Adsorption of aqueous Cd(II) over a Fe₃O₄/plant polyphenol magnetic material

Heavy metals have attracted great attention as they represent threats to public health and aquatic ecological systems (Singh et al. 2014). Cadmium is normally found in wastewater derived from industrial activities such as electroplating, solders, batteries, television sets, ceramics, photography, insecticides, electronics, metal-finishing, and metallurgy (Gautam et al. 2015). The harmful effects of exposure to certain concentrations of Cd(II) include acute and chronic disorders of the kidney and liver as well as the nervous and cardiovascular systems (Zeng et al. 2015). The world-shaking ‘itai-itai’ disease in Japan was caused by Cd ingestion resulting in multiple fractures due to osteomalacia (Gautam et al. 2014). Thus, Cd is considered a very toxic heavy metal.

In order to effectively remove Cd(II), a variety of methods such as chemical precipitation, ion exchange, flotation, reverse osmosis, and electrochemical treatment have been proposed (Shah et al. 2016). One of the most promising and frequently used techniques, the adsorption method, has received attention because of its high efficiency, economy, and effectiveness (Machida et al. 2012), and common adsorbents include activated carbons (Demiral & Güngör 2016), clays (Roque-Ruiz et al. 2016), and polymeric materials (Silva & Pissetti 2014).

Plant polyphenols (PP), or tannin, are inexpensive and ubiquitous natural polymers derived from vegetal secondary metabolites of higher plants (Morisada et al. 2011). PP can be found in almost every part of higher plants (e.g. bark, wood, leaves, fruit, and roots) (Wang et al. 2013). As a natural material, PP is produced in large quantities. PP has been manufactured as an important commercial material, with a large number of applications (e.g. tanning agents, adhesives, cosmetics, pharmaceuticals, and food) (Ping et al. 2011; Wang et al. 2013).

PP is usually classified into two groups, namely hydrolysable and condensed (also named non-hydrolysable) tannin (Morisada et al. 2011). Because of the presence of multiple adjacent phenolic hydroxyls in the molecular structure, PP exhibits a specific affinity towards numerous metal ions such as Cu(II), Pb(II), and Cr(VI) (Xu et al. 2017). Therefore, PP can remove metal ions and become an effective alternative adsorbent. However, the common condensed type of tannin, such as Larix gmelinii, bayberry, black wattle and persimmon tannin, is water soluble, which makes it difficult to separate from solution after interacting with metal ions (Huang et al. 2009; Xie et al. 2016; Xu et al. 2017). In order to remove aqueous heavy metals, PP requires immobilization or insolubilization processes before being used. PP can be prepared in the form of gels, resins, and immobilized on other matrices (Bacelo et al. 2016). Polyphenol–SiO₂ hybrid biosorbent, bayberry tannin-immobilized collagen fiber, persimmon waste gel, and modified quebracho tannin resin, which are all rich in phenolic hydroxyls, can be used to remove Hg(II), Pb(II), and Cr(VI) (Huang et al. 2009; Yurtsever & Sengil 2009; Inoue et al. 2010; Copello et al. 2013). Metal adsorption on PP normally involves physicochemical interactions (e.g. electrostatic force, ion exchange, metal-ion chelation, or complexation) between the adsorbate and the adsorbent (Dodson et al. 2015). However, separation of PP adsorbents (gel, resin, or immobilized on matrices) often needs filtration or centrifugation, which may result in adsorbent losses and secondary pollution.

Magnetic adsorbents such as iron oxide materials show excellent properties and have received considerable interest because of their easy separation and regeneration (Chang et al. 2016). Magnetite (Fe₃O₄) is a ferromagnetic compound with superior magnetic properties compared with other metallic compounds. With the aim of providing magnetic adsorbents with high adsorption capacity and specific metal complexation characteristics, a number of surface modification techniques have been developed via coating with different surface stabilizers. In this sense, a good number of materials including MnO₂, SiO₂, bentonite, and polymers have been reported to be suitable for modifying the surface of Fe₃O₄. In particular, natural polymers (e.g. chitosan, starch and its derivatives), which have received significant attention, have unique physicochemical characteristics, low cost, natural abundance, and multiple reactive groups in the molecule (Crini 2005).

