Hydroxyapatite-Based Adsorbent Materials from Aquaculture Waste for Remediation of Metal-Contaminated Waters: Investigation of Cadmium Removal

Abstract

Adsorption represents an effective strategy for water remediation applications, particularly when utilising eco-friendly materials in a circular economy framework. This approach offers significant advantages, including low cost, material availability, ease of operation, and high efficiency. Herein, the performance of cadmium ion adsorption onto hydroxyapatites, derived through a calcination-free process from shells of two mollusc species, Queen Scallop (Aequipecten opercularis) and Pacific Oyster (Magallana gigas), is examined. The phase and morphology of the synthesised adsorbents were investigated. The results showed that hydroxyapatites obtained from mollusc shells are characterised by high efficiency regarding cadmium removal from water, exhibiting rapid kinetics with equilibrium achieved within 5 min and high adsorption capacities up to 334.9 mg g⁻¹, much higher than many waste-based adsorbents reported in literature. Structural investigation revealed the presence of Cadmium Hydrogen Phosphate Hydrate in the hydroxyapatite derived from oyster shells loaded with Cd, indicating the formation of a solid solution. This finding suggests that the material not only has the capability to decontaminate but also to immobilise and store Cd. Overall, the results indicate that hydroxyapatites prepared via a synthetic route in mild conditions from waste shells are an economical and efficient sorbent for heavy metals encountered in wastewater.

1. Introduction

The introduction of toxic transition metals into the environment can arise from both natural and anthropogenic sources. Metal contamination is particularly relevant in freshwater ecosystems where it can be caused by atmospheric agents, or through the discharge of agricultural, domestic, and industrial waste [1]. Due to their toxicity and tendency to bioaccumulate, heavy metals represent a major concern for human health and ecosystems. They can cause adverse effects on plants, animals [2] and humans; indeed, toxic metals interfere with physiological processes, particularly hematopoietic, hepatic, and renal functions, and the nervous system [3]. Among heavy metals, cadmium (Cd) has an extremely long biological half-life, making it a cumulative toxic metal [4]. Furthermore, Cd has been identified as a human carcinogen by the World Health Organization’s International Agency for Research on Cancer and the United States [4]. Various strategies can be employed to remove metal pollutants from water [5]. Chemical precipitation is cost-effective, has low energy requirements and is safe and simple, but it produces high volumes of sludge, incurring associated disposal costs [5,6]. Coagulation and flocculation effectively reduce the settling time of suspended solids; however, the cost of sludge disposal remains a concern [6]. Membrane filtration is a highly efficient process that requires minimal operational space, but it is expensive and prone to fouling [5,6,7]. Electrochemical methods are highly efficient and selective processes, which do not produce sludge, but they are costly and characterised by high energy consumption [5,8]. Adsorption is another viable solution for the remediation of waters contaminated by metals [5,9,10], particularly when treating large volumes. It is characterised by simple design and operation, high efficiency and selectivity. Nonetheless, efficient adsorbents can be expensive [5], making it crucial to find economic and suitable alternatives. Therefore, developing solutions and evaluating adsorbent materials to remove pollutants within a circular economy context is a global challenge, and the valorisation of waste and biomass through recycling and transformation into new materials is a promising approach [11]. Recent studies on water remediation strategies offer various solutions for the reuse of waste materials to obtain eco-friendly adsorbents for heavy metals removal. Various bio-waste materials have been investigated including, for example, egg and mollusc shells [10,12,13], peanut shells [14,15], peanut shells biochar [16], and coconut husk [17]. Additionally, many calcium carbonate derived materials, such as apatite, have also been evaluated for the removal of Cd from aqueous matrix [18]. Among these materials, hydroxyapatite (HA, Ca₁₀(PO₄)₆(OH)₂) [19] is more stable than carbonates and is a low cost, largely available and biocompatible material [20]. Several studies have investigated the application of HA synthetised using pure reagents, or commercial HA for water remediation [21,22,23,24]. In the contest of a circular economy, biogenic calcium carbonate derived from seafood processing can indeed be valorised as a reagent for the synthesis of HA. However, recent literature, indicates that the synthesis of HA from biomass waste requires high temperatures to prepare the reagents through calcination [25,26,27,28], as well as a hydrothermal method for synthesis, both of which significantly impact the energy footprint of the process [28,29,30,31]. Well-known materials for the synthesis of hydroxyapatite (HA) are mollusc shells waste produced by aquaculture activities, which typically consists of 95–99% calcium carbonate (CaCO₃) by weight. These materials are widely available with an estimated global production of 10 million tons/annually and constitute critical wastes in areas where shellfish cultivation is a well-established activity. Indeed, aquaculture activities have to bear disposal costs to discard mollusc shells; the valorisation of this waste material through transformation into efficient adsorbents for water remediation can significantly reduce disposal costs [32]. Since shell morphology and composition play a critical role in influencing the interface properties and the chemical and mechanical behaviour [10,33] of the resulting material, HA prepared from shells of different species may show variations in the contaminant uptake capacity. For this reason, in the present study shells of two different species, namely Queen Scallop (Aequipecten opercularis) and Pacific Oyster (Magallana gigas), were evaluated as substrate for the synthesis under mild conditions of HA based adsorbent materials.

