An efficient method for removal of pentachlorophenol using adsorption and microwave regeneration with different magnetic carbon nanotubes

Polychlorinated organic compounds constitute a major group of environmental pollutants, with widespread application in the chemical, pharmaceutical, agricultural, and leather industries (Weavers et al. 2000). Polychlorinated organic compounds are highly toxic and persistent in the environment and highly resistant to environmental degradation. Therefore, it is highly desirable to develop an efficient treatment of polychlorinated organic compounds that involves no secondary pollution.

Some technologies, such as biodegradation processes (Zhu et al. 2012), chemical oxidation (Ye et al. 2010), adsorption (Peng et al. 2016), and electrochemical methods (Alfaya et al. 2015) have been proposed to treat polychlorinated organic compounds; however, each method has its own limitations and disadvantages. The coupling of technologies for treatment can overcome the weaknesses of one technology by acquiring the strong points of another technology to achieve the purpose of efficient treatment.

Adsorption is one of the most powerful techniques for the removal of organic compounds from wastewater because of its simple and easy operation and high efficiency (Mohammadi & Veisi 2018). However, adsorption is unable to degrade pollutants and so is only able to transfer the pollution. For treatment of pollutants without secondary contamination, adsorption therefore needs to be combined with another technology.

Microwave irradiation has been applied in the treatment of contaminated soil or wastewater in a process that is fast with no secondary pollution (Kahar et al. 2017; Qi et al. 2017). However, water can absorb microwaves and so most of the energy is absorbed by water, which means a lot of energy is wasted. It is therefore more energy efficient if the pollutant can be adsorbed onto microwave absorption materials for treatment.

In this study, a combination of adsorption and microwave technology is designed for the rapid treatment of chlorinated organic compounds. For this coupling of technologies, the key is the development of materials with high adsorption capacity, fast separation, and absorption of microwaves.

Carbon nanotubes (CNTs) are suitable as potential materials for the coupled technology treatment. Many studies have shown that CNTs have an excellent effect on the removal of organic pollutants from wastewater (Shao et al. 2010; Xiao et al. 2014; Bhanjana et al. 2017). To solve the problem of separating CNTs, there have been reports that the magnetic materials can load on the surface of CNTs, which facilitates efficient separation under the action of external magnetic fields (Peng et al. 2005; Gao et al. 2013; Wang et al. 2014).

The magnetic material ferrite is an excellent electromagnetic wave absorbing material (Kaiser 2012). The composite of ferrite magnetic materials and CNTs can therefore effectively improve the absorption microwave capabilities (Zhan et al. 2011; Wen et al. 2013). Among the various magnetic materials CoFe₂O₄ has good microwave absorption abilities (Li et al. 2015; Yan et al. 2015). If CoFe₂O₄ can be composited with other metals and loaded onto, CNT can increase its magnetism and microwave absorption abilities, which are more suitable for the adsorption–microwave combined treatment of organic matter in water. However, there has been little research in this area.

Based on the above research, in this study, Co_xM_y (M = Cu, Mn, Ni, Zn) Fe₂O₄ ferrite was selected and loaded onto the surface of CNTs for use in the adsorption–microwave irradiation coupled treatment for chlorinated organic compounds in water. The influencing factors and intrinsic relationship of the structure and composition of the adsorbed materials on the adsorption–microwave irradiation treatment are discussed.

The CNTs were synthesized by catalytic pyrolysis according to our previous report (Cui et al. 2008). They were grown at 900 °C for 2 h with a ferrocene–xylene mixture (Fe/C = 0.4%) as a catalyst and carbon source. The mixture gas was Ar and H₂ (9:1), which was flowed at a rate of 1,000 mL min⁻¹). The obtained CNTs were modified with a 4.0 M H₂SO₄–HNO₃ (1:1) mixture at 120 °C for 8 h under reflux. The modified CNTs were washed with deionized water and dried at 100 °C under vacuum and stored in a desiccator for use.

