Synthesis and application of surface-modified NiFe nanoparticles as a new magnetic nano adsorbent for the removal of nickel ions from aqueous solution

Heavy metal ions are widely produced in many industries such as galvanization, metallurgy, leather making, chemical industry and printing dyeing, etc. Discharge of wastewater containing heavy metal ions into the water system has caused serious environmental influence. Through drinking and crop irrigation, the polluted water has led to a risk to public health (Peng et al. 2004). Thus, effectively and thoroughly removing heavy metal ions from the water system is an important task with great challenges.

During the past decades, many methods have been developed to remove hazardous pollutants from wastewaters, such as coagulation/flocculation (Ahammed et al. 2014), chemical precipitation (Zhu et al. 2007), membrane filtration (Zheng et al. 2013), catalytic and photocatalytic oxidation (Copper & Burch 1999; Larachi et al. 2001), chlorination (Sinkkonen et al. 1997), reverse osmosis (Arzuada et al. 2008) and adsorption (Anbia & Lashgari 2009). Among these techniques, adsorption has attracted great interest and has become to the most common method for heavy metal removal because it is simple and cost-effective.

The traditional adsorbents, such as activated carbon (Guo et al. 2010), chitosan (Borsagli et al. 2015), zeolites (Nezamzadeh-Ejhieh & Kabiri-Samani 2013), and clays (Kaur et al. 2015), suffered from low efficiency. Nowadays, nanomaterials have emerged as a novel probe because of their excellent adsorption ability resulting from their small size and large surface area (Hua et al. 2012). For examples, Afkhami et al. (2010) developed 2,4-dinitrophenylhydrazine modified γ-Al₂O₃ for removal of Pb²⁺, Cd²⁺, Cr³⁺, Co²⁺, Ni²⁺ and Mn²⁺ where the adsorption capacity of Ni²⁺ was 18.18 mg/g; Tarasevich et al. (2001) also synthesized modified γ-Al₂O₃, achieving an adsorption capacity of 176.1 mg/g for Ni²⁺. More nanomaterials, such as amino-functionalized silica nano hollow spheres and silica gel (Najafi et al. 2012), nano-calcium titanate (Zhang et al. 2011), nano manganese dioxide (Fan et al. 2005; Su et al. 2010), nanosized titanium oxides (Engates & Shipley 2011), nanosized magnesium oxides (Gao et al. 2008), nanosized zinc oxides (Kikuchi et al. 2006) and carbon nanotubes (Chen & Wang 2006), have been reported as heavy metal ion adsorbents with high adsorption efficiency recently.

However, another challenge has to be faced. Nanosized adsorbents are hard to recycle, while magnetic adsorbents with the characteristic of easy separation are increasingly being recognized (Ambashta & Sillanpää 2010). The most common studies on magnetic adsorbents focus on ferric oxide (Gupta & Nayak 2012), ferriferous oxide (Yuan et al. 2010) and Fe (Boparai et al. 2011) because of their satisfactory capacity for adsorption of toxic matter from aqueous solution (for example, the adsorption of capacities of Ni²⁺ was 23.6 mg/g (Hu et al. 2006). Nevertheless, these particles easily lose their magnetism due to oxidization in air or in water.

Magnetic NiFe alloy nanoparticles exhibit improved oxidation resistance. Different to similar NiFe alloy particles that we reported previously (Liu et al. 2014), in this study the surface-modified magnetic NiFe alloy nanoparticles can effectively adsorb heavy metal ions, so they were chosen to study the mechanisms of adsorption and to analyse the adsorption isotherms. Because the novel particles can keep their magnetic property longer, they have more potential to be used repeatedly. Moreover, we are expecting to develop a better adsorbent by the combination of the magnetic particles and ordinary adsorbing materials.

The NiFe nanoparticles were synthesized using the hydrothermal method (Liu et al. 2015). Typically, 3:1 molar ratio of NiCl₂·6H₂O to FeSO₄·7H₂O and sodium citrate was dissolved into a specific volume of deionized water. 32 mL of 1 M NaOH was added to the solution to adjust the pH to around 12.5. An excess of hydrazine hydrate was dropped into the solution as a reducing agent. The mixture solution was transferred to an autoclave after being stirred well under N₂. The autoclave was sealed and put into a furnace, which was preheated to 120 °C. The autoclave was cooled to room temperature after heating for 19 h. NiFe nanoparticles were separated with a magnetic bar from the solution. The products obtained were cleaned and re-dispersed in deionized water and ethanol several times.

UV-vis spectrophotometry was used to determine the concentrations of nickel ions under the condition of diacetyldioxime as a chromogenic agent.

Another kind of NiFe particle was obtained by the above method by only changing the dosage of sodium citrate. In the paper, the NiFe particle was prepared under the condition of a molar ratio of the total of Ni and Fe salts to sodium citrate of 1:10, the NiFe particle reported before was of a molar ratio of 1:0.2 (Liu et al. 2014).

