A novel cost-effective and sustainable Zn2+-crosslinked alginate/nanohydroxyapatite (ZA/nHA) composite was developed based on the brown macroalga S. latifolium as a source of alginate and the calcified red macroalga T. fragilis as a material for hydroxyapatite synthesis (Fig. S1). The extracted alginate had a viscosity average molecular weight of 1.76 × 105 Da and mannuronic/guluronic acid ratio of 1.79. The pHpzc of the developed ZA/nHA beads was 6.55.

FT-IR analysis

The hydrogen bonded O–H and C-H stretching vibrations in the macroalgal-derived alginate were observed at 3427.52 and 2929.69 cm−1, respectively (Fig. 1a). The sharp bands centered at 1621.46 and 1411.24 cm−1 were attributed to the asymmetric and symmetric − COO− vibrations, respectively [30, 31]. While the C − C stretching vibration was located at 1452.25 cm−1. Furthermore, the stretching vibrations of C − O bonds were located at 1126.10 cm−1 and 1028.75 cm−1, while that at 1095.11 cm−1 indicated both C − O and C − C stretching vibrations [31]. The alginate spectrum exhibited a specific band of uronic acid at 947.44 cm−1 which was attributed to the C − O stretching vibrations (Fig. S1 b). Similarly, the bands at 904.65 and 839.97 cm−1 were assigned to the α-guluronic and β-mannuronic acid, respectively [31]. The band at 868.32 cm−1 was related to the C1-H deformation vibration of β-mannuronic acid [30].

Fig. 1

FT-IR spectra of (a) alginate from S. latifolium, (b) nanohydroxyapatite synthesized using T. fragilis biomass, (c) zinc alginate / nanohydroxyapatite composite, and (d) zinc alginate / nanohydroxyapatite composite after the adsorption of crystal violet

On the other hand, the FT-IR spectrum of the phyco-genic nHA (Fig. 1b) indicated the existence of strong and sharp peak centered at 1035.09 cm−1 (ν3), and two weak peaks at 1099.27 cm−1 (ν3) and 961.38 cm−1 (ν1) were attributed to PO43− ions (stretching vibration of the P − O − P bond). The splitting sharp peaks at 602.94 and 564.02 cm−1 (ν4) reflected bending vibration of the PO43− ions, which occupied two sites in the crystal lattice [32]. Furthermore, the asymmetric stretching vibration of PO43− was indicated by a weak peak at 472.01 cm−1 (ν2) [33]. The stretching vibrations of the OH− groups were assigned to the broad band at 3424.04 cm−1, with the band at 1640.71 cm−1 indicating the adsorbed or bounded H2O. Additionally, the presence of weak band at 873.79, 1413.89, and 1458.76 cm−1 may be assigned to CO32− anions substituting for the PO43− in the hydroxyapatite lattice (B-type hydroxyapatite) [34]. The carbonate that existed in the nHA was attributed to the adsorbed CO2 from the air, since the CaCO3 present in the seaweed biomass was solubilized using HCl during the first steps of nHA synthesis. Moreover, the existence of double sharp bands 564.02/602.94 and 1035.09/1099.27 cm−1 in the algal-derived n-HA reflected a high degree of crystallinity [33, 34].

The FT-IR spectrum of the ZA/nHA composite was depicted in Fig. 1c. New peaks located at 1624.48, 1423.72 and 1321.54 cm−1 were assigned to the COO− groups of alginate and the one at 817.11 cm−1 may be identified as the combination of three possible vibrational modes (τ(CO) + δ(CCO) + δ(CCH)) of Zn-alginate [16]. Moreover, the shift in the COO− groups in the composite compared to sodium alginate reflected the presence of electrostatic interactions and bond formation between the positively charged Ca2+ in the hydroxyapatite and COO− groups of alginate [35]. In general, the FT-IR analysis confirmed a successful utilization of macroalgae wastes as cost-effective and sustainable approach for the fabrication of nanocomposite based on alginate and hydroxyapatite.

