Activated carbon (BC) prepared from olive oil solid waste (olive husk) by slow pyrolysis was chemically activated using MgCl2 (BC-MgCl2). The BC and BC-MgCl2 were used as adsorbents for removal of three phenolic compounds, namely, phenol (P), p-methoxyphenol (PMP) and p-nitrophenol (PNP), from aqueous solution. The uptake of these three phenolic compounds by the BC and BC-MgCl2 was better expressed by the Langmuir and Dubinin–Radushkevich (D-R) isotherm models than by the Freundlich isotherm, and the kinetics of the adsorption process followed the pseudo-second order kinetic model. The maximum monolayer adsorption capacity of P, PMP and PNP were increased from 24.938, 45.455 and 61.728 on BC to 43.860, 98.039 and 121.951 mg/g on BC-MgCl2 by factors of 1.76, 2.16 and 1.98, respectively. Therefore, the chemical activation of BC by MgCl2 is indeed of importance for improving its adsorption performances. For both adsorbents, the adsorption phenomenon for different substituted phenols is a strong function of solubility, polarity, molecule structure, and size. At the tested temperatures (25, 35 and 45 °C), the negative values of ΔG° and positive values of ΔH° and ΔS° for the adsorption of P, PMP and PNP on BC and BC-MgCl2 demonstrated that the adsorption was a spontaneous, endothermic and entropy-increasing process.
Activated carbon (BC) was prepared from olive husk biomass by slow pyrolysis at 630°C.
BC was chemically activated with magnesium chloride to produce activated carbon.
Phenol, p-methoxyphenol and p-nitrophenol biosorption was performed in a batch method.
Biosorption was well described by the Langmuir and Dubinin–Radushkevich (D-R) isotherm models.
Biosorption kinetics of phenols was found to follow the pseudo-second order model.
The olive oil industry has witnessed rapid expansion over the past two decades (International Olive Council 2015). Jordan is the eighth largest producer of olive oil in the world. Olive oil farming provides vital economic benefits to the country and small farmers. Nevertheless, fast growth combined with unplanned and poor environmental management practices by olive-press factories has led to serious consequences on the surrounding environment. Therefore, it is important to find a low-cost, low-technology method to use treated wastewater and leachate from the olive mill. The olive cake is a lignocellulosic material that can be processed to generate useful products such as activated carbon (El Hanandeeh 2013). Interest in activated carbon production as a method for soil carbon sequestration as well as a soil amender has been growing (Lehmann et al. 2011). Activated carbon has also been suggested as a potential low-cost alternative to activated carbon for the removal of contaminants from water and wastewater (Tan et al. 2015).
Activated carbon is a low-cost material that can be produced from many organic sources, including agricultural waste (Zhang et al. 2014). It has great capacity to adsorb organic and inorganic sewage contaminants (Mohan et al. 2014; Tan et al. 2015). Activated carbon characteristics, however, are extremely dependent on the circumstances of parent biomass and preparation (Cha et al. 2016).
Many researchers studied the capacity of activated carbon prepared from olive oil solid waste (also known as olive husk and olive cake) to remove organic contaminants form wastewater. Baccar et al. (2012) studied the capacity of activated carbon prepared from olive-waste cakes to remove pharmaceutical compounds – ibuprofen, ketoprofen, naproxen and diclofenac – from water at 25 °C. They concluded that the equilibrium data best fit with the Langmuir isotherm model and the adsorption capacities of the activated carbon for the four drugs was reduced with increasing of pH and temperature. The removal capacity for dodecylbenzenesulfonic acid–sodium salt detergent (DBSNa) and methylene blue dye (MB) of olive cake was investigated (Cimino et al. 2005). The results showed that acidic treatment changes the surface properties of olive cake but does not enhance its adsorption capacity. Compared to commercial activated carbons the olive cake activated carbon generally is equally able to uptake MB and DBSNa from solution. Recently, the role of olive cake activated carbon to remove acid blue 80 dye, methylene blue and basic yellow 28 (BY28) was investigated by Toumi et al. (2018a, 2018b) who reported that the experimental adsorption isotherms show the presence of two steps, depending on these compounds' concentration. They further claimed that the interactions between activated carbon and these dyes were identified by molecular simulations as hydrogen, Van der Waals and electrostatic bonds. Weidemann et al. (2018) used olive-press wastes activated carbon to remove 10 compounds of environmental concern (octhilinone, triclosan, trimethoprim, sulfamethoxasole, ciprofloxacin, diclofenac, paracetamol, diphenhydramine, fluconazole, and bisphenol A) from aqueous solution. They have found that the adsorption capacity tests revealed that removal efficiencies varied substantially among different materials.
