membrane bioreactor system with iron-dosing https://doi.org/10.5004/dwt.2009.367. and side-stream co-fermentation for enhanced nutrient removal and recovery: European Commission, 2019. Regulation (EU) 2019/1009 laying down rules on the system performance and microbial community analysis. Bioresour. Technol. 260, making available on the market of EU fertilising products and amending regulations 248–255. https://doi.org/10.1016/j.biortech.2018.03.100. (EC) No 1069/2009 and (EC) No 1107/2009 and repealing regulation (EC) No 2003/ Li, R., Li, X., 2017. Recovery of phosphorus and volatile fatty acids from wastewater and 2003. Off. J. Eur. Union 2019 1–114. food waste with an iron-flocculation sequencing batch reactor and acidogenic co- Fang, L., Wang, Q., Li, J., Poon, C.S., Cheeseman, C.R., Donatello, S., Tsang, D.C.W., fermentation. Bioresour. Technol. 245, 615–624. https://doi.org/10.1016/j. 2021. Feasibility of wet-extraction of phosphorus from incinerated sewage sludge biortech.2017.08.199. ash (ISSA) for phosphate fertilizer production: a critical review. Crit. Rev. Environ. Li, T., Wang, H., Dong, W., Liu, T., Ouyang, F., Zhang, Q., Dong, X., 2014. Performance of Sci. Technol. 51, 939–971. https://doi.org/10.1080/10643389.2020.1740545. an anoxic reactor proposed before BAF: effect of ferrous sulfate on enhancing Franz, M., 2008. Phosphate fertilizer from sewage sludge ash (SSA). Waste Manag. 28, denitrification during simultaneous phosphorous removal. Chem. Eng. J. 248, 1809–1818. https://doi.org/10.1016/j.wasman.2007.08.011. 41–48. https://doi.org/10.1016/j.cej.2014.03.033. Gebremariam, S.Y., Beutel, M.W., Christian, D., Hess, T.F., 2011. Research advances and Liberti, L., Boari, G., Passino, R., 1979. Phosphates and ammonia recovery from challenges in the microbiology of enhanced biological phosphorus removal-a critical secondary effluents by selective ion exchange with production of a slow-release review. Water Environ. Res. 83, 195–219. https://doi.org/10.2175/ fertilizer. Water Res. 13, 65–73. https://doi.org/10.1016/0043-1354(79)90256-2. 106143010X12780288628534. Liberti, L., Petruzzelli, D., De Florio, L., 2001. REM NUT ion exchange plus struvite Guerrero, J., Guisasola, A., Baeza, J.A., 2011. The nature of the carbon source rules the precipitation process. Environ. Technol. 22, 1313–1324. https://doi.org/10.1080/ competition between PAO and denitrifiers in systems for simultaneous biological 09593330409355443. 13 C. Zhang et al. Water Research 212 (2022) 118102 Liu, J., Cheng, X., Qi, X., Li, N., Tian, J., Qiu, B., Xu, K., Qu, D., 2018. Recovery of Remy, C., Jossa, P., 2015. Life Cycle Assessment of selected processes for P recovery from phosphate from aqueous solutions via vivianite crystallization : thermodynamics and sewage sludge, sludge liquor, or ash. Deliverable 9.2 of P-REX project. Sustainable influence of pH Abstract: Chem. Eng. J. https://doi.org/10.1016/j.cej.2018.05.064. sewage sludge management fostering phosphorus recovery and energy efficiency. Lizarralde, I., Fernández-Arévalo, T., Manas, A., Ayesa, E., Grau, P., 2019. Model-based Rittmann, B.E., Mayer, B., Westerhoff, P., Edwards, M., 2011. Capturing the lost opti mization of phosphorus management strategies in Sur WWTP. Madrid. Water phosphorus. Chemosphere 84, 846–853. https://doi.org/10.1016/j. Res. 153, 39–52. https://doi.org/10.1016/j.watres.2018.12.056. chemosphere.2011.02.001. Luff, B.B., Reed, R.B., 1980. Thermodynamic properties of magnesium potassium Robles, Á., Aguado, D., Barat, R., Borrás, L., Bouzas, A., Giménez, J.B., Martí, N., orthophosphate hexahydrate. J. Chem. Eng. Data 25, 310–312. https://doi.org/ Ribes, J., Ruano, M.V., Serralta, J., Ferrer, J., Seco, A., 2020. New frontiers from 10.1021/je60087a028. removal to recycling of nitrogen and phosphorus from wastewater in the circular Lv, J.