3.1. Structure, morphology and physical properties
In Fig. 1, the diffraction peaks of (220), (311), (400), (442), (511), (440) of the Fe0.90Zn0.10Fe2O4 (S0) sample were completely fitted with the standard diffraction pattern of Fe3O4 as in our previous publication [31], and demonstrated the centered face cubic structure. Similarly, the XRD patterns of S2 and S3 samples showed the same diffraction peaks as that of Fe3O4 sample, which proved that the Zn2+ doping as well as the addition of polymer did not affect the crystal structure of materials. Thus, Fig. 1 presented typical S2 and S3 samples of all samples with different PANI amount. From diffraction conditions, 2dsinθ = nλ, the lattice constants of the nanoparticles in the Fe0.90Zn0.10Fe2O4, S2 and S3 samples were the same with a = 8.376 A0 which was determined by the relation: \(a={d}_{hkl}\sqrt{{h}^{2}+{k}^{2}+{l}^{2}}\); where dhkl is the distance of the lattice planes of Fe0.90Zn0.10Fe2O4. Because PANI polymer is an amorphous material and does not affect the crystal structure of Fe0.90Zn0.10Fe2O4, thus the calculated particle size from XRD patterns is also the particle size for S1, S2 and S3 samples. Thus, the average crystal particle sizes of Fe0.90Zn0.10Fe2O4 in all S0, S1, S2 and S3 samples were about 12 nm, calculated by the formula D = 0.9λ/βcosθ (λ = 1.5416 A0; β: full width at haft maximum of diffraction line). In Fig. 2a, the TEM image of S0 sample shows the grain sizes of ca. 10 -18 nm, however, Fig. 2b show the agglomeration in S1 sample with grain size of ca. 40 -50 nm. Meanwhile, in SEM images in Fig. 2c and Fig. 2d, the grain sizes in S2, S3 samples were about 35 - 50 nm and 40 – 50 nm, respectively. This result proved that PANI prevented the agglomeration of the magnetic nanoparticles. Simultaneously, it also suggested that S1, S2 and S3 samples have the core-shell structures created by the polymerization of PANI at the surface of Fe0.90Zn0.10Fe2O4 nanoparticles.
3.2. FTIR spectra analyses
Fig. 3a shows strong absorbed peaks of PANI in the region from 1149 to 1570 cm-1 with the highest intensity at 1149 cm-1 peak. These peaks were attributed to the absorption of the C = C (aromatic), C-N+, C = N groups of the PANI [37]. Meanwhile, Fig. 3b shows the strong adsorbed peak of Fe0.90Zn0.10Fe2O4 at 570 cm-1, and Fig. 3c shows some characteristic peaks of both samples.
The infrared spectrum of S2 (Fig. 3c) showed the absorption peaks appeared in the range of 1134 to 1173 cm-1 with much stronger intensity than some peaks of Fe0.90Zn0.10Fe2O4. This suggests that PANI is present in the component of S2.
3.3. TGA and DTG analyses
TGA curve in Fig. 4a showed that at temperature range below 80 0C, the volume reduction of PANI was due to the evaporation of water, corresponding to 13% with a sharp endothermic peak at 45 0C in the DTG curve. From 80 - 210 0C, the sample volume stays almost unchanged. However, from 210 - 320 0C, the reduction of sample volume in TGA curve happened due to the decomposition of PANI to form monomers, oligomers, dimers and trimers corresponding to a sharp endothermic peak at 292 0C in the DTG curve. At temperature higher than 320 0C, the weight loss caused by the thermal decomposition of the oligomers, dimers, trimers. This led to a complete decomposition of PANI at above 600 0C.
TGA curve of typical S2 sample (for S1 and S3) in Fig. 4b also showed that at temperature range below 80 0C, the 8.6% reduction of sample volume due to water evaporation corresponding to a sharp endothermic peak at 45.2 0C in the DTG curve. However, from 100 - 300 0C in the TGA curve, the 3.3% reduction of sample volume may be happened due to the decomposition of the residual monomers and oligomers in the sample. From 300 - 600 0C, the thermal decomposition of the oligomers, dimers, trimers caused a small and broad endothermic peak at 395 0C in the DTG curve. Thus, the sample mass remained only 30% at 600 0C. Because PANI is quite stable in the temperature range below 150 oC, it can be seen from above results that the S1, S2 and S3 nanocomposites have been successfully synthesized and the presence of PANI in these samples has improved the thermal stability of PANI-coated Fe0.90Zn0.10Fe2O4 in temperature range below 150 0C.