In this study, a magnetic Fe₃O₄ material coated with PP (Fe₃O₄/PP) was synthesized to overcome the difficulty in separating immobilized PP adsorbents and removed aqueous Cd(II). The combination of PP with magnetic Fe₃O₄ could induce effective functional groups and obtain rapid separation of the adsorbent from wastewater. The surface structure, morphology, and magnetic properties of Fe₃O₄/PP were analyzed using field emission scanning electron microscopy (FE-SEM), and vibrating sample magnetometry (VSM). The thermal stability was determined by thermogravimetric analysis (TGA). The effect of some factors, such as pH, contact time, initial metal ion concentration, adsorbent dosage, agitating speed, and coexisting ions, on the adsorption process was studied. The relevant mechanisms governing the adsorption of Cd(II) on Fe₃O₄/PP were investigated by kinetics (i.e. pseudo first-order and pseudo second-order), adsorption isotherm modelling (i.e. Langmuir and Freundlich), and thermodynamics analyses based on the experimental data. In particular, the surface properties of the Fe₃O₄/PP–Cd(II) composite formed after Cd(II) adsorption were analyzed by Fourier transform infrared spectroscopy (FT-IR) and X-ray photoelectron spectroscopy (XPS).

Raw PP powder (extracted from the bark of Larix gmelinii) was purchased from the Inner Mongolia Forestry Industrial Co., Ltd (Yakeshi, China). The naked Fe₃O₄ powder, produced from a natural magnetite ore, was obtained from Baotou Bally Ken Industrial Technology Co., Ltd (Baotou, China). Analytical-grade cadmium nitrate [Cd(NO₃)₂·4H₂O] was obtained from the Institute of Jinke Fine Chemicals (Tianjin, China). Other chemical reagents used herein were of analytical grade and used without further purification.

The PP powder was first purified following the method according to Wang et al. (2013). Briefly, 20 g of PP were dissolved in 300 mL of distilled water. The resulting PP solution was centrifuged, extracted, purified, and condensed via rotary evaporation. A purified PP powder was obtained by freeze-drying the above concentrated solution. After purification, the PP content increased from 34.9 to 44.2%. The naked Fe₃O₄ was washed with distilled water three times and subsequently dried at room temperature. The purified PP was dissolved in distilled water and the final concentration was 12.5 g/L. The Fe₃O₄/PP composite was synthesized by mixing 1.0 g of Fe₃O₄ particles with 40 mL of the PP solution for 2 h at room temperature. Using an external magnet, the Fe₃O₄/PP particles were separated from the solution and washed twice with 10 mL distilled water. The Fe₃O₄/PP composite was freeze-dried and stored at 4 °C.

The concentration of Cd(II) stock solution was 200 mg/L and dissolved [Cd(NO₃)₂·4H₂O] in ultrapure water (18 MΩ/cm) containing 0.03 mol/L of HNO₃. The Cd(II) concentrations needed in each adsorption experiment were achieved by diluting the above stock solution.

The mixture of PP and KBr at 1:100 (m:m) was pressed using a tableting machine for 15 s, and then was recorded on a Fourier transform infrared spectrometer (VERTEX 70, Bruker, Germany) from 400 to 4,000 cm⁻¹ with a resolution of 4 cm⁻¹. The morphology of the Fe₃O₄/PP composite, which was directly adhered to the sample stage with a conductive paste, was examined on a field emission scanning electron microscope (QUANTA 200 F, FEI Technologies, USA). The magnetic properties were determined with a vibrating sample magnetometer (LakeShore 7410, Lake Shore Cryotronics, USA) at room temperature under a magnetic field of ±2.0 T. The thermal stability of Fe₃O₄/PP was evaluated on a thermogravimetric analyzer (TGA Q50, TA Instruments-Waters LLC, USA) up to 800 °C and heating rate set to 10 °C/min. The X-ray diffraction (XRD) patterns were recorded on a Rigaku Dmax/2400 X-ray diffractometer (RINT 2000, Rigaku, Japan) equipped with a graphite monochromatized Cu Kα (λ = 0.1542 nm) source operating at 40 kV and 80 mA and with a scanning rate of 0.02°/min (2θ: 10–80°). The solid sample of XRD was added to an agate mortar and then placed in a sample stage for measurement after being uniformly ground. XPS analysis was carried out on an X-ray photoelectron spectroscope (PHI Quantera SXM, ULVAC-PHI, Japan) equipped with a monochromated X-ray source (Al-Kα, 1,486.6 eV photon energy) as the excitation source. The sample of XPS was spread on an aluminum foil with tape and covered with another piece of aluminum foil, and then pressed using a tableting machine for 10 s.

