Abstract

The study of the influence of operating parameters of dynamic adsorption on fixed-bed column of manganese oxoanions (MnO₄^-) in aqueous media on granular activated carbon (GAC), prepared from the shells palm nuts of Gabon, was carried out. The operating parameters studied were the particle size, the concentration of the initial solution (C₀) of MnO₄^-, the flow rate (D) and the pH of the media. The results obtained on the study of the influence of operating parameters show that the best adsorption capacities at saturation (Q_sat) of MnO₄^- ion on the CAG were obtained with particle size between 0.04 < x < 0.1 (5.90 mg.g^-1); with flow rate of 3 mL.min^-1 (8.36 mg.g^-1) and when the pH of the initial solution was equal to 3.5 (27.01 mg.g^-1). Also, these results showed that the bed of prepared GAC appeared more effective when C₀ was low (10 mg.L^-1). The kinetic models of the different studies carried out show that the pseudo-first-order kinetic model best describes the adsorption of MnO₄^- ions on the GAC. The results of the intraparticle diffusion model indicate that the adsorption of MnO₄^- follows a multi-step process and that the intraparticle diffusion is not the limiting step. In addition, the surface adsorption plays a predominant role in the adsorption mechanism of MnO₄^- ions on activated carbon studied in fixed-bed column dynamics.

1. Introduction

Water is a natural resource essential to the life of all living beings. Unfortunately, this vital resource is often exposed to natural and/or anthropogenic pollution. Water pollution is defined as the change in the physical, chemical and biological parameters of water quality ¹. The development of mining sectors in Africa resulted in continuous pollution of this resource due to the presence of heavy metals (Cr, Mn, Co, Fe ...) in the groundwater and rivers, used in the production of drinking water. In Gabon, for example, an intensification of manganese (Mn) production in the cities of Moanda and Franceville in recent years has been observed ². In 2019, Gabon achieved an annual production of 6.1 million tons compared to 5.2 million tons in 2018, resulting in 17% increase in manganese production ^{3, 4}. This strong mining activity resulted in increasing manganese concentrations in the soils and surrounding aqueous media ². This represents a major risk for the wellbeing of the population and the environment.

Manganese (Mn) is among inorganic pollutants sensitive to redox reactions and can form toxic MnO₄^- like oxoanions in aqueous media ⁵. These species can also be formed in the disinfection process of drinking water and the removal of organic substances from wastewater by oxidation processes ⁶. Although these oxidation processes are considered as promising technologies for removing organic micropollutants, they could also lead in the oxidation of inorganic ions such as chromium (Cr) and manganese, which form unwanted oxoanions ⁶. For example, manganese oxoanions, used as oxidant for liquid purification, can irritate the skin and eyes, and destroy the liver and kidneys if splashed or ingested orally ⁷. Considering the harmful properties of these metal oxoanions, their elimination from industrial wastewater becomes a major threat for environmental protection and public health ⁸. Thus, the World Health Organization (WHO) recommends a limit level of 0.5 mg.L^-1 of MnO₄^- ions in wastewater ⁹.

However, physicochemical techniques exist in the industrial sector to remove manganese oxoanions or metallic ions from industrial effluents ^{10, 11}. These include chemical precipitation ^{11, 13, 14}, chemical coagulation and flocculation ^{13, 14, 15}, ion exchange ^{13, 16, 17}, membrane filtration ¹⁸, electrochemical treatment ¹¹ and adsorption processes ^{14, 18, 19, 20}. Among all these techniques, adsorption processes using porous solids as adsorbents offer a simple design and operation, low and relatively inexpensive initial cost, wide adsorbents selection, and relatively inexpensive available adsorbents. This assures very accurate and efficient removal of metallic ions in wastewater ^{10, 11}.

Many studies using adsorption techniques in static (batch) for the trapping or removal of MnO₄^- ions using adsorbent materials prepared from corn cobs and animal bones ²¹, leaves’s powders of Azadirachta indica ²², Foeniculum vulgare ²³ and copper sulphide nanoparticles (CuS) ²⁴ were reported in the literature. Activated carbons, prepared from household waste, such as coconut shells ²⁵, are among the most widely used solid adsorbents in this type of study.

