Pilot-Scale Phosphate Recovery from Secondary ...

Environ. Process. DOI 10.1007/s40710-016-0139-1 O R I G I N A L A RT I C L E

Pilot-Scale Phosphate Recovery from Secondary Wastewater Effluents Kyriaki Kalaitzidou 1 & Manassis Mitrakas 1 & Christina Raptopoulou 2 & Athanasia Tolkou 2 & Panagiota-Aikaterini Palasantza 3 & Anastasios Zouboulis 2

Received: 7 December 2015 / Accepted: 16 February 2016 # Springer International Publishing Switzerland 2016

Abstract Earth’s phosphorus resources are being depleted at an alarming rate, while at the same time eutrophication caused by its uncontrolled disposal in surface waters is considered as a significant environmental problem. In order to achieve phosphate recovery from the secondary effluents of an urban wastewater (biological) treatment plant, the adsorption onto single iron (GFH, Bayoxide and FeOOH) and onto binary iron-manganese (AquAsZero) oxyhydroxides, as well as the ion exchange by using Purolite A200EMBCL resin, were investigated as post-treatment methods. Among them, laboratory batch experiments and dynamic Rapid Small Scale Column Tests (RSSCTs) evaluated AquAsZero, as the relatively better qualified material, presenting the higher efficiency. Based on these experimental results a pilotplant, utilizing AquAsZero, was constructed and operated, treating 200 L/h. The breakthrough curves of RSSCTs for AquAsZero showed an adsorption capacity of 33.6 mg PO43−/gads at the equilibrium concentration of 3 mg PO43−/L, whereas at pilot-scale application the respective breakthrough curve indicated a similar adsorption capacity (31.5 mg PO43−/gads). The regeneration process, by applying a NaOH solution at pH range 12.6–13, resulted in the efficient (>80 wt.%) phosphate desorption, which in turn allows the multiple reuse of adsorbent media. Subsequently, phosphate was recovered from the alkaline regeneration (concentrate) solution by precipitation with the appropriate Ca2+ addition, as the respective calcium salt (hydroxyapatite, HAP). Phosphate concentration in the finally collected amorphous (precipitated) solids from the laboratory scale experiments was around 51 wt.% and that of calcium was around Electronic supplementary material The online version of this article (doi:10.1007/s40710-016-0139-1) contains supplementary material, which is available to authorized users.

* Anastasios Zouboulis [email protected]

1

Department of Chemical Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece

2

Department of Chemistry, Aristotle University of Thessaloniki, Thessaloniki, Greece

3

AKTOR S.A., Wastewater Treatment Plant of Touristic Area of Thessaloniki BAINEIA^, N. Michaniona, Thessaloniki, Greece

K. Kalaitzidou et al.

19 wt.%, while the corresponding concentrations in the precipitated solids collected from the pilot-scale experiments were around 36 wt.% for phosphate and 33 wt.% for calcium. This high phosphate content of finally recovered solids indicates their potential utilization as efficient (alternative) fertilizers. Keywords Phosphate recovery . Secondary effluents . Adsorption . Desorption . Iron oxyhydroxides . HAP . Precipitation

