E ﬀ ective Adsorption of Hexavalent Chromium and Divalent Nickel Ions from Water through Polyaniline, Iron Oxide, and Their Composites

: Water pollution caused by industrial wastes containing heavy metals and dyes is a major environmental problem. This study reports on the synthesis, characterization, and utilizations of Polyaniline (PANI) and its composites with Fe 3 O 4 for the removal of hexavalent chromium Cr(VI) and divalent nickel Ni(II) ions from water. The adsorption data were ﬁtted in Freudlich, Langmuir, Tempkin, Dubbanin–Ruddishkawich (D–R), and Elovich adsorption isotherms. The Freundlich isotherm ﬁts more closely to the adsorption data with R 2 values of 0.9472, 0.9890, and 0.9684 for adsorption of Cr(VI) on Fe 3 O 4 , PANI, and PANI / Fe 3 O 4 composites, respectively, while for adsorption of Ni(II) these values were 0.9366, 0.9232, and 0.9307 respectively. The e ﬀ ects of solution pH, initial concentration, contact time, ionic strength, and adsorbent dosage on adsorption behavior were investigated. The adsorption ability of composites was compared with pristine PANI and Fe 3 O 4 particles. Activation energy and other thermodynamic properties such as changes in enthalpy, entropy, and Gibbs free energy indicated spontaneous and exothermic adsorption.


Introduction
Rapid increases in industrialization for the production of various products such as medicine, cloth, leather, paints, cement, and fertilizers are necessary to fulfill the basic needs of people in the modern world [1][2][3][4]. The effluents coming from these industries include several types of dyes and heavy metals. Mostly these effluents are added into nearby streams, canals, and rivers without any treatment. One of the very toxic elements in these effluents is chromium. It exists in various oxidation states such as Cr(II), Cr(III), Cr(IV), Cr(V), and Cr(VI) but trivalent Cr(III) and hexavalent Cr(VI) are environmentally stable [5].
Hexavalent chromium (Cr(VI)) is about 500 to 1000 times more toxic than its trivalent form, i.e., Cr(III), because of its high solubility and high mobility in almost the whole pH range of its solution [6]. Among toxic pollutants, Cr(VI) has been listed as a priority by the U.S. Environmental Protection Agency (EPA) [7]. It exists as oxyanions (CrO 4 2− , Cr 2 O 7 2− , HCrO 4 − ) in aqueous environments, and diffuses through cell membranes into living tissues and oxidize biological molecules. Accumulation of Cr(VI) in living tissues even in a very small concentration may cause severe neurological or physiological damage. The maximum permissible concentration of Cr(VI) in water is about 0.05 mg·L −1 as accepted by the World Health Organization (WHO) [8]. Perforation of the eardrums and the nasal septum, skin Adsorption data were fitted in the Freundlich, Langmuir, Dubinin-Radishkevitch (D-R), Tempkin, and Elovich isotherm models. Kinetics of adsorption and thermodynamic parameters were determined.

Batch Adsorption Study
PANI, Fe 3 O 4 and their composites were synthesized as per our previous reports [41,42]. Solutions of both Cr(VI) and Ni(II) in different concentrations in the range of 5-100 mg·L −1 were prepared in 20 mL volumes from stock solution. A known amount of PANI/Fe 3 O 4 composite was added to these solutions and shaken for about 90 min in a water bath shaker. These solutions were then filtered and the amount of Cr(VI) and Ni(II) ions in the filtrate were determined using an atomic absorption spectrophotometer by applying Equation (1) [40].
where C 0 (mg·L −1 ) represents initial concentration and C e (mg·L −1 ) is the equilibrium concentration of heavy metals. Similarly, q e (mg·g −1 ) is the heavy metals adsorbed at equilibrium. m (g) refers to the mass of the adsorbent and V is the volume of the solution in mL. The effects of different parameters such as temperature, initial concentration, contact time, ionic strength, and pH on adsorption were studied. The data obtained from adsorption were also used to calculate the kinetics and thermodynamic properties. The same procedure was also used to study removal of Cr(VI) and Ni(II) through Fe 3 O 4 and PANI. Desorption of Cr(VI) and Ni(II) were carried out by washing the adsorbents with double distilled water several time in a filter paper. Then 0.01 M sodium hydroxide was used to remove the remaining Cr(VI) while 0.01 M HCl was applied to remove Ni(II) ions.

Characterization
Surface morphologies of all adsorbents before and after adsorption were investigated with a scanning electron microscope (JSM-6490-JEOL, Tokyo, Japan). A Fourier transmission infrared spectrophotometer (Shimadzu, Tokyo, Japan) was used to record FTIR spectra of the PANI, and Fe 3 O 4 and Fe 3 O 4 /PANI composites in the spectral range of 400-4000 cm −1 . Elemental compositions of all adsorbents were determined through energy-dispersive X-ray (EDX) spectrophotometry (Oxford, UK, Inca 200). The amount of heavy metals adsorbed on the adsorbent materials was also determined with UV-Visible spectrophotometer (PerkinElmer, Buckinghamshire, UK). The amount of Cr(VI) and Ni(II) in water before and after adsorption was determined through atomic absorption spectrophotometry (A.Analyst 700, Perkin Elmer, Waltham, MA, USA).

Characterization
Morphological and structural features of all the adsorbents materials before and after adsorption were determined through various characterization techniques as discussed below.

