Adsorption of Phenol on Commercial Activated Carbons: Modelling and Interpretation

Adsorption by activated carbons (AC) is an effective option for phenolic wastewater treatment. Three commercial AC, including coal-derived granular activated carbons (GAC950), coal-derived powdered activated carbons (PAC800), and coconut shell-derived powdered activated carbons (PAC1000), were utilized as adsorbent to study its viability and efficiency for phenol removal from wastewater. Pseudo-first order, pseudo-second order, and the Weber–Morris kinetic models were used to find out the kinetic parameters and mechanism of adsorption process. Further, to describe the equilibrium isotherms, the experimental data were analyzed by the Langmuir and Freundlich isotherm models. According to the experimental results, AC presented a micro/mesoporous structure, and the removal of phenol by AC was affected by initial phenol concentration, contact time, pH, temperature, and humic acid (HA) concentration. The pseudo-second order kinetic and Langmuir models were found to fit the experimental data very well, and the maximum adsorption capacity was 169.91, 176.58, and 212.96 mg/g for GAC950, PAC800, and PAC1000, respectively, which was attributed to differences in their precursors and physical appearance. Finally, it was hard for phenol to be desorbed in a natural environment, which confirmed that commercial AC are effective adsorbents for phenol removal from effluent wastewater.


Introduction
Activated carbons (AC) are carbonaceous materials with large specific surface area, superior porosity, high physicochemical-stability, and excellent surface reactivity, extensively used for adsorption of several environmental contaminants, gas separation, heterogeneous catalysis, gas storage, and gas masks, among others [1]. The worldwide demand of AC reached 12,804,000 tons in 2015, and the research into AC has also had an important increase, yielding a production of 17,516 research papers from 1995 to 2016 [2]. Almost all carbon-rich precursors can be converted to AC through stabilization (if required), carbonization, and activation [3]. Selection of the raw material depends not only on the anticipated role of carbon-surface functionalities for given applications, but on the availability and low cost of raw material. In China, coal and coconut shell are the most common precursors for the largescale synthesis of commercial AC. Moreover, there are two most common physical forms, in which AC are used, these being granular active carbons (GAC) and powdered active carbons (PAC) [4]. Various commercial AC with different properties are manufactured for different applications.
Phenol is an organic compound found in wastewater disposed from many industries, such as refinery, petrochemical, coal processing, pharmaceutical, polymeric resin, and pesticide industries,

Physical and Chemical Characterization of AC
The surface morphologies of AC were observed by scanning electron microscope (SEM, Hitachi SU3500, Tokyo, Japan) after coating with gold for 30 s. The AC textural properties (specific surface area (As), average pore size, and total pore volume) were determined on the basis of nitrogen adsorption/desorption isotherms at 77 K, using a Micrometrics Gemini 2390 system. The specific surface area was determined with the standard Brunauer-Emmett-Teller (BET) method and cumulative pore volumes by Barrett-Joyner-Halenda analysis. Fourier transform infrared (FTIR) spectroscopy (Tracer-100, Tokyo, Japan) was applied to determine the surface chemistry of the AC. All the samples and the dried KBr were ground at a mass ratio of 1:100, and the spectra were recorded from 400 to 3 of 13 4000 cm −1 . The pH PZC (point of zero charge) of AC was determined by mixing 1 g of each AC with 20 mL of CO 2 -free deionized water, according to the procedure described by Moreno-Castilla et al. (2000) [18].

Phenol Adsorption on AC
Several key parameters, including the initial concentration (10-400 mg/L), contact time (0-400 min), pH (3)(4)(5)(6)(7)(8)(9)(10)(11), temperature (20-60 • C), and the concentration of humid acid (HA, 0-100 mg/L) were systematically taken into account to optimize the adsorption conditions. A total of 50 mg of AC and 80 mL of phenol solution were added to polyethylene tubes, which were placed onto a water bath oscillator at 150 r/min. Afterwards, the carbons were filtered and the concentration of phenol was analyzed by 4-aminoantipyrine spectrophotometric method. Phenol solution (1 mL) was taken in a clean dry colorimetric tube, then the volume was fixed to 50 mL with deionized water. The pH of solution was adjusted by adding 0.5 mL buffer solution (20% ammonia-ammonium chloride solution), and then 1 mL of 4-aminoantipyrine and 1 mL of potassium ferricyanide were added and mixed well. In this method, phenol reacts with 4-aminoantipyrine in the presence of potassium ferricyanide to form a colored antipyrine dye [19,20]. The colored sample was measured by UV-1100 UV-visible spectrophotometer (Shanghai Mapada Instrument Co. Ltd., Shanghai, China) at 510 nm. The assays were performed in replicate (n = 3) and blank tests were realized. The adsorption percentage (R, %) and the amount of phenols (q e , mg/g) on AC at equilibrium were calculated according to Equations (1) and (2): where C 0 (mg/L) and C e (mg/L) are the initial and equilibrium concentrations of phenol, respectively; V (L) is the volume of the phenol solution; and m (g) is the amount of AC applied for adsorption.