In this work, HA was synthesised from both raw and calcined shells (the latter requires an energetically demanding process). Cd adsorption experiments were conducted on both materials to compare the efficiency of the two adsorbents and to assess the possible use of the lower energy impact HA in water remediation applications. The mechanism of Cd uptake onto this novel promising material was investigated combining kinetic and isotherm adsorption experiments with structural and thermal analyses on pristine and Cd-loaded HAs. The features that influence the sorption capacity of the adsorbents were evaluated.

2. Materials and Methods

2.1. Materials

Pacific oyster (Magallana gigas) shells (OS), shown in Figure 1a, and Queen scallop (Aequipecten opercularis) shells (SS), depicted in Figure 1b, were provided from a shellfish farm and a fish products retailer located on the North Adriatic Sea coast (Naturedulis S.r.l., Goro (FE), Italy). The shells were thoroughly washed with ultrapure water (MilliQ^®, Merck KGaA, Darmstadt, Germany) and dried at 105 °C. The dried materials were then ground into a fine powder using a jade ball mortar (Retsch GmbH, Haan, Germany).

Cadmium, Calcium and Phosphorus standard solutions (1000 mg L⁻¹) for ICP were purchased from Carlo Erba Reagents S.r.l., Italy; Multielement standard solution (10–100 mg L⁻¹) for ICP, Sodium and Potassium standard solution (1000 mg L⁻¹) for AAS, were purchased from Merck, Darmstadt, Germany. Nitric Acid (69%, Suprapur^®); Cadmium nitrate tetrahydrate and Sodium phosphate dibasic were purchased from Merck, Darmstadt, Germany; Hydrochloric Acid (37%, Superpure) was obtained from Carlo Erba Reagents, Milan, Italy. Ultrapure water was used to prepare all solutions.

2.2. Preparation of the Adsorbents

2.2.1. Shells Calcination

The calcination is an endothermic reaction, as shown in Equation (1) [34], therefore is favoured by higher temperatures; in this work, powdered shells were calcined in a muffle at 900 °C for one hour.

CaCO₃ → CaO + CO₂ ΔH = 182.1 kJ mol⁻¹

(1)

The weight of both materials were monitored, and 69.1% by weight of calcined oyster shells (cOS) and 76.5% by weight of calcined scallop shells (cSS) were obtained.

2.2.2. Hydroxyapatite Synthesis

Hydroxyapatites (HAs) were synthesised from OS, SS, cOS and cSS following the method proposed by Zhang et al. [31]. 5 g of powdered shells were treated with 40 mL of 1.33 M Na₂HPO₄ solution to achieve a 1.67 Ca/P molar ratio; the mixture was kept under stirring at room temperature (22.0 ± 0.5 °C) for 48 h. Subsequently, the heterogeneous solution was filtered with ashless filter paper with pore size 4–12 μm (Whatman, Maidstone, UK). The obtained materials are hereafter referred as: HAOS (hydroxyapatite from oyster shells), HASS (hydroxyapatite from scallop shell), HAcOS (hydroxyapatite from calcined oyster shell) and HAcSS (hydroxyapatite from calcined scallop shell). All materials were dried in the oven at 105 °C for 24 h.

2.3. Characterisation

2.3.1. Scanning Electron Microscope and Transmission Electron Microscopy Measurements

Scanning Electron Microscopy (SEM) imaging was performed with a Zeiss Evo 40 electron microscope (Carl Zeiss AG, Oberkochen, Germany) equipped with Energy Dispersive x-ray Spectroscopy (EDS) (Oxford Instruments plc, Abingdon, UK). For SEM imaging acquisition the samples were coated with Au for 2 min (5–6 nm min⁻¹) with turbomolecular pump sputter coater system (Q150TS Plus, Quorum Technologies Ltd., Laughton, UK). The images were acquired using Backscattered electrons (BSE) and Secondary Electrons (SE) with an Electron High Tension (EHT) of 15 kVat magnifications from 1.00 K X to 11.00 K X.

Transmission electron microscopy (TEM) imaging was performed with Talos L120C (Thermo Fisher Scientific Inc., Waltham, MA, USA). The images were acquired at magnifications from 120 K X to 190 K X.