The adsorption kinetics were investigated in 500 mL conical flasks with 300 mL of Cd(II) solution at varying concentrations (5, 20, and 40 mg/L). After adjusting the pH to 7.0 by adding 0.1 M HNO₃ and NaOH, 3.0 g Fe₃O₄/PP were added to the Cd(II) solution. After a magnet was put at the bottom of the flasks, the aliquots were taken at determined time intervals (from 10 min to 1 h) and then filtered using a 0.45 μm water membrane. Finally, the filtrate concentration of Cd(II) was measured on an AA-6300 atomic absorption spectrophotometer (AAS, Shimadzu, Japan) provided with an air-acetylene burner.

The adsorption isotherm experiments set a series of different initial Cd concentrations and were carried out at 298 K. The dosage of Fe₃O₄/PP for varying concentrations of Cd(II) solution (50 mL, 2–40 mg/L) was fixed (0.5 g) and the pH value was adjusted to 7.0. The rotary shaker speed was fixed to 160 rpm and the adsorption system approached equilibrium after shaking for 2 h. All samples were filtered with a 0.45 μm water membrane and the Cd(II) concentration of the filtrates was analyzed by AAS. The adsorption thermodynamics study was conducted at three different temperatures (288, 298, and 308 K).

To investigate the influence of different parameters on the adsorption process, the effect of some operating parameters (e.g., pH, adsorbent dose, agitation speed, and common coexisting ions) was studied. The pH of the solution ranged from 3.0 to 8.0 by adding 0.1 M HNO₃ or NaOH aqueous solutions; 0.5 g of Fe₃O₄/PP were added to 50 mL Cd(II) solution and the concentration was 20 mg/L. The mixture was agitated at 298 K by a rotary shaker at 160 rpm for 8 h. The effect of adsorbent dosage was performed by using 50 mL of a 20 mg/L Cd(II) solution at pH 7.0. The adsorbent dosage varied from 0.5 to 3.0 g, while the agitation speed was maintained at 160 rpm for 8 h and a temperature of 298 K. To determine the effect of the agitation speed, 0.5 g of Fe₃O₄/PP were added to Cd(II) solution (50 mL, 20 mg/L) at pH = 7.0, and the agitation speed ranged from 50 to 250 rpm for 8 h at 298 K. Common coexisting ions were added at three concentration levels (5, 20, and 40 mg/L) by using Na⁺, K⁺, Mg²⁺, Cl⁻, NO₃⁻, and SO₄²⁻ ion solution. For these experiments with concentration levels, 50 mL of a 20 mg/L Cd(II) solution was used at constant Fe₃O₄/PP dosage (0.5 g), pH (7.0), agitation speed (160 rpm), and temperature (298 K) for 8 h. At each equilibrium time, Fe₃O₄/PP was separated from the upper solution with a magnet. The obtained sample solutions were filtered with a syringe filter containing a 0.45-μm water membrane. The Cd(II) concentration of the supernatant was measured by AAS. The Fe₃O₄/PP–Cd(II) complexes obtained after Cd(II) adsorption were analyzed by FT-IR and XPS spectra.

Analysis of variance (ANOVA) and least significant differences (LSD) were performed for statistical analysis using software SPSS 18.0 (SPSS Inc., Chicago, USA). The results were considered significant at a significance value (p-value) <0.05.

The surface morphology of the naked Fe₃O₄ and Fe₃O₄/PP was characterized by FE-SEM. As shown in Figure 1, both the naked Fe₃O₄ and Fe₃O₄/PP materials showed a very rough morphology. Fe₃O₄ did not show significant morphology and structural changes after PP coating. Large domains formed by blocky-shaped particles were observed, and this may result from the uneven particle size of natural magnetite. Overall, PP was bonded to the surface of Fe₃O₄ without changing the morphology and structure of Fe₃O₄.