-.H., Yuan, L.-.J., Chen, X., Liu, L., Luo, .D.-C., 2014. Phosphorus metabolism and economy. Bioresour. Technol. 300, 122673 https://doi.org/10.1016/j. population dynamics in a biological phosphate-removal system with simultaneous biortech.2019.122673. anaerobic phosphate stripping. Chemosphere 117C, 715–721. https://doi.org/ Ronteltap, M., Maurer, M., Gujer, W., 2007. Struvite precipitation thermodynamics in 10.1016/j.chemosphere.2014.10.018. source-separated urine. Water Res 41, 977–984. https://doi.org/10.1016/j. Lv, J., Yuan, L., 2015. Effects of chemical phosphate precipitation in the sidestream watres.2006.11.046. process on biological phosphorus removal at the anaerobic stage in an Roy, E.D., 2017. Phosphorus recovery and recycling with ecological engineering: a anaerobic—Aerobic sequencing batch reactor. Desalin. Water Treat. 54, 3011–3019. review. Ecol. Eng. https://doi.org/10.1016/j.ecoleng.2016.10.076. https://doi.org/10.1080/19443994.2014.904819. Rufí-Salís, M., Brunnhofer, N., Petit-Boix, A., Gabarrell, X., Guisasola, A., Villalba, G., Ma, P., Rosen, C., 2021. Land application of sewage sludge incinerator ash for 2020. Can wastewater feed cities? Determining the feasibility and environmental phosphorus recovery: a review. Chemosphere 274, 129609. https://doi.org/ burdens of struvite recovery and reuse for urban regions. Sci. Total Environ. 737, 10.1016/j.chemosphere.2021.129609. 139783 https://doi.org/10.1016/j.scitotenv.2020.139783. Mainstone, C.P., Parr, W., Day, M., 2000. Phosphorus and River Ecology: Tackling Salehi, S., Cheng, K.Y., Heitz, A., Ginige, M.P., 2019. Simultaneous nitrification, Sewage Inputs. English Nature. denitrification and phosphorus recovery (SNDPr) - An opportunity to facilitate full- Maity, J.P., Hsu, C.M., Lin, T.J., Lee, W.C., Bhattacharya, P., Bundschuh, J., Chen, C.Y., scale recovery of phosphorus from municipal wastewater. J. Environ. Manage. 238, 2018. Removal of fluoride from water through bacterial-surfactin mediated novel 41–48. https://doi.org/10.1016/j.jenvman.2019.02.063. hydroxyapatite nanoparticle and its efficiency assessment: adsorption isotherm, Sartorius, C., von Horn, J., Tettenborn, F., 2012. Phosphorus recovery from wastewater- adsorption kinetic and adsorption Thermodynamics. Environ. Nanotechnol., Monit. expert survey on present use and future potential. Water Environ. Res. 84, 313–322. Manag. 9, 18–28. https://doi.org/10.1016/j.enmm.2017.11.001. https://doi.org/10.2175/106143012X13347678384440. Marcelino, M., Wallaert, D., Guisasola, A., Baeza, J.A., 2011. A two-sludge system for Sekine, M., Akizuki, S., Kishi, M., Toda, T., 2018. Stable nitrification under sulfide supply simultaneous biological C, N and P removal via the nitrite pathway. Water Sci. in a sequencing batch reactor with a long fill period. J. Water Process Eng. 25, Technol. 64, 1142–1147. https://doi.org/10.2166/wst.2011.398. 190–194. https://doi.org/10.1016/j.jwpe.2018.05.012. Martin, B.D., De Kock, L., Gallot, M., Guery, E., Stanowski, S., MacAdam, J., McAdam, E. Sendrowski, A., Boyer, T.H., 2013. Phosphate removal from urine using hybrid anion J., Parsons, S.A., Jefferson, B., 2018. Quantifying the performance of a hybrid anion exchange resin. Desalination 322, 104–112. https://doi.org/10.1016/j. exchanger/adsorbent for phosphorus removal using mass spectrometry coupled with desal.2013.05.014. batch kinetic trials. Environ. Technol. 