Adsorption Kinetic Models
The kinetics of adsorption is an important characteristic to define the efficiency of adsorption. In order to investigate the mechanism of adsorption and kinetic parameters, sorption data was analyzed using pseudo-first order (Equation (3)), pseudo-second order (Equation (4)), and the Weber-Morris model (Equation (5)) [21]. Compared with the pseudo-first order and pseudo-second order kinetic models, the Weber-Morris model can generally identify the diffusion mechanism and rate controlling steps affecting the kinetics of adsorption. If an adsorption process is solely governed by intraparticle diffusion, the initial part of the Weber-Morris plot is a straight line passing through the origin. Otherwise, the intraparticle diffusion is not the only rate-controlling step but some degree of the boundary layer diffusion (or external mass transfer) also controls the adsorption [22].
where t (min) and q t (mg/g) are, respectively, the adsorption time and the amount of phenol adsorbed on AC at any time, t; q e,exp (mg/g) and q e (mg/g) are the amount of phenol adsorbed at experimental-equilibrium and calculated-equilibrium, respectively; k 1 (min −1 ), k 2 (g/mg/min), and k 3 (mg/g/min 1/2 ) are the rate constants of the pseudo-first order, pseudo-second order, and the Weber-Morris models, respectively.

Adsorption Equilibrium Models
The equilibrium adsorption data were fitted by Langmuir (Equation (6)) and Freundlich (Equation (7)) isotherm equations. The Langmuir isotherm is valid for monolayer adsorption onto a surface with a finite number of identical sites, and it is based on the assumption of adsorption homogeneity, such as equally available adsorption sites, monolayer surface coverage, and no interactions between adsorbed species. The Freundlich equation is an empirical relationship in which it is assumed that the non-ideal adsorption takes place on a heterogeneous surface with different adsorption energy and characters. It has been successfully applied for applications involving multilayer adsorption [23,24].
log q e = log K F + 1 n log C e where q e (mg/g) is the amount of phenol adsorbed at equilibrium, C e (mg/L) is the equilibrium concentration of phenol, q m (mg/g) is the maximum adsorption capacity, and K L (L/mg) is the Langmuir adsorption constant. K F (mg/g) and 1/n (unitless) are the Freundlich model constants related to adsorption capacity and adsorption intensity, respectively.

Phenol Desorption on AC
Desorption studies were conducted using various solution (deionized water/acid/alkali) and the equilibration time was 24 h. Nine samples of 50 mg each of AC saturated with phenol at initial concentration of 300 mg/L were prepared. Six AC samples were taken to be used at desorption experiment by NaOH (pH 12) and H 2 SO 4 (pH 3) solution, whereas the other samples were taken in polyethylene tubes, including only deionized water to determine amount of desorbed phenol by shaker. After equilibrium, the phenol concentration in the solution was measured. The assays were performed in replicate (n = 3) and the release percentage (R d ) was calculated through the following equation: where C d (mg/L) is the concentration of phenol in release solution; V d (L) is the volume of the release solution, and V d = 0.08 in this work; m (g) is the amount of AC applied for release; and q e (mg/g) is the amount of phenol on AC at equilibrium.

