Study of The Sorption of The Total Phenolic Compounds From Olive Mill Wastewater By Natural Soils: Conventional and Under Microwave Irradiation Kinetics.

The elimination of total phenolic compounds (TPC) from olive mill wastewater was studied by sorption under the conditions: conventional and under microwave irradiation on previously characterized soils. The sorption process has been studied in batch using inorganic materials in their natural states for sustainable development. The characterizations of the soils have shown variability in potential of hydrogen (4.6-8.9) in total nitrogen between 0.5 and 2.5% and in mineral matter which varies between 5.86 and 15.16%. On the other hand, the mineralogical characterization showed that the three soils are composed of several clay and non-clay minerals. The experimental data were analyzed using reaction models and diffusion models. The pseudo second order kinetic model provides the best correlation. It best represents the kinetics of adsorption by the natural adsorbents N 1 , N 3 and R. The sorption models of LANGMUIR, FREUNDLICH and DUBININ-RADUSHKEVICH were used for the mathematical description of the conventional adsorption equilibrium. The best correlations were obtained with the model of LANGMUIR (r 2 > 0.95) on soils N 1 and N 3 unlike the models of FREUNDLICH and DUBININ-RADUSHKEVICH (r 2 < 0.65). The soil R can be represented by the model of FREUNDLICH (r 2 ≥ 0.96) and the model of LANGMUIR (r 2 > 0.93). The latter is conrmed by the value of the dimensionless coecient R L . Removal rates of TPC were calculated. The value obtained (71 %) showed that the soil N 1 is a good adsorbent. The results are satisfactory and promising. The percentages of silica (SiO 2 ) and alumina (Al 2 O 3 ) are the highest in all soils. These results are in agreement with those of the XRD which revealed the richness of its soils in clay minerals (aluminosilicates) and quartz.


Introduction
Water is essential for life, but it is also very often at the heart of many economic activities. Oleic vegetation water is industrial wastewater commonly called "vegetable water" and comes from water during the olive oil extraction process (water from the pulp and added water). The quantity produced depends on the demand for olive oil and its extraction procedure which generates with e uents at concentration points pollutants (polyphenols) that cannot be absorbed by the purifying power of the media. They are considered to be one of the most harmful plant discharges by the agro-food industries (Cardinali et al. 2010) due to the toxicity of their pollutant load for the entire ecosystem (plants, microorganisms, aquatic and aerial organisms).
Phenolics, or polyphenols, are secondary metabolites ubiquitous in plants from roots to fruit. They represent organic pollutants very frequently encountered in e uents, coming from waste from olive oil production during winter. They refer to a family of mono or polycyclic aromatic compounds that contribute color and sensory properties such as bitterness and astringency (Dai et  Sorption is a very e cient and promising physical separation process used to lower the concentration of toxic organic pollutants in industrial e uents. It is a non-polluting treatment method, easy to apply and has already been proven in the treatment of wastewater (Meçabih et

Adsorbents
The adsorbents (soils) used in this work were collected near the sampling area of our two vegetation waters traditional and modern (Tizi-Ouzou). They are natural and differ from each other in color (red, brown, gray and black). Drying in the open air was preceded before being ground by a porcelain mortar, a sieving at different meshes and storage at room temperature in plastic boxes before characterizations and sorption experiments.

Characterization methods
Physicochemical characterization

Olive mill wastewater (OMWW) characterization
The pH was measured using a pH meter brand WTW multi 340i according to NF T 90-008 (ISO 2001). The electrical conductivity was obtained by means of a conductivity meter of the K 120 CONSORT digital conductivity type according to the French standard NF EN 27888 (1994). The results of the conductivity are expressed in mS/cm at ambient temperature. The TPC, expressed as gallic acid equivalents, were determined by the Folin-Ciocalteu colorimetric method (Box 1983) at the experimental wavelength 765 nm by a UV-visible spectrometer (T60, PG. Instruments), expressed in g/L.