SEM Study
Surface morphology and average size diameter of synthesized materials were investigated through SEM. Figure 1a-c shows SEM images of Fe 3 O 4 , PANI, and PANI/Fe 3 O 4 respectively before adsorption. Iron oxides have a spherical morphology and an average particles size of 0.07 µm (70 nm) and are highly porous. It can be observed that these particles are agglomerated to each other which may be due to van der Waals forces. Existence of such forces has been observed in the earlier work [43]. PANI shows cabbage leaf like morphology, in which the polymers chains are adhered to each other through hydrogen bonding. After formation of composites with Fe 3 O 4 particles, the morphology of PANI becomes sponge like and highly porous. Although visibility of Fe 3 O 4 particles in the composites is not clear due to PANI layer formation, on close observation some bright and dark images of the core-shell like microsphere composites structures can be seen in the composite texture [44]. After adsorption of Cr(VI), the surface morphology of Fe 3 O 4 slightly changes from spherical to rough due to interactions of Cr(VI) with active sites on the surface of Fe 3 O 4 . Figure 1d shows that the average diameter of the Fe 3 O 4 particles increased to 0.18 µm (180 nm) as reported elsewhere [45]. The visibility of sphericalness became clearer when Ni(II) adhered onto Fe 3 O 4 . The average diameter of Fe 3 O 4 particles increased to 0.09 µm (90 nm) after adsorption of Ni(II). Figure 1e,h shows SEM images of PANI after removal of Cr(VI) and Ni(II) respectively. As mentioned above, PANI showed cabbage leaf like morphology before adsorption which changed into cauliflower like shape when Cr(VI) and Ni(II) ions came in contact with the surface of PANI. Similarly cloudiness in the morphology of composites after adsorption of Cr(VI) and a cluster of ball like Ni(II) particles appeared after their adsorption as shown in Figure 1f,i. Similar results were observed in an earlier work where polypyrrole was used as an adsorbent to remove nickel from aqueous solution [46]. Surface morphology and average size diameter of synthesized materials were investigated through SEM. Figure 1a-c shows SEM images of Fe3O4, PANI, and PANI/Fe3O4 respectively before adsorption. Iron oxides have a spherical morphology and an average particles size of 0.07 µm (70 nm) and are highly porous. It can be observed that these particles are agglomerated to each other which may be due to van der Waals forces. Existence of such forces has been observed in the earlier work [43]. PANI shows cabbage leaf like morphology, in which the polymers chains are adhered to each other through hydrogen bonding. After formation of composites with Fe3O4 particles, the morphology of PANI becomes sponge like and highly porous. Although visibility of Fe3O4 particles in the composites is not clear due to PANI layer formation, on close observation some bright and dark images of the core-shell like microsphere composites structures can be seen in the composite texture [44]. After adsorption of Cr(VI), the surface morphology of Fe3O4 slightly changes from spherical to rough due to interactions of Cr(VI) with active sites on the surface of Fe3O4. Figure 1d shows that the average diameter of the Fe3O4 particles increased to 0.18 µm (180 nm) as reported elsewhere [45]. The visibility of sphericalness became clearer when Ni(II) adhered onto Fe3O4. The average diameter of Fe3O4 particles increased to 0.09 µm (90 nm) after adsorption of Ni(II). Figure  1e,h shows SEM images of PANI after removal of Cr(VI) and Ni(II) respectively. As mentioned above, PANI showed cabbage leaf like morphology before adsorption which changed into cauliflower like shape when Cr(VI) and Ni(II) ions came in contact with the surface of PANI. Similarly cloudiness in the morphology of composites after adsorption of Cr(VI) and a cluster of ball like Ni(II) particles appeared after their adsorption as shown in Figure 1f,i. Similar results were observed in an earlier work where polypyrrole was used as an adsorbent to remove nickel from aqueous solution [46].

EDX Study
EDX spectra of iron oxide, PANI, and PANI/Fe3O4 composites before adsorption are shown in Figure 2a-c. These spectra show that percentages of Fe and O by weight were 65.05% and 34.95%, respectively, in Fe3O4. PANI was composed of 66.17% carbon and 16.97% nitrogen by weight. Besides carbon and nitrogen some other elements like oxygen, sulfur, iron and chlorine were also present in the PANI texture. Existence of oxygen and sulfur indicate the presence of some moiety of DBSA in PANI, while iron and chlorine may be due to ferric chloride which was used as an oxidant. In the case of composites, percent weight of iron and oxygen were increased to 9.59% and 17.92%,

EDX Study
EDX spectra of iron oxide, PANI, and PANI/Fe 3 O 4 composites before adsorption are shown in Figure 2a-c. These spectra show that percentages of Fe and O by weight were 65.05% and 34.95%, respectively, in Fe 3 O 4 . PANI was composed of 66.17% carbon and 16.97% nitrogen by weight. Besides carbon and nitrogen some other elements like oxygen, sulfur, iron and chlorine were also present in the PANI texture. Existence of oxygen and sulfur indicate the presence of some moiety of DBSA in PANI, while iron and chlorine may be due to ferric chloride which was used as an oxidant. In the case of composites, percent weight of iron and oxygen were increased to 9.59% and 17.92%, respectively, while those of carbon and nitrogen were decreased to 61.45% and 6.60%. Figure  respectively, while those of carbon and nitrogen were decreased to 61.45% and 6.60%. Figure 2d-f represent spectra recorded after adsorption of Cr(VI) on Fe3O4, PANI, and PANI/Fe3O4. Appearance of Cr and Ni indicates adherence of these elements to the adsorbent materials (Figure 2g-i). On close observation one can find a higher weight percentage of Cr(VI) than Ni(II).