Characterization of AC
Activated carbon is generally described as an amorphous form of graphite with a random structure of graphite plates, having highly porous structure with a range of cracks and crevices reaching molecular dimensions. As shown in Figure 1, GAC 950 demonstrated a skeletal structure with pores, whereas PAC 800 and PAC 1000 had a uniform morphology. Further, the number of pores in PAC 1000 were found to be more than that of PAC 800 , and these were also relatively larger in size. The presence of cavities in the AC structure is favorable for the adsorption process because they enable the penetration of phenol molecules into the adsorbent [25,26].
Physisorption with inert gas (N 2 ) is one of the most common and effective ways to determine the specific surface area of porous materials. All the adsorption isotherms showed a mixture of type I and type IV isotherms. Thereafter, a hysteresis in the desorption of the N 2 isotherm appeared. This result was consistent with the pore size distribution curve and demonstrated the microporosity to mesoporosity (1.5-50 nm) of AC ( Figure 2) [27]. According to the nitrogen isotherms, GAC 950 , PAC 800 , and PAC 1000 showed high surface areas of 700.57 m 2 /g, 542.33 m 2 /g, and 1025.02 m 2 /g, respectively.
The average pore size defines the ability of the adsorbate molecules to penetrate inside the AC; only when the pores have a diameter larger than the effective molecular diameter of the adsorbate, the adsorbate molecules penetrate the adsorbent [28]. Phenol has an effective molecule diameter of 0.75 nm, and the average pore sizes presented by GAC 950 , PAC 800 , and PAC 1000 , were 2.22, 3.25, and 5.01 nm, respectively; all AC were suitable for phenol adsorption.
The functional groups on the AC (before adsorption and after adsorption) were measured by the FTIR spectroscopy, as depicted in Figure 3. The band at 3430 cm -1 referred to the O-H functional groups for all AC, and strong absorptions at around 1200 and 1500 cm -1 were observed, which might be assigned to C-H stretching of -CH 3 , -CH 2 groups, and C=O stretching of quinone groups [8,25]. The presence of these oxygenated groups confers a negative charge density in AC surface. The phenol adsorption onto activated carbon occurred through the formation of a donor-acceptor complex between the electron donor groups (e.g., carbonyls) at the AC surface and the aromatic ring of the phenol that acts as the acceptor [14,29]. In addition, the hydroxyl adsorption band widened and a diffuse broad spectrum appeared after adsorption; which suggested that hydrogen bond also favor the phenol adsorption [30]. and PAC1000 showed high surface areas of 700.57 m 2 /g, 542.33 m 2 /g, and 1025.02 m 2 /g, respectively.
The average pore size defines the ability of the adsorbate molecules to penetrate inside the AC; only when the pores have a diameter larger than the effective molecular diameter of the adsorbate, the adsorbate molecules penetrate the adsorbent [28]. Phenol has an effective molecule diameter of 0.75 nm, and the average pore sizes presented by GAC950, PAC800, and PAC1000, were 2.22, 3.25, and 5.01 nm, respectively; all AC were suitable for phenol adsorption. The functional groups on the AC (before adsorption and after adsorption) were measured by the FTIR spectroscopy, as depicted in Figure 3. The band at 3430 cm -1 referred to the O-H functional groups for all AC, and strong absorptions at around 1200 and 1500 cm -1 were observed, which might be assigned to C-H stretching of -CH3, -CH2 groups, and C=O stretching of quinone groups [8,25]. The presence of these oxygenated groups confers a negative charge density in AC surface. The phenol adsorption onto activated carbon occurred through the formation of a donor-acceptor complex between the electron donor groups (e.g., carbonyls) at the AC surface and the aromatic ring of the phenol that acts as the acceptor [14,29]. In addition, the hydroxyl adsorption band widened and a diffuse broad spectrum appeared after adsorption; which suggested that hydrogen bond also favor the phenol adsorption [30].    and PAC1000 showed high surface areas of 700.57 m 2 /g, 542.33 m 2 /g, and 1025.02 m 2 /g, respectively.
The average pore size defines the ability of the adsorbate molecules to penetrate inside the AC; only when the pores have a diameter larger than the effective molecular diameter of the adsorbate, the adsorbate molecules penetrate the adsorbent [28]. Phenol has an effective molecule diameter of 0.75 nm, and the average pore sizes presented by GAC950, PAC800, and PAC1000, were 2.22, 3.25, and 5.01 nm, respectively; all AC were suitable for phenol adsorption. The functional groups on the AC (before adsorption and after adsorption) were measured by the FTIR spectroscopy, as depicted in Figure 3. The band at 3430 cm -1 referred to the O-H functional groups for all AC, and strong absorptions at around 1200 and 1500 cm -1 were observed, which might be assigned to C-H stretching of -CH3, -CH2 groups, and C=O stretching of quinone groups [8,25]. The presence of these oxygenated groups confers a negative charge density in AC surface. The phenol adsorption onto activated carbon occurred through the formation of a donor-acceptor complex between the electron donor groups (e.g., carbonyls) at the AC surface and the aromatic ring of the phenol that acts as the acceptor [14,29]. In addition, the hydroxyl adsorption band widened and a diffuse broad spectrum appeared after adsorption; which suggested that hydrogen bond also favor the phenol adsorption [30].    and PAC1000 showed high surface areas of 700.57 m 2 /g, 542.33 m 2 /g, and 1025.02 m 2 /g, respectively.
The average pore size defines the ability of the adsorbate molecules to penetrate inside the AC; only when the pores have a diameter larger than the effective molecular diameter of the adsorbate, the adsorbate molecules penetrate the adsorbent [28]. Phenol has an effective molecule diameter of 0.75 nm, and the average pore sizes presented by GAC950, PAC800, and PAC1000, were 2.22, 3.25, and 5.01 nm, respectively; all AC were suitable for phenol adsorption. The functional groups on the AC (before adsorption and after adsorption) were measured by the FTIR spectroscopy, as depicted in Figure 3. The band at 3430 cm -1 referred to the O-H functional groups for all AC, and strong absorptions at around 1200 and 1500 cm -1 were observed, which might be assigned to C-H stretching of -CH3, -CH2 groups, and C=O stretching of quinone groups [8,25]. The presence of these oxygenated groups confers a negative charge density in AC surface. The phenol adsorption onto activated carbon occurred through the formation of a donor-acceptor complex between the electron donor groups (e.g., carbonyls) at the AC surface and the aromatic ring of the phenol that acts as the acceptor [14,29]. In addition, the hydroxyl adsorption band widened and a diffuse broad spectrum appeared after adsorption; which suggested that hydrogen bond also favor the phenol adsorption [30].

