Carbon Nanofibers : A New Adsorbent for Copper Removal from Wastewater

This research describes the adsorption of Cu2+ onto a helical ribbon carbon nanofiber. The characterization of carbon nanofiber by zeta potential showed an isoelectronic pH of 1.9. The influence of different adsorption factors, such as stirring speed, temperature, pH, adsorbent concentration, etc., on the Cu2+ adsorption capacity have been evaluated. The pH has a great influence on Cu2+ adsorption, with the maximum adsorption capacity reached at a pH of 10. The experimental data fit well to pseudo-second order kinetic and Langmuir isotherm models (qm = 8.80 mg·g−1) at T = 298 K and pH = 4. The Cu2+ adsorption could be explained by the particle diffusion model. Results showed that carbon nanofiber could be successfully used for the elimination of Cu2+ from wastewater.


Introduction
Heavy metals are characterized by their high relative densities (greater than 5 g•cm −3 ) and atomic weights (between 63.5 and 200.6) [1][2][3].Though many of these are necessary for life, higher amounts above required levels are hazardous [4].
Moreover, heavy metals are not biodegradable, being easily accumulated in living organisms, particularly in humans.These can cause serious illnesses, such as cancer, damage to the nervous system, increases in blood and respiratory pressure, kidney failure, and could become lethal at high concentrations [1,[5][6][7].For this reason, the pollution of water due to the presence of heavy metals is a major concern today world-wide.Each year large, volumes of wastewater are produced from different industrial sectors, such as mining, metallurgy and smelting, among others.The most common heavy metals present in wastewaters are nickel, zinc, lead, iron, copper, arsenic, cadmium, and uranium [8][9][10].
Among these, cooper is considered a harmful metal [9,11,12].There are three forms of copper: Cu 0 metal, Cu + cuprous ion, and Cu 2+ cupric ion, this last form normally appearing in the environment and considered to be the most toxic of the three [13,14].The main contributors of copper to the environment are mining industries, printing circuits, electroplating industries, paints, fertilizers, plastics and etching, etc. [15][16][17].Certain industries that release wastewater into the environment have reported Cu 2+ concentrations ranging from 2.0 mg•L −1 to 900 mg•L −1 [18].Therefore, the recovery of heavy metals, in general, and in particular of copper, is a subject of great social relevance due to the environmental and economic benefits of eliminating contaminates from industrial effluents.
A great number of methods can be used for the recovery or removal of heavy metals from liquid effluents [19].The choice of one method over another is based mainly on an economic question and metal concentration [1].Conventional methods include bioremediation [20,21], reverse osmosis [22], electrochemical treatment [23], coagulation [24], precipitation [25,26] solvent extraction [27][28][29], membrane filtration [30,31], electrodialysis [32], supercritical fluids extraction [33,34], ion exchange [35][36][37][38], and flotation [39,40].Adsorption is one of the most popular methods for eliminating heavy metals [41].It is a technology in which the pollutants, particularly in the case of heavy metals, can interact in a physical or chemical way with the active elements of the adsorbent.Adsorption is a technique that is easy to use and design, and it does not produce contaminants [42].Moreover, another advantage of the adsorption process is the possibility to regenerate the adsorbent by desorption, being therefore a reversible technique, so it is considered a technology that is not aggressive with the environment.These advantages have made adsorption a leading method for wastewater treatment [13].
Different types of adsorbent have been developed, such as active carbon, clay ore, chelating materials, natural chitosan/zeolites, etc. [19,41,43,44].In the last twenty years, the interest in nanotechnology has been growing.Materials such as fullerenes, carbon nanotubes, graphene and graphene oxide have a large number of applications in different fields, including the removal of heavy metals from wastewater [38,45].Their unique properties, including high resistance, electrical conductivity and thermal stability, as well as having large specific surface areas, make them sustainable adsorbents for wastewater treatment, and more specifically, for the elimination of heavy metals [46,47] This paper reports the results derived from Cu 2+ adsorption on new helix spiral carbon nanofibers (CNF).CNF are structures composed of a helical ribbons of graphene, spiraled about and angled to the fiber axis, similar to the structure of an Archimedes screw with flights inclined to the axis [48].
There are many factors that influence heavy metal adsorption efficiency, such as initial metal concentration, temperature, sorbate amount, pH, and contact time [42,43].Therefore, the objective of this research was to determine the effects of various experimental variables on Cu 2+ -CNF adsorption systems.Furthermore, kinetic mechanisms, adsorption isotherms models and thermodynamics parameters were investigated.

