Efficient Removal of Hazardous P-Nitroaniline from Wastewater by Using Surface-Activated and Modified Multiwalled Carbon Nanotubes with Mesostructure

P-nitroaniline (PNA) is an aniline compound with high toxicity and can cause serious harm to aquatic animals and plants. Multiwalled carbon nanotubes (MWCNTs) are a multifunctional carbon-based material that can be applied in energy storage and biochemistry applications and semiconductors as well as for various environmental purposes. In the present study, MWCNTs (CO2–MWCNTs and KOH–MWCNTs) were obtained through CO2 and KOH activation. ACID–MWCNTs were obtained through surface treatment with an H2SO4–HNO3 mixture. Herein, we report, for the first time, the various MWCNTs that were employed as nanoadsorbents to remove PNA from aqueous solution. The MWCNTs had nanowire-like features and different tube lengths. The nanotubular structures were not destroyed after being activated. The KOH–MWCNTs, CO2–MWCNTs, and ACID–MWCNTs had surface areas of 487, 484, and 80 m2/g, respectively, and pore volumes of 1.432, 1.321, and 0.871 cm3/g, respectively. The activated MWCNTs contained C–O functional groups, which facilitate PNA adsorption. To determine the maximum adsorption capacity of the MWCNTs, the influences of several adsorption factors—contact time, solution pH, stirring speed, and amount of adsorbent—on PNA adsorption were investigated. The KOH–MWCNTs had the highest adsorption capacity, followed by the CO2–MWCNTs, pristine MWCNTs, and ACID–MWCNTs. The KOH–MWCNTs exhibited rapid PNA adsorption (>85% within the first 5 min) and high adsorption capacity (171.3 mg/g). Adsorption isotherms and kinetics models were employed to investigate the adsorption mechanism. The results of reutilization experiments revealed that the MWCNTs retained high adsorption capacity after five cycles. The surface-activated and modified MWCNTs synthesized in this study can effectively remove hazardous pollutants from wastewater and may have additional uses.


Introduction
The discharge of wastewater from dye industries has caused water pollution and recently gained global attention.P-nitroaniline (PNA) is an important intermediate, which is used in the pesticide, antioxidant, fuel additive, and dye industries [1].In even small quantities, effluents containing toxic substances can have detrimental effects on aquatic environments [2].Several physicochemical techniques-including adsorption, photodegradation, bio-decomposition and electrochemical treatment-have been developed for treating PNA-containing wastewater [3].Of these techniques, adsorption is considered the most efficient approach due to its low cost and simplicity of operation, even for treating large quantities of wastewater.Various analytical techniques-including high-performance liquid chromatography, gas chromatography, gas chromatography-mass spectrometry, and liquid chromatography-mass spectrometry-have been employed to measure dye respectively.The purpose of the surface activation and modification was to improve the adsorption capacity of the nanotubes by creating more adsorption sites for the capture of organic molecule.This study is the first to compare the efficacy of these three types of MWCNTs in the removal of PNA from wastewater.The physicochemical characteristics of the synthesized MWCNTs-such as their crystalline phase, functional groups, elemental composition, thermal stability, graphitization, pore characteristics, and surface and internal morphologies-were extensively examined.Factors that influence the efficiency of PNA removal-contact time, stirring speed, pH, and adsorbent amount-were also examined to determine their effects on the adsorption capacity.The interaction between PNA and MWCNTs and the adsorption mechanism were investigated through isotherm and kinetics experiments.Reutilization tests were also conducted to evaluate the reusability of the synthesized MWCNTs.

MWCNT Synthesis
MWCNTs were synthesized using catalytic chemical vapor deposition [28,29].CH 4 was employed as a carbon source, and Fe 2 O 3 was employed as a catalyst.The reaction temperature was set to 800-1000 • C. The CNTs were grown in a horizontal reactor that consisted of a quartz tube with a diameter of 5 cm and a length of 100 cm.After the reaction, the synthesized CNTs were immersed in HCl solution to remove catalyst residues.The pristine nanotubes are referred to as COM-MWCNTs.

Modification of MWCNTs
ACID-MWCNTs, CO 2 -MWCNTs, and KOH-MWCNTs were synthesized as follows: (1) ACID-MWCNTs [30]: 1.0 g of MWCNTs was dispersed in 40 mL of H 2 SO 4 -HNO 3 solution (molar ratio = 3:1).The acidic solution was dropped slowly with 10 mL of anhydrous alcohol to increase the hydrophilicity of the carbon solid.The solution was subsequently heated to 70 • C and stirred continuously for 24 h, after which it was washed in water until it became neutral to remove any acidic residues.The solution then underwent drying and grinding processes until an acid-modified MWCNT solid was obtained.The obtained solid is referred to as the ACID-MWCNTs.(2) CO 2 -MWCNTs [31]: 1.0 g of MWCNTs was dried at 100 • C for 24 h in an air oven.
The dried solid was placed on a ceramic boat and then inserted into a quartz tubular reactor.The carbonaceous solid was subsequently heated at 850 • C for 4 h.CO 2 was employed as the reaction gas.The flow rate was 60 mL/min.The obtained solid is referred to as the CO 2 -MWCNTs.(3) KOH-MWCNTs [32,33]: 4.0 g of KOH was dissolved in 50 mL of distilled water.
MWCNTs (1.0 g) were then added to the solution.The solution was subsequently stirred continuously at 120 • C, left to dry for 24 h, and then heated to 800 • C (at a rate of 10 • C/min) for 1 h under pure N 2 atmosphere.The obtained solid was rinsed with HCl solution (1.0 M).Afterward, the solid was washed with deionized water and dried.The solid is referred to as the KOH-MWCNTs.

