Valorization of Golden Mussel Shells for Sustainable Phosphorus Recovery in Wastewater Treatment

Abstract

The golden mussel (Limnoperna fortunei) poses environmental and infrastructural challenges due to its ability to attach to various substrates and form dense colonies. These colonies are difficult to remove and threaten hydroelectric power stations, water treatment plants and fishing activities. However, the high calcium carbonate content of golden mussel shells (GMSs) presents an opportunity for phosphorus (P) recovery from wastewater, addressing both waste management and resource scarcity. This study evaluated the effectiveness of GS for P recovery from synthetic and real wastewater. Batch experiments were conducted to assess P recovery capacity under varying adsorbent dosages, pH levels, contact times and isotherm conditions (Langmuir, Freundlich and Temkin). Also, the chemical and physical analyses of GMSs were performed to elucidate the mechanisms of P recovery. The Freundlich isotherm model best describes the process, while the Langmuir model suggests a maximum recovery potential of approximately 59.9 mg P g⁻¹ of GMS, demonstrating a P recovery efficiency of up to 60.7% at a P concentration of 40–50 g L⁻¹ and a contact time of 3 h. Due to the predominance of negative charges, it was concluded that the precipitation was the major mechanism for P recovery by GS. This study highlights the potential of GMSs as a sustainable and low-cost material for phosphorus recovery in wastewater treatment, offering a promising solution for both waste valorization and environmental management contributing to a circular economy.

1. Introduction

The golden mussel (Limnoperna fortunei) is a bivalve mollusk native to freshwater environments in Southeast Asia and was accidentally introduced into South America in 1991 via ballast water from Asian ships, specifically in Rio de la Plata, Argentina [1,2]. In Brazil, the first recorded appearance of the species occurred between 1998 and 1999 in the Guaíba Lake Basin, Rio Grande do Sul. Its presence was reported in Itaipu Lake, Paraná River, in 2001, and later in the Ilha Solteira, Jupiá, and Porto Primavera hydroelectric plants, spreading to other regions of the country [3]. Since then, surveys indicate its occurrence in the states of Paraná, São Paulo, Mato Grosso do Sul, Minas Gerais and Goiás [4].

The capacity of this species to attach to different substrates promotes colony formation, leading to recurrent reports of damage to hydroelectric power plant cooling systems, water intake pipes, boat engines and fish farming net pens [3,5]. Furthermore, the increasing population of golden mussels in freshwater environments creates an environmental liability concerning the disposal of their shells. After removal from water bodies, the disposal of these residues in piles on land can be a source of contamination, as studies suggest that the shells have the potential to bioaccumulate Al, Zn, Cr and other inorganic contaminants [3,6]. Considering this, developing techniques to promote the reuse or proper disposal of these shells in the environment should be encouraged [7,8].

As a more environmentally sound alternative, there is the potential to recover phosphorus (P) from aqueous solutions through contact with golden mussels (GMs), given its composition rich in calcium, indicating the material’s potential for reducing the eutrophic power of wastewater in natural aquatic environments [9,10,11]. Recent studies emphasize the valorization of biogenic waste, such as cow bone, palm biochar and oyster shells, for P recovery through adsorption and precipitation mechanisms [12,13,14]. In this approach, combining chemical (i.e., precipitation) and physicochemical (i.e., adsorption) processes drives P removal. The former process arises from the addition of shells, rich in calcium carbonate, to liquid media containing phosphate, which would promote the formation of hydroxyapatite precipitate [Ca₁₀(PO₄)₆(OH)₂], a compound of great interest to the fertilizer industry [15,16]. On the other hand, the adsorption process relies on retaining P on the surface of the adsorbent material through chemical bonds or weak intermolecular attraction forces [17]. Commonly used to evaluate the adsorption capacity of the shells (or any other adsorbent material) at an equilibrium concentration, adsorption isotherms employ mathematical models such as Langmuir, Temkin and Freundlich [18].The scientific literature contains studies utilizing various alternative adsorbent materials for P recovery from wastewater [17]. These materials include rice husks [19], eggshells [14,20] cockle shells [21], iron-based metal–organic frameworks [22], cerium-based synthetic compounds [23], steel slag [24], biochar [16], water treatment plant sludge [25], among others. In turn, researchers have reported using mussel shells as a phosphate adsorbent medium in some experiments with synthetic aqueous solutions [9,11,26]. Additionally, three studies tested the efficiency of these shells in P recovery from wastewater—domestic sewage (treated or untreated) [27,28,29]; however, no studies were found on the specific application of Golden mussel shells for this purpose.