Materials and Methods
Carbon nanofibers donated by Antolin group S.A. were used as an adsorbent [49].Stock metal solution was prepared from CuSO 4 (Fluka, Munich, Germany).All other reagents used in the work were analytical reagent grade.

Characterization of Carbon Nanofibers
Carbon nanofibers were examined using a Hitachi S4800 (Tokyo, Japan) scanning electron apparatus equipped with a detectors of secondary, backscattered and transmission electrons, and an Oxford Inca Energy dispersion microanalysis system.Nanomaterials were dispersed in ethanol, and drops of the suspension material were deposited on TEM grating.
The zeta potential was measured using a Zetasizer Malvern Nano ZS (Malvern Panalytical Ltd., Worcestershire, UK) at 25 • C. To study the influence of pH, aqueous suspensions were prepared in pH solutions between 1 and 13 using solutions of 0.5 M of HCl and NaOH.

Batch Adsorption Experiments
The batch adsorption experiments were carried out in a 250 mL glass reactor capable of mechanical (impeller) shaking.
Solutions (100 mL) containing Cu 2+ were used, and weighed amounts of the adsorbent, carbon nanofibers (CNF), were added.The pH was adjusted using HCl and NaOH.The residual copper content in the aqueous solution was determined by AAS (Perkin Elmer 1100 B spectrophotometer, Waltham, MA, USA).These were filtered prior to solution analyses.The metal uptake capacity was calculated using Equation (1): where q t is the amount metals ions adsorbed (mg•Cu•g −1 sorbent), while C 0 and C t are the initial and contact time metal ion concentrations (mg•Cu•L −1 ) in the aqueous solution, respectively.V is the volume of the solution (0.1 L), and M is the mass of the sorbent (g).The removal efficiency was calculated using the following equation: where C e (mg•Cu•L −1 ) is the equilibrium metal concentration in solution.

Characterization of CNF
TEM micrographs of CNF are shown in Figure 1.The helix spiral structure is observed.The structure of CNF is a continuous cone-helix.The presence of pyrolytic amorphous carbon in the CNF is not observed [49].CNF structural details are presented in references [48][49][50].Table 1 shows the elemental composition.
where qt is the amount metals ions adsorbed (mg•Cu•g −1 sorbent), while C0 and Ct are the initial and contact time metal ion concentrations (mg•Cu•L −1 ) in the aqueous solution, respectively.V is the volume of the solution (0.1 L), and M is the mass of the sorbent (g).
The removal efficiency was calculated using the following equation: where Ce (mg•Cu•L −1 ) is the equilibrium metal concentration in solution.

Characterization of CNF
TEM micrographs of CNF are shown in Figure 1.The helix spiral structure is observed.The structure of CNF is a continuous cone-helix.The presence of pyrolytic amorphous carbon in the CNF is not observed [49].CNF structural details are presented in references [48][49][50].Table 1 shows the elemental composition.The isoelectric point (IEP) of CNF is obtained at a pH value of 1.9 (Figure 2).At a pH values around the pH(IEP), the sorbent surface charge is neutral.When the pH value is lower than the isoelectronic point, the sorbent surface has a positive charge, and a negative charge when the pH is greater than 1.9 [37,51].The surface charge could affect in the adsorption properties of the adsorbent [51].The isoelectric point (IEP) of CNF is obtained at a pH value of 1.9 (Figure 2).At a pH values around the pH(IEP), the sorbent surface charge is neutral.When the pH value is lower than the isoelectronic point, the sorbent surface has a positive charge, and a negative charge when the pH is greater than 1.9 [37,51].The surface charge could affect in the adsorption properties of the adsorbent [51].

Sorption Bach Experiment
To research the efficiency of the CNF in Cu 2+ adsorption, parameters such as contact time, stirring speed, temperature, pH, carbon nanomaterials concentration, and copper concentration have been studied.

Sorption Bach Experiment
To research the efficiency of the CNF in Cu 2+ adsorption, parameters such as contact time, stirring speed, temperature, pH, carbon nanomaterials concentration, and copper concentration have been studied.

Effect of the Contact Time and the Stirring Speed
Figure 3 shows the effect of the contact time on Cu 2+ adsorption.The adsorption kinetics was investigated for better understanding of the dynamics of adsorption of metal on the CNF.The removal of Cu 2+ between 0-30 min increased sharply, and then tended to became almost stable, which denoted accomplishment of equilibrium at 30 min.From this point, the qt values remained constant.This behavior could be due to the fact that, at the beginning, all of the active sites on the CNF surface were empty and the copper concentration high, and after 30 min, the surface active sites available diminish and the copper concentration lower, so the metal uptake remain constant [52].The copper adsorption was a quick process.