Figure 2 shows a typical S-type magnetic hysteresis of the naked Fe₃O₄ and Fe₃O₄/PP materials. The corresponding magnetization saturation (Ms) values of the naked Fe₃O₄ and Fe₃O₄/PP were 61.3 and 59.1 emu/g, respectively. PP, being a polymer, was diamagnetic and this explained the slightly lower saturation magnetization of the coated material. The remanent magnetization (Mr) and coercivity (Hc) of the naked Fe₃O₄ and Fe₃O₄/PP (right inset, Figure 2) were Mr = 6.22 and 5.09 emu/g and Hc = 67.22 and 70.21 Oe, respectively. Mr and Hc were non-zero values, revealing ferromagnetic properties for both Fe₃O₄ and Fe₃O₄/PP. The high magnetism showed by Fe₃O₄/PP after the coating process allowed quick separation from the liquid phase with an external magnet.

The TGA profiles of the naked Fe₃O₄ and Fe₃O₄/PP materials were obtained to investigate their thermal properties. The percentage of weight loss of different materials increases with increasing temperature. Therefore, the TGA curve can express the thermal stability of the sample and the coating amount of the material. As depicted in Figure 3, Fe₃O₄ showed a weight loss of only 0.13% from 28 to 200 °C due to release of physically adsorbed water, while the weight loss from 200 to 600 °C may result from the decomposition of the crystalline structure. A weight loss of 0.44% was obtained in the 600‒800 °C range, and this weight loss was related to the decomposition of the magnetite residue and the thermal decomposition of the crystal. This slight mass loss was indicative of the good thermal stability of Fe₃O₄. With regard to Fe₃O₄/PP, the initial weight loss (0.21%) in the 28‒200 °C range was produced by the release of water. An increase in temperature to 300 °C resulted in weight loss produced by the decomposition of organic compounds of the PP coating. The mass loss linearly decreased with temperature up to 800 °C, indicating that the PP was firmly anchored onto the Fe₃O₄ surface. As revealed by the weight of the material at 800 °C, PP was decomposed and carbonized at this temperature. The PP on the surface of the Fe₃O₄ also began to decompose when the temperature exceeded 120 °C (Wang et al. 2018). This indicated that the activity of the prepared Fe₃O₄/PP particles was not affected unless at high temperature. A comparison of the TGA curves of Fe₃O₄ and Fe₃O₄/PP revealed a PP coating loading of approximately 1.53%.

XRD was used to determine the crystalline structure and provided physical-chemical information on the PP coated on the magnetic material. Figure 4 shows that PP generated a broad peak at 21°, indicating its amorphous nature (He et al. 2012). For Fe₃O₄, the characteristic peaks agreed well with the standard XRD pattern of the JCPDS (No.74-0748) card. The XRD pattern of Fe₃O₄ showed seven peaks at 18.29, 30.08, 35.44, 43.06, 53.43, 56.96, and 62.54°, corresponding to Miller indices of (111), (220), (311), (400), (422), (333), and (440), respectively, typical of a cubic inverse spinel structure (Venkateswarlu & Yoon 2015). Similar peaks were observed for the Fe₃O₄/PP, suggesting that the coating of PP did not result in any phase change of Fe₃O₄.

The adsorption isotherms reflect the affinity between an adsorbent and an adsorbate at a particular temperature. Thus, it is necessary to study the adsorption at equilibrium via the isotherm equation. As shown in Figure 6(a), the removal capacity increased with the initial Cd(II) concentration (2–20 mg/L) and levelled off thereafter (above 25 mg/L), gradually reaching the saturation capacity of Fe₃O₄/PP. From statistical analysis, there was a significant difference when the concentration was less than 20 mg/L (p < 0.05). When the initial Cd(II) concentration was 20 mg/L, the saturated adsorption capacity of Fe₃O₄/PP was 0.722, 0.797, and 1.068 mg/g at 288, 298, and 308 K, respectively. Moreover, the three temperatures all had significant differences for the adsorption of Cd(II) (p < 0.05).

Studying the pH value of the metal solution was important and it had an obvious effect on the surface charge and ionic state of the adsorbent (Xu et al. 2016). Thus, the pH was investigated in the 3.0‒8.0 range. As shown in Figure 7, the adsorption of Cd(II) depended highly on the pH, which had a significant effect on the removal of Cd(II) (p < 0.05). Thus, the amount of Cd(II) adsorbed increased sharply from 0.0584 to 0.529 mg/g upon increasing the pH from 3.0 to 5.0 and increased slowly thereafter (pH = 5.0–7.0). The maximum adsorption was obtained at pH = 7.0 and decreased thereafter with pH increasing to 8.0. The K_sp of Cd(II) (4.47 × 10⁻¹⁵ at 25 °C) indicates that Cd(II) can form solid metal hydroxide precipitates at pH higher than 8.0. Thus, the experiments carried out herein were limited to the 3.0‒8.0 range of pH.