This work deals with the study of the operating parameters for optimal elimination in aqueous media of manganese oxoanions (MnO₄^-) on activated carbon, prepared from palm nuts shells collected in the city of Franceville in Gabon, in dynamic adsorption on fixed-bed column. Kinetic models of dynamic adsorption of MnO₄^- on prepared activated carbon were also carried out. The valorization of household waste, such as palm nuts shells, in the preparation of activated carbons for the elimination of mineral pollutants in aqueous media, is a second objective in this work.

2. Experimental

The MnO₄^- solution used in this study was obtained from the extrapure crystallized potassium permanganate (KMnO₄) from Scharlab SL. 1000 mg.L^-1 of MnO₄^- solution were prepared by dissolving in distilled water 1 g of crystallized KMnO₄ in a flask of 1 L. The solution was stirred during 15 minutes. Then by dissolution, solutions with concentrations 10; 25 and 50 mg.L^-1 were prepared.

For the preparation of activated carbon (AC), we used palm nut shells collected in the city of Franceville in the Haut Ogooué region of Gabon. The activated carbon was prepared using a chemical activation process ²⁶. The palm nut shells were scraped and washed several times with water, then dried in the open air during 7 days. The shells were then crushed to small particles, washed with distilled water and dried in an oven for 24 hours (h) at 383 K. These solids were then placed in a borosilicate glass crucible for preservation.

The dried palm nut shells were heat treated at 573 K for 2 h in a NABERTHEN oven before impregnation. After cooling, 50 g of the pretreated solid were impregnated with a solution of ZnCl₂ to 1 M (1:1 ratio) during 72 h. The impregnated solid was then dried in an oven for 24 h at 383 K. After drying, the impregnates were activated at 873 K for 1 h and 30 minutes (min) with a rise in temperature of 5 K.min^-1. The activated carbon obtained was cooled, rinsed with 0.1 M HCl, and then washed with distilled water until the pH of the residual water being equal to 6.5.

Finally, the activated carbon was dried in the oven for 48 h at 383 K and sieved to have granular activated carbons (GAC) with particle sizes 0.04 <x<0.1 mm; 0.1<x<0.25 mm and 0.25<x<0.5 mm (sieves used: TAMISAR (Norm: AFNOR_NF-X11-501).

Measurements of surface area and pore volume were carried out using a Micromeritics TRISTAR 3000 instrument following T. Belin method ²⁷. About 100 mg of the activated carbons sample was pretreated during 1 h at 363 K, and 10 h at 623 K successively. The adsorption/desorption isotherm of nitrogen was carried out at 77 K. The specific surface areas of samples were evaluated by means of the Brunauer, Emmett and Teller (BET) theory ²⁸. Microporous and mesoporous volumes were determined by t-plot method of De Boer and Dubinin-Radushkevich equations respectively ^{29, 30, 31}.

Quantification of surface function groups of adsorbent was performed using the Boehm method ³². 0.2 g of solid were placed in 4 erlenmeyers of 250 mL each containing 25 mL of solutions of NaHCO₃, Na₂CO₃, NaOH and HCl at 0.1 N respectively. Then, the mixtures were stirred at 200 rpm during 24 h at room temperature and filtered on Büchner. Then, 10 mL of each filtrate was titrated using 0.1 N HCl (for basidic solutions) and 0.1 N NaOH (for acidic solution) in the presence of phenolphthalein and methyl orange as indicators.

The number of equivalents (m_eq·g⁻¹) or concentrations (mmol.g^-1) of acid or basic functions were determined using the following equation ^{33, 34}:

Where, C is the concentration of NaOH or HCl (mol.L^-1); V_eq.B and V_eq,E are the equivalent volumes of the blank and sample (L) respectively; m_CA is the mass of activated carbon (g) and 1000 is the conversion factor in mmol.