First, 0.5 g of CNTs were fully dispersed in solutions of Fe(NO₃)₃·9H₂O, Co(NO₃)₂·6H₂O, Ni(NO₃)₂·6H₂O CuSO₄·5H₂O, MnSO₄·H₂O, ZnSO₄·7H₂O. The molar ratio of Co:M (M = Ni, Cu, Mn. Zn): Fe was 1:1:5 and the magnetic material load was 20%. Second, an ethylene glycol and polyethylene glycol (the volume ratio was 1:50) mixture as a sediment agent was added. The mixture was ultrasonicated and placed in an autoclave and heated for 12 h at 180 °C. After cooling to room temperature, the precipitate was isolated using a permanent magnet and was washed with absolute ethanol, filtered, dried in a drying oven at 45 °C for 24 h, and bottled for use.

The MCNTs (1 g) were added to a flask that contained 100 mL of PCP solution (250 mg L⁻¹) and this was shaken in a thermostatic shaker (25 °C, 180 rpm) for 2 h to achieve adsorption equilibrium. The saturated MCNTs were treated with microwave set to 850 W with a frequency of 2,450 MHz. A schematic diagram of the microwave experimental apparatus is shown in our previous work (Cui et al. 2015).

The morphology of MCNTs was characterized by transmission electron microscopy (TEM, HT7700, Japan) and X-ray diffractometry (XRD, LabxXRD-6000, Shimadzu, Japan). Sample magnetic properties were determined using a vibrating sample magnetometer (Riken Denshi, BHV-525). The Brunauer–Emmett–Teller (BET) specific surface area was determined through nitrogen adsorption–desorption measurements (Autosorb-iQ-MP-VP, USA). The microwave absorption performance was analyzed using a microwave network analyzer (VNA, N5244A PNA-X, Agilent, USA). The composition of the magnetic materials was determined using inductively coupled plasma atomic emission spectroscopy (Advantage, IRIS, USA).

The concentration of PCP was analyzed using HPLC (Agilent, 1200 USA) equipped with a C₁₈ reversed phase column (250 mm × 4.6 mm, 5 μm) and an ultraviolet detector (UV-1575). The degree of mineralization of PCP was determined using a total organic carbon (Liqui TOC trace, Elementar, Germany) analyzer.

The morphological structures and composition of the different MCNTs were detected by TEM and energy-dispersive X-ray spectroscopy (EDX) (Figure 1). As shown in Figure 1(a), the CNTs diameter was 60–100 nm with a smooth surface, and the Co_0.5Cu_0.5Fe₂O₄ nanoparticles were spherical in shape with a diameter of 30–50 nm and were distributed uniformly on the CNTs. The EDX spectrum exhibited peaks attributed to Co, Cu, Fe, and C. Figure 1(b) shows the TEM morphology and distribution of Co_0.5Mn_0.5Fe₂O₄ nanoparticles on the CNTs. The Co_0.5Mn_0.5Fe₂O₄ nanoparticles were dispersed uniformly on the CNTs with a diameter of 50–70 nm. The compositions of the Co_0.5Mn_0.5Fe₂O₄-CNT_S were determined from the EDX spectrum showing Co, Mn, Fe, and C. As shown in Figure 1(c), the Co_0.5Ni_0.5Fe₂O₄ nanoparticles were spherical in shape with an average diameter of 70 nm. The compositions of the Co_0.5Ni_0.5Fe₂O₄-CNT_S were determined by EDX experiments and Co, Ni, Fe, and C elements were mainly observed. The morphologies of the Co_0.5Zn_0.5Fe₂O₄-CNTs are shown in Figure 1(d). Co_0.5Zn_0.5Fe₂O₄ nanoparticles were dispersed uniformly on the CNTs with a diameter of 40–70 nm. Co, Zn, Fe, and C elements were mainly observed.

The XRD spectra of the CNTs Co_0.5Cu_0.5Fe₂O₄-CNTs, Co_0.5Mn_0.5Fe₂O₄-CNTs, Co_0.5Ni_0.5Fe₂O₄-CNTs, and Co_0.5Zn_0.5Fe₂O₄-CNTs are shown in Figure 2. A peak was observed at 26.03° from the XRD patterns in Figure 3(a), which corresponded to the (002) crystal of the CNTs. Compared with the CNTs, no new diffraction peaks were observed after loading with the magnetic material. As shown in the figure, the pure magnetic nanomaterials were amorphous. This proved that the magnetic material on CNT was amorphous.