UV-Visible Study
Electronic and optical properties of materials can be determined by UV-Vis spectroscopy. Figure  3a,d,g shows the UV-Visible spectra of Fe3O4, PANI, and PANI/Fe3O4 before adsorption in N-methyl pyrrolidone (NMP). In the spectrum of iron oxide, a peak at 386.89 nm is due to the interaction of incident radiation with the outer most shell electrons of Fe in iron oxide. As a result, these electrons start to oscillate with the frequency of incident radiations showing surface plasmon resonance phenomenon (SPR) [46]. Another peak at 535.42 nm shows the presence of the DBSA moiety which was used as an emulsifying agent as reported elsewhere [47].
The spectrum of PANI showed a peak at 316.11 nm which corresponds to the π−π transition of benzoinoid while another peak at 611.29 nm can be attributed to quinoid ring excitation as shown in Figure 3d. Both these peaks are the characteristics absorption bands of the emaraldine base [48]. Deprotonation of emaraldine salt may occur due to formation of hydrogen bonds between hydrogens attached to amine group of PANI and the C=O group of NMP [49].

UV-Visible Study
Electronic and optical properties of materials can be determined by UV-Vis spectroscopy. Figure 3a,d,g shows the UV-Visible spectra of Fe 3 O 4 , PANI, and PANI/Fe 3 O 4 before adsorption in N-methyl pyrrolidone (NMP). In the spectrum of iron oxide, a peak at 386.89 nm is due to the interaction of incident radiation with the outer most shell electrons of Fe in iron oxide. As a result, these electrons start to oscillate with the frequency of incident radiations showing surface plasmon resonance phenomenon (SPR) [46]. Another peak at 535.42 nm shows the presence of the DBSA moiety which was used as an emulsifying agent as reported elsewhere [47].
The spectrum of PANI showed a peak at 316.11 nm which corresponds to the π−π * transition of benzoinoid while another peak at 611.29 nm can be attributed to quinoid ring excitation as shown in Figure 3d. Both these peaks are the characteristics absorption bands of the emaraldine base [48]. Deprotonation of emaraldine salt may occur due to formation of hydrogen bonds between hydrogens attached to amine group of PANI and the C=O group of NMP [49].
Formation of PANI/Fe 3 O 4 composites can be characterized with two characteristics bands observed at 329.15 and 643.42 nm adsorption as shown in Figure 3g. Band at 643.42 nm can be assigned to the transfer of charge from the benzenoid to quinoid ring while the other band observed at 329.15 nm is attributed to the excitation of π electrons in the benzenoid ring. In the early research work, red shift is observed in the bipolaron band due to the doping of the amine group of benzenoid with Fe 3 O 4 [50].
Removal of Cr(VI) through Fe 3 O 4 , PANI, and PANI/Fe 3 O 4 composites was conducted by the UV-visible spectrum as shown in Figure 3b,e,h, while Figure 3c,f,i shows the same for Ni(II). The typical bands of the adsorbent materials were shifted to longer wavelengths, suggesting some physical interactions of these metals with active sites present on the surfaces of adsorbent materials. Cr(VI) cause more red shift as compared to Ni(II). Enhancement in Cr(VI) adsorption can be attributed to its larger size which can exist as HCrO 4 − in aqueous solution, thus causing greater London or dispersion forces [51]. The intensity of the peak at 378.11 nm increased when Cr(VI) came in contact with the surface of composite materials as shown in Figure 3h. In our already published work, it was explained that doping of PANI with Fe 3 O 4 caused an increase in the surface area of the composite resulting in the enhancement of adsorption [42].
Appl. Sci. 2020, 10, x 6 of 22 Formation of PANI/Fe3O4 composites can be characterized with two characteristics bands observed at 329.15 and 643.42 nm adsorption as shown in Figure 3g. Band at 643.42 nm can be assigned to the transfer of charge from the benzenoid to quinoid ring while the other band observed at 329.15 nm is attributed to the excitation of π electrons in the benzenoid ring. In the early research work, red shift is observed in the bipolaron band due to the doping of the amine group of benzenoid with Fe3O4 [50].
Removal of Cr(VI) through Fe3O4, PANI, and PANI/Fe3O4 composites was conducted by the UVvisible spectrum as shown in Figure 3b,e,h, while Figure 3c,f,i shows the same for Ni(II). The typical bands of the adsorbent materials were shifted to longer wavelengths, suggesting some physical interactions of these metals with active sites present on the surfaces of adsorbent materials. Cr(VI) cause more red shift as compared to Ni(II). Enhancement in Cr(VI) adsorption can be attributed to its larger size which can exist as HCrO4 − in aqueous solution, thus causing greater London or dispersion forces [51]. The intensity of the peak at 378.11 nm increased when Cr(VI) came in contact with the surface of composite materials as shown in Figure 3h. In our already published work, it was explained that doping of PANI with Fe3O4 caused an increase in the surface area of the composite resulting in the enhancement of adsorption [42].