The pH at the point of zero charge (pH_PZC) of activated carbon was obtained based on the method described by Amola L. A. et al. ²⁶. 20 mL of 0.1 M solution of KCl was introduced into 5 erlenmeyers of 50 mL. The pH of the solutions in each erlenmeyer was adjusted between 2 and 10 by adding 0.1 M HCl or NaOH solution. The pH of these solutions were identified as the initial pH (pH_i). Then, 0.04 g of activated carbon was contacted with solutions at different pH (2; 4; 6; 8 and 10) and stirred (200 rpm) during an average time of 24 h at room temperature. Then, final pH (pH_f) of the solutions were measured. The pH_PZC is given by the point of intersection between the experimental curve and the theoretical curve of pHf = f(pHi).

Dynamic adsorption experiments on fixed-bed column of MnO₄^- were carried out at room temperature (298 K) in a pyrex brand glass column with a length of 50 cm and diameter 2.4 cm. The adsorption column consisted of a bed of granular activated carbon (GAC) of height H, sandwiched between two layers of cotton (Figure 1).

MnO₄^- solutions with concentration C₀ were introduced into the column, and allowed to flow by gravity through the bed of GAC with defined flow rates (D). At the outlet of the column, residual solutions of MnO₄^- with concentrations C were collected at regular time intervals (2-10 min). The residual concentrations of MnO₄^- at column output C were determined using a UV-visible molecular absorption spectrophotometer, using a calibration curve previously established at 546 nm wavelength.

Experiments on the effect of flow were carried out with flow rates of 3; 6 and 9 mL.min^-1. For the effect of particle size, experiments were conducted with activated carbon grains with particle sizes 0.04<x<0.1 mm; 0.1<x<0.25 mm and 0.25 < x < 0.5 mm. Experiments on the effect of the pH of media and on the initial concentration C₀ were carried out with pH solutions of 3.5; 4.5 and 6.5 and C₀ of 10; 25 and 50 mg.L^-1 respectively.

Breakthrough curves of the dynamic adsorption of MnO₄^- ions on prepared activated carbon were plotted. These Breakthrough curves show the evolution of the ratio of concentrations at the outlet (C) and at the inlet (C₀) of the column in function of time. They give information on the efficiency of the bed to trap MnO₄^- ions, and on the adsorption mechanism of MnO₄^- on GAC.

The experiments were stopped when the beds were saturated, i.e. when the residual concentrations (C) were equal to the concentrations at the inlet of the column (C₀) (C/C₀ = 1). We then obtain an area (A) in minutes by integration according to the trapeze method ³⁵:

Where C₀ and C are the inlet and outlet concentrations (mg.L^-1) at times t_n and t_n+1 respectively.

The expression of the contact time (t_c) of the flow rate of MnO₄^- solution and the activated carbon bed is:

Where H is the height of the bed (m), d is the diameter of the column (m) and D is the flow rate (L.s^-1).

Adsorbed amounts of oxoanions (q_ads) and saturation adsorption capacities (Q_sat) are given by the following relations:

D, C₀, A and m_CA correspond to the flow rate (mL.min^-1), the concentration at the inlet of the column (mg.L^-1), the area surface corresponding to the adsorbed amount of MnO₄^- (min) and the activated carbon mass used (g) respectively.

Removal percentage (%E) of manganese oxoanions (MnO₄^-) on activated carbon is given by the relation:

Where t_sat is the saturation time (min).

Kinetic study of adsorption processes provides information on the adsorption mechanism and the mode of transfer of solutes from the liquid phase to the solid phase. In order to evaluate these parameters, the pseudo-first-order kinetic models developed by Lagergren ³⁴, the pseudo-second order developed by Blanchard and linearized by Ho ^{36, 37} and the Weber and Morris model ³⁸ were applied to experimental data of MnO₄^- breakthrough curves.

The rate law of a pseudo-first-order reaction is expressed by the following relation ³³:

After integration between t = 0 and t, and at Q_t = 0 and Q_t. We then obtain the following linear form:

Q_t and Q_e correspond to the adsorption capacities at time t and equilibrium (mg.g^-1) respectively, and k₁ the adsorption rate constant of the pseudo-first order kinetic model (min^-1).