39, 2304–2314. https://doi.org/10.1080/ Shewa, W.A., Dagnew, M., 2020. Revisiting chemically enhanced primary treatment of 09593330.2017.1354076. wastewater: a review. Sustainability 12, 5928. https://doi.org/10.3390/ Martin, B.D., Parsons, S.A., Jefferson, B., 2009. Removal and recovery of phosphate from su12155928. municipal wastewaters using a polymeric anion exchanger bound with hydrated Shi, J., Lu, X., Xu, Z., Fang, M., 2016. A novel anaerobic–anoxic/nitrifying-induced ferric oxide nanoparticles. Water Sci. Technol. 60, 2637–2645. https://doi.org/ crystallization sequence batch reactor (A 2 N-IC-SBR) process for enhancing 10.2166/wst.2009.686. phosphorus recovery and nutrient removal. Desalin. Water Treat. 57, 7358–7368. Mayer, B.K., Baker, L.A., Boyer, T.H., Drechsel, P., Gifford, M., Hanjra, M.A., https://doi.org/10.1080/19443994.2015.1015449. Parameswaran, P., Stoltzfus, J., Westerhoff, P., Rittmann, B.E., 2016. Total value of Shi, J., Lu, X., Yu, R., Zhu, W., 2012. Nutrient removal and phosphorus recovery phosphorus recovery. Environ. Sci. Technol. 50 https://doi.org/10.1021/acs. performances of a novel anaerobic-anoxic/nitrifying/induced crystallization process. est.6b01239. Bioresour. Technol. 121, 183–189. https://doi.org/10.1016/j.biortech.2012.06.064. Mcdowell, H., Gregory, T.M., Brown, W.E., 1977. Solubility of Ca5(P04)3OH in the Siciliano, A., Limonti, C., Curcio, G.M., Molinari, R., 2020. Advances in struvite System Ca(OH)2-H3P04-H20 at 5, 15, 25, and 37 ◦ C. J. Res. Natl. Bur. Stand. A. precipitation technologies for nutrients removal and recovery from aqueous waste Phys. Chem. 81A, 273–281. and wastewater. Sustain. https://doi.org/10.3390/su12187538. Mew, M.C., 2016. Phosphate rock costs, prices and resources interaction. Sci. Total Tchobanoglous, G., Stensel, H.D., Tsuchihashi, R., Burton, F., Abu-Orf, M., Bowden, G., Environ. 542, 1008–1012. https://doi.org/10.1016/j.scitotenv.2015.08.045. Pfrang, W., 2014. Wastewater Engineering: Treatment and Resource Recovery, 5th Nättorp, A., Remmen, K., Remy, C., 2017. Cost assessment of different routes for ed. Metcalf & Eddy, Inc. McGray-Hill Education. phosphorus recovery from wastewater using data from pilot and production plants. Tomei, M.C., Soria Pascual, J., Mosca Angelucci, D., 2016. Analysing performance of real Water Sci. Technol. 76, 413–424. https://doi.org/10.2166/wst.2017.212. textile wastewater bio-decolourization under different reaction environments. Ohlinger, K.N., Young, T.M., Schroeder, E.D., 1998. Predicting struvite formation in J. Clean. Prod. 129, 468–477. https://doi.org/10.1016/j.jclepro.2016.04.028. digestion. Water Res 32, 3607–3614. https://doi.org/10.1016/S0043-1354(98) Valverde-Pérez, B., Ramin, E., Smets, B.F., Plósz, B.G., 2015. EBP2R – an innovative 00123-7. enhanced biological nutrient recovery activated sludge system to produce growth Ottosen, L.M., Kirkelund, G.M., Jensen, P.E., 2013. Extracting phosphorous from medium for green microalgae cultivation. Water Res. 68, 821–830. https://doi.org/ incinerated sewage sludge ash rich in iron or aluminum. Chemosphere 91, 963–969. 10.1016/j.watres.2014.09.027. https://doi.org/10.1016/j.chemosphere.2013.01.101. van Loosdrecht, M.C.M., Brandse, F.A., de Vries, A.C., 1998. Upgrading of waste water Oubagha, N., Lemlikchi, W., Sharrock, P., Fiallo, M., Mecherri, M.O., 2017. treatment processes for integrated nutrient removal-the BCFS® process. Water Sci. Hydroxyapatite precipitation with hydron blue dye. J. Environ. Manage. 