1 Introduction Phosphorus (P) is considered as an essential element for all living organisms; however, accumulation of P in relatively high concentrations, that overpasses the sensitivity of receiving water bodies (e.g. >10 mg/L), could be detrimental for the aquatic life, human health and aquatic environment (Lalley et al. 2015). Since, eutrophication is considered as the major problem for surface (inland) waters, the legislative regulation regarding the maximum allowable concentration of phosphate in treated effluents of wastewater treatment plants is expected to be continuously stricter (Chamoglou et al. 2014). Phosphorus in wastewaters effluents is mostly present as ortho-phosphates (PO43−); therefore, the recovery of PO43− could be a viable resource for the alternative production of fertilizers (de-Bashan and Bashan 2004). Chemical precipitation is the leading supplementary technology, applied for the removal of phosphorus from the secondary (treated) effluents of a wastewater treatment plant, i.e. this process is usually applied as a post-treatment (tertiary) method. Biological phosphorus removal, or PO43− crystallization as struvite in certain side streams, such as sludge treatment, have also been established as important treatment processes for lowering the concentration of phosphates in the final (secondary) effluents of wastewater treatment plants. Nevertheless, the advanced phosphorus removal in order to approach near-zero residual levels cannot be achieved by the simple application of traditional biological treatment, crystallization and/or chemical precipitation processes. Therefore, alternative, more efficient and promising technologies are being developed, such as ion exchange or adsorption (Morse et al. 1998), presenting the additional advantage of phosphates’ recovery and potential reuse versus the simple removal. In particular, adsorption can offer several additional advantages, such as simple and continuous flow operation, compact facilities, easier handling of process-produced wastes, limited addition of chemicals and sludge production, lower labor cost, potential regeneration and reuse of adsorbent media, which in turn results in lower operational cost (Lalley et al. 2015). Various adsorbents and ion exchange resins have been examined for the removal of phosphates, such as waste materials or by-products, iron-based oxides and oxy-hydroxides, as well as other sorbents of different origin. It should be spotlighted, that all adsorbents used for the removal of arsenates (As (V)) could be effectively implemented also for the phosphate uptake, due to similar structure of these elements. Table 1 presents the application of several different adsorbents, as published in the literature and their phosphate adsorption capacities. Especially, the iron oxides have been found to exhibit strong affinity with phosphates. This affinity was mainly attributed to the specific exchange reaction of ligands on the surface of adsorbent, since the phosphate ions consistently being exchanged with the structural hydroxyl groups, existing on the surface (Zhong et al. 2007). This ion exchange process involves the replacement of one or more surface hydroxyl groups by phosphate ions, releasing the

Phosphate Recovery from Secondary Wastewater Effluents Table 1 Reported adsorption capacities of various sorbents applied for phosphates Adsorbent

Matrix

Capacity, pH (mg PO43−/g)

Tempera-ture Reference (°C)

Red mud

Distilled water

1.78

5.5

40

Huang et al. (2008)

Fly ash

Distilled water

32

9

25

Agyei et al. (2002)

Slag

Distilled water

60

9

25

Agyei et al. (2002)

Portland cement

Distilled water

75

9

25

Agyei et al. (2002)

Portland cement + slag

Distilled water

78

9

25

Agyei et al. (2002)

Portland cement + fly ash

Distilled water

63

9

25

Agyei et al. (2002)

Lewatit (FO36) + HFO

Milli-Q water

272

7.5

21

You et al. (2016)

La(III) zeolite adsorbent

Multi-component 24.6

6

20

Ping et al. (2008)

MgFe–Zr LDH @ magnetic particles

Wastewater

35

4.5

20

Drenkova-Tuhtan et al. (2013)

MOP700 Biochar

Wastewater

Chen et al. (2010)

S15-NN-Fe-0.5

1.2

–

25

62.1

5

35

Huang et al. (2013)

6

23

Sendrowski and Boyer (2013)

HAIX

Fresh urine

9.54

PuroliteFerrIX A33E

Distilled water

144

~7.4 24

ZBCS

Distilled water

15.13

4.8

–

Nur et al. (2014) Choi et al. (2014)

ACF-NanoHFO

Distilled water

12.86

2–6

RT

Zhou et al. (2012)

Modified Bayoxide E33

Milli-Q water

26.8

7

21

Lalley et al. (2015)

Modified Bayoxide E33/Mn

Milli-Q water

17.8

7

21


Modified Bayoxide E33/AgI

Milli-Q water

19.3

7

21


Modified Bayoxide E33/AgII Milli-Q water

28

7

21


Ferrihydrite

141

4.5

25

Wang et al. (2013)

Distilled water

Hematite

Distilled water

6.6

4.5

25

Wang et al. (2013)

Goethite

Distilled water

11.1

4.5

25

Wang et al. (2013)

Akaganéite

Distilled water

178.9

7

25

Deliyanni et al. (2007)

Zenith-N

Distilled water

4.166

7

25

Zamparas et al. (2012)

Zenith-Fe

Distilled water

11.151

7

25


Phoslock

Distilled water

11.159

7

25


Mg-Al calcined LDH

Distilled water

50

6

30

Das et al. (2006)

Hydrotalcite

Distilled water

141.9

8.6

25

Kuzawa et al. (2006)

