A water-stable functionalized NiCo-LDH/MOF nanocomposite: green synthesis, characterization, and its environmental application for heavy metals adsorption

2.1 Materials

Nickel(II) nitrate hexahydrate [Ni(NO₃)·6H₂O] (≥98.5%), cobalt(II) nitrate hexahydrate [Co(NO₃)₂·6H₂O] (98%), mercury(II) chloride (HgCl₂, ≥99.5%), 1,2,4,5-benzenetetracarboxylic acid (pyromellitic acid, 96%), zirconium(IV) chloride (ZrCl₄, ≥99.9%), and 3-[2-(2-aminoethylamino)ethylamino]propyltrimethoxysilane (triamino silane coupling agent, TASCA, technical grade) were purchased from Sigma-Aldrich (Spain). Ethylene glycol (EG, ≥98.0%), acetone (≥99.8%), sodium hydroxide pellets (NaOH, 99%), and hydrochloric acid (HCl, 36%) were purchased from Merck (Darmstadt, Germany). Ethanol (absolute and 96%) was purchased from Dr. Mojallali (Dr. Mojallali Chemical Complex Co. Tehran, Iran). Double-distilled water was used in the synthesis of the samples and adsorption experiments. All chemicals were used without further purifications.

2.2 Synthesis method

2.3 Characterization of the samples

The crystalline structure of the samples was determined by powder X-ray diffraction (XRD, AW-XDM300, Asenware, China) analysis. The functional groups in the powder samples were identified by Fourier transform infrared (FT-IR, Thermo Nicole, model AVATAR 370, USA) spectroscopy. Field emission scanning electron microscopy (FESEM, SIGMA HV, Zeiss, Germany) was used to observe the morphology of the samples. Nitrogen adsorption–desorption measurement at 77 K was conducted using a porosimetry analyzer (Belsorp-mini II, BEL, Japan) to analyze the pore structure and textural properties of the samples. A flame atomic absorption spectrometer (FAAS, PerkinElmer Model A300, Norwalk, USA) was used for the analysis of heavy metals concentrations in the solutions.

2.4 The batch adsorption experiments

Stock solutions of Hg(II) and Ni(II) (1000 mg L⁻¹) were prepared by dissolving an appropriate amount of HgCl₂ and Ni(NO₃)₂·6H₂O in double-distilled water. Metal solutions of different initial concentrations were prepared by dilution of the stock solutions. All batch adsorption experiments were carried out by adding a fixed amount of LDH/MOF NC (adsorbent dose ( $W$ )=2.0 mg) into 50-mL screw-cap polypropylene conical vials containing 20 mL of aqueous solutions of Hg(II) or Ni(II) (solution volume ( $V$ )=20 mL) and shaking the vials at 200 rpm in an orbital water bath shaker at a constant temperature ( $T$ =22 ± 3 °C) for certain contact time ( $t$ ). The solutions were centrifuged at 5000 rpm at the desired contact times to isolate the LDH/MOF NC particles and the supernatant solutions were analyzed using FAAS to measure the residual Hg(II) and Ni(II) concentrations. The pH of the solutions was adjusted with 0.1 mol L⁻¹ HCl or 0.1 mol L⁻¹ NaOH solutions by using a pH meter throughout the adsorption tests when desired.

Uptake experiments concerning the effect of solution pH on the adsorption efficiency of Hg(II) and Ni(II) were as follows: pH range: 2.0–8.0 for Hg(II) and 3.0–9.0 for Ni(II); initial metal concentration ( $C_{\mathrm{i}}$ )=25 mg L⁻¹; $W$ =2.0 mg; $V$ =20 mL; $T$ =22 ± 3 °C; $t$ =120 min; shaking speed=200 rpm. Uptake experiments concerning the effect of initial concentration were as follows: pH=3.0 for Hg(II) and pH=8.0 for Ni(II); $C_{\mathrm{i}}$ =0.5–250 mg L⁻¹; $W$ =2.0 mg; $V$ =20 mL; $T$ =22 ± 3 °C; $t$ =120 min; shaking speed=200 rpm. Uptake experiments concerning the effect of contact time were as follows: pH=3.0 for Hg(II) and pH=8.0 for Ni(II); $C_{\mathrm{i}}$ =100 mg L⁻¹; $W$ =2.0 mg; $V$ =20 mL; $T$ =22 ± 3 °C; $t$ =1–120 min; shaking speed=200 rpm. Metal removal percentages and adsorption capacities ( $Q$ ) were calculated by the following equations:

3.1 Synthesis of LDH/MOF NC

The LDH/MOF NC was synthesized by using a facile and “green“ synthesis protocol via in situ growth of UiO-66-(Zr)-(COOH)₂ nanoparticles on the TA/COOH-Ni₅₀Co₅₀-LDH ultrathin sheets. Among the various LDH nanoparticles, Ni₅₀Co₅₀-LDH was chosen as a scaffold because the strategy used for the preparation of Ni₅₀Co₅₀-LDH is an environmentally friendly protocol in which a non-toxic EG-water solvent system is used instead of common toxic and flammable solvents such as dimethylformamide, dimethyl sulfoxide, toluene, acetone, and 1-butanol. Besides, Ni₅₀Co₅₀-LDH possesses extended micrometer-size sheets that provide a suitable scaffold for the growth of MOF particles. To date, only a few MOFs have been synthesized under toxic solvent-free conditions (Yang et al., 2013; Sánchez-Sánchez et al., 2015). Among various kind of MOFs, UiO-66-(Zr)-(COOH)₂ was chosen because, in addition to good water resistance properties, hydrothermal stability, and the presence of abundant carboxylic groups within its structure as potential adsorption cites, a facile and “green” synthetic method is applied in which water is used as solvent under reflux condition. In order to chemically attach MOF nanoparticles to the LDH sheets, the LDH surface was first chemically functionalized with TASCA and pyromellitic acid molecules. By implementing this strategy, MOF nanoparticles grow chemically on the surface of TA/COOH-LDH sheets due to the presence of abundant carboxylic groups (pyromellitic acid molecules) grafted on the LDH surface. These carboxylic functional groups can act as potential coordination sites for nucleation and growth of clusters in MOF.

Due to the presence of abundant functional groups like —COOH, —NH—, —NH₂, —OH, and also aromatic rings in the structure of LDH/MOF NC, this material can be used as a potential adsorbent for the removal of heavy metals. It is reported that among the various kind of functional groups, the —COOH-metal ion interaction is probably the most significant for different applications, especially adsorption (Bala et al., 2007). Fig. 1 displays the overall synthesis procedure for the synthesis of LDH/MOF NC.

Close

Fig. 1

The overall synthesis process of the LDH/MOF NC.

Export to PPT

3.2 Characterization of the samples

The powder XRD patterns of the samples are shown in Fig. 2. The powder XRD patterns of the pure LDH (Ni₅₀Co₅₀-LDH) and the MOF (UiO-66-(Zr)-(COOH)₂) exhibited a similar pattern to those observed in previous works (Soltani et al., 2021; Yang et al., 2013; Li et al., 2016), indicating the successful preparation of these samples. The powder XRD profile of the LDH/MOF NC reveals that characteristic peaks representing both MOF and LDH phases are still present in their composite, implying the presence of both phases in the LDH/MOF NC.

Close

Fig. 2

XRD patterns of the samples.

Export to PPT

FT-IR spectra of the powder samples are shown in Fig. 3. The characteristic absorption bands related to the functional groups of pure Ni₅₀Co₅₀-LDH, pure UiO-66-(Zr)-(COOH)₂, TA/COOH-LDH, and the LDH/MOF NC were observed and the results are in line with previous reports (Soltani et al., 2021; Yang et al., 2013; Li et al., 2016). For pure Ni₅₀Co₅₀-LDH, the signal at 1630 cm⁻¹ is attributed to the bending vibration of H₂O molecules (interlayer and surface adsorbed water molecules). The FT-IR band at 3445 cm⁻¹ is assigned to O-H stretching vibration related to the metal-hydroxyl groups as well as water molecules. The signals at around 1378–1382 cm⁻¹ and 2850–2920 cm⁻¹ are attributed to the vibrations of interlayer anions ( ${\mathrm{C}\mathrm{O}}_3^{2-}$ and ${\mathrm{N}\mathrm{O}}_3^{-}$ ) and C—H vibration of trace EG molecules remained in the LDH, respectively. Additionally, the broad absorption band at 632 cm⁻¹ is related to the M—O—M, O—M—O, and M—O vibrations (M⚌Ni and Co). After the surface functionalization, the appearance of new FT-IR signals at 862 cm⁻¹,1492 cm⁻¹, and 1570 cm⁻¹ are attributed to the aromatic rings of pyromellitic acid molecules. The FT-IR shoulder at 1320–1340 cm⁻¹ is attributed to the C—N and C—O stretching modes of carboxyl (—COOH) and amine (—NH₂ and —NH—) groups. Another shoulder at approximately 1640 cm⁻¹ is assigned to the amide carbonyls as schematically shown in Fig. 3. In addition, the appearance of a broad signal in the range of 3000–3600 cm⁻¹ region is mainly attributed to the carboxyl groups of pyromellitic acid molecules.