Since Cd²⁺ is the predominant species in the 3.0–8.0 pH range, the distribution of Cd(II) species at different pH values had a negligible effect on the adsorption capacity (i.e. the surface charge of Fe₃O₄/PP controlled the effect of the pH). In addition, the pH observably determined the degree of deprotonation of the phenolic hydroxyls of PP. These groups can chelate with metal ions to form five-member chelating ring species (Kim et al. 2012). The degree of deprotonation of the adsorbent can be further explained by the surface charge of Fe₃O₄/PP and the point of zero charge (pzc). From the inset of Figure 7, Fe₃O₄/PP showed a pH_pzc of 5.3. When pH < pH_pzc, Fe₃O₄/PP was positively charged such that cationic Cd(II) species were electrostatically repelled. At lower pH values, H⁺ was in higher concentration and Cd(II) ions and H⁺ competed for the available adsorption sites, thereby additionally hindering Cd(II) adsorption. When pH > 5, the adsorption of Cd(II) was favored because the adsorbent was negatively charged (i.e. a higher number of phenolic hydroxyls on PP were ionized). As a result, the electrostatic attraction between Fe₃O₄/PP and Cd(II) ions increased, resulting in the formation of Cd–ligand magnetic complexes. The adsorption capacity of bayberry or black wattle tannin increased with the pH of the solution (Huang et al. 2009; Xu et al. 2017). The larch tannin (a condensed tannin) used herein also contains multiple adjacent phenolic hydroxyls, which acted as the main complexing sites for Cd(II).

To investigate the effect of the adsorbent dose, the adsorbent dosage varied from 0.5 to 3.0 g. As shown in Figure 8(a), the removal percentage increased gradually with the mass of adsorbent, mainly due to the higher number of active sites that were available for the removal of Cd(II) (El-Hamshary et al. 2014). It can be seen that, with the amount of Fe₃O₄/PP increasing from 0.5 to 3.0 g, the removal of Cd(II) significantly increased (p < 0.05) from 36.6 to 96.3% and the adsorption capacity significantly decreased (p < 0.05) from 0.734 to 0.400 mg/g. At lower adsorbent dosage, the number of active adsorption sites was limited and Cd(II) competed with the functional groups for these sites. When the amount of adsorbent was further increased to 3.0 g, the removal percentage reached a maximum and the number of active adsorption sites nearly reached saturation value. This phenomenon, whereby the removal percentage increases gradually to saturation with the increase of adsorbent dose, has also been observed for other adsorbents (e.g. cashew nut shell resin and hydrous manganese dioxide) (Zhai & Wang 2015; Devi et al. 2017).

The effect of the agitation speed was investigated by changing within the 50‒250 rpm range. As shown in Figure 8(b), the adsorption capacity increased from 0.267 to 0.665 mg/g upon increasing from 50 to 200 rpm. When setting the agitation speed at 250 rpm, the adsorption capacity decreased to 0.587 mg/g. Overall, there were significant differences (p < 0.05) in the range of 50‒200 rpm while 250 rpm exhibited no significant difference (p > 0.05) compared with other agitation speeds. This result can be explained in two aspects. On the one hand, increasing the agitation speed resulted in higher convective mass transfer of metal ions from the bulk solution. On the other hand, higher agitation speed decreased the boundary layer resistance close to the adsorbent surfaces (Murithi et al. 2014). In the case of the study herein, agitation made the contact of the metal ions with the surface of the adsorbent more sufficient, increasing the amount of Cd(II) adsorbed as a result.