Adsorption Experiment
PNA adsorption on the nanoadsorbents (COM-MWCNTs, ACID-MWCNTs, CO 2 -MWCNTs, and KOH-MWCNTs) was investigated using a bath apparatus.First, 10 mg of MWCNTs was added to 100 mL of PNA solution (PNA concentration = 150 mg/L).The solution was stirred for 2 h to ensure adsorption equilibrium.The effects of other factors-contact time, solution pH, stirring speed, and CNT amount-on the nanotubes' adsorption capacity were investigated.The PNA concentration was measured using a UV-Vis spectrophotometer (Genesys, Thermo Electron Corporation, Waltham, MA, USA).Adsorption capacity q t was calculated as follows: where V is the volume of the PNA solution (L), W is the CNT amount (g), and C 0 and C are the PNA concentrations at baseline and at time t, respectively.The reusability of the synthesized MWCNTs was assessed.First, the adsorption of PNA on the MWCNTs was conducted under optimal conditions.Once the adsorption equilibrium had been reached, desorption experiments were conducted by subjecting the MWCNTs to thermal treatment at 400 • C in N 2 atmosphere to achieve the maximum desorption efficiency.Adsorption-desorption experiments were performed at least five times, and PNA concentrations were measured using a UV-Vis spectrophotometer.

Characterization of CNT Materials
The pore volume, surface area, and pore diameter of the MWCNTs were evaluated using an ASAP 2020 adsorption analyzer (Micromeritics, Norcross, GA, USA) at −196 • C.An X-ray diffractometer (X'pert Pro System, PANalytical, Malvern, UK) with Cu Kα radiation was employed to observe the crystalline phase of the MWCNTs.A field-emission scanning electron microscope (JEOL JSM-6700F, Akishima, Tokyo, Japan) and a transmission electron microscope (JEOL JEM-1200CX II, Akishima, Tokyo, Japan) were used to examine their morphological features.A Fourier transmission infrared spectroscope (FTIR-8300, Shimadzu, Nakagyo-ku, Kyoto, Japan) was employed to examine the functional groups of the MWCNTs.Graphitization was assessed using a confocal Raman spectroscope (Renishaw, Gloucestershire, UK) with 632 nm He-Ne laser excitation.The stability of the MWCNTs before and after modification was examined using a thermogravimetric analyzer (Mettler Toledo, OH, USA, model TGA/SDTA851e).The surface elements on the MWCNTs were analyzed using an X-ray photoelectron spectroscope (Esca Lab 250Xi, Thermo Scientific, Waltham, MA, USA).The C1s peak at 284.60 eV was employed to calibrate the binding energy.The amounts of metallic impurities in the MWCNTs were determined using an inductively coupled plasma-mass spectrometer (ICP-MS) (Konton Plasmakon, Eching, Germany, model S-35).