Close

Fig. 3

FT-IR spectra of the Ni₅₀Co₅₀-LDH (a), TA/COOH-Ni₅₀Co₅₀-LDH (b), UiO-66-(Zr)-(COOH)₂ (c), and LDH/MOF NC (d).

Export to PPT

Pure UiO-66-(Zr)-(COOH)₂ showed two sharp signals with high intensity at 1414 cm⁻¹ and 1587 cm⁻¹ which related to symmetric and asymmetric stretching vibrations of coordinated carboxylic acid, respectively. The signal at 787 cm⁻¹ belongs to the γCH in aromatic rings and absorption bands approximately at 1498 cm⁻¹ and 802 cm⁻¹ are attributed to vibrational modes of the phenyl rings. The FT-IR band at 1715 cm⁻¹ is assigned to the carbonyl group of free carboxylic acid (linker). The broad absorption band in the spectral range of 3400–3600 belongs to acidic OH of free carboxylic groups as well as surface adsorbed water.

FT-IR spectrum of the LDH/MOF NC showed the characteristic absorption bands of both TA/COOH-LDH and UiO-66-(Zr)-(COOH)₂, indicating the simultaneous presence of both phases in its structure.

FESEM micrographs of the samples are shown in Fig. 4. The First row of Fig. 4 clearly shows the ultrathin 2D sheets of Ni₅₀Co₅₀-LDH. These extended 2D ultrathin sheets can act as a potential scaffold for the growth of various kinds of nanoparticles by implementing an appropriate synthesis strategy. The second row of Fig. 4 reveals the spherical nanoparticles of UiO-66-(Zr)-(COOH)₂. The FESEM micrographs taken after in situ growth of UiO-66-(Zr)-(COOH)₂ nanoparticles on the surface of ultrathin 2D sheets of TA/COOH-Ni₅₀Co₅₀-LDH clearly reveal that the nanoparticles are homogeneously distributed on the sheets (Fig. 4, third row). FESEM-EDX mapping analysis of LDH/MOF NC (Fig. 5, the first and second rows) revealed that the synthesized composite possesses a uniform surface structure consisted of a homogeneous distribution of the structural elements (Zr, Ni, Co, O, C, N, Si).

Close

Fig. 4

The FESEM images of pure Ni₅₀Co₅₀-LDH (the first row), UiO-66-(Zr)-(COOH)₂ (the second row), and LDH/MOF NC (the third row).

Export to PPT

Close

Fig. 5

FESEM-EDX mapping images (the first and second rows) of synthesized LDH/MOF NC adsorbent and the corresponding elements and its EDX spectrum (the third row).

Export to PPT

Besides, the EDX peaks corresponding to these elements are depicted in the third row of Fig. 5. The presence of silicon atoms in the structure of the LDH/MOF NC is because of the existence of TASCA molecules in the composite structure (Fig. 1). The origin of nitrogen atoms is also due to the presence of TASCA molecules and interlayer nitrate anions in the gallery space of the Ni₅₀Co₅₀-LDH. These observations, in addition to the results of powder XRD and FT-IR analyses, well confirmed the synthesis of the composite through in situ growth of the MOF nanoparticles on the surface of the LDH sheets.

In order to gain a better understanding of the textural properties (pore volume and surface area) of the samples, the adsorption–desorption isotherms of the samples were measured at 77 K (Fig. 6). The Ni₅₀Co₅₀-LDH reveals a Type III isotherm with a Type H1 hysteresis loop which is representative of the existence of mesopores and macropores (Li et al., 2016). As shown in Fig. 6, both UiO-66-(Zr)-(COOH)₂ and LDH/MOF NC reveal a combination of Type IV(a) and Type 1(b) isotherms, which is characteristic of mesoporous and microporous solids, respectively. The characteristic feature of the Type IV isotherm is the presence of a hysteresis loop and Type I(b) isotherm is generally found in solids with possibly narrow mesopores (pore size<2.5 nm) and micropores (pore size=1–2 nm) (Thommes et al., 2015). The Brunauer-Emmett-Teller surface area (S_BET), the Langmuir surface area (S_Langmuir), and total pore volume (PV) obtained for all samples are also shown in Fig. 6. It is clear that compared to pure Ni₅₀Co₅₀-LDH (S_Langmuir=27 m² g⁻¹ and PV=0.23 cm³ g⁻¹), the LDH/MOF NC possesses a higher surface area and pore volume (S_Langmuir=69 m² g⁻¹ and PV=0.35 cm³ g⁻¹) due to the growth of MOF nanoparticles on the LDH surface. Due to such a bi-/trimodal pore system in the synthesized LDH/MOF NC, the mass transfer through the pores can be dramatically improved (Marjani et al., 2020; Soltani et al., 2020a; Ng et al., 2018). This feature is especially effective in the adsorption, extraction, and catalysis applications where target molecules must reach adsorption sites more quickly.