From this standpoint, successful reports of employing alternative solid materials for P recovery from wastewater are pivotal in fostering and broadening research endeavors to develop innovative and environmentally suitable solutions for environmental sanitation. Hence, through kinetic analysis and construction of adsorption isotherms, this study aimed to assess the efficacy of ground GMSs in recovering P from university sewage in batch experiments conducted under varied experimental conditions.

2. Methods

2.1. GMS Collection and Preparation

GMS was provided by the Itaipu Hydroelectric Power Plant (of the Itaipu Binacional company), located in Foz do Iguaçu, Paraná, Brazil. The Itaipu reservoir, situated on the Paraná River, encompasses an experimental aquaculture area dedicated to fish farming in net cages within the Bela Vista Biological Refuge area, Foz do Iguaçu (25°26′49.09″ S; 54°32′56.68″ W). The net cage with encrusted mussels (Figure 1) was retrieved from the reservoir and placed under solar exposure in an open area for approximately 15 d, longer than the average survival period of golden mussels out of water [30]. Subsequently, the mussels were removed from their shells using a plastic spatula. After drying in an oven at 60 °C for 4 h, the material was ground using a hammer mill and a blender, and sieved to 0.425 mm.

2.2. Wastewater Collection

The wastewater used in the experiments came from the wastewater (sewage) treatment plant of the Federal University of Lavras (WWTP), Lavras, Minas Gerais, Brazil. It included screening, Parshall flume, grease trap, upflow anaerobic sludge blanket (UASB) reactor, submerged aerated biological filter (SABF) and sand filter. Effluent from the sand filter, devoid of the filtering substrate layer and filled only with rounded pebbles (during the period of the experiment), was collected. Two collections were made a month, apart during a period of fewer users on campus, when there were no classes at the institution, thus with reduced flow to the WWTP. The samples were stored in glass bottles at 4 ± 1 °C. The bottled samples were exposed to room temperature before starting the recovery experiments (Section 2.4).

2.3. GMS Characterization

A batch of collected GMSs underwent nitric–perchloric digestion and subsequent chemical characterization for total solution contents of Cu, Mn, Zn, Fe, N, P, K, Mg, Ca, and S using flame atomic absorption spectrophotometry [31], except for P, which was quantified via the colorimetric method based on ascorbic acid [32]. The remaining ground shells were utilized in the P recovery assays.

Additionally, the sample’s neutralization power (NP), apparent specific mass, and pH values in water (pH_H2O) and KCl (pH_KCl) were determined [33,34]. The purpose of pH determination in KCl is to compare it with values obtained in water, enabling the determination of delta pH (∆pH) (Equation (1)). Negative ∆pH values indicate the predominance of negative charges, i.e., higher cation exchange capacity (CEC) of the material. Conversely, positive values indicate higher anion exchange capacity (AEC). Meanwhile, a null ∆pH indicates a balance between positive and negative charges. Since P in solution exists in the form of phosphates, its recovery should be conducted on materials containing predominantly positive charges on their surface, thus yielding a positive ΔpH (pH_KCl > pH_H2O) [33].