Characterization of Nanotubes
Figure 1 depicts the X-ray diffraction patterns of the MWCNTs.The COM-MWCNT sample was not treated with chemical reagents.The peak at 2θ = 25.9 • corresponds to diffraction from the (002) plane, characteristic of hexagonal graphite structures [34].The signals detected at 2θ = 43  , and 53.5 • correspond to the (100), (101), and (004) planes, characteristic of graphite structures [35].For the ACID-MWCNT, CO 2 -MWCNT, and KOH-MWCNT samples, peaks corresponding to the (002), (100), (101), and (004) planes were again observed.However, the peaks in the spectra of the ACID-MWCNT, CO 2 -MWCNT, and KOH-MWCNT samples were less intense than those in the spectrum of the COM-MWCNT sample.Similar characteristics were also observed by Zhang and Chen [36].The hexagonal frame of the CNTs' graphite structure was evidently not destroyed by surface activation and modification.Table 1 shows the purity of the MWCNTs before and after modification [37].These nanotubes have the low concentration of metallic impurities.It can be ascribed to metals dissolving in the acidic solution after the HCl treatment.These metals were easily removed by leaching and filtration processes [12].Moreover, CO 2 -or KOH-activated samples had lower metal elements than pristine nanotubes.This decrease in impurities may be because metals were carried by the vaporization of carbonaceous matter during activation.
hydroxyl groups (OH) [38].The band at 2920 cm −1 corresponds to symmetric and asymmetric stretching of C-H groups.The peak at 1040 cm −1 corresponds to stretching of -COOH.In the case of the CO2-MWCNTs and KOH-MWCNTs, the peaks at 1375, 1625, and 1720 cm −1 represent C-O, C=C, and C=O stretching, respectively [10].These Ocontaining functional groups can effectively combine organic molecules through electrostatic attraction, van der Waals forces, and π-π interactions [39].These groups are thus favorable for the removal of contaminants.Functional groups were examined using Fourier transmission infrared spectroscopy (Figure 2).For all samples, the peak at 3400-3500 cm −1 corresponds to stretching of hydroxyl groups (OH) [38].The band at 2920 cm −1 corresponds to symmetric and asymmetric stretching of C-H groups.The peak at 1040 cm −1 corresponds to stretching of -COOH.In the case of the CO 2 -MWCNTs and KOH-MWCNTs, the peaks at 1375, 1625, and 1720 cm −1 represent C-O, C=C, and C=O stretching, respectively [10].These O-containing functional groups can effectively combine organic molecules through electrostatic attraction, van der Waals forces, and π-π interactions [39].These groups are thus favorable for the removal of contaminants.after modification [37].These nanotubes have the low concentration of metallic impurities.
It can be ascribed to metals dissolving in the acidic solution after the HCl treatment.These metals were easily removed by leaching and filtration processes [12].Moreover, CO2-or KOH-activated samples had lower metal elements than pristine nanotubes.This decrease in impurities may be because metals were carried by the vaporization of carbonaceous matter during activation.Functional groups were examined using Fourier transmission infrared spectroscopy (Figure 2).For all samples, the peak at 3400-3500 cm −1 corresponds to stretching of hydroxyl groups (OH) [38].The band at 2920 cm −1 corresponds to symmetric and asymmetric stretching of C-H groups.The peak at 1040 cm −1 corresponds to stretching of -COOH.In the case of the CO2-MWCNTs and KOH-MWCNTs, the peaks at 1375, 1625, and 1720 cm −1 represent C-O, C=C, and C=O stretching, respectively [10].These Ocontaining functional groups can effectively combine organic molecules through electrostatic attraction, van der Waals forces, and π-π interactions [39].These groups are thus favorable for the removal of contaminants.Figure 3 displays the Raman spectra of the unmodified and modified MWCNTs.The first peak at 1300-1340 cm −1 was attributable to disorder and defects (D band), and the second peak at 1570-1620 cm −1 was attributable to G band features [40].The ratio of D band intensity to G band intensity (I D /I G ) indicates the ratio of sp 3 to sp 2 carbon atom [41].The intensity ratio was highest for the CO 2 -MWCNTs (0.9859), followed by the KOH-MWCNTs (0.9606), COM-MWCNTs (0.9445), and ACID-MWCNTs (0.9348).More sp 3 carbon atoms were observed on the activated samples (CO 2 -MWCNTs and KOH-MWCNTs) than on the nonactivated samples.The activated MWCNTs displayed a high degree of oxidation, which suggested higher adsorption capacity [42].Figure 3 displays the Raman spectra of the unmodified and modified MWCNTs.The first peak at 1300-1340 cm −1 was attributable to disorder and defects (D band), and the second peak at 1570-1620 cm −1 was attributable to G band features [40].The ratio of D band intensity to G band intensity (ID/IG) indicates the ratio of sp 3 to sp 2 carbon atom [41].The intensity ratio was highest for the CO2-MWCNTs (0.9859), followed by the KOH-MWCNTs (0.9606), COM-MWCNTs (0.9445), and ACID-MWCNTs (0.9348).More sp 3 carbon atoms were observed on the activated samples (CO2-MWCNTs and KOH-MWCNTs) than on the nonactivated samples.The activated MWCNTs displayed a high degree of oxidation, which suggested higher adsorption capacity [42].The elements distributed on the surface of the unmodified and modified MWCNTs were inspected using X-ray photoelectron spectroscopy (Figure 4).Wide survey spectra confirmed that C and O were the main elements (Figure 4a) [43].In the C1s spectra (Figure 4b), the peak at approximately 284.0 eV was attributable to C-C and C=C groups, which came from the diamond-like sp 3 carbon and graphite-like sp 2 carbon, respectively.According to Zhao et al. [41], peaks at 287.5 and 291.0 eV are attributable to C=O and O-C=O bonds, respectively.However, the two peaks in our spectra were very weak.The O1s spectra (Figure 4c) contained peaks at 532.3 and 533.7 eV.These were associated with C=O and O-C=O bonds, respectively [44].The C-O functional groups in the MWCNTs would aid the π-π interaction between PNA molecules and the MWCNTs, thereby increasing the MWCNTs' adsorption activity.

CNTs
The thermal stability of the MWCNTs before and after modification was examined using thermogravimetric analysis (Figure 5).For all MWCNTs, weight loss could be divided into two zones (Figure 5a).The initial weight loss at 25-150 °C was due to moisture loss.The greatest weight loss at 150-800 °C may have been attributable to decomposition of the carbonaceous matters.The presence of ash residue (3.4-7.5 wt%) indicated that the MWCNTs did not completely decompose during oxidation.The CO2-MWCNTs had the highest purity.Peaks indicating the maximum rate of thermal decomposition are visible The elements distributed on the surface of the unmodified and modified MWCNTs were inspected using X-ray photoelectron spectroscopy (Figure 4).Wide survey spectra confirmed that C and O were the main elements (Figure 4a) [43].In the C1s spectra (Figure 4b), the peak at approximately 284.0 eV was attributable to C-C and C=C groups, which came from the diamond-like sp 3 carbon and graphite-like sp 2 carbon, respectively.According to Zhao et al. [41], peaks at 287.5 and 291.0 eV are attributable to C=O and O-C=O bonds, respectively.However, the two peaks in our spectra were very weak.The O1s spectra (Figure 4c) contained peaks at 532.3 and 533.7 eV.These were associated with C=O and O-C=O bonds, respectively [44].The C-O functional groups in the MWC-NTs would aid the π-π interaction between PNA molecules and the MWCNTs, thereby increasing the MWCNTs' adsorption activity.
The thermal stability of the MWCNTs before and after modification was examined using thermogravimetric analysis (Figure 5).For all MWCNTs, weight loss could be divided into two zones (Figure 5a).The initial weight loss at 25-150 • C was due to moisture loss.The greatest weight loss at 150-800 • C may have been attributable to decomposition of the carbonaceous matters.The presence of ash residue (3.4-7.5 wt%) indicated that the MWCNTs did not completely decompose during oxidation.The CO 2 -MWCNTs had the highest purity.Peaks indicating the maximum rate of thermal decomposition are visible in Figure 5b.The peak temperatures for the KOH-MWCNTs, CO 2 -MWCNTs, COM-MWCNTs, and ACID-MWCNTs were 467, 621, 692, and 705 • C, respectively.The ACID-MWCNTs had the highest thermal stability.The KOH-MWCNTs and CO 2 -MWCNTs had lower thermal decomposition temperatures than did the COM-MWCNTs and ACID-MWCNTs, possibly because the KOH-MWCNTs and CO 2 -MWCNTs were more porous.