Soils characterization
The soil pH was determined, according to ISO 10390 (2005), in a soil suspension diluted in water at 1:5 (V/V) and using a Microprocessor pH Meter WTW pH537. For the pH at the zero point charge, its measurement (Babić et al. 1999) consists in placing in a series of Erlenmeyer asks 40 mL of the KNO 3 solution (10 −1 M, Fluka), previously adjusted to the desired pH varying from 2 to 12 by addition of NaOH or HCl (10 −1 M), in contact with 0.2 g of soil (φ < 500 µm). The suspensions should be kept under constant stirring at room temperature for 48 h. The nal pH is then determined on the supernatant after separation of the two constituents by centrifugation at 3000 g and at 20°C for 15 min. The pHpzc is the point of intersection between the curve representing the nal pH as a function of the initial pH and the line nal pH = initial pH.
Moisture (H) was determined after drying of a mass (m) at 105°C for 24 hours. The rate, expressed as a percentage, was calculated using the formula: (1) where m 0 , m 1 are the mass of the sample before and after drying at 105°C expressed in grams, respectively. Soil densities were measured using a helium pycnometer (Micromeritics) connected to a vacuum system and to a helium cylinder via a pressure regulator (0.3-0.5 bar). They are calculated by the formula: d s = ρ s /ρ eau = (m s /(m L -m i ))/ρ eau (2) where ρ s is the density of the soil, ρ water is the density of water at a temperature of 4°C, m s is the mass of the soil (Kg), m L is the mass of liquid alone (Kg) and m i is the mass of the immersion liquid (Kg) in the pycnometer.
The functional groups were identi ed by Fourier transform infrared spectrometry (NICOLET 380) by preparing potassium bromide KBr (Merck, MM= 119.01 g. mole −1 ) pellets mixed with 1 % of soil powder.
The identi cation of the crystallized mineral species of the soils was carried out by X-ray diffraction (XRD) and this by comparison with a reference le called JCPDS sheets for "Joint Committee for Powder Diffraction Studies" (JCPDS 2000; Mathon 2008). The chemical composition of the soils was determined by the wavelength dispersive X-ray uorescence technique (WD-XRF). The microstructure of our adsorbents was observed by an FEI Quanta 650 type scanning electron microscope (SEM).
The organic and inorganic maters were determined according to the work of Wang et al. (2012). Organic matter, expressed as a percentage, was calculated using the formula: MOG(%) = (m 1 -m 2 /m 1 ).100 (3) where m 1 , m 2 are the mass of the sample after drying at 105°C and 375°C expressed in grams, respectively. On the other hand, the mineral material was calculated by the formula: MM(%) = (m 2 -m 3 /m 2 ).100 (4) where m 2 , m 3 are the mass of the sample after drying at 105°C and 800°C expressed in grams, respectively. The loss on ignition was determined after baking at 1000°C of our adsorbent for one hour in a mu e furnace. It is expressed as a percentage and evaluated by the following formula.
Pf(%) = (m 1 -m 4 /m 1 ).100 (5) where m 1 , m 4 are the mass of the sample after drying at 105°C and 1000°C expressed in grams, respectively. The cation exchange capacity (CEC) was measured at pH = 9.0 in a sandwich percolation tube between two layers of 10 g of sand (Mathieu, Pieltain 2003).
The speci c surface area, also known as mass area, was calculated using the so-called BET (Brunauer et al. 1938) calculation model (Brunauer, Emmett and Teller) at nitrogen at 77°K by a Micromeritics ASAP 2010, USA. The particle size distribution was carried out by sieving using a Retsch type vibrator at vibration amplitude xed at 40 for 30 minutes.

Sorption experimental test
The sorption tests for TPC were studied in static mode (batch). The overall experimental parameters chosen to carry out these tests are: time, weight and particle size of the adsorbent, the dilution factor (TPC concentration) and the pH of the e uent. The latter was adjusted to the required pH using either The content of the TPC was obtained in each case by UV-visible spectrometer at the experimental wavelength 765 nm, after application of the method of Box (1983) on the medium phase, by extrapolation on the calibration curve.
The adsorbed quantity (q e ) as well as the removal percentage (R %) of the PC were calculated as a function of the concentration of total phenolic compounds in both wastewaters before and after adsorption per mass unit of soil (0.2 g) and wastewater volume (20 mL) according to the equations, respectively: q e = (C 0 -C e ).V/w (6) R(%) = (C 0 -C t ).100/C 0 (7) where C 0 , C e , and C t are respectively the initial, the equilibrium and at time (t) the phenolic compounds concentrations in the solution during the adsorption process (g/L). V is the volume of the solution (L) and W is the weight of the sorbents (g).