2.3.2. Particle Size Distribution Measurements

Particle size distribution of HAs samples were evaluated with a Mastersizer 3000 laser diffraction particle size analyser (Malvern Panalytical Ltd., Malvern, UK); the analyses were performed in triplicate using 0.2 g of sample for each replicate [35].

2.3.3. Chemical Characterisation

The shells’ composition was determined after sample mineralisation. 0.3 g of HASS and HAOS were weighed in TMF (tetrafluoro-methoxyl) vessels, added with 7 mL of HNO₃ and 3 mL of HCl and mineralised with a microwave-assisted digestion system (details can be found in Supplementary Materials, Experimental S1). Samples were mineralised in triplicate.

Trace and major elements were quantified by inductively coupled plasma mass spectrometry and atomic spectroscopy; the instrumental conditions are reported in Supplementary Materials (Experimental S1).

2.3.4. X-Ray Powder Diffraction (XRPD)

As synthesised and Cd loaded HAOS and HASS were ground in an agate mortar with a pestle for X-Ray Powder Diffraction (XRPD) analyses. The details of the instrumental conditions are described in Supplementary Materials, in the Experimental S1 Section.

Data elaboration was performed with Bruker AXS EVA software (version 5) and TOPAS (version 5.0). All diffraction peak profiles were modelled employing Lorentzian broadening coefficients in conjunction with shifted Chebyshev polynomials. In each Rietveld structure refinement (RSR), we refined several parameters, including the background polynomial, scale factor, unit cell parameters, zero-point correction, and site occupancy factors, while atom positions and isotropic thermal parameters were held constant. The weight percentage (wt %) of each Bragg diffracting phase was calculated based on the refined scale factors.

The identified phases were further refined using the scale factor, unit-cell parameters, crystallinity percentage, and crystallite size (in nanometres). The crystallite size was determined utilising the Scherrer equation applied to the Full Width at Half Maximum (FWHM) of the hydroxyapatite characteristic peak within the 25–26° 2θ range. The detected phases were refined starting from the following space groups: R-3c (ICSD code 20179) for calcite, Pmcn (ICSD code 15198) for aragonite, P_6₃/m (ICSD code 26204) for hydroxyapatite, and C 2/c (ICSD code 10045) for cadmium hydrogen phosphate hydrate.

2.3.5. Thermal Analysis

Thermogravimetric (TGA) and Differential Thermal Analysis (DTA) were conducted on both bare and Cd-loaded HAOS and HASS samples under constant and controlled synthetic air flow conditions, utilising a STA 409 PC LUXX^® instrument from Netzch (Gerätebau, Germany). The heating rate was maintained at 10 °C min⁻¹, with temperatures ranging from room temperature (RT) to 1100 °C. Approximately 30 mg of each sample was employed for the analyses, with alumina (Al₂O₃) serving as the control and standard reference.

2.4. Batch Adsorption Experiments

Adsorption experiments were performed using batch method [36] in 20 mL glass vials equipped with PP screw caps with PTFE septa. All the tests were conducted at pH 7 to prevent the precipitation of Cd²⁺ as cadmium hydroxide [37,38]. Additionally, the pH value selected for the adsorption tests is within the pH range typical of surface waters [39,40,41]. To determine the adsorption kinetics, experiments were carried out using solutions with initial Cd concentration of 10 mg L⁻¹; the solution was mixed with the adsorbent material with a solid/liquid ratio of 5:1 (mg mL⁻¹) and kept under continuous stirring at 25.0 ± 0.5 °C. Cd²⁺ uptake was measured after contact times equal to 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 30, 40, 50, 60, 90, 120, 150 and 180 min, after each contact time the solution was filtered with a 25 mm PVDF syringe filter with a pore size of 0.45 µm (Agilent Technologies, Santa Clara, CA, USA). The pH value was measured at the beginning and at the end of each experiment, the recorded values were in the range 7.0 ± 0.5. Equilibrium adsorption tests were performed by dispersing HA in Cd²⁺ aqueous solution at different initial concentrations (5, 20, 40, 60, 100, 125, 150, 175, 200, 250, 300, 350 and 400 mg L⁻¹) using a solid to liquid ratio of 1:1 (mg mL⁻¹). The solutions were maintained at 25.0 ± 0.5 °C under stirring for 24 h, a contact time adequately longer compared to the equilibration time evaluated by kinetic experiments. The solutions were then filtered, and the pH was measured for each batch. All experiments were carried out in triplicate. ICP-OES was employed to quantify Cd and Ca in solution before and after the contact with HAs for both kinetic and equilibrium adsorption experiments, with conditions reported in Section 2.3.5.

