# Synthesis of diglycolic acid functionalized core-shell silica coated Fe3O4 nanomaterials for magnetic extraction of Pb(II) and Cr(VI) ions

Jun 22, 2020

### Structural elucidation and phase identification

The structural elucidation of Fe3O4, silica coated; their subsequent amine and acid functionalized magnetite nanoparticles were performed by fourier-transform infrared spectroscopy (FT-IR). The technique was used to analyse the nanoparticles and surface groups present on all the prepared samples as shown in Fig. 2. The band at 550 cm−1 assigned to the FeO stretching vibration indicating the formation of magnetite nanoparticles having no impurities in the sample. Silica coated Fe3O4 nanoparticles showed two characteristics vibrations at 790 and 1065 cm−1 due to the stretching of Si-O and Si-O-Si bonds respectively. Amine functionalized sample gave additional band at 1629 cm−1 due to N-H bending vibration along with Si-O-Si band ascertaining amine modification and a prominent carbonyl stretching vibration at 1697 cm−1 along with the amide carbonyl stretching vibration at 1626 cm−1 confirming the successful surface functionalization of nanomaterials with acid moieties. Moreover, additional bands can be seen at 690 cm−1, 1220 cm−1, 1425 cm−1 and in the region of 2900–3000 cm−1 due to N-H (oop) bending, ether C-O-C stretching, C-N stretching and CH2 stretching vibrations. The IR data proved the successful formation, coating and functionalization of the magnetite nanoparticles.

The acid moieties present on the surface of the nanomaterials played a vital role on heavy metal adsorption. Their quantification is usually carried out by spectroscopic and thermal analysis. The data obtained from these methods need careful analysis including baseline correction and deconvolution that can lead to misinterpretations and low reproducibility. However, chemical method that quantifies oxygen-containing functional groups on the surface of nanomaterials is Boehm titration (a back titration). The same titration was performed on the FGA-1 sample and the acidity was calculated as 0.32 mmol/g indicating the presence of the acid functionality on the surface of the nano-sorbents.

The post adsorption IR data was also recorded and presented in Figure S1 (see SI). Both the spectra clearly describe that surface acid functional groups illustrate a major contribution on heavy metal ions adsorption. The pre-adsorbed FGA-1 sample exhibited a band at 1697 cm−1 characteristic of carbonyl stretching of the acid moieties present on the surface. The same band was shifted to the lower wave number in the post adsorbed FGA-1 for both metals ions (Figure S1) i.e., 1554 and 1612 cm−1 for Pb(II) and Cr(VI) respectively. This observation clearly verified that acid functionalized magnetite nanoparticles interacted with metal ions and hence this nano-sorbent facilitated the removal of heavy metal ions from wastewater.

Crystallinity and phase identification of coated and functionalized magnetite nanoparticles were carried out by X-ray diffraction technique and XRD patterns recorded for 2θ in the range of 25° to 70° on the powdered samples are shown in Fig. 3. All the samples exhibited the same diffraction patterns at 2θ = 30.1°, 35.5°, 37.1°, 43.1°, 53.5°, 57.0° and at 62.6°, corresponding to [311], [222], [400], [422], [511] and [450] hkl values. These observations indicate the single phase materials and the XRD pattern of magnetite crystal structure is in good agreement with the JCPDS 01–075–0033 powder diffraction. It further suggest well retained magnetic cores present in all samples when compared with the initial Fe3O4 nanoparticles except for the distinct peak appeared at 2θ = 20° corresponding the diffraction peak for silica groups15. Moreover, the diffractograms show no peak shifting upon coating and functionalization meaning that the crystalline phase and stability of magnetite nanoparticles persevered. The crystallite size of magnetite nanoparticles was also determined by Scherrer formula using XRD data and the average particle size was 45.08 nm.