Conventional sorption test
The experiment was carried out at room temperature and protected from light. For this purpose, 20 mL of non-delipidated oleic e uents were brought into contact with a known weight of each soil separately in open reactors. The suspensions were shaken by a multi-station system (Gesellschaft für Labortechnik mbH, 3015). For each experiment, a control pot was produced under the same conditions without the solid.

Under microwave irradiation sorption test
These microwave irradiation sorption tests were performed in a modi ed SAMSUNG brand household microwave oven. For this, the natural soils (0.2 g) were placed separately in open reactors (diameter = 4.5 cm; length = 6.8 cm) with 10 mL of olive mill wastewater ([OMWW] 0 = 2.652 g/L). This mixture (adsorbent-adsorbate) was placed inside a microwave oven and was dispersed by stirring using a glass rod installed in a mechanical stirrer of the IKA RW 20n brand at 250 rpm. The power of the microwave oven was set at 180 W. The microwave irradiation was programmed for a time t with steps equal to 5 seconds.
Parametric study of the batch sorption

Effect of contact time
The contact time between the adsorbent (soils) and the adsorbate (TPC) is an important parameter which generally in uences the sorption process. It makes it possible to determine the rate of elimination of TPC and therefore to evaluate the effectiveness of the adsorbent. The sorption conditions applied in the conventional method were: 2 g of soil particle size equal to 400 µm, an initial concentration of TPC in OMWW equal to 2.910 g/L. The time was varied from 5 to 420 minutes and the stirring speed was set at 250 rpm. The experimental protocol which was carried out, in a reactor with a capacity of 50 mL, is as follows: 20 mL of diluted olive mill wastewater ([TPC] = 0.728 g/L) with deionized water are added to 2 g of soil of particle size equal to 400 µm. The solids load in the liquid is 100 g. L −1 . Several vials were thus prepared and the suspensions were shaken at room temperature (20 ± 5°C), at natural e uent and soil pH for variable times ranging from 5 to 420 minutes (contact time per bottle). After stirring, the suspension was thus centrifuged and ready for the determination of the TPC by the protocol of Box (1983).

Effect of the amount on the sorption capacity of TPC
This test has been studied in the range 0.5 -3.5 g with steps of 0.5 g for 24 h to ensure equilibrium.
Several asks were thus shaken, at room temperature (20 ± 5°C), at 250 rpm at natural e uent pH (4.55) and soil. The particle size of the latter was set at 400 µm. The content of TPC in the e uent, before adsorption, was measured and is equal to 0.728 g/L. After stirring, the suspension was centrifuged and ready for application of the Box assay protocol (1983).
3. Effect of the soil particle size of the adsorbent on sorption of TPC The effect of soil particle size was studied using different particle sizes ranging from 40 to 400 µm for the three adsorbents and taking into consideration the optimized parameter obtained previously (the oreviously optimized parameter obtained) such as the mass of the adsorbent (3.5 g). The other operating conditions were set at as follows: at 250 rpm for the stirring speed, natural pH of the e uent (4.55) and of the soil, ambient temperature close to 25°C and nally 0.728 mg/L for the content of TPC in the e uent before the application of sorption phenomenon. After stirring, the suspension was centrifuged and ready for application of the Box protocol (1983).

Effect of pH on sorption capacity of TPC
The in uence of pH on the adsorption of TPC by natural soils was studied using the mass (3.5 g) and grain size (40 µm) optimized in previous experiments. The initial concentration of the olive oil e uent in TPC used is 728 mg/L. The pH values explored range from 3 to 10. The suspensions were stirred at a speed of 250 rpm for a period of time required to reach equilibrium previously determined by the adsorption kinetics study (24 h) at room temperature (20 ± 5°C) and protected from light.