Figure 3 shows the magnetic response characterization of the different MCNTs. The MCNTs, as adsorbents, were well dispersed in water and maintained a stable state (Figure 3(a)). When an external magnetic field was applied, the different MCNTs were attracted toward the magnet in a very short time (Figure 3(b)), which demonstrated a high magnetic sensitivity. Figure 3(c) shows the hysteresis loops of four magnetic CNTs at room temperature. As revealed in the figure, the magnetization increased with an increasing magnetic field for the different MCNTs at 298 K. The saturation magnetization of Co_0.5Cu_0.5Fe₂O₄-CNTs, Co_0.5Mn_0.5Fe₂O₄-CNTs, Co_0.5Ni_0.5Fe₂O₄-CNTs, and Co_0.5Zn_0.5Fe₂O₄-CNTs were 0.42, 0.61, 0.61, and 0.40 emu g⁻¹, respectively. This indicated that the MCNTs could be collected successfully by applying a magnetic field after adsorption.

To test the adsorption rate of the different MCNTs, the adsorption of PCP from aqueous solutions was carried out using the batch technique. The different MCNTs (0.05 g) were added to bottles containing 50 mL of 50 mg L⁻¹ PCP solution and shaken in a constant temperature oscillator (25 °C, 180 rpm). Figure 4 shows the effect of contact time on the amount of PCP removed. The amount of PCP adsorbed by different MCNTs increased drastically in the range of 0–80 min, and finally reached equilibrium at 120 min, and the equilibrium times for the different MCNTs were not too different.

In this study, the pseudo-first-order and pseudo-second-order kinetic models were used to evaluate the mechanism of different MCNTs for the PCP.

In the equations, G_t is the adsorption capacity (mg g⁻¹) at time t; G_e is the saturated adsorption amount at equilibrium (mg g⁻¹); k₁ is the quasi-first-order adsorption rate constant (min⁻¹); and k₂ is the quasi-second-order adsorption rate constant (mg (g·min)⁻¹). The kinetic constants obtained by fitting are shown in Table 1. According to the correlation coefficients (R²) in Table 1, the pseudo-second-order model fitted the adsorption kinetics better than the pseudo-first-order for the investigated PCP. In addition, G_e calculated by the quasi-secondary kinetic equation was consistent with the experimental results. This indicates that physical and chemical interactions occurred between PCP and MCNTs. The adsorption capacity of the MCNTs for PCP was in the following order: Co_0.5Cu_0.5Fe₂O₄-CNTs > Co_0.5Mn_0.5Fe₂O₄-CNTs > Co_0.5Ni_0.5Fe₂O₄-CNTs > Co_0.5Nn_0.5Fe₂O₄-CNTs.

To investigate the adsorption isotherms, the different MCNTs (0.05 g) were added to bottles containing 50 mL of PCP solution with a certain concentration and shaken in a constant temperature oscillator (25 °C, 180 rpm).

According to the equilibrium concentration and the adsorption amount, the adsorption isotherms were fitted following the Langmuir and Freundlich adsorption isotherm models.

In Equations (5) and (6), G_e represents the equilibrium adsorption amount (mg g⁻¹); G_m represents the maximum adsorption amount (mg g⁻¹); C represents the equilibrium concentration (mg L⁻¹); b represents the adsorption equilibrium constant (L mg⁻¹); k_F is the adsorption capacity, and 1/n is the Freundlich constant, which indicates the rate of change with the amount of adsorption.

The obtained adsorption constants, according to the two adsorption isotherm models, are shown in Table 2.