203, Technol. 37, 209–217. https://doi.org/10.2166/wst.1998.0359. 807–810. https://doi.org/10.1016/j.jenvman.2016.09.039. Van Vuuren, D.P., Bouwman, A.F., Beusen, A.H.W., 2010. Phosphorus demand for the Parsons, S.A., Smith, J.A., 2008. Phosphorus removal and recovery from municipal 1970–2100 period: a scenario analysis of resource depletion. Glob. Environ. Chang. wastewaters. Elements 4, 109–112. https://doi.org/10.2113/GSELEMENTS.4.2.109. 20, 428–439. https://doi.org/10.1016/j.gloenvcha.2010.04.004. Petruzzelli, D., De Florio, L., Dell’Erba, A., Liberti, L., Notarnicola, M., Sengupta, A.K., Wang, D., Tooker, N.B., Srinivasan, V., Li, G., Fernandez, L.A., Schauer, P., Menniti, A., 2003. A new phosphate-selective sorbent for the Rem Nut(r) process. Laboratory Maher, C., Bott, C.B., Dombrowski, P., Barnard, J.L., Onnis-Hayden, A., Gu, A.Z., investigation and field experience at a medium size wastewater treatment plant. 2019. Side-stream enhanced biological phosphorus removal (S2EBPR) process Water Sci. Technol. 48, 179–184. https://doi.org/10.2166/wst.2003.0048. improves system performance - A full-scale comparative study. Water Res. 167, Pettersson, A., Åmand, L.E., Steenari, B.M., 2008. Leaching of ashes from co-combustion 115109 https://doi.org/10.1016/j.watres.2019.115109. of sewage sludge and wood-Part I: recovery of phosphorus. Biomass Bioenergy 32, Weigand, H., Bertau, M., Hübner, W., Bohndick, F., Bruckert, A., 2013. RecoPhos: full- 224–235. https://doi.org/10.1016/j.biombioe.2007.09.016. scale fertilizer production from sewage sludge ash. Waste Manag. 33, 540–544. Petzet, S., Peplinski, B., Cornel, P., 2012. On wet chemical phosphorus recovery from https://doi.org/10.1016/j.wasman.2012.07.009. sewage sludge ash by acidic or alkaline leaching and an optimized combination of Wilfert, P., Dugulan, A.I., Goubitz, K., Korving, L., Witkamp, G.J., Van Loosdrecht, M.C. both. Water Res. 46, 3769–3780. https://doi.org/10.1016/j.watres.2012.03.068. M., 2018. Vivianite as the main phosphate mineral in digested sewage sludge and its Pinelli, D., Bovina, S., Rubertelli, G., Martinelli, A., Guida, S., Soares, A., Frascari, D., role for phosphate recovery. Water Res 144, 312–321. https://doi.org/10.1016/j. 2022. Regeneration and modelling of a phosphorous removal and recovery hybrid watres.2018.07.020. ion exchange resin after long term operation with municipal wastewater. Williams, A.T., Zitomer, D.H., Mayer, B.K., 2015. Ion exchange-precipitation for nutrient Chemosphere 286, 131581. https://doi.org/10.1016/j.chemosphere.2021.131581. recovery from dilute wastewater. Environ. Sci. Water Res. Technol. 1, 832–838. Priambodo, R., Shih, Y.J., Huang, Y.H., 2017. Phosphorus recovery as ferrous phosphate https://doi.org/10.1039/C5EW00142K. (vivianite) from wastewater produced in manufacture of thin film transistor-liquid Wong, P.Y., Cheng, K.Y., Kaksonen, A.H., Sutton, D.C., Ginige, M.P., 2013. A novel post crystal displays (TFT-LCD) by a fluidized bed crystallizer (FBC). RSC Adv. 7, denitrification configuration for phosphorus recovery using polyphosphate 40819–40828. https://doi.org/10.1039/c7ra06308c. accumulating organisms. Water Res. 47, 6488–6495. https://doi.org/10.1016/j. Prot, T., Wijdeveld, W., Eshun, L.E., Dugulan, A.I., Goubitz, K., Korving, L., Loosdrecht watres.2013.08.023. Van, M.C.M., 2020. Full-scale increased iron dosage to stimulate the formation of Wu, Y., Luo, J., Zhang, Q., Aleem, M., Fang, F., Xue, Z., Cao, J., 2019. Potentials and vivianite and its recovery from digested sewage sludge. Water Res. 182, 115911 challenges of phosphorus recovery as vivianite from wastewater: a review. https://doi.org/10.1016/j.watres.2020.115911. Chemosphere 226, 246–258. https://doi.org/10.1016/j.chemosphere.2019.03.138. Xia, C.