Effect of the content of t TPC on the sorption capacity
This study consists in following the in uence of the initial concentration of the solute on the quantity adsorbed by the adsorbent ranging from the content 5 mg/L up to 189 mg/L and a natural pH of the e uent equal to 4.55. The mass (3.5 g) and the grain size (40 µm) used are those previously optimized. The suspensions were stirred at a speed equal to 250 rpm for a period of time equal to 24 hours at room temperature (20 ± 5°C) and protected from light.

Effect of irradiation microwave
The conditions for this effect were set at: a weight equal to 2 g of soil at 40 µm, a power of 180 W, a TPC concentration 663 mg/L, variable times ranging from 5 up to 50 seconds with steps, of 5 seconds, repeated several times in order to avoid boiling (one contact time per ask) and a stirring speed equal to 250 rpm.

Kinetic studies
The adsorption kinetics makes it possible to estimate the quantity of pollutants adsorbed as a function of time. It provides information on the adsorption mechanism and the transfer mode of solutes from the liquid phase to the solid phase. At equilibrium, the adsorption kinetics of a material can be modeled. For this purpose, the literature reports a number of models that can register as follows: Pseudo-rst order model (Srivastava et al. 2006) Ln(q e -q t ) = Ln q e -k 1  where k 1 (min −1 ) is the rate constant for the pseudo-rst-order model, k 2 (g/mg.min) is the rate constant for the pseudo-second-order model, k int (mg/g min 0,5 ) is the rate constant for the intra-particle-scattering model, k b (min −1 ) is the rate constant for the extern scattering model, q e and q t (mg/g) are the amount of solute absorbed per unit adsorbent at equilibrium and at time t, respectively. is the parameter that re ects the deviation of the adsorption process from the intra-particle scattering mechanism.

Isotherm studies
The equilibrium adsorption isotherms of the solid-liquid system of TPC were determined using the following models:

LANGMUIR model
The linearization if the equation of the LANGMUIR model is given by the relation: (12) where Q e and Q max (mg/L) are the adsorption capacity at equilibrium and maximum of the adsorbent, respectively. C e is the equilibrium concentration of TPC in the liquid phase (mg/L) and K L is the LANGMUIR equilibrium constant characteristic of the adsorbent (L/mg).
The plot (C e /Q e ) versus C e provides a line of slope (1/Q max ) and intercept (1/Q max . K L ). The essential characteristics of the LANGMUIR isotherm can be expressed by a dimensionless constant called separation factor or equilibrium parameter "R L " (Kestioğlu et al. 2005).
The value of this factor indicates the isotherm type: unfavorable if R L > 1; linear if R L = 1; favorable if 0 <

FREUNDLICH model
The linearization if the equation of the FREUNDLICH model is given by the relation: ln Q e = ln K F + (1/n) ln C e (14) This equation is that of a straight line with slope 1/n and intercept ln K F , In general, adsorption is favorable when 1/n is between 2 and 10, moderate for 1/n between 2 and 1, weak for 1/n less than 1 and linear when this constant tends towards 1 (Hamdaoui, Naffrechoux 2007).
where Q e and Q max (mg/L) are the adsorption capacity at equilibrium and maximum of the adsorbent, respectively. C e is the equilibrium concentration of TPC in the liquid phase (mg/L) and K f and 1/n are the FREUNDLICH constant.

DUBININ-RADUSHKEVICH (D-R) model
The linearization if the equation of the DUBININ-RADUSHKEVICH model is given by the relation: Ln q e = Ln q max -β'.R 2 .T 2 (ln(1 + (1/C e ))) 2 (15) where q max is the theoretical saturation capacity (mg/g), β' is a constant related to the adsorption energy (mole 2 /kJ 2 ), R is the perfect gas constant (8.314 10 −3 kJ/mole K), T is the solution temperature (K) and C e is the equilibrium concentration of TPC in the liquid (mg/L).
The constant β' is relayed to the free energy E D by the relation: If : E D < 8 kJ/mole, the adsorption process is physical; E D is between 8 and 16 kJ/mole, the adsorption is an ion exchange process; E D > 16 kJ/mole, the process is dominated by intra-particle diffusion (chemical).