The removal of phenols from aqueous solutions using activated carbon was investigated by several researchers (Hall et al. 2014; Sy & Yd 2016; Li et al. 2017). Activated carbon prepared at pyrolytic temperatures, 700 °C, from bamboo chips was used to remove polycyclic aromatic hydrocarbons, nitrobenzenes, phenols, and anilines from aqueous solution. The impacts on adsorption ability of organic molecular dimensions and melting points are attributed respectively to the molecular sieving impact and the holding effectiveness of organic molecules in the pores of the activated carbon (Yang et al. 2016). Han et al. (2013) investigated the efficacy of halogenated phenols using activated carbon prepared from bio-solids, fallen leaves, rice straw, corn stalk and used coffee grounds over a pyrolytic temperature range between 250 and 550 °C. They explained that the removal efficiency was limited because of the high pH and low surface area of activated carbon as well as the deprotonation of phenols in the activated carbon system. Nevertheless, the adsorption capacity of phenol on activated carbon was directly related to the surface area and surface charge. Recently, Vunain et al. (2018) reported the feasibility of remediation of catechol- and resorcinol-contaminated water using low-cost sunflower seed hull activated carbon. They have found that at the same experimental conditions, more catechol was adsorbed than resorcinol, which may be due to the catechol's affinity towards water and the position of the hydroxyl group on the benzene ring.
A limited number of studies on the capacity of activated carbon prepared from olive husk for the removal of phenolic compounds have been reported (Cimino et al. 2005; Michailof et al. 2008; Abdel-Ghani et al. 2016). For example, Cimino et al. (2005) reported the potential of activated carbon prepared from olive husk for the removal of phenol from aqueous solution under various conditions. The results showed that the activation of activated carbon by HCl changes the surface properties of olive cake and enhances its sorption capacity. Olive husk was used by chemical activation with KOH to prepare activated carbon. This activated carbon has been used to adsorb a polyphenol combination. The function of porosity and surface groups in the adsorption forces and the characteristics of adsorbed substances was discussed (Cimino et al. 2005). Abdel-Ghani et al. (2016) evaluated the removal of p-nitrophenol by adsorption onto olive cake based activated carbon having a high Brunauer–Emmett–Teller (BET) surface area. The results of this study proved the efficiency of using olive cake based activated carbon as a novel adsorbent for the removal of nitrophenol from aqueous solution.
Given the results of our previous study (El Hanandeh et al. 2016) on the effect of pyrolysis temperature on the preparation of activated carbon from solid waste from Jordanian olive oil processing (OOSW), the optimum temperature of 630 °C was selected. This sample was activated with MgCl2 in this work to study the effect of chemical activation. Thus, there are two main objectives of the present work. The first was to prepare activated carbon from OOSW by chemical activation with MgCl2 to obtain an abundant agricultural by-product, green and environmentally friendly low-cost activated carbon. The second was to reveal the effect of substituents on the adsorption of phenol (P), p-methoxyphenol (PMP) and p-nitrophenol (PNP) on BC and BC-MgCl2. Batch adsorption experiments were carried out under various operational conditions such as pH of phenols, ionic strength, contact time, adsorption temperature and adsorbent dosage.
Reagent-grade hexahydrate MgCl2, P, PMP and PNP were obtained from Sigma Aldrich (Buchs, Switzerland), with purity higher than 99%, and were used without further purification. Deionized water was used in the present study.