Surface Area and Pore Features
The pore structure and surface area of the CNTs before and after modi examined using an N2 sorption experiment.As shown in Figure 6a,c, simila obtained for all types of MWCNTs, indicating a type-IV isotherm (Internatio Pure and Applied Chemistry classification) [45].The four carbons containin teresis loop were typical mesoporous materials.The mesostructure came channels and the entanglement of nanotubes.The N-adsorbed volumes MWCNTs and KOH-MWCNTs were much higher than those of the COM-M ACID-MWCNT.The high adsorption volumes were mainly a result of the of MWCNT tissues during activation [46].The nanotubes' pore size distribut sented in Figure 6b,d.The pore size distribution of the COM-MWCNTs wa ering sizes between 2 and 120 nm.Intertube spaces formed through aggreg

Surface Area and Pore Features
The pore structure and surface area of the CNTs before and after modification were examined using an N2 sorption experiment.As shown in Figure 6a,c, similar loops were obtained for all types of MWCNTs, indicating a type-IV isotherm (International Union of Pure and Applied Chemistry classification) [45].The four carbons containing an H3 hysteresis loop were typical mesoporous materials.The mesostructure came from tubular channels and the entanglement of nanotubes.The N-adsorbed volumes of the CO2-MWCNTs and KOH-MWCNTs were much higher than those of the COM-MWCNTs and ACID-MWCNT.The high adsorption volumes were mainly a result of the vaporization of MWCNT tissues during activation [46].The nanotubes' pore size distributions are presented in Figure 6b,d.The pore size distribution of the COM-MWCNTs was broad, covering sizes between 2 and 120 nm.Intertube spaces formed through aggregation of long tubes may have led to this broad pore size distribution [36].By contrast, the pore size

Surface Area and Pore Features
The pore structure and surface area of the CNTs before and after modification were examined using an N 2 sorption experiment.As shown in Figure 6a,c, similar loops were obtained for all types of MWCNTs, indicating a type-IV isotherm (International Union of Pure and Applied Chemistry classification) [45].The four carbons containing an H3 hysteresis loop were typical mesoporous materials.The mesostructure came from tubular channels and the entanglement of nanotubes.The N-adsorbed volumes of the CO 2 -MWCNTs and KOH-MWCNTs were much higher than those of the COM-MWCNTs and ACID-MWCNT.The high adsorption volumes were mainly a result of the vaporization of MWCNT tissues during activation [46].The nanotubes' pore size distributions are presented in Figure 6b,d.The pore size distribution of the COM-MWCNTs was broad, covering sizes between 2 and 120 nm.Intertube spaces formed through aggregation of long tubes may have led to this broad pore size distribution [36].By contrast, the pore size distributions of the ACID-MWCNTs, CO 2 -MWCNTs, and KOH-MWCNTs were narrow, indicating a uniform pore structure.The average pore diameters of the ACID-MWCNTs, CO 2 -MWCNTs, and KOH-MWCNTs were 36.29,14.87, and 10.70 nm, respectively.Notably, the KOH-MWCNTs had smaller pores than did the CO 2 -MWCNTs.Chemical activation may have more drastic effects than does physical activation.The surface area, pore volume, and pore fraction values are listed in Table 2.The pristine nanotubes (COM-MWCNTs) had a surface area of 189 m 2 /g and a pore volume of 1.039 cm 3 /g.The surface area and pore volume of the ACID-MWCNTs, which were treated with H 2 SO 4 -HNO 3 , were 80 m 2 /g and 0.871 cm 3 /g, respectively.The surface areas and pore volumes of the CO 2 -MWCNTs and KOH-MWCNTs, which were physically and chemically activated, were 484 m 2 /g and 1.321 cm 3 /g, respectively, and 487 m 2 /g and 1.432 cm 3 /g, respectively.An increase in mesoporous volume may cause an increase in the textural parameters of activated samples.The KOH-MWCNTs had the highest surface area and pore volume, consisting with the observation in Figure 6a,c.The mesopore fraction of the ACID-MWCNTs (91.50%) was lower than that of the COM-MWCNTs (98.84%), indicating that strong acid etching on the carbon surface resulted in the formation of a microporous structure.The mesopore fractions of the CO 2 -MWCNTs and KOH-MWCNTs were 99.32% and 97.63%, respectively.The physically activated MWCNTs had the highest mesoporosity.
distributions of the ACID-MWCNTs, CO2-MWCNTs, and KOH-MWCNTs were n indicating a uniform pore structure.The average pore diameters of the ACID-MW CO2-MWCNTs, and KOH-MWCNTs were 36.29,14.87, and 10.70 nm, respective tably, the KOH-MWCNTs had smaller pores than did the CO2-MWCNTs.Chemic vation may have more drastic effects than does physical activation.The surface are volume, and pore fraction values are listed in Table 2.The pristine nanotubes ( MWCNTs) had a surface area of 189 m 2 /g and a pore volume of 1.039 cm 3 /g.The area and pore volume of the ACID-MWCNTs, which were treated with H2SO4were 80 m 2 /g and 0.871 cm 3 /g, respectively.The surface areas and pore volumes CO2-MWCNTs and KOH-MWCNTs, which were physically and chemically act were 484 m 2 /g and 1.321 cm 3 /g, respectively, and 487 m 2 /g and 1.432 cm 3 /g, respe An increase in mesoporous volume may cause an increase in the textural parame activated samples.The KOH-MWCNTs had the highest surface area and pore v consisting with the observation in Figure 6a,c.The mesopore fraction of the MWCNTs (91.50%) was lower than that of the COM-MWCNTs (98.84%), indicati strong acid etching on the carbon surface resulted in the formation of a microporou ture.The mesopore fractions of the CO2-MWCNTs and KOH-MWCNTs were 99.32 97.63%, respectively.The physically activated MWCNTs had the highest mesopor