The pseudo-second order model is expressed by the following relation ³⁶:

After integration between t = 0 and t, then when Q_t = 0 and Q_t, we obtain the following linear form ³⁷:

Q_t and Q_e correspond to the adsorption capacities at time t and equilibrium (mg.g^-1) respectively, and k₂ the adsorption rate constant of the pseudo-first order kinetic model (g.mg^-1.min^-1).

According to Webber and Morris ³⁷, the kinetic expression of intraparticle diffusion is showed by:

Where k_d is the intraparticle diffusion rate constant (mg.g^-1.min^1/2) and C_d, intercept of the curve. This gives an indication of the thickness of the boundary layer and/or adhesion to the surface.

3. Results et Discussion

In order to evaluate the textural and chemical characteristics of the prepared activated carbon, textural analysis by nitrogen (N₂) physisorption was performed. Determination of pH at point of zero charge (pH_PZC) and the quantification of surface function groups were also performed.

The N₂ physisorption isotherm obtained on the activated carbon studied is shown in Figure 2. The form of the isotherm obtained describes a type I isotherm according to the IUPAC classification ³⁹. This isotherm is characteristic of microporous materials with a hysteresis loop of type H4, which indicates the presence of pores in the form of a slit. The specific surface area BET (S_BET) and the microporous volume obtained after exploitation of the isotherm of N₂ physisorption are 116.2 m².g^-1 and 0.046 cm³.g^-1 respectively (Table 1).

The results in Table 1 also confirm the absence of mesopores. Indeed, the mesoporous volume (0.002 cm³.g^-1) is insignificant compared to microporous volume (0.046 cm³.g^-1) obtained on the AC studied.

According to Figure 3, the pH at point of zero charge (pH_PZC) of AC studied is equal to 6.7. This result indicates that the surface of activated carbon is positively charged when the pH of the solution (pH_sol) is less than 6.7 and is negatively charged when the pH is above 6.7. Thus, interactions between the surface of activated carbon and manganese oxoanions (MnO₄^-) will be more favorable or important at pH_sol < pH_PZC.

Results of the determination of concentrations of surface functions of the activated carbon studied are summarized in Table 1. These results show that the amounts of acid groups (3.2 mmol.g^-1) are greater than those of basic groups (0.3 mmol.g^-1). They indicate also that phenol groups are predominant among acid oxygen functions (carboxyl) followed by lactone groups on the surface of prepared activated carbon.

The objective of this study was to evaluate the influence of operating parameters on the adsorption capacities of granular activated carbon (GAC) that constitute the bed under dynamic conditions. Thus, studies on the effect of particle size (x), initial solution concentration (C₀) of MnO₄^-, flow rate (D) and media pH were performed. This in order to find the optimal operating conditions for a better removal of manganese oxoanions on the prepared activated carbon.

Since adsorption is a surface phenomenon and the efficiency of its process should be proportional to the available specific surface area of the adsorbent, particle size is therefore an important factor in adsorption processes. Results of the study of the influence of particle size on the adsorption of MnO₄^- ions are grouped in Table 2.

Results show that the more the particle size decreases, the more the adsorbed amounts of MnO₄^- (q_ads) and bed saturation capacities (Q_sat) increase. Indeed, Q_sat of GAC bed of particle sizes 0.04 <x< 0.1 mm (5.9 mg.g^-1) is higher than those of particle sizes 0.1 <x< 0.25 mm (3.1 mg.g^-1) and 0.25 <x< 0.5 mm (1.3 mg.g^-1). This increase in adsorption capacities is related to the fact that beds composed of small particles have larger surface areas and consequently, greater distribution of adsorption sites.

Figure 4 shows the breakthrough curves of MnO₄^- obtained on the GAC with different particle sizes. All breakthrough curves obtained show a multi-step adsorption mode (3 zones) in the form of staircase, with different adsorption kinetics and diffusions according to form of slopes of each adsorption domain.