Activated carbon preparation was conducted using a Protherm-PC 402 tube furnace model, Turkey. Sample weighing was performed using the analytical balance Precisa 410AM-FR. Using a EuroVector elemental analyzer, elemental analysis of coal, hydrogen and nitrogen was performed. The spectrum of Fourier transform infrared (FTIR) spectroscopy was registered using the Thermo Nicolet NEXUS 670 FTIR spectrophotometer. Using PANalytical's X'Pert PRO X-ray diffraction (XRD) system at 40 kV and 30 mA with a step of 0.02° over the range of 4–60°, the samples were examined using an X-ray powder diffractometer with Cu K range radiation ( = 1.5418 Å). Using FEI Inspect F50 scanning electron microscopy (SEM), the sample shape and surface morphology were examined. The sampling was performed using the GFL 1083 shaker. Shaking was done using the thermostat-equipped GFL 1083 shaker. Using the UV-Vis Cary Varian spectrophotometer, phenol levels were determined.
Preparation of activated carbon
The activated carbon was prepared in the laboratory from olive husk obtained from a modern olive press in northern Jordan. Pyrolysis of olive husk was carried out at 630 °C pyrolysis temperature under oxygen-limited conditions for the duration of 1 h and labeled as BC. The prepared BC was stored in air-tight containers for further uses.
where is the weight of crucible, is the initial weight of crucible with sample and is the final weight of crucible with a sample. The percentage yield, moisture and ash content of activated carbon were expressed as the average of three experiments by standard deviation.
The pH of BC and BC-MgCl2 was measured by a pH meter (HI9025) and the measurements were carried out in triplicate at 25 °C. The pH at point of zero charge (pHpzc) of the BC and BC-MgCl2 adsorbent was determined by solid addition method (Mall et al. 2006) by transferring 50 mL KNO3 to a series of 100 mL conical flasks. This solution's original pH (pHi) was adapted approximately from 2 to 10 with either 0.1 M HNO3 or 0.1 M NaOH solutions. The solution's pHi was correctly measured and 0.1 g of BC or BC-MgCl2 was added to each flask. The flasks were permitted to balance with intermittent manual shaking for 24 hours. The supernatant liquid's final pH (pHf) values were measured. The difference between pHi and pHf values was plotted against pHi. The intersection of the resulting curve at the point where the pH difference = 0 is the pHpzc.
The physicochemical parameters, volatile matter content (%), acid extractable content and fixed carbon (%), were determined according to the procedures given in our previous publication (El Hanandeh et al. 2016).
Preparation of P, PMP and PNP solution
Stock solution (500 mg /L) was prepared individually for P, PMP and PNP by dissolving the necessary amount in 0.01 M NaCl. The stock solutions were used to prepare solutions with distinct levels (10–100 mg/L); the dilution was achieved by using 0.01 M NaCl solution (to maintain the ionic strength constant for all the different concentrations). The solution's pH was adapted with solutions of 0.1 M HCl and NaOH.
Determination of specific surface area
The sample surface area of the pore structure of BC and BC-MgCl2 was assessed using a Nova 2200e surface area and pore size analyzer (Quantachrome Corp., Boynton Beach, FL, USA) from nitrogen gas adsorption isotherms analysis at 77 K. Each sample was degassed at 105 °C for 8 hours. BET equations determined the particular surface area (SABET) and calculated the complete pore quantity from the near-saturation uptake (relative pressure (pressure of the adsorbate relative to its saturation pressure) = 0.99). The mesopore volume, mesopore surface area (SAmeso), and pore size distribution were calculated by the Barret, Joyner, and Halenda method.
Adsorption equilibrium and kinetics models
Three most popular isothermic adsorption models portraying the adsorption isotherms of P, PMP and PNP on BC and BC-MgCl2 adsorbent, Langmuir, Freundlich and Dubinin–Radushkevich (D-R), were used.