XRD analysis

The XRD spectrum (Fig. 2a) of Na-alginate exhibited a distinctive peak between (2θ) 20° to 30°, which could be related to the amorphous nature of the polymer network [19, 36, 37]. The algal-derived nHA exhibited numerous peaks in the XRD spectrum (Fig. 2b), which were perfectly matched with standard card JCPDS: 01–074-0565 and indicates the formation of single phase of hydroxyapatite with the chemical formula Ca10(PO4)6(OH)2. Furthermore, the developed nHA showed hexagonal structure with a space group number 176 and a space group category P63/m. The average crystallite size of nHA was 41.88 nm based on the Debye–Scherrer equation [16]. This size is quite similar to the nHA prepared from other natural sources such as bone [17], and egg shell [38].

Fig. 2

XRD spectra of (a) alginate, (b) nanohydroxyapatite, and (c) zinc alginate/nanohydroxyapatite composite

On the other side, the spectrum of ZA/nHA beads exhibited characteristic peaks for both alginate and nHA (Fig. 2c). However, the spectrum of ZA/nHA indicated a shift and decrease in intensity of the major peaks for pure nHA and alginate, which implied the successful interaction between the nHA particles and alginate network.

TEM analysis

Figure 3a depicts the TEM of the algal-derived hydroxyapatite. It was observed that the nHA particles had elongated ellipsoid-like shape. The crystallite size from TEM images using ImageJ software was 42.46 ± 17.92 nm. The TEM image of ZA/nHA beads showed the presence of agglomeration of nHA crystallites within the alginate matrix (Fig. 3b).

Fig. 3

TEM images of (a) nanohydroxyapatite and (b) zinc alginate/ nanohydroxyapatite composite

Leaching of Zn2+ from the nanocomposite

The leaching of Zn2+ from the developed nanocomposite is of great importance, since it could influence its stability and application in water treatment as well as its environmental and health effects. As depicted in Fig. S2, the amount of Zn2+ released from the ZA/nHA composite was 1.50 and 2.16 mg L−1 after 6 and 24 h, respectively. The content of Zn2+ released remained relatively constant after 24 h. These values were generally lower than the maximum allowable concentration of 3 – 5 Zn2+ mg L−1 in drinking water suggested by WHO [1]. This result implied the safe use of the developed ZA/nHA composite for the treatment of water.

Bacterial removal from contaminated water

The antibacterial activity of the developed ZA/nHA nanocomposite was tested against E. coli, a common pathogen in contaminated water (Fig. S3). The results depicted in Fig. 4 indicated that ZA/nHA exhibited concentration-dependent antibacterial activity. The effective disinfection of E. coli was related to the initial concentration of bacterial cells as well as the nanocomposite dosage.

Fig. 4

Kinetics of E. coli disinfection using different concentrations of zinc alginate/ nanohydroxyapatite beads at different initial bacterial concentrations of (a) 104, (b) 105, and (c) 106 CFU mL.-1

The survival curves of E. coli in the presence of ZA/nHA nanocomposite were fitted to the Log-linear and the Weibull models (Table 1). Based on the R2 and RMSE values, the Weibull equation exhibited better fitting to the experimental results than the Log-linear equation. The inactivation kinetics of E. coli by the ZA/nHA are in good consistent with those of thermally-inactivated bacteria, which can also best described by the Weibull equation [39, 40]. The shape parameter (P) of the Weibull equation reflected downward concavity (P > 1), when the initial population of E. coli was ~ 104 and 105 CFU mL−1 (Table 1). Conversely, when the bacterial concentration was increased to ~ 106 CFU mL−1, an upward concavity (P < 1) of the survival curve was evident except at 1% w/v of Zn/nHA, the curve was linear (P = 1) (Table 1). These empirical values could also reflect physiological effects of the nanocomposite on bacterial cells. At P > 1, the remaining cells of E. coli were increasingly damaged, while at P < 1, the cells may adapt to the applied stress [39]. These observations implied the effectiveness of the developed ZA/nHA composite in the disinfection of E. coli-contaminated water at populations ≤ 105 CFU mL−1. Higher bacterial populations require prolonged treatment and/or higher composite dosage for effective disinfection.

Table 1 Kinetic modelling of bacterial disinfection under different treatments using Log-linear and Weibull modelsCyanobacterial and algal removal from water

The effective removal of Chroococcus sp. and Chlorella sp. from water by ZA/nHA beads was monitored daily by measuring the concentration of Chl. a. The results depicted in Fig. 5 clearly demonstrated that ZA/nHA eliminated nearly 100% of Chroococcus sp. cells after 2 days of treatment. Conversely, the removal of Chlorella sp. reached ~ 90% after 4 days of treatment. More importantly, neither Chroococcus sp. nor Chlorella sp. was able to regrow after removing the beads and cultivation in a new medium.