Close

Fig. 6

N₂ adsorption–desorption isotherms of the samples at 77 K and corresponding calculated values for the surface area and the total pore volume.

Export to PPT

3.3 Adsorption studies

To evaluate the capability of LDH/MOF NC for the removal of heavy metals, a set of adsorption tests were conducted and some adsorption factors were optimized. Also, several nonlinear isotherm and kinetic models were used to find the plausible adsorption mechanism involved in the removal process. Determining the quality of the fit is achieved by $R_{\mathrm{A}\mathrm{d}\mathrm{j}.}^2$ , $X_{\mathrm{r}\mathrm{e}\mathrm{d}.}^2$ , and SD statistics as mentioned before.

The effect of solution pH on the adsorption performance of LDH/MOF NC was investigated (Fig. 7a). For Hg(II), the adsorption rate reached the maximum at the pH value of 3.0 (%Removal ∼ 99) and then decreased successively. The lower adsorption rate at pH<3.0 is due to the high concentration of hydrogen ions (H₃O⁺) in the solution and competition with Hg(II) and Ni(II) cations for adsorption sites on the adsorbent surface. At low pHs, the LDH/MOF NC was protonated and the complexation between functional groups of the adsorbent and Hg(II) cations was inhibited because of the electrostatic repulsion between positively charged LDH/MOF NC and Hg(II) cations, leading to a decrease in adsorption rate. Furthermore, it is reported that there would be mercury hydroxide species, Hg(OH)⁺ or Hg(OH)₂, in an aqueous solution when the pH values are more than 3.0 and 4.0 (Fu et al., 2019; Lin et al., 2018), leading to a decrease in adsorption rate. For Ni(II), the maximum removal percentage was observed when the pH value is 8.0. The Ni(II) ion removal pattern for pH changes follows the IMD model (increasing-maximum-decreasing model), which has already been demonstrated by Soltani and co-workers (Soltani et al., 2020e; Soltani et al., 2019). Accordingly, the optimum pH values were 3.0 and 8.0 for adsorption of Hg(II) and Ni(II), respectively, by the LDH/MOF NC from the aqueous solution, and all the next adsorption experiments were carried out under these pH values.

Close

Fig. 7

The effect of pH (a), initial metal concentration (b), and contact time (c) on the adsorption of Hg(II) and Ni(II) by the LDH/MOF NC. The non-linear fitting results for the isotherm and kinetic models are shown in Fig. 7a and Fig. 7b, respectively.

Export to PPT

The effect of initial metal ion concentration on the adsorption capacity of the LDH/MOF NC was investigated and the results are shown in Fig. 7b. The adsorption capacities for Hg(II) and Ni(II) continually increase when the initial concentrations of the heavy metals increase from 0.5 mg L⁻¹ to 50 mg L⁻¹ (the X-axis in Fig. 7b shows the corresponding equilibrium concentrations). The adsorption equilibrium for both Hg(II) and Ni(II) was almost achieved when the initial concentration was 100 mg L⁻¹. After this point, the adsorption capacity for Hg(II) and Ni(II) reached an almost constant value. The experimental maximum adsorption capacities of Hg(II) and Ni(II) by LDH/MOF NC at an initial concentration of 250 mg L⁻¹ were 530 mg g⁻¹ and 439 mg g⁻¹, respectively. To gain the theoretical maximum adsorption capacity of the adsorbent toward Hg(II) and Ni(II), as well as the plausible adsorption mechanism involved in the adsorption process, three isotherm models, namely the Langmuir (Eq. 7) (Langmuir, 1916), the Freundlich (Eq. 8) (Freundlich, 1906), and the Redlich-Peterson (R-P, Eq. 9) (Redlich and Peterson, 1959), were used. The non-linear form of these equations and their corresponding parameter values after the non-linear fitting method are given in Table 1. Also, the Langmuir, Freundlich, and R-P isotherm curves for adsorption of Hg(II) and Ni(II) ions are shown in Fig. 7b. For the results presented in Table 1, among the two-parameter isotherm models (the Langmuir and the Freundlich), the Langmuir isotherm provides a better fit based on $R_{\mathrm{A}\mathrm{d}\mathrm{j}.}^2$ , $X_{\mathrm{r}\mathrm{e}\mathrm{d}.}^2$ , and SD values. In addition, theoretical maximum adsorption capacities ( $Q_{\mathrm{m},\mathrm{c}\mathrm{a}\mathrm{l}.}$ ) of Hg(II) and Ni(II) obtained by the Langmuir equation were close to the corresponding experimental maximum adsorption capacities ( $Q_{\mathrm{m},\mathrm{e}\mathrm{x}\mathrm{p}.}$ ) as shown in Table 1.