$$∆\mathrm{p}\mathrm{H}={\mathrm{p}\mathrm{H}}_{\mathrm{K}\mathrm{C}\mathrm{l}}-{\mathrm{p}\mathrm{H}}_{{\mathrm{H}}_2\mathrm{O}}$$

(1)

2.4. Batch P Recovery Experiments

Batch experiments were conducted using three different liquid solutions to assess the use of GMSs as a P adsorbent material: (i) effluent from the WWTP (1st collection) with pH 3.0 (after HCl addition); (ii) synthetic solution with 10 mgP L⁻¹ (similar concentration of WWTP effluent) and pH 7.0 prepared by diluting KH₂PO₄ in distilled water; and (iii) effluent from the WWTP (2nd collection) with natural pH (~6.8). The experiments were carried out in triplicates in 125 mL Erlenmeyer flasks, with 100 mL of aqueous medium (WWTP effluent or synthetic solution) and ground GMSs at concentrations of 0, 5, 10, 15, 20, 30, 40 and 50 g of shells L⁻¹. The reaction flasks were stirred at 80 rpm on a magnetic stirrer, and the solution temperature was constantly monitored. At defined time intervals of 30 min, 1, 2 and 3 h, 4 mL of supernatant were withdrawn from each flask and transferred to individual test tubes. After centrifugation of the tubes (10 min at 2500 rpm), the P concentration in the supernatant was measured by colorimetry [32].

For the construction of the isotherms, 0.5 g of ground GMSs and 10 mL of a synthetic solution based on KH₂PO₄ (5, 20, 50, 80, 100, 155, 215, 530, 980 and 1230 mgP L⁻¹) were added to 50 mL beakers. These adsorbate concentrations were determined based on studies assessing the maximum recovery capacity of P using steel slag powder [35] and clamshell [36]. The suspensions were placed on a vertical shaker at 80 rpm for 24 h, then centrifuged at 1500 rpm for 10 min. The supernatant was collected to quantify the P concentration in the equilibrium solution by colorimetry [32].

All batch experiments were conducted in triplicate, with mean values employed for result analysis. If the relative difference between replicates exceeded 5%, additional tests were carried out. The experimental data were fitted to nonlinear kinetics and isotherm models using the nlstools package in R [37]. To determine the best-fitting model in each scenario, the coefficient of determination (R²) and the Akaike information criterion (AIC) were evaluated [38].

3. Results and Discussion

3.1. Physical and Chemical Characteristics of the Shells

Table 1. Physical and chemical characterization of ground GMSs.

ρ	NP	pHH2O	pHKCl	∆pH	K	Ca	Fe	P	N	Mg	Cu	Zn	S	Mn
g cm−3	%	–	–	–	------------------------------ g kg−1 ----------------------------------
1.35	85.65	8.35	8.30	−0.05	0.1	379.3	3.06	0.6	9.7	0.2	0.012	0.016	0.1	0.053

Notes: ρ = specific mass; NP = neutralization power; pH_H2O = pH in water; pH_KCl = pH in KCl; ∆pH = delta pH (see Equation (1)).

presents the results of the physical and chemical characterization of the ground golden mussel shells used in the tests. The material exhibited a higher specific mass than that reported by [12] for oyster shells (1.03 g cm⁻³). Furthermore, it exceeded the range reported for dewatered sewage treatment plant sludge, which typically falls between 1.05 and 1.10 g cm⁻³ [39]. These comparisons suggest that GMSs constitute a dense waste product characterized by reduced porosity and compact composition, likely attributable to the presence of carbonates. The chemical composition analysis in Table 1 reveals calcium as the predominant constituent of the shells, accounting for the material’s alkaline pH. However, the pH value observed in this study is lower than that reported for the shells of other mussel species. In this regard, ref. [26] documented a pH in water for Mediterranean mussel shells of 9.4.