The amount of Cd²⁺ adsorbed per gram at equilibrium, defined as q (mg g⁻¹) was calculated with the mass balance Equation (2) [42]:

$$q={\displaystyle \frac{\left(C_0-C_e\right)V}{m}}$$

(2)

where C₀ (mg L⁻¹) is the initial solution concentration, C_e (mg L⁻¹) is the concentration at the equilibrium, V (L) is the solution volume of the batch and m (g) the mass of adsorbent.

3. Results and Discussion

3.1. Materials Characterisation

Unprocessed grounded shells of each species (SS, OS, cSS, cOS) and synthesised HAs (HAOS, HAcOS and HASS, HAcSS) were firstly characterised by SEM to obtain information on the morphology. Subsequently, the particle size distribution was evaluated for each sample of HAs synthesised. Structural investigation and elemental bulk composition of HAOS and HASS were carried out before the adsorption tests to support the results obtained; the results are reported in Section 3.3 and Supplementary Discussion S1.

3.1.1. SEM/EDS and TEM Imaging

Surface morphology of the synthesised HAs and unprocessed shells were analysed by scanning electron microscopy (SEM). The images collected for each material are reported in Figure 2 and Figures S1–S8. Specifically, HAOS (Figure 2a) and HASS (Figure 2b) exhibited particles characterised by irregular shape and variable size. Additionally, in Figures S1–S8 the elements distributions determined by scanning electron microscopy (SEM) equipped with energy dispersive X-ray spectrometry (EDS) are reported. The calcium is homogeneously distributed in each shell sample (Figures S1–S4), which is reasonable considering that calcium carbonate constitutes 95% of the mollusc shell weight [43]. Furthermore, the elemental maps of HAs (Figures S5–S8), showed a high concentration and homogeneous distribution of phosphorus, oxygen and calcium, which are the main components of HAs.

Transmission electron microscopy analyses were performed on all sythesised materials, the images were acquired on the smaller particles to obtain maximum resolution (Figures S9 and S10). Differences can be observed between the raw materials (i.e., CaCO₃ from SS and OS) and HASS and HAOS. Moreover, the comparison between HAs from raw powdered shells and HAs from calcined powdered shows that the materials synthesised from CaCO₃ present a more laminar structure, whilst those prepared from CaO have a more compacted structure.

3.1.2. Particle Size Distribution

Figure 3 shows the particle size distribution of each HA. The data obtained suggest that the particle size dimensions of HAs span over a broad range, with HAs from calcined shells more homogeneously distributed than HASS and HAOS, despite a similar average distribution (Dv 50) as shown in

Table 1. Particle size distribution of HASS, HAOS, HAcSS and HAcOS.

Parameters	HASS	HAOS	HAcSS	HAcOS
D [4;3] (µm)	22.389 ± 1.009	17.291 ± 0.765	13.519 ± 0.696	13.578 ± 0.779
D [3;4] (µm)	4.895 ± 0.145	6.588 ± 0.301	4.695 ± 0.325	4.883 ± 0.344
Dv (10) (µm)	1.913 ± 0.096	3.063 ± 0.153	2.017 ± 0.110	2.023 ± 0.125
Dv (50) (µm)	9.582 ± 0.379	10.113 ± 0.406	10.330 ± 0.617	9.506 ± 0.467
Dv (90) (µm)	67.212 ± 3.261	43.153 ± 1.878	29.910 ± 1.669	31.418 ± 1.171
Dv (98) (µm)	110.072 ± 5.306	78.134 ± 3.709	45.799 ± 2.156	51.669 ± 2.283
Dv (100) (µm)	162.552 ± 7.128	125.641 ± 6.582	66.712 ± 3.036	75.878 ± 3.294
Volume Below 1 μm (%)	2.93	0.75	3.21	2.67
Volume Below 10 μm (%)	51.26	49.54	48.66	52.02
Volume Above 45 μm (%)	16.80	9.30	2.22	3.64

. Furthermore, about 50% of each material is less than 10 µm.

3.2. Batch Adsorption Results

3.2.1. Effect of Calcination in Adsorption Experiments

To compare the adsorption properties of the adsorbent materials, preliminary tests were performed in triplicate at three different initial concentrations (5, 20 and 40 mg L⁻¹), following the method reported in paragraph 2.4. The results reported in Figure 4 show that all the materials exhibit similar adsorption quantities of Cd²⁺ in the examined concentration range. Since, HAs synthesised from both uncalcined mollusc shells have a lower environmental impact than those obtained from calcined shells, only the most sustainable materials, namely HASS and HAOS were evaluated in the following.

3.2.2. Kinetics Experiments

To gain information on the equilibrium time and the kinetics behaviour of the adsorption process, the Cd²⁺ uptake onto HASS and HAOS was investigated. Batch experiments were conducted as indicated in Section 2.4.