The R-P model can be applied to ascertain whether the adsorption behavior is the Langmuir or the Freundlich. In this equation, when $g$ =1 and $g$ =0 the equation reduces to the Langmuir isotherm and becomes linear at low surface coverage, respectively. It was observed that for both Hg(II) and Ni(II) adsorption, the values obtained for exponent ( $g$ =0.9684 for Hg(II) adsorption and $g$ =0.9936 for Ni(II) adsorption) were very close to unity (Table 1), indicating that the adsorption of Hg(II) and Ni(II) ions on the surface of the LDH/MOF NC fits more readily to the Langmuir model than the Freundlich model. Consequently, the adsorption process takes place without any interactions between adsorbates and via monomolecular coverage of adsorbate ions on the surface of the LDH/MOF NC adsorbent with a limited number of localized adsorption sites (Lima et al., 2015).

The equilibrium parameter ( $R_{\mathrm{L}}$ , Table 1) is a dimensionless factor that can be applied to suggest the adsorption behavior (irreversible as $R_{\mathrm{L}}$ =0; unfavorable as $R_{\mathrm{L}}$ >1; linear as $R_{\mathrm{L}}$ =1; favorable as 0< $R_{\mathrm{L}}$  < 1). The $R_{\mathrm{L}}$ values obtained for Hg(II) and Ni(II) (Table 1) at the highest initial metal concentration (250 mg L⁻¹) lie between zero and unity, indicating that the adsorption of Hg(II) and Ni(II) by the LDH/MOF NC is favorable.

The effect of contact time on the removal of Hg(II) and Ni(II) by the adsorbent was also investigated and three kinetic models, namely pseudo-first-order (PFO, Eq. 10), pseudo-second-order (PSO, Eq. 11), and Avrami (Eq. 12) were used to fit the experimental adsorption data and to interpret the adsorption behavior.

As shown in Fig. 7c, with increasing contact time from 1 min to 20 min, the adsorption capacity increases sharply, from ∼ 80 mg g⁻¹ to ∼ 500 mg g⁻¹ for Hg(II) and from ∼ 50 mg g⁻¹ to ∼ 410 mg g⁻¹ for Ni(II), until it reaches almost equilibrium. After 20 min, the adsorption curve reaches a constant state. This relatively rapid adsorption time might be attributed to the presence of abundant adsorption sites (organic functional groups such as —COOH, —NH₂, —NH—, —OH, and aromatic rings) in the adsorbent structure that possesses a high tendency to capture heavy metals from the aqueous medium. The non-linear forms of the three aforementioned kinetic models are given in Table 1. The kinetic parameters after non-linear fitting are also given in Table 1. By comparing the obtained kinetic data after the non-linear fitting method, it can be seen that the Avrami kinetic model gives the best fit to the experimental data than the other two models. The theoretical equilibrium adsorption capacities ( $Q_{\mathrm{e},\mathrm{c}\mathrm{a}\mathrm{l}.}$ =498.4 mg g⁻¹ for Hg(II) and 424.1 mg g⁻¹ for Ni(II)) obtained by the Avrami model are in close agreement with the corresponding experimental adsorption capacities at equilibrium ( $Q_{\mathrm{e},\mathrm{e}\mathrm{x}\mathrm{p}.}$ =500.0 mg g⁻¹ for Hg(II and 425.2 mg g⁻¹ for Ni(II)). Also, for both Hg(II) and Ni(II) adsorption, the Avrami model shows a higher $R_{\mathrm{A}\mathrm{d}\mathrm{j}.}^2$ and lower $X_{\mathrm{r}\mathrm{e}\mathrm{d}.}^2$ and SD values than those parameters obtained by PFO and PSO models.