### Morphology and elemental composition

Morphological investigations on the powdered samples of magnetite, silica coated and acid functionalized nanoparticles were carried out by scanning electron microscopy and SEM images are presented in Fig. 4. The micrographs show a quasi-spherical shape with particles showing little agglomeration having an average particle size 55.5 ± 5.5 nm (Fig. 4(a)), which is close to the crystallite size obtained from the XRD data. After silica modification, particle size is increased with a cloudy shell-like appearance on the surface providing evidence about the presence of amorphous functionalities around the nanoparticles (Fig. 4(b)). This cloudy shell becomes thicker with increasing number of functionalities on the surface. Figure 4(c) shows more thickened cloudy shell on the surface of acid functionalized nanoparticles without affecting the crystalline phase and stability of magnetite nanoparticles. Additionally, these particles also appeared to be agglomerated due to the nanosized interacting surfaces.

Energy-dispersive X-ray technique usually coupled with SEM give valuable information regarding the chemical composition of the materials. The elemental composition of magnetite nanoparticles was monitored by recording the EDS spectra as given in Fig. 5. The pristine magnetite nanoparticles gave two peaks for Fe and O elements showing the purity of synthesized nanoparticles (Fig. 5(a)). Figure 5(b) indicated EDS spectrum of diglycolic acid functionalized nanoparticles with additional peaks of carbon, nitrogen and silicon.

In acid functionalized spectrum, more elements are expected relative to pure magnetite nanoparticles due to silica coating and functionalization of surface with amine and acid moieties. Moreover, oxygen percentage also increased in the later spectrum due to the surface bearing one ether (-O-) and two acid groups (-COOH, -NHCO).

The internal morphology of diglycolic acid functionalized nanoparticles was also monitored by transmission electron microscopy and the TEM image is presented in Fig. 6. The micrograph depicted the presence of cloudy shell of acid functionalized nanoparticles showing an average particle size of core approximately 60 nm and 20 nm shell around on the surface. Furthermore, it confirmed that functionalization did not affect the crystallinity and stability of magnetic core.

### Magnetic measurements

VSM method is used to determine the magnetic properties of nanomaterials based on the change of magnetic field produced by the magnetometer’s vibrating component. The magnetic behaviour and hysteresis loops of pure and acid functionalized magnetite samples as measured by VSM is shown in Fig. 7. The mass saturation magnetization value (Ms) obtained for pristine Fe3O4 nanoparticles was 65.96 emu/g, which was further reduced to 31.65 emu/g for acid functionalized nanoparticles due to the incorporation of non-magnetic shell (silica/amino/acid) around the surface. These observations are in good agreement with the previous studies that functionalization causes a decrease in saturation magnetization33,34,35. However, these samples showed large values of remanence (Mr) and coercivity (Hc) favouring ferromagnetic behaviour (Table 1). The other evidence in favour of ferrimagnetic nature comes from the particle size i.e., more than 20 nm excluding the paramagnetic anticipation in the nanoparticles. These measurements are in line with reported values in the literature36,37,38.

### Zeta potential, surface area and porosity measurements

Zeta potential is the measure of potential difference across phase boundaries between solids and liquids. Zeta potential measurement and the surface charge on FGA-1 at pH 7 is given in Fig. 8. The zeta potential is negative (−7.39) as expected due to presence of carboxylate anions on the surface at pH 7. This relates with the adsorption studies carried out on metal ions at various pH values. The surface of nano-sorbent is negatively charged at this pH value due to ionization of carboxylic acid, adsorption of Pb(II) ions is preferred relative to Cr(VI) ions. Slight change for Pb(II) adsorption was observed at higher pH values, however, the adsorption of Cr(VI) was decreased. At lower pH values, the surface is assumed to be protonated facilitating the adsorption of Cr(VI) ions more relative to Pb(II) cations. The same trend has been observed while moving from higher to lower pH values.

Moreover, BET measurements on FSA-1 and FGA-1 samples were performed using N2 adsorption-desorption isotherms (Fig. 9) to evaluate textural parameters such as surface area, pore volume and pore size of nano-sorbents as described in Table 2. The BET linear plots of FSA-1 and FGA-1 from N2 isotherms at 77 K are shown in Figure S2. The obtained parameters showing a mesoporous structure with narrow range of pore size distribution. These values are in favour of electrostatic adsorption phenomena, as surface area of nanoparticles is less due to more size and more functionality on the surface. However, physisorption cannot be neglected completely and seems more favourable in case of Cr(VI) ion.