From the correlation coefficient R² in Table 2, it can be seen that the Langmuir equation and the Freundlich equation described the adsorption of PCP on MCNTs. The results showed that the monolayer and multilayer adsorption were simultaneously occurring adsorption. From the fitting, the maximum adsorption amount (G_m) of Co_0.5Cu_0.5Fe₂O₄-CNTs, Co_0.5Mn_0.5Fe₂O₄-CNTs, Co_0.5Ni_0.5Fe₂O₄-CNTs, and Co_0.5Zn_0.5Fe₂O₄-CNTs was 76.02, 75.80, 75.83, and 74.50 mg g⁻¹, respectively. In addition, 1/n is related to the strength of the adsorption driving force. When 0.1 < 1/n< 0.5, adsorption was easy; when 0.5 < 1/n ≤ 1, the adsorption process was a little difficult; and when 1/n> 1, the adsorption process was very difficult (Yu et al. 2018). The values of 1/n for PCP on the different MCNTs were less than 0.5, which demonstrated that PCP could be quite easily adsorbed on the MCNTs.

To investigate the pH effect on adsorption, the different MCNTs were added to a PCP solution with pH values of 2–10 and shaken for 2 h to reach adsorption equilibrium. Figure 5 shows that the adsorption capacity of PCP was remarkably dependent on the solution pH and this trend was the same for the different MCNTs. When the solution pH was increased from 2 to 5, the adsorption capacity of PCP remained stable, but at a pH 5–12, the adsorption capacity was markedly reduced. We attributed this result to when pH ≤ 4.9, the MCNT's surface exhibited a positive charge and the PCP mostly existed in an un-ionized molecular state (C₆CI₅OH). The pH change, therefore, did not affect the adsorption capacity. When pH > 5, the MCNT's surface exhibited a negative charge and the ionized state content of C₆Cl₅O⁻ increased significantly with an increase of pH. When pH > 5, electrostatic repulsion existed between the MCNT's surface and the C₆Cl₅O⁻, and so the adsorption capacity was significantly reduced.

To study the effect of temperature on adsorption, the adsorption experiments were performed at different temperatures. The different MCNTs (0.03 g) were added to bottles containing 50 mL of PCP and shaken at temperatures of 20 °C, 30 °C, 40 °C, and 50 °C. Figure 6 shows the adsorption capacities of the Co_0.5Cu_0.5Fe₂O₄-CNTs, Co_0.5Mn_0.5Fe₂O₄-CNTs, Co_0.5Ni_0.5Fe₂O₄-CNTs, and Co_0.5Zn_0.5Fe₂O₄-CNTs were 43.2, 40.8, 35.1, and 33.9 mg g⁻¹ at 20 °C, respectively. When the temperature was increased from 20 °C to 50 °C, the adsorption capacities of the MCNTs for PCP were noticeably reduced. The adsorption capacities of the Co_0.5Cu_0.5Fe₂O₄-CNTs, Co_0.5Mn_0.5Fe₂O₄-CNTs, Co_0.5Ni_0.5Fe₂O₄-CNTs, and Co_0.5Zn_0.5Fe₂O₄-CNTs were reduced to 26.5, 25.9, 22.8, and 20.7 mg g⁻¹ at 50 °C, respectively.

The thermodynamic relative parameters are summarized in Table 3. The negative values of enthalpy changes (ΔH) indicated that the adsorption process was exothermic. The negative values of entropy changes (ΔS) indicated that the randomness at the interface between the adsorbent and the solution was reduced due to adsorption, and there was no significant change in the internal structure of the adsorbent after adsorption (Liu et al. 2015). All results clearly indicate that the reaction was mainly based on physical adsorption, and a high adsorption capacity can be achieved at room temperature with no additional conditions required, which saves energy for practical applications and allowing this process to be widely applied.

Figure 7 shows complex permittivity and complex magnetic permeability of the different MCNTs in the 2–18 GHz band.