Purolite

NSF water

1.6

7

20

This study

GFH

NSF water

34

7

20

This study

Bayoxide

NSF water

25

7

20

This study

FeOOH

NSF water

50

7

20

This study

AquAsZero

NSF water

60

7

20

This study

respective surface structural OH− species into the bulk solution. The affinity of phosphates with the iron oxy-hydroxides surfaces depends mainly on the complexing capacity of anions, due to exchange of ligands, as well as on the attractive electrostatic interactions between the charged surfaces. Generally, the iron-based adsorbents, such as the iron oxy-hydroxides goethite (α-FeOOH), akaganéite (β-FeOOH) and lepidocrocite (γ -FeOOH), are considered as very useful due to their economic (low production cost) and safety characteristics. It has


also been reported that other iron oxy-hydroxides, such as schwertmannite, have a tunnel structure similar to that of akaganéite, which favor arsenate (and in turn phosphate) uptake, due to ion exchange with sulfate (Bigham et al. 1996; Tresintsi et al. 2012). Recently, the use of polymeric materials/organic resins is emerging as ion exchange materials, since they offer durability, desired mechanical strength and have the ability to remove specifically several anions from water. Because of their similar chemical properties and dissociation constants, the commercially available resins that are commonly used for arsenic removal can also be practiced for the removal of phosphates. However, their main disadvantage, regarding phosphates uptake, is that most resins luck the property of specific selectivity, especially when other competing ions (e.g. sulfates) are present in water or wastewater (Blaney et al. 2007; Wu et al. 2007). Recent research focuses on the combination of polymeric resins with iron oxide particles that creates a durable and reusable material, which is efficient for phosphates removal, because of the specific strong binding that occurs between iron oxides and phosphates (Sendrowski and Boyer 2013). This study aiming to investigate the simultaneous removal and recovery of phosphates from the secondary effluents of a wastewater (biological) treatment plant (WWTP). The major challenge of this study was to scale-up the appropriate removal/recovery treatment scheme of phosphorus, from the laboratory to the pilot-scale. Batch adsorption experiments, as well as Rapid Small Scale Column Tests (RSSCTs) were initially conducted. Based on these experimental results a pilot-plant was subsequently constructed to evaluate the adsorption/ regeneration process and was used on site to treat the effluents of a conventional WWTP.

2 Materials and Methods 2.1 WWTP The conventional WWTP of BΑΙΝΕΙΑ^ was selected in the present investigation, which is located near N. Michaniona, southern of Thessaloniki, Greece. AINEIA is a rather typical plant, treating about 8 × 103 m3/d of mostly urban, along with septage wastewaters. This WWTP consists of a combination of primary and secondary (biological) treatment processes, followed by ozone disinfection of secondary effluent. The treated effluent is discharged into the nearby sea, i.e. to Thermaikos Gulf. This effluent usually contains 12–17 mg PO43−/L, which exceeds the upcoming phosphorus discharge limit of 3 mg PO43−/L (i.e. 1 mg P-PO43−/L) that is expected to be applied in the near future. However, it must be underlined that the current disposal limit for phosphates in the case of AINEIA is 36 mg PO43−/L (12 mg P- PO43−/L). The main chemical characteristics of WWTP BAINEIA^ secondary effluents are presented in Table 2.

2.2 Adsorbent Media The examined adsorbents in this study, regarding their phosphate adsorption efficiency, were either commercially available or laboratory synthesized, iron-based oxy-hydroxides (FeOOH), selected according to their physic-chemical characteristics and previous experience. The evaluated commercially available adsorbents were the iron oxy-hydroxides GFH and Bayoxide, which nominally consist of akaganeite and goethite Fe-forms, respectively, the binary iron/manganese oxy-hydroxide (Fe-MnOOH) AquAsZero, nominally consisting of the