Fig. 5

Time-dependent variations in Chl. a concentration of (a) Chlorella sp. and (b) Chroococcus sp. after treatment with 1% w/v of zinc alginate/ nanohydroxyapatite beads in relation to the untreated control

Adsorption of CV using ZA/nHA nanocomposite

The developed ZA/nHA nanocomposite exhibited fast adsorption properties towards CV (Fig. 6a). Accordingly, the nanocomposite adsorbed 0.11 mg g−1 of CV after 0.5 min and the maximum adsorbed amount was 0.17 mg g−1 at 7 min. However, a small degree of desorption was observed at prolonged contact time (Fig. 6a). Similarly, the adsorption of methylene blue using sodium alginate/hydroxyapatite composite was fast and the equilibrium was attained within short period (30 min) [16].

Fig. 6

a Kinetics of crystal violet (CV) removal and fitting to pseudo-first order (PFO), pseudo-second order (PSO) and Elovich equations. b Kinetics modelling of CV adsorption using intra-particle diffusion model. c Effect of different initial CV concentrations on the adsorption process and fitting to different isotherm models. d Effect of pH on CV adsorption. e Effect of temperature on CV adsorption. f Thermodynamic plot for CV adsorption

The kinetic mechanism of CV adsorption was evaluated using different kinetic equations and the results are listed in Table 2. The best fitting of the models was arbitrated based on high coefficient of determination (R2), and low error (%ARE). The results indicated that the PSO and Elovich equations better described the kinetics of CV adsorption on the surface of ZA/nHA than the PFO model (Fig. 6a, Table 2). The calculated equilibrium adsorption capacity (qe = 0.178 mg g−1) by the PSO exhibited satisfactory fitting the experimental value (qe = 0.171 mg g−1). The PSO model assumes that chemical adsorption is a rate-controlling mechanism for the removal of CV by ZA/nHA from aqueous solution. Similarly, Elovich model describes chemisorption and assumes that the adsorbent have heterogenous binding sites with various binding energies [41].

Table 2 Different parameters for the pseudo-first order (PFO), pseudo-second order (PSO), and Elovich models for the biosorption of crystal violet using zinc alginate/nanohydroxyapatite beads

The kinetics of CV removal were also fitted to the intra-particle diffusion model to determine the rate limiting step, and the results were depicted in Fig. 6b. The high R2 of the first linear region in the plot (qtvs t0.5) deviated from the origin (Table 2). This implied that the intra-particle diffusion was prominent, but the deviation of the plot from the origin indicated a significant effect of the external mass transfer. The second linear region was related to the desorption of CV. The intra-particle rate constant (Ki) was low in the second linear region, which reflected a slow desorption process in relation to the fast adsorption process.

The adsorption process of CV was also evaluated using different isotherms (Table 2, Fig. 6c). The investigated models exhibited satisfactory fitting to the experimental data, but the best fit was related to the Sips model, owing to relatively low %ARE values (Table 3). In general, the Sips model combines the assumptions of both Langmuir and Freundlich models [28]. The Langmuir model assumes the formation of a monolayer of the adsorbate molecules, which is chemically adsorbed at energetically homogenous sites [42]. However, the presence of different functional groups on the surface of ZA/nHA indicated their heterogenous behavior towards CV, which violates the basic principle of Langmuir model. Conversely, the assumption of the Freundlich model is related to the formation of multilayer of the adsorbate molecules at energetically different binding sites. At low pollutant concentrations, the Sips equation relatively reduces to the Freundlich model, but it can describe the monolayer coverage of the Langmuir model at high pollutant concentrations [28].