Additional analysis regarding the pH in solution of the shells can be conducted using the calculation resulting from Equation (1). With a higher pH in water than in a KCl-containing medium, the residue exhibits a negative ΔpH (−0.05). Therefore, the CEC surpasses the AEC of the material due to the predominance of negative charges on its surface [33]. Thus, while phosphate (i.e., anions) recovery may occur, it is unlikely that the adsorption be the predominant mechanism for promoting P recovery from wastewater by the GMSs [40]. Furthermore, the residue has an NP of 85.65% CaCO₃, indicating a high potential for acidity neutralization. In other words, it suggests that 100 g of ground GMSs may neutralize acidity equivalent to 85.65 g of CaCO₃ [34]. The methodology used for grinding and crushing the shells can yield different NP results. For instance, ref. [41] obtained an NP equivalent to 73.00% CaCO₃ after grinding the GMSs and sieving them to 2.0 mm. Thus, using a finer mesh sieve (<0.425 mm) led to higher NP in the present study. Conversely, ref. [36] found a higher NP (95.40% CaCO₃) in clamshell meal with a particle size of 0.3 mm, indicating a material with a larger surface area and, thus, more reactive. On the other hand, ref. [42] reported an NP of 16.48% CaCO₃ for blast furnace slag with a particle size of approximately 10 mm present in constructed wetlands (CWs), suggesting that both the chemical composition and particle size can influence the material’s neutralization capacity. The significance of NP analysis of residues lies, for example, in the potential use of alternative materials to correct the pH of acidic soils.

From another perspective, the high concentrations of Ca and Fe in the residue, measuring 379.3 and 3.06 g kg⁻¹, exceeded those found by [3], at 110.54 and 1.27 g kg⁻¹, respectively. However, ref. [3] ground golden mussels together with their shells, which may account for the differences in this study. Additionally, the chemical composition of the residue may vary among colonies, depending on the season, location, or even the presence of anthropogenic pollutant sources in the watercourse where they proliferated, given that mussels are filter feeders and have a high bioaccumulation capacity [6,43,44].

3.2. Recovery Kinetics, and Efficiency

The high concentration of CaCO₃ in the residue, along with its excellent neutralization capacity (high NP value), and higher CEC than AEC suggest that precipitation was the predominant mechanism in the tested solutions. Figure 2 shows the effect of adsorbent dose on P recovery from synthetic real wastewater (WWTP). The P concentration in both effluent samples was approximately 10 mg L⁻¹, with slight variation only in temperature of 23 °C for the first collection and 21 °C for the second.

P recovery efficiency was negligible at shell doses up to 10 g L⁻¹, Figure 2a. This could be due to the competing anions in WWTP compared to the synthetic effluent. P recovery in the WWTP began at 15 g L⁻¹ and reached its peak (60%) at 50 g L⁻¹ of GMSs. Higher P recovery could be achieved in acidic conditions due to the increased AEC of the shells [7,20,45,46]. For instance, [47] reported higher P recovery efficiency using mussel shells in a pH 1.5 medium. Two factors may have favored low P recovery at the first doses: (i) the sand filter of WWTP did not contain the filter substrate layer at the time of effluent collection, thus impairing the retention of solids released from the previous unit (i.e., SABF), which may have promoted the release of P into the low-pH solution; and (ii) GMSs contain about 0.6 g P kg⁻¹ (Table 1), which might have released additional P into the solution. Longer contact times and higher adsorbent concentrations improved P recovery [48,49,50,51,52].

Figure 2b, shows the P recovery over time at a GMS dose of 40 g L⁻¹. The highest P recovery efficiency was achieved after 3 h of mixing. The primary mechanism of P recovery could be explained by the precipitation of soluble phosphates with calcium from GMSs. As explained in Section 2.3, in solutions with higher pH, there may be a predominance of negative charges (∆pH negative—see Equation (1)—and consequently, higher CEC), which may have resulted in a lower potential for P recovery by adsorption. For a more in-depth discussion in future studies, it is recommended to quantify the point of zero charge (PZC), the pH at which there is no predominance of charges, of the residue.