The first step (0 - 50 min) seems to indicate an adsorption kinetic and a rapid diffusion of the aqueous phase containing MnO₄^- on the surface of the adsorbent and on easily accessible sites. This reveals weak interactions between the structure of the adsorbent and the MnO₄^- ion according to the slopes of the curves of this domain (Figure 4). In this zone (0 - 50 min), the 0.04<x<0.1 mm particle size bed, compared to the other two beds, appears to show slow diffusion and/or stronger adsorbate-adsorbent interactions. Indeed, the slope obtained is more inclined compared to the other two beds of GAC (Figure 4).

The second zone (50 - 150 min) and the third (> 150 min) are zones corresponding respectively to intraparticle and surface adsorption of the oxoanions (MnO₄^-) on hardfull access sites of the prepared GAC. The slopes of these adsorption zones tend toward horizontal, indicating slow adsorption kinetics and diffusion of MnO₄^- at the surface or between the grains of the adsorbent with strong adsorbent-adsorbent interactions greater than those observed in zone 1 (0 - 50 min).

These breakthrough curves show also a very low affinity of MnO₄^- with the CAG under our operating conditions (Figure 4). On the other hand, the saturation times obtained (Table 2) indicate that the prepared GAC has adsorption sites favorable to interactions of MnO₄^- with the GAC structure. The saturation time of the bed composed of particle size 0.04<x<0.1 mm is 270 min, which is higher compared to those of GAC beds of particle size 0.1<x<0.25 mm and 0.25<x<0.5 mm (155 and 100 min respectively). This is mainly due to the fact that smaller particles have higher flow resistance and a shorter scattering path in the pores, allowing MnO₄^- to easily access the surface of the GAC particle ³⁹.

Finally, we realize that the variation in particle size influences the adsorption capacity of MnO₄^- on GAC in our operating conditions. According to the results obtained, the activated carbon bed of low particle size 0.04<x<0.1 mm is the most likely for the removal of MnO₄^- ions. Indeed, this bed shows a more accessible surface and adsorption sites and consequently, a better affinity between MnO₄^- and the adsorbent structure.

The adsorbate concentration is also important in fixed-bed adsorption system because it determines the mass transfer in the solid phase. Results obtained from the study of the influence of the initial concentration of MnO₄^- on GAC are presented in Table 3.

Results show that q_ads and Q_sat decrease when the concentration of the initial solution of MnO₄^- decreases. On the other hand, the percentage of elimination (%E) increases. Indeed, at C₀ = 50 mg.L^-1, the Q_sat value (8.4 mg.g^-1) is higher than those obtained with solutions at 25 mg.L^-1 and 10 mg.L^-1 (6.9 and 6.5 mg.g^-1 respectively). The percentage of removal (% E) of MnO₄^- from a solution at 50 mg.L^-1 (17.2%) is lower than that obtained with solutions at 25 and 10 mg.L^-1 (22.9% and 34.5% respectively).

The decrease in adsorption capacities when the concentration decreases would be due to the effect of the concentration gradient considered to be the driving force of the adsorption process ⁴⁰. When this gradient decreases, the mass transfer is less important. Thus, the amount adsorbed on the surface and the adsorption capacity of the bed decrease. This phenomenon can also be attributed to a major occupation of adsorption sites. The more there are MnO₄^- ions in the media and in contact with GAC particles, the more the surface sites will be occupied and consequently a significant capacity of the adsorbed species bed ⁴⁰. On the other hand, the increase in the percentage of removal (%E) when C₀ decreases would be due to the ratio of MnO₄^- species / GAC particles, which is in favor of GAC because of low amounts of MnO₄^- present in solution ⁴⁰.

However, the breakthrough curves of MnO₄^- obtained (Figure 5) show that the bed is more effective when the concentration of the initial solution is 10 mg.L^-1. According to these curves, breakthrough and saturation times increase when the concentration decreases as seen in Figure 4 and Table 3.