Surface Morphology
The morphologies of the MWCNTs before and after modification were examined using scanning electron microscopy (Figure 7).Compared with the diameter of the pristine MWC-NTs (Figure 7a), the MWCNTs modified using H 2 SO 4 -HNO 3 was coarser (Figure 7b), indicating that the MWCNTs were inflated by the acidic treatment.After activation (Figure 7c,d), the MWCNTs were shorter, indicating that the MWCNTs were sectioned into smaller MWCNTs as the carbon skeleton was etched with CO 2 or KOH [36].The drastic gasification during physical and chemical activation resulted in pore opening on the inner surface of MWCNTs, increasing the microporous and mesoporous volumes and thereby increasing the nanotubes' adsorption capacity.

Surface Morphology
The morphologies of the MWCNTs before and after modification were examined using scanning electron microscopy (Figure 7).Compared with the diameter of the pristine MWCNTs (Figure 7a), the MWCNTs modified using H2SO4-HNO3 was coarser (Figure 7b), indicating that the MWCNTs were inflated by the acidic treatment.After activation (Figure 7c,d), the MWCNTs were shorter, indicating that the MWCNTs were sectioned into smaller MWCNTs as the carbon skeleton was etched with CO2 or KOH [36].The drastic gasification during physical and chemical activation resulted in pore opening on the inner surface of MWCNTs, increasing the microporous and mesoporous volumes and thereby increasing the nanotubes' adsorption capacity.The internal morphologies of the MWCNTs before and after modification were examined using transmission electron microscopy (Figure 8).All of the MWCNTs were hollow and tubular.The acid and alkali treatments did not affect the internal morphologies of the MWCNTs [47].The COM-MWCNT had a glossy and smooth surface (Figure 8a).The acidified nanotubes had a relatively coarse surface (Figure 8b).The ACID-MWCNTs were tied in knots and twisted, resulting in less surface area being available than for the pristine nanotubes.The observation consisted with the results of surface area analysis in Table 2.The activated nanotubes were shorter than the nonactivated nanotubes (Figure 8c,d) The surface of the activated nanotubes was coarser than that of the pristine nanotubes due to the drastic oxidation reaction during activation.These conditions are helpful for increasing the surface area of the nanotubes (Table 2).The internal morphologies of the MWCNTs before and after modification were examined using transmission electron microscopy (Figure 8).All of the MWCNTs were hollow and tubular.The acid and alkali treatments did not affect the internal morphologies of the MWCNTs [47].The COM-MWCNT had a glossy and smooth surface (Figure 8a).The acidified nanotubes had a relatively coarse surface (Figure 8b).The ACID-MWCNTs were tied in knots and twisted, resulting in less surface area being available than for the pristine nanotubes.The observation consisted with the results of surface area analysis in Table 2.The activated nanotubes were shorter than the nonactivated nanotubes (Figure 8c,d) The surface of the activated nanotubes was coarser than that of the pristine nanotubes due to the drastic oxidation reaction during activation.These conditions are helpful for increasing the surface area of the nanotubes (Table 2).

Adsorption Performance of CNTs
The PNA adsorption capabilities of the COM-MWCNTs, ACID-MWCNTs, CO2 MWCNTs, and KOH-MWCNTs were evaluated.The ACID-MWCNTs had lower adsorp tion capacity than did the COM-MWCNTs, indicating that acid treatment did not im prove the nanotubes' adsorption capacity (Figure 9a).Corrosion due to acid on the nano tube surface may cause the collapse of pore walls.The reduction in surface area and por volume for the ACID-MWCNT sample (Table 2) resulted in lower adsorption capability The CO2-MWCNTs and KOH-MWCNTs had high adsorption capabilities compared wit the COM-MWCNTs because their surface area and pore volume were higher than thos of the pristine nanotubes.In addition, the presence of oxygen-containing functiona groups in the activated nanotubes (Figure 2) promoted the interactions between PNA an the nanotubes, leading to an enhancement of PNA adsorption [48].The KOH-MWCNT had higher adsorption capacity than did the CO2-MWCNTs.This was because the KOH MWCNTs had a higher pore volume (Table 2).Despite the higher adsorption capacity o the KOH-MWCNTs, the CO2-MWCNTs exhibited adsorption ability significantly highe than that of the KOH-MWCNTs in the initial 2 min, possibly because the pores of th KOH-MWCNTs were narrower than those of the CO2-MWCNTs (Table 2), which cause long-distance diffusion of PNA into the nanotube pores.The relatively large diameter o pores in the CO2-MWCNTs meant that the pores did not become blocked, which wa beneficial to PNA adsorption.The maximum adsorption capacities of the ACID MWCNTs, COM-MWCNTs, CO2-MWCNTs, and KOH-MWCNTs were 26.8, 49.6, 123.4 and 171.3 mg/g, respectively.The results indicate that activated nanotubes are excellen adsorbent materials for recovering organic contaminants, such as PNA, from wastewate