Results And Discussion
Olive vegetation water characterization The main characteristics of the vegetation water used to carry out the adsorption tests are summarized in Table 1. They show that the olive oil discharges are acidic and rich in polluting organic matter. The main chemical properties and composition of the studied soils are given in Table 2. They show that the N 3 soil is acidic (pH = 4.6) which is due to the presence of humic and fulvic substances and to the oxidation of inorganic compounds such as iron sulphide (Kedi 2011;Oertli 2008  The particle size distribution (Fig. 1) shows that the soils are characterized by a high percentage of the greater fraction than 40 µm (> 97%) for all the studied soils. The lutite fraction (< 40 µm) is also present, it respectively represents: 0.42, 2.09 and 2.31 % for R, N 3 and N 1 soils.
The physical and chemical characteristics of soils are different. This is shown by the pH values at the zero point charge (pHpzc) of the three soils N 3 , R and N 1 which are respectively: 4.11, 6.39 and 8.17 (Fig.   2).
For the mineralogical composition of each soil type in the raw state (Fig. 3), we can say that it is complex, varied and speci c. According to the JCPDS (2000) sheets, it corresponds to different minerals constituting the soils (Table 3).
Each soil is made up of several clay minerals and various other no clay minerals. Some components are found in all the soils studied such as quartz and muscovite. Kaolinite is also present in two soil samples, notably N 1 and R.
The metallic element composition of the three sols was determined by the wavelength dispersive x-ray uorescence (WD-XRF) technique. It is expressed in oxide equivalent and presented in mass percentages in Table 4. Table 3 Mineralogical composition of soils according to X-ray diffractograms

Soil
Clay mineral composition/ formulas Various mineral composition/ formulas  These soils are also rich in iron; the high value in sample R may explain its red color. The high content of calcium oxide in soil N 1 also con rms the results of the XRD which revealed the presence of calcite only in this sample.
The metals Mg, Mn, Fe, Al and Si, generally constitute the tetrahedral and octahedral layers of clays. Sodium, potassium and calcium are also known as exchangeable cations incorporated into the interfoliar space of clay sheets.
The results of this analysis are in conformity with those of the XRD by con rming the composition, in clays and quartz, of the raw soils studied.
The IRTF spectra of the adsorbents N 1 , N 3 and R, shown in Fig. 4 reveal the vibration bands of the bonds constituting the minerals that make the studied soils.  Table 5. The sign * indicates that the band is not observed in the IR spectrum of the corresponding soil.
These results show that the vibration bands observed correspond well to those of aluminosilicate minerals links which generally make up clays. These links are: Si-O, O-H, Al-O-H which are found, in particular in montmorillonite, muscovite, kaolinite, vermiculite, clinochlor and quartz. This further con rms that these minerals make up the studied soils and thus joins the results revealed by XRD and WD-XRF.
Observations with a scanning electron microscope show that the adsorbents N 1 and N 3 appear to have a more porous surface than the third adsorbent R which appears with a rough image (Fig. 5). The grain size ranges between 0.89 and 5.559 µm.
The EDX analysis perfectly con rms the results obtained by the mineralogical and chemical analyzes. It demonstrates the intensity of the calcium peak, characteristic of carbonates which appears very high only in soil N 1 (Fig. 6a). The high calcium content in soil N 1 is also con rmed by the results of other analyzes, including: XRD and X-ray uorescence.
The EDAX spectra also reveal the presence of silicates, characteristic of quartz and clays that make up the three soils.
It is important to note, according to the EDAX spectra, that the chemical composition of the soils is predominated by the elements: Si, Al and Fe with in addition the element S for the soil N 3 and Ca, Mg for the soil N 1 which contains calcite and dolomite based on XRD analyzes.
The EDAX spectra also reveal the presence of other low intensity peaks, re ecting the presence of other elements with lower contents, namely: phosphorus, sulfur, titanium in soil N 1 ; calcium in soil N 3 ; potassium, calcium and titanium in soil R.