The experimental data collected were fitted using pseudo second order (PSO) kinetic model [10], given by Equation (3):

$$q_t={\displaystyle \frac{k_2 q_{eq }^{2}t}{1+k_2 q_{eq} t}}$$

(3)

where $q_t$ and $q_{eq}$ are the adsorbed quantity per unit mass after a specific contact time t (min) and at the equilibrium respectively, and $k_2$ (g mg⁻¹ min⁻¹) is the second order adsorption rate constant. The PSO model has already proved to be suitable to describe the adsorption of heavy metals [10]. The adsorption kinetics of HASS and HAOS are reported in Figure 5; it can be seen that for both materials the adsorption equilibrium was reached in about 5 min. The kinetic parameters obtained by nonlinear fitting regression of the experimental data using Equation (3) are reported in

Table 2. Parameters obtained by non-linear fitting of the uptake data using PSO model. The error is given as confidence interval at 95% of probability.

Materials	k2 (g mg−1 min−1)	qe (mg L−1)	R2
HASS	25.8 ± 1.6	2.0170 ± 0.0005	1
HAOS	189.9 ± 8.4	2.0870 ± 0.0001	1

. From the observed value of $k_2$, the uptake of Cd²⁺ onto HAOS is faster than HASS. Furthermore, HAs synthesised in this work showed faster kinetics compared to those of HAs reported in literature. For example, Li et al. [22] reported values of k₂ of 1.64 and 4.28 g mg⁻¹·min⁻¹ for HAs synthesised from pure reagents, while in Silva-Holguín et al. [21] HA supported on alumina spheres showed a k₂ of 0.012 g mg⁻¹·min⁻¹ [21]. Moreover, the obtained $k_2$ values are much higher compared to the PSO rate constants of the precursor materials; in particular, HASS and HAOS showed an extremely faster Cd²⁺ adsorption compared to SS [10] and OS [44,45]. Indeed, at the same Cd²⁺ initial concentration, Chenet et al. [10] reported a k₂ for SS of 0.87 g mg⁻¹·min⁻¹, [10] while for OS values of k₂ of 0.0015 and 0.018 g mg⁻¹·min⁻¹ are indicated [44,45]. Therefore, the HAs derived from mollusc shells enabled an improvement in the kinetics of the adsorption process when compared to the starting materials intended for the same application. This very fast Cd²⁺ uptake suggests that both HASS and HAOS are suitable materials for water remediation applications; the very short time required to reach the equilibrium indicates a suitability also for continuous flow systems [46].

3.2.3. Adsorption Isotherms

The adsorption capacities of the adsorbent materials were obtained by batch experiments. The experimental data collected were fitted using the Langmuir isotherm model, Equation (4):

$$q_e={\displaystyle \frac{q_s b{ C}_e}{1+b{ C}_e}}$$

(4)

where q_s (mg g⁻¹) is the saturation capacity, C_e (mg L⁻¹) is the Cd concentration at the equilibrium, and b (L mg⁻¹) is the Langmuir constant which represents the affinity between adsorbent material and adsorbate.

This isotherm model describes a homogeneous monolayer adsorption of the adsorbate onto the material surface, without adsorbate-solvent and lateral interactions between adsorbate molecules. The Langmuir model is suitable to describe adsorption isotherms characterised by a concave curve with a steep first interval followed by a plateau which represents the saturation point, q_s [42,47,48].

Isotherms data were fitted also with the Freundlich model, expressed by Equation (5):

$$q_e=k_F{C}_e^{{\displaystyle \frac{1}{n}}}$$

(5)

where k_F ((mg g⁻¹) (L mg⁻¹)^1/n) is the Freundlich constant which reflects the affinity between adsorbent and adsorbate and n is related to the sorption intensity, particularly, values of n > 1 indicate favourable adsorption conditions. The Freundlich isotherm model is suitable for describing heterogeneous adsorption with an increase of the uptake onto the adsorbent when the species concentration in the aqueous phase is increased [47,48].

Figure 6 shows the experimental data fitted with Equations (4) and (5), and the estimated parameters are reported in

Table 3. Isotherm parameters obtained from Langmuir and Freundlich isotherm models for HASS and HAOS, with confidence intervals at 95% of probability.

Isotherm	HASS	HAOS
Langmuir
qs (mg g−1)	173.70 ± 16.95	334.30 ± 46.20
b (L mg−1)	2.04 ± 1.58	0.31 ± 0.10
R2	0.9154	0.9443
Freundlich
kF (mg g−1) (L mg−1)1/n	112.90 ± 14.87	105.90 ± 12.19
n	10.63 ± 3.72	3.97 ± 0.26
R2	0.9577	0.9219

. The correlation coefficient, R², allows to compare the goodness of the fitting between models; the R² values obtained show that each isotherm model selected is suitable for describing the experimental data; however, Langmuir model describes better the adsorption onto HAOS, whereas the uptake onto HASS is better described by the Freundlich model.