FTIR Study
A FTIR study was conducted to detect the characteristic groups of PANI and its composites with Fe3O4 before and after removal of Cr(VI) and Ni(II) as shown in Figure 4a-i. Significant peaks observed at 536.6 and 855.9 cm −1 corresponded to the Fe−O bond stretching vibrations in the spectrum of Fe3O4, thus confirming its formation. Similarly, a peak at 3466.4 cm −1 was attributed to the stretching vibrations due to the -OH group attached to the surface of Fe3O4. The presence of DBSA in the vicinity of Fe3O4 was also confirmed by the peak at 1649.6 cm −1 which showed the stretching

FTIR Study
A FTIR study was conducted to detect the characteristic groups of PANI and its composites with  [52].
A FTIR spectrum of pristine PANI before adsorption is shown in Figure 4d. A characteristic peak showing the C−N−C bending mode in the aromatic ring appears at 574.7 cm −1 . Another peak at 667.9 cm −1 can be assigned to bending mode vibrations due to the C−C bond in the aromatic ring. Peaks at 819.6 and 1121.3 cm −1 were attributed to the bending vibrations due to out plane and in plane C−H bonds in the benzenoid ring [53]. Similarly, a peak appearing at 1309.5 cm −1 shows stretching vibrations due to −C−N + in the secondary aromatic amine. Thao et al. have synthesized polypyrrole/polyaniline composite through the interfacial polymerization method and observed stretching vibrations due to −C−N + of the secondary aromatic amine in the pristine PANI at 1297 cm −1 [54]. Peaks at 1461.4 and 1592.9 cm −1 can be assigned to C=N and C=C stretching vibrations of the quinoid and benzenoid rings of the PANI texture. The presence of DBSA moiety in the PANI backbone was also confirmed by the peak appearing at 1005.8 cm −1 , which was attributed to the stretching vibration due to S=O bond of -SO 3 H group [55]. Symmetric and asymmetric stretching vibrations due to the C-H bonds in PANI were represented respectively by two peaks at 2840. 3 [52]. A FTIR spectrum of pristine PANI before adsorption is shown in Figure 4d. A characteristic peak showing the C−N−C bending mode in the aromatic ring appears at 574.7 cm −1 . Another peak at 667.9 cm −1 can be assigned to bending mode vibrations due to the C−C bond in the aromatic ring. Peaks at 819.6 and 1121.3 cm −1 were attributed to the bending vibrations due to out plane and in plane C−H bonds in the benzenoid ring [53]. Similarly, a peak appearing at 1309.5 cm −1 shows stretching vibrations due to −C−N + in the secondary aromatic amine. Thao et al. have synthesized polypyrrole/polyaniline composite through the interfacial polymerization method and observed stretching vibrations due to −C−N + of the secondary aromatic amine in the pristine PANI at 1297 cm −1 [54]. Peaks at 1461.4 and 1592.9 cm −1 can be assigned to C=N and C=C stretching vibrations of the quinoid and benzenoid rings of the PANI texture. The presence of DBSA moiety in the PANI backbone was also confirmed by the peak appearing at 1005.8 cm −1 , which was attributed to the stretching vibration due to S=O bond of -SO3H group [55]. Symmetric and asymmetric stretching vibrations due to the C-H bonds in PANI were represented respectively by two peaks at 2840.3 and 2933.3 cm −1 . Formation of PANI/Fe3O4 composites were confirmed by a weak peak at 511.2 cm −1 which corresponded to the stretching vibration due to the Fe−O bond in Fe3O4 as shown in Figure 4g. Moreover, the intensity of peaks at 2813.1 and 2926.5 cm −1 decreased, showing some interactions between the active sites present on the PANI surface and Fe3O4 particles. Interactions between PANI and Fe3O4 particles in composite materials were also observed in the slight shift in % transmittance towards longer wavelength regions as reported elsewhere [56].  After adsorption of Cr(VI), a slight shift towards low frequency range was observed in all peaks in the spectra of Fe3O4, PANI, and composite materials thus indicating some coating of chromium on adsorbents. Moreover, in the spectrum of Fe3O4 after adsorption of Cr(VI), peaks at 3404.9 and 2969.5 cm −1 were attributed to the stretching and bending vibrations due to the O−H bond of water molecules [43]. A peak present at 1724.5 cm −1 was associated to hydroxyl group deformation After adsorption of Cr(VI), a slight shift towards low frequency range was observed in all peaks in the spectra of Fe 3 O 4 , PANI, and composite materials thus indicating some coating of chromium on adsorbents. Moreover, in the spectrum of Fe 3 O 4 after adsorption of Cr(VI), peaks at 3404.9 and 2969.5 cm −1 were attributed to the stretching and bending vibrations due to the O−H bond of water molecules [43]. A peak present at 1724.5 cm −1 was associated to hydroxyl group deformation vibrations. Similarly a peak appeared at 1044.2 cm −1 which indicated intrinsic vibrations of the Cr−O bond, thus confirming the adherence of Cr(VI) on to Fe 3 O 4 . Adsorption of Cr(VI) onto PANI was determined by the increase in intensity of the peak at 1005.2 cm −1 while in composites the same was observed at 1039.6 cm −1 as reported elsewhere [57,58]. Similar results were obtained after adsorption of Ni(II) on all adsorbent materials, however; shifting of % transmittance was less as compared to Cr(VI), indicating stronger interactions between Cr(VI) and adsorbent materials as compared to Ni(II) [59]. FTIR results correlated to the results obtained from EDX and UV-visible spectroscopy.