Kinetics of Phenol Adsorption on AC
According to previous reports, q e of any adsorbent is highly dependent on the contact time with the adsorbate. The determination of equilibrium time of phenol adsorption was conducted at the initial phenol concentration range of 300 mg/L under neutral pH and temperature of 25 • C, and the experimental result is given in Figure 4. The rate of phenol adsorption was found to be very rapid during the initial 100 min. Thereafter, the adsorption capacity reached a plateau as the contact time increased further and exhibited adsorption at equilibrium after 180 min. This behavior can be explained because a large number of vacant surface sites are available for adsorption during the initial stage, and after a lapse of time, the remaining vacant surface sites are difficult to occupy due to repulsive forces between the solute molecules on the solid and bulk phases [31]. At equilibrium, the q e,exp of GAC 950 , PAC 800 , and PAC 1000 were 169.91, 176.58, and 212.96 mg/g, respectively. It was also found that the removal of phenol by GAC 950 was only slightly less than that by PAC 800 at any contact time and that the removal of phenol by PAC 800 was around 17% less than that by PAC 1000 .
stage, and after a lapse of time, the remaining vacant surface sites are difficult to occupy due to repulsive forces between the solute molecules on the solid and bulk phases [31]. At equilibrium, the qe,exp of GAC950, PAC800, and PAC1000 were 169.91, 176.58, and 212.96 mg/g, respectively. It was also found that the removal of phenol by GAC950 was only slightly less than that by PAC800 at any contact time and that the removal of phenol by PAC800 was around 17% less than that by PAC1000.
Pseudo-first order, pseudo-second order, and the Weber-Morris model kinetic parameters are represented in Table 1. The correlation coefficient for the pseudo-second order was higher than that of pseudo-first order kinetic model for both AC, and the calculated values of qe were extremely close to the experimental data (qe,exp), whereas the gap between calculated values and experimental data for pseudo-first-order model was more obvious. This indicated that the adsorption perfectly complied with pseudo-second order reaction and the adsorption of phenol appeared to be controlled by the chemisorption process simultaneously. In other words, the adsorption of phenol happened via surface exchange reactions until the surface functional sites were fully occupied; thereafter, phenol molecules diffused into the AC network for further interactions (such as inclusion complex, hydrogen bonding, hydrogen-phobic interactions) [32]. This result was consistent with the kinetics behavior of phenol adsorption on other AC samples [9,25]. On the other hand, the Weber-Morris plots were shown in Figure 5. The plots were not linear over the whole time range, implying that more than one process affected the adsorption, and the adsorption process contained both the surface adsorption and intraparticle diffusion. These behaviors can be explained due to the presence of micropores in the adsorbent.  Pseudo-first order, pseudo-second order, and the Weber-Morris model kinetic parameters are represented in Table 1. The correlation coefficient for the pseudo-second order was higher than that of pseudo-first order kinetic model for both AC, and the calculated values of q e were extremely close to the experimental data (q e,exp ), whereas the gap between calculated values and experimental data for pseudo-first-order model was more obvious. This indicated that the adsorption perfectly complied with pseudo-second order reaction and the adsorption of phenol appeared to be controlled by the chemisorption process simultaneously. In other words, the adsorption of phenol happened via surface exchange reactions until the surface functional sites were fully occupied; thereafter, phenol molecules diffused into the AC network for further interactions (such as inclusion complex, hydrogen bonding, hydrogen-phobic interactions) [32]. This result was consistent with the kinetics behavior of phenol adsorption on other AC samples [9,25]. On the other hand, the Weber-Morris plots were shown in Figure 5. The plots were not linear over the whole time range, implying that more than one process affected the adsorption, and the adsorption process contained both the surface adsorption and intraparticle diffusion. These behaviors can be explained due to the presence of micropores in the adsorbent.