The parameters obtained with the Freundlich isotherm model show n values higher than 1, particularly 10.63 ± 3.72 and 3.97 ± 0.26 for HASS and HAOS, respectively. These results suggest that adsorption is achieved under favourable conditions, confirmed by high K_F values of 112.90 ± 14.87 and 105.90 ± 12.19 for HASS and HAOS, respectively (Table 3). Indeed, in the literature there are many works that demonstrate the high affinity between metal cations and HA adsorbents. For example, Núñez at al. (2019) [49] obtained an K_F value of 15.86 (mg g⁻¹) (L mg⁻¹)^1/n for HA from clam shells, and Li et al. (2017) [22] found a K_F of 1.04 (mg g⁻¹) (L mg⁻¹)^1/n from HA synthesised from pure reagents.

Langmuir isotherm model allows to obtain the values of q_s (mg g⁻¹); the results indicate high saturation capacities for the materials investigated, especially for HAOS with a q_s of 334.3 ± 46.20 mg g⁻¹, which is higher than that obtained with other adsorbent materials reported in the literature (Table 4). Due to the biogenic origin of the starting materials used for the synthesis of HASS and HAOS, and the consequent complexity of the adsorbent structure which will be discussed in Section 3.3, the Cd adsorption may involve more than one mechanism.

A mechanism for the metal uptake onto hydroxyapatites was proposed by Marchat et al. [50], suggesting that the adsorption of Cd onto HA occurs through two mechanisms: adsorption and co-precipitation.

At the beginning, Cd²⁺ occupies individual sites on the HA surface through exchange with Ca²⁺ or by forming complexes. Then, residual Cd²⁺ may diffuse into the adsorbent structure or form a Cd containing HA phase [50]. Considering the adsorption results described above, the HAs synthesised from scallop and oyster shells are indeed very promising materials for water remediation in the case of heavy metals contamination.

Even though a specific cost analysis is beyond the objective of the present work, the high adsorption capacities and fast adsorption kinetics of HASS and HAOS indeed indicate their suitability for water remediation in a cost-effective way since the materials show very high saturation capacities compared to other materials.

Moreover, the high efficiency demonstrated by the HAs synthetised directly from mollusc shells and skipping the costly calcination step, suggest a lower synthesis cost since it utilises “zero-cost” raw materials; the major contributions to the total cost would be given by transportation and reagents [51].

Table 4. Adsorption capacity (q_s (mg g⁻¹)) towards Cd²⁺ of different adsorbent materials fitted with Langmuir model.

Adsorbents	qs (mg g−1)	References
Bamboo charcoal	12.08	[52]
Oyster shells (OS)	15.03	[45]
Cashew nutshell	21.11	[53]
Sugar cane	48.31	[54]
Scallop shells (SS)	55 ± 17.4	[10]
Orange peel	56.5 ± 0.50	[55]
Mussel shells HA	62.5	[25]
Clam shells HA	62.5	[49]
Small surface area HA	71.94	[22]
Pine bark biochar (600 °C)	85.5	[56]
Alumina-HA Spheres	89.37	[21]
Canna indica biochar (600 °C)	140.01 ± 14.42	[57]
HASS	173.7 ± 16.95	Present study
Large surface area HA	207.97	[22]
nano-HA	243.90	[18]
HAOS	334.3 ± 46.20	Present study

Table 4. Adsorption capacity (q_s (mg g⁻¹)) towards Cd²⁺ of different adsorbent materials fitted with Langmuir model.

Adsorbents	qs (mg g−1)	References
Bamboo charcoal	12.08	[52]
Oyster shells (OS)	15.03	[45]
Cashew nutshell	21.11	[53]
Sugar cane	48.31	[54]
Scallop shells (SS)	55 ± 17.4	[10]
Orange peel	56.5 ± 0.50	[55]
Mussel shells HA	62.5	[25]
Clam shells HA	62.5	[49]
Small surface area HA	71.94	[22]
Pine bark biochar (600 °C)	85.5	[56]
Alumina-HA Spheres	89.37	[21]
Canna indica biochar (600 °C)	140.01 ± 14.42	[57]
HASS	173.7 ± 16.95	Present study
Large surface area HA	207.97	[22]
nano-HA	243.90	[18]
HAOS	334.3 ± 46.20	Present study

3.3. Structural and Thermal Characterisation

The X-ray diffraction patterns of pristine and Cd-loaded HAOS and HASS are given in Figure 7. The pattern profiles are characterised by broad peaks, which can be attributed to the diffracting particles poor crystallinity and high defect density (crystallites). A close examination of the diffraction profiles present in Figure 7 reveals a pronounced anisotropy in the peak broadening.