The influence of the presence of coexisting cations (e.g. Na⁺, K⁺, and Mg²⁺) on the adsorption process is summarized in Figure 9(a). In summary, the presence of Na⁺, K⁺, and Mg²⁺ hindered the adsorption of Cd(II) at the three ion concentration levels studied herein, which had a significant influence (p < 0.05) on the removal percentage compared with the control, this being particularly true for Mg²⁺. Na⁺ significantly inhibited the adsorption of Cd(II) at low concentration (5 mg/L), whereas K⁺ and Mg²⁺ reduced the Cd(II) removal percentage at higher concentration (40 mg/L). The influence of the cations followed the trend Mg²⁺ > K⁺ > Na⁺. The removal efficiency decreased upon increasing the K⁺ and Mg²⁺ concentrations, since these metal cations competed for the active sites of the adsorbent with Cd²⁺ (Chen et al. 2017). This competition was particularly strong for divalent metal ions (Ca²⁺ or Mg²⁺), which have same charge as Cd(II) (Li et al. 2012), explaining the higher hindering effect of Mg²⁺ compared with K⁺ and Na⁺.

From Figure 9(b), the presence of Cl⁻, NO₃⁻, and SO₄²⁻ all made a notable difference (p < 0.05) compared with the control at different concentration levels. Cl⁻ hindered Cd(II) adsorption at the three ion concentration levels used herein, whereas NO₃⁻ favored Cd(II) adsorption when present at concentrations of 5 and 40 mg/L. SO₄²⁻ increased the Cd(II) removal percentage at lower concentration (5 mg/L). Cl⁻ is a strong binding ligand in solution via formation of CdCl₂ and CdCl⁺ species (Yao et al. 2014). Thus, the adsorption of Cd ions was suppressed in the presence of Cl⁻. The enhanced Cd(II) adsorption at low NO₃⁻ concentration (5 mg/L) can be explained by the anion serving as a ligand and presenting a co-adsorptive effect on the adsorbent (Barot & Bagla 2012). Anions such as NO₃⁻ and SO₄²⁻ hindered the adsorption of Cd(II) at higher concentration, and this was produced by the higher anionic strength of the metal solution. Thus, the higher anionic strength of the solution affected the activity coefficient of metal ions, limiting the transfer of Cd(II) to the surface of the adsorbent (Chen et al. 2009).

FT-IR can obtain information on certain functional groups present in the structure of the materials and investigate the possible adsorption mechanism. The FT-IR spectra of Fe₃O₄, Fe₃O₄/PP, Fe₃O₄/PP-Cd, and PP are shown in Figure 10. The peak at 570 cm⁻¹ belonged to the Fe–O stretching vibration, which was also observed for Fe₃O₄/PP and Fe₃O₄/PP–Cd. PP, a natural polymer, showed a peak at about 2,927 cm⁻¹ belonging to O–H stretching. The peak corresponding to C–O stretching appeared at 1,100 and 1,200 cm⁻¹. The spectrum of Fe₃O₄/PP (after PP coating) revealed the presence of C = C vibration (resonance) from the benzene ring. The peak corresponding to this vibration shifted from 1,619 to 1,632 cm⁻¹ after PP coating. In addition, the large and intense band of O–H from PP shifted from 3,408 to 3,439 cm⁻¹ after PP coating. Therefore, Fe₃O₄ was successfully coated with PP, revealing that the hydroxyl functional groups of PP were covalently bonded to the surface of the magnetite. A new peak could be seen from the spectrum of Fe₃O₄/PP–Cd (after adsorption of Cd(II)) at 1,058 cm⁻¹, which was tentatively ascribed to the Cd–OH band. It is concluded that the formation of coordination bonds with the hydroxyl functional groups was the main interaction for removal of Cd(II).

XPS was used to identify the surface elements' composition and investigate the adsorption mechanism by analyzing the binding energy changes. The XPS spectra of Fe₃O₄/PP and Fe₃O₄/PP–Cd were obtained to further illustrate the interaction between Cd(II) and Fe₃O₄/PP. The wide scans of Figure 11(a) clearly show Cd 3d peaks at 411.6 and 404.7 eV, also indicating successful adsorption of Cd(II) by Fe₃O₄/PP (Wu et al. 2015). The detailed O 1s spectra of Fe₃O₄/PP were deconvoluted into two peaks at 530.0 and 531.5 eV (Figure 11(b)). This is mainly due to metal oxide lattice (Fe–O) and surface hydroxyl (H–O) oxygen species, respectively (Cao et al. 2012). After the Cd(II) was adsorbed on Fe₃O₄/PP, the metal oxide lattice oxygen shifted from 530.0 eV to 530.2 eV, while surface hydroxyl groups shifted slightly down by 0.1 eV to 531.4 eV (Figure 11(c)). The decreased binding energy of H–O and the increase of the binding energy of lattice O were indicative of a complexation process between the surface hydroxyl groups and Cd(II) (Zhong et al. 2016). As shown in Figure 11(d), the deconvolution of the Cd 3d band resulted in two peaks at 411.3 and 404.6 eV, which were tentatively ascribed to CdO, CdO₂, Cd(OH)₂, or CdCO₃ complex species (Ramrakhiani et al. 2017). Based on the above analysis, Cd(II) was adsorbed on Fe₃O₄/PP via complexation with hydroxyl groups. In summary, electrostatic interaction (attraction and repulsion) and complexation processes may be involved in the adsorption of Cd(II) on Fe₃O₄/PP.