Isotherm of Phenol Adsorption on AC
We investigated the impact of initial phenol concentration (C0) by fixing adsorbent dose (50 mg) and contact time (≥420 min) under neutral pH and at ambient temperature (25 °C). As shown by the results in Figure 6, the amount of phenol adsorbed per unit weight of adsorbent increased with increasing C0. The C0 provided the necessary driving force to overcome the resistances to the mass transfer of phenol between aqueous and solid phases, and the increase in C0 also enhanced the

Isotherm of Phenol Adsorption on AC
We investigated the impact of initial phenol concentration (C 0 ) by fixing adsorbent dose (50 mg) and contact time (≥420 min) under neutral pH and at ambient temperature (25 • C). As shown by the results in Figure 6, the amount of phenol adsorbed per unit weight of adsorbent increased with increasing C 0 . The C 0 provided the necessary driving force to overcome the resistances to the mass transfer of phenol between aqueous and solid phases, and the increase in C 0 also enhanced the interaction between phenol and AC [31]. Therefore, an increase in C 0 of phenol enhanced the adsorption uptake of phenol. On the other hand, the adsorption efficiency (R%) of AC increased gradually at the beginning with increasing concentration of phenol. It decreased when the concentration increased further. This typical trend has been previously reported, and the reason can be mainly attributed to the saturated occupation of the adsorption sites of AC. In other words, at a lower concentration, the higher adsorption may be due to the presence of more available sites on the adsorbent than the number of phenol ions which are available in the solution. However, at higher concentrations, the number of phenol ions is relatively higher than available sites for adsorption [24]. The maximum adsorption percentage was determined at 30 mg/L as 69% for GAC 950 , 83% for PAC 800 , and 87% for PAC 1000 .
The linear form of Langmuir and Freundlich isotherms of phenol on AC are shown in Figure 7 and the parameters of models are listed in Table 2. By comparing the coefficient of determination (R 2 ), it can be concluded that the Langmuir model was more suitable to describe these sorption processes than the Freundlich model over the whole range of phenol concentrations; this indicated the homogeneous (monolayer) adsorption characteristics. In other words, AC had a uniform interfacial adsorption site for phenol [33]. PAC 1000 had the highest adsorption capacity for phenol, and the maximum adsorption capacities calculated by the Langmuir isotherm were 246.31 mg/g. The next in line with a q max of 232.02 mg/g was PAC 800 , and GAC 950 was the lowest with an adsorption capacity of 214.13 mg/g. This can be explained by the fact that PAC 1000 had the highest specific surface area and pore diffusion. Further, a possible explanation for the fact that the adsorption capacity of GAC 950 was lower than PAC 800 was the higher significance of pore diffusion and pore mouth closure in GAC [8,25].  Table 2. By comparing the coefficient of determination (R 2 ), it can be concluded that the Langmuir model was more suitable to describe these sorption processes than the Freundlich model over the whole range of phenol concentrations; this indicated the homogeneous (monolayer) adsorption characteristics. In other words, AC had a uniform interfacial adsorption site for phenol [33]. PAC1000 had the highest adsorption capacity for phenol, and the maximum adsorption capacities calculated by the Langmuir isotherm were 246.31 mg/g. The next in line with a qmax of 232.02 mg/g was PAC800, and GAC950 was the lowest with an adsorption capacity of 214.13 mg/g. This can be explained by the fact that PAC1000 had the highest specific surface area and pore diffusion. Further, a possible explanation for the fact that the adsorption capacity of GAC950 was lower than PAC800 was the higher significance of pore diffusion and pore mouth closure in GAC [8,25].   The linear form of Langmuir and Freundlich isotherms of phenol on AC are shown in Figure 7 and the parameters of models are listed in Table 2. By comparing the coefficient of determination (R 2 ), it can be concluded that the Langmuir model was more suitable to describe these sorption processes than the Freundlich model over the whole range of phenol concentrations; this indicated the homogeneous (monolayer) adsorption characteristics. In other words, AC had a uniform interfacial adsorption site for phenol [33]. PAC1000 had the highest adsorption capacity for phenol, and the maximum adsorption capacities calculated by the Langmuir isotherm were 246.31 mg/g. The next in line with a qmax of 232.02 mg/g was PAC800, and GAC950 was the lowest with an adsorption capacity of 214.13 mg/g. This can be explained by the fact that PAC1000 had the highest specific surface area and pore diffusion. Further, a possible explanation for the fact that the adsorption capacity of GAC950 was lower than PAC800 was the higher significance of pore diffusion and pore mouth closure in GAC [8,25].