Effects of contact time
These curves show two phases and show that the time to reach almost adsorption equilibrium is about 2 h for the N 1 soil and 4 h for the N 3 and R soils (Fig. 7). The rst phase, during which adsorption of almost of the TPC occurs, is characterized by a relatively short duration (1 h) and a rapid adsorption rate. In the second phase, which ends in equilibrium, the adsorption rate and the adsorbed amount are low. The adsorption is faster in the case of soil N 1 and the adsorption capacity of the latter, at equilibrium, is the highest (70.1 %); it is double that of soil R (45.5 %). This can be explained by the low porosity of the sol R (see SEM image).
By comparing our results with those of the literature, we deduce that the adsorption process of PC on our adsorbents is also slow and agrees with the results (

Effects of amount on the adsorption capacity of TPC
The results show that when removing TPC by adsorption on natural soils, the rate of removal of TPC increases with the weight of adsorbent used (Fig. 8). It reaches 69, 70 and 73 %, respectively on soils N 3 , N 1 and R. According to the work of Garg et al. (2004), the increase in adsorption percentage with soil mass may be due to the increase in adsorbent area and adsorption sites (Aarfane et al. 2014;Uddin et al. 2007). The best removal rate is obtained for weights ranging from 3 to 4 g for all soils.

Effects of the soil particle size of the adsorbent on adsorption of TPC
The results show that the increase in particle size make (Fig. 9) decreases the adsorption capacity of TPC due to the decrease in the speci c surface area, so the adsorption capacity. The best removal rate is obtained for the ne fractions (φ = 40 µm).
Effects of e uent pH on adsorption capacity of TPC The effect of pH on the adsorption capacity of TPC from the e uent by natural adsorbents (N 1 , N 3 and R) shows that the adsorption of TPC is largely affected by the pH of the solution (Fig. 10).

Effects of the initial content of TPC on the adsorption capacity
The adsorption capacities of TPC increase with the initial concentration of the latter until the saturation of the adsorption sites, indicated by a plateau (Fig. 11). The maximum adsorbed quantity Q max is approximately 0.1 mg of PC/g of adsorbent, or exactly: 0.098; 0.103 and 0.108 mg/g for N 3 , N 1 and R, respectively. It is reached from the initial TPC concentration of 45.5 mg/L.

Effect of irradiation microwave
1. Temperature at the end reaction of microwave irradiation The temperatures at the end of the reaction of the adsorbents studied are variable (Fig. 12). It varies between 36.81°C (soil N 3 ) and 33.10°C (soil N 1 ).

Effect of the contact time
Analysis of the curves (Fig. 13) shows that the rate of adsorption of TPC by the three adsorbents (N 1 , N 3 and R) changes rapidly until it reaches saturation. We can say that microwave irradiations generate the very rapid rotation of molecules. As a result, it creates instantaneous heating of the medium and additional shock movement between molecules, which increases the probability of interaction and the speed of reactions.
The retention by the adsorbent R is faster than in the case of other adsorbents N 1 then N 3 . This difference in adsorption retention comes from the speci c surface area of soil R which is relatively greater (37.8 m 2 /g) than those of other soils N 1 (33.18 m 2 /g) and N 3 (24.23 m 2 /g).
Indeed, according to the literature, microwave irradiations accelerate reaction speeds and increase adsorption capacity [Foo et al. 2009] while preserving the pore structure and active adsorption sites [Foo et al. 2012].
By comparing our results with similar work in the literature (Table 6, Fig. 7, Fig. 13: effect of contact time on the adsorption of TPC), we notice that our soil samples are effective in removing the total phenolic compounds in aqueous solutions under the effect of microwave irradiation. Finally, the use of microwave activation reduced the time required to reach equilibrium. The plots of the sorptions kinetics of the TPC (Fig. 14 and 15) show that the pseudo-second order model is suitable for describing the adsorption reactions of the TPC by the three sols (N 1 , N 3 and R) for the two processes because the determination coe cients are much closer to unity, so this model describes the empirical results well. The kinetic parameters (k 2 , q e and r 2 ) deduced from the applied kinetic models are summarized in Table 7.
The coe cients of determination r 2 obtained for the pseudo-second order model are equal in decreasing order to 1.000; 0.975 and 0.971 on soils named respectively N 1 , R and N 3 for the conventional adsorption and 0.99 for adsorption under microwave irradiation. According to Ho and McKay (2000), the adsorption is of the chemisorption type with the involvement of valence forces by sharing or exchange of electrons between the adsorbent and the adsorbate. The sorption isotherms of the TPC on the three natural soils show a classic type L appearance, subgroup 1 for soils N 1 , N 3 and subgroup 2 for soil R. The plots obtained are illustrated in Fig. 16.
The maximum amount adsorbed at equilibrium Q max is approximately 0.1 mg of TPC/ g of adsorbent, or exactly: 0.098, 0.103 and 0.108 mg/g for N 3 , N 1 and R, respectively. These amounts correspond to equilibriums concentrations of TPC of 22.5 mg/L for N 1 and N 3 soils and 20 mg/L for R soil. These results show that the equilibriums concentrations of TPC in the aqueous phase and in the three adsorbents are close.
The models of these adsorption isotherms, illustrated in their linear forms, are illustrated by the plots in Fig. 17, 18 and 19. The parameters obtained from these models (K F , 1/n, Q max , K L , β, E D and r 2 ) for the conventional adsorption of TPC on the three natural soils are determined graphically and listed in Table  9.
According to the values of the determination coe cient r 2 , it is found that the Langmuir model gives a good representation of the adsorption of TPC 0.957 (N 1 ); 0.972 (N 3 ) and 0.936 (R). The dimensionless separation factor R L con rms that this isotherm is favorable because its measured value is between zero and one. Namely that the Langmuir model is established on the following assumptions: the equivalence of all the adsorption sites, the non-dependence of the adsorption energy with the coverage rate of the surface, the absence of interactions between the adsorbed and adjacent species on the surface, the reversibility of the adsorption, which is therefore essentially physical, and the uniformity of the surface of the solid. However, the Freundlich isotherm provided a better match for the adsorption of TPC to soil R because the value of the determination coe cient r 2 is closest to unity (0.957). The classi cation of these isotherms (L) con rms the slowness of the process because the attractions forces between the adsorbed molecules are weak.