Sorption Bach Experiment
To research the efficiency of the CNF in Cu 2+ adsorption, parameters such as contact time, stirring speed, temperature, pH, carbon nanomaterials concentration, and copper concentration have been studied.

Effect of the Contact Time and the Stirring Speed
Figure 3 shows the effect of the contact time on Cu 2+ adsorption.The adsorption kinetics was investigated for better understanding of the dynamics of adsorption of metal on the CNF.The removal of Cu 2+ between 0-30 min increased sharply, and then tended to became almost stable, which denoted accomplishment of equilibrium at 30 min.From this point, the qt values remained constant.This behavior could be due to the fact that, at the beginning, all of the active sites on the CNF surface were empty and the copper concentration high, and after 30 min, the surface active sites available diminish and the copper concentration lower, so the metal uptake remain constant [52].The copper adsorption was a quick process.The equilibrium time could be decreased by increasing the stirring speed, so the appropriate hydrodynamic conditions were investigated.Table 2 shows that an increase in the stirring speed increased qt values, reaching a maximum between 500-2000 rpm.In this stirring speed interval, the thickness of the aqueous film diffusion layer and the aqueous resistance to mass transfer are at a minimum.An increase in the stirring speed to higher than 2000 rpm decreases the Cu 2+ adsorption.This could be due the formation of local equilibrium between the Cu 2+ solution and the surrounding CNF and/or the agglomeration of the nanoparticles.

Stirring Speed (rpm)
q e (mg

Adsorption Kinetic and Diffusion Mechanism
To understand the process of Cu 2+ adsorption onto CNF, three kinetic adsorption models have been assessed [53]: (a) pseudo-first order [51] ln(q e − q t ) = ln q e − k 1 t (b) pseudo-second order [54] (c) Elovich model [55] where q e and q t (mg•g −1 ) are the sorption capacities at equilibrium and at time t (min), respectively, k 1 is first-order adsorption rate constant (min −1 ), and k 2 (g•mg −1 •min −1 ) is the pseudo-second order constant.α is the initial adsorption rate (mg•min −1 ) at contact time t = 0 min, and β is the extent of surface coverage and activated energy (g•mg −1 ).
According to the calculated data, among the different adsorption reaction models, the higher correlation coefficient, R 2 , was obtained for the pseudo-second order kinetics model (Equation ( 4) and Table 3).The q e values derived from the model and the experiment are really similar: q e (experimental) 6.51 mg•g −1 [56].

Kinetic reaction models
Pseudo-first 0.0594 min −1 q e = 21.04 mg•g −1 0.9357 Pseudo-second 0.0574 g•mg −1 •min −1 q e = 6.62 mg•g −1 0.9985 Elovich α = 6.18 mg•g −1 •min −1 β = 8.12 g•mg −1 0.9588 Three adsorption mechanisms have been considered to evaluate the adsorption of Cu 2+ on CNF [57,58]: (a) Film diffusion controlled process: The metal species diffused from the aqueous solution to the CNF surface (b) Particle diffusion controlled process: The ion diffused inside CNF ln 1 − q t q e 2 = −Kt (c) Moving boundary process where K (min −1 ) is the rate constant model, q e and q t (mg•g −1 ) are the sorption capacities at equilibrium and contact time, respectively, and t (min).
The results obtained by the kinetic diffusion models show that the Cu 2+ adsorption onto the CNF could be better explained by the particle diffusion model, where K = 0.0525 min −1 (Table 4).As seen in Figure 4, at a low pH of two, the adsorption of Cu 2+ onto CNF is very low.This is due to the pH value being very near the zero potential pH = 1.9, the charge in the surface being 0 (neutral), and the Cu 2+ adsorption being nil or very low.At pH values lower than two, the potential of the surface charge of the CNF is positive, the cations in solution are repealed from the adsorbent, and it is not adsorbed.When the pH increases to between 4 and 10, the cooper adsorption increased to its maximum at pH 10 (q e = 8.88 mg•g −1 ).Above the isoelectronic pH, the nanomaterial surfaces have a negative charge and attract the metal ions facilitating the adsorption [18,37].
where K (min −1 ) is the rate constant model, qe and qt (mg•g −1 ) are the sorption capacities at equilibrium and contact time, respectively, and t (min).
The results obtained by the kinetic diffusion models show that the Cu 2+ adsorption onto the CNF could be better explained by the particle diffusion model, where K = 0.0525 min −1 (Table 4).As seen in Figure 4, at a low pH of two, the adsorption of Cu 2+ onto CNF is very low.This is due to the pH value being very near the zero potential pH = 1.9, the charge in the surface being 0 (neutral), and the Cu 2+ adsorption being nil or very low.At pH values lower than two, the potential of the surface charge of the CNF is positive, the cations in solution are repealed from the adsorbent, and it is not adsorbed.When the pH increases to between 4 and 10, the cooper adsorption increased to its maximum at pH 10 (qe = 8.88 mg•g −1 ).Above the isoelectronic pH, the nanomaterial surfaces have a negative charge and attract the metal ions facilitating the adsorption [18,37].