Adsorption Performance of CNTs
The PNA adsorption capabilities of the COM-MWCNTs, ACID-MWCNTs, CO 2 -MWCNTs, and KOH-MWCNTs were evaluated.The ACID-MWCNTs had lower adsorption capacity than did the COM-MWCNTs, indicating that acid treatment did not improve the nanotubes' adsorption capacity (Figure 9a).Corrosion due to acid on the nanotube surface may cause the collapse of pore walls.The reduction in surface area and pore volume for the ACID-MWCNT sample (Table 2) resulted in lower adsorption capability.The CO 2 -MWCNTs and KOH-MWCNTs had high adsorption capabilities compared with the COM-MWCNTs because their surface area and pore volume were higher than those of the pristine nanotubes.In addition, the presence of oxygen-containing functional groups in the activated nanotubes (Figure 2) promoted the interactions between PNA and the nanotubes, leading to an enhancement of PNA adsorption [48].The KOH-MWCNTs had higher adsorption capacity than did the CO 2 -MWCNTs.This was because the KOH-MWCNTs had a higher pore volume (Table 2).Despite the higher adsorption capacity of the KOH-MWCNTs, the CO 2 -MWCNTs exhibited adsorption ability significantly higher than that of the KOH-MWCNTs in the initial 2 min, possibly because the pores of the KOH-MWCNTs were narrower than those of the CO 2 -MWCNTs (Table 2), which caused long-distance diffusion of PNA into the nanotube pores.The relatively large diameter of pores in the CO 2 -MWCNTs meant that the pores did not become blocked, which was beneficial to PNA adsorption.The maximum adsorption capacities of the ACID-MWCNTs, COM-MWCNTs, CO 2 -MWCNTs, and KOH-MWCNTs were 26.8, 49.6, 123.4,and 171.3 mg/g, respectively.The results indicate that activated nanotubes are excellent adsorbent materials for recovering organic contaminants, such as PNA, from wastewater.The effects of pH on PNA adsorption were investigated (Figure 9b).For the COM-MWCNTs and CO2-MWCNTs, the qt values were almost constant at pH = 2-11, indicating that their adsorption ability was not affected by the pH level.For the KOH-MWCNTs, qt was lowest at pH = 6.The low adsorption efficiency was because the surface of the KOH-MWCNTs contained OH − ions.The electrostatic attraction between the adsorbent and adsorbate may have been smaller when OH − ions were neutralized with weak H + solution.For the ACID-MWCNT sample, qt was constant at pH = 2-9 and then decreased as the pH was increased to 11.The surface of the ACID-MWCNTs contained H + ions after acid treatment.When the medium was strongly basic, the nanotubes lost H + ions and did not easily interact with PNA molecules through electrostatic force [49].The effect of the stirring speed on PNA adsorption was investigated.The samples were stirred at speeds of 50-150 The effects of pH on PNA adsorption were investigated (Figure 9b).For the COM-MWCNTs and CO 2 -MWCNTs, the q t values were almost constant at pH = 2-11, indicating that their adsorption ability was not affected by the pH level.For the KOH-MWCNTs, q t was lowest at pH = 6.The low adsorption efficiency was because the surface of the KOH-MWCNTs contained OH − ions.The electrostatic attraction between the adsorbent and adsorbate may have been smaller when OH − ions were neutralized with weak H + solution.For the ACID-MWCNT sample, q t was constant at pH = 2-9 and then decreased as the pH was increased to 11.The surface of the ACID-MWCNTs contained H + ions after acid treatment.When the medium was strongly basic, the nanotubes lost H + ions and did not easily interact with PNA molecules through electrostatic force [49].The effect of the stirring speed on PNA adsorption was investigated.The samples were stirred at speeds of 50-150 rpm.The adsorption capacity was found to be positively correlated with the stirring speed (Figure 9c).A high stirring speed may promote interaction between PNA and the nanotubes and decrease the thickness of the diffusion layer on the nanotube surface.However, the change in the adsorption capacity was not obvious when the stirring speed was faster than 100 rpm, indicating that mass transfer resistance could be ignored.The effect of the CNT amount on the adsorption capacity was also investigated.A CNT amount of 10-20 mg (Figure 9d) was employed.The adsorption capacity decreased slightly as the CNT amount was increased from 10 to 15 mg.This decrease in the adsorption capacity was due to more adsorption sites being provided when the amount of nanotubes was increased [50].A further increase in the CNT amount from 15 to 20 mg did not influence the adsorption capacity because the majority of the PNA could already be adsorbed and more PNA was not available for adsorption.Photographs of the residual solutions after degradation of PNA for 150 min are shown in Figure 9e.Visual observation confirmed that the KOH-MWCNTs were a better PNA adsorbent than the CO 2 -MWCNTs or ACID-MWCNTs.