Phosphate Recovery from Secondary Wastewater Effluents

Table 2 Chemical characteristics of WWTP BAINEIA^ secondary effluent

Parameter

Range

pH

7.4 ± 0.3

Conductivity Hardness

4400 ± 200 μS/cm 750 ± 50 mg CaCO3/L

HCO3−

500 ± 50 mg/L

NΟ2−

0.45 ± 0.1 mg/L

ΝΗ4+

0.6 ± 0.2 mg/L

PO43−

12-17 mg/L

Ca2+

120 ± 10 mg/L

Mg2+

100 ± 10 mg/L

Na+ K+

585 ± 10 mg/L 37 ± 5 mg/L

feroxyhyte of Mn(IV) (Tresintsi et al. 2013), as well as the organic resin (used for the ion exchange study) PuroliteA200 EMBCL. Moreover, the laboratory synthesized single iron oxy-hydroxide (FeOOH), nominally consisting from schwertmannite (Tresintsi et al. 2012), was also investigated. This was synthesized by the oxidation-precipitation of FeSO4·H2O, using H2O2 in a two-step continuous flow process, as described in detail by Tresintsi et al. (2012). Table 3 summarizes the main physic-chemical parameters of examined adsorbents, which are expected to influence the removal efficiency for PO43−.

2.3 Chemical Reagents Α 300 mg/L PO43− stock solution was prepared by diluting 429.5 mg anhydrous KH2PO4 in 1 L of distilled water. Working standard solutions, used for laboratory preliminary experiments, were accordingly prepared by proper dilution of stock solution. The pH values of phosphate solutions in distilled water were adjusted either by NaOH or HCl addition, while 2 mM N,N-Bis(2-hydroxyethyl)-2-aminoethanesulfonic acid (BES) was added to facilitate pH control. The natural test water was prepared, according to National Sanitation Foundation (NSF) standard and contained 252 mg NaHCO3, 12.14 mg NaNO3, 0.178 mg NaH2PO4 H2O, 2.21 mg NaF, 70.6 mg NaSiO3·5H2O, 147 mg CaCl2·2H2O and 128.3 mg MgSO4·7H2O dissolved in 1 L of distilled water (National Sanitation Foundation International Standard, 13/ 4/2015) and used for certain laboratory experiments. Calcium stock solution was prepared by the dilution of proper amount of CaCl2 in distilled water. Table 3 Main physic-chemical characteristics of examined adsorbents for phosphates Adsorbent

Fe (wt.%)

Mn (wt.%)

ΒΕΤ (m2/g)

IEP (mV)

ZPC (mV)

Ctotal (mmol[ΟΗ−]/g)

FeOOH AquAsZero

44.9 38

– 11.5

53 187

7.2 7.35

2.9 3.2

3.2 2.7

GFH

54.2

–

237

7.15

5.2

0.9

Bayoxide

52

–

135

7.4

7.8

0.3


2.4 Laboratory Scale Experiments Several batch adsorption experiments, as well as Rapid Small Scale Column Tests (RSSCTs) were conducted to study the removal of PO43− at the temperature of 20 ± 1o C and pH = 6–8, by using distilled, NSF water, as well as the secondary effluent from BAINEIA^ WWTP and applying the aforementioned iron-based adsorbents. These adsorbents were evaluated, according to their adsorption capacities in repeated cycles of adsorption/regeneration operation, as well as according to their mechanical durability and stability. Batch adsorption experiments were initially carried out in order to record the respective isotherms for the preliminary evaluation of adsorbent efficiency. An amount of 15–60 mg of fine powdered adsorbent samples was dispersed in 200 mL of phosphate-containing solutions in 300 mL conical flasks. The flasks were placed in an orbital shaker and stirred for 24 h at 20 oC. The examined pH range was 6–8, which is commonly encountered in most wastewaters. For the RSSCTs experiments, glass columns of 1.1 cm diameter and 50 cm length with the appropriate polytetrafluoroethylene (PTFE) valves and caps and a glass frit in the bottom of each column, were used. The treated effluent from the WWTP was fed to this column from the top with a dosing pump at flow rate of 380 mL/h, resulting in an Empty Bed Contact Time (EBCT) of 3 min at 20o C. The columns were filled with oxy-hydroxide adsorbents at the height of 20 cm. The examined wastewater was collected from the secondary effluent of WWTP BΑΙΝΕΙΑ^, which as mentioned, presented a concentration range of 12–17 mg PO43−/L. The samples were regularly collected from the effluent of RSSCTs and analyzed for the residual phosphate concentration. As the main criterion for the evaluation of examined adsorbents was set the adsorption capacity of them at the equilibrium concentration equal to the regulation limit for wastewaters disposal in environmentally sensitive areas, i.e. 3 mg PO43−/L, or 1 mg P-PO43−/L, which was set according to the respective Council of the European Communities (1991) and is abbreviated as Q3, henceforth. The regeneration by using NaOH solution is considered as a common practice for the case of iron based (FeOOH) spent adsorbents (Kunaschk et al. 2015; Tresintsi et al. 2014). The onsite regeneration of adsorbent can lower the operating cost of the process by reusing the adsorbent, while at the same time, the higher pH value, as well as the higher concentration of phosphates in the regeneration solution can favor their subsequent potential recovery as calcium or magnesium salts. Based upon relevant preliminary experiments, the regeneration of adsorption columns was achieved by pumping at up-flow configuration 380 mL/h of 0.015 N NaOH solution (pH = 12.5 ± 0.1) for the adsorbents Bayoxide, AquAsZero and FeOOH, and 1 N NaOH solution (pH = 13 ± 0.1) for the case of GFH. The desorbed (leached) phosphates in the regeneration solution were subsequently precipitated (batch process) as calcium salt by the addition of Ca2+ solution.