Adsorption Isotherm
An adsorption study mainly depends upon the interaction between adsorbents and pollutants. These interactions can be clearly understood by considering the adsorption isotherms. Various types of isotherms like Dubinin-Radushkevitch (D-R), Langmuir, Freundlich, and Tempkin isotherm models have been successfully applied by earlier researchers [60][61][62]. In the present study, the data were fitted into Langmuir, Freundlich, Tempkin, Dubinin-Radushkevitch (D-R), and Elovich isotherm models shown in Figure 5. Various adsorption parameters were calculated from these models shown in Table 1. The Freundlich isotherm is the most common adsorption model and describes the formation of multilayers by the molecules of pollutants on the surface of adsorbent materials. It also explains heterogeneity of adsorption due to different interactions between pollutants and adsorbents. A linearized form of Freundlich adsorption isotherm equation has been expressed by the following Equation (2).
where K f is constant known as the Freundlich constant, 1/n is the slope referring to the adsorbent intensity, C e (mg·L −1 ) is the remaining concentration of heavy metals at equilibrium, and q e (mg·g −1 ) shows adsorption of heavy metals by one gram of adsorbent at equilibrium. Figure 5a shows the Freundlich isotherm graph obtained by plotting lnq e vs. lnC e . From the intercept and slope of this graph the constants K F and 1/n were calculated, which relate respectively to the adsorption capacity and the adsorption intensity. reported that when the value of 1/n is smaller than one but more than zero (1 > 1/n > 0) the adsorption is highly favorable and heterogeneous, but when its values are more than unity (1 < 1/n) adsorption is unfavorable. However, if 1/n = 0, adsorption is irreversible. Similarly, when the value of 1/n is unity then system is at equilibrium and homogeneous [43].
Appl. Sci. 2020, 10, x 9 of 22 graph the constants KF and 1/n were calculated, which relate respectively to the adsorption capacity and the adsorption intensity. Table 1 shows that 1/n values for adsorption of Ni(II) were 0.2987, 0.3545, and 0.2139 while for adsorption of Cr(VI) these values were 0.3688, 0.4716, and 0.3905 respectively, on Fe3O4, PANI, and PANI/Fe3O4 composites. These results showed high favorability of adsorption over the entire range of concentrations and closely resembled already published research works [58]. Moreover, by comparing the values of 1/n after adsorption of Cr(VI) and Ni(II) ions over all three adsorbents (Fe3O4, PANI, and PANI/ Fe3O4 composites), one can deduce that Cr(VI) shows more heterogeneity as compared to Ni(II). Ren et al. has reported that when the value of 1/n is smaller than one but more than zero (1 > 1/n > 0) the adsorption is highly favorable and heterogeneous, but when its values are more than unity (1 ˂ 1/n) adsorption is unfavorable. However, if 1/n = 0, adsorption is irreversible. Similarly, when the value of 1/n is unity then system is at equilibrium and homogeneous [43]. The Langmuir adsorption isotherm is another important model which expresses that adsorption of heavy metals occurs homogeneously resulting in the formation of a monolayer without interaction between the absorbed materials. It is expressed by the following Equation (3): The Langmuir adsorption isotherm is another important model which expresses that adsorption of heavy metals occurs homogeneously resulting in the formation of a monolayer without interaction between the absorbed materials. It is expressed by the following Equation (3): where K L is constant called the Langmuir constant, q max (mg·g −1 ) refers to maximum adsorption, q e (mg·g −1 ) is the amount of heavy metals adsorbed at equilibrium, and C e (mg·L −1 ) is the remaining concentration of heavy metals when equilibrium is established. Figure 5b shows a Langmuir isotherm plot between C e /q e and C e . in aqueous environments which may cause greater interactions with positively charged sites present on the surface of adsorbents. Similar results were shown in the already published work [63].
A dimensionless constant, also called separation factors (R L ), can be calculated by using the initial concentrations of pollutants (C 0 ) and constant K L according to Equation (4) shown below: The results closely resembled previously published reports which demonstrated that adsorption was favorable if the R L value was more than zero but less than unity (1 > R L > 0), unfavorable (1 < R L ), irreversible (R L = 0), and linear (R L = 1) [64].
Adsorption data were also fitted in the Tempkin isotherm to express the binding interactions of Cr(VI) and Ni(II) with adsorbent material shown by Figure 5c. It is expressed by the following Equation (5): From the intercept of this equation, the binding constant (K T ) was calculated, which corresponds to the maximum binding energy of Cr(VI) and Ni(II) with adsorbent materials. Its values shown in the Tempkin isotherm constant b (J·mol −1 K −1 ) expresses heat of adsorption. The negative sign with the values of β showed that adsorption of Cr(VI) and Ni(II) on Fe 3 O 4 , PANI, and PANI/Fe 3 O 4 composites is exothermic (Table 1) [65,66].
The Dubinin-Radushkevich (D-R) isotherm model is also very important and has been successfully applied by the earlier researchers. Its unique feature is differentiation between physical and chemical adsorption and estimation of the energy of adsorption. The linear form of the D-R adsorption isotherm is described by the following Equation (7): lnq e = lnq s − Bε 2 (7) where q s (mg·g −1 ) represents the saturation capacity of adsorption. Figure 5d shows the D-R isotherm graph, obtained by plotting lnq e vs. ε 2 , where ε is polyanyi potential, which can be calculated according to Equation (8): where R (J·mol −1 K −1 ) is the general gas constant, C e (mg·L −1 ) refers to the equilibrium concentration, and T is the Kelvin scale temperature. Energy of adsorption can be calculated by Equation (9) as expressed below: Using Equations (7) and (9), values of q s and E ads for Cr(VI) and Ni(II) adsorbed on Fe 3 O 4 , PANI, and PANI/Fe 3 O 4 composite were calculated which indicated physical adsorption (Table 1). Similar results were shown by the already reported work which suggested that if the values of E ads are less than 40 kJ·mol −1 , the adsorption is physical, but values higher than 40 kJ·mol −1 suggest chemical adsorption [67]. Moreover, the typical range of ion-exchange process is 8-16 kJ·mol −1 [60]. In the present study, the E ads value for adsorption of both Cr(VI) and Ni(II) were not in this range, which indicate that the ion-exchange process was not involved.
The Elovich isotherm linear form is represented as follow: ln q e C e = ln(k E q max ) − 1 q max q e (10) where q max is the maximum capacity of adsorption, K E (L·mg -1 ) refers to the Elovich constant and q e is the amount of heavy metals adsorbed at equilibrium. From the slope and intercept of Figure 5e, values of q max and K E can be calculate (Table 1) [68]. R 2 values show that the Freundlich isotherm fit more closely than all other isotherms to the adsorption data.