3.4. Magnetization and chemical instability
In the air environment, Fe3O4 is easy to be oxidized to give γ-Fe2O3. Thus, the saturation magnetization of new synthesized Fe3O4 samples decreased from 63.13 emu/g to 56 emu/g after 2 months of synthesis [30]. This phenomenon related the oxidation of Fe3O4 into γ-Fe2O3 due to oxygen in air according to the equation:
4Fe3O4 + O2 → 6Fe2O3
The change of magnetization and chemical instability of the Fe3O4 sample led to a limitation of their application ability. Thus, the Zn2+ substitution for Fe2+ ions and the polymer coating outside nanoparticles are the optimal methods to stabilize chemical properties and magnetization of Fe0.90Zn0.10Fe2O4 material.
For the Fe0.90Zn0.10Fe2O4 sample in Fig. 5, the saturation magnetization of 71.5 emu/g is higher than that of Fe3O4 [36] (ca. 63 emu/g). Unlike the replacement of transition metal elements Cu, Mn, Ni for Fe2+ into octahedral B-site [24, 25, 27], preferably the replacement of 𝑍n2+ ions for Fe3+ at tetrahedron A-site occurs [19, 26], simultaneously 𝐹𝑒2+ amount at the octahedral B-site gives electrons to become Fe3+ in order to balance the charge of cell network [19, 26]. Since the ionic radius of the Fe3+ ion is 0.64 Å and the ionic radius of the Zn2+ ion is 0.74 Å, the substitution of Zn with content x = 0.10 caused a slight increasing of the lattice constant (a = 8.386 A0) in comparison with Fe3O4 (a = 8.3760 A0) [30, 36]. This was due to the Zn2+ size larger than Fe3+. This substitution and transformation caused a change in ions distribution of the network subdivisions to [Znx2+Fe1-x3+]A[Fe3+Fex3+Fe1-x2+]B.
The magnetic moment depends on the content of x that is considered due to the decision of spin direction of Fex3+ in B-site, but do not depends on the presence of non-magnetic Zn ions [19, 31]. This has been clearly explained in our recent publication [31]. Therefore, the saturation magnetization of Fe0.90Zn0.10Fe2O4 is higher than Fe3O4 [30, 31].
On the other hand, the saturation magnetizations of S1, S2 and S3 samples decreased from 65 emu/g to 43 emu/g when the non-magnetic PANI content increased from 5–15% (Table 2). However, due to the PANI coating, the magnetization of nanocomposite materials is more stable over time.
Table 2
Lattice constant, grain size, Ms of S0, S1, S2 and S3
|
Lattice constant (A0)
|
Dx-ray (nm)
|
DTEM (nm)
|
DSEM (nm)
|
Ms (Oe)
|
S0
|
8.385
|
11.81
|
10-18
|
|
71.5
|
S1
|
|
|
|
40-50
|
65
|
S2
|
8.386
|
11.80
|
|
35-50
|
53
|
S3
|
8.386
|
11.81
|
|
40-50
|
43
|
3.5. Adsorption kinetic, porous properties and arsenic adsorption ability
The adsorption kinetic of S0, S1, S2 and S3 nanocomposites can be explained [30, 36] by the relation of adsorption ability based on the inelastic exchange interaction between specific surface area of nanoparticles and adsorbed materials. The surface and structure of nanoparticle mesopores were studied by the nitrogen adsorption-desorption isotherms of 0.54 g for S0, S1, S2 and S3 samples at 77 K [30, 36].
Collision of N2 gas molecules with nanoparticles is considered to be inelastic, so that the N2 gas molecules remain in contact with the nanoparticles for a time before returning to the gas phase. This time delay is taken as responsible for the phenomenon of adsorption that demonstrated by equation: P/Va(P0− P) = (1/Vm)(P/P0) [38]. Here, Va is the quantity of N2 gas adsorbed at pressure P and Vm is the quantity of gas adsorbed when the entire surface is covered with a mono-molecular layer. The N2 adsorption–desorption isotherm curves of S0 and S1 samples at 77 K were presented in Fig. 6. It can be clearly seen that the adsorption and desorption ability of S1 sample is higher than that of S0. By the BET (Brunauer, Emmett, and Taller) theory [38], the pore size distribution at relative pressure P/P0 ≈ 1 and the specific surface area at low P/P0 were calculated as shown in Table 3.