2.5 Pilot-Scale Configuration Based on the experimental results of RSSCTs, a pilot-plant for the treatment of 100– 300 L/h was subsequently constructed to evaluate on site (AINEIA) the adsorption/ regeneration process to recover phosphates from the effluents of this WWTP. A detailed illustration of the overall process is presented in Fig. 1. The first step of the process is the filtration (Fig. 1a) of the secondary (treated) wastewater, by practicing commercially available hollow fiber membranes with pore size 0.4 μm. The removal of residual


(b)

(d)

(e)

(i) (g)

(f) (a) (c) (h)

Fig. 1 Pilot-plant located (and used) in BAINEIA^ WWTP. (a) Membrane filtration tank, (b) Adsorption bed, (c) NaOH and Ca2+ solution tanks, (d) NaOH and Ca2+ solution dosing pumps, (e) Pipe flocculation, (f) Precipitation/sludge storage tank, (g) Regeneration solution buffer tank, (h) Collection of precipitate, (i) Programmable Logic Controllers

suspended solids aimed to limit the adsorption process obstruction due to subsequent fouling of adsorption column. Then, the filtered wastewater is pumped to the adsorption bed/column (Fig. 1b) and the effluent of this bed is disposed of. When the adsorption bed is saturated, the process shifts to the regeneration mode. The regeneration solution from the buffer tank (Fig. 1g) is pumped (at up-flow configuration) to the adsorption bed and is being collected in another tank (Fig. 1f). Then, in the regeneration solution, which contains much higher phosphate concentrations in comparison with the secondary effluent, the calcium solution (Fig. 1c) is pumped (Fig. 1d), mixed afterwards with the recirculated sludge of phosphate from the bottom of tank (Fig. 1f) and flocculated by an appropriated in-pipe configuration (Fig. 1e). Moreover, pH is regulated by NaOH addition. The aforementioned sludge contains the precipitates/salts of phosphate (Fig. 1f), while the regeneration solution is filtered again through another set of hollow fiber membranes with pore size 0.4 μm and eventually, collected to the regeneration solution buffer tank for recycling/repeated use (Fig. 1g). The concentration of phosphate and calcium was monitored both in the regeneration solution and in the exit of the adsorption bed, so as the molar ratio of [Ca]/[PO43−] was regulated by calcium addition to 2.5. The adsorption column had a height of 2 m and a diameter of 0.184 m. The column was filled up to 1.20 m bed depth with AquAsZero granules (25 kg) of particle size of 0.2– 2 mm, which was selected due to its highest phosphate adsorption capacity, as observed during the preliminary lab-experiments. The separated precipitates/sludge of Ca-phosphates was further dehydrated by using a polyethylene bag filtration system, dried in the air and physic-chemically characterized.

2.6 Chemical Determinations The initial and residual concentration of phosphates during the adsorption process, as well as their concentration in the regeneration solution and in the precipitated Ca-phosphate salts (HAP) were determined colorimetrically by the stannous chloride method (Clesceri et al.