Effect of the Copper Concentration
The adsorption capacity depends on Cu 2+ concentration; Table 5 shows an increase in copper concentration in the aqueous phase resulting in a reduction of the recovery percentage.This may be associated with a competitive effect of copper by the active position of the adsorbent.The adsorption equilibrium data of the Cu 2+ -CNF system were checked in terms of three isotherm models: Langmuir, Freundlich, Dubinin-Radushkevich and Temkin [59].The following equations have been used: (a) Langmuir [60,61]: (b) Freundlich [62]: where q e (mg•g −1 ) is the sorption capacities at equilibrium time, C e (mg•L −1 ) is the metal ion concentration in the aqueous phase at equilibrium, q m (mg•g −1 ) is maximum sorption capacity, and b (L•mg −1 ) is the Langmuir sorption constant.R L is a dimensional constant separation [63].C 0 (mg•L −1 ) is the maximun initial copper concentration.K F is the Freundlich constant (L•g −1 ), and 1/n is a heterogeneity factor constant, their values must be less than one for favorable adsorption.(c) Dubinin-Radushkevich [64]: ln where: where q e are mg•g −1 in equilibrium, q d is the theoretical adsorption (mg•g −1 ), K is a constant related to the mean free energy in the adsorption (mol 2 •J −2 ) and E (KJ•mol −1 ) [63].ε is the Polanyi adsorption potential, R is the universal gas constant (8.314J•K −1 •mol −1 ), T is the temperature in K, and C e is the metal ion concentration in the aqueous phase at equilibrium in mg•L −1 .(d) Temkin [65]: where q e is the sorption capacities at equilibrium time in mol•g −1 , C e is the metal ion concentration in the aqueous phase at equilibrium in mol•L −1 , and K T is the isotherm equilibrium binding constant (L•g −1 ); B is a constant related to the sorption heat in J•mol −1 .
According to the values of the correlation coefficient, R 2 , the experimental results fit well to the studied isotherms, Table 6.However, the best fit is obtained for the Langmuir model.The q m value obtained from this model is similar to the experimental one, 8.77 mg•g −1 .The separation factor, R L , shows that the adsorption process is favorable if the R L value is 0 < R L < 1, unfavorable when R L > 1, linear when R L = 1, and irreversible at values of R L = 0.The Cu 2+ adsorption on CNF is favorable with a R L value between zero and one, the value here being 5.6 × 10 −3 .The 1/n value obtained from adjusting the data to the Freundlich isotherm is 9.12 × 10 −2 , also indicating that the copper adsorption on CNF is favorable.Table 6.Calculated isotherms parameters.Experimental conditions: CNF = 0.1 g, V = 100 mL of [Cu 2+ ] = variable pH = 4, T = 298 K, θ = 500 rpm.

Langmuir Model
Freundlich Model The energy value calculated using the Dubinin-Raduskevich model was E = 7.34 (KJ•mol −1 ).This is lower than 8 KJ•mol −1 , which indicates that the mechanism of adsorption has a physical nature [63].