The pseudo-first order and pseudo-second order kinetic models were used to explore the best fit of the experimental information extracted from the adsorption of P, PMP and PNP.
RESULTS AND DISCUSSION
Characterization of BC and BC-MgCl2
The effects of impregnation on the physicochemical properties including activated carbon yield, moisture content, volatile matter content%, acid extractable sample, ash content, pH, pHpzc and elemental analysis are listed in Table 1. The pH value of the BC decreased from 10.22 to 8.69, which is consistent with the depressed ash content (84.95–57.34%) as the result of impregnation. The pHpzc of the activated carbons was found to be lowered from 9.71 to 6.92. The carbon content of BC-MgCl2 substantially increased from 62.45 to 76.29 wt.% after the activation process. Hydrogen and oxygen contents showed the opposite change trend, as expected due to the release of volatiles during carbonization, which results in the elimination of non-carbon species and enrichment of carbon. These results are consistent with the findings of other research studies using MgCl2 or ZnCl2 as a chemical activated agent (Angin et al. 2013; Cardoso & Ataide 2015; Tazibet et al. 2018).
The effects of impregnation with MgCl2 on the surface area and pore size analysis of the activated carbons were shown in Table 2. The results reveal that the SABET and SAmeso of the BC before impregnation were 53 and 14 m2/g, while after impregnation they were raised to 267 and 109 m2/g, respectively, which indicates that the surface area of the BC increased after impregnation. Total pore (Vtotal) and mesopore volumes (Vmeso) and particle diameter (Dv) of BC-MgCl2 also increased with the impregnation. The method of activation appears to be improving pore growth and creating fresh pores. Similar findings have been noted by Ucar et al. (2009) and Angın et al. (2013), where the rise in surface region and impregnated pore volumes indicates that porosity produced by MgCl2 may be due to pore widening and mesopore formation left by magnesium chloride following acid and water washing.
The XRD for BC and BC-MgCl2 are shown in Figure 1. The BC pattern displayed a very noisy, amorphous structure with some peaks appearing at 2 = 40.99, 42.82, 43.61, 47.51 and 57.38o. Extra peaks appeared at 2 = 14.61, 35.10o, and 50.54o, which belong to MgCl2 in BC-MgCl2 where the peak resulted from enhancement to the crystallinity. The surface morphological structure of activated carbons before and after activation was examined using SEM. Figure 2 shows that BC has smooth surface without distinctive pores (Figure 2(a)). Activation of the BC produced an irregular and heterogeneous surface morphology with a well-developed porous structure in various sizes as shown in Figure 2(b). Thus, it seems that the cavities resulted from the evaporation of MgCl2 residues and other impurities such as MgO during carbonization, leaving the space previously occupied by the MgCl2 and MgO. Similar results have also been reported by other researchers (Yang & Lua 2006; Ucar et al. 2009).
It is well known that the surface functional groups would give insight into the adsorption capability of the produced activated carbon. The FTIR spectra obtained for BC and BC–MgCl2 (Figure 3) were determined to be similar, which indicates that the same surface functional groups and structures are present in the BC and BC–MgCl2 prepared at the same temperature. However, significant differences concerning the relative intensity of bands are observed between both spectra. The broad absorption band between 3,200 and 3,550 cm−1 shows the existence of OH groups and the vibration of N-H stretching. This band is stronger in BC than in the BC–MgCl2, which could be attributed to the molecular interaction between MgCl2 and activated carbon BC. The band located at about 2,300 cm−1 was ascribed to triple bonds such as nitriles and carbene groups. The bands located at 1,600 and 1,450 cm−1 correspond to olefinic C,C stretching and C-H in-plane bending vibrations in methyl and methylene groups. The intense band at 1,050 cm−1 belongs to C-O stretching vibrations in alcohols, phenols, and ether or ester groups. The band around 860 cm−1 may be attributed to the presence of carboxylic acid or alkyl halides.