1989), using the spectrometer UV–VIS HITACHI U–5100. The concentration of calcium in the stock, as well as in the regeneration solution, was determined by flame atomic absorption spectroscopy (Perkin Elmer AAnalyst 800).

2.6.1 Characterization of Precipitated Ca-Phosphate Salts Dissolution A 0.25 g dry sample was placed in a 100 mL PTFE beaker, 20 mL of 5 N H2SO4 were added, heated on a hot plate and boiled for about 2 h until the residue had been completely dissolved. The solution was cooled and transferred to a 250 mL volumetric flask. Metals Metals concentrations were determined by atomic absorption spectrophotometry, using a Perkin Elmer AAnalyst800 instrument, either by flame or graphite furnace. Carbonate was determined in the solid samples by a volumetric-calcimeter method. Total Carbon (TC) The organic carbon of samples was determined by using a ThermoFinnigan Flash EA 1112 CHNS Analyser.

3 Results and Discussion 3.1 Adsorption Batch Experiments (Laboratory-Scale) Adsorption isotherms describe the relationship between the concentrations of phosphate in the solution and onto the surface of the adsorbent at a given temperature. They are usually evaluated through the fitting according to Langmuir or Freundlich equations. Langmuir model indicates the monolayer coverage of adsorption sites and is described by Eq. (1): Qe ¼

Qm K L C e 1 þ K LCe

ð1Þ

where: KL is the adsorption equilibrium constant; Ce and Qe are the phosphate equilibrium concentrations in the solution and onto the adsorbent surface, respectively; and Qm denotes the maximum adsorption capacity. The empirical Freundlich model, which assumes a multilayer coverage of a heterogeneous adsorbent’s surface with unlimited adsorption sites of unequal energies, is described by Eq. (2): 1 = Qe ¼ K F C e n

ð2Þ

where: Qe is the equilibrium phosphate adsorption capacity; Ce denotes the residual phosphate concentration at equilibrium; and KF and n are the Freundlich constants, related to the adsorption capacity and intensity, respectively.


3.1.1 Effect of pH Since pH is an important parameter that generally affects substantially the adsorption process, the phosphate adsorption capacity was studied at pH range 6 to 8 which is commonly encountered in most wastewaters. Figure 2 shows the significant effect of pH on the adsorption capacity of AquAsZero. Moreover, all examined adsorbents were found to exhibit similar behavior, regarding the pH effect (Online supplementary Resource 1, A1). Generally, phosphate adsorption onto the iron oxy-hydroxides is favored at lower pH values (Table 4), which is attributed to the fact that as pH increases: (a) ΗPO42− anion that requires two surface sites for adsorption dominates in aqueous solutions, instead of Η2PO4−, and (b) the concentration of OH− that strongly compete with PO43− for adsorption sites is also increased. In particular, by rising solution pH from 6 to 8 the adsorption capacity of AquAsZero was found to decrease by 9.8 mg PO43−/gads, i.e. more than 20 % (Table 4). Furthermore, the data of Table 4 indicate that the obtained adsorption data are better fitted to the Freundlich model, while it is noteworthy that the adsorption capacities at equilibrium concentration equal to the regulation limit of 3 mg PO43−/L (Q3) are similar to Qm values.

3.1.2 Effect of Water Matrix The adsorption isotherms are of great importance, since their shapes provide fundamental information about the adsorption capacity, as well as about the sorption mechanism. Figure 3 shows the adsorption isotherms for AquAsZero at pH 7 in deionized, as well as in NSF water matrix. It is obvious that the adsorption capacity of phosphates onto AquAsZero is strongly (and positively) affected by the presence of other ions commonly encountered in water matrix, while the adsorption data are better fitted to the Freundlich model, especially when using distilled water matrix, which indicates heterogeneous adsorption; noting also that similar results were observed for all the examined adsorbents. Table 5 demonstrates the fitted parameters for phosphate adsorption onto AquAsZero in distilled, as well as in NSF water matrices at pH 7. The Q3 value increases at NSF water matrix (57.4 mg PO43−/gads) by more than 30 %, as compared to the corresponding of distilled water matrix (43.3 mg PO43−/gads), which should be attributed to the presence of calcium in NSF water (Sperlich 2010). Finally, since adsorption capacity is influenced by the co-existing ions, Fig. 2 Adsorption isotherms of phosphate for the case of AquAsZero at pH values 6, 7 and 8, fitted by the application of Freundlich or Langmuir models (distilled water, T = 20 oC)