For the synthetic solution (Figure 2b), where a medium temperature of 24 °C was recorded at the time of testing, there was a decrease in P recovery at the beginning of sample agitation, possibly due to the release of P present in the GMSs (see Table 1). Throughout the experimental time and with the increase in shell concentration in the medium, the recovery efficiency increased to up to 23%, with a contact time of 3 h and shell dosage of 50 g L⁻¹. When compared to the values found for the effluent at pH 6.8 (close to the pH 7.0 of the synthetic solution), the efficiency was lower, probably because the synthetic solution does not contain residues of Ca or Mg that may be present in effluents from WWTPs, since these ions react with phosphate and precipitate [53,54]. The expectation, however, was that the laboratory solution would present higher P recovery efficiencies for not containing interfering substances (competitors to recovery sites) as in the effluent samples. This set of factors, therefore, highlights that the precipitation mechanism played a relevant role in the P recovery in the conducted tests.

The fact that ground GMSs showed a reasonable capacity to remove P from treated sewage—about 60% under conditions commonly found in WWTPs (Figure 2b)—indicates that there is a potential use of this waste as a simple and low-cost technique for P recovery from sewage. Alternative solutions for utilizing the material for wastewater treatment can be explored in future research.

Ref. [13], for instance, used ground oyster shells as a supporting medium in a system integrating CWs and filters and reported P recovery from low-strength wastewater of up to 95.5%.

3.3. Recovery Isotherms

The equations derived from the Freundlich, Langmuir and Temkin models and their respective coefficients of determination appear in

Table 2. Equations fitted to the Freundlich, Langmuir and Temkin models.

Fitting Model	Equation Coefficients	Akaike Information Criterion (AIC)	Coefficient of Dettermination (R2)
Freundlich	$S=0.0442\ast C_e^{1/1.124 *}$	31.38	0.92
Langmuir	$S={\displaystyle \frac{{59.9\ast C}_e}{1+0.0004\ast C_e}}$	31.43	0.90
Temkin	$S={\displaystyle \frac{RT}{b_{T}0.0732}}\ast 1.014\ast C_e$	51.51	0.85

Notes: * C_e is the equilibrium concentration; T is the temperature (K); R is the universal gas constant 0.082 L atm mol⁻¹ K⁻¹.

. Additionally, Figure 3 illustrates the P recovery isotherms generated from these models. We assessed the isotherm equations and observed that the Freundlich model demonstrated a superior fit to the experimental data, exhibiting a lower AIC value. Furthermore, the shape analysis of the Freundlich model curve indicated favorable recovery, with a value of n less than 1 (n = 1/1.124) [55]. However, this model does not specify the maximum recovery capacity of the tested material. From this perspective, the Langmuir model revealed a maximum recovery capacity of approximately 59.9 mg P g⁻¹ and a recovery energy constant (K_L) of 0.0004 L mg⁻¹.

Table 3. Literature review on experiments involving P adsorption from aqueous media using mussel shells from different species.

Mussel Species	Pretreatment of the Shells	Aqueous Medium	Contact Time	Adsorbent Concentration in the Liquid Medium (g L−1)	Maximum Adsorption Capacity * (mgP g−1)	Maximum P Recovery Efficiency (%)	Reference
Mediterranean mussels (Mytilus galloprovincialis)	Grinding to <2 mm	Synthetic solution (46.5 mmolP L−1, pH 8.5, and 20 °C)	24 h	13.3	18.23	60	[26]
	Calcination				38.75	78
Not specified	Synthesis of nano-calcium hydroxide	Synthetic solution (20 mgP L−1, pH 7.5, and 21 °C)	10 min	0.08	–	99.3	[11]
Not specified	Grinding to 0.60–1.18 mm	Treated sewage (7 mgP L−1 and pH 7.66)	5 d	100	0.248	66.2	[29]
Not specified	Incorporation of Fe	Treated sewage (7 mgP L−1 and pH 7.1)	120 h	100	3.14	95.7	[28]
Not specified	Grinding to 1.18 mm	Synthetic solution	1.000 min	20	0.04	15	[9]
	Calcination		10 min		1.32	97
Blue mussel (Mytilus edulis)	Thermochemical calcination with KOH	Treated sewage (20 mgP L−1, pH 7.32, and 22 °C)	2 h	4	12.44	99	[27]
Golden mussel (Golden mussel)	Grinding to <0.425 mm	Treated university sewage (10 mgP L−1, pH 3.0, and 23 °C)	3 h	40–50	59.9	60	Present study