Contact time and adsorbent dosage studies
It is well known that adsorption efficiency is heavily dependent on adsorption time. The effect of contact time in P, PMP and PNP adsorption was therefore observed. Figure 4 shows the effect of contact time on the adsorption of P, PMP and PNP by BC and BC-MgCl2. As illustrated in Figure 4, P, PMP and PNP adsorption increased significantly up to the first 60 min, resulting in a total adsorption efficiency of qe. A general observation was that the removal efficiency of P, PMP and PNP with increasing contact time, which is generally true for good adsorbents, is increasing before equilibrium (Liao et al. 2011). During the preliminary stage of adsorption, rapid adsorption was attributed to the availability of vacant surface sites and, after a certain period of time, the vacant sites are occupied by phenol molecules, which leads to a decrease in adsorption capacity due to a decrease in the adsorption sites available (Garba & Abdul 2016). The BC and BC-MgCl2qe values for P, PMP and PNP removal are calculated as 22.989, 45.240, 58.310, 37.356, 98.253 and 117.126 mg/g, respectively.
It is well known that in a given initial concentration of adsorbent molecules in aqueous solution, adsorbent dose plays a very important role in determining the adsorption capacity. The effect of BC and BC-MgCl2 dosages on the percentage removal (%Removal) of P, PMP and PNP was investigated for a range of adsorbent concentration (0.01–0.35 g) in 50 mL of 50 mg/L adsorbent concentration solution and findings are shown in Figure 5. The percentage of P, PMP and PNP removal increased with a rising adsorbent dose and reached a peak removal of approximately 74, 81, 88, 94, 89 and 98%, respectively, for BC and BC-MgCl2. This may be due to an increase in the adsorbent surface area, increasing the number of adsorption sites available for adsorption. It can be seen that the peak percentage of removal of the BC and BC-MgCl2 was usually highest at the beginning (0.20 g/50 mL), and thereafter no significant amount of these phenols was extracted from the solution, suggesting the saturation of the adsorption sites. These findings align well with what our recent previous research has stated (Mohammed et al. 2018; Hamadneh et al. 2019). Additional adsorption tests were performed with the 0.20 g/50 mL dose of BC and BC-MgCl2 from the experimental results obtained.
Effects of pH and ionic strength
It is well known that solution pH influences both the adsorbent surface charge and the dissociation status of the phenolic species correlated with their constants of dissociation (pKa). Phenols are ionizable and the degree of adsorption of phenolate anions to the geometric surface is primarily determined by the adsorbent's surface load, which in turn is correlated with the pHpzc. Therefore, it is necessary to determine the pHpzc of the adsorbent to understand the adsorption process. It determines how quickly pollutants are adsorbed by an adsorbent. PH > pHpzc is favored for cationic adsorption and pH < pHpzc is recommended for anionic adsorption. The BC (9.71) and BC-MgCl2 (6.92) pHpzc were lower than the BC (10.22) and BC-MgCl2 (8.69) pH suggesting abundance of negative charges on the surface of the activated carbon. The phenolic compounds examined in the present work have different values of pKa, 9.89, 10.17 and 7.15, respectively, for P, PMP and PNP (Table 3).
To examine effect of pH on the adsorption of P, PMP and PNP, the batch equilibrium experiments were carried out at different pH values of 2–11, and results are as shown in Figure 6. The figure revealed that the qe values of P, PMP and PNP were increasing slightly with increasing pH up to pH = pKa for each phenol and then decreased with further increasing pH when pH > pKa. At lower pH, i.e., from pH 2 to pKa values, the hydrophobic effects (Table 3), –* interaction and hydrogen-bonding interaction played an essential role in the adsorption of P, PMP and PNP on the BC and BC-MgCl2 surfaces. This could be because phenols exist as a neutral molecule at pH < pKa, and the deprotonation process starts at around pKa. The decreasing of BC and BC-MgCl2 adsorption of P, PMP and PNP at higher pH could be explained on the assumption that the negative charge covering the activated carbon surfaces increased the electrochemical repulsion forces between phenol anionic form and the activated carbon surfaces. Similar results have been found by our previous work and other researchers (Soto et al. 2011; Parker et al. 2013; Mohammed et al. 2018; Houari et al. 2014).