K. Kalaitzidou et al. Table 4 Fitting parameters to Freundlich and Langmuir models of phosphate adsorption onto AquAsZero at pH values 6, 7 and 8 (distilled water, T = 20 oC) Freundlich

Langmuir

pH

KF (mg PO43−/gads)/ (mg/L)1/n

1/n

R2

KL (L/mg PO43−)

6

43.4

0.063

0.981

7

38.2

0.114

0.981

8

29.3

0.206

0.983

4.7

Qm (mg PO43−/gads)

R2

Q3 (mg PO43−/gads)

87.2

44.6

0.825

46.5

10.5

43

0.933

43.3

37.4

0.953

36.7

the evaluation of adsorbents was further examined by using NSF water, as well as the secondary (treated) effluent during batch and dynamic (RSSCTs) experiments, respectively. Consequently, the adsorption isotherm data were acquired in NSF water matrix at pH 7, fitted to the Freundlich and Langmuir equations and the Q3 values were calculated. The data in Table 6 shows the following ranking of adsorbents, according to the respective Q3 value: AquAsZero (57.4 mg PO43−/gads), FeOOH (55.8 mg PO43−/gads), GFH (34.7 mg PO43−/gads), Bayoxide (mg 26.28 PO43−/gads) and Purolite (mg 0.5 PO43−/gads). Conclusively, AquAsZero showed the highest adsorption capacity and was evaluated as the qualified material. In contrast, the adsorption capacity of Purolite was found extremely low and any further evaluation of this material was cancelled.

3.2 Adsorption – Desorption Experiments Using RSSCTs (Laboratory-Scale) Bearing in mind the influence of water matrix on the uptake capacity of adsorbents, RSSCTs were subsequently implemented by using the secondary effluent from the WWTP BAINEIA^ in order to acquire certain information for the subsequent scale-up (pilot) process. The pH value of secondary effluent was 7.5 ± 0.3 and the concentration of phosphates ranged between 12 and 17 mg PO43−/L (Table 2). The breakthrough and the regeneration curves of RSSCTs for each adsorbent are presented in Fig. 4. The highest Q3 value (i.e. 34 mg PO43−/gads) was also achieved by the qualified adsorbent AquAsZero, which, however, was significantly lower to Fig. 3 Adsorption isotherms fitted by Freundlich and Langmuir models in distilled and NSF water matrices (T = 20 oC) for the case of AquaAsZero adsorbent

Phosphate Recovery from Secondary Wastewater Effluents Table 5 Fitting parameters for phosphate adsorption onto AquAsZero adsorbent in distilled and in NSF water matrices (pH 7, T = 20 oC) Freundlich

Langmuir

KF (mg PO43−/gads)/ (mg/L)1/n

1/n

R2

KL (L/mg PO43−)

Qm (mg PO43−/gads)

R2

Q3 (mgPO43−/gads)

0.114

0.981

10.5

43

0.933

43.3

0.255

0.984

3.3

60

0.973

57.4

Distilled water 38.2 NSF water 43.4

the corresponding value that was calculated from the adsorption isotherms (i.e. 57 mg PO43 /gads), when using the NSF water matrix. Similarly, significantly lower adsorption capacity was observed for all the other examined adsorbents: FeOOH (31 mg PO43−/gads), GFH (15 mg PO43−/gads) and Bayoxide (19 mg PO43−/gads). The Q3 value for Purolite was estimated to be 1.7 mg PO43−/gads (Online Supplementary Resource 1, A2). The phosphate desorption procedure was practically integrated within 3 h for all the examined adsorbents, as concluded from the results of Fig. 4. Preliminary experiments proved that 1 N NaOH solution (having pH 13.5 ± 0.1) is the optimum concentration for GFH regeneration, while 0.015 M NaOH solution (of pH 12.5 ± 0.2) was proved effective for the regeneration of all the other examined adsorbents. The uptake capacity of all adsorbents was step-wise lowered after each regeneration procedure, but remained satisfactory for at least three subsequent regeneration cycles. More specifically, the breakthrough curves of RSSCTs before and after the regeneration process (Fig. 4), showed that the adsorption capacity decreased by almost 20 % after the first regeneration cycle, e.g. the initial Q3 value of 34 mg PO43−/gads of AquAsZero was decreased to around 27 mg PO43−/gads. This loss of efficiency should be mainly attributed to the remaining phosphate content (2–4 mg PO43−/gads), which cannot be desorbed due to strongly bound onto iron oxy-hydroxides, as well as to other interfering parameters, such as the partial destruction of adsorption sites during the regeneration process. It is estimated that the latter contributes equally to the strongly bound phosphate, since the loss of adsorption capacity during the subsequent adsorption-regeneration cycles were (more or less) around 10 %. It must be spotlighted also the much higher phosphate concentration in the regeneration solution, which favors its subsequent recovery as calcium salt. −