Note: * = value extracted from the Langmuir model-based curve from adsorption tests in a synthetic solution containing P.

presents the results obtained in studies on P adsorption using mussel shells from various species. Close to the values obtained in the present study, ref. [26] achieved adsorption of 18.23 mg P g⁻¹ and 60% recovery of P using Mediterranean mussel shells with a contact time of 24 h. This prolonged contact time can be justified by the larger particle size of the shells used (<2 mm compared to <0.425 mm in the present study), which reduces the available surface area for adsorption. Conversely, ref. [27] reported a lower adsorption capacity (12.44 mg P g⁻¹) for blue mussel shells, but the contact time (2 h) was shorter than that of the present study (3 h). After calcination of the mussel shells, ref. [26] recorded a 112.5% increase in adsorption capacity and an 18% increase in P recovery, while [9] achieved recovery rates of up to 97%. Moreover, advanced shell pretreatment strategies can result in higher P recovery efficiencies. For instance, refs. [11,27,28] achieved recovery rates above 90% by adopting strategies such as nano-calcium hydroxide synthesis, thermochemical calcination with KOH, and Fe incorporation.

It is important to emphasize that the results of P adsorption capacity and recovery presented in Table 3 may vary due to factors such as (i) mussel species, (ii) mussel collection time, (iii) material pretreatment strategy, (iv) operational conditions of the adsorption tests, e.g., contact time, adsorbent dosage, mixing speed, and (v) physical and chemical properties of the medium, e.g., temperature, pH, P concentration. Furthermore, it is worth noting that part of the P recovery in the present study may have originated from calcium phosphate precipitation, as GMSs have a high concentration of Ca in their composition (see Table 1). Given these conditions, it would be prudent to state that the isotherms in Figure 3 may have encompassed combined adsorption and precipitation mechanisms to promote P recovery from the aqueous medium.

4. Conclusions

The results reported in this study demonstrate that GMSs can reduce sewage’s eutrophication potential. The maximum values achieved in the experimental tests were approximately 59.9 mgP g⁻¹ retention capacity, according to the Langmuir model, and 60% P recovery during 3 h in contact with sewage at pH 3.0. The results are similar to others reported in the literature for mussel shells from other species. Additionally, the Freundlich model better fits the experimental data and indicates a favorable adsorption curve. The predominance of negative charges and the high neutralizing power of GMSs suggest that precipitation was likely the primary mechanism of P recovery rather than adsorption in natural conditions (without pH adjustment).

As suggestions for future work, evaluating the point of zero charge (PZC) of GMSs to better understand their P adsorption potential and quantify the difference between adsorption and precipitation mechanism of P recovery. Moreover, we suggest adopting pretreatment strategies for the material, such as calcination and metal incorporation, as these techniques can significantly increase their adsorption capacity and P recovery, even in non-acidified sewage.

In summary, this study demonstrates how alternative materials can be reused and applied to sewage treatment. In this regard, lab-scale experiments are necessary for the planning and optimizing of large-scale treatment units that allow for the adoption of realistic operating conditions, such as a reasonable dosage of shells and contact times close to the hydraulic retention times of sewage treatment plants.