Effect of pH on Adsorption
The pH of a solution greatly influences the behavior of adsorbents towards pollutants due to different interactions with them. The effect of pH on the removal of Cr(VI) and Ni(II) through  [70]. The effect of pH on the removal of Ni(II) ions having 50 mg·L −1 concentrations was also studied from pH 2 to 12 at 30 • C shown in Figure 6b. The graph shows that adsorption of Ni(II) was less in acidic medium, because at lower pH, hydrogen ions may start to compete with Ni(II) ions for the surface active sites of adsorbents.
But as the pH increases, the amine and imine groups in PANI as well as in composites become free due to un-doping, hence resulting in the enhancement of adsorption. The highest adsorption of Ni(II) occurred at pH 7.5, but as the pH further increased, adsorption once again decreased.

Effect of Time and Temperature on Adsorption
Contact time between adsorbents and pollutants is very important for finding the equilibrium time during adsorption experiments. In the present work, the amount of Cr(VI) and Ni(II) adsorbed as a function of time on Fe3O4, PANI, and PANI/Fe3O4 composites were determined as shown in the Figure 7a. All the experiments were carried out by adding 0.035-0.036 g of each Fe3O4, PANI, and PANI/Fe3O4 composites to Cr(VI) and Ni(II) ion solutions. The pH of the Cr(VI) solution was adjusted at 3-4 while that of Ni(II) was adjusted to 7.5. The results show that initially the rate of adsorption was very high on all the adsorbent materials due to the presence of a large number of active sites. With the passage of time, these active sites were filled with Cr(VI) and Ni(II) ions which repelled further adsorption. As a result adsorption slowed down. By close observation of Figure 7a, it is clear that point of equilibrium was reached within 60-90 min for adsorbent materials [73]. Moreover, adsorption capacity of hybrid materials was higher than pure Fe3O4 and PANI. Figure 7b shows that adsorption was high at low temperatures and goes on to decrease as the temperature increased. This behavior showed the spontaneous and exothermic nature of adsorption and strongly correlated with the results shown by the Tempkin adsorption isotherm [74].

Effect of Time and Temperature on Adsorption
Contact time between adsorbents and pollutants is very important for finding the equilibrium time during adsorption experiments. In the present work, the amount of Cr(VI) and Ni(II) adsorbed as a function of time on Fe 3 O 4 , PANI, and PANI/Fe 3 O 4 composites were determined as shown in the Figure 7a. All the experiments were carried out by adding 0.035-0.036 g of each Fe 3 O 4 , PANI, and PANI/Fe 3 O 4 composites to Cr(VI) and Ni(II) ion solutions. The pH of the Cr(VI) solution was adjusted at 3-4 while that of Ni(II) was adjusted to 7.5. The results show that initially the rate of adsorption was very high on all the adsorbent materials due to the presence of a large number of active sites. With the passage of time, these active sites were filled with Cr(VI) and Ni(II) ions which repelled further adsorption. As a result adsorption slowed down. By close observation of Figure 7a, it is clear that point of equilibrium was reached within 60-90 min for adsorbent materials [73]. Moreover, adsorption capacity of hybrid materials was higher than pure Fe 3 O 4 and PANI. Figure 7b shows that adsorption was high at low temperatures and goes on to decrease as the temperature increased. This behavior showed the spontaneous and exothermic nature of adsorption and strongly correlated with the results shown by the Tempkin adsorption isotherm [74].

Effect of Adsorbent Dosage
The amount of adsorbent dosage in the solution of heavy metals also affects the adsorption process to great extant because of different surface areas. The effect of the adsorbent dose was determined by adding 0.034, 0.054, 075, 0.1, 0.13, and 0.16 g of Fe3O4, PANI, and PANI/Fe3O4 composites to 50 mg·L −1 solutions of Cr(VI) and Ni(II) ions separately. The pH of Cr(VI) was adjusted to 3-4 while that of Ni(II) was adjusted to 7.5 and the solutions were shaken for their corresponding equilibrium time. The results obtained are shown in Figure 8, which shows that removal efficiency of both Cr(VI) and Ni(II) ions depended upon the amount of adsorbent added. Enhancement in the removal of Cr(VI) and Ni(II) ions from the aqueous solution with adsorbent doses was attributed to the larger surface area and the large number of adsorption sites [75]. The results also show that removal efficiency of both Cr(VI) and Ni(II) was same up to 0.1 g of adsorbents but when the amount of adsorbent dose exceeded 0.1 g, the adsorption of Ni(II) became constant while that of Cr(VI) still increased and became constant after 0.14 g of adsorbent. This was due to hydration; Cr(VI) exists as CrO4 −2 , Cr2O7 −2 , or HCrO −4 in aqueous solution, which causes an increase in its size or volume. Due to this reason it needs a relatively large surface area and a large number of adsorbent sites as compared to nickel which exist as Ni +2 ions predominantly. As the concentrations of Cr(VI) and Ni(II) ions are constant, Ni(II) ions finish earlier than Cr(VI) in the aqueous solution [76].