Conclusion
The objective of this study was to nd the optimal conditions for the elimination of TPC from olive mill wastewater. Our experimental tests were aimed at testing the adsorption power of natural soils under conventional and microwave conditions. They were chosen in order to reduce costs and simplify the process of treating of olive waste water.
The results showed that the retention of the TPC is faster under the conditions under microwave irradiation than under the conventional conditions. Equilibrium is reached after 5-10 seconds unlike conventional sorption where the shortest equilibrium time is 2 hours.
The results of the kinetic study for all the pollutants show that the retention is very rapid under the conditions under microwave irradiation, it is almost instantaneous. The linear regressions have shown that the kinetics are controlled by the pseudo-second order model. This is clearly con rmed by the values of determination coe cients corresponding to each model.
The study of the sorption isotherm has shown that the experimental data are well reproduced by the Langmuir model for the adsorbents N 1 , N 3 (r 2 > 0.95) and by the Freundlich model for the adsorbent R (r 2 ≥ 0.96). The Langmuir model can also be described on the soil R because r 2 = 0.94. The dimensionless separation factor (R L ) con rms the application of this last model since it is between zero and one, therefore favorable.
In view of these results, it is concluded that the conventional sorption process is slow and better compared to the sorption under the effect of microwave irradiation which is more than fast. The TPC removal rate is acceptable under conventional conditions because it reaches 71 % on soil N 1 . The removal of TPC by conventional sorption is better than sorption under microwave irradiation.
Declarations Figure 2 Curves: nal pH = f (initial pH), the line : nal pH = initial pH and the points pHpzc of the three soils Kinetic linearization of the conventional adsorption of TPC on natural soils a pseudo-rst order, b pseudosecond order, c intra-particle diffusion and d external diffusion Figure 15 Kinetic linearization of adsorption under microwave activation of TPC on natural soils a pseudo-rst order, b pseudo-second order, c intra-particle diffusion and d external diffusion Linearization of the Freundlich isotherm for conventional sorption of TPC on N1, N3 and R soils at room temperature Figure 18 Linearization of the Langmuir isotherm for conventional sorption of TPC on N1, N3 and R soils at room temperature Figure 19 Linearization of the Dubinin-Radushkevich isotherm for conventional sorption of TPC on N1, N3 and R soils at room temperature