The breakthrough start and saturation times of the bed at C₀ = 10 mg.L^-1 (30 and 1180 min respectively) are superior to those obtained with solutions at C₀ = 25 mg.L^-1 (4 and 600 min respectively) and 50 mg.L^-1 (2 and 490 min respectively). It appears that at low concentrations, adsorption sites on the surface of the prepared GAC are slowly occupied by MnO₄^-. Contrarily, high concentrations of MnO₄^- rapidly saturate the bed of the adsorbent.

Slopes of breakthrough curves obtained (Figure 5) reflect this slow sites occupation when C₀ = 10 mg.L^-1 and rapid saturation at C₀ = 50 mg.L^-1. Indeed, the more C₀ increases, the more the diffusion processes are fast and / or the adsorbent interactions /MnO₄^- are weak.

The variation in the initial concentration of the manganese oxoanion (MnO₄^-) solution influences the adsorption capacity of the bed. The GAC bed is more effective for a MnO₄^- solution at C₀ = 10 mg.L^-1 in our operating conditions.

The flow rate affects the efficiency of the fixed-bed adsorption system. Indeed, it determines the sufficient contact time between the adsorbate and the adsorbent. That contact time should be responsible of better elimination of pollutants within the column. Results obtained from the study of the influence of flow rate on adsorption capacities of the prepared GAC are grouped in Table 4.

These results show that the percentage of elimination (%E), q_ads and Q_sat decrease when the flow rate increases. Indeed, the saturation capacity of GAC, for a flow rate of 3 mL.min^-1 (8.4 mg.g^-1) is higher than those obtained with flow rates 6 and 9 mL.min^-1 (3.8 and 5.6 mg.g^-1 respectively). This decrease in adsorption capacity when flow rate increases appears to be related to rapid movement of the mass transfer zone when flow rate increases ⁴⁰.

Figure 6 shows the breakthrough curves of MnO₄^- obtained. As in previous cases, the multi-step adsorption process of MnO₄^- on the prepared GAC is always observable regardless of the flow rate applied. This process is more marked on the breakthrough curve of MnO₄^- obtained with a flow rate of 3 mL.min^-1. For high flow rates, the diffusion of the aqueous phase seems to be very low, thus limiting the interactions between the adsorption sites and MnO₄^-, as shown on the slopes of breakthrough curves obtained with flow rates 6 and 9 mL.min^-1 (Figure 6).

From breakthrough curves obtained, bed saturation occurs earlier with flow rates 9 and 6 mL.min^-1 (140 min and 240 min respectively). This indicates that the mass transfer zone flows through the GAC bed faster when the flow rate increases. This leads to a reduction of breakthrough start and saturation times of bed. The amounts of MnO₄^- ions adsorbed are also reduced (Table 4). In addition, an effect of contact time is also observed. The more important the contact time, better are the amounts adsorbed and also the saturation capacity of the bed and the percentage of elimination (Table 4).

Thus, the variation in flow rate influences the adsorption capacity of the prepared GAC bed. Based on results obtained, 3 mL.min^-1 is the most suitable flow rate for better interactions between GAC and the MnO₄^- ions in our operating conditions.

pH is an important factor in all adsorption processes. This parameter influences both the overall charge of the adsorbent and the structure of the adsorbate, as well as the adsorption mechanism. Results of the study of the effect of pH on MnO₄^- adsorption capacities of the prepared GAC are grouped in Table 5.

Results show that when the pH decreases, the percentage of removal (%E), adsorbed amounts (q_ads) and adsorption capacities at saturation of GAC bed (Q_sat) increase. Indeed, Q_sat at pH 3.5 (27 mg.g^-1) is higher than those obtained at pH 4.5 and 6.5 (20.3 and 8.4 mg.g^-1 respectively). This indicates that the oxoanion MnO₄^- in acidic media, undergoes a strong attraction of surface functions characterized previously, which are affected by the following protonation effect ⁴¹:

This protonation of the adsorbate surface enhances the attraction between MnO₄^- and the prepared GAC surface and leads to significant adsorbed amounts and adsorption capacity ⁴². This hypothesis suggests physical interactions (physisorption) between MnO₄^- and the prepared GAG. Also in acidic media, the central ion Mn^VII of the anion MnO₄^- can be reduced to Mn²⁺ in the presence of electro-donor groups. In other terms, surface functions possessing the oxygen atom, nitrogen and others depend on the reaction:

The resulting Mn²⁺ ions would then react with the acid functions that are ionized in an aqueous media (-COOH = -COO-, C-OH = C-O-, etc.) ^{43, 44}. In this case, the MnO₄^- adsorption on the prepared GAC would be a chemisorption with modification of the structure of the adsorbate (Mn^VII = > Mn²⁺) when the media is highly acidified.