The influence of ionic strength on the P, PMP and PNP adsorption on the BC and BC-MgCl2 is shown in Figure 7. It can be seen that the increase in ionic strength from 0.01 to 0.20 M NaCl resulted in the slight decrease in the amount of P, PMP and PNP adsorbed onto the BC and BC-MgCl2. The deceased adsorption capacity of BC and BC-MgCl2 with an increase in the NaCl concentration is possibly attributed to electrostatic attraction between the negatively charged activated carbon surface and Na+. These electrostatic attraction forces prevent activated carbon particles and phenols from being very close to each other; as a consequence, the adsorption capacity of the activated carbon for phenols will decrease. This outcome demonstrates the increasing competition between phenol molecules and Na+ ions for available adsorbing sites on the activated carbon surface when the salt concentration increases.
Two kinetic models were used to study the adsorption kinetics of P, PMP and PNP on BC and BC-MgCl2 experimental outcomes: Lagergren pseudo-first order (which was the most accurate and appropriate model for original quick reaction) and pseudo-second pseudo ist also given in figs order (which was presumed to depend on the amount of active locations on the adsorbent surface as a rate control phase. The kinetic information acquired from the nonlinear regression assessment appropriate for BC and BC-MgCl2 together with R2, χ2 and Ferror% are provided in Figure 8 and listed in Table 4. The greater value of R2 (>0.999), which was nearer to unity, and reduced χ2 and Ferror% values for P, PMP, and PNP adsorption on BC and BC-MgCl2 (Table 4) for the pseudo-second order kinetic model show better information fitness, compared with the pseudo-first order model. It can also be seen that the adsorption capacity calculated by the dynamic model of the pseudo-second order () is much closer to the experiment values (). Thus, the above results indicate that the adsorption process follows the pseudo-second order kinetic model better than the pseudo-first order, suggesting that the P, PMP and PNP rates were proportional to the number of active sites on the BC and BC-MgCl2 surfaces. For the adsorption of phenols on graphene oxide, Wang et al. (2014) recorded comparable observations.
Effect of phenolic molecular structure on adsorption isotherms
The adsorption data were analyzed using the two-parameter isotherm models of Langmuir, Freundlich, and D-R isotherm equations (Equations (6)–(9)). The three models were assessed by nonlinear regression analysis in the study. The adsorption equilibrium isotherms are shown in Figure 9, and the constant parameters and linear correlation coefficient (R2), chi-squared (χ2) and error function at 25, 35 and 45 °C are tabulated in Tables 5 and 6. From these results, it can be concluded that Langmuir and D-R models are more suitable for the adsorption process than Freundlich model by comparing the values of R2, χ2 and .
The model of Langmuir adsorption is based on the assumption of homogeneous surface locations for monolayer adsorption. The D–R isotherm is more general than the Langmuir isotherm because it assumes no homogeneous surface or potential for continuous adsorption. The applicability of the isothermic models Langmuir and D-R to the six structures means that both monolayer adsorption and heterogeneous surface conditions exist. The monolayer adsorption capacities for P, PMP and PNP onto BC and BC-MgCl2 at 25 °C were 24.938, 45.455, 61.728, 43.860, 98.039 and 121.951 mg/g, respectively.
The Freundlich constant n is a measure of the deviation from linearity of the adsorption. If a value for n is above unity, adsorption is favorable. In particular, the value of n obtained for the six systems is significantly higher than unity at all the temperatures studied (Tables 5 and 6). The mean adsorption free energy (kJ/mol) can be calculated from D-R isotherm constant (mol2/kJ2) using Equation (9), and provides information on the physical or chemical nature of the adsorption mechanism: within the range of 1–8 kJ/mol indicates that physical adsorption occurs; between 8 and 16 kJ/mol, ion exchange adsorption; and when EDR is between 20 and 40 kJ/mol, chemisorption is indicated (Dubinin et al. 1947). For the six systems, the calculated was 2.168 to 6.521 kJ/mol (Tables 5 and 6), indicating that P, PMP and PNP are likely to be removed primarily through physical adsorption.