### Metal ions uptake

The synthesized nanosorbents comprising terminal acid moieties on their surfaces can be exploited as excellent chelating sites for heavy metal ions, thus removing metal cations from wastewater and effluents of various industries. In the present study, lead (II) and chromium (VI) were preferred for uptake measurements because these metals ions have excellent coordination with acid groups present on the surface of nanoparticles and the results were evaluated by Langmuir and Freundlich models. To achieve an optimal adsorption, certain parameters (time, pH and adsorbate concentration) were optimised over a constant dosage of adsorbent at room temperature. For each experiment, 10 mg of adsorbent dosage was used as previously optimized 32 over the variable range of one parameter to get the maximum adsorption limit and after that the same procedure was repeated with all other parameters. Average values of three runs of each parameter optimized were taken with ± 0.01 standard deviation.

#### Effect of time

Time was the first parameter to be optimised inceptively for both the metal ions. The solution of each ion (10 ppm) was agitated with 10 mg of acid functionalized nanomaterials at neutral pH and room temperature over a variable range of time intervals (15, 30, 60, 90 and 120 min). Figure 10 summarises optimum adsorption with time for both the metal ions up to 60 min at equilibrium point and then it almost level off with further increase in time. However, nanomaterials showed more adsorption for Pb(II) cations than Cr(VI) ions due to negatively charged surface groups, favoured the adsorption of cations and the maximum adsorption for Pb(II) ions were ~60% and ~50% for Cr(VI) ions. The theoretical values of qe(Theoretical) and also agree with the experimental values qe(Experimental) exhibiting that the present adsorption system follows to the pseudo-second-order mechanism and the adsorption rate is controlled by chemical sorption (Table S2 and S4).

Kinetic modelling not only allows estimation of sorption rates but also leads to suitable rate expressions characteristic of possible reaction mechanisms. Therefore, pseudo-first order and pseudo-second order kinetics models were applied on the adsorption data as a function of time for heavy metal removal. The respective kinetics and kinetic parameters of these two models investigated on the adsorption data of both metal ions can be seen from the Figures S3S6 and Tables S1S4. Pseudo-first order was found to be inapplicable while pseudo-second order kinetic model was applicable with R2 values 0.99 for both the metal ions. In this model, the rate-limiting step is the surface adsorption that involves chemisorption, where the removal from a solution is due to physicochemical interactions between the two phases.

#### Effect of pH

pH has vigorous effect on metal uptake as varying pH values causes a tremendous change for the adsorption of both the metal ions. The effect of pH was optimised at constant adsorbent concentration (10 mg), time (1 h) at room temperature. Since Pb(II) ions are positively charged species and inherently attractive towards the negative charge. So, pH was varied from acidic to basic and the adsorption behaviour is depicted in Fig. 11. The uptake was maximum at neutral pH, tangible evident showing the tendency of active sites to attract positively charged species Pb(II) when these are more likely to be unoccupied. This is because at lower pH values, more protonated surfaces block the binding of Pb(II) cations and reduces the adsorption rate while at higher pH values, concentration of OH ions competes for Pb(II) cations and reduces the adsorption rate. The maximum pH observed was at 7 pH of the solution and almost negligible at acidic or basic pH values. Cr(VI) showed opposite trend and surprisingly more adsorption was observed at lower pH values i.e. up to ~85 percent adsorption at acidic pH = 3. This is because the more protonated surfaces tends to attract more Cr(VI) ions towards themselves causing the more adsorption of Cr(VI) ions that decreases further increasing the pH value, possibly due to protonation of chromate ions.

The effect of metal ions concentration was studied to describe the adsorption kinetics at maximum adsorption capacity of the adsorbents. Figure 12 explains the adsorption behaviour at optimised values of time (1 hr), pH = 7 for Pb(II) cations and pH = 3 for Cr(VI) ions at room temperature per 10 mg of acid functionalized magnetite. While the concentration of metal ions was varied from 10–50 ppm and saturating the solution with metal ions have adverse effect on the adsorption rate. Because all the available active sites are already occupied that decreased the tendency to pick up more metal ions in the solutions. However, the maximum adsorption was observed at 10 ppm solution, optimised value over 10 mg of acid functionalized adsorbent’s concentration beyond further increase in concentration was unfavourable for maximum adsorption capacity.