Hydroxyapatite’s 001 reflections exhibit significantly sharper peaks compared to other reflections within the patterns. This striking anisotropy in the profile function of the hydroxyapatite samples indicates a presence of defects in the a, b planes and/or an anisotropic shape of the crystallites, with coherent domains that are elongated along the c axis. Such features are commonly observed in hydroxyapatite derived from biomass waste [58], suggesting that the coherent diffraction domains preferentially extend along the crystallographic c direction within the hydroxyapatite crystals.

Quantitative phase analysis (QPA) was conducted by the Rietveld method, wherein the weight fraction wi of all ith crystalline component in the multiphase system is calculated from the corresponding refined scale parameter Si, according to the Equation (6):

$$wi=\frac{SiMiVi}{jSjMjVj} with the normalisation condition \Sigma wi = 1.0$$

(6)

where Mi and Vi are the unit cell mass and volume, respectively.

Calcite was observed at approximately 23% weight, alongside hydroxyapatite (HA), as indicated in

Table 5. Crystallite size (nm) Crystallinity degree (%) and weighted amount (wt%) of phases obtained using RSR for bare and Cd-loaded HAOS and HASS samples.

	HAOS	HASS	Cd-HAOS	Cd-HASS
Calcite (wt%)	24.5 (4)	22.8 (2)	18.4 (4)	17.0 (4)
Aragonite (wt%)	-	4.2 (2)	-	4.2 (5)
Hydroxyapatite (wt%)	75.5 (4)	73.0 (3)	68.1 (5)	78.8 (6)
Cadmium Hydrogen Phosphate Hydrate (wt%)	-	-	13.5 (3)	-
Crystallite size (nm)	24.8	25.2	31.9	32.0
Crystallinity (%)	63.2	65.4	63.1	62.1

. This observation suggests an incomplete transition to HA in both the HAOS and HASS samples. Additionally, a low fraction of aragonite, approximately 4% weight, was detected in the HASS sample, also detailed in Table 5. The diffraction peaks corresponding to these secondary phases are distinctly marked in Figure 7.

Commonly, cadmium fixation in hydroxyapatite occurs in accordance with the following mechanisms [50,59]: Cd²⁺ adsorption on specific P-OH groups of HA surface, diffusion of the excess Cd²⁺ into the apatite structure, ion exchange, and heterogeneous dissolution and precipitation.

The exchange reaction impacts the structural integrity of hydroxyapatite and its chemical composition, leading to a solid solution incorporating Cd²⁺. This incorporation can alter the crystallinity and porosity of HA (

Table 6. Unit-cell parameters for bare and Cd-loaded HAOS and HASS samples.

Calcite CaCO3
R-3 c H	a = b (Å)	c (Å)	Volume (Å3)
HAOS	4.9989 (7)	17.111 (6)	370.3 (3)
Cd-HAOS	4.9927 (8)	17.082 (3)	368.8 (1)
HAAS	4.9917 (1)	17.0846 (4)	368.6 (1)
Cd-HAAS	4.9890 (4)	17.0746 (2)	368.0 (1)
Hydroxyapatite Ca10(PO4)6(OH)2
P 63/m	a = b (Å)	c (Å)	Volume (Å3)
HAOS	9.404 (5)	6.891 (4)	527.8 (6)
Cd-HAOS	9.455 (7)	6.717 (6)	520.1 (9)
HAAS	9.396 (1)	6.877 (1)	525.8 (3)
Cd-HAAS	9.400 (3)	6.805 (4)	520.7 (4)
Aragonite CaCO3
P mcn	a (Å)	b (Å)	c (Å)	Volume (Å3)
HAAS	4.968 (3)	7.962 (7)	5.752 (4)	227.5 (3)
Cd-HAAS	4.964 (2)	7.955 (3)	5.745 (2)	226.9 (1)
Cadmium Hydrogen Phosphate Hydrate Cd5H2(PO4)4(H2O)4
C 1 2/c 1	a (Å)	b (Å)	c (Å)	Volume (Å3)	β (°)
Cd-HAOS	17.9553 (1)	9.4258 (4)	9.7121 (4)	1632.9 (4)	96.57 (1)

), which are critical factors in its adsorption capacity. Additionally, the presence of carbonate ions in the solution can facilitate the formation of mixed carbonate-phosphate phases, further influencing the removal efficiency of cadmium.

Following cadmium adsorption, the cell volume of hydroxyapatite decreased in both HASS and HAOS. This reduction can be attributed to the smaller ionic radii of cadmium (0.95 Å and 1.10 Å for six- and eight-coordination, respectively) compared to calcium (1.00 Å and 1.12 Å for six- and eight-coordination). These findings indicate an exchange reaction between the two elements, resulting in the formation of a Ca_1−xCd_xCO₃ and Ca(_5−x)Cd_x(PO₄)₃ (OH) solid solution [59].