Figure 7 shows the breakthrough curves of MnO₄^- obtained on the prepared GAC. These curves also show, as in previous studies, a multi-steps adsorption profile. The diffusion processes and/or interactions between MnO₄^- and activated carbon surface are comparable, regarding breakthrough curves at pH 3.5 and 4.5, especially after a breakthrough time of more than 100 minutes. The diffusion between the grains or on the adsorbent surface of the aqueous phase containing MnO₄^- seems to be slow with strong adsorbent/adsorbate interactions depending on the slopes obtained in these regions.

The hypothesis of a MnO₄^- chemisorption seems to be verified in an acidic media. In addition, slower MnO₄^- adsorptions at pH 3.5 and 4.5 are observable compared to that at pH 6.5 (Figure 7), marked by longer bed saturation times (Table 5). This could be explained by the increase in electrostatic attraction forces due to positive charged surface functions, thus reflecting an increase in interactions between GAC and MnO₄^- particles.

Based on the results, it appears that the prepared GAC bed has more affinity with MnO₄^- ions when the media is acidic. The saturation capacities of prepared GAC at pH 3.5 are equal to 27 mg.g^-1, greater than 8.4 and 20.2 mg.g^-1 corresponding respectively to saturation capacities of GAC with MnO₄^- solutions of pH 6.5 and 4.5.

The objective of this part is to perform kinetic modeling of MnO₄^- adsorption on the prepared GAC in fixed-bed dynamics in order to identify the kinetics that control the adsorption mechanism of MnO₄^- ions on GAC in our operating conditions. Pseudo-first and pseudo-second order kinetics and intraparticle diffusion models were applied to experimental data from the various studies previously performed.

Figure 8 shows the linearization curves of pseudo-first-order (Figure 8a) and pseudo-second-order kinetic models (Figure 8b) applied to experimental data from the above study concerning the influence of pH on the MnO₄^- adsorption capacities in fixed-bed dynamics. From these curves, it seems that the pseudo-first-order kinetics model shows better linearization of experimental points compared to the pseudo-second-order kinetic model for adsorption times up to 300 minutes.

Table 6 summarizes the studied operating parameters, the experimental values of equilibrium adsorption capacities (Q_{e, exp}.) and the characteristic parameters (Q_e,cal., rate constants (k) and linear regression coefficient (R²)) of pseudo-first and pseudo-second order kinetic models related to the different studies carried out.

Results obtained indicate that the pseudo-first-order kinetic model is the model that best describes the MnO₄^- ions adsorption in fixed-bed dynamics on the prepared GAG. Indeed, contrary to the results of the pseudo-second-order model (Table 6), the calculated equilibrium adsorption capacities (Q_{e cal.}) are very close to those obtained experimentally (Q_{e, exp.}). Moreover, the correlation coefficients (R²) obtained for the pseudo-first order is also very close to the unit for all results (R² = 0.999). These results are different from those obtained by Al-Aoh ⁴⁵ on the kinetic study for potassium permanganate adsorption by Neem leaves powder where the adsorption of MnO₄^- ions in static mode on Neem leaves powder follows a pseudo-second-order kinetic.

In addition, results in Table 6 show that the rate constant k₁ increases when D, C₀ and particle size increase. This indicates that when D, C₀ and particle size increase, we have rapid bed saturation and therefore reduced bed efficiency. On the other hand, k₁ decreases when the pH of the initial solution decreases (Table 6). This means that we have a slower saturation of bed due to the increase of adsorption sites by protonation of surface functions when the solution is acidic and therefore, an increase in the efficiency of bed. These results are in agreement with those obtained during the study of the influence of these parameters on the MnO₄^- adsorption capacities of prepared GAC.