X-ray diffraction (XRD) patterns were recorded by a Rigaku D/MAX 2500 diffractometer. Scanning electron microscopy (SEM) was performed on a Zeiss Ultra Plus model. Magnetization measurement (VSM) was undertaken using a Lakeshore Model 7407 Vibrating Sample Magnetometer at room temperature. Thermogravimetric analysis (TG) was performed on a TA Instruments SDT Q600. The Fourier transform infrared spectroscopy (FTIR) measurements were performed on a Perkin–Elmer Spectrum 100.

For the equilibrium experiment, 100 mg Ni_2.33Fe nanoparticles were added into 30 mL Ni²⁺ solution at a series of known initial concentrations (C₀), where the pH value of the Ni²⁺ solution was controlled at 5.0–5.5. After stirring the samples for 24 h with an agitator, the magnetic particles were separated from the solution using a magnet.

For kinetic experiments, zero time was taken as when 200 mg of Ni_2.33Fe nanoparticles was added into 200 mL of Ni²⁺ solution with different initial concentrations. The solution was stirred and sampled at appropriate time intervals.

Compared with the particles we synthesized previously (Liu et al. 2014), the dosage of sodium citrate in the synthesis as a capping agent showed great influence on the composition, morphology and heat stability of NiFe alloy particles.

The compositions of Ni and Fe in the as-synthesized NiFe nanoparticles were determined using the inductively coupled plasma optical emission spectrometry (ICP-OES) method. The ratio of Ni to Fe is 2.33:1, so the particles can be expressed as Ni_2.33Fe. The other particles we reported before (Liu et al. 2014) were Ni_3.54Fe. This demonstrated that the dosage of sodium citrate had significant impact on the compositions of the bimetallic particles. The basic characterizations for the alloys, including XRD, SEM, transmission electron microscopy (TEM) and magnetic property, are in the Supporting Information (available with the online version of this paper).

To prove sodium citrate modified to the surface of NiFe alloy particles, TG and IR were tested.

The TG and DTA data for Ni_2.33Fe and Ni_3.54Fe are shown in Figure 1. For Ni_3.54Fe, its baseline is flat below 400 °C, explaining that there is no weight loss in the temperature range. The further information deduced is that the surface of the Ni_3.54Fe particles does not adsorb volatile molecules, such as water or ethanol. A sharp weight increase started at 400 °C, originating from the oxidation of Fe and Ni at 450 °C and 656 °C (inset curve b) (Liu et al. 2014).

An obvious difference for Ni_2.33Fe particles is that a weight loss of about 4.7% is observed below 400 °C. The weight loss comes from the volatilization of the water or ethanol bound to the Ni_2.33Fe particles. The same as Ni_3.54Fe particles, a sharp weight increase started at 400 °C, but the weight increase comes from the oxidation of Fe and Ni at 448 °C and 582 °C. Another difference appeared at 718 °C, with Ni_2.33Fe particles exhibiting weight loss while Ni_3.54Fe particles showed a weight increase. The weight loss of Ni_2.33Fe particles is attributed to the decomposition of citrate bound to the surface of the particles.

Further evidence about the existence of sodium citrate bound to the surface of Ni_2.33Fe particles was based on the characteristics of IR spectra (Figure 2). Comparison of the FTIR spectra of nano Ni_3.54Fe, nano Ni_2.33Fe and sodium citrate, it can be obviously seen from Figure 2 that there was a strong stretching vibration peak of C = O at 1,623 cm⁻¹ in curve b; the same peak can be found in curve c, but the peak did not appear in curve a. The clear evidence shows that the surface of Ni_2.33Fe was bound by sodium citrate, but Ni_3.54Fe was not.

The magnetic properties were characterized using a vibrating sample magnetometer at room temperature. The magnetization hysteresis curves of two samples show S-shapes, characteristic of magnetically-soft materials. For Ni_2.33Fe, the values of the remanence (Mr), saturation magnetization (Ms) and coercivity (Hc) are 2.91 emu·g⁻¹, 75.44 emu·g⁻¹ and 45.9 Oe; they are 1.0 emu·g⁻¹, 56.3 emu·g⁻¹ and 14 Oe for Ni_3.54Fe, respectively.

If the second order kinetic equation fits the data well, q_e and k′ can be determined through the slop and intercept of the plot because of a linear relationship of t/q_e against t of Equation (3).

Adsorption dynamics of Ni²⁺ by nanomagnetic particles of Ni_2.33Fe were investigated at different initial concentrations of Ni²⁺ and the results are shown in Figure 6. It is clear that the q_t has a rapid increase for contact time within 30 min, and sorption equilibration was achieved by 270 min, followed by a constant adsorption rate with further shaking time.

The values of log(q_e – q_t) were calculated from the kinetic data of Figure 3 and plotted versus t of Equation (2). The adsorption parameters and the plots from pseudo-first-order kinetic model are presented in Table 2.