Adsorption Isotherm, Kinetics, and Mechanism
The Langmuir and Freundlich models were employed to acquire information on the PNA distribution between solid and liquid phases at equilibrium.The linear Langmuir and Freundlich equations are represented as follows [51]: ) where n, K L , and K F are the Langmuir and Freundlich constants; C e (mg/L) is the PNA equilibrium concentration; and q e and q L (mg/g) are the equilibrium and maximum adsorption capacity, respectively.The two isotherm models, before and after modification of nanotubes, are plotted in Figure 10.The calculated parameters are listed in Table 3.The determination coefficients R 2 indicate that the best fit for the COM-MWCNTs and ACID-MWCNTs was the Langmuir model.The model revealed that a single PNA layer uniformly covered the surface of the nanotubes [52].Conversely, the Freundlich model was the best fit for the CO 2 -MWCNTs and KOH-MWCNTs.The model assumed multilayer adsorption of PNA on the nanotubes with a heterogeneous surface [53].The separation factors R L of the CNTs were in the range 0.1304-0.6329.These values are less than one, suggesting that nanotubes were beneficial to PNA adsorption [54].
(Figure 9c).A high stirring speed may promote interaction between PNA and the nan tubes and decrease the thickness of the diffusion layer on the nanotube surface.Howeve the change in the adsorption capacity was not obvious when the stirring speed was fast than 100 rpm, indicating that mass transfer resistance could be ignored.The effect of th CNT amount on the adsorption capacity was also investigated.A CNT amount of 10-2 mg (Figure 9d) was employed.The adsorption capacity decreased slightly as the CN amount was increased from 10 to 15 mg.This decrease in the adsorption capacity was du to more adsorption sites being provided when the amount of nanotubes was increase [50].A further increase in the CNT amount from 15 to 20 mg did not influence the adsorp tion capacity because the majority of the PNA could already be adsorbed and more PN was not available for adsorption.Photographs of the residual solutions after degradatio of PNA for 150 min are shown in Figure 9e.Visual observation confirmed that the KOH MWCNTs were a better PNA adsorbent than the CO2-MWCNTs or ACID-MWCNTs.

Adsorption Isotherm, Kinetics, and Mechanism
The Langmuir and Freundlich models were employed to acquire information on th PNA distribution between solid and liquid phases at equilibrium.The linear Langmu and Freundlich equations are represented as follows [51]: where n, KL, and KF are the Langmuir and Freundlich constants; Ce (mg/L) is the PN equilibrium concentration; and qe and qL (mg/g) are the equilibrium and maximu adsorption capacity, respectively.The two isotherm models, before and after modificatio of nanotubes, are plotted in Figure 10.The calculated parameters are listed in Table 3. Th determination coefficients R 2 indicate that the best fit for the COM-MWCNTs and ACID MWCNTs was the Langmuir model.The model revealed that a single PNA lay uniformly covered the surface of the nanotubes [52].Conversely, the Freundlich mod was the best fit for the CO2-MWCNTs and KOH-MWCNTs.The model assume multilayer adsorption of PNA on the nanotubes with a heterogeneous surface [53].Th separation factors RL of the CNTs were in the range 0.1304-0.6329.These values are le than one, suggesting that nanotubes were beneficial to PNA adsorption [54].To understand the PNA adsorption process, pseudo-first-order, pseudo-second-order, and intraparticle diffusion models were used to fit the experimental data.The three models were represented as follows [55]: where k 1 , k 2 , and k i are rate constants; q t and q e (mg/g) are the adsorption capacity at time t and at equilibrium, respectively; and I is the boundary thickness.Linear plots of the pseudo-first-order and pseudo-second-order models are shown in Figure 11.For all types of nanotube, the R 2 values clearly indicate that the pseudo-second-order model gives the best fit to the adsorption data (Table 4), indicating chemisorption between PNA and the nanotubes [56].The experimental q e values were consistent with the calculated q e values.The diffusion mechanism was observed by using an intraparticle diffusion model (Figure 11).Two linearity curves were obtained, indicating that diffusion occurred in two stages.A similar mechanism was discovered by Guo et al. [57] for malachite green adsorption on magnetically activated carbons, which were obtained from peanut and FeCl 3 •6H 2 O by CO 2 activation.In the first stage, PNA molecules diffused into the boundary layer of the nanotubes.In the second stage, the PNA molecules diffused into the pores and then arrived at the surface of the CNTs.The k i values for the first stage were larger than those for the second stage (Table 4).This was because, in the first stage, the mass transfer resistance was not high enough in the bulk fluid, enhancing PNA diffusion.Over time, the diffusion rate decreased, which may be because diffusion of PNA into the nanotube pores needed a longer duration.The CO 2 -MWCNTs and KOH-MWCNTs had a higher PNA diffusion rate than did the COM-MWCNTs and ACID-MWCNTs (Table 4), indicating less competition for adsorption sites in the activated nanotubes with a high surface area.
The activated nanotubes (CO 2 -MWCNT and KOH-MWCNT) exhibited higher adsorption capabilities than did the pristine and acid-modified nanotubes (COM-MWCNT and ACID-MWCNT, respectively), possibly because the activated nanotubes had a high surface area (Table 2) and more sites available for capturing PNA molecules.Furthermore, PNA and the activated nanotubes had C=C double bonds that contained π electrons [10].The π electrons favored π-π interactions between the nanotubes and aromatic PNA rings [58].The activated nanotubes also had a large pore volume, which facilitated electrostatic attraction and van der Waals forces on the surface of the nanotubes and effective interaction between PNA molecules and the nanotubes, thereby increasing the adsorption capacity.