-.W., Ma, Y.-.J., Zhang, F., Lu, Y.-.Z., Zeng, R.J., 2014. A novel approach for phosphorus recovery and no wasted sludge in enhanced biological phosphorus 14 C. Zhang et al. Water Research 212 (2022) 118102 removal process with external COD addition. Appl. Biochem. Biotechnol. 172, Zhang, L., Keller, J., Yuan, Z., 2009. Inhibition of sulfate-reducing and methanogenic 820–828. https://doi.org/10.1007/s12010-013-0575-6. activities of anaerobic sewer biofilms by ferric iron dosing. Water Res 43, Xu, M., Li, P., Tang, T., Hu, Z., 2015. Roles of SRT and HRT of an algal membrane 4123–4132. https://doi.org/10.1016/j.watres.2009.06.013. bioreactor system with a tanks-in-series configuration for secondary wastewater Zhang, Z., Wang, X., Wang, H., Zhao, J., 2018. Removal of Pb(II) from aqueous solution effluent polishing. Ecol. Eng. 85, 257–264. https://doi.org/10.1016/j. using hydroxyapatite/calcium silicate hydrate (HAP/C-S-H) composite adsorbent ecoleng.2015.09.064. prepared by a phosphate recovery process. Chem. Eng. J. 344, 53–61. https://doi. Xu, M., Xu, S., Bernards, M., Hu, Z., 2016. Evaluation of high density algal cultivation for org/10.1016/j.cej.2018.03.066. secondary wastewater polishing. Water Environ. Res. 88, 47–53. https://doi.org/ Zou, H., Lu, X., 2016. An innovative continuous flow BNR-IC process for nutrients 10.2175/106143015X14362865227599. removal and phosphorus recovery from synthetic and real domestic wastewater. Yang, Y., Shi, X., Ballent, W., Mayer, B.K., 2017. Biological phosphorus recovery: review J. Environ. Eng. Landsc. Manag. 21, 153–162. https://doi.org/10.3846/ of current progress and future needs. Water Environ. Res. 89 https://doi.org/ 16486897.2012.721784. 10.2175/106143017x15054988926424. Zou, H., Wang, Y., 2017. Optimization of induced crystallization reaction in a novel Yu, X., Tian, W., Deng, Y., Cai, Y., Wang, Y., Wang, Z., Li, J., Ma, J., 2021. Nutrient process of nutrients removal coupled with phosphorus recovery from domestic removal and phosphorus recovery performance of an anaerobic side-stream wastewater. Arch. Environ. Prot. 43, 33–38. https://doi.org/10.1515/aep-2017- extraction based enhanced biological phosphorus removal subjected to low dissolved 0037. oxygen. J. Water Process Eng. 42, 101861 https://doi.org/10.1016/j. Zou, H., Wang, Y., 2016. Phosphorus removal and recovery from domestic wastewater in jwpe.2020.101861. a novel process of enhanced biological phosphorus removal coupled with Yuan, Z., Pratt, S., Batstone, D.J., 2012. Phosphorus recovery from wastewater through crystallization. Bioresour. Technol. 211, 87–92. https://doi.org/10.1016/j. microbial processes. Curr. Opin. Biotechnol. 23, 878–883. https://doi.org/10.1016/ biortech.2016.03.073. j.copbio.2012.08.001. Zu, X., Nan, J., He, L., Liu, B., 2020. Effects of various side stream phosphorus recovery Zhang, C., Guisasola, A., Baeza, J.A., 2020. Achieving simultaneous biological COD and volume on the performance and microbial structures of mainstream biological phosphorus removal in a continuous anaerobic/aerobic A-stage system. Water Res. system. Environ. Technol. 1–10. https://doi.org/10.1080/09593330.2020.1736645. 190, 116703 https://doi.org/10.1016/j.watres.2020.116703. 15