The Effect of pH and Temperature on Phenol Adsorption
The pH of the solution is a critical parameter affecting the adsorption process because it affects the surface charge of the adsorbents as well as the degree of ionization and speciation of pollutants. It is a common observation that the surface adsorbs anions favorably at lower pH due to the presence of H + ions, whereas the surface is active for the adsorption of cations at higher pH due to the deposition of OH − ions [34]. In this work, the effect of pH on the adsorption of phenol was investigated under five individual pH values of 3.0, 5.0, 7.0, 9.0, and 11.0. The initial concentration of phenol was 300 mg/L, the contact time was ≥420 min, and the adsorption temperature was controlled at 25 • C accordingly. As shown in Figure 8, the maximum adsorption was determined at pH 7 as 169.72 mg/g for GAC 950 and 183.72 mg/g for PAC 800 . The maximum for PAC 1000 was 200.96 mg/g at pH 5. The decrease in phenol adsorption as the pH dropped from 7(5) to 3 was mainly due to electrostatic repulsion. In detail, under acidic conditions, the pH PZC (point of zero charge) of AC (GAC 950 7.4, PAC 800 7.6, and PAC 1000 5.7) were higher than the pH of solution, and there were many positive charges on the surface of AC. Further, the phenolic compounds were in the non-ionized forms, and the surface groups were either neutral or positively charged. The adsorption of water and phenol was competitive and the adsorbability of phenol was low. At pH >7(5), the decrease of phenol adsorption may have resulted from two reasons. First, the negative charges on the surface of AC increased with pH and phenol changed from molecular state to ionic state, which made the repulsion force between phenol ions and AC significant. Second, phenolate anions were more soluble in the aqueous solution, and stronger adsorbate-water bonds must be broken before adsorption can take place [7,20]. The pH of the solution is a critical parameter affecting the adsorption process because it affects the surface charge of the adsorbents as well as the degree of ionization and speciation of pollutants. It is a common observation that the surface adsorbs anions favorably at lower pH due to the presence of H + ions, whereas the surface is active for the adsorption of cations at higher pH due to the deposition of OH − ions [34]. In this work, the effect of pH on the adsorption of phenol was investigated under five individual pH values of 3.0, 5.0, 7.0, 9.0, and 11.0. The initial concentration of phenol was 300 mg/L, the contact time was ≥420 min, and the adsorption temperature was controlled at 25 °C accordingly. As shown in Figure 8, the maximum adsorption was determined at pH 7 as 169.72 mg/g for GAC950 and 183.72 mg/g for PAC800. The maximum for PAC1000 was 200.96 mg/g at pH 5. The decrease in phenol adsorption as the pH dropped from 7(5) to 3 was mainly due to electrostatic repulsion. In detail, under acidic conditions, the pHPZC (point of zero charge) of AC (GAC950 7.4, PAC800 7.6, and PAC1000 5.7) were higher than the pH of solution, and there were many positive charges on the surface of AC. Further, the phenolic compounds were in the non-ionized forms, and the surface groups were either neutral or positively charged. The adsorption of water and phenol was competitive and the adsorbability of phenol was low. At pH >7(5), the decrease of phenol adsorption may have resulted from two reasons. First, the negative charges on the surface of AC increased with pH and phenol changed from molecular state to ionic state, which made the repulsion force between phenol ions and AC significant. Second, phenolate anions were more soluble in the aqueous solution, and stronger adsorbate-water bonds must be broken before adsorption can take place [7,20]. Temperature is another important parameter for removing phenol from wastewater, and it is desirable to obtain a high qe under room temperature without additional energy consumption. Figure  9 represents the experimental results of temperature effect on phenol adsorption within the range of 20-60 °C. The adsorption capacity of AC for phenol increased with the increase of temperature when the temperature was lower than 30 °C; however, when the temperature was higher than 30 °C, the adsorption capacity of AC decreased with the increase of temperature. At 30 °C, the qe values of GAC950, PAC800, and PAC1000, were 170.48, 182.77, and 200.49 mg/g, respectively. Because sorption is an exothermic process, it would be expected that an increase in temperature of the adsorbateadsorbent system would result in decreased sorption capacity. However, if the adsorption process is controlled by the diffusion process (intraparticle transport-pore diffusion), the sorption capacity will show an increase with an increase in temperatures [31]. As has been shown earlier, the diffusion of adsorbate into pores of the sorbent was also a rate-controlling step. Therefore, the increase in sorption capacity with an increase in low temperature may be attributed to the diffusion process playing an important role. On the other hand, the active binding sites of AC were damaged and the binding forces between the phenol molecules and AC were weakened at a higher temperature; thus, the trend of desorption of phenol from the interface to the solution was increased [14,34]. Temperature is another important parameter for removing phenol from wastewater, and it is desirable to obtain a high q e under room temperature without additional energy consumption. Figure 9 represents the experimental results of temperature effect on phenol adsorption within the range of 20-60 • C. The adsorption capacity of AC for phenol increased with the increase of temperature when the temperature was lower than 30 • C; however, when the temperature was higher than 30 • C, the adsorption capacity of AC decreased with the increase of temperature. At 30 • C, the q e values of GAC 950 , PAC 800 , and PAC 1000 , were 170.48, 182.77, and 200.49 mg/g, respectively. Because sorption is an exothermic process, it would be expected that an increase in temperature of the adsorbate-adsorbent system would result in decreased sorption capacity. However, if the adsorption process is controlled by the diffusion process (intraparticle transport-pore diffusion), the sorption capacity will show an increase with an increase in temperatures [31]. As has been shown earlier, the diffusion of adsorbate into pores of the sorbent was also a rate-controlling step. Therefore, the increase in sorption capacity with an increase in low temperature may be attributed to the diffusion process playing an important role. On the other hand, the active binding sites of AC were damaged and the binding forces between the phenol molecules and AC were weakened at a higher temperature; thus, the trend of desorption of phenol from the interface to the solution was increased [14,34].