The adsorption mechanism was studied by Dubinin-Raduchkevick, Langmuir and Freundlich isotherm models. The parameters derived from these three models provide much information about the affinity of adsorbent, surface properties and the nature of adsorption. Dubinin-Raduchkevick isotherm model was applied on the adsorption data for both the metal ions and this model was found to be not applicable as can be seen from Figures S7 and S8 and D-R isotherm parameters (Tables S5 and S6).

Langmuir adsorption isotherm states that the adsorption phenomena are monolayered because of the limited number of homogeneous active sites on adsorbent’s surface, where if one site is occupied by a particle, no further adsorption is possible. Langmuir adsorption isotherm’s linear form is given by the following equation.

$$frac{1}{{{rm{q}}}_{{rm{e}}}}=frac{1}{{{rm{Q}}}_{0}}+frac{1}{{{rm{bQ}}}_{0}{{rm{C}}}_{{rm{e}}}}$$

(5)

where, qe = equilibrium concentration of adsorbed ions, Q0 = maximum removal capacity, Ce = equilibrium concentration of adsorbate

Graphical representation of Langmuir adsorption isotherm for Pb(II) and Cr(VI) adsorption at the surface of acid functionalized nanoparticles was obtained by plotting of 1/Ce against 1/qe using the above straight-line equation as shown in Figs. 13 and 14. The parameters of Langmuir adsorption isotherm are given in the Table 3.

$${R}_{L}=frac{1}{1+b{Q}_{O}}$$

(6)

The value of RL from the above equation is found to be higher than 0 but less than 1, depicting that adsorption is favourable in both cases. The Langmuir process thermodynamic was ensured by calculating of ΔG values for Pb(II) = ‒1.59 KJmol−1 and for Cr(VI) = ‒2.78 KJmol−1. These negative values indicated that sorption of both the metal ions is spontaneous and thermodynamically feasible.

Freundlich isotherm is known to describe the non-ideal and reversible adsorption behaviour in favour of the development of multilayers. This observational model is used to describe the heterogeneous surface for non-uniform distribution adsorption affinities. This isotherm says that the occupation of active sites depends upon their strength, as most strong sites will be occupied first. Freundlich adsorption isotherm’s linear form is given as

$$log ,{{rm{q}}}_{{rm{e}}}=,log ,{{rm{k}}}_{{rm{F}}}+frac{1}{{rm{n}}},log ,{rm{Ce}}$$

(7)

where, KF = Freundlich constant, n = Heterogeneity factor

Linear Freundlich isotherm plots are given below in Figs. 15 and 16. The parameters of this isotherm are given in Table 4. The n value denotes the degree of nonlinearity between solution concentration and adsorption as follows: if n = 1, then adsorption is considered as linear; if n < 1, then adsorption is through a chemical process; if n > 1, then adsorption is via a physical process. The values of n = 1.35 and 1.40 for Pb(II) and Cr(VI) respectively is the most common feature of this model yielding from the distribution of surface sites or any factor that causes a decrease in sorbent-sorbate interaction with increasing density.

Both the models are in favour of adsorption for metal ions with slight preference of Freundlich adsorption isotherm for acid functionalized nanoparticles, as metals can be adsorbed by forming multi-layered surfaces on the nanosorbents due to the presence of heterogeneous active sites. It showed that adsorption process is a complex process comprising two kinds of interactions i.e. electrostatic attraction between adsorbent surfaces and metal ions and chemical bonding between the surface groups and metal ions.

The recycling of the nanosorbents makes them cost-effective in the practical applications. The recovery of nanosorbents was evaluated till four consecutive cycles and after each cycle, adsorption efficiency was decreased for both metal ions. However, still it can be used up to 4–5 cycles to remove these toxic metal ions from wastewater. The reusability results are given in Fig. 17. The possible reduction in adsorption efficiency might be either due to incomplete stripping of metal ions as evident from Langmuir and Freundlich isotherms that adsorption involves both physio-chemical interactions. Another possible reason for decreased efficiency in each cycle might be partial erosion of adsorbent surface in strong acidic or basic media used during regeneration.