Effect of Adsorbent Dosage
The amount of adsorbent dosage in the solution of heavy metals also affects the adsorption process to great extant because of different surface areas. The effect of the adsorbent dose was determined by adding 0.034, 0.054, 075, 0.1, 0.13, and 0.16 g of Fe 3 O 4 , PANI, and PANI/Fe 3 O 4 composites to 50 mg·L −1 solutions of Cr(VI) and Ni(II) ions separately. The pH of Cr(VI) was adjusted to 3-4 while that of Ni(II) was adjusted to 7.5 and the solutions were shaken for their corresponding equilibrium time.
The results obtained are shown in Figure 8, which shows that removal efficiency of both Cr(VI) and Ni(II) ions depended upon the amount of adsorbent added. Enhancement in the removal of Cr(VI) and Ni(II) ions from the aqueous solution with adsorbent doses was attributed to the larger surface area and the large number of adsorption sites [75]. The results also show that removal efficiency of both Cr(VI) and Ni(II) was same up to 0.1 g of adsorbents but when the amount of adsorbent dose exceeded 0.1 g, the adsorption of Ni(II) became constant while that of Cr(VI) still increased and became constant after 0.14 g of adsorbent. This was due to hydration; Cr(VI) exists as CrO 4 −2 , Cr 2 O 7 −2 , or HCrO −4 in aqueous solution, which causes an increase in its size or volume. Due to this reason it needs a relatively large surface area and a large number of adsorbent sites as compared to nickel which exist as Ni +2 ions predominantly. As the concentrations of Cr(VI) and Ni(II) ions are constant, Ni(II) ions finish earlier than Cr(VI) in the aqueous solution [76].

Effect of Co-Existing Ions on Adsorption
The effluents coming out from industries contain several types of ions. These ions affect the surface properties of adsorbent materials. Sodium sulfate was chosen as an electrolyte to find the ionic strength effect on the removal of Cr(VI) and Ni(II) by Fe3O4, PANI, and PANI/Fe3O4 composites from aqueous environments. The concentration of both Cr(VI) and Ni(II) were kept constant (50 mg·L −1 ) at a 20 mL volume. The pH ofthe Cr(VI) solution was adjusted to 2-4 while that of Ni(II) was adjusted to 7.5. Figure 9 shows the continuous decrease in the removal of both Cr(VI) and Ni(II) as the ionic strength increased from 0.01 to 0.1 then became constant [77]. In an acidic environment all

Effect of Co-Existing Ions on Adsorption
The effluents coming out from industries contain several types of ions. These ions affect the surface properties of adsorbent materials. Sodium sulfate was chosen as an electrolyte to find the ionic strength effect on the removal of Cr(VI) and Ni(II) by Fe 3 O 4 , PANI, and PANI/Fe 3 O 4 composites from aqueous environments. The concentration of both Cr(VI) and Ni(II) were kept constant (50 mg·L −1 ) at a 20 mL volume. The pH ofthe Cr(VI) solution was adjusted to 2-4 while that of Ni(II) was adjusted to 7.5. Figure 9 shows the continuous decrease in the removal of both Cr(VI) and Ni(II) as the ionic strength increased from 0.01 to 0.1 then became constant [77]. In an acidic environment all active sites, like the oxygen of Fe 3 O 4 , the nitrogen present in amine, and the imine group of PANI and composites, become protonated. Under this condition chromium as well as sulfate ions also become hydrated as:  [78]. Adsorption of Ni(II) was high at pH 7.5. At this pH, all the active sites present on the surface of all adsorbents become deprotonated and behave as negatively charged sites due to presence of a lone pair of electrons, hence sulfate ions have no effect on adsorption of Ni(II) ions. But under this condition, Na + ions start to compete with Ni(II) ions thus causing a decrease in its adsorption. Similar results have also been shown by already reported work [79]. From Figure 9, one can also deduce that initially the decrease in adsorption is fast, but as the ionic strength became 0.1, the decrease became relatively less and became nearly constant. This may be due to filling of all active sites by electrolyte ions as well pollutants.
Appl. Sci. 2020, 10, x 15 of 22 ions thus causing a decrease in its adsorption. Similar results have also been shown by already reported work [79]. From Figure 9, one can also deduce that initially the decrease in adsorption is fast, but as the ionic strength became 0.1, the decrease became relatively less and became nearly constant. This may be due to filling of all active sites by electrolyte ions as well pollutants.

Kinetics of Adsorption
Knowledge of adsorption kinetics is a necessity to design kinetics models and to determine the adsorption mechanism. A number of kinetics models such as Lagergren's pseudo first order, Ho and McKay's pseudo second order, and Elovich and Weber and Morris's intra particle diffusion models have been tested by previous researchers [80].
In the present study, the adsorption data were also fitted in the pseudo first order and second order kinetics models. These models are expressed by the following Equations (11) and (12) respectively. 1 K t Figure 9. Effect of other ions on adsorption of Cr(VI) and Ni(II).