Table 3 Calculated parameter for different isotherm models for the adsorption of crystal violet using zinc alginate/nanohydroxyapatite beads

An important parameter of the Langmuir isotherm is a dimensionless separation factor (RL), which is obtained as follows:

where KL is the Langmuir constant and C0 is the initial CV concentration. This parameter signifies adsorption as unfavorable (RL > 1), linear (RL = 1), favorable (0 < RL < 1), or irreversible (RL = 0) [10]. The calculated RL values in the present study fluctuated between 0.07 and 0.6 (Table 3), which implied favorable CV adsorption. In addition, the lower RL values were observed at high CV concentration, which reflected that the adsorption process became more favorable at high CV concentrations. The dimensionless exponent of Freundlich model (n) can also be used to describe the nature of the adsorption as (n = 1), physical (n > 1), or chemical (n < 1) [41]. The n value for CV adsorption was 2.25 (Table 3), which demonstrated the heterogenous nature of the developed adsorbent and signified the adsorption of CV as favorable physical process. The existence of various binding sites with different binding energies was also supported by the satisfactory fitting of the Temkin isotherm. The bT values of the Temkin model was 12.11 kJ mol-1, which is lower than 20 kJ mol-1, which is a characteristic feature of exothermic physical adsorption [43].

On the other side, the D-R isotherm showed the highest %ARE among the tested isotherms, suggesting a poor fitting to the experimental data. An essential property of the D-R isotherm is the calculation of mean free energy (E, kJ mol-1) of the adsorption using the following expression:

The E values signifies the adsorption process as chemical (E > 16 kJ mol-1), or physical (E < 8 kJ mol-1). The estimated E value for CV adsorption in the present study is 0.33 kJ mol-1 (Table 3), which indicated that the process is more inclined into physisorption.

On the other hand, pH of the solution during the adsorption process is a crucial parameter since it directly influences the surface charges of both adsorbent and adsorbate molecules. The effects of pH on the adsorption process of CV were depicted in Fig. 6d. The results indicated a maximum adsorption of CV at pH 6–7. The pHpzc of the developed ZA/nHA beads was 6.55, which indicated its existence as a positively charged adsorbent at pH < pHpzc. Accordingly, increasing the pH enables the adsorption of −OH ions on the surface of ZA/nHA beads, making it negatively charged and thus the removal of CV increased as a result of electrostatic interactions. However, the slight increase of CV adsorption at pH 3 may indicate that the adsorption is hydrogen-bonded controlled. The enhancement in the adsorption capacity of cationic dyes with the increasing pH agreed with previous studies [44].

Increasing temperature during adsorption markedly decreases the ability of the ZA/nHA to remove CV (Fig. 6e, f). The orientation and mechanism of CV adsorption was further investigated using thermodynamic parameters. The ΔHº value obtained in the current study for CV adsorption was negative, which is related to exothermic adsorption. The decrease of randomness at the adsorbent/adsorbate interface was indicated by the negative ΔSº value. Besides, the negative ΔSº suggested an associative adsorption mechanism without changes in the internal structure of the nanocomposite [29]. While the negative values of ΔGº signified the spontaneity of the thermodynamically favorable adsorption. The magnitude of ΔGº can also be used to identify adsorption mechanism as physisorption (ΔGº: − 20 to 0 kJ mol-1) or chemisorption (ΔGº: − 80 to − 400 kJ mol-1). Accordingly, the listed ΔGº values in Table S1 indicated that the adsorption of CV on the surface of ZA/nHA is physical, which agreed with the results of the D-R model. In a similar study, the adsorption of methylene blue was predominately physical on the surface of sodium alginate/hydroxyapatite material [16].

The FT-IR analysis of the nanocomposite beads after the adsorption of CV indicated a marked shift in the O–H stretching vibrations into lower wavenumber (3423.55 cm−1). Similarly, the peak of symmetric − COO− vibrations was slightly shifted into lower wavenumber (1420.26 cm−1). Furthermore, A slight shift to the higher wavenumber was observed in the peaks of C-H (2928.09 cm−1), PO43− (603.51 cm−1), and the three vibrational modes of τ(CO) + δ(CCO) + δ(CCH) (819.21 cm−1). The shift in these functional groups in the nanocomposite implied their potential role in the removal of CV. The hydrogen-donating groups such as OH and COOH in the ZA/nHA adsorbent may form H-bonding interactions with the hydrogen accepting group (nitrogen) of CV. Furthermore, another type of H-bonding may occur between the H-donating groups of the adsorbent and the aromatic rings of CV, which is known as Yoshida bonding [10, 11]. Similarly, the presence of electron- donating oxygen groups in the ZA/nHA may induce the formation of n-π interaction with aromatic rings of CV as electron-acceptor. Furthermore, the electrostatic interactions were also evident as indicated by the effects of pH on the removal process.