This process introduces strain into the HA network structure and increases the HA decomposition [60].

The dissolution of CaCO₃ and subsequent Cd adsorption on the surface progressively altered the regular rhombohedral etch pits of calcite, as cadmium ions pinned the acute angles. This process effectively passivated the calcite surface through the epitaxial growth of a less soluble (Ca,Cd)CO₃ layer [61,62]. After Cd²⁺ adsorption, the crystallite size increased in both HAOS and HASS samples, ranging from approximately 25 to 32 nm, while the crystallinity percentage remained nearly unchanged (Table 5).

In the Cd-treated HAOS sample, a new solid phase has been identified as Cadmium Hydrogen Phosphate Hydrate, Cd₅H₂(PO₄)₄·4H₂O, constituting approximately 13% in weight. This finding suggests a dissolution-precipitation mechanism involving hydroxyapatite. The formation of this new phase is influenced by multiple factors, including the liquid-to-solid ratio, pH, ionic strength, temperature, duration, concentration of the aqueous phase, and the presence of other ions. Additionally, the types and pathways of geochemical reactions play a significant role in this process.

The formation of this phase in Cd-treated HAOS could be due to the different composition of oyster and scallop shells [10,63] (see Discussion S1) and it could explain the higher saturation capacity of HAOS compared to HASS, despite the fact that roughly 4% of aragonite was detected in HASS, but not in HAOS. Aragonite, indeed, facilitates additional Cd sequestration during the Ca/Cd replacement reaction compared to calcite, leading to fracturing and interfacial stress, as well as the incorporation of a greater Cd content [61,64]. It has been already pointed out that biogenic carbonates showed higher adsorbent capacities than non-biogenic one’s due to structural and compositional features. The results of the present study also suggest that hydroxyapatites synthesised from different biogenic carbonates have different adsorption performances.

The thermal characterisation of bare and loaded samples of HAOS and HASS revealed comparable weight loss, approximately 20% by weight, at the maximum investigated temperature of 1100 °C. Throughout the heating process, several distinct steps were observed (see Figure 8 and Figure 9). Initially, the release of physically adsorbed water from hydroxyapatite occurs between 30 and 200 °C, associated with an endothermic reaction evident in the differential thermal analysis (DTA) curve.

At temperatures ranging from 200 to 400 °C, a second thermal event occurs, characterised by an exothermic effect in the DTA, indicating the loss of structural water from HA. As the temperature increases to approximately 600 °C, hydroxyapatite undergoes gradual dehydration, resulting in the release of hydroxide ions (OH⁻) and subsequently transforming into oxyapatite. Above this temperature, significant weight loss is primarily attributed to the thermal decomposition of residual CaCO₃ and potentially amorphous carbonate. According to the literature, calcite is recognised as the most thermodynamically stable polymorph of calcium carbonate under ambient conditions [65]. Consequently, the decarbonation of the less stable polymorph, aragonite, begins at around 450 °C, while the decomposition of calcite occurs at elevated temperatures ranging from 720 to 875 °C. It is important to note that the temperature at which the exothermic peak occurs can vary depending on the sample’s composition [66]. Furthermore, cadmium ions bound to HA and CaCO₃ can induce a shift in the decarboxylation and dihydroxylation reactions to lower temperatures [67]. The differential thermal gravimetry (DTG) curve of Cd-loaded hydroxyapatite (HASS) exhibited two prominent peaks at 430 °C and 600 °C. In contrast, the incorporation of Cd²⁺ into the hydroxyapatite structure, along with the mineralisation of Cd₅H₂(PO₄)₄·4H₂O, resulted in a shift of the curves to 368 °C, 460 °C, and 520 °C. This was accompanied by three exothermic reactions observed in the differential thermal analysis (DTA).

4. Conclusions

The Hydroxyapatites studied in this work were synthesised from not calcined shells and without any thermal process. The results showed higher saturation capacities towards the adsorption of Cd²⁺ onto HAs and faster kinetics than those reported for mollusc shells as adsorbents. The adsorption performances of the synthesised hydroxyapatites are strongly dependent on the type of the shell employed for the synthesis; indeed, saturation capacities of 173.7 ± 16.95 mg g⁻¹ and 334.3 ± 46.2 mg g⁻¹ were obtained for HASS and HAOS, respectively. These differences may be attributed to variations in the composition and structure of the starting materials, as evidenced from diffractometric and bulk composition characterisation. These findings suggest that hydroxyapatites from mollusc shells are suitable materials for water remediation applications.