Table 3
Porous properties and maximum arsenic adsorption ability
Sample
|
BET pore size (nm)
at P/P0 ≈ 1
|
BET specific surface area
± 0.02 (m2/g) at P/Po = 0.294
|
Maximum adsorption capacity
qmax (mg/g)
|
S0
|
9.16 ± 0.02
|
64.83 m²/g
|
41.49 ± 0.02
|
S1
|
7.88 ± 0.05
|
105.89 m²/g
|
43.48 ± 0.03
|
S2
|
8.42 ± 0.03
|
98.25 m²/g
|
40.06 ± 0.02
|
S3
|
8.36 ± 0.04
|
99.67 m²/g
|
34.48 ± 0.03
|
3.5. Arsenic adsorption ability
In order to study As adsorption ability of the nanoparticles, the effects of pH in environment and As maximum adsorption capacity also investigate at room temperature.
Effects of pH to arsenic adsorption
To study the pH effect on the adsorption ability of the nanocomposites, the As(III) solutions with different pH in range of 1-14 were prepared. The As(III) adsorption results were presented in Fig. 7. It can be seen from Fig. 7 that the remaining arsenic content was a function of the pH. For all nanocomposite samples, the arsenic adsorption capacity increased as the pH was increased from 1 to 7. Then, the adsorption capacity decreased as the pH was increased above 7. The highest adsorption capacities were obtained in the range of pH 5-9. In both strong acidic and basic solution, the adsorption capacity decreased.
This trend can be explained by the speciation of arsenic (III) at different pH media and the surface charge status of the nanocomposites [39] in strong acidic environment. The surface of nanocomposites will be negatively charged at pH higher than pHpzc (~7). Meanwhile, in neutral media, the un-charge in surface of nanocomposites and As(III) occurs mostly in neutral state (H3AsO3 at pH below 9.2) [40], so at pH 7, the highest As(III) adsorption occurred, due to the electrostatic interaction is not feasible under this condition. Therefore, the As(III) adsorption was controlled by the surface complexation rather than the electrostatic interactions. Similar results were observed with other magnetic nanocomposites as reported earlier [39, 40].
When the pH elevates, As(III) exists mainly in form of H2AsO3−, HAsO32− and AsO33− anions while the surface charge of the nanoparticles is negative. Thus, the sharp decrease in the arsenic adsorption capacity at pH range of 11-14 is likely due to the electrostatic repulsion between the negatively charged surface of the nanocomposites and the deprotonated anionic arsenic. At extreme high pH 14, due to strong electrostatic repulsive force, the material has depicted no arsenic adsorption ability.
Moreover, at very low pH levels (1-2), a decomposition of the Fe0.90Zn0.10Fe2O4 nanocomposites was observed proving by the presence of iron and zinc in the solution. In neutral and alkaline media, the nanocomposites were stable with no iron and zinc ions detected in the solution. Thus, it’s suggested that the de-adsorption process of S0, S1, S2 and S3 should be conducted in solution at pH 14.
As above analyzed results, the best As adsorption occurs in a neutral environment (pH7), where the inelastic exchange interaction takes place, which has the source of the Van der Waal interaction between the magnetic nanoparticles and adsorbent. Thus, maximum arsenic adsorption capacity was investigated by the Langmuir isotherm mode at 300 K in environment with pH 7.
Maximum arsenic adsorption capacity
Equilibrium time of As(III) adsorption was analyzed by measuring the remaining arsenic content in solution pH 7. As shown in Fig. 8, the arsenic contents remain in the equilibrium state when the adsorption time is 20 minutes at room temperature. The maximum arsenic adsorption capacity qmax of unit volume of adsorbent (mg/g) is calculated by the Langmuir isotherm equation at pH 7 and 300 K [6, 10, 17] with the first order linear relation:
where Cf: the remaining arsenic content (mg/L) at equilibrium; q: the arsenic adsorption ability at equilibrium for a unit volume of adsorbent (mg/g); b: constants attributed to the interaction of the adsorbent and adsorbed compounds.
The calculated values of qmax for S0, S1, S2 and S3 are presented in Table 3.
As in Table 3, the arsenic adsorption capability of the materials, qmax of S1 was better than that of S0 (Fe0.90Zn0.10Fe2O4) sample and Fe3O4, that reported in [36] with the same condition. The results in Table 2 also are in good agreement with those discussed in Fig. 8.