Figure 9 shows the kinetic model of Weber and Morris (intraparticle diffusion model) applied to the experimental data of the study concerning the influence of pH on the adsorption capacities of the prepared GAC in fixed-bed dynamics adsorption. The curves obtained show multilinear regions (2 regions) for the adsorbent studied. This result confirms the hypothesis that fixed-bed adsorption of MnO₄^- ions on the prepared GAC follows a multi-step adsorption process. Al-Aoh ⁴⁵ also obtained similar results in his work.

The values of the constants C_d and intraparticle diffusion rates k_d, as well as the linear regression coefficients R² of the different zones, obtained after applying the intraparticle diffusion model on the experimental data of the different studies, are presented in Table 7. Results show that linear regression lines have good correlation coefficients but do not pass through the origin (C_d ≠ 0), which means that intraparticle diffusion is involved in the mechanism of MnO₄^- adsorption on the prepared GAG, but is not the limiting step and the only process governing the adsorption of manganese oxoanions on GAC in dynamic adsorption on fixed-bed column.

The intercept curve C_d of different linear domains (Table 7) represents the thickness of the boundary layer (C_d). The greater the C_d value, the greater the contribution to surface adhesion in the speed limit step ³⁷. C_d2 values are particularly high. This confirms the fact that intraparticle diffusion is not the limiting step, and rather that surface adsorption plays a predominant role in the adsorption of MnO₄^- ions on the prepared GAC in our operating conditions.

Also, in an acidic medium (pH = 3.5) the surface adhesion is very high, especially in zone 2 (C_d2 = 9.151). This result confirms the protonation of surface functions on the prepared GAC in acidic media thus creating new adsorption sites and/or strong interactions between MnO₄^- / GAC on the surface. The hypothesis in acidic media of a chemisorption of MnO₄^- ions still seems to be valid.

4. Conclusion

Objective of these experiments was to study the influence of operating parameters such as particle size, concentration of the initial solution (C₀) of manganese oxoanion (MnO₄^-), flow rate (D) and pH of the media on the adsorption capacities of granular activated carbon (GAC), prepared from palm nuts shells, in dynamic adsorption on fixed-bed column. Also, the adsorption kinetics of MnO₄^- on prepared GAC were studied.

The prepared activated carbon has been characterized texturally and chemically. Characterization by physisorption of nitrogen (N₂) reveals a mainly microporous structure with a specific surface area (S_BET) of 116.2 m².g^-1 and a microporous volume equal to 0.046 cm^-3.g^-1. The determination of surface functions indicates that the concentration of acid groups (3.2 mmol.g^-1) is greater than that of basic groups (0.3 mmol.g^-1). The pH at the point of zero charge (pH_PZC) of prepared activated carbon was evaluated at 6.7.

Study of the influence of operating parameters on the saturation adsorption capacities (Q_sat) of prepared activated carbon in fixed-bed dynamics shows an increase of adsorption capacities when the particle size decreases, the flow rate (D) decreases, the concentration of MnO₄^- ions (C₀) increases and when the pH decreases. Best Q_sat of manganese oxoanion on GAC were obtained when, following the study, the MnO₄^- adsorption was carried out with a particle size 0.04 < x < 0.1 (5.90 mg.g^-1); a flow rate of 3 mL.min^-1 (8.36 mg.g^-1) and at pH 3.5 (27.01 mg.g^-1). Also, the bed of prepared GAC appears more effective at C₀ = 10 mg.L^-1.

Kinetic models of the different studies carried out show that the pseudo-first-order kinetic model best describes the adsorption of MnO₄^- ions on the GAC. The results of the intraparticle diffusion model indicate that the adsorption of MnO₄^- follows a multi-step process and that the intraparticle diffusion is not the limiting step. In addition, the surface adsorption plays a predominant role in the adsorption mechanism of MnO₄^- ions on activated carbon studied in fixed-bed column dynamics.

References