The Effect of Humic Acid (HA) on Phenol Adsorption
The humic acid was combined with phenol to investigate the adsorption between humic acid (HA) and phenol for ACs in our work. As is shown in Figure 10, when the concentration of HA was 10 mg/L, qe decreased from 170.39 to 146.96 mg/g, 180.58 to 159.15 mg/g, and 210.01 to 186.87 mg/g for GAC950, PAC800, and PAC1000, respectively, with decrease rates of 13.75%, 11.87%, and 11.02%, respectively. Then, as the concentrations of HA increased, the qe value increased slowly; however, it was still lower than the qe without HA. The adsorption capacity of phenol on AC was decreased in the presence of HA because of competition by HA with phenol, either site competition or pore/interstice blockage [35]. In this case, the humic acid had an average molecular weight of 1716 ± 478 (by number) or 3303 ± 904 g/mol (by mass), whereas phenol had a molecular weight of 94 g/mol. It was unlikely that the humic acid would be capable of occupying the small pores energetically preferred by the phenol molecules. It was speculated that hydrophilic groups such as carboxyl group, hydroxyl group, and phenolic hydroxyl group in HA molecular structure were easy to unite with the oxygen-containing groups of AC, forming complex on the AC surface and thus blocking the adsorption site of AC [36]. Moreover, HA behaved as polyelectrolytes in solution and changed the surface charge of AC by making AC more negative, which increase adsorption of phenol reasonably. Some researchers have also concluded that the increased adsorption is due to desorption of poreblocking phenol [37].

The Effect of Humic Acid (HA) on Phenol Adsorption
The humic acid was combined with phenol to investigate the adsorption between humic acid (HA) and phenol for ACs in our work. As is shown in Figure 10, when the concentration of HA was 10 mg/L, q e decreased from 170.39 to 146.96 mg/g, 180.58 to 159.15 mg/g, and 210.01 to 186.87 mg/g for GAC 950 , PAC 800 , and PAC 1000 , respectively, with decrease rates of 13.75%, 11.87%, and 11.02%, respectively. Then, as the concentrations of HA increased, the q e value increased slowly; however, it was still lower than the q e without HA. The adsorption capacity of phenol on AC was decreased in the presence of HA because of competition by HA with phenol, either site competition or pore/interstice blockage [35]. In this case, the humic acid had an average molecular weight of 1716 ± 478 (by number) or 3303 ± 904 g/mol (by mass), whereas phenol had a molecular weight of 94 g/mol. It was unlikely that the humic acid would be capable of occupying the small pores energetically preferred by the phenol molecules. It was speculated that hydrophilic groups such as carboxyl group, hydroxyl group, and phenolic hydroxyl group in HA molecular structure were easy to unite with the oxygen-containing groups of AC, forming complex on the AC surface and thus blocking the adsorption site of AC [36]. Moreover, HA behaved as polyelectrolytes in solution and changed the surface charge of AC by making AC more negative, which increase adsorption of phenol reasonably. Some researchers have also concluded that the increased adsorption is due to desorption of poreblocking phenol [37].