To study the effect of molecular structure on adsorption, three phenolic compounds were considered (Table 3): P, which has no substituent on phenol; PMP, which has electron donating group on phenol; and PNP, which has electron withdrawing group on phenol. The BC and BC-MgCl2 adsorption capacity increases in the order of P < PMP < PNP. This trend can be interpreted by two factors: electron density on the phenol ring and the solubility of phenolic compounds in water (Table 3). The order of electron density on the phenol ring, P (−0.508) > PMP (−0.397) > PNP (−0.086) (Parker et al. 2013), demonstrates a close correlation to the above trend. In the case of PNP, by the addition of electron with drawing groups, there is less electron density on the phenol ring leading to better uptake, indicative of adsorption through π–π interactions between the BC and BC-MgCl2 surface and the PNP (Dabrowski et al. 2005). Since the solubility of the phenolics is a measure of the adsorbed solvent forces of the attraction, any change in the solubility of a molecule can impact target forces that adsorb. If we consider the solubility of the phenolics used in the present work, the trend is PNP (16 g/L) < PMP (40 g/L) < P (67 g/L), which is in accordance with the literature (Wang et al. 2014) stating that a hydrophobic substance would be more likely to be adsorbed from aqueous solution.
Comparison of P, PMP and PNP adsorption on BC and BC-MgCl2 with other data from the literature are presented in Table 7. Comparing the qmax values with those reported earlier for adsorption of P, PMP and PNP showed that the qmax values for phenolic compounds studied here are 1–2 orders larger than those of activated carbon studied previously. From these results it may be concluded that activated carbon derived from olive waste and modified with MgCl2 is an effective adsorbent for removal of phenolic compounds from aqueous solution.
At each temperature, the adsorption of P, PMP and PNP on BC and BC-MgCl2 had negative , showing that the adsorption phenomenon occurred favorably and spontaneously with minimal requirements of the adsorption energies. This result was in good agreement with the conclusions drawn from the Freundlich exponent n obtained from the isotherm study described in the above section.
A physisorption process usually yields values between −0 and −40 kJ/mol, while chemisorption yields values between −800 and −400 kJ/mol (Yousef et al. 2011). The values of in the present work ranged from −19.077 to −36.997 kJ/mol. The adsorption of P, PMP and PNP on BC and BC-MgCl2 was therefore a process of physisorption.
When the temperature decreased from 45 to 25 °C, the magnitude of free energy transition () moved to the higher negative value, indicating that the adsorption at low temperature is more spontaneous. Similar results have been published for the free energy of extracting phenolic compounds by several researchers (Khenniche & Benissad 2010; Salam & Burk 2010).
The positive value of indicates that BC and BC-MgCl2 adsorption of P, PMP and PNP is an endothermic reaction consistent with temperature influence. Positive values of affirm the solution's dissociative mechanism and the increasing degree of phenolic freedom. Therefore, the arrangement of phenolics on the solid/solution interfaces becomes more random during the adsorption process.
In this study, adsorption of three phenolic compounds by low-cost activated carbon produced from olive oil solid waste (BC) and by BC chemically activated using MgCl2 (BC-MgCl2) using slow pyrolysis at 630 °C was characterized and investigated for efficient and effective removal of P, PMP and PNP. The study on adsorption mechanism indicated that P, PMP and PNP adsorption was mainly enhanced by introducing electron donating and withdrawing functional groups on the benzene ring. In the adsorption process, P, PMP and PNP adsorption on BC and BC-MgCl2 was a spontaneous, endothermic reaction and entropy-increasing process which was more in line with pseudo-second order kinetic model and Langmuir and D-R isotherm models. The maximum adsorption capacity for P, PMP and PNP onto BC and BC-MgCl2 was in the order of PNP > PMP > P.