Kinetics of Adsorption
Knowledge of adsorption kinetics is a necessity to design kinetics models and to determine the adsorption mechanism. A number of kinetics models such as Lagergren's pseudo first order, Ho and McKay's pseudo second order, and Elovich and Weber and Morris's intra particle diffusion models have been tested by previous researchers [80].
In the present study, the adsorption data were also fitted in the pseudo first order and second order kinetics models. These models are expressed by the following Equations (11) and (12) respectively. log (q e − q t ) = logq e − K 1 t 2.303 (11) where q t (mg·g −1 ) and q e (mg·g −1 ) are the amount of heavy metals adsorbed after time "t" and equilibrium time respectively, K 1 and K 2 are the rate constants of pseudo first order (min −1 ) and pseudo second order (g· mg −1 min −1 ) respectively. "t" is the time in minutes and "C" is constant (mg· g −1 ). Figure 10a shows a pseudo first order graph obtained by plotting ln(qe − qt) vs.  [81]. Although all plots of Cr(VI) and Ni(II) adsorbed on all the three adsorbents have good linearity, the values of R 2 indicate that pseudo first order equations do not fit closely to the adsorption data. Therefore, a pseudo-second order kinetics model was tested for the adsorption data. Figure 10b shows the pseudo-second order plot obtained by plotting t/qt vs t. Values of the rate constant (K2) and equilibrium adsorption (q e ) were calculated from the intercept and slope respectively. The pseudo-second order model explained the behavior of adsorption over the whole range of time and closely agreed with the chemisorption mechanism. Values of R 2 indicated that the pseudo-second order kinetics model fit more closely to the adsorption data than the pseudo-first order model [82]. for the adsorption data. Figure 10b shows the pseudo-second order plot obtained by plotting t/qt vs t. Values of the rate constant (K2) and equilibrium adsorption (qe) were calculated from the intercept and slope respectively. The pseudo-second order model explained the behavior of adsorption over the whole range of time and closely agreed with the chemisorption mechanism. Values of R 2 indicated that the pseudo-second order kinetics model fit more closely to the adsorption data than the pseudofirst order model [82].

Thermodynamic of Adsorption
Adsorption behavior of Cr(VI) and Ni(II) ions on adsorbents can be investigated by studying the thermodynamic properties such as Gibb's Free energy, enthalpy, and entropy changes. In the present work, the change in the Gibb's Free energy (∆G) was calculated with the help of Equation (13) as

Thermodynamic of Adsorption
Adsorption behavior of Cr(VI) and Ni(II) ions on adsorbents can be investigated by studying the thermodynamic properties such as Gibb's Free energy, enthalpy, and entropy changes. In the present work, the change in the Gibb's Free energy (∆G) was calculated with the help of Equation (13) as shown below: Enthalpy (∆H) and entropy changes (∆S) were calculated from the slope and intercept of the Van't Hoff graph obtained by plotting lnq e /c e vs. 1/T as shown in Figure 11b. The Van't Hoff equation is expressed below: where T is the absolute temperature and R (J·mol −1 k −1 ) is the general gas constant. The results obtained are summarized in Table 3. The negative sign with values of Gibb's free energy change and enthalpy and entropy change indicates the spontaneous and exothermic nature of adsorption. Moreover, the values of Gibb's Free energy change and enthalpy change show the existence of some electrostatic interactions between pollutants (Cr +6 and Ni +2 ) and adsorbents thus showing physicochemical adsorption. This is in correlation with the previously reported work, which explain that for physical adsorption, the values of Gibb's free energy should range from −20 to zero kJ·mol −1 , but when these values are more than this limit, the adsorption will be chemical [80]. Similarly Debnath et al. demonstrated that if the values of enthalpy change range from 84 to 420 kJ·mol −1 , the adsorption will be chemical while when values of enthalpy change less than 84 kJ·mol −1 it indicates physical adsorption [83]. Besides these properties, activation energy also indicates the chemical and physical nature of adsorption. Its values have been calculated from the slope of Figure 11a obtained by plotting lnK vs. 1/T according to the Arrhenius equation as expressed below: The values of activation energy shown in the Table 3 indicate physical adsorption. This result closely agrees with already reported works which indicate that if the values of activation energy range from 5 to 40 kJ·mol −1 , the adsorption is physical. However, values of activation energy higher than 60 kJ·mol −1 indicate chemical adsorption. In the present work, the values activation energy for adsorption of Cr(VI) and Ni(II) on all the three adsorbents were less than 40 kJ·mol −1 , thus adsorption was physical. values of Gibb's Free energy change and enthalpy change show the existence of some electrostatic interactions between pollutants (Cr +6 and Ni +2 ) and adsorbents thus showing physicochemical adsorption. This is in correlation with the previously reported work, which explain that for physical adsorption, the values of Gibb's free energy should range from −20 to zero kJ·mol −1 , but when these values are more than this limit, the adsorption will be chemical [80]. Similarly Debnath et al. demonstrated that if the values of enthalpy change range from 84 to 420 kJ·mol −1 , the adsorption will be chemical while when values of enthalpy change less than 84 kJ·mol −1 it indicates physical adsorption [83]. Besides these properties, activation energy also indicates the chemical and physical nature of adsorption. Its values have been calculated from the slope of Figure 11a obtained by plotting lnK vs. 1/T according to the Arrhenius equation as expressed below: The values of activation energy shown in the Table 3 indicate physical adsorption. This result closely agrees with already reported works which indicate that if the values of activation energy range from 5 to 40 kJ·mol −1 , the adsorption is physical. However, values of activation energy higher than 60 kJ·mol −1 indicate chemical adsorption. In the present work, the values activation energy for adsorption of Cr(VI) and Ni(II) on all the three adsorbents were less than 40 kJ·mol −1 , thus adsorption was physical.

Conclusions
Fe 3 O 4 , PANI, and their composite can be applied as effective adsorbents for the removal of hexavalent chromium and divalent nickel from water. The adsorption of both Cr(VI) and Ni(II) was higher for composite material than