The Effect of Humic Acid (HA) on Phenol Adsorption
The humic acid was combined with phenol to investigate the adsorption between humic acid (HA) and phenol for ACs in our work. As is shown in Figure 10, when the concentration of HA was 10 mg/L, qe decreased from 170.39 to 146.96 mg/g, 180.58 to 159.15 mg/g, and 210.01 to 186.87 mg/g for GAC950, PAC800, and PAC1000, respectively, with decrease rates of 13.75%, 11.87%, and 11.02%, respectively. Then, as the concentrations of HA increased, the qe value increased slowly; however, it was still lower than the qe without HA. The adsorption capacity of phenol on AC was decreased in the presence of HA because of competition by HA with phenol, either site competition or pore/interstice blockage [35]. In this case, the humic acid had an average molecular weight of 1716 ± 478 (by number) or 3303 ± 904 g/mol (by mass), whereas phenol had a molecular weight of 94 g/mol. It was unlikely that the humic acid would be capable of occupying the small pores energetically preferred by the phenol molecules. It was speculated that hydrophilic groups such as carboxyl group, hydroxyl group, and phenolic hydroxyl group in HA molecular structure were easy to unite with the oxygen-containing groups of AC, forming complex on the AC surface and thus blocking the adsorption site of AC [36]. Moreover, HA behaved as polyelectrolytes in solution and changed the surface charge of AC by making AC more negative, which increase adsorption of phenol reasonably. Some researchers have also concluded that the increased adsorption is due to desorption of poreblocking phenol [37].

Desorption Study
Desorption studies help elucidate the mechanism of adsorption and recover the precious phenols, water, and adsorbent. Attempts were made to desorb phenol from the spent carbons using various solutions (deionized water/acid/alkali), and the results are shown in Figure 11. The release rate of AC in alkali was higher than that of acid and deionized water. This may be attributed to the formation of salt of phenol, which may have facilitated desorption of phenol from the carbon surfaces and weakened the interaction between phenol and AC [19]. Further, the desorption rates of phenol from PAC 800 and PAC 1000 in deionized water were 13% and 12%, respectively, whereas for GAC 950 it was only 9%, which may have been caused by the small pore size of GAC 950 . Phenol has the high affinity of the compounds to AC, and it's hard to desorption after it penetrates into the smaller pores of AC [38]. Desorption studies help elucidate the mechanism of adsorption and recover the precious phenols, water, and adsorbent. Attempts were made to desorb phenol from the spent carbons using various solutions (deionized water/acid/alkali), and the results are shown in Figure 11. The release rate of AC in alkali was higher than that of acid and deionized water. This may be attributed to the formation of salt of phenol, which may have facilitated desorption of phenol from the carbon surfaces and weakened the interaction between phenol and AC [19]. Further, the desorption rates of phenol from PAC800 and PAC1000 in deionized water were 13% and 12%, respectively, whereas for GAC950 it was only 9%, which may have been caused by the small pore size of GAC950. Phenol has the high affinity of the compounds to AC, and it's hard to desorption after it penetrates into the smaller pores of AC [38].

Conclusions
The results concluded that the feedstock and physical appearance had a great effect on the properties of commercial AC and that they would impact the capacity and mechanism of phenol adsorption on commercial AC. Although the surface area values of GAC950 were higher than PAC800, both adsorption rate and the adsorption capacity of GAC950 were found to be lower than PAC800. A possible explanation for this observation was the higher pore diffusion in PAC800. Compared with PAC800, PAC1000 was derived from coconut shell, possessing higher surface area and pore diffusion, and had the highest adsorption capacity for phenol. Additionally, the effects of several parameters, including the pH, temperature, initial phenol concentration, and contact time, were systematically investigated, and the desirable adsorption conditions, adsorption kinetics model, and equilibrium model for phenol were obtained. More importantly, the three commercial AC are suitable to be used for emergency treatment of phenol accidents because of their low desorption ratio in possible release conditions.

Conclusions
The results concluded that the feedstock and physical appearance had a great effect on the properties of commercial AC and that they would impact the capacity and mechanism of phenol adsorption on commercial AC. Although the surface area values of GAC 950 were higher than PAC 800 , both adsorption rate and the adsorption capacity of GAC 950 were found to be lower than PAC 800 . A possible explanation for this observation was the higher pore diffusion in PAC 800 . Compared with PAC 800 , PAC 1000 was derived from coconut shell, possessing higher surface area and pore diffusion, and had the highest adsorption capacity for phenol. Additionally, the effects of several parameters, including the pH, temperature, initial phenol concentration, and contact time, were systematically investigated, and the desirable adsorption conditions, adsorption kinetics model, and equilibrium model for phenol were obtained. More importantly, the three commercial AC are suitable to be used for emergency treatment of phenol accidents because of their low desorption ratio in possible release conditions.

Conflicts of Interest:
The authors declare no conflict of interest.