It can be found the corresponding regression coefficient (R²) based on the pseudo-first-order model is low (R² < 0.85), and the theoretical q_e values predicted by this model are very different to the experimental q_e values at different initial concentrations of Ni²⁺.

Figure 4 shows the plots of the pseudo-second-order kinetic model for adsorption on magnetic nano-Ni_2.33Fe, where t/q_e is plotted against t. And adsorption kinetic parameters based on the pseudo-second-order equation (Equation (3)) are presented in Table 1. The theoretical q_e values predicted by this model agree very closely to the experimental q_e values at different initial adsorbate concentrations. Further, a good linearity with R² above 0.999 also indicates the adsorption kinetics follows the pseudo-second-order model.

The results obtained using the pseudo-second-order model are not enough to predict the diffusion mechanism (Akar et al. 2008). Therefore, the intraparticle diffusion model is investigated to analyze the adsorption kinetic data.

The diffusion model plots shown in Figure 5 are related by two straight lines during the adsorption process, indicating the adsorption was regulated by both the surface and intraparticle diffusion processes. The first linear portion of the plot indicates the boundary layer effect, i.e. surface adsorption while the second linear portion is due to the intraparticle diffusion. Moreover, no line passes through the original point. All these facts prove that intraparticle diffusion is not the only controlling factor. The slope of the second linear portion of the plot has been defined as a rate parameter (k_d), which characterizes the rate of adsorption in the region where pore diffusion is rate limiting. Thus, all these suggest that in the adsorption of Ni²⁺ over the nano Ni_2.33Fe was controlled by external mass transfer followed by intraparticle diffusion mass transfer.

The values of K_id and C_i are listed in Table 2. It can be found that K_i1 is larger than K_i2, while C₁ is smaller than C₂, demonstrating that the global adsorption process is controlled by intraparticle diffusion.

The values of q_m and K_L were determined from the slope and intercept of the linear plots of C_e/q_e versus C_e, and values of K_F and 1/n were determined from the slope and intercept of the linear plot of ln q_e versus lnC_e.

The equilibrium adsorption capacities of Ni²⁺ are fitted by Langmuir and Freundlich isotherm equations (Equations (5) and (6)), and the results of the data fitting are shown in Figure 6. In the Langmuir model, the regression coefficient, R² = 0.8025, is low, and the slope of the line is a negative value. It will provide an unreasonable negative value of q_m. This implies that the Langmuir model could not represent the data reasonably well. The Langmuir describes monolayer adsorption, and this means monolayer adsorption is not suitable for representing the adsorbing of Ni²⁺ over Ni_2.33Fe.

The regression coefficient, R², is 0.9923 based on fitting the data on the Freundlich model. It obviously fits better to the Freundlich equation than to the Langmuir equation. Thus, the adsorption of Ni²⁺ on Ni_2.33Fe obeyed the Freundlich adsorption isotherm. The values of K_F and 1/n were determined from the slope and intercept of the linear plot of ln q_e versus lnC_e, and are 0.0177 and 1.1905, respectively. The value of 1/n between 1 and 2 indicates that Ni²⁺ can be adsorbed by the magnetic Ni_2.33Fe. The Freundlich model, as an empirical model, is particularly suitable to describe the interaction of multi-layer ions and chemical adsorption.

Based on all that has been discussed above, the models of the surface-modified NiFe nanoparticles in the water and the adsorption for Ni²⁺ can be deduced to be as in Figure 7. Citrates are densely arranged on the surface of the basic unit of NiFe particles by combination with the O^2– in one of three carboxylate radicals due to the excessive dosage of sodium citrate during the preparation process. Hence, the existence of sodium citrate in the NiFe magnetic particles can be detected from IR and TG. Further, the wires seen in SEM are grown by the O²⁻ in the second carboxylate radical attached to another NiFe particle. For the adsorption, the chelation between Ni²⁺ and the carboxylate must be generated during the adsorbing process. There are two carboxylates in the sodium citrate to chelate Ni²⁺, illustrating that the adsorption is via the multisite adsorption Freundlich model, the same as the experimental results. Some analogous results (Zhong et al. 2013; Li et al. 2015) have been obtained on the effects of sodium citrate on the adsorption. Results from Zhong et al. showed that the carboxyl content gave a great enhancement in the adsorption properties of modified activated carbon for copper. Hence, it can be sure that sodium citrate plays a key role in the adsorbing ability of NiFe particles.

The NiFe bimetallic alloy particles were successfully modified by sodium citrate. The surface-modified magnetic NiFe alloy nanoparticles exhibited the ability to adsorb nickel ions. By comparing the pseudo-first and pseudo-second-order kinetic models, it was found that the pseudo-second-order kinetic model successfully described the kinetic data. The Freundlich and Langmuir adsorption models were adopted to describe the adsorption isotherms, and it was found that the experimental data were best modeled to the Freundlich model.

The research work was supported by the National Natural Science Foundation of China (NSFC, Grant No. 51574084).