Effect of the CNF Dosage
The CNF dosage was modified in order to evaluate the effect of this parameter.It was expected that an increase in the CNF dosage will increase the adsorption percentage and diminish the adsorption capacity values.
The objective of this research was to achieve the highest possible metal recovery.The best result was obtained using 0.2 g of CNF, with 97% of the copper recovered (see Table 7).The influence of the temperature (293-333 K) on Cu 2+ adsorption on CNF have been studied.The assay conditions were 0.01 g•L −1 of Cu(II), 4 pH, 0.1 g of CNF adsorbent, and a stirring speed of 500 rpm. Figure 5 shows the variation of q t vs. t.An increase of the qe value with temperature was observed, having reached a maximum at 333 K, 8 mg•g −1 .Therefore, the process is considered endothermic.The increase of adsorption with temperature can be due to different factors, such as a change in the pore size of the nanofibers and/or an increase of the active sites.In addition, the collisions between the metal and the surface of the nanofibers are greater, thereby contributing to a rise in the adsorption capacity [18].
The experimental data obtained at different temperatures (293-333 K) was used to calculate the thermodynamics parameters, ΔH°, ΔG°, and ΔS°, using the following equation: where ΔS° is the entropy of the reaction, ΔH° is the enthalpy, T is the temperature in Kelvin, and R is the universal gas constant (8.314J•K −1 •mol −1 ).Kd is defined as follows: where qe, (mg•g −1 ) and Ce (mg•L −1 ) are the equilibrium capacity on the solid and the liquid phases, An increase of the q e value with temperature was observed, having reached a maximum at 333 K, 8 mg•g −1 .Therefore, the process is considered endothermic.The increase of adsorption with temperature can be due to different factors, such as a change in the pore size of the nanofibers and/or an increase of the active sites.In addition, the collisions between the metal and the surface of the nanofibers are greater, thereby contributing to a rise in the adsorption capacity [18].
The experimental data obtained at different temperatures (293-333 K) was used to calculate the thermodynamics parameters, ∆H • , ∆G • , and ∆S • , using the following equation: where ∆S • is the entropy of the reaction, ∆H • is the enthalpy, T is the temperature in Kelvin, and R is the universal gas constant (8.314J•K −1 •mol −1 ).K d is defined as follows: Metals 2018, 8, 914 9 of 13 where q e , (mg•g −1 ) and C e (mg•L −1 ) are the equilibrium capacity on the solid and the liquid phases, respectively.Plotting the linear fit ln C e , CNF C e vs. 1 T to the slope could be used to calculate the ∆H • value (Table 8).For the entire temperature range studied, the ∆G • values were negative, this indicating that the adsorption process is spontaneous.These values are lower than 18 KJ•mol −1 , characteristic of a physisorption [18], and this agrees with the result obtained from R L Langmuir parameter (see Table 6).
The positive enthalpy confirms the endothermic behavior of Cu 2+ adsorption on the CNF.The ∆S • value is positive, which indicates an increase in the randomness in the CNF/Cu 2+ system.In the temperature range studied, the ∆H • is lower than T•∆S • , so the sorption process was dominated by entropic rather than enthalpic changes [51,62].

Cu 2+ Elution
According to the results for Cu 2+ adsorption on CNF, it is expected that an increase in the acid concentration changes the adsorption conditions, and Cu 2+ could then be eluted.The effect of the acid concentration in elution was studied.For 30 min, 0.05 g of CNF load with 4.83 mg•g −1 of Cu 2+ were put in contact with 20 mL of H 2 SO 4 in solutions of different concentrations at a stirring speed of 500 rpm.
An increase in the acid concentration increased the elution percentage (see Table 9), reaching a percentage of around 96% with a H 2 SO 4 concentration of 4 M.The kinetic elution study was performed using a solution of 2 M sulfuric acid, with equilibrium reached after 15 min.The elution kinetics was faster than the adsorption.

Conclusions
The results obtained in this research prove that helical carbon nanofibers are a potential adsorbent of Cu 2+ from wastewaters.Of the parameters studied, pH has the greatest influence on copper adsorption capacity, reaching a maximum q m value of around 8.80 mg•g −1 at a pH value of 10.An increase in the temperature does not influence the adsorption capacity; however, the equilibrium is reached in less time.
The Cu 2+ kinetics fitted a pseudo-second order (0.0574 g•mg −1 •min −1 ).The Cu 2+ adsorption is considered a particle diffusion controlled process.The experimental data adjusted to a Langmuir model, the q m = 8.80 mg•g −1 , is a favorable, endothermic (∆H • = 26.76KJ•mol −1 ) and physical nature adsorption.Metal elution was carried out with a H 2 SO 4 4 M solution, at 30 min, reaching an elution percentage of 96%.

Figure 5 .
Figure5.Temperature effect on the variation of q t (mg•g −1 ) vs. time.

Table 2 .
Effect of the stirring speed on the Cu 2+ adsorption onto the CNF.Experimental conditions: CNF