The real part of the complex permittivity (e′) and the real part of the complex permeability (μ′) represent the ability to store electric energy and magnetic energy, and the imaginary part of the complex permittivity (e″) and the imaginary part of the complex permeability (μ″) represent the ability to lose electric energy and magnetic energy (Zhang et al. 2013). Figure 7 shows that the e′ and e″ of the four MCNTs in the range of 2–10 GHz gradually decreased with increasing frequency. At 10–18 GHz, the real and imaginary parts of Co_0.5Ni_0.5Fe₂O₄-CNTs and Co_0.5Zn_0.5Fe₂O₄-CNTs exhibited a stable trend. Co_0.5Cu_0.5Fe₂O₄-CNTs showed a resonance peak in the range of 10–12 GHz, and Co_0.5Mn_0.5Fe₂O₄-CNTs showed a resonance peak in the range of 12–15 GHz. This phenomenon was attributed to the hysteresis response of the dipole polarization under the high frequency variation of the electric field. In addition, the μ′ and μ″ of the four kinds of MCNTs fluctuated greatly, and the μ′ and μ″ of Co_0.5Mn_0.5Fe₂O₄-CNTs and Co_0.5Zn_0.5Fe₂O₄-CNTs reached a maximum near 12 GHz and reached a minimum near 14 GHz. The μ′ of Co_0.5Cu_0.5Fe₂O₄-CNTs showed a resonance peak near 12–15 GHz, and the μ″ value showed a resonance peak near 14–16 GHz. The above phenomena were due to the combination of the dispersion effect and the polarization effect.

If the e″ value is greater than the μ″ value, it indicates that the adsorbent is an electric loss-absorbing medium; if the μ″ value is greater than the e″ value, it indicates that the adsorbent is a magnetic loss-absorbing medium (Hou et al. 2013). Figure 7 shows that the e″ value was much larger than the μ″ value, which indicated that these four magnetic carbon nanotubes are electric loss-type microwave absorbing media (Zeng et al. 2012). In summary, the four MCNTs exhibit good wave absorbing properties at low frequencies. Therefore, adsorption saturation of PCP on different MCNTs is suitable for microwave regeneration. At 850 W, the temperature of MCNTs quickly reaches 1,100 °C within 100 s.

In this study, we selected microwave irradiation at 850 W for 180 s for the degradation of the adsorption-saturated PCP on different MCNTs. According to the analysis of the combined substances of the MCNT extraction and distillation, the degradation and mineralization efficiencies were over 90%. The relationship between the regeneration times and the regeneration efficiency is shown in Figure 8. The regeneration efficiency of MCNTs was basically stable with the increase of regeneration times, and had high regeneration efficiency. Under the same conditions, the regeneration efficiency of Co_0.5Cu_0.5 Fe₂O₄-CNTs, Co_0.5Mn_0.5Fe₂O₄-CNTs, Co_0.5Ni_0.5 Fe₂O₄-CNTs, and Co_0.5Zn_0.5Fe₂O₄-CNTs were still 98.1%, 95.8%, 93.1%, and 94.7%, respectively, after six cycles of microwave regeneration. In addition, after reuse six times, the loss of magnetic materials was less than 7%.

In this study, Co_0.5M_0.5Fe₂O₄ (M = Ni, Cu, Mn, Zn)-CNTs were successfully prepared using the solvothermal method with high specific surface area, good magnetic properties, and absorption microwave characteristics. The adsorption kinetics and isotherms of PCP on Co_0.5M_0.5Fe₂O₄-CNTs demonstrated that Co_0.5M_0.5Fe₂O₄ has the advantages of fast adsorption equilibrium and large adsorption capacity, and the adsorption capacity is sensitive to pH and temperature. The influencing factors of the acidity condition and low temperature are most favorable for adsorption. In addition, Co_0.5M_0.5Fe₂O₄-CNTs exhibits good microwave absorption properties at low frequencies. After magnetic recovery of the saturated Co_0.5M_0.5Fe₂O₄-CNTs, they can be rapidly regenerated by microwave regeneration with no secondary pollution. After six adsorption and microwave treatment cycles, the regeneration efficiency remains at 98% and no apparent damage is observed to the structure. Therefore, these Co_0.5M_0.5Fe₂O₄-CNTs are suitable for the combination of adsorption, magnetic recovery, and microwave degradation to treat organic pollutants in water.

This work was supported financially by the Nature Science Foundation (No. 51678323) and Shandong Provincial Science Foundation (No. ZR2017MEE013) of China. We thank Liwen Bianji, Edanz Group China (www.liwenbianji.cn/ac), for editing the English text of a draft of this manuscript.