Table 6 Fitting parameters for phosphate adsorption in NSF water matrix at pH 7 and 20o C for all the examined adsorbents Adsorbent

Freundlich

Langmuir

ΚF (mg PO43−/gads)/ 1/n (mg/L)1/n

R2

KL Qm R2 (L/mg PO43−) (mg PO43−/gads)

Q3 (mgPO43−/gads)

AquAsZero 43.4

0.255 0.984

3.3

60

0.953 57.4

FeOOH GFH

44.5 27.7

0.207 0.986 10.7 0.206 0.98 7

50 34

0.932 55.8 0.939 34.7

Bayoxide

22.6

0.133 0.984 16

25

Purolite

0.006

3.943 0.924

3.3

1.6

0.936 26.2 0.272

0.5


Fig. 4 Adsorption and desorption breakthrough curves of RSSCTs for (a) GFH, (b) Bayoxide, (c) FeOOH, and (d) AquAsZero


3.3 Recovery of Phosphates from the Regeneration Solution Preliminary experiments showed that in order to decrease the phosphate concentration below the regulation limit of 3 mg PO43−/L by applying precipitation as calcium salt, a water pH greater than 10 has to be applied (Fig. 5). Consequently, the pH of regeneration solution, which is greater than 12, favors the desorption of phosphate as well as the precipitation-recovery of calcium phosphate. However, for the effective phosphate precipitation a molar ratio [Ca2+]/[PO43−] around 2.5 should be also ensured. In the regeneration solution the appropriate volume of calcium stock solution was added under intensive stirring for 10 min. Afterwards, the suspension was stirred gently for 1 h for flocculation of sludge, which was subsequently separated by sedimentation. A typical composition of dried sludge is shown in Table 7. At this point the high phosphate content (50 wt.%) must be underlined, since it ensures its potential further utilization as fertilizer. The results from the CHNS Analyser indicate the low content of co-existing total carbon (i.e. less than 1 wt.%), which is also a significant parameter for the potential use of this precipitate as fertilizer. The XRD diagram (Fig. 6) indicates that the precipitate contains mainly Ca3(PO4)2 mH2O, which is a precursor of hydroxy-apatite (HAP) (Liu et al. 2001), as well as around 6 wt.% CaCO3.

3.4 Pilot-Plant Operation Based on the results of RSSCTs experiments, a pilot-plant, operating in continuous flow, was designed and constructed (Fig. 1) to treat 100–300 L/h of secondary effluent from the WWTP BAINEIA^. The treatment process of the pilot-plant differs from the corresponding of lab-scale experiments in two points: (1) the precipitate formed during the regeneration process was recirculated through an in-pipe flocculation mode, in order to be further enriched with phosphates and to increase the effectiveness of sludge separation; (2) the regeneration solution was reused after an appropriate membrane filtration treatment. Fig. 5 The residual phosphate concentration as a function of solution pH during the precipitation with calcium (initial concentrations: 40 mg Ca2+/L, 10 mg PO43−/L, experiments performed in distilled water)


Table 7 Chemical composition of calcium phosphate precipitate obtained from the RSSCTs

Main elements

Concentration wt.%

Ortho-phosphates

50 ± 1

Calcium Magnesium

12 ± 1 0.7 ± 0.2

Iron

4 ± 0.3

Total C