The raw material of PP powder used in this study is extracted from bark, which is widely distributed in nature. The yield of PP as a kind of natural polymer is large in developing countries. Furthermore, PP is inexpensive as a natural biomass and contains many specific reactive functional groups in its structure. Therefore, it is possible to use PP as a heavy metal adsorbent and it is economically feasible in some developing countries.

The composite adsorbent uses Fe₃O₄ as the core, which gives it superior magnetic properties and can be rapidly separated and recovered from the solution. In general, the separation of common adsorbents from metal solution requires a series of operations (such as filtration, centrifugation, etc.) which increases the treatment cost and the adsorbents cannot be recycled for further use. With the combination of PP and Fe₃O₄, effective adsorption capacity can be obtained while achieving rapid and easy separation. In addition, compared with other methods, the adsorption method has high efficiency, low cost, and is environmentally friendly.

In this study, the amount of PP coated on the surface of Fe₃O₄ was slightly low and it is necessary to increase the load mass of PP in future research. Fe₃O₄ is obtained from natural magnetite ore and particle size distribution is uneven. Chemical reactions can be considered to control the synthesis process of Fe₃O₄ and reduce particle size, then increase the adsorption efficiency. Actual industrial wastewater is a mixed solution with various heavy metal ions, therefore the removal of multicomponent heavy metals using Fe₃O₄/PP needs to be further explored. In the desorption experiment, the desorption efficiency achieved 75.20% after using 1.0 M HNO₃ and ultrasonication. The degree of corrosion of Fe₃O₄ in 1.0 M HNO₃ was negligible (0.023%), which may result from the property of the natural Fe₃O₄. In order to reuse the Fe₃O₄ it was recoated with PP and the adsorption capacity of Fe₃O₄/PP slightly decreased from 0.70 mg/g to 0.55 mg/g, but stable desorption regeneration was still realized. The slight decrease may be due to the Cd(II) adsorbed on Fe₃O₄. Moreover, the desorbed compound of PP-Cd(II) can be further applied due to the decomposition of PP at high temperature. The concentrated Cd(II) eluent can be used in further industrial processes. The Fe₃O₄/PP achieved good adsorption of Cd(II) in this study and can be recycled easily, which is important from the perspective of cost and practical application. It also indicates the practical application of heavy metal removal from wastewater in developing countries.

A Fe₃O₄/PP magnetic adsorbent was successfully synthesized by a simple impregnation method. PP was coated on the surface of Fe₃O₄ with an approximate loading of 1.5%. The presence of multiple adjacent phenolic hydroxyls on PP as functional groups allowed effective removal of Cd(II). The adsorption kinetics followed a pseudo second-order model, indicating that the Cd(II) adsorption was controlled by chemical processes. The adsorption isotherm fitted well with the Langmuir model. Thus, the adsorption of Cd(II) was favorable and was carried out by forming a homogeneous monolayer. The adsorption capacity of Cd(II) on Fe₃O₄/PP depended highly on the pH (i.e., the adsorption capacity increased with the pH in the 3.0‒8.0 range). The adsorption of Cd(II) was suppressed in the presence of cations and Cl⁻, whereas NO₃⁻ and SO₄²⁻ favored the adsorption of Cd(II) at low concentration (5 mg/L). Thermodynamic studies demonstrated that the Cd(II) adsorption process was spontaneous and endothermic. The adsorption mechanism was mainly controlled by the formation of Cd(II) complexes.

This work was financially supported by the Chinese National Natural Science Foundation (51672028), Fundamental Research Funds for the Central Universities (2015ZCQ-HJ-02), and Major Science and Technology Program for Water Pollution Control and Treatment (2018ZX07105-001, 2018ZX07110-004).