Reutilization Experiments
The cost of nanotube-based wastewater treatment can be reduced by reusing the nanotubes.The adsorption capacities of the COM-MWCNTs, ACID-MWCNTs, CO 2 -MWCNTs, and KOH-MWCNTs after regeneration processes were found to be 86.22%,89.86%, 79.35%, and 82.58%, respectively (Figure 12).The nanotubes retained very high adsorption capacity after five cycles.The experimental results thus confirmed that the MWCNTs are stable even when reused.
The PNA adsorption performance of the activated nanotubes was compared with that of other materials (Table 5) [1,[59][60][61][62][63][64][65][66][67].Different adsorbents have distinct functional groups.Therefore, a comparison is useful.The activated nanotubes had higher adsorption capacities and shorter equilibrium time in our experiments than did adsorbents reported in other studies.Despite the higher adsorption capacity of the single-walled CNT and activated carbon fiber in the literature, their equilibrium times are longer than that of activated nanotubes in the present study.The mesoporous structure of activated nanotubes can provide relatively large pore size for rapid mass transfer.The combined influence of a high surface area and C-O functional groups on the nanotube surface promoted PNA adsorption.
in other studies.Despite the higher adsorption capacity of the single-walled activated carbon fiber in the literature, their equilibrium times are longer tha activated nanotubes in the present study.The mesoporous structure of nanotubes can provide relatively large pore size for rapid mass transfer.The influence of a high surface area and C-O functional groups on the nanotub promoted PNA adsorption.

Conclusions
This study investigated the use of surface-activated and modified MWCN cient nanoadsorbents for the removal of PNA from aqueous solutions.The su and pore volume were greatest for the KOH-MWCNTs, followed by the CO2-M COM-MWCNTs, and ACID-MWCNTs.The appearance of the nanotubes was ously changed after acid or alkali modification.The ACID-MWCNTs exhibited

Conclusions
This study investigated the use of surface-activated and modified MWCNTs as efficient nanoadsorbents for the removal of PNA from aqueous solutions.The surface area and pore volume were greatest for the KOH-MWCNTs, followed by the CO 2 -MWCNTs, COM-MWCNTs, and ACID-MWCNTs.The appearance of the nanotubes was not obviously changed after acid or alkali modification.The ACID-MWCNTs exhibited the highest thermal stability.The CO 2 -MWCNTs and KOH-MWCNTs exhibited higher PNA adsorption capacities (123.4 and 171.3 mg/g, respectively) than did the pristine and acid-treated nanotubes (49.6 and 26.8 mg/g, respectively).The enhancement of adsorption capacity primarily depends on the increase of pore volume and surface area after surfaceactivated MWCNTs.The adsorption capacity of the CO 2 -MWCNTs in the initial 2 min was higher than that of the KOH-MWCNTs due to the wide pores of the CO 2 -MWCNTs.The equilibrium data were fitted using the Langmuir model for the COM-MWCNTs and ACID-MWCNTs and the Freundlich model for the CO 2 -MWCNTs and KOH-MWCNTs.For all types of nanotube, the best fit to the kinetic data was achieved using a pseudosecond-order model.CO 2 -MWCNTs and KOH-MWCNTs were found to have several advantages over the other nanoadsorbents, including a high surface area and abundant C-O functional groups.According to our reutilization experiments, the synthesized nanotubes have excellent reusability, with high adsorption capacity (>79%) after five cycles.The activated MWCNTs synthesized in this study have great potential for effectively removing both organic and inorganic pollutants and can be applied on a large scale for wastewater treatment purposes.

Figure 1 .
Figure 1.XRD patterns of MWCNTs before and after modification.

Figure 2 .
Figure 2. FTIR spectra of MWCNTs before and after modification.

Figure 1 .
Figure 1.XRD patterns of MWCNTs before and after modification.

Figure 1 .
Figure 1.XRD patterns of MWCNTs before and after modification.

Figure 2 .
Figure 2. FTIR spectra of MWCNTs before and after modification.

Figure 2 .
Figure 2. FTIR spectra of MWCNTs before and after modification.

Figure 3 .
Figure 3. Raman spectra of MWCNTs before and after modification.

Figure 3 .
Figure 3. Raman spectra of MWCNTs before and after modification.

Figure 9 .
Figure 9.Effect of different parameters on the PNA adsorption onto MWCNTs: (a) nanotube types, (b) solution pH, (c) stirring speed, (d) amount of nanotube, and (e) optical photographs of residual solutions.

Figure 9 .
Figure 9.Effect of different parameters on the PNA adsorption onto MWCNTs: (a) nanotube types, (b) solution pH, (c) stirring speed, (d) amount of nanotube, and (e) optical photographs of residual solutions.

Table 2 .
Surface area and pore textural parameters of MWCNTs.

Table 2 .
Surface area and pore textural parameters of MWCNTs.

Table 3 .
Parameters of isotherm models for PNA adsorption on MWCNTs.

Table 4 .
Parameters of kinetic models for PNA adsorption on MWCNTs.

Table 5 .
Comparison of adsorption capacity for PNA adsorption onto different nanoads

Table 5 .
Comparison of adsorption capacity for PNA adsorption onto different nanoadsorbents.